Large epidemiologic studies have shown disparities in COVID-19 outcomes by race, ethnicity, and socioeconomic status (SES). Racial and ethnic minorities and individuals of lower SES have experienced disproportionately higher rates of intensive care unit (ICU) admission and death. In Washington, DC, Black individuals (47% of the population) accounted for 51% of COVID-19 cases and 75% of deaths. In comparison, White individuals (41% of the population) accounted for 21% of cases and 11% of deaths.¹ Place of residence, such as living in socially vulnerable communities, has also been shown to be associated with higher rates of COVID-19 mortality and lower vaccination rates.^2-4 Social and structural inequities, such as limited access to health care services and mistrust of the health care system, may explain some of the observed disparities.⁵ However, data are limited regarding COVID-19 outcomes for individuals with equal access to care.

The Veterans Health Administration (VHA) is the largest integrated US health care system and operates 123 acute care hospitals. Previous research has demonstrated that disparities in outcomes for other diseases are attenuated or erased among veterans receiving VHA care.^6,7 Based on literature from the pandemic, markers of health care inequity relating to SES (eg, place of residence, median income) are expected to impact the outcomes of patients acutely hospitalized with COVID-19.⁴ We hypothesized that the impact on clinical outcomes of infection would be mitigated for veterans receiving VHA care.

This retrospective cohort study included veterans who presented to Washington Veterans Affairs Medical Center (WVAMC) with the goal of determining whether place of residence as a marker of SES, health care access, and median income were predictive of COVID-19 disease severity.

Methods

The WVAMC serves about 125,000 veterans across the metropolitan area, including parts of Maryland and Virginia. It is a high-complexity hospital with 164 acute care beds, 30 psychosocial residential rehabilitation beds, and an adjacent 120-bed community living center providing long-term, hospice, and palliative care.⁸

The WVAMC developed a dashboard that tracked patients with COVID-19 through on-site testing by admission date, ward, and other key demographics (PowerBi, Corporate Data Warehouse). All patients admitted to WVAMC with a diagnosis of COVID-19 between March 1, 2020, and June 30, 2021, were included in this retrospective review. Using the Computerized Patient Record System (CPRS) and the dashboard, we collected demographic information, baseline clinical diagnoses, laboratory results, and clinical interventions for all patients with documented COVID-19 infection as established by laboratory testing methods available at the time of diagnosis. Veterans treated exclusively outside the WVAMC were excluded. Hospitalization was defined as any acute inpatient admission or transfer recorded within 5 days before and 30 days after the laboratory collection of a positive COVID-19 test. Home testing kits were not widely available during the study period. An ICU stay was defined as any inpatient admission or transfer recorded within 5 days before or 30 days after the laboratory collection of a positive COVID-19 test for which the ward location had the specialty of medical or surgical ICU. Death due to COVID-19 was defined as occurring within 42 days (6 weeks) of a positive COVID-19 test.⁹ This definition assumed that during the peak of the pandemic, COVID-19 was the attributable cause of death, despite the possible contribution of underlying health conditions.

Patients’ admission periods were based on US Centers for Disease Control and Prevention (CDC) national data and classified as early 2020 (January 2020–April 2020), mid-2020 (May 2020–August 2020), late 2020 (September 2020–December 2020), and early 2021 (January 2021–April 2021).¹⁰ We chose to use these time periods as surrogates for the frequent changes in circulating COVID-19 variants, surges in case numbers, therapies and interventions available during the pandemic. The dominant COVID-19 variant during the study period was Alpha (B.1.17). Beta (B.1.351) variants were circulating infrequently, and Delta and Omicron appeared after the study period.¹¹ Treatment strategies evolved rapidly with emerging evidence, including the use of dexamethasone, beginning in June 2020.¹² WVAMC followed the Advisory Committee on Immunization Practices guidance on vaccination rollout beginning in December 2020.¹³

Patients' income was estimated by the median household income of the zip code residence based on US Census Bureau 2021 estimates and was assessed as both a continuous and categorical variable.¹⁴ The Charlson Comorbidity Index (CCI) was included in models as a continuous variable.¹⁵ Variables contributing to the CCI include myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, dementia, hemiplegia or paraplegia, ulcer disease, hepatic disease, diabetes (with or without end-organ damage), chronic obstructive pulmonary disease (COPD), connective tissue disease, leukemia, lymphoma, moderate or severe renal disease, solid tumor (with or without metastases), and HIV/AIDS. The WVAMC Institutional Review Board approved this study (IRB #1573071).

Variables

This study assessed 3 primary outcomes as indicators of disease severity during hospitalization: need for high-flow oxygen (HFO), intubation, and presumed mortality at any time during hospitalization. The following variables were collected as potential social determinants or clinical risk-adjustment predictors of disease severity outcomes: age; sex; race and ethnicity; median income for patient’s zip code residence, state, and county; wards within Washington, DC; comorbidities, CCI; tobacco use; and body mass index.¹⁵ Although medications at baseline, treatments during hospitalization for COVID-19, and laboratory parameters during hospitalization are shown in eAppendices 1 and 2, they are beyond the scope of this analysis.

Statistical Analysis

Three types of logistic regression models were calculated for predicting the disease severity outcomes: (1) simple unadjusted models; (2) models predicting from single variables plus age (age-adjusted); and (3) multivariable models using all nonredundant potential predictors with adequate sample sizes (multivariable). Variables were considered to have inadequate sample sizes if there was nontrivial missing data or small numbers within categories, (eg, AIDS, connective tissue disease). Potential predictors for the multivariable model included age, sex, race, median income by zip code residence, CCI, CDC admission period, obesity, hypertension, chronic kidney disease, obstructive sleep apnea (OSA), diabetes, COPD or asthma, liver disease, antibiotics, and acute kidney injury.

For the multivariable models, the following modifications were made to avoid unreliable parameter estimation and computation problems (quasi-separation): age and CCI were included as continuous rather than categorical variables. Race was recoded as a 2-category variable (Black vs other [White, Hispanic, American Indian, Alaska Native, Asian, Native Hawaiian, and Pacific Islander]), and ethnicity was excluded because of the small number of patients in this group (n = 16). Admission period was included. Predicted probability plots were generated for each outcome with continuous independent predictors (income and CCI), both unadjusted and adjusted for age as a continuous covariate. All analyses were performed using SAS version 9.4.

Heat Maps

Heat maps were generated to visualize the geospatial distribution of COVID-19 cases and median incomes across zip codes in the greater Washington, DC area. Patient case data and median income, aggregated by zip code, were imported using ArcGIS Online. A zip code boundary layer from Esri (United States Zip Code Boundaries) was used to spatially align the case data. Data were joined by matching zip codes or median incomes in the patient dataset to those in the boundary layer. The resulting polygon layer was styled using the Counts and Amounts (Color) symbology in ArcGIS Online, with case counts or median income determining the intensity of the color gradient.

Results

Between March 1, 2020, and June 30, 2021, 348 patients were hospitalized with COVID-19 (Table 1). The mean (SD) age was 68.4 (13.9) years, 313 patients (90.2%) were male, 281 patients (83.4%) were Black, 47 patients (13.6%) were White, and 16 patients (4.8%) were Hispanic. One hundred forty patients (40.2%) resided in Washington, DC, 151 (43.4%) in Maryland, and 19 (5.5%) in Virginia. HFO was received by 86 patients (24.7%), 33 (9.5%) required intubation and mechanical ventilation, and 57 (16.4%) died. All intubations and deaths occurred among patients aged > 50 years, with death occurring in 17.8% of patients aged > 50 years.

Demographic characteristics and baseline comorbidities associated with COVID-19 disease severity can be found in eAppendix 2. In unadjusted analyses, age was significantly associated with the risk of HFO, with a mean (SD) age of 72.5 (11.7) years among those requiring HFO and 67.1 (14.4) years among patients without HFO (odds ratio [OR], 1.03; 95% CI, 1.01-1.05; P = .002). Although age was not associated with the risk of intubation, it was significantly associated with mortality. Patients who died had a mean (SD) age of 76.8 (11.8) years compared with 66.8 (13.7) years among survivors (OR, 1.06; 95% CI, 1.04-1.09; P < .001).

Compared with patients with no comorbidities, CCI categories of mild, moderate, and severe were associated with increased risk of requiring HFO (eAppendix 3). The adjusted OR (aOR) was highest among patients with severe CCI (aOR, 7.00; 95% CI, 2.42-20.32; P = .0007). In age-adjusted analyses, CCI was not associated with intubation or mortality.

Geospatial Analyses

State of residence, county of residence, and geographic area (including Washington, DC wards, and geographic divisions within counties of residence in Maryland and Virginia) were not associated with the clinical outcomes studied (eAppendix 4). However, zip code-based median income, analyzed as a continuous variable, was associated with a reduced likelihood of receiving HFO (aOR, 0.91; 95% CI, 0.84-0.99; P = .03). Income was not significantly associated with intubation or mortality.

The majority of patients hospitalized for COVID-19 at WVAMC resided in zip codes in eastern Washington, DC, inclusive of wards 7 and 8, and Prince George’s County, Maryland (Figure 1). These areas also corresponded to the lowest median household income by zip code (Figure 2).

Multivariable Analysis

Significant predictors of HFO requirement included comorbid diabetes (OR, 2.42; 95% CI, 1.27-4.61; P = .006) and liver disease or cirrhosis (OR, 2.19; 95% CI, 1.09-4.39; P = .02) (Table 2). CDC admission period was also associated with HFO need. Patients admitted after early 2020 had lower odds of receiving HFO. Race and median income based on zip code residence were not associated with HFO requirement.

Comorbid liver disease or cirrhosis was a significant predictor of intubation (OR, 2.81; 95% CI, 1.07-7.40; P = .03). CDC admission period was associated with intubation with lower odds of intubation for patients admitted after early 2020. Race and median income by zip code were not associated with intubation.

Significant predictors of mortality included age (OR, 2.20; 95% CI, 1.55-3.14; P = .0001), comorbid liver disease or cirrhosis (OR, 2.97; 95% CI, 1.31-6.74; P = .008), and OSA (OR, 3.45; 95% CI, 1.49-7.97; P = .003). CDC admission period was associated with mortality, with lower odds of intubation for patients admitted in mid- and late 2020. Race and median income by zip code residence were not associated with intubation.

Discussion

In this study of COVID-19 disease severity at a large integrated health care system that provides equal access to care, race, ethnicity, and geographic location were not associated with the need for HFO, intubation, or presumed mortality. Median income by zip code residence was associated with reduced HFO use in univariable analyses but not in multivariable models.

These findings support existing literature suggesting that race and ethnicity alone do not explain disparities in COVID-19 outcomes. Multiple studies have demonstrated that disparities in health outcomes have been reduced for patients receiving VHA care.^6,16-19 However, even within a health care system with assumed equal access, the finding of an association between income and need for HFO in the univariable analysis may reflect a greater likelihood of delays in care due to structural barriers. Multiple studies suggest low SES may be an independent risk factor for severe COVID-19 disease. Individuals with low SES have higher rates of chronic diseases of obesity, diabetes, heart disease, and lung disease; thus, they are also at greater risk of serious illness with COVID-19.^20-24 Socioeconomic disadvantage may also have limited individuals’ ability to engage in protective behaviors to reduce COVID-19 infection risk, including food stockpiling, social distancing, avoidance of public transportation, and refraining from working in “essential jobs.”²¹

Beyond SES, place of residence also influences health outcomes. Prior literature supports using zip codes to assess area-based SES status and monitor health disparities.²⁵ The Social Vulnerability Index incorporates SES factors for communities and measures social determinates of health at a zip code level exclusive of race and ethnicity.²⁶ Socially vulnerable communities are known to have higher rates of chronic diseases, COVID-19 mortality, and lower vaccination rates.³ Within a defined geographic area, an individual’s outcome for COVID-19 can be influenced by individual resources such as access to care and median income. Disposable income may mitigate COVID-19 risk by facilitating timely care, reducing occupational exposure, improving housing stability, and supporting health-promoting behaviors.²¹

Limitations

Due to the evolving nature of the COVID-19 pandemic, variants, treatments, and interventions varied throughout the study period and are not included in this analysis. In late December 2020, COVID-19 vaccination was approved with a tiered allocation for at-risk patients and direct health care professionals. Three of the 4 study periods analyzed in this study were prior to vaccine rollout and therefore vaccination history was not assessed. However, we tried to capture the evolving changes in COVID-19 variants, treatments and interventions, and skill in treating the disease through use of CDC-defined time frames. Another limitation is that some studies have shown that use of median income by zip code residence can underestimate mortality.²⁷ Also, shared resources and access to other sources of disposable income can impact the immediate attainment of social needs. For example, during the COVID-19 pandemic, health care systems in Washington, DC assisted vulnerable individuals by providing food, housing, and other resources.^28,29 Finally, the modest sample size limits generalizability and power to detect differences for certain variables, including Hispanic ethnicity.

Conclusions

There have been widely described disparities in disease severity and death during the COVID-19 pandemic. In this urban veteran cohort of hospitalized patients, there was no difference in the need for intubation or mortality associated with race. The findings suggest that a lower median income by zip code residence may be associated with greater disease severity at presentation, but do not predict severe outcomes and mortality overall. VHA care, which provides equal access to care, may mitigate the disparities seen in the private sector.

Large epidemiologic studies have shown disparities in COVID-19 outcomes by race, ethnicity, and socioeconomic status (SES). Racial and ethnic minorities and individuals of lower SES have experienced disproportionately higher rates of intensive care unit (ICU) admission and death. In Washington, DC, Black individuals (47% of the population) accounted for 51% of COVID-19 cases and 75% of deaths. In comparison, White individuals (41% of the population) accounted for 21% of cases and 11% of deaths.¹ Place of residence, such as living in socially vulnerable communities, has also been shown to be associated with higher rates of COVID-19 mortality and lower vaccination rates.^2-4 Social and structural inequities, such as limited access to health care services and mistrust of the health care system, may explain some of the observed disparities.⁵ However, data are limited regarding COVID-19 outcomes for individuals with equal access to care.

The Veterans Health Administration (VHA) is the largest integrated US health care system and operates 123 acute care hospitals. Previous research has demonstrated that disparities in outcomes for other diseases are attenuated or erased among veterans receiving VHA care.^6,7 Based on literature from the pandemic, markers of health care inequity relating to SES (eg, place of residence, median income) are expected to impact the outcomes of patients acutely hospitalized with COVID-19.⁴ We hypothesized that the impact on clinical outcomes of infection would be mitigated for veterans receiving VHA care.

This retrospective cohort study included veterans who presented to Washington Veterans Affairs Medical Center (WVAMC) with the goal of determining whether place of residence as a marker of SES, health care access, and median income were predictive of COVID-19 disease severity.

Methods

The WVAMC serves about 125,000 veterans across the metropolitan area, including parts of Maryland and Virginia. It is a high-complexity hospital with 164 acute care beds, 30 psychosocial residential rehabilitation beds, and an adjacent 120-bed community living center providing long-term, hospice, and palliative care.⁸

The WVAMC developed a dashboard that tracked patients with COVID-19 through on-site testing by admission date, ward, and other key demographics (PowerBi, Corporate Data Warehouse). All patients admitted to WVAMC with a diagnosis of COVID-19 between March 1, 2020, and June 30, 2021, were included in this retrospective review. Using the Computerized Patient Record System (CPRS) and the dashboard, we collected demographic information, baseline clinical diagnoses, laboratory results, and clinical interventions for all patients with documented COVID-19 infection as established by laboratory testing methods available at the time of diagnosis. Veterans treated exclusively outside the WVAMC were excluded. Hospitalization was defined as any acute inpatient admission or transfer recorded within 5 days before and 30 days after the laboratory collection of a positive COVID-19 test. Home testing kits were not widely available during the study period. An ICU stay was defined as any inpatient admission or transfer recorded within 5 days before or 30 days after the laboratory collection of a positive COVID-19 test for which the ward location had the specialty of medical or surgical ICU. Death due to COVID-19 was defined as occurring within 42 days (6 weeks) of a positive COVID-19 test.⁹ This definition assumed that during the peak of the pandemic, COVID-19 was the attributable cause of death, despite the possible contribution of underlying health conditions.

Patients’ admission periods were based on US Centers for Disease Control and Prevention (CDC) national data and classified as early 2020 (January 2020–April 2020), mid-2020 (May 2020–August 2020), late 2020 (September 2020–December 2020), and early 2021 (January 2021–April 2021).¹⁰ We chose to use these time periods as surrogates for the frequent changes in circulating COVID-19 variants, surges in case numbers, therapies and interventions available during the pandemic. The dominant COVID-19 variant during the study period was Alpha (B.1.17). Beta (B.1.351) variants were circulating infrequently, and Delta and Omicron appeared after the study period.¹¹ Treatment strategies evolved rapidly with emerging evidence, including the use of dexamethasone, beginning in June 2020.¹² WVAMC followed the Advisory Committee on Immunization Practices guidance on vaccination rollout beginning in December 2020.¹³

Patients' income was estimated by the median household income of the zip code residence based on US Census Bureau 2021 estimates and was assessed as both a continuous and categorical variable.¹⁴ The Charlson Comorbidity Index (CCI) was included in models as a continuous variable.¹⁵ Variables contributing to the CCI include myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, dementia, hemiplegia or paraplegia, ulcer disease, hepatic disease, diabetes (with or without end-organ damage), chronic obstructive pulmonary disease (COPD), connective tissue disease, leukemia, lymphoma, moderate or severe renal disease, solid tumor (with or without metastases), and HIV/AIDS. The WVAMC Institutional Review Board approved this study (IRB #1573071).

Variables

This study assessed 3 primary outcomes as indicators of disease severity during hospitalization: need for high-flow oxygen (HFO), intubation, and presumed mortality at any time during hospitalization. The following variables were collected as potential social determinants or clinical risk-adjustment predictors of disease severity outcomes: age; sex; race and ethnicity; median income for patient’s zip code residence, state, and county; wards within Washington, DC; comorbidities, CCI; tobacco use; and body mass index.¹⁵ Although medications at baseline, treatments during hospitalization for COVID-19, and laboratory parameters during hospitalization are shown in eAppendices 1 and 2, they are beyond the scope of this analysis.

Statistical Analysis

Three types of logistic regression models were calculated for predicting the disease severity outcomes: (1) simple unadjusted models; (2) models predicting from single variables plus age (age-adjusted); and (3) multivariable models using all nonredundant potential predictors with adequate sample sizes (multivariable). Variables were considered to have inadequate sample sizes if there was nontrivial missing data or small numbers within categories, (eg, AIDS, connective tissue disease). Potential predictors for the multivariable model included age, sex, race, median income by zip code residence, CCI, CDC admission period, obesity, hypertension, chronic kidney disease, obstructive sleep apnea (OSA), diabetes, COPD or asthma, liver disease, antibiotics, and acute kidney injury.

For the multivariable models, the following modifications were made to avoid unreliable parameter estimation and computation problems (quasi-separation): age and CCI were included as continuous rather than categorical variables. Race was recoded as a 2-category variable (Black vs other [White, Hispanic, American Indian, Alaska Native, Asian, Native Hawaiian, and Pacific Islander]), and ethnicity was excluded because of the small number of patients in this group (n = 16). Admission period was included. Predicted probability plots were generated for each outcome with continuous independent predictors (income and CCI), both unadjusted and adjusted for age as a continuous covariate. All analyses were performed using SAS version 9.4.

Heat Maps

Heat maps were generated to visualize the geospatial distribution of COVID-19 cases and median incomes across zip codes in the greater Washington, DC area. Patient case data and median income, aggregated by zip code, were imported using ArcGIS Online. A zip code boundary layer from Esri (United States Zip Code Boundaries) was used to spatially align the case data. Data were joined by matching zip codes or median incomes in the patient dataset to those in the boundary layer. The resulting polygon layer was styled using the Counts and Amounts (Color) symbology in ArcGIS Online, with case counts or median income determining the intensity of the color gradient.

Results

Between March 1, 2020, and June 30, 2021, 348 patients were hospitalized with COVID-19 (Table 1). The mean (SD) age was 68.4 (13.9) years, 313 patients (90.2%) were male, 281 patients (83.4%) were Black, 47 patients (13.6%) were White, and 16 patients (4.8%) were Hispanic. One hundred forty patients (40.2%) resided in Washington, DC, 151 (43.4%) in Maryland, and 19 (5.5%) in Virginia. HFO was received by 86 patients (24.7%), 33 (9.5%) required intubation and mechanical ventilation, and 57 (16.4%) died. All intubations and deaths occurred among patients aged > 50 years, with death occurring in 17.8% of patients aged > 50 years.

Demographic characteristics and baseline comorbidities associated with COVID-19 disease severity can be found in eAppendix 2. In unadjusted analyses, age was significantly associated with the risk of HFO, with a mean (SD) age of 72.5 (11.7) years among those requiring HFO and 67.1 (14.4) years among patients without HFO (odds ratio [OR], 1.03; 95% CI, 1.01-1.05; P = .002). Although age was not associated with the risk of intubation, it was significantly associated with mortality. Patients who died had a mean (SD) age of 76.8 (11.8) years compared with 66.8 (13.7) years among survivors (OR, 1.06; 95% CI, 1.04-1.09; P < .001).

Compared with patients with no comorbidities, CCI categories of mild, moderate, and severe were associated with increased risk of requiring HFO (eAppendix 3). The adjusted OR (aOR) was highest among patients with severe CCI (aOR, 7.00; 95% CI, 2.42-20.32; P = .0007). In age-adjusted analyses, CCI was not associated with intubation or mortality.

Geospatial Analyses

State of residence, county of residence, and geographic area (including Washington, DC wards, and geographic divisions within counties of residence in Maryland and Virginia) were not associated with the clinical outcomes studied (eAppendix 4). However, zip code-based median income, analyzed as a continuous variable, was associated with a reduced likelihood of receiving HFO (aOR, 0.91; 95% CI, 0.84-0.99; P = .03). Income was not significantly associated with intubation or mortality.

The majority of patients hospitalized for COVID-19 at WVAMC resided in zip codes in eastern Washington, DC, inclusive of wards 7 and 8, and Prince George’s County, Maryland (Figure 1). These areas also corresponded to the lowest median household income by zip code (Figure 2).

Multivariable Analysis

Significant predictors of HFO requirement included comorbid diabetes (OR, 2.42; 95% CI, 1.27-4.61; P = .006) and liver disease or cirrhosis (OR, 2.19; 95% CI, 1.09-4.39; P = .02) (Table 2). CDC admission period was also associated with HFO need. Patients admitted after early 2020 had lower odds of receiving HFO. Race and median income based on zip code residence were not associated with HFO requirement.

Comorbid liver disease or cirrhosis was a significant predictor of intubation (OR, 2.81; 95% CI, 1.07-7.40; P = .03). CDC admission period was associated with intubation with lower odds of intubation for patients admitted after early 2020. Race and median income by zip code were not associated with intubation.

Significant predictors of mortality included age (OR, 2.20; 95% CI, 1.55-3.14; P = .0001), comorbid liver disease or cirrhosis (OR, 2.97; 95% CI, 1.31-6.74; P = .008), and OSA (OR, 3.45; 95% CI, 1.49-7.97; P = .003). CDC admission period was associated with mortality, with lower odds of intubation for patients admitted in mid- and late 2020. Race and median income by zip code residence were not associated with intubation.

Discussion

In this study of COVID-19 disease severity at a large integrated health care system that provides equal access to care, race, ethnicity, and geographic location were not associated with the need for HFO, intubation, or presumed mortality. Median income by zip code residence was associated with reduced HFO use in univariable analyses but not in multivariable models.

These findings support existing literature suggesting that race and ethnicity alone do not explain disparities in COVID-19 outcomes. Multiple studies have demonstrated that disparities in health outcomes have been reduced for patients receiving VHA care.^6,16-19 However, even within a health care system with assumed equal access, the finding of an association between income and need for HFO in the univariable analysis may reflect a greater likelihood of delays in care due to structural barriers. Multiple studies suggest low SES may be an independent risk factor for severe COVID-19 disease. Individuals with low SES have higher rates of chronic diseases of obesity, diabetes, heart disease, and lung disease; thus, they are also at greater risk of serious illness with COVID-19.^20-24 Socioeconomic disadvantage may also have limited individuals’ ability to engage in protective behaviors to reduce COVID-19 infection risk, including food stockpiling, social distancing, avoidance of public transportation, and refraining from working in “essential jobs.”²¹

Beyond SES, place of residence also influences health outcomes. Prior literature supports using zip codes to assess area-based SES status and monitor health disparities.²⁵ The Social Vulnerability Index incorporates SES factors for communities and measures social determinates of health at a zip code level exclusive of race and ethnicity.²⁶ Socially vulnerable communities are known to have higher rates of chronic diseases, COVID-19 mortality, and lower vaccination rates.³ Within a defined geographic area, an individual’s outcome for COVID-19 can be influenced by individual resources such as access to care and median income. Disposable income may mitigate COVID-19 risk by facilitating timely care, reducing occupational exposure, improving housing stability, and supporting health-promoting behaviors.²¹

Limitations

Due to the evolving nature of the COVID-19 pandemic, variants, treatments, and interventions varied throughout the study period and are not included in this analysis. In late December 2020, COVID-19 vaccination was approved with a tiered allocation for at-risk patients and direct health care professionals. Three of the 4 study periods analyzed in this study were prior to vaccine rollout and therefore vaccination history was not assessed. However, we tried to capture the evolving changes in COVID-19 variants, treatments and interventions, and skill in treating the disease through use of CDC-defined time frames. Another limitation is that some studies have shown that use of median income by zip code residence can underestimate mortality.²⁷ Also, shared resources and access to other sources of disposable income can impact the immediate attainment of social needs. For example, during the COVID-19 pandemic, health care systems in Washington, DC assisted vulnerable individuals by providing food, housing, and other resources.^28,29 Finally, the modest sample size limits generalizability and power to detect differences for certain variables, including Hispanic ethnicity.

Conclusions

There have been widely described disparities in disease severity and death during the COVID-19 pandemic. In this urban veteran cohort of hospitalized patients, there was no difference in the need for intubation or mortality associated with race. The findings suggest that a lower median income by zip code residence may be associated with greater disease severity at presentation, but do not predict severe outcomes and mortality overall. VHA care, which provides equal access to care, may mitigate the disparities seen in the private sector.

Large epidemiologic studies have shown disparities in COVID-19 outcomes by race, ethnicity, and socioeconomic status (SES). Racial and ethnic minorities and individuals of lower SES have experienced disproportionately higher rates of intensive care unit (ICU) admission and death. In Washington, DC, Black individuals (47% of the population) accounted for 51% of COVID-19 cases and 75% of deaths. In comparison, White individuals (41% of the population) accounted for 21% of cases and 11% of deaths.¹ Place of residence, such as living in socially vulnerable communities, has also been shown to be associated with higher rates of COVID-19 mortality and lower vaccination rates.^2-4 Social and structural inequities, such as limited access to health care services and mistrust of the health care system, may explain some of the observed disparities.⁵ However, data are limited regarding COVID-19 outcomes for individuals with equal access to care.

The Veterans Health Administration (VHA) is the largest integrated US health care system and operates 123 acute care hospitals. Previous research has demonstrated that disparities in outcomes for other diseases are attenuated or erased among veterans receiving VHA care.^6,7 Based on literature from the pandemic, markers of health care inequity relating to SES (eg, place of residence, median income) are expected to impact the outcomes of patients acutely hospitalized with COVID-19.⁴ We hypothesized that the impact on clinical outcomes of infection would be mitigated for veterans receiving VHA care.

This retrospective cohort study included veterans who presented to Washington Veterans Affairs Medical Center (WVAMC) with the goal of determining whether place of residence as a marker of SES, health care access, and median income were predictive of COVID-19 disease severity.

Methods

The WVAMC serves about 125,000 veterans across the metropolitan area, including parts of Maryland and Virginia. It is a high-complexity hospital with 164 acute care beds, 30 psychosocial residential rehabilitation beds, and an adjacent 120-bed community living center providing long-term, hospice, and palliative care.⁸

The WVAMC developed a dashboard that tracked patients with COVID-19 through on-site testing by admission date, ward, and other key demographics (PowerBi, Corporate Data Warehouse). All patients admitted to WVAMC with a diagnosis of COVID-19 between March 1, 2020, and June 30, 2021, were included in this retrospective review. Using the Computerized Patient Record System (CPRS) and the dashboard, we collected demographic information, baseline clinical diagnoses, laboratory results, and clinical interventions for all patients with documented COVID-19 infection as established by laboratory testing methods available at the time of diagnosis. Veterans treated exclusively outside the WVAMC were excluded. Hospitalization was defined as any acute inpatient admission or transfer recorded within 5 days before and 30 days after the laboratory collection of a positive COVID-19 test. Home testing kits were not widely available during the study period. An ICU stay was defined as any inpatient admission or transfer recorded within 5 days before or 30 days after the laboratory collection of a positive COVID-19 test for which the ward location had the specialty of medical or surgical ICU. Death due to COVID-19 was defined as occurring within 42 days (6 weeks) of a positive COVID-19 test.⁹ This definition assumed that during the peak of the pandemic, COVID-19 was the attributable cause of death, despite the possible contribution of underlying health conditions.

Patients’ admission periods were based on US Centers for Disease Control and Prevention (CDC) national data and classified as early 2020 (January 2020–April 2020), mid-2020 (May 2020–August 2020), late 2020 (September 2020–December 2020), and early 2021 (January 2021–April 2021).¹⁰ We chose to use these time periods as surrogates for the frequent changes in circulating COVID-19 variants, surges in case numbers, therapies and interventions available during the pandemic. The dominant COVID-19 variant during the study period was Alpha (B.1.17). Beta (B.1.351) variants were circulating infrequently, and Delta and Omicron appeared after the study period.¹¹ Treatment strategies evolved rapidly with emerging evidence, including the use of dexamethasone, beginning in June 2020.¹² WVAMC followed the Advisory Committee on Immunization Practices guidance on vaccination rollout beginning in December 2020.¹³

Patients' income was estimated by the median household income of the zip code residence based on US Census Bureau 2021 estimates and was assessed as both a continuous and categorical variable.¹⁴ The Charlson Comorbidity Index (CCI) was included in models as a continuous variable.¹⁵ Variables contributing to the CCI include myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, dementia, hemiplegia or paraplegia, ulcer disease, hepatic disease, diabetes (with or without end-organ damage), chronic obstructive pulmonary disease (COPD), connective tissue disease, leukemia, lymphoma, moderate or severe renal disease, solid tumor (with or without metastases), and HIV/AIDS. The WVAMC Institutional Review Board approved this study (IRB #1573071).

Variables

This study assessed 3 primary outcomes as indicators of disease severity during hospitalization: need for high-flow oxygen (HFO), intubation, and presumed mortality at any time during hospitalization. The following variables were collected as potential social determinants or clinical risk-adjustment predictors of disease severity outcomes: age; sex; race and ethnicity; median income for patient’s zip code residence, state, and county; wards within Washington, DC; comorbidities, CCI; tobacco use; and body mass index.¹⁵ Although medications at baseline, treatments during hospitalization for COVID-19, and laboratory parameters during hospitalization are shown in eAppendices 1 and 2, they are beyond the scope of this analysis.

Statistical Analysis

Three types of logistic regression models were calculated for predicting the disease severity outcomes: (1) simple unadjusted models; (2) models predicting from single variables plus age (age-adjusted); and (3) multivariable models using all nonredundant potential predictors with adequate sample sizes (multivariable). Variables were considered to have inadequate sample sizes if there was nontrivial missing data or small numbers within categories, (eg, AIDS, connective tissue disease). Potential predictors for the multivariable model included age, sex, race, median income by zip code residence, CCI, CDC admission period, obesity, hypertension, chronic kidney disease, obstructive sleep apnea (OSA), diabetes, COPD or asthma, liver disease, antibiotics, and acute kidney injury.

For the multivariable models, the following modifications were made to avoid unreliable parameter estimation and computation problems (quasi-separation): age and CCI were included as continuous rather than categorical variables. Race was recoded as a 2-category variable (Black vs other [White, Hispanic, American Indian, Alaska Native, Asian, Native Hawaiian, and Pacific Islander]), and ethnicity was excluded because of the small number of patients in this group (n = 16). Admission period was included. Predicted probability plots were generated for each outcome with continuous independent predictors (income and CCI), both unadjusted and adjusted for age as a continuous covariate. All analyses were performed using SAS version 9.4.

Heat Maps

Heat maps were generated to visualize the geospatial distribution of COVID-19 cases and median incomes across zip codes in the greater Washington, DC area. Patient case data and median income, aggregated by zip code, were imported using ArcGIS Online. A zip code boundary layer from Esri (United States Zip Code Boundaries) was used to spatially align the case data. Data were joined by matching zip codes or median incomes in the patient dataset to those in the boundary layer. The resulting polygon layer was styled using the Counts and Amounts (Color) symbology in ArcGIS Online, with case counts or median income determining the intensity of the color gradient.

Results

Between March 1, 2020, and June 30, 2021, 348 patients were hospitalized with COVID-19 (Table 1). The mean (SD) age was 68.4 (13.9) years, 313 patients (90.2%) were male, 281 patients (83.4%) were Black, 47 patients (13.6%) were White, and 16 patients (4.8%) were Hispanic. One hundred forty patients (40.2%) resided in Washington, DC, 151 (43.4%) in Maryland, and 19 (5.5%) in Virginia. HFO was received by 86 patients (24.7%), 33 (9.5%) required intubation and mechanical ventilation, and 57 (16.4%) died. All intubations and deaths occurred among patients aged > 50 years, with death occurring in 17.8% of patients aged > 50 years.

Demographic characteristics and baseline comorbidities associated with COVID-19 disease severity can be found in eAppendix 2. In unadjusted analyses, age was significantly associated with the risk of HFO, with a mean (SD) age of 72.5 (11.7) years among those requiring HFO and 67.1 (14.4) years among patients without HFO (odds ratio [OR], 1.03; 95% CI, 1.01-1.05; P = .002). Although age was not associated with the risk of intubation, it was significantly associated with mortality. Patients who died had a mean (SD) age of 76.8 (11.8) years compared with 66.8 (13.7) years among survivors (OR, 1.06; 95% CI, 1.04-1.09; P < .001).

Compared with patients with no comorbidities, CCI categories of mild, moderate, and severe were associated with increased risk of requiring HFO (eAppendix 3). The adjusted OR (aOR) was highest among patients with severe CCI (aOR, 7.00; 95% CI, 2.42-20.32; P = .0007). In age-adjusted analyses, CCI was not associated with intubation or mortality.

Geospatial Analyses

State of residence, county of residence, and geographic area (including Washington, DC wards, and geographic divisions within counties of residence in Maryland and Virginia) were not associated with the clinical outcomes studied (eAppendix 4). However, zip code-based median income, analyzed as a continuous variable, was associated with a reduced likelihood of receiving HFO (aOR, 0.91; 95% CI, 0.84-0.99; P = .03). Income was not significantly associated with intubation or mortality.

The majority of patients hospitalized for COVID-19 at WVAMC resided in zip codes in eastern Washington, DC, inclusive of wards 7 and 8, and Prince George’s County, Maryland (Figure 1). These areas also corresponded to the lowest median household income by zip code (Figure 2).

Multivariable Analysis

Significant predictors of HFO requirement included comorbid diabetes (OR, 2.42; 95% CI, 1.27-4.61; P = .006) and liver disease or cirrhosis (OR, 2.19; 95% CI, 1.09-4.39; P = .02) (Table 2). CDC admission period was also associated with HFO need. Patients admitted after early 2020 had lower odds of receiving HFO. Race and median income based on zip code residence were not associated with HFO requirement.

Comorbid liver disease or cirrhosis was a significant predictor of intubation (OR, 2.81; 95% CI, 1.07-7.40; P = .03). CDC admission period was associated with intubation with lower odds of intubation for patients admitted after early 2020. Race and median income by zip code were not associated with intubation.

Significant predictors of mortality included age (OR, 2.20; 95% CI, 1.55-3.14; P = .0001), comorbid liver disease or cirrhosis (OR, 2.97; 95% CI, 1.31-6.74; P = .008), and OSA (OR, 3.45; 95% CI, 1.49-7.97; P = .003). CDC admission period was associated with mortality, with lower odds of intubation for patients admitted in mid- and late 2020. Race and median income by zip code residence were not associated with intubation.

Discussion

In this study of COVID-19 disease severity at a large integrated health care system that provides equal access to care, race, ethnicity, and geographic location were not associated with the need for HFO, intubation, or presumed mortality. Median income by zip code residence was associated with reduced HFO use in univariable analyses but not in multivariable models.

These findings support existing literature suggesting that race and ethnicity alone do not explain disparities in COVID-19 outcomes. Multiple studies have demonstrated that disparities in health outcomes have been reduced for patients receiving VHA care.^6,16-19 However, even within a health care system with assumed equal access, the finding of an association between income and need for HFO in the univariable analysis may reflect a greater likelihood of delays in care due to structural barriers. Multiple studies suggest low SES may be an independent risk factor for severe COVID-19 disease. Individuals with low SES have higher rates of chronic diseases of obesity, diabetes, heart disease, and lung disease; thus, they are also at greater risk of serious illness with COVID-19.^20-24 Socioeconomic disadvantage may also have limited individuals’ ability to engage in protective behaviors to reduce COVID-19 infection risk, including food stockpiling, social distancing, avoidance of public transportation, and refraining from working in “essential jobs.”²¹

Beyond SES, place of residence also influences health outcomes. Prior literature supports using zip codes to assess area-based SES status and monitor health disparities.²⁵ The Social Vulnerability Index incorporates SES factors for communities and measures social determinates of health at a zip code level exclusive of race and ethnicity.²⁶ Socially vulnerable communities are known to have higher rates of chronic diseases, COVID-19 mortality, and lower vaccination rates.³ Within a defined geographic area, an individual’s outcome for COVID-19 can be influenced by individual resources such as access to care and median income. Disposable income may mitigate COVID-19 risk by facilitating timely care, reducing occupational exposure, improving housing stability, and supporting health-promoting behaviors.²¹

Limitations

Due to the evolving nature of the COVID-19 pandemic, variants, treatments, and interventions varied throughout the study period and are not included in this analysis. In late December 2020, COVID-19 vaccination was approved with a tiered allocation for at-risk patients and direct health care professionals. Three of the 4 study periods analyzed in this study were prior to vaccine rollout and therefore vaccination history was not assessed. However, we tried to capture the evolving changes in COVID-19 variants, treatments and interventions, and skill in treating the disease through use of CDC-defined time frames. Another limitation is that some studies have shown that use of median income by zip code residence can underestimate mortality.²⁷ Also, shared resources and access to other sources of disposable income can impact the immediate attainment of social needs. For example, during the COVID-19 pandemic, health care systems in Washington, DC assisted vulnerable individuals by providing food, housing, and other resources.^28,29 Finally, the modest sample size limits generalizability and power to detect differences for certain variables, including Hispanic ethnicity.

Conclusions

There have been widely described disparities in disease severity and death during the COVID-19 pandemic. In this urban veteran cohort of hospitalized patients, there was no difference in the need for intubation or mortality associated with race. The findings suggest that a lower median income by zip code residence may be associated with greater disease severity at presentation, but do not predict severe outcomes and mortality overall. VHA care, which provides equal access to care, may mitigate the disparities seen in the private sector.

To improve safety performance, many health care organizations have embarked on the journey to becoming high reliability organizations (HROs). HROs operate in complex, high-risk, constantly changing environments and avoid catastrophic events despite the inherent risks.¹ HROs maintain high levels of safety and reliability by adhering to core principles, foundational practices, rigorous processes, a strong organizational culture, and continuous learning and process improvement.^1-3

Becoming an HRO requires understanding what makes systems safer for patients and staff at all levels by taking ownership of 5 principles: (1) sensitivity to operations (increased awareness of the current status of systems); (2) reluctance to simplify (avoiding oversimplification of the cause[s] of problems); (3) preoccupation with failure (anticipating risks that might be symptomatic of a larger problem); (4) deference to expertise (relying on the most qualified individuals to make decisions); and (5) commitment to resilience (planning for potential failure and being prepared to respond).^1,2,4 In addition to these, the Veterans Health Administration has identified 3 pillars of HROs: leadership commitment (safety and reliability are central to leadership vision, decision-making, and action-oriented behaviors), safety culture (across the organization, safety values are key to preventing harm and learning from mistakes), and continuous process improvement (promoting constant learning and improvement with evidence-based tools and methodologies).⁵

Implementing these principles is not enough to achieve high reliability. This transition requires significant change, which can be met with resistance. Without attending to organizational change, implementation of HRO principles can be superficial, scattered, and isolated.⁶ Large organizations often struggle with change as it conflicts with the fundamental human need for stability and security.⁷ Consequently, the journey to becoming an HRO requires an understanding of the reasons for resistance to change (RtC) as well as evidence-based strategies.

REASONS FOR RESISTANCE TO CHANGE

RtC is the informal and covert behavior of an individual or group to a particular change. RtC is commonly recognized as the failure of employees to do anything requested by managers and is a main reason change initiatives fail.⁸ While some staff see change as opportunities for learning and growth, others resist based on uncertainty about how the changes will impact their current work situation, or fear, frustration, confusion, and distrust.^8,9 Resistance can overtly manifest with some staff publicly expressing their discontent in public without offering solutions, or covertly by ignoring the change or avoiding participation in any aspect of the change process. Both forms of RtC are equally detrimental.⁸

Frequent changes in organizations can also cause cynicism. Employees will view the change as something initially popular, but will only last until another change comes along.^8,9 Resistance can result in the failure to achieve desired objectives, wasted time, effort, and resources, decreased momentum, and loss of confidence and trust in leaders to effectively manage the change process.⁹ To understand RtC, 3 main factors must be considered: individual, interpersonal, and organizational.

Individual

An individual’s personality can be an important indicator for how they will respond to change. Some individuals welcome and thrive on change while others resist in preference for the status quo.^8,10 Individuals will also resist change if they believe their position, power, or prestige within the organization are in jeopardy or that the change is contrary to current personal or organizational values, principles, and objectives.^8-12 Resistance can also be the result of uncertainty about what the change means, lack of information regarding the change, or questioning motives for the change.⁹

Interpersonal

Another influence on RtC is the interpersonal factors of employees. The personal satisfaction individuals receive from their work and the type of interactions they experience with colleagues can impact RtC. When communication with colleagues is lacking before and during change implementation, negative reactions to the change can fuel resistance.¹¹ Cross-functional and bidirectional communication is vital; its absence can leave staff feeling inadequately informed and less supportive of the change.⁸ Employees’ understanding of changes through communication between other members of the organization is critical to success.¹¹

Organizational

How organizational leaders introduce change affects the extent to which staff respond.¹⁰ RtC can emerge if staff feel change is imposed on them. Change is better received when people are actively engaged in the process and adopt a sense of ownership that will ultimately affect them and their role within the organization.^12,13 Organizations are also better equipped to address potential RtC when leadership is respected and have a genuine concern for the overall well-being of staff members. Organizational leaders who mainly focus on the bottom line and have little regard for staff are more likely to be perceived as untrustworthy, which contributes to RtC.^9,13 Lack of proper education and guidance from organizational leaders, as well as poor communication, can lead to RtC.^8,13

MANAGING RESISTANCE TO CHANGE

RtC can be a significant factor in the success or failure of the change process. Poorly managed change can exponentially increase resistance, necessitating a multifaceted approach to managing RtC, while well-managed change can result in a high success rate. Evidence-based strategies to counter RtC focus on communication, employee participation, education and training, and engaging managers.⁸

Communication

Open and effective communication is critical to managing RtC, as uncertainty often exaggerates the negative aspects of change. Effective communication involves active listening, with leadership and management addressing employee concerns in a clear and concise manner. A psychologically safe culture for open dialogue is essential when addressing RtC.^9,14,15 Psychological safety empowers staff to speak up, ask questions, and offer ideas, forming a solid basis for open and effective communication and participation. Leaders and managers should create opportunities for open dialogue for all members of the organization throughout the process. This can be accomplished with one-on-one meetings, open forums, town hall meetings, electronic mail, newsletters, and social media. Topics should cover the reasons for change, details of what is changing, the individual, organizational, and patient risks of not changing, as well as the benefits of changing.⁹ Encouraging staff to ask questions and provide feedback to promote bidirectional and closed-loop communication is essential to avoid misunderstandings.^9,15 While open communication is essential, leaders must carefully plan what information to share, how much to share, and how to avoid information overload. Information about the change should be timely, adequate, applicable, and informative.¹⁵ The HRO practice of leader rounding for high reliability can be instrumental to ensure effective, bidirectional communication and collaboration among all disciplines across a health care organization through improving leadership visibility during times of change and enhancing interactions and communication with staff.³

Employee Participation

Involving staff in the change process significantly reduces RtC. Engagement fosters ownership in the change process, increasing the likelihood employees will support and even champion it. Health care professionals welcome opportunities to be involved in helping with aspects of organizational change, especially when invited to participate in the change early in the process and throughout the course of change.^7,14,15

Leaders should encourage staff to provide feedback to understand the impact the change is having on them and their roles and responsibilities within the organization. This exemplifies the HRO principle of deference to expertise as the employee often has the most in-depth knowledge of their work setting. Employee perspectives can significantly influence the success of change initatives.^7,14 Participation is impactful in providing employees with a sense of agency facilitating acceptance and improving desire to adopt the change.¹⁴

Tiered safety huddles and visual management systems (VMSs) also can engage staff. Tiered safety huddles provide a forum for transparent communication, increasing situational awareness, and improving a health care organization’s ability to appropriately respond to staff questions, suggestions, and concerns. VMSs display the status and progress toward organizational goals during the change process, and are highly effective in creating environments where staff feel empowered to voice concerns related to the change process.³

Education and Training

Educating employees on the value of change is crucial to overcome RtC. RtC often stems from employees not feeling prepared to adapt or adopt new processes. Health care professionals who do not receive information about change are less likely to support it.^7,12,15 Staff are more likely to accept change when they understand why it is needed and how it impacts the organization’s long-term mission.^11,15 Timely, compelling, and informative education on how to adapt to the change will promote more positive appraisal of the change and reduce RtC.^8,15 Employees must feel confident they will receive the appropriate training, resources, and support to successfully adapt to the change. This requires leaders and managers taking time to clarify expectations, conduct a gap analysis to identify the skills and knowledge needed to support the planned change, and provide sufficient educational opportunities to fill those gaps.⁸ For example, the US Department of Veterans Affairs offers classes to employees on the Prosci ADKAR (Awareness, Desire, Knowledge, Ability, and Reinforcement) Model. This training provides individuals with the information and skills needed for change to be successful.¹⁶

Safety forums can be influential and allow leadership to educate staff on updates related to change processes and promote bidirectional communication.³ In safety forums, staff have an opportunity to ask questions, especially as they relate to learning about available resources to become more informed about the organizational changes.

Engaging Managers

Managers are pivotal to the successful implementation of organizational change.⁸ They serve as the bridge between senior leadership and frontline employees and are positioned to influence the adoption and success of change initiatives. Often the first point of contact for employees, managers can effectively communicate the need for change, and act as the liaison to align it with individual employee motivations. Since they are often the first to encounter resistance among employees, managers serve as advocates through the process. Through a coaching role, managers can help employees develop the knowledge and ability to be successful and thrive in the new environment. The Table summarizes the evidence-based strategies.

CONCLUSIONS

Implementing change in health care organizations can be challenging, especially on the journey to high reliability. RtC is the result of factors at the individual, interpersonal, and organizational levels that leaders must address to increase chances for success. Organizational changes in health care are more likely to succeed when staff understand why the change is needed through open and continuous communication, can influence the change by sharing their own perspectives, and have the knowledge, skills, and resources to prepare for and participate in the process.

To improve safety performance, many health care organizations have embarked on the journey to becoming high reliability organizations (HROs). HROs operate in complex, high-risk, constantly changing environments and avoid catastrophic events despite the inherent risks.¹ HROs maintain high levels of safety and reliability by adhering to core principles, foundational practices, rigorous processes, a strong organizational culture, and continuous learning and process improvement.^1-3

Becoming an HRO requires understanding what makes systems safer for patients and staff at all levels by taking ownership of 5 principles: (1) sensitivity to operations (increased awareness of the current status of systems); (2) reluctance to simplify (avoiding oversimplification of the cause[s] of problems); (3) preoccupation with failure (anticipating risks that might be symptomatic of a larger problem); (4) deference to expertise (relying on the most qualified individuals to make decisions); and (5) commitment to resilience (planning for potential failure and being prepared to respond).^1,2,4 In addition to these, the Veterans Health Administration has identified 3 pillars of HROs: leadership commitment (safety and reliability are central to leadership vision, decision-making, and action-oriented behaviors), safety culture (across the organization, safety values are key to preventing harm and learning from mistakes), and continuous process improvement (promoting constant learning and improvement with evidence-based tools and methodologies).⁵

Implementing these principles is not enough to achieve high reliability. This transition requires significant change, which can be met with resistance. Without attending to organizational change, implementation of HRO principles can be superficial, scattered, and isolated.⁶ Large organizations often struggle with change as it conflicts with the fundamental human need for stability and security.⁷ Consequently, the journey to becoming an HRO requires an understanding of the reasons for resistance to change (RtC) as well as evidence-based strategies.

REASONS FOR RESISTANCE TO CHANGE

RtC is the informal and covert behavior of an individual or group to a particular change. RtC is commonly recognized as the failure of employees to do anything requested by managers and is a main reason change initiatives fail.⁸ While some staff see change as opportunities for learning and growth, others resist based on uncertainty about how the changes will impact their current work situation, or fear, frustration, confusion, and distrust.^8,9 Resistance can overtly manifest with some staff publicly expressing their discontent in public without offering solutions, or covertly by ignoring the change or avoiding participation in any aspect of the change process. Both forms of RtC are equally detrimental.⁸

Frequent changes in organizations can also cause cynicism. Employees will view the change as something initially popular, but will only last until another change comes along.^8,9 Resistance can result in the failure to achieve desired objectives, wasted time, effort, and resources, decreased momentum, and loss of confidence and trust in leaders to effectively manage the change process.⁹ To understand RtC, 3 main factors must be considered: individual, interpersonal, and organizational.

Individual

An individual’s personality can be an important indicator for how they will respond to change. Some individuals welcome and thrive on change while others resist in preference for the status quo.^8,10 Individuals will also resist change if they believe their position, power, or prestige within the organization are in jeopardy or that the change is contrary to current personal or organizational values, principles, and objectives.^8-12 Resistance can also be the result of uncertainty about what the change means, lack of information regarding the change, or questioning motives for the change.⁹

Interpersonal

Another influence on RtC is the interpersonal factors of employees. The personal satisfaction individuals receive from their work and the type of interactions they experience with colleagues can impact RtC. When communication with colleagues is lacking before and during change implementation, negative reactions to the change can fuel resistance.¹¹ Cross-functional and bidirectional communication is vital; its absence can leave staff feeling inadequately informed and less supportive of the change.⁸ Employees’ understanding of changes through communication between other members of the organization is critical to success.¹¹

Organizational

How organizational leaders introduce change affects the extent to which staff respond.¹⁰ RtC can emerge if staff feel change is imposed on them. Change is better received when people are actively engaged in the process and adopt a sense of ownership that will ultimately affect them and their role within the organization.^12,13 Organizations are also better equipped to address potential RtC when leadership is respected and have a genuine concern for the overall well-being of staff members. Organizational leaders who mainly focus on the bottom line and have little regard for staff are more likely to be perceived as untrustworthy, which contributes to RtC.^9,13 Lack of proper education and guidance from organizational leaders, as well as poor communication, can lead to RtC.^8,13

MANAGING RESISTANCE TO CHANGE

RtC can be a significant factor in the success or failure of the change process. Poorly managed change can exponentially increase resistance, necessitating a multifaceted approach to managing RtC, while well-managed change can result in a high success rate. Evidence-based strategies to counter RtC focus on communication, employee participation, education and training, and engaging managers.⁸

Communication

Open and effective communication is critical to managing RtC, as uncertainty often exaggerates the negative aspects of change. Effective communication involves active listening, with leadership and management addressing employee concerns in a clear and concise manner. A psychologically safe culture for open dialogue is essential when addressing RtC.^9,14,15 Psychological safety empowers staff to speak up, ask questions, and offer ideas, forming a solid basis for open and effective communication and participation. Leaders and managers should create opportunities for open dialogue for all members of the organization throughout the process. This can be accomplished with one-on-one meetings, open forums, town hall meetings, electronic mail, newsletters, and social media. Topics should cover the reasons for change, details of what is changing, the individual, organizational, and patient risks of not changing, as well as the benefits of changing.⁹ Encouraging staff to ask questions and provide feedback to promote bidirectional and closed-loop communication is essential to avoid misunderstandings.^9,15 While open communication is essential, leaders must carefully plan what information to share, how much to share, and how to avoid information overload. Information about the change should be timely, adequate, applicable, and informative.¹⁵ The HRO practice of leader rounding for high reliability can be instrumental to ensure effective, bidirectional communication and collaboration among all disciplines across a health care organization through improving leadership visibility during times of change and enhancing interactions and communication with staff.³

Employee Participation

Involving staff in the change process significantly reduces RtC. Engagement fosters ownership in the change process, increasing the likelihood employees will support and even champion it. Health care professionals welcome opportunities to be involved in helping with aspects of organizational change, especially when invited to participate in the change early in the process and throughout the course of change.^7,14,15

Leaders should encourage staff to provide feedback to understand the impact the change is having on them and their roles and responsibilities within the organization. This exemplifies the HRO principle of deference to expertise as the employee often has the most in-depth knowledge of their work setting. Employee perspectives can significantly influence the success of change initatives.^7,14 Participation is impactful in providing employees with a sense of agency facilitating acceptance and improving desire to adopt the change.¹⁴

Tiered safety huddles and visual management systems (VMSs) also can engage staff. Tiered safety huddles provide a forum for transparent communication, increasing situational awareness, and improving a health care organization’s ability to appropriately respond to staff questions, suggestions, and concerns. VMSs display the status and progress toward organizational goals during the change process, and are highly effective in creating environments where staff feel empowered to voice concerns related to the change process.³

Education and Training

Educating employees on the value of change is crucial to overcome RtC. RtC often stems from employees not feeling prepared to adapt or adopt new processes. Health care professionals who do not receive information about change are less likely to support it.^7,12,15 Staff are more likely to accept change when they understand why it is needed and how it impacts the organization’s long-term mission.^11,15 Timely, compelling, and informative education on how to adapt to the change will promote more positive appraisal of the change and reduce RtC.^8,15 Employees must feel confident they will receive the appropriate training, resources, and support to successfully adapt to the change. This requires leaders and managers taking time to clarify expectations, conduct a gap analysis to identify the skills and knowledge needed to support the planned change, and provide sufficient educational opportunities to fill those gaps.⁸ For example, the US Department of Veterans Affairs offers classes to employees on the Prosci ADKAR (Awareness, Desire, Knowledge, Ability, and Reinforcement) Model. This training provides individuals with the information and skills needed for change to be successful.¹⁶

Safety forums can be influential and allow leadership to educate staff on updates related to change processes and promote bidirectional communication.³ In safety forums, staff have an opportunity to ask questions, especially as they relate to learning about available resources to become more informed about the organizational changes.

Engaging Managers

Managers are pivotal to the successful implementation of organizational change.⁸ They serve as the bridge between senior leadership and frontline employees and are positioned to influence the adoption and success of change initiatives. Often the first point of contact for employees, managers can effectively communicate the need for change, and act as the liaison to align it with individual employee motivations. Since they are often the first to encounter resistance among employees, managers serve as advocates through the process. Through a coaching role, managers can help employees develop the knowledge and ability to be successful and thrive in the new environment. The Table summarizes the evidence-based strategies.

CONCLUSIONS

Implementing change in health care organizations can be challenging, especially on the journey to high reliability. RtC is the result of factors at the individual, interpersonal, and organizational levels that leaders must address to increase chances for success. Organizational changes in health care are more likely to succeed when staff understand why the change is needed through open and continuous communication, can influence the change by sharing their own perspectives, and have the knowledge, skills, and resources to prepare for and participate in the process.

To improve safety performance, many health care organizations have embarked on the journey to becoming high reliability organizations (HROs). HROs operate in complex, high-risk, constantly changing environments and avoid catastrophic events despite the inherent risks.¹ HROs maintain high levels of safety and reliability by adhering to core principles, foundational practices, rigorous processes, a strong organizational culture, and continuous learning and process improvement.^1-3

Becoming an HRO requires understanding what makes systems safer for patients and staff at all levels by taking ownership of 5 principles: (1) sensitivity to operations (increased awareness of the current status of systems); (2) reluctance to simplify (avoiding oversimplification of the cause[s] of problems); (3) preoccupation with failure (anticipating risks that might be symptomatic of a larger problem); (4) deference to expertise (relying on the most qualified individuals to make decisions); and (5) commitment to resilience (planning for potential failure and being prepared to respond).^1,2,4 In addition to these, the Veterans Health Administration has identified 3 pillars of HROs: leadership commitment (safety and reliability are central to leadership vision, decision-making, and action-oriented behaviors), safety culture (across the organization, safety values are key to preventing harm and learning from mistakes), and continuous process improvement (promoting constant learning and improvement with evidence-based tools and methodologies).⁵

Implementing these principles is not enough to achieve high reliability. This transition requires significant change, which can be met with resistance. Without attending to organizational change, implementation of HRO principles can be superficial, scattered, and isolated.⁶ Large organizations often struggle with change as it conflicts with the fundamental human need for stability and security.⁷ Consequently, the journey to becoming an HRO requires an understanding of the reasons for resistance to change (RtC) as well as evidence-based strategies.

REASONS FOR RESISTANCE TO CHANGE

RtC is the informal and covert behavior of an individual or group to a particular change. RtC is commonly recognized as the failure of employees to do anything requested by managers and is a main reason change initiatives fail.⁸ While some staff see change as opportunities for learning and growth, others resist based on uncertainty about how the changes will impact their current work situation, or fear, frustration, confusion, and distrust.^8,9 Resistance can overtly manifest with some staff publicly expressing their discontent in public without offering solutions, or covertly by ignoring the change or avoiding participation in any aspect of the change process. Both forms of RtC are equally detrimental.⁸

Frequent changes in organizations can also cause cynicism. Employees will view the change as something initially popular, but will only last until another change comes along.^8,9 Resistance can result in the failure to achieve desired objectives, wasted time, effort, and resources, decreased momentum, and loss of confidence and trust in leaders to effectively manage the change process.⁹ To understand RtC, 3 main factors must be considered: individual, interpersonal, and organizational.

Individual

An individual’s personality can be an important indicator for how they will respond to change. Some individuals welcome and thrive on change while others resist in preference for the status quo.^8,10 Individuals will also resist change if they believe their position, power, or prestige within the organization are in jeopardy or that the change is contrary to current personal or organizational values, principles, and objectives.^8-12 Resistance can also be the result of uncertainty about what the change means, lack of information regarding the change, or questioning motives for the change.⁹

Interpersonal

Another influence on RtC is the interpersonal factors of employees. The personal satisfaction individuals receive from their work and the type of interactions they experience with colleagues can impact RtC. When communication with colleagues is lacking before and during change implementation, negative reactions to the change can fuel resistance.¹¹ Cross-functional and bidirectional communication is vital; its absence can leave staff feeling inadequately informed and less supportive of the change.⁸ Employees’ understanding of changes through communication between other members of the organization is critical to success.¹¹

Organizational

How organizational leaders introduce change affects the extent to which staff respond.¹⁰ RtC can emerge if staff feel change is imposed on them. Change is better received when people are actively engaged in the process and adopt a sense of ownership that will ultimately affect them and their role within the organization.^12,13 Organizations are also better equipped to address potential RtC when leadership is respected and have a genuine concern for the overall well-being of staff members. Organizational leaders who mainly focus on the bottom line and have little regard for staff are more likely to be perceived as untrustworthy, which contributes to RtC.^9,13 Lack of proper education and guidance from organizational leaders, as well as poor communication, can lead to RtC.^8,13

MANAGING RESISTANCE TO CHANGE

RtC can be a significant factor in the success or failure of the change process. Poorly managed change can exponentially increase resistance, necessitating a multifaceted approach to managing RtC, while well-managed change can result in a high success rate. Evidence-based strategies to counter RtC focus on communication, employee participation, education and training, and engaging managers.⁸

Communication

Open and effective communication is critical to managing RtC, as uncertainty often exaggerates the negative aspects of change. Effective communication involves active listening, with leadership and management addressing employee concerns in a clear and concise manner. A psychologically safe culture for open dialogue is essential when addressing RtC.^9,14,15 Psychological safety empowers staff to speak up, ask questions, and offer ideas, forming a solid basis for open and effective communication and participation. Leaders and managers should create opportunities for open dialogue for all members of the organization throughout the process. This can be accomplished with one-on-one meetings, open forums, town hall meetings, electronic mail, newsletters, and social media. Topics should cover the reasons for change, details of what is changing, the individual, organizational, and patient risks of not changing, as well as the benefits of changing.⁹ Encouraging staff to ask questions and provide feedback to promote bidirectional and closed-loop communication is essential to avoid misunderstandings.^9,15 While open communication is essential, leaders must carefully plan what information to share, how much to share, and how to avoid information overload. Information about the change should be timely, adequate, applicable, and informative.¹⁵ The HRO practice of leader rounding for high reliability can be instrumental to ensure effective, bidirectional communication and collaboration among all disciplines across a health care organization through improving leadership visibility during times of change and enhancing interactions and communication with staff.³

Employee Participation

Involving staff in the change process significantly reduces RtC. Engagement fosters ownership in the change process, increasing the likelihood employees will support and even champion it. Health care professionals welcome opportunities to be involved in helping with aspects of organizational change, especially when invited to participate in the change early in the process and throughout the course of change.^7,14,15

Leaders should encourage staff to provide feedback to understand the impact the change is having on them and their roles and responsibilities within the organization. This exemplifies the HRO principle of deference to expertise as the employee often has the most in-depth knowledge of their work setting. Employee perspectives can significantly influence the success of change initatives.^7,14 Participation is impactful in providing employees with a sense of agency facilitating acceptance and improving desire to adopt the change.¹⁴

Tiered safety huddles and visual management systems (VMSs) also can engage staff. Tiered safety huddles provide a forum for transparent communication, increasing situational awareness, and improving a health care organization’s ability to appropriately respond to staff questions, suggestions, and concerns. VMSs display the status and progress toward organizational goals during the change process, and are highly effective in creating environments where staff feel empowered to voice concerns related to the change process.³

Education and Training

Educating employees on the value of change is crucial to overcome RtC. RtC often stems from employees not feeling prepared to adapt or adopt new processes. Health care professionals who do not receive information about change are less likely to support it.^7,12,15 Staff are more likely to accept change when they understand why it is needed and how it impacts the organization’s long-term mission.^11,15 Timely, compelling, and informative education on how to adapt to the change will promote more positive appraisal of the change and reduce RtC.^8,15 Employees must feel confident they will receive the appropriate training, resources, and support to successfully adapt to the change. This requires leaders and managers taking time to clarify expectations, conduct a gap analysis to identify the skills and knowledge needed to support the planned change, and provide sufficient educational opportunities to fill those gaps.⁸ For example, the US Department of Veterans Affairs offers classes to employees on the Prosci ADKAR (Awareness, Desire, Knowledge, Ability, and Reinforcement) Model. This training provides individuals with the information and skills needed for change to be successful.¹⁶

Safety forums can be influential and allow leadership to educate staff on updates related to change processes and promote bidirectional communication.³ In safety forums, staff have an opportunity to ask questions, especially as they relate to learning about available resources to become more informed about the organizational changes.

Engaging Managers

Managers are pivotal to the successful implementation of organizational change.⁸ They serve as the bridge between senior leadership and frontline employees and are positioned to influence the adoption and success of change initiatives. Often the first point of contact for employees, managers can effectively communicate the need for change, and act as the liaison to align it with individual employee motivations. Since they are often the first to encounter resistance among employees, managers serve as advocates through the process. Through a coaching role, managers can help employees develop the knowledge and ability to be successful and thrive in the new environment. The Table summarizes the evidence-based strategies.

CONCLUSIONS

Implementing change in health care organizations can be challenging, especially on the journey to high reliability. RtC is the result of factors at the individual, interpersonal, and organizational levels that leaders must address to increase chances for success. Organizational changes in health care are more likely to succeed when staff understand why the change is needed through open and continuous communication, can influence the change by sharing their own perspectives, and have the knowledge, skills, and resources to prepare for and participate in the process.

Epilepsy affects about 4.5 million people in the United States and 150,000 new individuals are diagnosed each year.^1,2 In 2019, epilepsy-attributable health care spending for noninstitutionalized people was around $5.4 billion and total epilepsy-attributable and epilepsy or seizure health care-related costs totaled $54 billion.³

Accurate surveillance of epilepsy in large health care systems can potentially improve health care delivery and resource allocation. A 2012 Institute of Medicine (IOM) report identified 13 recommendations to guide public health action on epilepsy, including validation of standard definitions for case ascertainment, identification of epilepsy through screening programs or protocols, and expansion of surveillance to better understand disease burden.⁴

A systematic review of validation studies concluded that it is reasonable to use administrative data to identify people with epilepsy in epidemiologic research. Combining The International Classification of Diseases (ICD) codes for epilepsy (ICD-10, G40-41; ICD-9, 345) with antiseizure medications (ASMs) could provide high positive predictive values (PPVs) and combining symptoms codes for convulsions (ICD-10, R56; ICD-9, 780.3, 780.39) with ASMs could lead to high sensitivity.⁵ However, identifying individuals with epilepsy from administrative data in large managed health care organizations is challenging.⁶ The IOM report noted that large managed health care organizations presented varying incidence and prevalence estimates due to differing methodology, geographic area, demographics, and definitions of epilepsy.

The Veterans Health Administration (VHA) is the largest integrated US health care system, providing care to > 9.1 million veterans.⁷ To improve the health and well-being of veterans with epilepsy (VWEs), a network of sites was established in 2008 called the US Department of Veterans Affairs (VA) Epilepsy Centers of Excellence (ECoE). Subsequent to the creation of the ECoE, efforts were made to identify VWEs within VHA databases.^8,9 Prior to fiscal year (FY) 2016, the ECoE adopted a modified version of a well-established epilepsy diagnostic algorithm developed by Holden et al for large managed care organizations.¹⁰ The original algorithm identified patients by cross-matching ASMs with ICD-9 codes for an index year. But it failed to capture a considerable number of stable patients with epilepsy in the VHA due to incomplete documentation, and had false positives due to inclusion of patients identified from diagnostic clinics. The modified algorithm the ECoE used prior to FY 2016 considered additional prior years and excluded encounters from diagnostic clinics. The result was an improvement in the sensitivity and specificity of the algorithm. Researchers evaluating 500 patients with epilepsy estimated that the modified algorithm had a PPV of 82.0% (95% CI, 78.6%-85.4%).¹¹

After implementation of ICD-10 codes in the VHA in FY 2016, the task of reliably and efficiently identifying VWE led to a 3-tier algorithm. This article presents a validation of the different tiers of this algorithm after the implementation of ICD-10 diagnosis codes and summarizes the surveillance data collected over the years within the VHA showing the trends of epilepsy.

Methods

The VHA National Neurology office commissioned a Neurology Cube dashboard in FY 2021 in collaboration with VHA Support Service Center (VSSC) for reporting and surveillance of VWEs as a quality improvement initiative. The Neurology Cube uses a 3-tier system for identifying VWE in the VHA databases. VSSC programmers extract data from the VHA Corporate Data Warehouse (CDW) and utilize Microsoft SQL Server and Microsoft Power BI for Neurology Cube reports. The 3-tier system identifies VWE and divides them into distinct groups. The first tier identifies VWE with the highest degree of confidence; Tiers 2 and 3 represent identification with successively lesser degrees of confidence (Figure 1).

Tier 1

Definition. For a given index year and the preceding 2 years, any of following diagnosis codes on ≥ 1 clinical encounter are considered: 345.xx (epilepsy in ICD-9), 780.3x (other convulsions in ICD-9), G40.xxx (epilepsy in ICD-10), R40.4 (transient alteration of awareness), R56.1 (posttraumatic seizures), or R56.9 (unspecified convulsions). To reduce false positive rates, EEG clinic visits, which may include long-term monitoring, are excluded. Patients identified with ICD codes are then evaluated for an ASM prescription for ≥ 30 days during the index year. ASMs are listed in Appendix 1.

Validation. The development and validation of ICD-9 diagnosis codes crossmatched with an ASM prescription in the VHA has been published elsewhere.¹¹ In FY 2017, after implementation of ICD-10 diagnostic codes, Tier 1 development and validation was performed in 2 phases. Even though Tier 1 study phases were conducted and completed during FY 2017, the patients for Tier 1 were identified from evaluation of FY 2016 data (October 1, 2015, to September 30, 2016). After the pilot analysis, the Tier 1 definition was implemented, and a chart review of 625 randomized patients was conducted at 5 sites for validation. Adequate preliminary data was not available to perform a sample size estimation for this study. Therefore, a practical target of 125 patients was set for Tier 1 from each site to obtain a final sample size of 625 patients. This second phase validated that the crossmatch of ICD-10 diagnosis codes with ASMs had a high PPV for identifying VWE.

Tiers 2 and 3

Definitions. For an index year, Tier 2 includes patients with ≥ 1 inpatient encounter documentation of either ICD-9 345.xx or ICD-10 G40.xxx, excluding EEG clinics. Tier 3 Includes patients who have had ≥ 2 outpatient encounters with diagnosis codes 345.xx or G40.xxx on 2 separate days, excluding EEG clinics. Tiers 2 and 3 do not require ASM prescriptions; this helps to identify VWEs who may be getting their medications outside of VHA or those who have received a new diagnosis.

Validations. Tiers 2 and 3 were included in the epilepsy identification algorithm in FY 2021 after validation was performed on a sample of 8 patients in each tier. Five patients were subsequently identified as having epilepsy in Tier 2 and 6 patients were identified in Tier 3. A more comprehensive validation of Tiers 2 and 3 was performed during FY 2022 that included patients at 5 sites seen during FY 2019 to FY 2022. Since yearly trends showed only about 8% of total patients were identified as having epilepsy through Tiers 2 and 3 we sought ≥ 20 patients per tier for the 5 sites for a total of 200 patients to ensure representation across the VHA. The final count was 126 patients for Tier 2 and 174 patients for Tier 3 (n = 300).

Gold Standard Criteria for Epilepsy Diagnosis

We used the International League Against Epilepsy (ILAE) definition of epilepsy for the validation of the 3 algorithm tiers. ILAE defines epilepsy as ≥ 2 unprovoked (or reflex) seizures occurring > 24 hours apart or 1 unprovoked (or reflex) seizure and a probability of further seizures similar to the general recurrence risk (≥ 60%) after 2 unprovoked seizures, occurring over the next 10 years.¹²

A standard protocol was provided to evaluators to identify patients using the VHA Computerized Patient Record System (Appendix 1). After review, evaluators categorized each patient in 1 of 4 ways: (1) Yes, definite: The patient’s health care practitioner (HCP) believes the patient has epilepsy and is treating with medication; (2) Yes, uncertain: The HCP has enough suspicion of epilepsy that a medication is prescribed, but uncertainty is expressed of the diagnosis; (3) No, definite: The HCP does not believe the patient has epilepsy and is therefore not treating with medication for seizure; (4) No, uncertain: The HCP is not treating with medication for epilepsy, because the diagnostic suspicion is not high enough, but there is suspicion for epilepsy.

As a quality improvement operational project, the Epilepsy National Program Office approved this validation project and determined that institutional review board approval was not required.

Statistical Analysis

Counts and percentages were computed for categories of epilepsy status. PPV of each tier was estimated with asymptotic 95% CIs.

Results

ICD-10 codes for 480 patients were evaluated in Tier 1 phase 1; 13.8% were documented with G40.xxx, 27.9% with R56.1, 34.4% with R56.9, and 24.0% with R40.4 (Appendix 2). In total, 68.1% fulfilled the criteria of epilepsy, 19.2% did not, and 12.7% were uncertain). From the validation of Tier 1 phase 2 (n = 625), the PPV of the algorithm for patients presumed to have epilepsy (definite and uncertain) was 85.1% (95% CI, 82.1%-87.8%) (Table).

Of 300 patients evaluated, 126 (42.0%) were evaluated for Tier 2 with a PPV of 61.9% (95% CI, 53.4%-70.4%), and 174 (58.0%) patients were evaluated for Tier 3 with a PPV of 59.8% (95% CI, 52.5%-67.1%. The PPV of the algorithm for patients presumed to have epilepsy (definite and uncertain) were combined to calculate the PPV. Estimates of VHA VWE counts were computed for each tier from FY 2014 to FY 2023 using the VSSC Neurology Cube (Figure 2). For all years, > 92% patients were classified using the Tier 1 definition.

Discussion

The development and validation of the 3-tier diagnostic algorithm represents an important advancement in the surveillance and management of epilepsy among veterans within the VHA. The validation of this algorithm also demonstrates its practical utility in a large, integrated health care system.

Specific challenges were encountered when attempting to use pre-existing algorithms; these challenges included differences in the usage patterns of diagnostic codes and the patterns of ASM use within the VHA. These challenges prompted the need for a tailored approach, which led to the development of this algorithm. The inclusion of additional ICD-10 codes led to further revisions and subsequent validation. While many of the basic concepts of the algorithm, including ICD codes and ASMs, could work in other institutions, it would be wise for health care organizations to develop their own algorithms because of certain variables, including organizational size, patient demographics, common comorbidities, and the specific configurations of electronic health records and administrative data systems.

Studies have shown that ICD-10 codes for epilepsy (G40.* and/or R56.9) perform well in identifying epilepsy whether they are assigned by neurologists (sensitivity, 97.7%; specificity, 44.1%; PPV, 96.2%; negative predictive value, 57.7%), or in emergency department or hospital discharges (PPV, 75.5%).^13,14 The pilot study of the algorithm’s Tier 1 development (phase 1) evaluated whether the selected ICD-10 diagnostic codes accurately included the VWE population within the VHA and revealed that while most codes (eg, epilepsy [G40.xxx]; posttraumatic seizures [R56.1]; and unspecified convulsions [R56.9]), had a low false positive rate (< 16%), the R40.4 code (transient alteration of awareness) had a higher false positivity of 42%. While this is not surprising given the broad spectrum of conditions that can manifest as transient alteration of awareness, it underscores the inherent challenges in diagnosing epilepsy using diagnosis codes.

In phase 2, the Tier 1 algorithm was validated as effective for identifying VWE in the VHA system, as its PPV was determined to be high (85%). In comparison, Tiers 2 and 3, whose criteria did not require data on VHA prescribed ASM use, had lower tiers of epilepsy predictability (PPV about 60% for both). This was thought to be acceptable because Tiers 2 and 3 represent a smaller population of the identified VWEs (about 8%). These VWEs may otherwise have been missed, partly because veterans are not required to get ASMs from the VHA.

Upon VHA implementation in FY 2021, this diagnostic algorithm exhibited significant clinical utility when integrated within the VSSC Neurology Cube. It facilitated an efficient approach to identifying VWEs using readily available databases. This led to better tracking of real-time epilepsy cases, which facilitated improving current resource allocation and targeted intervention strategies such as identification of drug-resistant epilepsy patients, optimizing strategies for telehealth and patient outreach for awareness of epilepsy care resources within VHA. Meanwhile, data acquired by the algorithm over the decade since its development (FY 2014 to FY 2023) contributed to more accurate epidemiologic information and identification of historic trends. Development of the algorithm represents one of the ways ECoEs have led to improved care for VWEs. ECoEs have been shown to improve health care for veterans in several metrics.¹⁵

A strength of this study is the rigorous multitiered validation process to confirm the diagnostic accuracy of ICD-10 codes against the gold standard ILAE definition of epilepsy to identify “definite” epilepsy cases within the VHA. The use of specific ICD codes further enhances the precision of epilepsy diagnoses. The inclusion of ASMs, which are sometimes prescribed for conditions other than epilepsy, could potentially inflate false positive rates.¹⁶

This study focused exclusively on the identification and validation of definite epilepsy cases within the VHA VSSC database, employing more stringent diagnostic criteria to ensure the highest level of certainty in ascertaining epilepsy. It is important to note there is a separate category of probable epilepsy, which involves a broader set of diagnostic criteria. While not covered in this study, probable epilepsy would be subject to future research and validation, which could provide insights into a wider spectrum of epilepsy diagnoses. Such future research could help refine the algorithm’s applicability and accuracy and potentially lead to more comprehensive surveillance and management strategies in clinical practice.

This study highlights the inherent challenges in leveraging administrative data for disease identification, particularly for conditions such as epilepsy, where diagnostic clarity can be complex. However, other conditions such as multiple sclerosis have noted similar success with the use of VHA administrative data for categorizing disease.¹⁷

Limitations

The algorithm discussed in this article is, in and of itself, generalizable. However, the validation process was unique to the VHA patient population, limiting the generalizability of the findings. Documentation practices and HCP attitudes within the VHA may differ from those in other health care settings. Identifying people with epilepsy can be challenging because of changing definitions of epilepsy over time. In addition to clinical evaluation, EEG and magnetic resonance imaging results, response to ASM treatment, and video-EEG monitoring of habitual events all can help establish the diagnosis. Therefore, studies may vary in how inclusive or exclusive the criteria are. ASMs such as gabapentin, pregabalin, carbamazepine, lamotrigine, topiramate, and valproate are used to treat other conditions, including headaches, generalized pain, and mood disorders. Consequently, including these ASMs in the Tier 1 definition may have increased the false positive rate. Additional research is needed to evaluate whether excluding these ASMs from the algorithm based on specific criteria (eg, dose of ASM used) can further refine the algorithm to identify patients with epilepsy.

Further refinement of this algorithm may also occur as technology changes. Future electronic health records may allow better tracking of different epilepsy factors, the integration of additional diagnostic criteria, and the use of natural language processing or other forms of artificial intelligence.

Conclusions

This study presents a significant step forward in epilepsy surveillance within the VHA. The algorithm offers a robust tool for identifying VWEs with good PPVs, facilitating better resource allocation and targeted care. Despite its limitations, this research lays a foundation for future advancements in the management and understanding of epilepsy within large health care systems. Since this VHA algorithm is based on ASMs and ICD diagnosis codes from patient records, other large managed health care systems also may be able to adapt this algorithm to their data specifications.

Epilepsy affects about 4.5 million people in the United States and 150,000 new individuals are diagnosed each year.^1,2 In 2019, epilepsy-attributable health care spending for noninstitutionalized people was around $5.4 billion and total epilepsy-attributable and epilepsy or seizure health care-related costs totaled $54 billion.³

Accurate surveillance of epilepsy in large health care systems can potentially improve health care delivery and resource allocation. A 2012 Institute of Medicine (IOM) report identified 13 recommendations to guide public health action on epilepsy, including validation of standard definitions for case ascertainment, identification of epilepsy through screening programs or protocols, and expansion of surveillance to better understand disease burden.⁴

A systematic review of validation studies concluded that it is reasonable to use administrative data to identify people with epilepsy in epidemiologic research. Combining The International Classification of Diseases (ICD) codes for epilepsy (ICD-10, G40-41; ICD-9, 345) with antiseizure medications (ASMs) could provide high positive predictive values (PPVs) and combining symptoms codes for convulsions (ICD-10, R56; ICD-9, 780.3, 780.39) with ASMs could lead to high sensitivity.⁵ However, identifying individuals with epilepsy from administrative data in large managed health care organizations is challenging.⁶ The IOM report noted that large managed health care organizations presented varying incidence and prevalence estimates due to differing methodology, geographic area, demographics, and definitions of epilepsy.

The Veterans Health Administration (VHA) is the largest integrated US health care system, providing care to > 9.1 million veterans.⁷ To improve the health and well-being of veterans with epilepsy (VWEs), a network of sites was established in 2008 called the US Department of Veterans Affairs (VA) Epilepsy Centers of Excellence (ECoE). Subsequent to the creation of the ECoE, efforts were made to identify VWEs within VHA databases.^8,9 Prior to fiscal year (FY) 2016, the ECoE adopted a modified version of a well-established epilepsy diagnostic algorithm developed by Holden et al for large managed care organizations.¹⁰ The original algorithm identified patients by cross-matching ASMs with ICD-9 codes for an index year. But it failed to capture a considerable number of stable patients with epilepsy in the VHA due to incomplete documentation, and had false positives due to inclusion of patients identified from diagnostic clinics. The modified algorithm the ECoE used prior to FY 2016 considered additional prior years and excluded encounters from diagnostic clinics. The result was an improvement in the sensitivity and specificity of the algorithm. Researchers evaluating 500 patients with epilepsy estimated that the modified algorithm had a PPV of 82.0% (95% CI, 78.6%-85.4%).¹¹

After implementation of ICD-10 codes in the VHA in FY 2016, the task of reliably and efficiently identifying VWE led to a 3-tier algorithm. This article presents a validation of the different tiers of this algorithm after the implementation of ICD-10 diagnosis codes and summarizes the surveillance data collected over the years within the VHA showing the trends of epilepsy.

Methods

The VHA National Neurology office commissioned a Neurology Cube dashboard in FY 2021 in collaboration with VHA Support Service Center (VSSC) for reporting and surveillance of VWEs as a quality improvement initiative. The Neurology Cube uses a 3-tier system for identifying VWE in the VHA databases. VSSC programmers extract data from the VHA Corporate Data Warehouse (CDW) and utilize Microsoft SQL Server and Microsoft Power BI for Neurology Cube reports. The 3-tier system identifies VWE and divides them into distinct groups. The first tier identifies VWE with the highest degree of confidence; Tiers 2 and 3 represent identification with successively lesser degrees of confidence (Figure 1).

Tier 1

Definition. For a given index year and the preceding 2 years, any of following diagnosis codes on ≥ 1 clinical encounter are considered: 345.xx (epilepsy in ICD-9), 780.3x (other convulsions in ICD-9), G40.xxx (epilepsy in ICD-10), R40.4 (transient alteration of awareness), R56.1 (posttraumatic seizures), or R56.9 (unspecified convulsions). To reduce false positive rates, EEG clinic visits, which may include long-term monitoring, are excluded. Patients identified with ICD codes are then evaluated for an ASM prescription for ≥ 30 days during the index year. ASMs are listed in Appendix 1.

Validation. The development and validation of ICD-9 diagnosis codes crossmatched with an ASM prescription in the VHA has been published elsewhere.¹¹ In FY 2017, after implementation of ICD-10 diagnostic codes, Tier 1 development and validation was performed in 2 phases. Even though Tier 1 study phases were conducted and completed during FY 2017, the patients for Tier 1 were identified from evaluation of FY 2016 data (October 1, 2015, to September 30, 2016). After the pilot analysis, the Tier 1 definition was implemented, and a chart review of 625 randomized patients was conducted at 5 sites for validation. Adequate preliminary data was not available to perform a sample size estimation for this study. Therefore, a practical target of 125 patients was set for Tier 1 from each site to obtain a final sample size of 625 patients. This second phase validated that the crossmatch of ICD-10 diagnosis codes with ASMs had a high PPV for identifying VWE.

Tiers 2 and 3

Definitions. For an index year, Tier 2 includes patients with ≥ 1 inpatient encounter documentation of either ICD-9 345.xx or ICD-10 G40.xxx, excluding EEG clinics. Tier 3 Includes patients who have had ≥ 2 outpatient encounters with diagnosis codes 345.xx or G40.xxx on 2 separate days, excluding EEG clinics. Tiers 2 and 3 do not require ASM prescriptions; this helps to identify VWEs who may be getting their medications outside of VHA or those who have received a new diagnosis.

Validations. Tiers 2 and 3 were included in the epilepsy identification algorithm in FY 2021 after validation was performed on a sample of 8 patients in each tier. Five patients were subsequently identified as having epilepsy in Tier 2 and 6 patients were identified in Tier 3. A more comprehensive validation of Tiers 2 and 3 was performed during FY 2022 that included patients at 5 sites seen during FY 2019 to FY 2022. Since yearly trends showed only about 8% of total patients were identified as having epilepsy through Tiers 2 and 3 we sought ≥ 20 patients per tier for the 5 sites for a total of 200 patients to ensure representation across the VHA. The final count was 126 patients for Tier 2 and 174 patients for Tier 3 (n = 300).

Gold Standard Criteria for Epilepsy Diagnosis

We used the International League Against Epilepsy (ILAE) definition of epilepsy for the validation of the 3 algorithm tiers. ILAE defines epilepsy as ≥ 2 unprovoked (or reflex) seizures occurring > 24 hours apart or 1 unprovoked (or reflex) seizure and a probability of further seizures similar to the general recurrence risk (≥ 60%) after 2 unprovoked seizures, occurring over the next 10 years.¹²

A standard protocol was provided to evaluators to identify patients using the VHA Computerized Patient Record System (Appendix 1). After review, evaluators categorized each patient in 1 of 4 ways: (1) Yes, definite: The patient’s health care practitioner (HCP) believes the patient has epilepsy and is treating with medication; (2) Yes, uncertain: The HCP has enough suspicion of epilepsy that a medication is prescribed, but uncertainty is expressed of the diagnosis; (3) No, definite: The HCP does not believe the patient has epilepsy and is therefore not treating with medication for seizure; (4) No, uncertain: The HCP is not treating with medication for epilepsy, because the diagnostic suspicion is not high enough, but there is suspicion for epilepsy.

As a quality improvement operational project, the Epilepsy National Program Office approved this validation project and determined that institutional review board approval was not required.

Statistical Analysis

Counts and percentages were computed for categories of epilepsy status. PPV of each tier was estimated with asymptotic 95% CIs.

Results

ICD-10 codes for 480 patients were evaluated in Tier 1 phase 1; 13.8% were documented with G40.xxx, 27.9% with R56.1, 34.4% with R56.9, and 24.0% with R40.4 (Appendix 2). In total, 68.1% fulfilled the criteria of epilepsy, 19.2% did not, and 12.7% were uncertain). From the validation of Tier 1 phase 2 (n = 625), the PPV of the algorithm for patients presumed to have epilepsy (definite and uncertain) was 85.1% (95% CI, 82.1%-87.8%) (Table).

Of 300 patients evaluated, 126 (42.0%) were evaluated for Tier 2 with a PPV of 61.9% (95% CI, 53.4%-70.4%), and 174 (58.0%) patients were evaluated for Tier 3 with a PPV of 59.8% (95% CI, 52.5%-67.1%. The PPV of the algorithm for patients presumed to have epilepsy (definite and uncertain) were combined to calculate the PPV. Estimates of VHA VWE counts were computed for each tier from FY 2014 to FY 2023 using the VSSC Neurology Cube (Figure 2). For all years, > 92% patients were classified using the Tier 1 definition.

Discussion

The development and validation of the 3-tier diagnostic algorithm represents an important advancement in the surveillance and management of epilepsy among veterans within the VHA. The validation of this algorithm also demonstrates its practical utility in a large, integrated health care system.

Specific challenges were encountered when attempting to use pre-existing algorithms; these challenges included differences in the usage patterns of diagnostic codes and the patterns of ASM use within the VHA. These challenges prompted the need for a tailored approach, which led to the development of this algorithm. The inclusion of additional ICD-10 codes led to further revisions and subsequent validation. While many of the basic concepts of the algorithm, including ICD codes and ASMs, could work in other institutions, it would be wise for health care organizations to develop their own algorithms because of certain variables, including organizational size, patient demographics, common comorbidities, and the specific configurations of electronic health records and administrative data systems.

Studies have shown that ICD-10 codes for epilepsy (G40.* and/or R56.9) perform well in identifying epilepsy whether they are assigned by neurologists (sensitivity, 97.7%; specificity, 44.1%; PPV, 96.2%; negative predictive value, 57.7%), or in emergency department or hospital discharges (PPV, 75.5%).^13,14 The pilot study of the algorithm’s Tier 1 development (phase 1) evaluated whether the selected ICD-10 diagnostic codes accurately included the VWE population within the VHA and revealed that while most codes (eg, epilepsy [G40.xxx]; posttraumatic seizures [R56.1]; and unspecified convulsions [R56.9]), had a low false positive rate (< 16%), the R40.4 code (transient alteration of awareness) had a higher false positivity of 42%. While this is not surprising given the broad spectrum of conditions that can manifest as transient alteration of awareness, it underscores the inherent challenges in diagnosing epilepsy using diagnosis codes.

In phase 2, the Tier 1 algorithm was validated as effective for identifying VWE in the VHA system, as its PPV was determined to be high (85%). In comparison, Tiers 2 and 3, whose criteria did not require data on VHA prescribed ASM use, had lower tiers of epilepsy predictability (PPV about 60% for both). This was thought to be acceptable because Tiers 2 and 3 represent a smaller population of the identified VWEs (about 8%). These VWEs may otherwise have been missed, partly because veterans are not required to get ASMs from the VHA.

Upon VHA implementation in FY 2021, this diagnostic algorithm exhibited significant clinical utility when integrated within the VSSC Neurology Cube. It facilitated an efficient approach to identifying VWEs using readily available databases. This led to better tracking of real-time epilepsy cases, which facilitated improving current resource allocation and targeted intervention strategies such as identification of drug-resistant epilepsy patients, optimizing strategies for telehealth and patient outreach for awareness of epilepsy care resources within VHA. Meanwhile, data acquired by the algorithm over the decade since its development (FY 2014 to FY 2023) contributed to more accurate epidemiologic information and identification of historic trends. Development of the algorithm represents one of the ways ECoEs have led to improved care for VWEs. ECoEs have been shown to improve health care for veterans in several metrics.¹⁵

A strength of this study is the rigorous multitiered validation process to confirm the diagnostic accuracy of ICD-10 codes against the gold standard ILAE definition of epilepsy to identify “definite” epilepsy cases within the VHA. The use of specific ICD codes further enhances the precision of epilepsy diagnoses. The inclusion of ASMs, which are sometimes prescribed for conditions other than epilepsy, could potentially inflate false positive rates.¹⁶

This study focused exclusively on the identification and validation of definite epilepsy cases within the VHA VSSC database, employing more stringent diagnostic criteria to ensure the highest level of certainty in ascertaining epilepsy. It is important to note there is a separate category of probable epilepsy, which involves a broader set of diagnostic criteria. While not covered in this study, probable epilepsy would be subject to future research and validation, which could provide insights into a wider spectrum of epilepsy diagnoses. Such future research could help refine the algorithm’s applicability and accuracy and potentially lead to more comprehensive surveillance and management strategies in clinical practice.

This study highlights the inherent challenges in leveraging administrative data for disease identification, particularly for conditions such as epilepsy, where diagnostic clarity can be complex. However, other conditions such as multiple sclerosis have noted similar success with the use of VHA administrative data for categorizing disease.¹⁷

Limitations

The algorithm discussed in this article is, in and of itself, generalizable. However, the validation process was unique to the VHA patient population, limiting the generalizability of the findings. Documentation practices and HCP attitudes within the VHA may differ from those in other health care settings. Identifying people with epilepsy can be challenging because of changing definitions of epilepsy over time. In addition to clinical evaluation, EEG and magnetic resonance imaging results, response to ASM treatment, and video-EEG monitoring of habitual events all can help establish the diagnosis. Therefore, studies may vary in how inclusive or exclusive the criteria are. ASMs such as gabapentin, pregabalin, carbamazepine, lamotrigine, topiramate, and valproate are used to treat other conditions, including headaches, generalized pain, and mood disorders. Consequently, including these ASMs in the Tier 1 definition may have increased the false positive rate. Additional research is needed to evaluate whether excluding these ASMs from the algorithm based on specific criteria (eg, dose of ASM used) can further refine the algorithm to identify patients with epilepsy.

Further refinement of this algorithm may also occur as technology changes. Future electronic health records may allow better tracking of different epilepsy factors, the integration of additional diagnostic criteria, and the use of natural language processing or other forms of artificial intelligence.

Conclusions

This study presents a significant step forward in epilepsy surveillance within the VHA. The algorithm offers a robust tool for identifying VWEs with good PPVs, facilitating better resource allocation and targeted care. Despite its limitations, this research lays a foundation for future advancements in the management and understanding of epilepsy within large health care systems. Since this VHA algorithm is based on ASMs and ICD diagnosis codes from patient records, other large managed health care systems also may be able to adapt this algorithm to their data specifications.

Epilepsy affects about 4.5 million people in the United States and 150,000 new individuals are diagnosed each year.^1,2 In 2019, epilepsy-attributable health care spending for noninstitutionalized people was around $5.4 billion and total epilepsy-attributable and epilepsy or seizure health care-related costs totaled $54 billion.³

Accurate surveillance of epilepsy in large health care systems can potentially improve health care delivery and resource allocation. A 2012 Institute of Medicine (IOM) report identified 13 recommendations to guide public health action on epilepsy, including validation of standard definitions for case ascertainment, identification of epilepsy through screening programs or protocols, and expansion of surveillance to better understand disease burden.⁴

A systematic review of validation studies concluded that it is reasonable to use administrative data to identify people with epilepsy in epidemiologic research. Combining The International Classification of Diseases (ICD) codes for epilepsy (ICD-10, G40-41; ICD-9, 345) with antiseizure medications (ASMs) could provide high positive predictive values (PPVs) and combining symptoms codes for convulsions (ICD-10, R56; ICD-9, 780.3, 780.39) with ASMs could lead to high sensitivity.⁵ However, identifying individuals with epilepsy from administrative data in large managed health care organizations is challenging.⁶ The IOM report noted that large managed health care organizations presented varying incidence and prevalence estimates due to differing methodology, geographic area, demographics, and definitions of epilepsy.

The Veterans Health Administration (VHA) is the largest integrated US health care system, providing care to > 9.1 million veterans.⁷ To improve the health and well-being of veterans with epilepsy (VWEs), a network of sites was established in 2008 called the US Department of Veterans Affairs (VA) Epilepsy Centers of Excellence (ECoE). Subsequent to the creation of the ECoE, efforts were made to identify VWEs within VHA databases.^8,9 Prior to fiscal year (FY) 2016, the ECoE adopted a modified version of a well-established epilepsy diagnostic algorithm developed by Holden et al for large managed care organizations.¹⁰ The original algorithm identified patients by cross-matching ASMs with ICD-9 codes for an index year. But it failed to capture a considerable number of stable patients with epilepsy in the VHA due to incomplete documentation, and had false positives due to inclusion of patients identified from diagnostic clinics. The modified algorithm the ECoE used prior to FY 2016 considered additional prior years and excluded encounters from diagnostic clinics. The result was an improvement in the sensitivity and specificity of the algorithm. Researchers evaluating 500 patients with epilepsy estimated that the modified algorithm had a PPV of 82.0% (95% CI, 78.6%-85.4%).¹¹

After implementation of ICD-10 codes in the VHA in FY 2016, the task of reliably and efficiently identifying VWE led to a 3-tier algorithm. This article presents a validation of the different tiers of this algorithm after the implementation of ICD-10 diagnosis codes and summarizes the surveillance data collected over the years within the VHA showing the trends of epilepsy.

Methods

The VHA National Neurology office commissioned a Neurology Cube dashboard in FY 2021 in collaboration with VHA Support Service Center (VSSC) for reporting and surveillance of VWEs as a quality improvement initiative. The Neurology Cube uses a 3-tier system for identifying VWE in the VHA databases. VSSC programmers extract data from the VHA Corporate Data Warehouse (CDW) and utilize Microsoft SQL Server and Microsoft Power BI for Neurology Cube reports. The 3-tier system identifies VWE and divides them into distinct groups. The first tier identifies VWE with the highest degree of confidence; Tiers 2 and 3 represent identification with successively lesser degrees of confidence (Figure 1).

Tier 1

Definition. For a given index year and the preceding 2 years, any of following diagnosis codes on ≥ 1 clinical encounter are considered: 345.xx (epilepsy in ICD-9), 780.3x (other convulsions in ICD-9), G40.xxx (epilepsy in ICD-10), R40.4 (transient alteration of awareness), R56.1 (posttraumatic seizures), or R56.9 (unspecified convulsions). To reduce false positive rates, EEG clinic visits, which may include long-term monitoring, are excluded. Patients identified with ICD codes are then evaluated for an ASM prescription for ≥ 30 days during the index year. ASMs are listed in Appendix 1.

Validation. The development and validation of ICD-9 diagnosis codes crossmatched with an ASM prescription in the VHA has been published elsewhere.¹¹ In FY 2017, after implementation of ICD-10 diagnostic codes, Tier 1 development and validation was performed in 2 phases. Even though Tier 1 study phases were conducted and completed during FY 2017, the patients for Tier 1 were identified from evaluation of FY 2016 data (October 1, 2015, to September 30, 2016). After the pilot analysis, the Tier 1 definition was implemented, and a chart review of 625 randomized patients was conducted at 5 sites for validation. Adequate preliminary data was not available to perform a sample size estimation for this study. Therefore, a practical target of 125 patients was set for Tier 1 from each site to obtain a final sample size of 625 patients. This second phase validated that the crossmatch of ICD-10 diagnosis codes with ASMs had a high PPV for identifying VWE.

Tiers 2 and 3

Definitions. For an index year, Tier 2 includes patients with ≥ 1 inpatient encounter documentation of either ICD-9 345.xx or ICD-10 G40.xxx, excluding EEG clinics. Tier 3 Includes patients who have had ≥ 2 outpatient encounters with diagnosis codes 345.xx or G40.xxx on 2 separate days, excluding EEG clinics. Tiers 2 and 3 do not require ASM prescriptions; this helps to identify VWEs who may be getting their medications outside of VHA or those who have received a new diagnosis.

Validations. Tiers 2 and 3 were included in the epilepsy identification algorithm in FY 2021 after validation was performed on a sample of 8 patients in each tier. Five patients were subsequently identified as having epilepsy in Tier 2 and 6 patients were identified in Tier 3. A more comprehensive validation of Tiers 2 and 3 was performed during FY 2022 that included patients at 5 sites seen during FY 2019 to FY 2022. Since yearly trends showed only about 8% of total patients were identified as having epilepsy through Tiers 2 and 3 we sought ≥ 20 patients per tier for the 5 sites for a total of 200 patients to ensure representation across the VHA. The final count was 126 patients for Tier 2 and 174 patients for Tier 3 (n = 300).

Gold Standard Criteria for Epilepsy Diagnosis

We used the International League Against Epilepsy (ILAE) definition of epilepsy for the validation of the 3 algorithm tiers. ILAE defines epilepsy as ≥ 2 unprovoked (or reflex) seizures occurring > 24 hours apart or 1 unprovoked (or reflex) seizure and a probability of further seizures similar to the general recurrence risk (≥ 60%) after 2 unprovoked seizures, occurring over the next 10 years.¹²

A standard protocol was provided to evaluators to identify patients using the VHA Computerized Patient Record System (Appendix 1). After review, evaluators categorized each patient in 1 of 4 ways: (1) Yes, definite: The patient’s health care practitioner (HCP) believes the patient has epilepsy and is treating with medication; (2) Yes, uncertain: The HCP has enough suspicion of epilepsy that a medication is prescribed, but uncertainty is expressed of the diagnosis; (3) No, definite: The HCP does not believe the patient has epilepsy and is therefore not treating with medication for seizure; (4) No, uncertain: The HCP is not treating with medication for epilepsy, because the diagnostic suspicion is not high enough, but there is suspicion for epilepsy.

As a quality improvement operational project, the Epilepsy National Program Office approved this validation project and determined that institutional review board approval was not required.

Statistical Analysis

Counts and percentages were computed for categories of epilepsy status. PPV of each tier was estimated with asymptotic 95% CIs.

Results

ICD-10 codes for 480 patients were evaluated in Tier 1 phase 1; 13.8% were documented with G40.xxx, 27.9% with R56.1, 34.4% with R56.9, and 24.0% with R40.4 (Appendix 2). In total, 68.1% fulfilled the criteria of epilepsy, 19.2% did not, and 12.7% were uncertain). From the validation of Tier 1 phase 2 (n = 625), the PPV of the algorithm for patients presumed to have epilepsy (definite and uncertain) was 85.1% (95% CI, 82.1%-87.8%) (Table).

Of 300 patients evaluated, 126 (42.0%) were evaluated for Tier 2 with a PPV of 61.9% (95% CI, 53.4%-70.4%), and 174 (58.0%) patients were evaluated for Tier 3 with a PPV of 59.8% (95% CI, 52.5%-67.1%. The PPV of the algorithm for patients presumed to have epilepsy (definite and uncertain) were combined to calculate the PPV. Estimates of VHA VWE counts were computed for each tier from FY 2014 to FY 2023 using the VSSC Neurology Cube (Figure 2). For all years, > 92% patients were classified using the Tier 1 definition.

Discussion

The development and validation of the 3-tier diagnostic algorithm represents an important advancement in the surveillance and management of epilepsy among veterans within the VHA. The validation of this algorithm also demonstrates its practical utility in a large, integrated health care system.

Specific challenges were encountered when attempting to use pre-existing algorithms; these challenges included differences in the usage patterns of diagnostic codes and the patterns of ASM use within the VHA. These challenges prompted the need for a tailored approach, which led to the development of this algorithm. The inclusion of additional ICD-10 codes led to further revisions and subsequent validation. While many of the basic concepts of the algorithm, including ICD codes and ASMs, could work in other institutions, it would be wise for health care organizations to develop their own algorithms because of certain variables, including organizational size, patient demographics, common comorbidities, and the specific configurations of electronic health records and administrative data systems.

Studies have shown that ICD-10 codes for epilepsy (G40.* and/or R56.9) perform well in identifying epilepsy whether they are assigned by neurologists (sensitivity, 97.7%; specificity, 44.1%; PPV, 96.2%; negative predictive value, 57.7%), or in emergency department or hospital discharges (PPV, 75.5%).^13,14 The pilot study of the algorithm’s Tier 1 development (phase 1) evaluated whether the selected ICD-10 diagnostic codes accurately included the VWE population within the VHA and revealed that while most codes (eg, epilepsy [G40.xxx]; posttraumatic seizures [R56.1]; and unspecified convulsions [R56.9]), had a low false positive rate (< 16%), the R40.4 code (transient alteration of awareness) had a higher false positivity of 42%. While this is not surprising given the broad spectrum of conditions that can manifest as transient alteration of awareness, it underscores the inherent challenges in diagnosing epilepsy using diagnosis codes.

In phase 2, the Tier 1 algorithm was validated as effective for identifying VWE in the VHA system, as its PPV was determined to be high (85%). In comparison, Tiers 2 and 3, whose criteria did not require data on VHA prescribed ASM use, had lower tiers of epilepsy predictability (PPV about 60% for both). This was thought to be acceptable because Tiers 2 and 3 represent a smaller population of the identified VWEs (about 8%). These VWEs may otherwise have been missed, partly because veterans are not required to get ASMs from the VHA.

Upon VHA implementation in FY 2021, this diagnostic algorithm exhibited significant clinical utility when integrated within the VSSC Neurology Cube. It facilitated an efficient approach to identifying VWEs using readily available databases. This led to better tracking of real-time epilepsy cases, which facilitated improving current resource allocation and targeted intervention strategies such as identification of drug-resistant epilepsy patients, optimizing strategies for telehealth and patient outreach for awareness of epilepsy care resources within VHA. Meanwhile, data acquired by the algorithm over the decade since its development (FY 2014 to FY 2023) contributed to more accurate epidemiologic information and identification of historic trends. Development of the algorithm represents one of the ways ECoEs have led to improved care for VWEs. ECoEs have been shown to improve health care for veterans in several metrics.¹⁵

A strength of this study is the rigorous multitiered validation process to confirm the diagnostic accuracy of ICD-10 codes against the gold standard ILAE definition of epilepsy to identify “definite” epilepsy cases within the VHA. The use of specific ICD codes further enhances the precision of epilepsy diagnoses. The inclusion of ASMs, which are sometimes prescribed for conditions other than epilepsy, could potentially inflate false positive rates.¹⁶

This study focused exclusively on the identification and validation of definite epilepsy cases within the VHA VSSC database, employing more stringent diagnostic criteria to ensure the highest level of certainty in ascertaining epilepsy. It is important to note there is a separate category of probable epilepsy, which involves a broader set of diagnostic criteria. While not covered in this study, probable epilepsy would be subject to future research and validation, which could provide insights into a wider spectrum of epilepsy diagnoses. Such future research could help refine the algorithm’s applicability and accuracy and potentially lead to more comprehensive surveillance and management strategies in clinical practice.

This study highlights the inherent challenges in leveraging administrative data for disease identification, particularly for conditions such as epilepsy, where diagnostic clarity can be complex. However, other conditions such as multiple sclerosis have noted similar success with the use of VHA administrative data for categorizing disease.¹⁷

Limitations

The algorithm discussed in this article is, in and of itself, generalizable. However, the validation process was unique to the VHA patient population, limiting the generalizability of the findings. Documentation practices and HCP attitudes within the VHA may differ from those in other health care settings. Identifying people with epilepsy can be challenging because of changing definitions of epilepsy over time. In addition to clinical evaluation, EEG and magnetic resonance imaging results, response to ASM treatment, and video-EEG monitoring of habitual events all can help establish the diagnosis. Therefore, studies may vary in how inclusive or exclusive the criteria are. ASMs such as gabapentin, pregabalin, carbamazepine, lamotrigine, topiramate, and valproate are used to treat other conditions, including headaches, generalized pain, and mood disorders. Consequently, including these ASMs in the Tier 1 definition may have increased the false positive rate. Additional research is needed to evaluate whether excluding these ASMs from the algorithm based on specific criteria (eg, dose of ASM used) can further refine the algorithm to identify patients with epilepsy.

Further refinement of this algorithm may also occur as technology changes. Future electronic health records may allow better tracking of different epilepsy factors, the integration of additional diagnostic criteria, and the use of natural language processing or other forms of artificial intelligence.

Conclusions

This study presents a significant step forward in epilepsy surveillance within the VHA. The algorithm offers a robust tool for identifying VWEs with good PPVs, facilitating better resource allocation and targeted care. Despite its limitations, this research lays a foundation for future advancements in the management and understanding of epilepsy within large health care systems. Since this VHA algorithm is based on ASMs and ICD diagnosis codes from patient records, other large managed health care systems also may be able to adapt this algorithm to their data specifications.

Total hip arthroplasty (THA) is among the most common elective orthopedic procedures performed annually in the United States, with an estimated 635,000 to 909,000 THAs expected each year by 2030.¹ Consequently, complication rates and revision surgeries related to THA have been increasing, along with the financial burden on the health care system.^2-4 Optimizing outcomes for patients undergoing THA and identifying risk factors for treatment failure have become areas of focus.

Over the last decade, there has been a renewed interest in the effect of previous lumbar spine fusion (LSF) surgery on THA outcomes. Studies have explored the rates of complications, postoperative mobility, and THA implant impingement.^5-8 However, the outcome receiving the most attention in recent literature is the rate and effect of dislocation in patients with lumbar fusion surgery. Large Medicare database analyses have discovered an association with increased rates of dislocations in patients with lumbar fusion surgeries compared with those without.^9,10 Prosthetic hip dislocation is an expensive complication of THA and is projected to have greater impact through 2035 due to a growing number of THA procedures.¹¹ Identifying risk factors associated with hip dislocation is paramount to mitigating its effect on patients who have undergone THA.

Recent research has found increased rates of THA dislocation and revision surgery in patients with LSF, with some studies showing previous LSF as the strongest independent predictor.^6-16 However, controversy surrounds this relationship, including the sequence of procedures (LSF before or after THA), the time between procedures, and involvement of the sacrum in LSF. One study found that patients had a 106% increased risk of dislocation when LSF was performed before THA compared with patients who underwent LSF 5 years after undergoing THA, while another study showed no significant difference in dislocations pre- vs post-LSF.^16,17 An additional study showed no significant difference in the rate of dislocation in patients without sacral involvement in the LSF, while also showing significantly higher rates of dislocation in LSF with sacral involvement.¹² The researchers also found a trend toward more dislocations in longer lumbosacral fusions. Recent studies have also examined dislocation rates with lumbar fusion in patients treated with dual-mobility liners.^18-20 The consensus from these studies is that dual-mobility liners significantly decrease the rate of dislocation in primary THAs with lumbar fusion.

The present study sought to determine the rates of hip dislocations in a US Department of Veterans Affairs (VA) hospital setting. To the authors’ knowledge, no retrospective study focusing on THAs in the veteran population has been performed. This study benefits from controlling for various surgeon techniques and surgical preferences when compared to large Medicare database studies because the orthopedic surgeon (ABK) only performed the posterior approach for all patients during the study period.

The primary objective of this study was to determine whether the rates of hip dislocation would, in fact, be higher in patients with lumbar fusion surgery, as recent database studies suggest. Secondary objectives included determining whether patient characteristics, comorbidities, number of levels fused, or inclusion of the sacrum in the fusion construct influenced dislocation rates. Furthermore, VA Dayton Healthcare System (VADHS) began routine use of dual-mobility liners for lumbar fusion patients in 2018, allowing for examination of these patients.

Methods

The Wright State University and VADHS Institutional Review Board approved this study design. A retrospective review of all primary THAs at VADHS was performed to investigate the relationship between previous lumbar spine fusion and the incidence of THA revision. Manual chart review was performed for patients who underwent primary THA between January 2003, and December 2022. One surgeon performed all surgeries using only the posterior approach. Patients were not excluded if they had bilateral procedures and all eligible hips were included. Patients with a concomitant diagnosis of fracture of the femoral head or femoral neck at the time of surgery were excluded. Additionally, only patients with ≥ 12 months of follow-up data were included.

The primary outcome was dislocation within 12 months of THA; the primary independent variable was LSF prior to THA. Covariates included patient demographics (age, sex, body mass index [BMI]) and Charlson Comorbidity Index (CCI) score, with additional data collected on the number of levels fused, sacral spine involvement, revision rates, and use of dual-mobility liners. Year of surgery was also included in analyses to account for any changes that may have occurred during the study period.

Statistical Analysis

Statistical analyses were performed in SAS 9.4. Patients were grouped into 2 cohorts, depending on whether they had received LSF prior to THA. Analyses were adjusted for repeated measures to account for the small percentage of patients with bilateral procedures.

Univariate comparisons between cohorts for covariates, as well as rates of dislocation and revision, were performed using the independent samples t test for continuous variables and the Fisher exact test for dichotomous categorical variables. Significant comorbidities, as well as age, sex, BMI, liner type, LSF cohort, and surgery year, were included in a logistic regression model to determine what effect, if any, they had on the likelihood of dislocation. Variables were removed using a backward stepwise approach, starting with the nonsignificant variable effect with the lowest χ² value, and continuing until reaching a final model where all remaining variable effects were significant. For the variables retained in the final model, odds ratios (ORs) with 95% CIs were derived, with dislocation designated as the event. Individual comorbidity subcomponents of the CCI were also analyzed for their effects on dislocation using backward stepwise logistic regression. A secondary analysis among patients with LSF tested for the influence of the number of vertebral levels fused, the presence or absence of sacral involvement in the fusion, and the use of dual-mobility liners on the likelihood of hip dislocation.

Results

The LSF cohort included 39 patients with THA and prior LSF, 3 of whom had bilateral procedures, for a total of 42 hips. The non-LSF cohort included 813 patients with THA, 112 of whom had bilateral procedures, for a total of 925 hips. The LSF and non-LSF cohorts did not differ significantly in age, sex, BMI, CCI, or revision rates (Table). The LSF cohort included a significantly higher percentage of hips receiving dual-mobility liners than did the non-LSF cohort (23.8% vs 0.6%; P < .001) and had more than twice the rate of dislocation (4 of 42 hips [9.5%] vs 35 of 925 hips [3.8%]), although this difference was not statistically significant (P = .08).

The final logistic regression model with dislocation as the outcome was statistically significant (χ², 17.47; P < .001) and retained 2 significant predictor variables: LSF cohort (χ², 4.63; P = .03), and sex (χ², 18.27; P < .001). Females were more likely than males to experience dislocation (OR, 5.84; 95% CI, 2.60-13.13; P < .001) as were patients who had LSF prior to THA (OR, 3.42; 95% CI, 1.12-10.47; P = .03) (Figure). None of the CCI subcomponent comorbidities significantly affected the probability of dislocation (myocardial infarction, P = .46; congestive heart failure, P = .47; peripheral vascular disease, P = .97; stroke, P = .51; dementia, P = .99; chronic obstructive pulmonary disease, P = .95; connective tissue disease, P = .25; peptic ulcer, P = .41; liver disease, P = .30; diabetes, P = .06; hemiplegia, P = .99; chronic kidney disease, P = .82; solid tumor, P = .90; leukemia, P = .99; lymphoma, P = .99; AIDS, P = .99). Within the LSF cohort, neither the number of levels fused (P = .83) nor sacral involvement (P = .42), significantly affected the probability of hip dislocation. None of the patients in either cohort who received dual-mobility liners subsequently dislocated their hips, nor did any of them require revision surgery.

Discussion

Spinopelvic biomechanics have been an area of increasing interest and research. Spinal fusion has been shown to alter the mobility of the pelvis and has been associated with decreased stability of THA implants.²¹ For example, in the setting of a fused spine, the lack of compensatory changes in pelvic tilt or acetabular anteversion when adjusting to a seated or standing position may predispose patients to impingement because the acetabular component is not properly positioned. Dual-mobility constructs mitigate this risk by providing an additional articulation, which increases jump distance and range of motion prior to impingement, thereby enhancing stability.

The use of dual-mobility liners in patients with LSF has also been examined.^18-20 These studies demonstrate a reduced risk of postoperative THA dislocation in patients with previous LSF. The rate of postoperative complications and revisions for LSF patients with dual-mobility liners was also found to be similar to that of THAs without dual-mobility in patients without prior LSF. This study focused on a veteran population to demonstrate the efficacy of dual-mobility liners in patients with LSF. The results indicate that LSF prior to THA and female sex were predictors for prosthetic hip dislocations in the 12-month postoperative period in this patient population, which aligns with the current literature.

The dislocation rate in the LSF-THA group (9.5%) was higher than the dislocation rate in the control group (3.8%). Although not statistically significant in the univariate analysis, LSF was shown to be a significant risk factor after controlling for patient sex. Other studies have found the dislocation rate to be 3% to 7%, which is lower than the dislocation rate observed in this study.^8,10,16

The reasons for this higher rate of dislocation are not entirely clear. A veteran population has poorer overall health than the general population, which may contribute to the higher than previously reported dislocation rates.²² These results can be applied to the management of veterans seeking THA.

There have been conflicting reports regarding the impact a patient’s sex has on THA outcomes in the general population.^23-26 This study found that female patients had higher rates of dislocation within 1 year of THA than male patients. This difference, which could be due to differences in baseline anatomic hip morphology between the sexes; females tend to have smaller femoral head sizes and less offset compared with males.^27,28 However, this finding could have been confounded by the small number of female veterans in the study cohort.

A type 2 diabetes mellitus (T2DM) diagnosis, which is a component of CCI, trended toward increased risk of prosthetic hip dislocation. Multiple studies have also discussed the increased risk of postoperative infections and revisions following THA in patients with T2DM.^29-31 One study found T2DM to be an independent risk factor for immediate in-hospital postoperative complications following hip arthroplasty.³²

Another factor that may influence postoperative dislocation risk is surgical approach. The posterior approach has historically been associated with higher rates of instability when compared to anterior or lateral THA.³³ Researchers have also looked at the role that surgical approach plays in patients with prior LSF. Huebschmann et al confirmed that not only is LSF a significant risk factor for dislocation following THA, but anterior and laterally based surgical approaches may mitigate this risk.³⁴

Limitations

As a retrospective cohort study, the reliability of the data hinges on complete documentation. Documentation of all encounters for dislocations was obtained from the VA Computerized Patient Record System, which may have led to some dislocation events being missed. However, as long as there was adequate postoperative follow-up, it was assumed all events outside the VA were included. Another limitation of this study was that male patients greatly outnumbered female patients, and this fact could limit the generalizability of findings to the population as a whole.

Conclusions

This study in a veteran population found that prior LSF and female sex were significant predictors for postoperative dislocation within 1 year of THA surgery. Additionally, the use of a dual-mobility liner was found to be protective against postoperative dislocation events. These data allow clinicians to better counsel veterans on the risk factors associated with postoperative dislocation and strategies to mitigate this risk.

Total hip arthroplasty (THA) is among the most common elective orthopedic procedures performed annually in the United States, with an estimated 635,000 to 909,000 THAs expected each year by 2030.¹ Consequently, complication rates and revision surgeries related to THA have been increasing, along with the financial burden on the health care system.^2-4 Optimizing outcomes for patients undergoing THA and identifying risk factors for treatment failure have become areas of focus.

Over the last decade, there has been a renewed interest in the effect of previous lumbar spine fusion (LSF) surgery on THA outcomes. Studies have explored the rates of complications, postoperative mobility, and THA implant impingement.^5-8 However, the outcome receiving the most attention in recent literature is the rate and effect of dislocation in patients with lumbar fusion surgery. Large Medicare database analyses have discovered an association with increased rates of dislocations in patients with lumbar fusion surgeries compared with those without.^9,10 Prosthetic hip dislocation is an expensive complication of THA and is projected to have greater impact through 2035 due to a growing number of THA procedures.¹¹ Identifying risk factors associated with hip dislocation is paramount to mitigating its effect on patients who have undergone THA.

Recent research has found increased rates of THA dislocation and revision surgery in patients with LSF, with some studies showing previous LSF as the strongest independent predictor.^6-16 However, controversy surrounds this relationship, including the sequence of procedures (LSF before or after THA), the time between procedures, and involvement of the sacrum in LSF. One study found that patients had a 106% increased risk of dislocation when LSF was performed before THA compared with patients who underwent LSF 5 years after undergoing THA, while another study showed no significant difference in dislocations pre- vs post-LSF.^16,17 An additional study showed no significant difference in the rate of dislocation in patients without sacral involvement in the LSF, while also showing significantly higher rates of dislocation in LSF with sacral involvement.¹² The researchers also found a trend toward more dislocations in longer lumbosacral fusions. Recent studies have also examined dislocation rates with lumbar fusion in patients treated with dual-mobility liners.^18-20 The consensus from these studies is that dual-mobility liners significantly decrease the rate of dislocation in primary THAs with lumbar fusion.

The present study sought to determine the rates of hip dislocations in a US Department of Veterans Affairs (VA) hospital setting. To the authors’ knowledge, no retrospective study focusing on THAs in the veteran population has been performed. This study benefits from controlling for various surgeon techniques and surgical preferences when compared to large Medicare database studies because the orthopedic surgeon (ABK) only performed the posterior approach for all patients during the study period.

The primary objective of this study was to determine whether the rates of hip dislocation would, in fact, be higher in patients with lumbar fusion surgery, as recent database studies suggest. Secondary objectives included determining whether patient characteristics, comorbidities, number of levels fused, or inclusion of the sacrum in the fusion construct influenced dislocation rates. Furthermore, VA Dayton Healthcare System (VADHS) began routine use of dual-mobility liners for lumbar fusion patients in 2018, allowing for examination of these patients.

Methods

The Wright State University and VADHS Institutional Review Board approved this study design. A retrospective review of all primary THAs at VADHS was performed to investigate the relationship between previous lumbar spine fusion and the incidence of THA revision. Manual chart review was performed for patients who underwent primary THA between January 2003, and December 2022. One surgeon performed all surgeries using only the posterior approach. Patients were not excluded if they had bilateral procedures and all eligible hips were included. Patients with a concomitant diagnosis of fracture of the femoral head or femoral neck at the time of surgery were excluded. Additionally, only patients with ≥ 12 months of follow-up data were included.

The primary outcome was dislocation within 12 months of THA; the primary independent variable was LSF prior to THA. Covariates included patient demographics (age, sex, body mass index [BMI]) and Charlson Comorbidity Index (CCI) score, with additional data collected on the number of levels fused, sacral spine involvement, revision rates, and use of dual-mobility liners. Year of surgery was also included in analyses to account for any changes that may have occurred during the study period.

Statistical Analysis

Statistical analyses were performed in SAS 9.4. Patients were grouped into 2 cohorts, depending on whether they had received LSF prior to THA. Analyses were adjusted for repeated measures to account for the small percentage of patients with bilateral procedures.

Univariate comparisons between cohorts for covariates, as well as rates of dislocation and revision, were performed using the independent samples t test for continuous variables and the Fisher exact test for dichotomous categorical variables. Significant comorbidities, as well as age, sex, BMI, liner type, LSF cohort, and surgery year, were included in a logistic regression model to determine what effect, if any, they had on the likelihood of dislocation. Variables were removed using a backward stepwise approach, starting with the nonsignificant variable effect with the lowest χ² value, and continuing until reaching a final model where all remaining variable effects were significant. For the variables retained in the final model, odds ratios (ORs) with 95% CIs were derived, with dislocation designated as the event. Individual comorbidity subcomponents of the CCI were also analyzed for their effects on dislocation using backward stepwise logistic regression. A secondary analysis among patients with LSF tested for the influence of the number of vertebral levels fused, the presence or absence of sacral involvement in the fusion, and the use of dual-mobility liners on the likelihood of hip dislocation.

Results

The LSF cohort included 39 patients with THA and prior LSF, 3 of whom had bilateral procedures, for a total of 42 hips. The non-LSF cohort included 813 patients with THA, 112 of whom had bilateral procedures, for a total of 925 hips. The LSF and non-LSF cohorts did not differ significantly in age, sex, BMI, CCI, or revision rates (Table). The LSF cohort included a significantly higher percentage of hips receiving dual-mobility liners than did the non-LSF cohort (23.8% vs 0.6%; P < .001) and had more than twice the rate of dislocation (4 of 42 hips [9.5%] vs 35 of 925 hips [3.8%]), although this difference was not statistically significant (P = .08).

The final logistic regression model with dislocation as the outcome was statistically significant (χ², 17.47; P < .001) and retained 2 significant predictor variables: LSF cohort (χ², 4.63; P = .03), and sex (χ², 18.27; P < .001). Females were more likely than males to experience dislocation (OR, 5.84; 95% CI, 2.60-13.13; P < .001) as were patients who had LSF prior to THA (OR, 3.42; 95% CI, 1.12-10.47; P = .03) (Figure). None of the CCI subcomponent comorbidities significantly affected the probability of dislocation (myocardial infarction, P = .46; congestive heart failure, P = .47; peripheral vascular disease, P = .97; stroke, P = .51; dementia, P = .99; chronic obstructive pulmonary disease, P = .95; connective tissue disease, P = .25; peptic ulcer, P = .41; liver disease, P = .30; diabetes, P = .06; hemiplegia, P = .99; chronic kidney disease, P = .82; solid tumor, P = .90; leukemia, P = .99; lymphoma, P = .99; AIDS, P = .99). Within the LSF cohort, neither the number of levels fused (P = .83) nor sacral involvement (P = .42), significantly affected the probability of hip dislocation. None of the patients in either cohort who received dual-mobility liners subsequently dislocated their hips, nor did any of them require revision surgery.

Discussion

Spinopelvic biomechanics have been an area of increasing interest and research. Spinal fusion has been shown to alter the mobility of the pelvis and has been associated with decreased stability of THA implants.²¹ For example, in the setting of a fused spine, the lack of compensatory changes in pelvic tilt or acetabular anteversion when adjusting to a seated or standing position may predispose patients to impingement because the acetabular component is not properly positioned. Dual-mobility constructs mitigate this risk by providing an additional articulation, which increases jump distance and range of motion prior to impingement, thereby enhancing stability.

The use of dual-mobility liners in patients with LSF has also been examined.^18-20 These studies demonstrate a reduced risk of postoperative THA dislocation in patients with previous LSF. The rate of postoperative complications and revisions for LSF patients with dual-mobility liners was also found to be similar to that of THAs without dual-mobility in patients without prior LSF. This study focused on a veteran population to demonstrate the efficacy of dual-mobility liners in patients with LSF. The results indicate that LSF prior to THA and female sex were predictors for prosthetic hip dislocations in the 12-month postoperative period in this patient population, which aligns with the current literature.

The dislocation rate in the LSF-THA group (9.5%) was higher than the dislocation rate in the control group (3.8%). Although not statistically significant in the univariate analysis, LSF was shown to be a significant risk factor after controlling for patient sex. Other studies have found the dislocation rate to be 3% to 7%, which is lower than the dislocation rate observed in this study.^8,10,16

The reasons for this higher rate of dislocation are not entirely clear. A veteran population has poorer overall health than the general population, which may contribute to the higher than previously reported dislocation rates.²² These results can be applied to the management of veterans seeking THA.

There have been conflicting reports regarding the impact a patient’s sex has on THA outcomes in the general population.^23-26 This study found that female patients had higher rates of dislocation within 1 year of THA than male patients. This difference, which could be due to differences in baseline anatomic hip morphology between the sexes; females tend to have smaller femoral head sizes and less offset compared with males.^27,28 However, this finding could have been confounded by the small number of female veterans in the study cohort.

A type 2 diabetes mellitus (T2DM) diagnosis, which is a component of CCI, trended toward increased risk of prosthetic hip dislocation. Multiple studies have also discussed the increased risk of postoperative infections and revisions following THA in patients with T2DM.^29-31 One study found T2DM to be an independent risk factor for immediate in-hospital postoperative complications following hip arthroplasty.³²

Another factor that may influence postoperative dislocation risk is surgical approach. The posterior approach has historically been associated with higher rates of instability when compared to anterior or lateral THA.³³ Researchers have also looked at the role that surgical approach plays in patients with prior LSF. Huebschmann et al confirmed that not only is LSF a significant risk factor for dislocation following THA, but anterior and laterally based surgical approaches may mitigate this risk.³⁴

Limitations

As a retrospective cohort study, the reliability of the data hinges on complete documentation. Documentation of all encounters for dislocations was obtained from the VA Computerized Patient Record System, which may have led to some dislocation events being missed. However, as long as there was adequate postoperative follow-up, it was assumed all events outside the VA were included. Another limitation of this study was that male patients greatly outnumbered female patients, and this fact could limit the generalizability of findings to the population as a whole.

Conclusions

This study in a veteran population found that prior LSF and female sex were significant predictors for postoperative dislocation within 1 year of THA surgery. Additionally, the use of a dual-mobility liner was found to be protective against postoperative dislocation events. These data allow clinicians to better counsel veterans on the risk factors associated with postoperative dislocation and strategies to mitigate this risk.

Total hip arthroplasty (THA) is among the most common elective orthopedic procedures performed annually in the United States, with an estimated 635,000 to 909,000 THAs expected each year by 2030.¹ Consequently, complication rates and revision surgeries related to THA have been increasing, along with the financial burden on the health care system.^2-4 Optimizing outcomes for patients undergoing THA and identifying risk factors for treatment failure have become areas of focus.

Over the last decade, there has been a renewed interest in the effect of previous lumbar spine fusion (LSF) surgery on THA outcomes. Studies have explored the rates of complications, postoperative mobility, and THA implant impingement.^5-8 However, the outcome receiving the most attention in recent literature is the rate and effect of dislocation in patients with lumbar fusion surgery. Large Medicare database analyses have discovered an association with increased rates of dislocations in patients with lumbar fusion surgeries compared with those without.^9,10 Prosthetic hip dislocation is an expensive complication of THA and is projected to have greater impact through 2035 due to a growing number of THA procedures.¹¹ Identifying risk factors associated with hip dislocation is paramount to mitigating its effect on patients who have undergone THA.

Recent research has found increased rates of THA dislocation and revision surgery in patients with LSF, with some studies showing previous LSF as the strongest independent predictor.^6-16 However, controversy surrounds this relationship, including the sequence of procedures (LSF before or after THA), the time between procedures, and involvement of the sacrum in LSF. One study found that patients had a 106% increased risk of dislocation when LSF was performed before THA compared with patients who underwent LSF 5 years after undergoing THA, while another study showed no significant difference in dislocations pre- vs post-LSF.^16,17 An additional study showed no significant difference in the rate of dislocation in patients without sacral involvement in the LSF, while also showing significantly higher rates of dislocation in LSF with sacral involvement.¹² The researchers also found a trend toward more dislocations in longer lumbosacral fusions. Recent studies have also examined dislocation rates with lumbar fusion in patients treated with dual-mobility liners.^18-20 The consensus from these studies is that dual-mobility liners significantly decrease the rate of dislocation in primary THAs with lumbar fusion.

The present study sought to determine the rates of hip dislocations in a US Department of Veterans Affairs (VA) hospital setting. To the authors’ knowledge, no retrospective study focusing on THAs in the veteran population has been performed. This study benefits from controlling for various surgeon techniques and surgical preferences when compared to large Medicare database studies because the orthopedic surgeon (ABK) only performed the posterior approach for all patients during the study period.

The primary objective of this study was to determine whether the rates of hip dislocation would, in fact, be higher in patients with lumbar fusion surgery, as recent database studies suggest. Secondary objectives included determining whether patient characteristics, comorbidities, number of levels fused, or inclusion of the sacrum in the fusion construct influenced dislocation rates. Furthermore, VA Dayton Healthcare System (VADHS) began routine use of dual-mobility liners for lumbar fusion patients in 2018, allowing for examination of these patients.

Methods

The Wright State University and VADHS Institutional Review Board approved this study design. A retrospective review of all primary THAs at VADHS was performed to investigate the relationship between previous lumbar spine fusion and the incidence of THA revision. Manual chart review was performed for patients who underwent primary THA between January 2003, and December 2022. One surgeon performed all surgeries using only the posterior approach. Patients were not excluded if they had bilateral procedures and all eligible hips were included. Patients with a concomitant diagnosis of fracture of the femoral head or femoral neck at the time of surgery were excluded. Additionally, only patients with ≥ 12 months of follow-up data were included.

The primary outcome was dislocation within 12 months of THA; the primary independent variable was LSF prior to THA. Covariates included patient demographics (age, sex, body mass index [BMI]) and Charlson Comorbidity Index (CCI) score, with additional data collected on the number of levels fused, sacral spine involvement, revision rates, and use of dual-mobility liners. Year of surgery was also included in analyses to account for any changes that may have occurred during the study period.

Statistical Analysis

Statistical analyses were performed in SAS 9.4. Patients were grouped into 2 cohorts, depending on whether they had received LSF prior to THA. Analyses were adjusted for repeated measures to account for the small percentage of patients with bilateral procedures.

Univariate comparisons between cohorts for covariates, as well as rates of dislocation and revision, were performed using the independent samples t test for continuous variables and the Fisher exact test for dichotomous categorical variables. Significant comorbidities, as well as age, sex, BMI, liner type, LSF cohort, and surgery year, were included in a logistic regression model to determine what effect, if any, they had on the likelihood of dislocation. Variables were removed using a backward stepwise approach, starting with the nonsignificant variable effect with the lowest χ² value, and continuing until reaching a final model where all remaining variable effects were significant. For the variables retained in the final model, odds ratios (ORs) with 95% CIs were derived, with dislocation designated as the event. Individual comorbidity subcomponents of the CCI were also analyzed for their effects on dislocation using backward stepwise logistic regression. A secondary analysis among patients with LSF tested for the influence of the number of vertebral levels fused, the presence or absence of sacral involvement in the fusion, and the use of dual-mobility liners on the likelihood of hip dislocation.

Results

The LSF cohort included 39 patients with THA and prior LSF, 3 of whom had bilateral procedures, for a total of 42 hips. The non-LSF cohort included 813 patients with THA, 112 of whom had bilateral procedures, for a total of 925 hips. The LSF and non-LSF cohorts did not differ significantly in age, sex, BMI, CCI, or revision rates (Table). The LSF cohort included a significantly higher percentage of hips receiving dual-mobility liners than did the non-LSF cohort (23.8% vs 0.6%; P < .001) and had more than twice the rate of dislocation (4 of 42 hips [9.5%] vs 35 of 925 hips [3.8%]), although this difference was not statistically significant (P = .08).

The final logistic regression model with dislocation as the outcome was statistically significant (χ², 17.47; P < .001) and retained 2 significant predictor variables: LSF cohort (χ², 4.63; P = .03), and sex (χ², 18.27; P < .001). Females were more likely than males to experience dislocation (OR, 5.84; 95% CI, 2.60-13.13; P < .001) as were patients who had LSF prior to THA (OR, 3.42; 95% CI, 1.12-10.47; P = .03) (Figure). None of the CCI subcomponent comorbidities significantly affected the probability of dislocation (myocardial infarction, P = .46; congestive heart failure, P = .47; peripheral vascular disease, P = .97; stroke, P = .51; dementia, P = .99; chronic obstructive pulmonary disease, P = .95; connective tissue disease, P = .25; peptic ulcer, P = .41; liver disease, P = .30; diabetes, P = .06; hemiplegia, P = .99; chronic kidney disease, P = .82; solid tumor, P = .90; leukemia, P = .99; lymphoma, P = .99; AIDS, P = .99). Within the LSF cohort, neither the number of levels fused (P = .83) nor sacral involvement (P = .42), significantly affected the probability of hip dislocation. None of the patients in either cohort who received dual-mobility liners subsequently dislocated their hips, nor did any of them require revision surgery.

Discussion

Spinopelvic biomechanics have been an area of increasing interest and research. Spinal fusion has been shown to alter the mobility of the pelvis and has been associated with decreased stability of THA implants.²¹ For example, in the setting of a fused spine, the lack of compensatory changes in pelvic tilt or acetabular anteversion when adjusting to a seated or standing position may predispose patients to impingement because the acetabular component is not properly positioned. Dual-mobility constructs mitigate this risk by providing an additional articulation, which increases jump distance and range of motion prior to impingement, thereby enhancing stability.

The use of dual-mobility liners in patients with LSF has also been examined.^18-20 These studies demonstrate a reduced risk of postoperative THA dislocation in patients with previous LSF. The rate of postoperative complications and revisions for LSF patients with dual-mobility liners was also found to be similar to that of THAs without dual-mobility in patients without prior LSF. This study focused on a veteran population to demonstrate the efficacy of dual-mobility liners in patients with LSF. The results indicate that LSF prior to THA and female sex were predictors for prosthetic hip dislocations in the 12-month postoperative period in this patient population, which aligns with the current literature.

The dislocation rate in the LSF-THA group (9.5%) was higher than the dislocation rate in the control group (3.8%). Although not statistically significant in the univariate analysis, LSF was shown to be a significant risk factor after controlling for patient sex. Other studies have found the dislocation rate to be 3% to 7%, which is lower than the dislocation rate observed in this study.^8,10,16

The reasons for this higher rate of dislocation are not entirely clear. A veteran population has poorer overall health than the general population, which may contribute to the higher than previously reported dislocation rates.²² These results can be applied to the management of veterans seeking THA.

There have been conflicting reports regarding the impact a patient’s sex has on THA outcomes in the general population.^23-26 This study found that female patients had higher rates of dislocation within 1 year of THA than male patients. This difference, which could be due to differences in baseline anatomic hip morphology between the sexes; females tend to have smaller femoral head sizes and less offset compared with males.^27,28 However, this finding could have been confounded by the small number of female veterans in the study cohort.

A type 2 diabetes mellitus (T2DM) diagnosis, which is a component of CCI, trended toward increased risk of prosthetic hip dislocation. Multiple studies have also discussed the increased risk of postoperative infections and revisions following THA in patients with T2DM.^29-31 One study found T2DM to be an independent risk factor for immediate in-hospital postoperative complications following hip arthroplasty.³²

Another factor that may influence postoperative dislocation risk is surgical approach. The posterior approach has historically been associated with higher rates of instability when compared to anterior or lateral THA.³³ Researchers have also looked at the role that surgical approach plays in patients with prior LSF. Huebschmann et al confirmed that not only is LSF a significant risk factor for dislocation following THA, but anterior and laterally based surgical approaches may mitigate this risk.³⁴

Limitations

As a retrospective cohort study, the reliability of the data hinges on complete documentation. Documentation of all encounters for dislocations was obtained from the VA Computerized Patient Record System, which may have led to some dislocation events being missed. However, as long as there was adequate postoperative follow-up, it was assumed all events outside the VA were included. Another limitation of this study was that male patients greatly outnumbered female patients, and this fact could limit the generalizability of findings to the population as a whole.

Conclusions

This study in a veteran population found that prior LSF and female sex were significant predictors for postoperative dislocation within 1 year of THA surgery. Additionally, the use of a dual-mobility liner was found to be protective against postoperative dislocation events. These data allow clinicians to better counsel veterans on the risk factors associated with postoperative dislocation and strategies to mitigate this risk.