User login
Risk Score Personalizes CRC Screening for Veterans
Risk Score Personalizes CRC Screening for Veterans
TOPLINE:
A recalibrated environmental risk score for colorectal cancer (CRC) shows improved predictive performance in a study of 227,504 male veterans. The veteran-tailored score could help personalize screening better than previous models.
METHODOLOGY:
- Demographic, lifestyle, and CRC data from 2011 to 2022 were abstracted from survey responses and health records of 227,504 male Million Veteran Program (MVP) participants, with complete data needed to construct the environmental risk score (e-Score).
- Researchers randomly split the male sample into 2 halves to produce training and validation samples (each n = 113,752; CRC cases n = 590) using simple random sampling with strata based on the CRC variable.
- Weighting for each environmental factor's effect size was recalculated using US Department of Veterans Affairs training data to create a recalibrated e-Score, which was compared with the original weighted e-Score in the validation sample.
- Analysis included nested multiple logistic regression models testing associations between quintiles for recalibrated and original e-Scores, with likelihood ratio tests used to compare model performance.
- Factors used to construct the e-Score included BMI, height, diabetes diagnosis, aspirin use, nonsteroidal anti-inflammatory drug use, educational attainment, physical activity level, smoking status, alcohol use, and dietary intake of fiber, calcium, folate, processed meats, red meat, fruits, vegetables, and total energy.
TAKEAWAY:
- The recalibrated e-Score showed a significant association with CRC, with higher quintiles indicating increased risk.
- In the validation sample, the recalibrated e-Score model significantly improved the base model performance (P < .001), while the original GECCO e-Score model did not show significant improvement (P = .07).
- The recalibrated e-Score model quintile 5 was associated with significantly higher odds for CRC compared with quintile 1 (odds ratio [OR], 1.79; 95% CI, 1.38-2.33; P for trend < .001).
- Black participants had higher odds for CRC compared with the White reference group across all models (base model OR, 1.46; 95% CI, 1.13-1.92; GECCO e-Score model OR, 1.44; 95% CI, 1.09-1.88; and recalibrated e-Score model OR, 1.38; 95% CI, 1.05-1.82).
IN PRACTICE:
"Despite the robust methods used in the work by the GECCO study upon which our study was based, an e-Score using their study’s weighting was not significantly associated with colorectal cancer among the male veteran sample. However, data from nearly a quarter million (n = 227,504) male US veteran participants of the MVP were used to recalibrate the e-Score to be veteran specific, and the recalibrated e-Score validation showed that it was significantly associated with colorectal cancer," wrote the authors of the study.
SOURCE:
The study was led by April R. Williams, US Department of Veterans Affairs Million Veteran Program Coordinating Center in Boston. It was published online in Cancer Epidemiology, Biomarkers & Prevention.
LIMITATIONS:
The study's limitations include potential recall and self-selection bias due to the use of self-reported data from the MVP. The generalizability of the findings may be limited to the veteran population, and the sample of Black veterans may have been insufficient for conclusive analysis. Additionally, the study did not include female participants due to insufficient data for a veteran-specific e-Score.
DISCLOSURES:
B.A. Sullivan disclosed receiving grants from the American Gastroenterological Association. D. Lieberman reported support from Geneoscopy, UDX, and ColoWrap. Additional disclosures are noted in the original article.
This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.
TOPLINE:
A recalibrated environmental risk score for colorectal cancer (CRC) shows improved predictive performance in a study of 227,504 male veterans. The veteran-tailored score could help personalize screening better than previous models.
METHODOLOGY:
- Demographic, lifestyle, and CRC data from 2011 to 2022 were abstracted from survey responses and health records of 227,504 male Million Veteran Program (MVP) participants, with complete data needed to construct the environmental risk score (e-Score).
- Researchers randomly split the male sample into 2 halves to produce training and validation samples (each n = 113,752; CRC cases n = 590) using simple random sampling with strata based on the CRC variable.
- Weighting for each environmental factor's effect size was recalculated using US Department of Veterans Affairs training data to create a recalibrated e-Score, which was compared with the original weighted e-Score in the validation sample.
- Analysis included nested multiple logistic regression models testing associations between quintiles for recalibrated and original e-Scores, with likelihood ratio tests used to compare model performance.
- Factors used to construct the e-Score included BMI, height, diabetes diagnosis, aspirin use, nonsteroidal anti-inflammatory drug use, educational attainment, physical activity level, smoking status, alcohol use, and dietary intake of fiber, calcium, folate, processed meats, red meat, fruits, vegetables, and total energy.
TAKEAWAY:
- The recalibrated e-Score showed a significant association with CRC, with higher quintiles indicating increased risk.
- In the validation sample, the recalibrated e-Score model significantly improved the base model performance (P < .001), while the original GECCO e-Score model did not show significant improvement (P = .07).
- The recalibrated e-Score model quintile 5 was associated with significantly higher odds for CRC compared with quintile 1 (odds ratio [OR], 1.79; 95% CI, 1.38-2.33; P for trend < .001).
- Black participants had higher odds for CRC compared with the White reference group across all models (base model OR, 1.46; 95% CI, 1.13-1.92; GECCO e-Score model OR, 1.44; 95% CI, 1.09-1.88; and recalibrated e-Score model OR, 1.38; 95% CI, 1.05-1.82).
IN PRACTICE:
"Despite the robust methods used in the work by the GECCO study upon which our study was based, an e-Score using their study’s weighting was not significantly associated with colorectal cancer among the male veteran sample. However, data from nearly a quarter million (n = 227,504) male US veteran participants of the MVP were used to recalibrate the e-Score to be veteran specific, and the recalibrated e-Score validation showed that it was significantly associated with colorectal cancer," wrote the authors of the study.
SOURCE:
The study was led by April R. Williams, US Department of Veterans Affairs Million Veteran Program Coordinating Center in Boston. It was published online in Cancer Epidemiology, Biomarkers & Prevention.
LIMITATIONS:
The study's limitations include potential recall and self-selection bias due to the use of self-reported data from the MVP. The generalizability of the findings may be limited to the veteran population, and the sample of Black veterans may have been insufficient for conclusive analysis. Additionally, the study did not include female participants due to insufficient data for a veteran-specific e-Score.
DISCLOSURES:
B.A. Sullivan disclosed receiving grants from the American Gastroenterological Association. D. Lieberman reported support from Geneoscopy, UDX, and ColoWrap. Additional disclosures are noted in the original article.
This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.
TOPLINE:
A recalibrated environmental risk score for colorectal cancer (CRC) shows improved predictive performance in a study of 227,504 male veterans. The veteran-tailored score could help personalize screening better than previous models.
METHODOLOGY:
- Demographic, lifestyle, and CRC data from 2011 to 2022 were abstracted from survey responses and health records of 227,504 male Million Veteran Program (MVP) participants, with complete data needed to construct the environmental risk score (e-Score).
- Researchers randomly split the male sample into 2 halves to produce training and validation samples (each n = 113,752; CRC cases n = 590) using simple random sampling with strata based on the CRC variable.
- Weighting for each environmental factor's effect size was recalculated using US Department of Veterans Affairs training data to create a recalibrated e-Score, which was compared with the original weighted e-Score in the validation sample.
- Analysis included nested multiple logistic regression models testing associations between quintiles for recalibrated and original e-Scores, with likelihood ratio tests used to compare model performance.
- Factors used to construct the e-Score included BMI, height, diabetes diagnosis, aspirin use, nonsteroidal anti-inflammatory drug use, educational attainment, physical activity level, smoking status, alcohol use, and dietary intake of fiber, calcium, folate, processed meats, red meat, fruits, vegetables, and total energy.
TAKEAWAY:
- The recalibrated e-Score showed a significant association with CRC, with higher quintiles indicating increased risk.
- In the validation sample, the recalibrated e-Score model significantly improved the base model performance (P < .001), while the original GECCO e-Score model did not show significant improvement (P = .07).
- The recalibrated e-Score model quintile 5 was associated with significantly higher odds for CRC compared with quintile 1 (odds ratio [OR], 1.79; 95% CI, 1.38-2.33; P for trend < .001).
- Black participants had higher odds for CRC compared with the White reference group across all models (base model OR, 1.46; 95% CI, 1.13-1.92; GECCO e-Score model OR, 1.44; 95% CI, 1.09-1.88; and recalibrated e-Score model OR, 1.38; 95% CI, 1.05-1.82).
IN PRACTICE:
"Despite the robust methods used in the work by the GECCO study upon which our study was based, an e-Score using their study’s weighting was not significantly associated with colorectal cancer among the male veteran sample. However, data from nearly a quarter million (n = 227,504) male US veteran participants of the MVP were used to recalibrate the e-Score to be veteran specific, and the recalibrated e-Score validation showed that it was significantly associated with colorectal cancer," wrote the authors of the study.
SOURCE:
The study was led by April R. Williams, US Department of Veterans Affairs Million Veteran Program Coordinating Center in Boston. It was published online in Cancer Epidemiology, Biomarkers & Prevention.
LIMITATIONS:
The study's limitations include potential recall and self-selection bias due to the use of self-reported data from the MVP. The generalizability of the findings may be limited to the veteran population, and the sample of Black veterans may have been insufficient for conclusive analysis. Additionally, the study did not include female participants due to insufficient data for a veteran-specific e-Score.
DISCLOSURES:
B.A. Sullivan disclosed receiving grants from the American Gastroenterological Association. D. Lieberman reported support from Geneoscopy, UDX, and ColoWrap. Additional disclosures are noted in the original article.
This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.
Risk Score Personalizes CRC Screening for Veterans
Risk Score Personalizes CRC Screening for Veterans