Reuters Health

NEW YORK  - Health care professionals should be highly suspicious of megacolon in Clostridium difficile-infected patients and have a low threshold for transferring infected patients to intensive care units, Veterans Affairs researchers warn.

"The incidence of Clostridium difficile-associated megacolon has nearly tripled and mortality has nearly doubled over the past decade," Dr. SreyRam Kuy from the Overton Brooks VA Medical Center in Shreveport, Louisiana, said by email.

"It could be argued that this increased incidence may be due in part to improvements in detection of Clostridium difficile," she said. "However, this increase over the past decade . . . correlates with prior work by the Agency for Healthcare Research and Quality, which showed a 74% increase in the overall number of hospital discharges with Clostridium difficile infections from 1993-2001, in the decade prior to our study."

Dr. Kuy and colleagues analyzed records in the Nationwide Inpatient Sample (2000-2010) and identified patients with both C. difficile infection and megacolon.

They identified 28,219 cases of C. difficile infection, or 0.38% of all hospitalized patients, in 2000. That grew to more than 68,600 cases, or 0.88% of hospitalized patients, in 2010.

While the overall incidence of megacolon remained steady at 0.02% of hospitalized patients from 2000 to 2010, the rate of megacolon cases tied to C. diff infection increased from 3.61% in 2000 to 9.39% in 2010 (p<0.05).

"Compared with patients with megacolon but without C. difficile infection, patients with C. difficile-associated megacolon are significantly older, are more likely to have an urgent or emergent admission, are more likely to be admitted from the emergency department or transferred from another hospital, and are more likely to be treated at large, urban, teaching hospitals," the researchers write in an article online Oct. 7 in JAMA Surgery.

They report that the mean length of hospital stay for patients with C. difficile-associated megacolon was 16.13 days, the mean cost of hospitalization came to $41,968, and 50.7% required transitional care after hospital discharge. The mortality among these patients went up from 13.56% in 2000 to 24.45% in 2010, and peaked at 30.03% in 2007 (p<0.05).

"The rise in mortality is also alarming," Dr. Kuy said. "Potentially, this rise in mortality could be attributed to changes in virulence of Clostridium difficile strains, though we are unable to determine this due to limitations of the dataset utilized for this research. Other factors that can affect mortality are antibiotic regimen, severity of disease, physical exam findings, immunosuppression, presence of end organ failure, patient frailty status, APACHE (Acute Physiology and Chronic Health Evaluation) score, signs of sepsis, and leukocytosis, which have been identified as factors associated with mortality in prior published reports."

"This study draws attention to the tremendous burden of Clostridium difficile on patient mortality, health care costs and resources, transitional care utilization, and the extremely important need for aggressive prevention of this iatrogenic disease," she concluded.

The authors reported no funding or disclosures.

NEW YORK  - Health care professionals should be highly suspicious of megacolon in Clostridium difficile-infected patients and have a low threshold for transferring infected patients to intensive care units, Veterans Affairs researchers warn.

"The incidence of Clostridium difficile-associated megacolon has nearly tripled and mortality has nearly doubled over the past decade," Dr. SreyRam Kuy from the Overton Brooks VA Medical Center in Shreveport, Louisiana, said by email.

"It could be argued that this increased incidence may be due in part to improvements in detection of Clostridium difficile," she said. "However, this increase over the past decade . . . correlates with prior work by the Agency for Healthcare Research and Quality, which showed a 74% increase in the overall number of hospital discharges with Clostridium difficile infections from 1993-2001, in the decade prior to our study."

Dr. Kuy and colleagues analyzed records in the Nationwide Inpatient Sample (2000-2010) and identified patients with both C. difficile infection and megacolon.

They identified 28,219 cases of C. difficile infection, or 0.38% of all hospitalized patients, in 2000. That grew to more than 68,600 cases, or 0.88% of hospitalized patients, in 2010.

While the overall incidence of megacolon remained steady at 0.02% of hospitalized patients from 2000 to 2010, the rate of megacolon cases tied to C. diff infection increased from 3.61% in 2000 to 9.39% in 2010 (p<0.05).

"Compared with patients with megacolon but without C. difficile infection, patients with C. difficile-associated megacolon are significantly older, are more likely to have an urgent or emergent admission, are more likely to be admitted from the emergency department or transferred from another hospital, and are more likely to be treated at large, urban, teaching hospitals," the researchers write in an article online Oct. 7 in JAMA Surgery.

They report that the mean length of hospital stay for patients with C. difficile-associated megacolon was 16.13 days, the mean cost of hospitalization came to $41,968, and 50.7% required transitional care after hospital discharge. The mortality among these patients went up from 13.56% in 2000 to 24.45% in 2010, and peaked at 30.03% in 2007 (p<0.05).

"The rise in mortality is also alarming," Dr. Kuy said. "Potentially, this rise in mortality could be attributed to changes in virulence of Clostridium difficile strains, though we are unable to determine this due to limitations of the dataset utilized for this research. Other factors that can affect mortality are antibiotic regimen, severity of disease, physical exam findings, immunosuppression, presence of end organ failure, patient frailty status, APACHE (Acute Physiology and Chronic Health Evaluation) score, signs of sepsis, and leukocytosis, which have been identified as factors associated with mortality in prior published reports."

"This study draws attention to the tremendous burden of Clostridium difficile on patient mortality, health care costs and resources, transitional care utilization, and the extremely important need for aggressive prevention of this iatrogenic disease," she concluded.

The authors reported no funding or disclosures.

NEW YORK  - Health care professionals should be highly suspicious of megacolon in Clostridium difficile-infected patients and have a low threshold for transferring infected patients to intensive care units, Veterans Affairs researchers warn.

"The incidence of Clostridium difficile-associated megacolon has nearly tripled and mortality has nearly doubled over the past decade," Dr. SreyRam Kuy from the Overton Brooks VA Medical Center in Shreveport, Louisiana, said by email.

"It could be argued that this increased incidence may be due in part to improvements in detection of Clostridium difficile," she said. "However, this increase over the past decade . . . correlates with prior work by the Agency for Healthcare Research and Quality, which showed a 74% increase in the overall number of hospital discharges with Clostridium difficile infections from 1993-2001, in the decade prior to our study."

Dr. Kuy and colleagues analyzed records in the Nationwide Inpatient Sample (2000-2010) and identified patients with both C. difficile infection and megacolon.

They identified 28,219 cases of C. difficile infection, or 0.38% of all hospitalized patients, in 2000. That grew to more than 68,600 cases, or 0.88% of hospitalized patients, in 2010.

While the overall incidence of megacolon remained steady at 0.02% of hospitalized patients from 2000 to 2010, the rate of megacolon cases tied to C. diff infection increased from 3.61% in 2000 to 9.39% in 2010 (p<0.05).

"Compared with patients with megacolon but without C. difficile infection, patients with C. difficile-associated megacolon are significantly older, are more likely to have an urgent or emergent admission, are more likely to be admitted from the emergency department or transferred from another hospital, and are more likely to be treated at large, urban, teaching hospitals," the researchers write in an article online Oct. 7 in JAMA Surgery.

They report that the mean length of hospital stay for patients with C. difficile-associated megacolon was 16.13 days, the mean cost of hospitalization came to $41,968, and 50.7% required transitional care after hospital discharge. The mortality among these patients went up from 13.56% in 2000 to 24.45% in 2010, and peaked at 30.03% in 2007 (p<0.05).

"The rise in mortality is also alarming," Dr. Kuy said. "Potentially, this rise in mortality could be attributed to changes in virulence of Clostridium difficile strains, though we are unable to determine this due to limitations of the dataset utilized for this research. Other factors that can affect mortality are antibiotic regimen, severity of disease, physical exam findings, immunosuppression, presence of end organ failure, patient frailty status, APACHE (Acute Physiology and Chronic Health Evaluation) score, signs of sepsis, and leukocytosis, which have been identified as factors associated with mortality in prior published reports."

"This study draws attention to the tremendous burden of Clostridium difficile on patient mortality, health care costs and resources, transitional care utilization, and the extremely important need for aggressive prevention of this iatrogenic disease," she concluded.

The authors reported no funding or disclosures.

NEW YORK - Most, but not all, guideline-recommended drugs and their combinations yield modest survival benefits in older adults with multiple chronic medical conditions, researchers report.

"Until there is better evidence in older adults with multiple chronic conditions, clinicians need to take a more thoughtful and nuanced approach to medication prescribing," Dr. Mary E. Tinetti from Yale School of Medicine, New Haven, Connecticut said by email. "This is particularly supported by the fact that outside of perhaps anticoagulation for atrial fibrillation, the magnitude of benefit for most of these medications is quite modest."

Nearly 40% of adults 65 years and older take at least five prescription medications, but the benefits of drugs prescribed for a single condition are difficult to ascertain in the

presence of multiple conditions and drugs.

Dr. Tinetti and colleagues used data from the Medicare Current Beneficiary Survey to estimate the association between nine guideline-recommended and commonly prescribed drugs and death in more than 8,500 Medicare participants (mean age, 77.4 years).

The nine drugs included beta-blockers; calcium channel blockers; clopidogrel; metformin; renin-angiotensin system (RAS) blockers; selective serotonin reuptake inhibitors (SSRIs) and selective serotonin norepinephrine reuptake inhibitors (SNRIs); statins; thiazide diuretics; and warfarin.

The most common medical conditions included hypertension, hyperlipidemia, diabetes, and coronary artery disease, the researchers report in The BMJ, online Oct. 2. More than half of the participants took at least three of the nine study drugs, and the mean number of total drugs was 10.0.

The benefits on survival of beta-blockers, calcium channel blockers, RAS blockers, statins, and warfarin were comparable to those reported in randomized controlled trials, although for beta-blockers and warfarin they varied according to coexisting conditions.

Clopidogrel, metformin, and SSRIs/SNRIs, however, were not associated with survival benefits.

The association between drug use and mortality risk was generally similar across patterns of coexisting conditions, suggesting to the researchers "that benefits often remain despite comorbidity."

"Research focused on identifying the medications that have the greatest benefit (defined by patient's outcomes priority such as improved symptoms, optimal physical or cognitive function, or survival) and least harm for key subpopulations of individuals with varying combinations of coexisting conditions should be a top priority," Dr. Tinetti said. "Ironically, although these are the major users of health care, they have largely been ignored in research."

"Many individuals, although they have many conditions, only one or two are really affecting the outcomes that matter most to them," Dr. Tinetti said. "Therefore, medication regimens can be simplified by eliminating medications for conditions that are not likely to benefit the individuals outcome priority, such as improved symptoms, optimal physical or cognitive function, or simplified treatment regimens."

"We know that about one in five individuals are currently receiving guideline-recommended medications for one condition that may be harming another," Dr. Tinetti concluded. "A careful medication review will review these potential offending medications. Given the marginal benefit of each individual medication, and the importance of avoiding harm, potentially offending medications should be discontinued."

Dr. Una Makris from UT Southwestern Medical Center and Dallas VA Medical Center has reported on high-risk medication use among older veterans with chronic pain. She said by email, "As clinicians we need to understand that not all of the outcomes we measure and not all of the medical conditions a patient has will be perceived as equally important (by the patient), so involving patients in the decision of which medications (even if guideline concordant) to add or remove is integral. How we communicate with and educate our older patients about their chronic conditions and the risks/benefits of each medication is evolving; this often differs between specialty providers and primary care providers."

Dr. Makris added, "This publication should heighten our awareness that patients, especially older adults, often have multiple comorbidities that can be treated with multiple guideline driven therapies and that our goal is really to work with patients to determine which combination of drugs and for which condition optimizes the risks/benefits for that individual."

NEW YORK - Most, but not all, guideline-recommended drugs and their combinations yield modest survival benefits in older adults with multiple chronic medical conditions, researchers report.

"Until there is better evidence in older adults with multiple chronic conditions, clinicians need to take a more thoughtful and nuanced approach to medication prescribing," Dr. Mary E. Tinetti from Yale School of Medicine, New Haven, Connecticut said by email. "This is particularly supported by the fact that outside of perhaps anticoagulation for atrial fibrillation, the magnitude of benefit for most of these medications is quite modest."

Nearly 40% of adults 65 years and older take at least five prescription medications, but the benefits of drugs prescribed for a single condition are difficult to ascertain in the

presence of multiple conditions and drugs.

Dr. Tinetti and colleagues used data from the Medicare Current Beneficiary Survey to estimate the association between nine guideline-recommended and commonly prescribed drugs and death in more than 8,500 Medicare participants (mean age, 77.4 years).

The nine drugs included beta-blockers; calcium channel blockers; clopidogrel; metformin; renin-angiotensin system (RAS) blockers; selective serotonin reuptake inhibitors (SSRIs) and selective serotonin norepinephrine reuptake inhibitors (SNRIs); statins; thiazide diuretics; and warfarin.

The most common medical conditions included hypertension, hyperlipidemia, diabetes, and coronary artery disease, the researchers report in The BMJ, online Oct. 2. More than half of the participants took at least three of the nine study drugs, and the mean number of total drugs was 10.0.

The benefits on survival of beta-blockers, calcium channel blockers, RAS blockers, statins, and warfarin were comparable to those reported in randomized controlled trials, although for beta-blockers and warfarin they varied according to coexisting conditions.

Clopidogrel, metformin, and SSRIs/SNRIs, however, were not associated with survival benefits.

The association between drug use and mortality risk was generally similar across patterns of coexisting conditions, suggesting to the researchers "that benefits often remain despite comorbidity."

"Research focused on identifying the medications that have the greatest benefit (defined by patient's outcomes priority such as improved symptoms, optimal physical or cognitive function, or survival) and least harm for key subpopulations of individuals with varying combinations of coexisting conditions should be a top priority," Dr. Tinetti said. "Ironically, although these are the major users of health care, they have largely been ignored in research."

"Many individuals, although they have many conditions, only one or two are really affecting the outcomes that matter most to them," Dr. Tinetti said. "Therefore, medication regimens can be simplified by eliminating medications for conditions that are not likely to benefit the individuals outcome priority, such as improved symptoms, optimal physical or cognitive function, or simplified treatment regimens."

"We know that about one in five individuals are currently receiving guideline-recommended medications for one condition that may be harming another," Dr. Tinetti concluded. "A careful medication review will review these potential offending medications. Given the marginal benefit of each individual medication, and the importance of avoiding harm, potentially offending medications should be discontinued."

Dr. Una Makris from UT Southwestern Medical Center and Dallas VA Medical Center has reported on high-risk medication use among older veterans with chronic pain. She said by email, "As clinicians we need to understand that not all of the outcomes we measure and not all of the medical conditions a patient has will be perceived as equally important (by the patient), so involving patients in the decision of which medications (even if guideline concordant) to add or remove is integral. How we communicate with and educate our older patients about their chronic conditions and the risks/benefits of each medication is evolving; this often differs between specialty providers and primary care providers."

Dr. Makris added, "This publication should heighten our awareness that patients, especially older adults, often have multiple comorbidities that can be treated with multiple guideline driven therapies and that our goal is really to work with patients to determine which combination of drugs and for which condition optimizes the risks/benefits for that individual."

NEW YORK - Most, but not all, guideline-recommended drugs and their combinations yield modest survival benefits in older adults with multiple chronic medical conditions, researchers report.

"Until there is better evidence in older adults with multiple chronic conditions, clinicians need to take a more thoughtful and nuanced approach to medication prescribing," Dr. Mary E. Tinetti from Yale School of Medicine, New Haven, Connecticut said by email. "This is particularly supported by the fact that outside of perhaps anticoagulation for atrial fibrillation, the magnitude of benefit for most of these medications is quite modest."

Nearly 40% of adults 65 years and older take at least five prescription medications, but the benefits of drugs prescribed for a single condition are difficult to ascertain in the

presence of multiple conditions and drugs.

Dr. Tinetti and colleagues used data from the Medicare Current Beneficiary Survey to estimate the association between nine guideline-recommended and commonly prescribed drugs and death in more than 8,500 Medicare participants (mean age, 77.4 years).

The nine drugs included beta-blockers; calcium channel blockers; clopidogrel; metformin; renin-angiotensin system (RAS) blockers; selective serotonin reuptake inhibitors (SSRIs) and selective serotonin norepinephrine reuptake inhibitors (SNRIs); statins; thiazide diuretics; and warfarin.

The most common medical conditions included hypertension, hyperlipidemia, diabetes, and coronary artery disease, the researchers report in The BMJ, online Oct. 2. More than half of the participants took at least three of the nine study drugs, and the mean number of total drugs was 10.0.

The benefits on survival of beta-blockers, calcium channel blockers, RAS blockers, statins, and warfarin were comparable to those reported in randomized controlled trials, although for beta-blockers and warfarin they varied according to coexisting conditions.

Clopidogrel, metformin, and SSRIs/SNRIs, however, were not associated with survival benefits.

The association between drug use and mortality risk was generally similar across patterns of coexisting conditions, suggesting to the researchers "that benefits often remain despite comorbidity."

"Research focused on identifying the medications that have the greatest benefit (defined by patient's outcomes priority such as improved symptoms, optimal physical or cognitive function, or survival) and least harm for key subpopulations of individuals with varying combinations of coexisting conditions should be a top priority," Dr. Tinetti said. "Ironically, although these are the major users of health care, they have largely been ignored in research."

"Many individuals, although they have many conditions, only one or two are really affecting the outcomes that matter most to them," Dr. Tinetti said. "Therefore, medication regimens can be simplified by eliminating medications for conditions that are not likely to benefit the individuals outcome priority, such as improved symptoms, optimal physical or cognitive function, or simplified treatment regimens."

"We know that about one in five individuals are currently receiving guideline-recommended medications for one condition that may be harming another," Dr. Tinetti concluded. "A careful medication review will review these potential offending medications. Given the marginal benefit of each individual medication, and the importance of avoiding harm, potentially offending medications should be discontinued."

Dr. Una Makris from UT Southwestern Medical Center and Dallas VA Medical Center has reported on high-risk medication use among older veterans with chronic pain. She said by email, "As clinicians we need to understand that not all of the outcomes we measure and not all of the medical conditions a patient has will be perceived as equally important (by the patient), so involving patients in the decision of which medications (even if guideline concordant) to add or remove is integral. How we communicate with and educate our older patients about their chronic conditions and the risks/benefits of each medication is evolving; this often differs between specialty providers and primary care providers."

Dr. Makris added, "This publication should heighten our awareness that patients, especially older adults, often have multiple comorbidities that can be treated with multiple guideline driven therapies and that our goal is really to work with patients to determine which combination of drugs and for which condition optimizes the risks/benefits for that individual."

NEW YORK  - The five-factor ORBIT bleeding score accurately predicts major bleeding risk in patients with atrial fibrillation (AF) who are taking oral anticoagulants (OACs), researchers report.

"The ORBIT score highlights modifiable factors that increase bleeding risk and can help providers identify high-risk AF patients for closer monitoring," Dr. Emily C. O'Brien, from Duke Clinical Research Institute, Durham, North Carolina, said by email. "Along with clinical judgment, the ORBIT score can be used to give an estimate of bleeding risk for any AF patient considering OAC treatment."

Two existing bleeding scores - HAS-BLED and ATRIA - are based on small numbers of events and have shown inconsistent performance. They also may require elements that are not available for all OAC users, the researchers wrote.

Dr. O'Brien's team developed a five-element bleeding score and compared its performance with those of HAS-BLED and ATRIA using data from the ORBIT-AF and ROCKET-AF studies.

The numerical score included the five strongest predictors of bleeding:

-Older age (75 years and above): one point

-Reduced hemoglobin, hematocrit, or history of anemia: two points

-Bleeding history: two points

-Insufficient kidney function (eGFR below 60 mL/min/1.73 m2): one point

-Treatment with an antiplatelet agent: one point

Observed bleeding rates in the ORBIT-AF participants increased with increasing ORBIT bleeding score: from 2.4 per 100 patient-years in the low-risk group (scores 0-2) to 4.7 per 100 patient-years in the medium-risk group (score 3) to 8.1 per 100 patient-years in the high-risk group (scores 4-7), according to the Sept. 30 European Heart Journal online report.

In both the ORBIT-AF and ROCKET-AF cohorts, the ORBIT bleeding score showed better discrimination than the HAS-BLED and ATRIA scores.

Model calibration analysis also showed superior calibration for the ORBIT bleeding score. The HAS-BLED score showed relatively poor calibration for low-risk score strata, whereas the ATRIA score showed poor calibration for most risk groups.

"The ORBIT score is a simple, useful tool that predicts bleeding as well as other, more complicated scores and can be used in any AF patient regardless of the type of OAC he or she is taking," Dr. O'Brien said.

"For chronic conditions like AF, periodic assessment of risk for adverse events is important to support clinical decision-making," Dr. O'Brien explained. "Risk factors may

change over time particularly as patients get older. Therefore, incorporating new data on these factors into longitudinal risk assessment provides an optimal framework for ongoing AF management."

"While bleeding risk estimation can be helpful in identifying high-risk AF patients for closer monitoring, it is important to note that prior work has demonstrated a net clinical benefit of OAC even in patients with high estimated bleeding risk," the researchers wrote. "Further, while risk scores provide important information to the clinician for estimating risk of adverse events, they represent only one consideration relevant to therapeutic decision making."

Janssen Scientific Affairs sponsors ORBIT-AF; the Agency for Healthcare Research and Quality partially supported this research. Ten coauthors reported relevant relationships.

NEW YORK  - The five-factor ORBIT bleeding score accurately predicts major bleeding risk in patients with atrial fibrillation (AF) who are taking oral anticoagulants (OACs), researchers report.

"The ORBIT score highlights modifiable factors that increase bleeding risk and can help providers identify high-risk AF patients for closer monitoring," Dr. Emily C. O'Brien, from Duke Clinical Research Institute, Durham, North Carolina, said by email. "Along with clinical judgment, the ORBIT score can be used to give an estimate of bleeding risk for any AF patient considering OAC treatment."

Two existing bleeding scores - HAS-BLED and ATRIA - are based on small numbers of events and have shown inconsistent performance. They also may require elements that are not available for all OAC users, the researchers wrote.

Dr. O'Brien's team developed a five-element bleeding score and compared its performance with those of HAS-BLED and ATRIA using data from the ORBIT-AF and ROCKET-AF studies.

The numerical score included the five strongest predictors of bleeding:

-Older age (75 years and above): one point

-Reduced hemoglobin, hematocrit, or history of anemia: two points

-Bleeding history: two points

-Insufficient kidney function (eGFR below 60 mL/min/1.73 m2): one point

-Treatment with an antiplatelet agent: one point

Observed bleeding rates in the ORBIT-AF participants increased with increasing ORBIT bleeding score: from 2.4 per 100 patient-years in the low-risk group (scores 0-2) to 4.7 per 100 patient-years in the medium-risk group (score 3) to 8.1 per 100 patient-years in the high-risk group (scores 4-7), according to the Sept. 30 European Heart Journal online report.

In both the ORBIT-AF and ROCKET-AF cohorts, the ORBIT bleeding score showed better discrimination than the HAS-BLED and ATRIA scores.

Model calibration analysis also showed superior calibration for the ORBIT bleeding score. The HAS-BLED score showed relatively poor calibration for low-risk score strata, whereas the ATRIA score showed poor calibration for most risk groups.

"The ORBIT score is a simple, useful tool that predicts bleeding as well as other, more complicated scores and can be used in any AF patient regardless of the type of OAC he or she is taking," Dr. O'Brien said.

"For chronic conditions like AF, periodic assessment of risk for adverse events is important to support clinical decision-making," Dr. O'Brien explained. "Risk factors may

change over time particularly as patients get older. Therefore, incorporating new data on these factors into longitudinal risk assessment provides an optimal framework for ongoing AF management."

"While bleeding risk estimation can be helpful in identifying high-risk AF patients for closer monitoring, it is important to note that prior work has demonstrated a net clinical benefit of OAC even in patients with high estimated bleeding risk," the researchers wrote. "Further, while risk scores provide important information to the clinician for estimating risk of adverse events, they represent only one consideration relevant to therapeutic decision making."

Janssen Scientific Affairs sponsors ORBIT-AF; the Agency for Healthcare Research and Quality partially supported this research. Ten coauthors reported relevant relationships.

NEW YORK  - The five-factor ORBIT bleeding score accurately predicts major bleeding risk in patients with atrial fibrillation (AF) who are taking oral anticoagulants (OACs), researchers report.

"The ORBIT score highlights modifiable factors that increase bleeding risk and can help providers identify high-risk AF patients for closer monitoring," Dr. Emily C. O'Brien, from Duke Clinical Research Institute, Durham, North Carolina, said by email. "Along with clinical judgment, the ORBIT score can be used to give an estimate of bleeding risk for any AF patient considering OAC treatment."

Two existing bleeding scores - HAS-BLED and ATRIA - are based on small numbers of events and have shown inconsistent performance. They also may require elements that are not available for all OAC users, the researchers wrote.

Dr. O'Brien's team developed a five-element bleeding score and compared its performance with those of HAS-BLED and ATRIA using data from the ORBIT-AF and ROCKET-AF studies.

The numerical score included the five strongest predictors of bleeding:

-Older age (75 years and above): one point

-Reduced hemoglobin, hematocrit, or history of anemia: two points

-Bleeding history: two points

-Insufficient kidney function (eGFR below 60 mL/min/1.73 m2): one point

-Treatment with an antiplatelet agent: one point

Observed bleeding rates in the ORBIT-AF participants increased with increasing ORBIT bleeding score: from 2.4 per 100 patient-years in the low-risk group (scores 0-2) to 4.7 per 100 patient-years in the medium-risk group (score 3) to 8.1 per 100 patient-years in the high-risk group (scores 4-7), according to the Sept. 30 European Heart Journal online report.

In both the ORBIT-AF and ROCKET-AF cohorts, the ORBIT bleeding score showed better discrimination than the HAS-BLED and ATRIA scores.

Model calibration analysis also showed superior calibration for the ORBIT bleeding score. The HAS-BLED score showed relatively poor calibration for low-risk score strata, whereas the ATRIA score showed poor calibration for most risk groups.

"The ORBIT score is a simple, useful tool that predicts bleeding as well as other, more complicated scores and can be used in any AF patient regardless of the type of OAC he or she is taking," Dr. O'Brien said.

"For chronic conditions like AF, periodic assessment of risk for adverse events is important to support clinical decision-making," Dr. O'Brien explained. "Risk factors may

change over time particularly as patients get older. Therefore, incorporating new data on these factors into longitudinal risk assessment provides an optimal framework for ongoing AF management."

"While bleeding risk estimation can be helpful in identifying high-risk AF patients for closer monitoring, it is important to note that prior work has demonstrated a net clinical benefit of OAC even in patients with high estimated bleeding risk," the researchers wrote. "Further, while risk scores provide important information to the clinician for estimating risk of adverse events, they represent only one consideration relevant to therapeutic decision making."

Janssen Scientific Affairs sponsors ORBIT-AF; the Agency for Healthcare Research and Quality partially supported this research. Ten coauthors reported relevant relationships.

When patients are hospitalized more than once in the same month, it may have more to do with their income or education levels than the quality of care they received, a U.S. study suggests.

Perhaps unsurprisingly, patients 85 and older are more likely to return to the hospital within 30 days of being sent home than people a decade or two younger, according to the analysis of data from Medicare, the U.S. health program for the elderly and disabled.

But patients also have higher odds of returning soon after discharge if they lack a high school diploma, have limited income and assets or have health benefits from Medicaid, the U.S. health program for the poor.

The findings suggest that Medicare penalties for what's known as readmissions under the Affordable Care Act may in some instances mete out punishment for outcomes that are beyond doctors' control, said lead study author Dr. Michael Barnett and senior author Dr. Michael McWilliams, colleagues at Harvard Medical School and Brigham and Women's Hospital in Boston.

"Hospitals are being penalized to a large extent based on the patients they serve," the doctors said by email. "Patients admitted to hospitals with higher readmission rates are sicker and more socially disadvantaged in a variety of ways than patients admitted to hospitals with lower readmission rates."

Under the current penalty system, Medicare deducts 3 percent from inpatient payments to hospitals with higher than expected readmission rates, the researchers report in JAMA Internal Medicine. Expected rates are only adjusted for patients' age, sex and recent diagnoses including the one from their hospital stay.

In 2014, the second year of the program, about 2,600 hospitals were fined a combined $428 million for excessive readmissions, the authors report.

To get a better understanding of how individual patient characteristics might influence repeat hospitalizations, the researchers examined several other variables Medicare doesn't consider in determining expected readmission rates - such as education and income levels, marital status, employment, race and ethnicity, smoking status and drinking habits.

They linked records from a nationwide health and retirement survey of Americans over 50 collected between 2000 and 2010 to data from Medicare claims from 2000 to 2012. The combined analysis assessed more than 8,000 hospital admissions.

The researchers sorted hospitals into quintiles based on readmission rates. They found that at least half of the observed difference in the probability of repeat hospitalizations between hospitals with the highest and lowest readmission rates might be accounted for by patient characteristics not currently considered by Medicare.

When researchers only used Medicare's criteria comparing readmission rates, they found the probability of repeat hospitalization was about 15 percent at facilities with the lowest rates and about 19.5 percent at hospitals with the highest rates.

But when they took another look using more criteria on patients' medical, social and economic characteristics, the gap between hospitals with the lowest and the highest readmission rates narrowed to 16 percent and 18.4 percent, respectively, odds of repeat hospitalization.

One limitation of the study, the authors acknowledge, is the data didn't allow them to calculate how considering individual patient characteristics might impact readmission rates at specific hospitals.

Even so, the findings suggest that the current Medicare penalty system for repeat hospitalizations may put facilities serving poor communities at a distinct financial disadvantage, Dr. Carl van Walraven, a senior scientist at the Ottawa Hospital Research Institute in Canada, noted in an accompanying editorial.

"Differences between hospitals in readmissions may be due to who is treated rather than how they're treated," van Walraven said by email.

When patients are hospitalized more than once in the same month, it may have more to do with their income or education levels than the quality of care they received, a U.S. study suggests.

Perhaps unsurprisingly, patients 85 and older are more likely to return to the hospital within 30 days of being sent home than people a decade or two younger, according to the analysis of data from Medicare, the U.S. health program for the elderly and disabled.

But patients also have higher odds of returning soon after discharge if they lack a high school diploma, have limited income and assets or have health benefits from Medicaid, the U.S. health program for the poor.

The findings suggest that Medicare penalties for what's known as readmissions under the Affordable Care Act may in some instances mete out punishment for outcomes that are beyond doctors' control, said lead study author Dr. Michael Barnett and senior author Dr. Michael McWilliams, colleagues at Harvard Medical School and Brigham and Women's Hospital in Boston.

"Hospitals are being penalized to a large extent based on the patients they serve," the doctors said by email. "Patients admitted to hospitals with higher readmission rates are sicker and more socially disadvantaged in a variety of ways than patients admitted to hospitals with lower readmission rates."

Under the current penalty system, Medicare deducts 3 percent from inpatient payments to hospitals with higher than expected readmission rates, the researchers report in JAMA Internal Medicine. Expected rates are only adjusted for patients' age, sex and recent diagnoses including the one from their hospital stay.

In 2014, the second year of the program, about 2,600 hospitals were fined a combined $428 million for excessive readmissions, the authors report.

To get a better understanding of how individual patient characteristics might influence repeat hospitalizations, the researchers examined several other variables Medicare doesn't consider in determining expected readmission rates - such as education and income levels, marital status, employment, race and ethnicity, smoking status and drinking habits.

They linked records from a nationwide health and retirement survey of Americans over 50 collected between 2000 and 2010 to data from Medicare claims from 2000 to 2012. The combined analysis assessed more than 8,000 hospital admissions.

The researchers sorted hospitals into quintiles based on readmission rates. They found that at least half of the observed difference in the probability of repeat hospitalizations between hospitals with the highest and lowest readmission rates might be accounted for by patient characteristics not currently considered by Medicare.

When researchers only used Medicare's criteria comparing readmission rates, they found the probability of repeat hospitalization was about 15 percent at facilities with the lowest rates and about 19.5 percent at hospitals with the highest rates.

But when they took another look using more criteria on patients' medical, social and economic characteristics, the gap between hospitals with the lowest and the highest readmission rates narrowed to 16 percent and 18.4 percent, respectively, odds of repeat hospitalization.

One limitation of the study, the authors acknowledge, is the data didn't allow them to calculate how considering individual patient characteristics might impact readmission rates at specific hospitals.

Even so, the findings suggest that the current Medicare penalty system for repeat hospitalizations may put facilities serving poor communities at a distinct financial disadvantage, Dr. Carl van Walraven, a senior scientist at the Ottawa Hospital Research Institute in Canada, noted in an accompanying editorial.

"Differences between hospitals in readmissions may be due to who is treated rather than how they're treated," van Walraven said by email.

When patients are hospitalized more than once in the same month, it may have more to do with their income or education levels than the quality of care they received, a U.S. study suggests.

Perhaps unsurprisingly, patients 85 and older are more likely to return to the hospital within 30 days of being sent home than people a decade or two younger, according to the analysis of data from Medicare, the U.S. health program for the elderly and disabled.

But patients also have higher odds of returning soon after discharge if they lack a high school diploma, have limited income and assets or have health benefits from Medicaid, the U.S. health program for the poor.

The findings suggest that Medicare penalties for what's known as readmissions under the Affordable Care Act may in some instances mete out punishment for outcomes that are beyond doctors' control, said lead study author Dr. Michael Barnett and senior author Dr. Michael McWilliams, colleagues at Harvard Medical School and Brigham and Women's Hospital in Boston.

"Hospitals are being penalized to a large extent based on the patients they serve," the doctors said by email. "Patients admitted to hospitals with higher readmission rates are sicker and more socially disadvantaged in a variety of ways than patients admitted to hospitals with lower readmission rates."

Under the current penalty system, Medicare deducts 3 percent from inpatient payments to hospitals with higher than expected readmission rates, the researchers report in JAMA Internal Medicine. Expected rates are only adjusted for patients' age, sex and recent diagnoses including the one from their hospital stay.

In 2014, the second year of the program, about 2,600 hospitals were fined a combined $428 million for excessive readmissions, the authors report.

To get a better understanding of how individual patient characteristics might influence repeat hospitalizations, the researchers examined several other variables Medicare doesn't consider in determining expected readmission rates - such as education and income levels, marital status, employment, race and ethnicity, smoking status and drinking habits.

They linked records from a nationwide health and retirement survey of Americans over 50 collected between 2000 and 2010 to data from Medicare claims from 2000 to 2012. The combined analysis assessed more than 8,000 hospital admissions.

The researchers sorted hospitals into quintiles based on readmission rates. They found that at least half of the observed difference in the probability of repeat hospitalizations between hospitals with the highest and lowest readmission rates might be accounted for by patient characteristics not currently considered by Medicare.

When researchers only used Medicare's criteria comparing readmission rates, they found the probability of repeat hospitalization was about 15 percent at facilities with the lowest rates and about 19.5 percent at hospitals with the highest rates.

But when they took another look using more criteria on patients' medical, social and economic characteristics, the gap between hospitals with the lowest and the highest readmission rates narrowed to 16 percent and 18.4 percent, respectively, odds of repeat hospitalization.

One limitation of the study, the authors acknowledge, is the data didn't allow them to calculate how considering individual patient characteristics might impact readmission rates at specific hospitals.

Even so, the findings suggest that the current Medicare penalty system for repeat hospitalizations may put facilities serving poor communities at a distinct financial disadvantage, Dr. Carl van Walraven, a senior scientist at the Ottawa Hospital Research Institute in Canada, noted in an accompanying editorial.

"Differences between hospitals in readmissions may be due to who is treated rather than how they're treated," van Walraven said by email.

NEW YORK - Chemotherapy-related hospitalization happens much more often in the real world than in drug trials, according to a new study.

Patients with advanced lung cancer receiving chemotherapy in real-world settings were almost eight times more likely to be hospitalized during treatment than those participating in clinical trials.

What's more, very few clinical trials even report how often participants are hospitalized during the research, the study authors found.

"Clinical trials should be routinely reporting their hospitalization rates so we know what to expect," said senior author Dr. Monika Krzyzanowska of the Princess Margaret Cancer Center in Toronto, Canada. "I think that (hospitalization is) actually much more common than we ever anticipated," Krzyzanowska said.

For the new meta-analysis, released online September 17 in JAMA Oncology, the researchers looked at data on patients receiving chemotherapy for metastatic non-small-cell lung cancer, from five reports of clinical trials with a total of 3962 people that specified how many hospitalizations occurred, and five studies involving 8624 people receiving chemotherapy in real-world settings.

Overall, 51% of the real-world patients were hospitalized during their treatments, compared to 16% of those in clinical trials.

Some of the research looked at factors related to the risk of hospitalization like the type of chemotherapy used and hospital performance measures, but results varied from study to study and Krzyzanowska said that she can't say with confidence which factors may be tied to an increased risk of being hospitalized.

But, she said, similar patterns of high hospitalizations are likely to be found among people with other cancers and on other types of treatments.

"I think this is unfortunately a common phenomenon across disease site and treatment regimen," Krzyzanowska said.

Knowing how much time patients may spend in hospitals during chemotherapy might help them and their doctors in deciding which treatment is right, Krzyzanowska said.

"I think the low-hanging fruit is that clinical trials should start reporting hospitalizations," she said of the findings.

With that kind of data, the researchers suggest, scientists can calculate the risk of hospitalization per month of chemotherapy and ultimately provide that to patients.

Krzyzanowska also said she'd like to look at what factors drive hospitalizations among cancer patients receiving chemotherapy.

"I definitely think there is a substantial portion of people whose symptoms can be managed earlier so they don't end up in the hospital," she said.

NEW YORK - Chemotherapy-related hospitalization happens much more often in the real world than in drug trials, according to a new study.

Patients with advanced lung cancer receiving chemotherapy in real-world settings were almost eight times more likely to be hospitalized during treatment than those participating in clinical trials.

What's more, very few clinical trials even report how often participants are hospitalized during the research, the study authors found.

"Clinical trials should be routinely reporting their hospitalization rates so we know what to expect," said senior author Dr. Monika Krzyzanowska of the Princess Margaret Cancer Center in Toronto, Canada. "I think that (hospitalization is) actually much more common than we ever anticipated," Krzyzanowska said.

For the new meta-analysis, released online September 17 in JAMA Oncology, the researchers looked at data on patients receiving chemotherapy for metastatic non-small-cell lung cancer, from five reports of clinical trials with a total of 3962 people that specified how many hospitalizations occurred, and five studies involving 8624 people receiving chemotherapy in real-world settings.

Overall, 51% of the real-world patients were hospitalized during their treatments, compared to 16% of those in clinical trials.

Some of the research looked at factors related to the risk of hospitalization like the type of chemotherapy used and hospital performance measures, but results varied from study to study and Krzyzanowska said that she can't say with confidence which factors may be tied to an increased risk of being hospitalized.

But, she said, similar patterns of high hospitalizations are likely to be found among people with other cancers and on other types of treatments.

"I think this is unfortunately a common phenomenon across disease site and treatment regimen," Krzyzanowska said.

Knowing how much time patients may spend in hospitals during chemotherapy might help them and their doctors in deciding which treatment is right, Krzyzanowska said.

"I think the low-hanging fruit is that clinical trials should start reporting hospitalizations," she said of the findings.

With that kind of data, the researchers suggest, scientists can calculate the risk of hospitalization per month of chemotherapy and ultimately provide that to patients.

Krzyzanowska also said she'd like to look at what factors drive hospitalizations among cancer patients receiving chemotherapy.

"I definitely think there is a substantial portion of people whose symptoms can be managed earlier so they don't end up in the hospital," she said.

NEW YORK - Chemotherapy-related hospitalization happens much more often in the real world than in drug trials, according to a new study.

Patients with advanced lung cancer receiving chemotherapy in real-world settings were almost eight times more likely to be hospitalized during treatment than those participating in clinical trials.

What's more, very few clinical trials even report how often participants are hospitalized during the research, the study authors found.

"Clinical trials should be routinely reporting their hospitalization rates so we know what to expect," said senior author Dr. Monika Krzyzanowska of the Princess Margaret Cancer Center in Toronto, Canada. "I think that (hospitalization is) actually much more common than we ever anticipated," Krzyzanowska said.

For the new meta-analysis, released online September 17 in JAMA Oncology, the researchers looked at data on patients receiving chemotherapy for metastatic non-small-cell lung cancer, from five reports of clinical trials with a total of 3962 people that specified how many hospitalizations occurred, and five studies involving 8624 people receiving chemotherapy in real-world settings.

Overall, 51% of the real-world patients were hospitalized during their treatments, compared to 16% of those in clinical trials.

Some of the research looked at factors related to the risk of hospitalization like the type of chemotherapy used and hospital performance measures, but results varied from study to study and Krzyzanowska said that she can't say with confidence which factors may be tied to an increased risk of being hospitalized.

But, she said, similar patterns of high hospitalizations are likely to be found among people with other cancers and on other types of treatments.

"I think this is unfortunately a common phenomenon across disease site and treatment regimen," Krzyzanowska said.

Knowing how much time patients may spend in hospitals during chemotherapy might help them and their doctors in deciding which treatment is right, Krzyzanowska said.

"I think the low-hanging fruit is that clinical trials should start reporting hospitalizations," she said of the findings.

With that kind of data, the researchers suggest, scientists can calculate the risk of hospitalization per month of chemotherapy and ultimately provide that to patients.

Krzyzanowska also said she'd like to look at what factors drive hospitalizations among cancer patients receiving chemotherapy.

"I definitely think there is a substantial portion of people whose symptoms can be managed earlier so they don't end up in the hospital," she said.

NEW YORK  - Older adults living in long-term care are more than twice as likely as their peers living at home to suffer a fracture, and a new guideline endorsed by the Scientific Advisory Council of Osteoporosis Canada explains how to reduce their risk.

Residents of long-term care tend to be frailer and have more health problems than similar people who live on their own, which explains the higher risk of fractures in long-term care facilities, said lead author Dr. Alexandra Papaioannou of McMaster University and Hamilton Health Sciences in Hamilton, Ontario.

"Up to a third of seniors in long-term care suffer a fracture," often of the hip or spine, she said. For these residents, "long-term care is their home, the nurses know them, and acute care can be a frightening traumatic experience for residents."

The authors developed the new guideline based on input from older people and their families, who most wanted to avoid pain, loss of activity, and hospitalization, Papaioannou said. They also studied published literature on the risks and benefits of strategies to prevent fracture.

The guideline strongly recommends calcium supplementation of 1200 mg or three servings of dairy daily for people older than 70. These calcium levels reduce hip fracture risk and slightly reduce the risk of other fractures, but they may also cause gastrointestinal side effects. For residents who want to avoid these, supplementation may not be a good option, the authors write.

Residents at high risk of fracture, i.e., those with prior fracture of the hip or spine, more than one prior fracture, or one prior fracture and recent use of glucocorticoids, should also take daily vitamin D3 supplements, which are more affordable than vitamin D2, the authors wrote online September 14 in CMAJ.

They also recommend that high-risk residents take alendronate weekly, or risedronate weekly or monthly, as first-line therapy to prevent fractures, as long as they do not have difficulty swallowing and can remain upright for 30 minutes after administration, and they do not have severe renal insufficiency.

For residents who are at high risk of fractures and who have difficulty taking oral medications, they recommend zoledronic acid or denosumab as first-line therapy.

High-risk residents who are mobile should wear hip-protectors, which can protect against fracture in the event of a fall. Low-risk residents who are mobile may wear the devices, depending on their values and preferences.

Balance, strength, and functional exercise can help prevent falls for low-risk residents, and may be useful for high-risk residents, but the exercise itself increases the risk of fall slightly.

Lastly, the authors recommend that all residents have "multifactorial interventions" that are tailored to each individual and include medication reviews, environmental hazard assessment, assistive device use, exercise, and educational interventions for staff.

"Many residents have multiple medical conditions and we need to make sure that they include their lifespan and goals of care in the assessment," Papaioannou said. "The goals of those with short lifespans may be very different from those with longer lifespans."

These recommendations are similar to those for residential care facilities in Australia and the ones made by the Society for Post-Acute and Long-Term Care Medicine in the U.S., the authors point out.

"The document is an excellent guide on how to identify patients at risk, who should be treated and how," said Dr. Gustavo Duque, director of the Musculoskeletal Ageing Research Program at the University of Sydney in Australia.

Doctors often reduce medications for residents who are admitted to nursing homes without realizing that those with osteoporosis or previous fractures are more likely to suffer a fracture at their new residence than in the community, Duque, who was not involved in writing the new guideline, said by email.

"Ceasing osteoporosis treatment has demonstrated to increase the risk of fractures," Duque said. "Unfortunately we see that situation every day."

NEW YORK  - Older adults living in long-term care are more than twice as likely as their peers living at home to suffer a fracture, and a new guideline endorsed by the Scientific Advisory Council of Osteoporosis Canada explains how to reduce their risk.

Residents of long-term care tend to be frailer and have more health problems than similar people who live on their own, which explains the higher risk of fractures in long-term care facilities, said lead author Dr. Alexandra Papaioannou of McMaster University and Hamilton Health Sciences in Hamilton, Ontario.

"Up to a third of seniors in long-term care suffer a fracture," often of the hip or spine, she said. For these residents, "long-term care is their home, the nurses know them, and acute care can be a frightening traumatic experience for residents."

The authors developed the new guideline based on input from older people and their families, who most wanted to avoid pain, loss of activity, and hospitalization, Papaioannou said. They also studied published literature on the risks and benefits of strategies to prevent fracture.

The guideline strongly recommends calcium supplementation of 1200 mg or three servings of dairy daily for people older than 70. These calcium levels reduce hip fracture risk and slightly reduce the risk of other fractures, but they may also cause gastrointestinal side effects. For residents who want to avoid these, supplementation may not be a good option, the authors write.

Residents at high risk of fracture, i.e., those with prior fracture of the hip or spine, more than one prior fracture, or one prior fracture and recent use of glucocorticoids, should also take daily vitamin D3 supplements, which are more affordable than vitamin D2, the authors wrote online September 14 in CMAJ.

They also recommend that high-risk residents take alendronate weekly, or risedronate weekly or monthly, as first-line therapy to prevent fractures, as long as they do not have difficulty swallowing and can remain upright for 30 minutes after administration, and they do not have severe renal insufficiency.

For residents who are at high risk of fractures and who have difficulty taking oral medications, they recommend zoledronic acid or denosumab as first-line therapy.

High-risk residents who are mobile should wear hip-protectors, which can protect against fracture in the event of a fall. Low-risk residents who are mobile may wear the devices, depending on their values and preferences.

Balance, strength, and functional exercise can help prevent falls for low-risk residents, and may be useful for high-risk residents, but the exercise itself increases the risk of fall slightly.

Lastly, the authors recommend that all residents have "multifactorial interventions" that are tailored to each individual and include medication reviews, environmental hazard assessment, assistive device use, exercise, and educational interventions for staff.

"Many residents have multiple medical conditions and we need to make sure that they include their lifespan and goals of care in the assessment," Papaioannou said. "The goals of those with short lifespans may be very different from those with longer lifespans."

These recommendations are similar to those for residential care facilities in Australia and the ones made by the Society for Post-Acute and Long-Term Care Medicine in the U.S., the authors point out.

"The document is an excellent guide on how to identify patients at risk, who should be treated and how," said Dr. Gustavo Duque, director of the Musculoskeletal Ageing Research Program at the University of Sydney in Australia.

Doctors often reduce medications for residents who are admitted to nursing homes without realizing that those with osteoporosis or previous fractures are more likely to suffer a fracture at their new residence than in the community, Duque, who was not involved in writing the new guideline, said by email.

"Ceasing osteoporosis treatment has demonstrated to increase the risk of fractures," Duque said. "Unfortunately we see that situation every day."

NEW YORK  - Older adults living in long-term care are more than twice as likely as their peers living at home to suffer a fracture, and a new guideline endorsed by the Scientific Advisory Council of Osteoporosis Canada explains how to reduce their risk.

Residents of long-term care tend to be frailer and have more health problems than similar people who live on their own, which explains the higher risk of fractures in long-term care facilities, said lead author Dr. Alexandra Papaioannou of McMaster University and Hamilton Health Sciences in Hamilton, Ontario.

"Up to a third of seniors in long-term care suffer a fracture," often of the hip or spine, she said. For these residents, "long-term care is their home, the nurses know them, and acute care can be a frightening traumatic experience for residents."

The authors developed the new guideline based on input from older people and their families, who most wanted to avoid pain, loss of activity, and hospitalization, Papaioannou said. They also studied published literature on the risks and benefits of strategies to prevent fracture.

The guideline strongly recommends calcium supplementation of 1200 mg or three servings of dairy daily for people older than 70. These calcium levels reduce hip fracture risk and slightly reduce the risk of other fractures, but they may also cause gastrointestinal side effects. For residents who want to avoid these, supplementation may not be a good option, the authors write.

Residents at high risk of fracture, i.e., those with prior fracture of the hip or spine, more than one prior fracture, or one prior fracture and recent use of glucocorticoids, should also take daily vitamin D3 supplements, which are more affordable than vitamin D2, the authors wrote online September 14 in CMAJ.

They also recommend that high-risk residents take alendronate weekly, or risedronate weekly or monthly, as first-line therapy to prevent fractures, as long as they do not have difficulty swallowing and can remain upright for 30 minutes after administration, and they do not have severe renal insufficiency.

For residents who are at high risk of fractures and who have difficulty taking oral medications, they recommend zoledronic acid or denosumab as first-line therapy.

High-risk residents who are mobile should wear hip-protectors, which can protect against fracture in the event of a fall. Low-risk residents who are mobile may wear the devices, depending on their values and preferences.

Balance, strength, and functional exercise can help prevent falls for low-risk residents, and may be useful for high-risk residents, but the exercise itself increases the risk of fall slightly.

Lastly, the authors recommend that all residents have "multifactorial interventions" that are tailored to each individual and include medication reviews, environmental hazard assessment, assistive device use, exercise, and educational interventions for staff.

"Many residents have multiple medical conditions and we need to make sure that they include their lifespan and goals of care in the assessment," Papaioannou said. "The goals of those with short lifespans may be very different from those with longer lifespans."

These recommendations are similar to those for residential care facilities in Australia and the ones made by the Society for Post-Acute and Long-Term Care Medicine in the U.S., the authors point out.

"The document is an excellent guide on how to identify patients at risk, who should be treated and how," said Dr. Gustavo Duque, director of the Musculoskeletal Ageing Research Program at the University of Sydney in Australia.

Doctors often reduce medications for residents who are admitted to nursing homes without realizing that those with osteoporosis or previous fractures are more likely to suffer a fracture at their new residence than in the community, Duque, who was not involved in writing the new guideline, said by email.

"Ceasing osteoporosis treatment has demonstrated to increase the risk of fractures," Duque said. "Unfortunately we see that situation every day."

After a heart attack, black patients typically don't live as long as whites - a racial difference that is starkest among the affluent - according to a new U.S. study.

Researchers evaluated data on more than 132,000 white heart attack patients and almost 9,000 black patients covered by Medicare, the government health program for the elderly and disabled. They used postal codes to assess income levels in patients' communities.

After 17 years of follow-up, the overall survival rate was 7.4 percent for white patients and 5.7 percent for black patients, according to the results published in Circulation, the journal of the American Heart Association.

On average, across all ages, white patients in low-income areas lived longer after a heart attack - about 5.6 years compared with 5.4 years for black patients. But in high-income communities, the gap widened to a life expectancy of seven years for white people and 6.3 years for black individuals.

"We found that socioeconomic status did not explain the racial disparities in life expectancy after a heart attack," lead study author Dr. Emily Bucholz of Boston Children's Hospital said by email.

"Contrary to common belief, this suggests that improving socioeconomic standing may improve outcomes for black and white patients globally but is unlikely to eliminate racial disparities in health," Bucholz added.

To see how race and class impact heart attack outcomes, Bucholz and colleagues reviewed health records collected from 1994 to 1996 for patients aged 65 to 90 years.

Just 6.3 percent of the patients were black, and only 6.8 percent lived in low-income communities, based on the typical household income in their postal codes.

Among white patients under 80, life expectancy was longest for patients in the most affluent neighborhoods and it got progressively shorter for middle-income and poor communities, the study found.

By contrast, life expectancy was similar for black patients residing in poor and middle-income communities across all ages. Only black patients under age 75 living in affluent areas had a survival advantage compared with their peers in less wealthy neighborhoods.

One shortcoming of the study is that it included a small proportion of black and poor patients, the authors acknowledge. It's also possible that using postal codes to assess income may have led to some instances where income levels were inflated or underestimated, the authors note.

It's possible that black patients living in affluent areas don't fare as well as white patients because they don't have the same amount of social support from their peers, said Dr. Joaquin Cigarroa, a cardiovascular medicine researcher at Oregon Health & Science University in Portland.

In poor neighborhoods, black patients may face additional challenges to surviving a heart attack, added Cigarroa, who wasn't involved in the study.

"They more often live in low socioeconomic segments of our community that often have less access to health care resources and less access to stores with good nutrition," Cigarroa said by email. "In addition, these segments of our community are often not ideally configured for promoting physical activity with parks, sidewalks, bike lanes, etc."

The study findings highlight a need to improve outcomes among poor and black patients and suggest some differences in heart attack survival may come down to disparities in quality of care, said senior study author Dr. Harlan Krumholz of Yale University School of Medicine in New Haven, Connecticut.

Because black patients have a greater burden of heart disease than white people, doctors may also need to focus more on prevention in this community, Krumholz said by email.

"Healthy heart habits may be even more important for African-Americans, for whom avoiding a heart attack is even more important given their worse outcomes after the event," Krumholz said.

After a heart attack, black patients typically don't live as long as whites - a racial difference that is starkest among the affluent - according to a new U.S. study.

Researchers evaluated data on more than 132,000 white heart attack patients and almost 9,000 black patients covered by Medicare, the government health program for the elderly and disabled. They used postal codes to assess income levels in patients' communities.

After 17 years of follow-up, the overall survival rate was 7.4 percent for white patients and 5.7 percent for black patients, according to the results published in Circulation, the journal of the American Heart Association.

On average, across all ages, white patients in low-income areas lived longer after a heart attack - about 5.6 years compared with 5.4 years for black patients. But in high-income communities, the gap widened to a life expectancy of seven years for white people and 6.3 years for black individuals.

"We found that socioeconomic status did not explain the racial disparities in life expectancy after a heart attack," lead study author Dr. Emily Bucholz of Boston Children's Hospital said by email.

"Contrary to common belief, this suggests that improving socioeconomic standing may improve outcomes for black and white patients globally but is unlikely to eliminate racial disparities in health," Bucholz added.

To see how race and class impact heart attack outcomes, Bucholz and colleagues reviewed health records collected from 1994 to 1996 for patients aged 65 to 90 years.

Just 6.3 percent of the patients were black, and only 6.8 percent lived in low-income communities, based on the typical household income in their postal codes.

Among white patients under 80, life expectancy was longest for patients in the most affluent neighborhoods and it got progressively shorter for middle-income and poor communities, the study found.

By contrast, life expectancy was similar for black patients residing in poor and middle-income communities across all ages. Only black patients under age 75 living in affluent areas had a survival advantage compared with their peers in less wealthy neighborhoods.

One shortcoming of the study is that it included a small proportion of black and poor patients, the authors acknowledge. It's also possible that using postal codes to assess income may have led to some instances where income levels were inflated or underestimated, the authors note.

It's possible that black patients living in affluent areas don't fare as well as white patients because they don't have the same amount of social support from their peers, said Dr. Joaquin Cigarroa, a cardiovascular medicine researcher at Oregon Health & Science University in Portland.

In poor neighborhoods, black patients may face additional challenges to surviving a heart attack, added Cigarroa, who wasn't involved in the study.

"They more often live in low socioeconomic segments of our community that often have less access to health care resources and less access to stores with good nutrition," Cigarroa said by email. "In addition, these segments of our community are often not ideally configured for promoting physical activity with parks, sidewalks, bike lanes, etc."

The study findings highlight a need to improve outcomes among poor and black patients and suggest some differences in heart attack survival may come down to disparities in quality of care, said senior study author Dr. Harlan Krumholz of Yale University School of Medicine in New Haven, Connecticut.

Because black patients have a greater burden of heart disease than white people, doctors may also need to focus more on prevention in this community, Krumholz said by email.

"Healthy heart habits may be even more important for African-Americans, for whom avoiding a heart attack is even more important given their worse outcomes after the event," Krumholz said.

After a heart attack, black patients typically don't live as long as whites - a racial difference that is starkest among the affluent - according to a new U.S. study.

Researchers evaluated data on more than 132,000 white heart attack patients and almost 9,000 black patients covered by Medicare, the government health program for the elderly and disabled. They used postal codes to assess income levels in patients' communities.

After 17 years of follow-up, the overall survival rate was 7.4 percent for white patients and 5.7 percent for black patients, according to the results published in Circulation, the journal of the American Heart Association.

On average, across all ages, white patients in low-income areas lived longer after a heart attack - about 5.6 years compared with 5.4 years for black patients. But in high-income communities, the gap widened to a life expectancy of seven years for white people and 6.3 years for black individuals.

"We found that socioeconomic status did not explain the racial disparities in life expectancy after a heart attack," lead study author Dr. Emily Bucholz of Boston Children's Hospital said by email.

"Contrary to common belief, this suggests that improving socioeconomic standing may improve outcomes for black and white patients globally but is unlikely to eliminate racial disparities in health," Bucholz added.

To see how race and class impact heart attack outcomes, Bucholz and colleagues reviewed health records collected from 1994 to 1996 for patients aged 65 to 90 years.

Just 6.3 percent of the patients were black, and only 6.8 percent lived in low-income communities, based on the typical household income in their postal codes.

Among white patients under 80, life expectancy was longest for patients in the most affluent neighborhoods and it got progressively shorter for middle-income and poor communities, the study found.

By contrast, life expectancy was similar for black patients residing in poor and middle-income communities across all ages. Only black patients under age 75 living in affluent areas had a survival advantage compared with their peers in less wealthy neighborhoods.

One shortcoming of the study is that it included a small proportion of black and poor patients, the authors acknowledge. It's also possible that using postal codes to assess income may have led to some instances where income levels were inflated or underestimated, the authors note.

It's possible that black patients living in affluent areas don't fare as well as white patients because they don't have the same amount of social support from their peers, said Dr. Joaquin Cigarroa, a cardiovascular medicine researcher at Oregon Health & Science University in Portland.

In poor neighborhoods, black patients may face additional challenges to surviving a heart attack, added Cigarroa, who wasn't involved in the study.

"They more often live in low socioeconomic segments of our community that often have less access to health care resources and less access to stores with good nutrition," Cigarroa said by email. "In addition, these segments of our community are often not ideally configured for promoting physical activity with parks, sidewalks, bike lanes, etc."

The study findings highlight a need to improve outcomes among poor and black patients and suggest some differences in heart attack survival may come down to disparities in quality of care, said senior study author Dr. Harlan Krumholz of Yale University School of Medicine in New Haven, Connecticut.

Because black patients have a greater burden of heart disease than white people, doctors may also need to focus more on prevention in this community, Krumholz said by email.

"Healthy heart habits may be even more important for African-Americans, for whom avoiding a heart attack is even more important given their worse outcomes after the event," Krumholz said.

NEW YORK - Endovascular repair (EVAR) of abdominal aortic aneurysms (AAAs) is associated with an initial survival advantage over open repair, according to a study of "real-world" data from California.

However, the difference disappears in the long term, researchers report in JAMA Surgery, online September 2.

Dr. David C. Chang of Harvard Medical School in Boston said by email that the study "highlights the importance of looking at real-world data in evaluating surgical options. Clinically, our study found that the survival advantage for EVAR repairs is maintained until 3 years, after which mortality was higher for patients who had EVAR repairs."

The team studied more than 23,000 patients who underwent AAA repair between 2001 and 2009. Just over half had EVAR while the remaining patients underwent open repair. Median follow-up was for three years.

EVAR was associated with improved 30-day all-cause mortality (1.54% vs. 4.74%) and significantly improved survival until three years postoperatively. After that mortality rose, and the researchers found no difference in long-term mortality for the entire cohort after adjusting for confounders (hazard ratio, 0.99; p=0.64).

EVAR was linked with a significantly higher rate of reinterventions and AAA late ruptures. At five years, for instance, the reintervention rate was 6.59% in the EVAR group vs. 1.48% in the open group.

"This is different from data from clinical trials," Dr. Chang pointed out. "The short-term survival advantage of EVAR from clinical trials data likely eroded as patient risk factors exact their toll over time. These are real-world issues and concerns that are often not captured in idealized clinical trials."

Senior author Dr. Samuel E. Wilson, of the University of California-Irvine Medical Center, added by email that EVAR is safer than the open procedure, which it has replaced. The mortality advantage last for three years, "then other morbidity, especially effects of smoking, even out survival."

Dr. Chang went on to note that "our use of data from the State of California also has an important policy implication: That many states actually have better and more complete population data than the federal government when it comes to healthcare quality. While research and policies related to healthcare quality are driven mostly by the federal government

currently (through Medicare), the federal government has limited data on patient care outside of Medicare."

"Therefore," Dr. Chang concluded, "an argument can be made that the federal government should delegate healthcare research and quality improvement responsibilities to individual states, and support state-level efforts to examine and improve healthcare quality. Healthcare, like politics, is all local."

In an accompanying editorial, Drs. Jamie E. Anderson and James W. Holcroft, of the University of California Davis Medical Center, Sacramento, observe that the study "offers a glimpse into the future of population-based health services research methods."

In a joint email, they said, "Harnessing information already captured for patient care or billing purposes to advance medical research makes sense."

NEW YORK - Endovascular repair (EVAR) of abdominal aortic aneurysms (AAAs) is associated with an initial survival advantage over open repair, according to a study of "real-world" data from California.

However, the difference disappears in the long term, researchers report in JAMA Surgery, online September 2.

Dr. David C. Chang of Harvard Medical School in Boston said by email that the study "highlights the importance of looking at real-world data in evaluating surgical options. Clinically, our study found that the survival advantage for EVAR repairs is maintained until 3 years, after which mortality was higher for patients who had EVAR repairs."

The team studied more than 23,000 patients who underwent AAA repair between 2001 and 2009. Just over half had EVAR while the remaining patients underwent open repair. Median follow-up was for three years.

EVAR was associated with improved 30-day all-cause mortality (1.54% vs. 4.74%) and significantly improved survival until three years postoperatively. After that mortality rose, and the researchers found no difference in long-term mortality for the entire cohort after adjusting for confounders (hazard ratio, 0.99; p=0.64).

EVAR was linked with a significantly higher rate of reinterventions and AAA late ruptures. At five years, for instance, the reintervention rate was 6.59% in the EVAR group vs. 1.48% in the open group.

"This is different from data from clinical trials," Dr. Chang pointed out. "The short-term survival advantage of EVAR from clinical trials data likely eroded as patient risk factors exact their toll over time. These are real-world issues and concerns that are often not captured in idealized clinical trials."

Senior author Dr. Samuel E. Wilson, of the University of California-Irvine Medical Center, added by email that EVAR is safer than the open procedure, which it has replaced. The mortality advantage last for three years, "then other morbidity, especially effects of smoking, even out survival."

Dr. Chang went on to note that "our use of data from the State of California also has an important policy implication: That many states actually have better and more complete population data than the federal government when it comes to healthcare quality. While research and policies related to healthcare quality are driven mostly by the federal government

currently (through Medicare), the federal government has limited data on patient care outside of Medicare."

"Therefore," Dr. Chang concluded, "an argument can be made that the federal government should delegate healthcare research and quality improvement responsibilities to individual states, and support state-level efforts to examine and improve healthcare quality. Healthcare, like politics, is all local."

In an accompanying editorial, Drs. Jamie E. Anderson and James W. Holcroft, of the University of California Davis Medical Center, Sacramento, observe that the study "offers a glimpse into the future of population-based health services research methods."

In a joint email, they said, "Harnessing information already captured for patient care or billing purposes to advance medical research makes sense."

NEW YORK - Endovascular repair (EVAR) of abdominal aortic aneurysms (AAAs) is associated with an initial survival advantage over open repair, according to a study of "real-world" data from California.

However, the difference disappears in the long term, researchers report in JAMA Surgery, online September 2.

Dr. David C. Chang of Harvard Medical School in Boston said by email that the study "highlights the importance of looking at real-world data in evaluating surgical options. Clinically, our study found that the survival advantage for EVAR repairs is maintained until 3 years, after which mortality was higher for patients who had EVAR repairs."

The team studied more than 23,000 patients who underwent AAA repair between 2001 and 2009. Just over half had EVAR while the remaining patients underwent open repair. Median follow-up was for three years.

EVAR was associated with improved 30-day all-cause mortality (1.54% vs. 4.74%) and significantly improved survival until three years postoperatively. After that mortality rose, and the researchers found no difference in long-term mortality for the entire cohort after adjusting for confounders (hazard ratio, 0.99; p=0.64).

EVAR was linked with a significantly higher rate of reinterventions and AAA late ruptures. At five years, for instance, the reintervention rate was 6.59% in the EVAR group vs. 1.48% in the open group.

"This is different from data from clinical trials," Dr. Chang pointed out. "The short-term survival advantage of EVAR from clinical trials data likely eroded as patient risk factors exact their toll over time. These are real-world issues and concerns that are often not captured in idealized clinical trials."

Senior author Dr. Samuel E. Wilson, of the University of California-Irvine Medical Center, added by email that EVAR is safer than the open procedure, which it has replaced. The mortality advantage last for three years, "then other morbidity, especially effects of smoking, even out survival."

Dr. Chang went on to note that "our use of data from the State of California also has an important policy implication: That many states actually have better and more complete population data than the federal government when it comes to healthcare quality. While research and policies related to healthcare quality are driven mostly by the federal government

currently (through Medicare), the federal government has limited data on patient care outside of Medicare."

"Therefore," Dr. Chang concluded, "an argument can be made that the federal government should delegate healthcare research and quality improvement responsibilities to individual states, and support state-level efforts to examine and improve healthcare quality. Healthcare, like politics, is all local."

In an accompanying editorial, Drs. Jamie E. Anderson and James W. Holcroft, of the University of California Davis Medical Center, Sacramento, observe that the study "offers a glimpse into the future of population-based health services research methods."

In a joint email, they said, "Harnessing information already captured for patient care or billing purposes to advance medical research makes sense."

Diabetes affects up to 14 percent of the U.S. population - an increase from nearly 10 percent in the early 1990s - yet over a third of cases still go undiagnosed, according to a new analysis.

Screening seems to be catching more cases, accounting for the general rise over two decades, the study authors say, but mainly whites have benefited; for Hispanic and Asian people in particular, more than half of cases go undetected.

"We need to better educate people on the risk factors for diabetes - including older age, family history and obesity - and improve screening for those at high risk," lead study author Andy Menke, an epidemiologist at Social and Scientific Systems in Silver Spring, Maryland, said by email.

Globally, about one in nine adults has diagnosed diabetes, and the disease will be the seventh leading cause of death by 2030, according to the World Health Organization.

Most of these people have Type 2, or adult-onset, diabetes, which happens when the body can't properly use or make enough of the hormone insulin to convert blood sugar into energy. Left untreated, diabetes can lead to nerve damage, amputations, blindness, heart disease and strokes.

Average blood sugar levels over the course of several months can be estimated by measuring changes to the hemoglobin molecule in red blood cells. The hemoglobin A1c test measures the percentage of hemoglobin - the protein in red blood cells that

carries oxygen - that is coated with sugar, with readings of 6.5 percent or above signaling diabetes.

People with A1c levels between 5.7 percent and 6.4 percent aren't diabetic, but because this is considered elevated it is sometimes called "pre-diabetes" and considered a risk factor for going on to develop full-blown diabetes.

Menke and colleagues estimated the prevalence of diabetes and pre-diabetes using data from the National Health and Nutrition Examination Survey (NHANES) collected on 2,781 adults in 2011 to 2012 and an additional 23,634 adults from 1988 to 2010.

While the prevalence of diabetes increased over time in the overall population, gains were more pronounced among racial and ethnic minorities, the study found.

About 11 percent of white people have diabetes, the researchers calculated, compared with 22 percent of non-Hispanic black participants, 21 percent of Asians and 23 percent of Hispanics.

Among Asians, 51 percent of those with diabetes were unaware of it, and the same was true for 49 percent of Hispanic people with the condition.

An additional 38 percent of adults fell into the pre-diabetes category. Added to the prevalence of diabetes, that means more than half of the U.S. population has diabetes or is at increased risk for it, the authors point out.

The good news, however, is fewer people are undiagnosed than in the past, Dr. William Herman and Dr. Amy Rothberg of the University of Michigan in Ann Arbor note in commentary accompanying the study in JAMA.

In it, they note that the increase in diabetes prevalence between 1988 and 2012 seen in the study was due to an increase in diagnosed cases, and that overall undiagnosed cases fell from 40 percent in 1988-1994 to 31 percent in 2008-2012.

This "likely reflects increased awareness of the problem of undiagnosed diabetes and increased testing," they said by email.

The drop in undiagnosed cases, they added, may be due in part to the newer, simpler A1c test, which doesn't require fasting or any advance preparation.

It's also possible that new cases of diabetes are starting to fall for the first time in decades because more people are getting the message about lifestyle choices that can contribute to diabetes, noted Dr. David Nathan, director of the diabetes center at Massachusetts General Hospital in Boston and a professor at Harvard Medical School.

In particular, more patients now understand that being overweight or obese increases the risk for diabetes, Nathan, author of a separate report in JAMA on advances in diagnosis and treatment, said by email.

"Behavioral changes, including healthy eating and more activity can prevent, or at least ameliorate, the diabetes epidemic," Nathan said.

Diabetes affects up to 14 percent of the U.S. population - an increase from nearly 10 percent in the early 1990s - yet over a third of cases still go undiagnosed, according to a new analysis.

Screening seems to be catching more cases, accounting for the general rise over two decades, the study authors say, but mainly whites have benefited; for Hispanic and Asian people in particular, more than half of cases go undetected.

"We need to better educate people on the risk factors for diabetes - including older age, family history and obesity - and improve screening for those at high risk," lead study author Andy Menke, an epidemiologist at Social and Scientific Systems in Silver Spring, Maryland, said by email.

Globally, about one in nine adults has diagnosed diabetes, and the disease will be the seventh leading cause of death by 2030, according to the World Health Organization.

Most of these people have Type 2, or adult-onset, diabetes, which happens when the body can't properly use or make enough of the hormone insulin to convert blood sugar into energy. Left untreated, diabetes can lead to nerve damage, amputations, blindness, heart disease and strokes.

Average blood sugar levels over the course of several months can be estimated by measuring changes to the hemoglobin molecule in red blood cells. The hemoglobin A1c test measures the percentage of hemoglobin - the protein in red blood cells that

carries oxygen - that is coated with sugar, with readings of 6.5 percent or above signaling diabetes.

People with A1c levels between 5.7 percent and 6.4 percent aren't diabetic, but because this is considered elevated it is sometimes called "pre-diabetes" and considered a risk factor for going on to develop full-blown diabetes.

Menke and colleagues estimated the prevalence of diabetes and pre-diabetes using data from the National Health and Nutrition Examination Survey (NHANES) collected on 2,781 adults in 2011 to 2012 and an additional 23,634 adults from 1988 to 2010.

While the prevalence of diabetes increased over time in the overall population, gains were more pronounced among racial and ethnic minorities, the study found.

About 11 percent of white people have diabetes, the researchers calculated, compared with 22 percent of non-Hispanic black participants, 21 percent of Asians and 23 percent of Hispanics.

Among Asians, 51 percent of those with diabetes were unaware of it, and the same was true for 49 percent of Hispanic people with the condition.

An additional 38 percent of adults fell into the pre-diabetes category. Added to the prevalence of diabetes, that means more than half of the U.S. population has diabetes or is at increased risk for it, the authors point out.

The good news, however, is fewer people are undiagnosed than in the past, Dr. William Herman and Dr. Amy Rothberg of the University of Michigan in Ann Arbor note in commentary accompanying the study in JAMA.

In it, they note that the increase in diabetes prevalence between 1988 and 2012 seen in the study was due to an increase in diagnosed cases, and that overall undiagnosed cases fell from 40 percent in 1988-1994 to 31 percent in 2008-2012.

This "likely reflects increased awareness of the problem of undiagnosed diabetes and increased testing," they said by email.

The drop in undiagnosed cases, they added, may be due in part to the newer, simpler A1c test, which doesn't require fasting or any advance preparation.

It's also possible that new cases of diabetes are starting to fall for the first time in decades because more people are getting the message about lifestyle choices that can contribute to diabetes, noted Dr. David Nathan, director of the diabetes center at Massachusetts General Hospital in Boston and a professor at Harvard Medical School.

In particular, more patients now understand that being overweight or obese increases the risk for diabetes, Nathan, author of a separate report in JAMA on advances in diagnosis and treatment, said by email.

"Behavioral changes, including healthy eating and more activity can prevent, or at least ameliorate, the diabetes epidemic," Nathan said.

Diabetes affects up to 14 percent of the U.S. population - an increase from nearly 10 percent in the early 1990s - yet over a third of cases still go undiagnosed, according to a new analysis.

Screening seems to be catching more cases, accounting for the general rise over two decades, the study authors say, but mainly whites have benefited; for Hispanic and Asian people in particular, more than half of cases go undetected.

"We need to better educate people on the risk factors for diabetes - including older age, family history and obesity - and improve screening for those at high risk," lead study author Andy Menke, an epidemiologist at Social and Scientific Systems in Silver Spring, Maryland, said by email.

Globally, about one in nine adults has diagnosed diabetes, and the disease will be the seventh leading cause of death by 2030, according to the World Health Organization.

Most of these people have Type 2, or adult-onset, diabetes, which happens when the body can't properly use or make enough of the hormone insulin to convert blood sugar into energy. Left untreated, diabetes can lead to nerve damage, amputations, blindness, heart disease and strokes.

Average blood sugar levels over the course of several months can be estimated by measuring changes to the hemoglobin molecule in red blood cells. The hemoglobin A1c test measures the percentage of hemoglobin - the protein in red blood cells that

carries oxygen - that is coated with sugar, with readings of 6.5 percent or above signaling diabetes.

People with A1c levels between 5.7 percent and 6.4 percent aren't diabetic, but because this is considered elevated it is sometimes called "pre-diabetes" and considered a risk factor for going on to develop full-blown diabetes.

Menke and colleagues estimated the prevalence of diabetes and pre-diabetes using data from the National Health and Nutrition Examination Survey (NHANES) collected on 2,781 adults in 2011 to 2012 and an additional 23,634 adults from 1988 to 2010.

While the prevalence of diabetes increased over time in the overall population, gains were more pronounced among racial and ethnic minorities, the study found.

About 11 percent of white people have diabetes, the researchers calculated, compared with 22 percent of non-Hispanic black participants, 21 percent of Asians and 23 percent of Hispanics.

Among Asians, 51 percent of those with diabetes were unaware of it, and the same was true for 49 percent of Hispanic people with the condition.

An additional 38 percent of adults fell into the pre-diabetes category. Added to the prevalence of diabetes, that means more than half of the U.S. population has diabetes or is at increased risk for it, the authors point out.

The good news, however, is fewer people are undiagnosed than in the past, Dr. William Herman and Dr. Amy Rothberg of the University of Michigan in Ann Arbor note in commentary accompanying the study in JAMA.

In it, they note that the increase in diabetes prevalence between 1988 and 2012 seen in the study was due to an increase in diagnosed cases, and that overall undiagnosed cases fell from 40 percent in 1988-1994 to 31 percent in 2008-2012.

This "likely reflects increased awareness of the problem of undiagnosed diabetes and increased testing," they said by email.

The drop in undiagnosed cases, they added, may be due in part to the newer, simpler A1c test, which doesn't require fasting or any advance preparation.

It's also possible that new cases of diabetes are starting to fall for the first time in decades because more people are getting the message about lifestyle choices that can contribute to diabetes, noted Dr. David Nathan, director of the diabetes center at Massachusetts General Hospital in Boston and a professor at Harvard Medical School.

In particular, more patients now understand that being overweight or obese increases the risk for diabetes, Nathan, author of a separate report in JAMA on advances in diagnosis and treatment, said by email.

"Behavioral changes, including healthy eating and more activity can prevent, or at least ameliorate, the diabetes epidemic," Nathan said.

NEW YORK - A variant of the CHADS2 score that's used to estimate ischemic stroke risk in patients with atrial fibrillation (AF) is also modestly accurate in heart failure patients, even in those without AF, researchers say. The variant, CHA2DS2-VASc, calculates stroke risk based on 10 possible points with higher scores indicating higher risk.

Line Melgaard from Aalborg University in Denmark and colleagues used three Danish nationwide registries to investigate whether the CHA2DS2-VASc score could predict ischemic stroke, thromboembolism, and death in patients with heart failure without AF as effectively as it does in patients with AF.

Patients with heart failure had a high risk of all three outcomes, whether or not AF was present, and the CHA2DS2-VASc score modestly predicted these endpoints at one-year and five-year follow-up (C statistics, 0.67 and 0.69, respectively).

Heart failure patients without AF whose CHA2DS2-VASc score was 4 or higher had increased risks of ischemic stroke, thromboembolism, and death in a manner comparable to patients with AF, according to the August 30 JAMA online report.

The negative predictive value (NPV) was around 90% at one-year follow-up for all three outcomes, although NPVs were strongly attenuated by the five-year follow-up.

"In our study, one of our principal findings was that the absolute risk of ischemic stroke among patients without AF was about 1.5% per year or higher with CHA2DS2-VASc scores of 2 or higher, with associated five-year absolute ischemic stroke risks in excess of 4% or more," the researchers noted. This risk level would be sufficient to prompt initiation of long-term anticoagulation in patients with AF, they say.

"The poor prognosis of atrial fibrillation for ischemic stroke and death in patients with heart failure was evident in our study and expected," Melgaard said. "But the observation that additional risk factors in patients with heart failure are particularly significant among those without atrial fibrillation is an important and (to some extent) unexpected result."

"I hope physicians will recognize that patients with heart failure and sinus rhythm have an increased risk of ischemic stroke, and that some subgroups within this population most likely need thromboprophylaxis," Melgaard concluded. "Especially patients with multiple comorbidities (high CHA2DS2-VASc score) need attention in the clinic."

NEW YORK - A variant of the CHADS2 score that's used to estimate ischemic stroke risk in patients with atrial fibrillation (AF) is also modestly accurate in heart failure patients, even in those without AF, researchers say. The variant, CHA2DS2-VASc, calculates stroke risk based on 10 possible points with higher scores indicating higher risk.

Line Melgaard from Aalborg University in Denmark and colleagues used three Danish nationwide registries to investigate whether the CHA2DS2-VASc score could predict ischemic stroke, thromboembolism, and death in patients with heart failure without AF as effectively as it does in patients with AF.

Patients with heart failure had a high risk of all three outcomes, whether or not AF was present, and the CHA2DS2-VASc score modestly predicted these endpoints at one-year and five-year follow-up (C statistics, 0.67 and 0.69, respectively).

Heart failure patients without AF whose CHA2DS2-VASc score was 4 or higher had increased risks of ischemic stroke, thromboembolism, and death in a manner comparable to patients with AF, according to the August 30 JAMA online report.

The negative predictive value (NPV) was around 90% at one-year follow-up for all three outcomes, although NPVs were strongly attenuated by the five-year follow-up.

"In our study, one of our principal findings was that the absolute risk of ischemic stroke among patients without AF was about 1.5% per year or higher with CHA2DS2-VASc scores of 2 or higher, with associated five-year absolute ischemic stroke risks in excess of 4% or more," the researchers noted. This risk level would be sufficient to prompt initiation of long-term anticoagulation in patients with AF, they say.

"The poor prognosis of atrial fibrillation for ischemic stroke and death in patients with heart failure was evident in our study and expected," Melgaard said. "But the observation that additional risk factors in patients with heart failure are particularly significant among those without atrial fibrillation is an important and (to some extent) unexpected result."

"I hope physicians will recognize that patients with heart failure and sinus rhythm have an increased risk of ischemic stroke, and that some subgroups within this population most likely need thromboprophylaxis," Melgaard concluded. "Especially patients with multiple comorbidities (high CHA2DS2-VASc score) need attention in the clinic."

NEW YORK - A variant of the CHADS2 score that's used to estimate ischemic stroke risk in patients with atrial fibrillation (AF) is also modestly accurate in heart failure patients, even in those without AF, researchers say. The variant, CHA2DS2-VASc, calculates stroke risk based on 10 possible points with higher scores indicating higher risk.

Line Melgaard from Aalborg University in Denmark and colleagues used three Danish nationwide registries to investigate whether the CHA2DS2-VASc score could predict ischemic stroke, thromboembolism, and death in patients with heart failure without AF as effectively as it does in patients with AF.

Patients with heart failure had a high risk of all three outcomes, whether or not AF was present, and the CHA2DS2-VASc score modestly predicted these endpoints at one-year and five-year follow-up (C statistics, 0.67 and 0.69, respectively).

Heart failure patients without AF whose CHA2DS2-VASc score was 4 or higher had increased risks of ischemic stroke, thromboembolism, and death in a manner comparable to patients with AF, according to the August 30 JAMA online report.

The negative predictive value (NPV) was around 90% at one-year follow-up for all three outcomes, although NPVs were strongly attenuated by the five-year follow-up.

"In our study, one of our principal findings was that the absolute risk of ischemic stroke among patients without AF was about 1.5% per year or higher with CHA2DS2-VASc scores of 2 or higher, with associated five-year absolute ischemic stroke risks in excess of 4% or more," the researchers noted. This risk level would be sufficient to prompt initiation of long-term anticoagulation in patients with AF, they say.

"The poor prognosis of atrial fibrillation for ischemic stroke and death in patients with heart failure was evident in our study and expected," Melgaard said. "But the observation that additional risk factors in patients with heart failure are particularly significant among those without atrial fibrillation is an important and (to some extent) unexpected result."

"I hope physicians will recognize that patients with heart failure and sinus rhythm have an increased risk of ischemic stroke, and that some subgroups within this population most likely need thromboprophylaxis," Melgaard concluded. "Especially patients with multiple comorbidities (high CHA2DS2-VASc score) need attention in the clinic."