Siontis et al empirically evaluated the ability of validated predictive tools to predict death in diverse settings based on the area under the receiver operating characteristic curve (AUC). Across 240 assessments of predictive tools, AUC values ranged from no discrimination (AUC = 0.43) to excellent discrimination (AUC = 0.98). Known and widely validated tools showed considerable variability in their predictive performance. None of the tools examined in many studies had consistently high performance based on the AUC. The majority of the assessed tools did not have sufficiently documented clinical utility.
The Computerization of Medical Practices for the Enhancement of Therapeutic Effectiveness (COMPETE) III pragmatic randomized trial evaluated a complex intervention based on shared electronic decision support for reducing vascular risk in 1109 high-risk older adults in community-based primary care. While the processes of vascular care and patient satisfaction significantly improved, vascular events and quality of life were not improved over the 12-month follow-up period. This trial supports the growing concern over the lack of evidence of benefit for health information technologies on patient outcomes.
Cardiovascular disease (CVD) is the leading cause of morbidity and mortality in patients with diabetes. Current guidelines classify all diabetic individuals as high risk (10-year CVD risk of at least 20%) for incident CVD. Paynter et al compared this strategy with one using traditional risk factors and hemoglobin A1c for predicting the 10-year CVD risk of diabetic men and women enrolled in the Women's Health Study and Physician's Health Study II. In both men and women with diabetes, prediction of CVD risk was substantially improved by using models with risk factors and hemoglobin A1c. Larger improvements were seen in women, a population with a lower overall 10-year risk of CVD, compared with men.
In a retrospective cohort study of 83 625 persons with colorectal cancer (CRC) within the Medicare–Surveillance, Epidemiology, and End Results (SEER) database, Ferrante et al examined the association between the number of primary care visits in the 3- to 27-month period prior to CRC diagnosis and colorectal cancer outcomes. They found that a greater number of visits to primary care in the prediagnostic period was associated with CRC screening, earlier stage diagnosis, lower CRC mortality, and lower overall mortality. The authors suggest that revitalizing primary care in the United States may help strengthen the national efforts on reducing the burden of colorectal cancer.
This study evaluated sudden cardiac death (SCD) risk among the 2763 women with CAD who were enrolled in the Hormone and Estrogen Replacement Study (HERS). Over a mean follow-up of 6.8 years, SCD comprised 54% (136 events) of the cardiac-related deaths and 27% of all deaths. The following 6 predictors were independently associated with SCD and predicted incident events: myocardial infarction, heart failure, estimated glomerular filtration rate lower than 40 mL/min/1.73 m2, atrial fibrillation, physical inactivity, and diabetes. The incidences of SCD among women with 0 (n = 683), 1 (n = 1224), 2 (n = 610), and 3 or more (n = 246) risk factors at baseline were 0.3%, 0.5%, 1.2%, and 2.9% per year, respectively. The combination of clinical risk factors and left ventricular ejection function (C-index, 0.681) were better predictors of SCD than left ventricular ejection function alone (C-index, 0.600) and resulted in a 20% improvement in the net reclassification index (P < .001).
In This Issue of Archives of Internal Medicine. Arch Intern Med. 2011;171(19):1700. doi:10.1001/archinternmed.2011.487