Journal of clinical epidemiology
-
To assess the accuracy of administrative claims data for measuring antibiotic prescribing behavior, we conducted a stratified randomized medical record review of office visits by children with pharyngitis, and adults with acute bronchitis, to primary care physicians in Colorado in 1998. The diagnoses of pharyngitis (n = 422) and acute bronchitis (n = 497) based on administrative data were verified in 83% and 79%, respectively, of medical records. The sensitivity, specificity, and positive predictive value of administrative data in identifying antibiotic treatment for pharyngitis was 68%, 91%, and 90%, respectively, and for bronchitis was 79%, 84% and 98%, respectively. ⋯ Absence of testing in administrative data (when present in the medical record) was more frequent among visits to physicians associated with a capitated health plan. We conclude that administrative claims data are accurate sources for measuring and profiling antibiotic prescribing practices in ambulatory practice, although they underestimate actual antibiotic treatment decisions by individual physicians. Measuring and profiling antibiotic prescribing behavior in relation to group A streptococcal test utilization may overestimate inappropriate antibiotic treatment by physicians enrolled in capitated contracts.
-
Publication and selection biases in meta-analysis are more likely to affect small studies, which also tend to be of lower methodological quality. This may lead to "small-study effects," where the smaller studies in a meta-analysis show larger treatment effects. Small-study effects may also arise because of between-trial heterogeneity. ⋯ Tests for small-study effects should routinely be performed in meta-analysis. Their power is however limited, particularly for moderate amounts of bias or meta-analyses based on a small number of small studies. When evidence of small-study effects is found, careful consideration should be given to possible explanations for these in the reporting of the meta-analysis.
-
Neck pain is the most frequently reported feature in connection with whiplash injury, but it is also a common complaint in the general population. Therefore it is crucial to include an unexposed comparison group when evaluating the association between neck pain and a previous motor vehicle crash (MVC). To determine whether exposure to a rear-end collision, without or with whiplash injury, is associated with future neck or shoulder pain, a cohort study was conducted. ⋯ The corresponding relative risk in subjects with whiplash injury was 2.7 (95% CI 2.1-3. 5). We conclude that there is no increased risk of future neck or shoulder pain in drivers who did not report whiplash injury in connection with a rear-end collision 7 years earlier. In drivers with reported whiplash injury, the risk of neck or shoulder pain 7 years after the collision was increased nearly three-fold compared with that in unexposed subjects.
-
There are no reliable, valid, and responsive scales to measure the quality of sedation in adult critically ill patients. Our objective was to develop a summated rating scale with these properties and to define the minimal clinically important difference (MCID). We developed and tested the scale in an 18-bed medical-surgical intensive care unit (ICU) (12-bed acute and 6-bed subacute unit). ⋯ To test construct validity, MCID, and responsiveness 302 observations were made on 54 patients. Construct validity: calmness score vs. need for further intervention to make the patient calm (R = -0.82, P < 0.001); interaction score discriminated between acute vs. subacute units, mean scores 15.28 +/- 8.26 vs. 23.54 +/- 7.42, mean difference 8.27 (95% CI - 10.32 to -6.22); MCID - 2.2 and 2.5 for the calmness and interaction subscales; Guyatt's responsiveness statistics - 1.4 and 2.3. The Vancouver Interaction and Calmness Scale (VICS) is reliable, valid, and responsive.