Journal of clinical epidemiology
-
There is no empirical evidence on the sensitivity and specificity of methods to identify the possible overuse and underuse of medical procedures. To estimate the sensitivity and specificity of the RAND/UCLA Appropriateness Method. Parallel three-way replication of the RAND/UCLA Appropriateness Method for each of two procedures, coronary revascularization and hysterectomy. ⋯ The sensitivity and specificity of detecting the underuse of coronary revascularization were 94% (92-95%) and 97% (96-98%), respectively. Past applications of the appropriateness method have overestimated the prevalence of the overuse of hysterectomy, underestimated the prevalence of the overuse of the coronary revascularization, and provided true estimates of the underuse of revascularization. The sensitivity and specificity of the RAND/UCLA Appropriateness Method vary according to the procedure assessed and appear to estimate the underuse of procedures more accurately than their overuse.
-
To link hospital administrative data and an electronic medical record at a children's hospital in order to identify children with cancer admitted for fever and neutropenia. Hospital administrative data concerning 13,374 inpatient and outpatient encounters were validated against and linked to clinical data stored in an electronic medical record. Queries of the linked databases identified children with fever and neutropenia. ⋯ The experimental strategy for case finding had a sensitivity of 73.1% (95% CI: 58.1, 88.3), specificity 99.6% (95% CI: 99.1, 100). If only administrative data such as diagnosis-related group and hospital service were used for case finding, both the sensitivity (P < 0.01) and specificity (P < 0.01) were significantly lower. Linking a children's hospital administrative data system with clinical data is feasible and can be utilized for specific case finding for a common and costly condition in children.
-
Asymmetry in funnel plots may indicate publication bias in meta-analysis, but the shape of the plot in the absence of bias depends on the choice of axes. We evaluated standard error, precision (inverse of standard error), variance, inverse of variance, sample size and log sample size (vertical axis) and log odds ratio, log risk ratio and risk difference (horizontal axis). Standard error is likely to be the best choice for the vertical axis: the expected shape in the absence of bias corresponds to a symmetrical funnel, straight lines to indicate 95% confidence intervals can be included and the plot emphasises smaller studies which are more prone to bias. ⋯ We found similar evidence for asymmetry and between trial variation in a sample of 78 published meta-analyses whether odds ratios or risk ratios were used on the horizontal axis. Different conclusions were reached for risk differences and this was related to increased between-trial variation. We conclude that funnel plots of meta-analyses should generally use standard error as the measure of study size and ratio measures of treatment effect.