Measuring quality for public reporting of health provider quality: making it meaningful to patients.

Health Policy Research Institute, University of California-Irvine, 100 Theory, Suite 110, Irvine, CA 92697-5800, USA.
American Journal of Public Health (Impact Factor: 4.23). 12/2009; 100(2):264-9. DOI: 10.2105/AJPH.2008.153759
Source: PubMed

ABSTRACT Public quality reports of hospitals, health plans, and physicians are being used to promote efficiency and quality in the health care system. Shrinkage estimators have been proposed as superior measures of quality to be used in these reports because they offer more conservative and stable quality ranking of providers than traditional, nonshrinkage estimators. Adopting the perspective of a patient faced with choosing a local provider on the basis of publicly provided information, we examine the advantages and disadvantages of shrinkage and nonshrinkage estimators and contrast the information made available by them. We demonstrate that 2 properties of shrinkage estimators make them less useful than nonshrinkage estimators for patients making choices in their area of residence.


Available from: Laurent Glance, Jun 02, 2015
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Background Inappropriate use and overuse of antibiotics is a serious concern in the treatment of upper respiratory tract infections (URTIs), especially in developing countries. In recent decades, information disclosure and public reporting (PR) has become an instrument for encouraging good practice in healthcare. This study evaluated the impact of PR on antibiotic prescribing for URTIs in a sample of primary care institutions in China. Methods A matched-pair cluster-randomized trial was undertaken in QJ city, with 20 primary care institutions participating in the trial. Participating institutions were matched into pairs before being randomly assigned into a control and an intervention group. Prescription statistics were disclosed to patients, health authorities, and health workers monthly within the intervention group, starting from October 2013. Outpatient prescriptions for URTIs were collected from both groups before (1st March to 31st May, 2013) and after the intervention (1st March to 31st May, 2014). A total of 34,815 URTI prescriptions were included in a difference-in-difference analysis using multivariate linear or logistic regression models, controlling for patient attributes as well as institutional characteristics. Results Overall, 90% URTI prescriptions required antibiotics and 21% required combined use of antibiotics. More than 77% of URTI prescriptions required intravenous (IV) injection or infusion of drugs. PR resulted in a 9 percentage point (95% CI -17 to -1) reduction in the use of oral antibiotics (adjusted RR =39%, P =0.027), while the use of injectable antibiotics remained unchanged. PR led to a 7 percentage point reduction (95% CI -14 to 0; adjusted RR =36%) in combined use of antibiotics (P =0.049), which was largely driven by a significant reduction in male patients (-7.5%, 95% CI -14 to -1, P =0.03). The intervention had little impact on the use of IV injections or infusions, or the total prescription expenditure. Conclusions The results suggest that PR could improve prescribing practices in terms of reducing oral antibiotics and combined use of antibiotics; however, the impacts were limited. We suggest that PR would probably be enhanced by provider payment reform, management and training for providers, and health education for patients.
    Health Research Policy and Systems 10/2014; 12(1):61. DOI:10.1186/1478-4505-12-61 · 1.86 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: One of every 150 hospitalized patients experiences a lethal adverse event; nearly half of these events involves surgical patients. Although variations in surgeon performance and quality have been reported in the literature, less is known about the influence of anesthesiologists on outcomes after major surgery. Our goal of this study was to determine whether there is significant variation in outcomes between anesthesiologists after controlling for patient case mix and hospital quality. Using clinical data from the New York State Cardiac Surgery Reporting System, we conducted a retrospective observational study of 7920 patients undergoing isolated coronary artery bypass graft surgery. Multivariable logistic regression modeling was used to examine the variation in death or major complications (Q-wave myocardial infarction, renal failure, stroke) across anesthesiologists, controlling for patient demographics, severity of disease, comorbidities, and hospital quality. Anesthesiologist performance was quantified using fixed-effects modeling. The variability across anesthesiologists was highly significant (P < 0.001). Patients managed by low-performance anesthesiologists (corresponding to the 25th percentile of the distribution of anesthesiologist risk-adjusted outcomes) experienced nearly twice the rate of death or serious complications (adjusted rate 3.33%; 95% confidence interval [CI], 3.09%-3.58%) as patients managed by high-performance anesthesiologists (corresponding to the 75th percentile) (adjusted rate 1.82%; 95% CI, 1.58%-2.10%). This performance gap was observed across all patient risk groups. The rate of death or major complications among patients undergoing coronary artery bypass graft surgery varies markedly across anesthesiologists. These findings suggest that there may be opportunities to improve perioperative management to improve outcomes among high-risk surgical patients.
    Anesthesia & Analgesia 03/2015; 120(3):526-33. DOI:10.1213/ANE.0000000000000522 · 3.42 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Background Risk adjusted mortality for intensive care units (ICU) is usually estimated via logistic regression. Random effects (RE) or hierarchical models have been advocated to estimate provider risk-adjusted mortality on the basis that standard estimators increase false outlier classification. The utility of fixed effects (FE) estimators (separate ICU-specific intercepts) has not been fully explored. Methods Using a cohort from the Australian and New Zealand Intensive Care Society Adult Patient Database, 2009–2010, the model fit of different logistic estimators (FE, random-intercept and random-coefficient) was characterised: Bayesian Information Criterion (BIC; lower values better), receiver-operator characteristic curve area (AUC) and Hosmer-Lemeshow (H-L) statistic. ICU standardised hospital mortality ratios (SMR) and 95%CI were compared between models. ICU site performance (FE), relative to the grand observation-weighted mean (GO-WM) on odds ratio (OR), risk ratio (RR) and probability scales were assessed using model-based average marginal effects (AME). Results The data set consisted of 145355 patients in 128 ICUs, years 2009 (47.5%) & 2010 (52.5%), with mean(SD) age 60.9(18.8) years, 56% male and ICU and hospital mortalities of 7.0% and 10.9% respectively. The FE model had a BIC = 64058, AUC = 0.90 and an H-L statistic P-value = 0.22. The best-fitting random-intercept model had a BIC = 64457, AUC = 0.90 and H-L statistic P-value = 0.32 and random-coefficient model, BIC = 64556, AUC = 0.90 and H-L statistic P-value = 0.28. Across ICUs and over years no outliers (SMR 95% CI excluding null-value = 1) were identified and no model difference in SMR spread or 95%CI span was demonstrated. Using AME (OR and RR scale), ICU site-specific estimates diverged from the GO-WM, and the effect spread decreased over calendar years. On the probability scale, a majority of ICUs demonstrated calendar year decrease, but in the for-profit sector, this trend was reversed. Conclusions The FE estimator had model advantage compared with conventional RE models. Using AME, between and over-year ICU site-effects were easily characterised.
    PLoS ONE 07/2014; 9(7):e102297. DOI:10.1371/journal.pone.0102297 · 3.53 Impact Factor