To Register      SMDM Homepage

Sunday, 17 October 2004

This presentation is part of: Poster Session - Public Health; Methodological Advances

SYSTEMATIC META-ANALYSIS ON THE EFFICACY OF DISEASE MANAGEMENT PROGRAMS IN PATIENTS WITH CONGESTIVE HEART FAILURE

Alexander Gohler, MD, Charitè Campus Virchow Klinikum, Humboldt Medical School, Division of Cardiology, Berlin, Germany, Rainer Dietz, MD, Charitè Campus Virchow Klinikum, Humboldt Medical School, Divison of Cardiology, Berlin, Germany, Karl Josef Osterziel, MD, Charitè Campus Virchow Klinikum, Humboldt Medical School, Divison in Cardiology, Berlin, Germany, and Uwe Siebert, MD, MPH, MSc, Massachusetts General Hospital, Harvard Medical School, Institute for Technology Assessment and Department of Radiology, Boston, MA.

Purpose: Hospital admission for CHF is an important public health problem. Although several randomized controlled trials (RCT) have successfully linked DMPs to improved outcomes and reduced readmission rates, most effects are statistically not significant and vary regarding their magnitude. We sought to (1) systematically combine the evidence on efficacy of disease management programs (DMP) in the treatment of congestive heart failure (CHF), (2) to identify reasons for the existing heterogeneity and (3) identify publication bias.

Methods: We performed a systematic MEDLINE research on RCTs investigating DMPs for CHF treatment from 1966-05/2004. We included all studies that were performed randomized, included the core curriculum of a DMP (i.e., patient education, medication optimization, follow-up after discharge), and reported mortality and hospitalization as outcomes. We performed a meta-analysis using random or fixed effects models depending on the statistical heterogeneity of effects and estimated the pooled relative risk (RR) with 95% confidence intervals (95%CI). We assessed effect heterogeneity using meta-regressions to identify the impact of covariates on the DMP effect size. Publication bias was assessed by inspection of funnel plots.

Results: Our analysis included 16 studies from 5 different countries with data from 2868 patients. A random effects model which compared DMPs vs. control groups yielded a pooled RR of 0.79 (95%CI 0.65-0.97) for mortality and of 0.87 (95%CI 0.79-0.95) for rehospitalization during DMP (see Fig. 1). Meta-regression analysis identified mean age, severity of disease (NYHA distribution), and duration of intervention as statistically significant variables explaining the heterogeneity. Funnel plot was asymmetric indicating a bias towards positive studies.

Conclusions: DMPs in the CHF treatment lead to a clinically relevant and statistically significant reduction of mortality and rehospitalization. Heterogeneity across studies could largely be explained by age, severity of disease, and duration of DMP. Our analysis may overestimate the true DMP effect, because of a potential publication bias.

 

Fig. 1: Forest plot for DMP effect on hospitalization, sorted by study weight


See more of Poster Session - Public Health; Methodological Advances
See more of The 26th Annual Meeting of the Society for Medical Decision Making (October 17-20, 2004)