Aims: To systematically evaluate the published evidence regarding the effectiveness of disease management programmes (DMPs) reducing hospital re-admissions among elderly patients with heart failure (HF).
Methods and results: Computerised search of MEDLINE (1966 to 31 August 2003) and EMBASE (1966 to 31 August 2003). The Cochrane Library was also searched, and reference lists of review articles on the topic, and of all relevant studies identified, were scanned. Search and selection of studies, data-extraction using standardised forms, and assessment of study quality was performed by two reviewers. The end-point was the proportion of persons who underwent hospital re-admission, and pooled relative risks (RR) were used to summarise the effectiveness of DMPs. The meta-analysis included 54 articles, comprising 27 randomised and 27 non-randomised controlled studies. Randomised studies consistently suggested that, in comparison with usual care, DMP reduced the frequency of re-admission for HF or cardiovascular disease by 30% (pooled RR 0.70; confidence interval (CI) 95% 0.62-0.79), all-cause re-admission by 12% (pooled RR 0.88, 95% CI: 0.79-0.97), and the combined event of re-admission or death by 18% (pooled RR 0.82, 95% CI: 0.72-0.94). The results displayed no substantial variation when only DMPs with home visits, out-patient visits to a clinic, or patient follow-up longer than 6 months were included. For DMPs with out-patient clinical visits, however, the reduction in re-admission for HF or cardiovascular disease, and for all causes, did not attain statistical significance. The magnitude of DMP benefits reported by non-randomised studies was more than double that reported by randomised studies. Practically all the non-randomised studies failed to control for confounding factors, such as severity, co-morbidity and drug therapy.
Conclusion: DMPs are effective at reducing re-admissions among elderly patients with HF. Their effectiveness is close to that observed in clinical trials evaluating drugs for HF, such as angiotensin-converting enzyme inhibitors, beta-blockers or digoxin. However, since none of the DMP studies compared different interventions directly, we do not know the relative effectiveness of types of healthcare delivery within the DMP.