Reducing unnecessary hospital days to improve quality of care through physician accountability: a cluster randomised trial

Background: Over 20% of hospital bed use is inappropriate, implying a waste of resources and the increase of patient iatrogenic risk. Methods: This is a cluster, pragmatic, randomised controlled trial, carried out in a large University Hospital of Northern Italy, aiming to evaluate the effect of a strategy to reduce unnecessary hospital days. The primary outcome was the percentage of patient-days compatible with discharge. Among secondary objectives, to describe the strategy's effect in the long-term, as well as on hospital readmissions, considered to be a marker of the quality of hospital care. The 12 medical wards with the longest length of stay participated. Effectiveness was measured at the individual level on 3498 eligible patients during monthly index days. Patients admitted or discharged on index days, or with stay >90 days, were excluded. All ward staff was blinded to the index days, while staff in the control arm and data analysts were blinded to the trial's objectives and interventions. The strategy comprised the distribution to physicians of the list of their patients whose hospital stay was compatible with discharge according to a validated Delay Tool, and of physician length of stay profiles, followed by audits managed autonomously by the physicians of the ward. Results: During the 12 months of data collection, over 50% of patient-days were judged to be compatible with discharge. Delays were mainly due to problems with activities under medical staff control. Multivariate analysis considering clustering showed that the strategy reduced patient-days compatible with discharge by 16% in the intervention vs control group, (OR=0.841; 95% CI, 0.735 to 0.963; P=0.012). Follow-up at 1 year did not yield a statistically significant difference between the percentages of patient-days judged to be compatible with discharge between the two arms (OR=0.818; 95% CI, 0.476 to 1.405; P=0.47). There was no significant difference in 30-day readmission and mortality rates for all eligible patients (N=3498) between the two arms. Conclusions: Results indicate that a strategy, involving physician direct accountability, can reduce unnecessary hospital days. Relatively simple interventions, like the one assessed in this study, should be implemented in all hospitals with excessive lengths of stay, since unnecessary prolongation may be harmful to patients. Trial registration: ClinicalTrials.gov, identifier NCT01422811.