Related Experiment Videos
[Description of long-term stays, based on PMSI data, in hospital specializing in short-term stays]
O Dalco1, P Gillois, G Chatellier
1Service d'Informatique médicale, Hôpital Broussais, Paris.
Objective:
In a context of organization of care where the budget of hospitals is a function of the number and of the severity of the in-patients and not of the duration of stay, stays of long duration (SLD) in short-stay hospitals represent a problem of both medical and administrative management. To identify the characteristics of long-duration stays.
Methods:
The data used in this retrospective study were drawn from the Standardized Discharge Summaries of the year 1997 of a University Hospital of the Paris area (France). A stay of long duration (SLD) was defined in an arbitrary way as a stay exceeding 30 days. The stays of long duration (> 30 days) were subdivided in "long stays" (from 31 to 60 days) and "very long stays" (more than 60 days).
Results:
The SLDs represent 3.7% of the discharge summaries of our hospital, among them, 40% are medical DRGs and 60% surgical DRGs. The patients in SLD more often come from other structures of care than the patients having a short duration stay of (SDS) coming mainly from their residence and were also hospitalised in several different units during their stay. Patients having a long stay were more often classified in DRGs outside the principal activity of this hospital (i.e. cardiovascular diseases).
Conclusion:
This first approach suggests that a set of simple descriptive variables (pre-existing and acquired co-morbidity, admission in surgical ward, multi-unit stay...) makes it possible to identify the patients likely to have a long duration stay. Simple variables added to the current hospital minimum medical record would make it possible to consider a predictive approach.