[Relationship between in-hospital mortality rate and patient volume for total hip arthroplasty in the Rhône-Alpes public hospitals].

BACKGROUND

Comparisons between hospitals using surgical volume, among other criteria can be made using large hospital databases. A relationship between hospital volume and care results, particularly mortality, is assumed to justify the use of activity volume as a comparison criterion. We aimed to assess a relationship between hospital volume and mortality after total hip replacement using the data available in the French Diagnosis Related Groups (Programme de Médicalisation des Systèmes d'Information, PMSI). We also searched for a threshold associated with mortality.

METHODS

We included all stays with a surgical procedure for total hip replacement in public hospitals in the Rhône-Alpes region in France in 1997 (n=5521). We examined the relationship between mortality and the number of procedures per hospital and assessed activity threshold using logistic regression.

RESULTS

Using the number of procedures as the continuous variable, we observed a relationship with mortality (OR=0.94 [0.91; 0.96] for an increase in activity of 100 total hip replacements). We found no evidence of an activity threshold. The reason for performing total hip replacement (OR=6.36 [2.78; 14.55] for trauma compared with rheumatology diseases) and patient age (OR=1.76 [1.31; 2.36] for a 10-year increase in age) were strongly related to mortality.

CONCLUSIONS

PMSI only collects in-hospital mortality, limiting the impact of our findings. The relationship between mortality and hospital volume is significant, but too small and consequently of little use. We found no activity threshold. It would be difficult to recommend surgical volume as a criterion for comparing hospitals.