Hoe vaak incontinentie na radicale prostatectomie?

Incidentie bepaald op basis van declaratie incontinentiemateriaal
Onderzoek
10-04-2018
Maike H.J. Schepens, Cathelijne M.P. Ziedses des Plantes, D.M. (Rik) Somford, Judith A. van Erkelens, Ruben G. Cremers, Sytske de Vries, Katja K.H. Aben, René Hoekstra,, Jozette J.C. Stienen, Bart P. Wijsman, Martijn B. Busstra en Jacques van Limbeek

Incidence of incontinence after radical prostatectomy using claims-based data

Objective

To determine the effect of radical prostatectomy (RP) hospital volume on the probability of post-RP incontinence.

Design

Retrospective research based on claims-based data of health insurers.

Method

For every patient with RP the probability of incontinence was determined, based on the definition of claims of one or more incontinence pads per day. Casemix corrections were made based on indicators available in claims-data: age, lymph node dissection, and radiotherapy. No casemix corrections could be made for tumour stage and surgical technique.

Predicted numbers of incontinent patients per hospital were calculated by adding the probability of incontinence for all patients. Per hospital, a comparison was made between the predicted incontinence rate and the actual incontinence rate. Finally, hospitals were categorised by the volume of RP conducted. We investigated which hospital category was associated with a significantly lower than expected risk of incontinence based on their casemix.

Results

A total of 1590 patients were included in this study; for 26.0% of these patients, an average of one or more incontinence pads per day were claimed for. A significant relation between the volume of RP per hospital and the claims of incontinence material was observed. The probability of incontinence was significantly lower in hospitals with a volume of more than 100 RP patients per year when compared to hospitals with less than 100 RP patients per year.

Conclusion

The probability of post-RP incontinence decreases as hospitals conduct more RP procedures. The casemix factors included in the analysis only had a limited impact on this observation.

Conflict of interest and financial support: none declared.