The effect of surgical volume, age and comorbidities on 30-day mortality after radical prostatectomy: a population-based analysis of 9208 consecutive cases.

Link:
Autor/in:
Erscheinungsjahr:
2008
Medientyp:
Text
Beschreibung:
  • OBJECTIVES: To examine the effect of surgical volume (SV) on 30-day mortality after radical prostatectomy (RP; reportedly 0.1-0.6% and influenced by age and comorbidities) and to explore the most informative SV, age and comorbidity thresholds to distinguish between high- and low-risk men. PATIENTS AND METHODS: Between 1989 and 2000, 9208 consecutive patients were treated with RP. The effects on 30-day mortality of (either continuously coded or categorized) patient age, comorbidities (Charlson Comorbidity Index, CCI) and SV were tested in multivariable logistic regression models. The models were corrected for overfit bias using 200 bootstrap re-samples and were displayed graphically as nomograms. RESULTS: The overall 30-day mortality was 0.52%; being younger (27 RPs, 0.07 vs 0.6% otherwise, P = 0.049) had a protective effect and represented independent predictors of 30-day mortality. After correction for overfit bias, their combined input was 72.3% accurate in predicting 30-day mortality, vs 67.1% (P <0.001) when the same variables were used in continuously coded (uncategorized) format. A model-derived probability threshold of 27 RPs) can accurately identify patients at negligible risk of 30-day mortality.
  • OBJECTIVES: To examine the effect of surgical volume (SV) on 30-day mortality after radical prostatectomy (RP; reportedly 0.1-0.6% and influenced by age and comorbidities) and to explore the most informative SV, age and comorbidity thresholds to distinguish between high- and low-risk men. PATIENTS AND METHODS: Between 1989 and 2000, 9208 consecutive patients were treated with RP. The effects on 30-day mortality of (either continuously coded or categorized) patient age, comorbidities (Charlson Comorbidity Index, CCI) and SV were tested in multivariable logistic regression models. The models were corrected for overfit bias using 200 bootstrap re-samples and were displayed graphically as nomograms. RESULTS: The overall 30-day mortality was 0.52%; being younger (27 RPs, 0.07 vs 0.6% otherwise, P = 0.049) had a protective effect and represented independent predictors of 30-day mortality. After correction for overfit bias, their combined input was 72.3% accurate in predicting 30-day mortality, vs 67.1% (P <0.001) when the same variables were used in continuously coded (uncategorized) format. A model-derived probability threshold of 27 RPs) can accurately identify patients at negligible risk of 30-day mortality.
Lizenz:
  • info:eu-repo/semantics/restrictedAccess
Quellsystem:
Forschungsinformationssystem des UKE

Interne Metadaten
Quelldatensatz
oai:pure.atira.dk:publications/658ab3c3-2e50-41e7-86c5-925e545aeeee