Case-mix adjustment to compare hospital performances regarding complications after cytoreductive surgery for ovarian cancer: a nationwide population-based study

Abstract

ObjectiveComplication rates after cytoreductive surgery are important quality indicators for hospitals that treat patients with advanced-stage ovarian cancer. Case-mix factors are patient and tumor characteristics that may influence hospital outcomes such as the complication rates. Currently, no case-mix adjustment model exists for complications after cytoreductive surgery; therefore, it is unclear whether hospitals are being compared correctly. This study aims to develop the first case-mix adjustment model for complications after surgery for advanced-stage ovarian cancer, enabling an accurate comparison between hospitals.MethodsThis population-based study included all patients undergoing cytoreductive surgery for advanced-stage ovarian cancer registered in the Netherlands in 2017–2019. Case-mix variables were identified and assessed using logistic regressions. The primary outcome was the composite outcome measure ‘complicated course’. Patients had a complicated course when at least one of the following criteria were met: (1) any complication combined with a prolonged length of hospital stay; (2) complication requiring reintervention; (3) any complication with a prolonged length of stay in the intensive care unit; or (4) 30-day mortality or in-hospital mortality during admission following surgery. Inter-hospital variation was analyzed using univariable and multivariable logistic regressions and visualized using funnel plots.ResultsA total of 1822 patients were included, of which 10.7% (n=195) had a complicated course. Comorbidity and tumor stage had a significant impact on complicated course rates in multivariable logistic regression. Inter-hospital variation was not significant for case-mix factors. Complicated course rates ranged between 2.2% and 29.1%, and case-mix adjusted observed/expected ratios ranged from 0.20 to 2.67 between hospitals. Three hospitals performed outside the confidence intervals for complicated course rates. These hospitals remained outliers after case-mix adjustment.ConclusionThere is variation between hospitals regarding complicated course rates after cytoreductive surgery for ovarian cancer in the Netherlands. While comorbidity and tumor stage significantly affected the complicated course rates, adjusting for case-mix factors did not significantly affect hospital outcomes. The limited impact of case-mix adjustment could be a result of the Dutch centralized healthcare model.

Population-Based Clinical Cancer Registration in Germany

“…However, recent publications indicate that the use of CR data for quality assessment at the HCP level is just beginning. In the Netherlands, hospital comparisons based on quality indicators have already been published for gynaecological cancers [29][30][31], lung cancer [32] or rectal cancer [33]. Such analyses, like ours, show that cancer care can be made measurable at the HCP level and that fair benchmarking of specific clinical aspects is possible.…”

Quality indicators in surgical oncology: systematic review of measures used to compare quality across hospitals

“…The authors of 22 studies (25%) reported feeding the results back to the groups (that is hospitals or regions) they studied (see Table S5 for details) 14 , 21 , 22 , 33 , 37 , 42–44 , 46 , 47 , 49 , 50 , 52 , 53 , 57 , 64 , 71 , 73 , 78 , 81 , 83 , 84 , whereas 67 studies (75%) made no mention of feeding the results back to the hospitals or regions studied. Those authors who did report providing feedback fed the results into quality improvement efforts at the regional (9; 39%) or national (6; 27%) level, sent the results to all studied groups (4; 18%), or both sent the results to all studied groups and fed the results into quality improvement efforts (3; 13%).…”