Measurement of cervical cancer quality indicators using a population-based linked dataset in New South Wales, Australia

Abstract

The aims of this study were to assess feasibility of measuring cervical cancer quality indicators (QIs) using a population-based linked dataset, assess the quality of care in the treatment of cervical cancer at a population level and to evaluate whether any sociodemographic variables were associated with quality of care. The primary data source was the NSW Clinical Cancer Registry, linked to NSW Cancer Registry, NSW Admitted Patient Data Collection and NSW Registry of Births, Deaths and Marriages. Published QIs in cervical cancer were identified through a literature search. Feasibility of measurement was assessed according to available variables. QIs were measured where possible. Logistic regression analyses were performed to assess the impact of sociodemographic variables on QIs. There were 527 patients, of whom 160 (30%) had surgery, 259 (49%) radiotherapy and 153 (29%) chemotherapy. Only 16 of 93 (17%) published QIs could be measured. Only two of eight QIs with published benchmarks were met – postoperative deaths and duration of radiotherapy. The proportion alive at 1,3 & 5 years were 87%, 71% and 64%. Socio-demographic factors had limited associations with process QIs but were associated with outcome QIs. Increasing age, Australian country-of-birth, higher stage and squamous histology were associated with poorer survival. Population-based linked datasets are of limited value in measuring quality of care. Granular data collection at the institutional or multidisciplinary team level would allow improved measurement of quality indicators. Despite some QIs not meeting target benchmarks, survival from cervical cancer in NSW was comparable to other jurisdictions. Only 17% of published cervical cancer quality indicators could be measured from population-level linked datasets. Socio-demographic factors had limited impact on process quality indicators but were associated with outcomes. Survival from cervical cancer was comparable to published literature. Population-based linked datasets are of limited value in measuring quality of care.