Monitoring quality of care in men diagnosed with prostate cancer: Developing consensus quality indicators using modified-Delphi methodology

Sue M. Evans, Denise Lin, Dragan Ilic, Jeremy Millar, Declan Murphy, Joanne Dean

Research output: Contribution to journalArticleResearch


Objective: To develop a core set of clinical indicators to measure the quality of care provided to men with prostate cancer.
Design: A modified Delphi study involving interviews of key informants and two rounds of survey to obtain consensus on the indicator set.
Setting: Melbourne, Australia
Participants: n=20, including specialists involved in prostate cancer management (urology, radiation oncology, medical oncology, nursing,psychology, palliative care) epidemiologists, scientists, consumers and a policy advisor.
Intervention(s): A literature review was undertaken to identify potential quality indicators. Interviews were undertaken to ensure completeness of the set and explore potential for inclusion of novel indicators. Survey of Delphi panel participants was conducted to refine the list.
Main Outcome Measure(s): Items with panel agreement ≥ 60%for reliability and capacity to be objectively assessed and with a median validity score of ≥ 8 (scale ranged from 1 (not important) -9 (very important)).Results: Of the total 104 proposed indicators, the panel retained 4/20 structural indicators, 15/46 process measures and 7/37 outcome indicators.
Conclusions: Indicators that scored highly in validity, reliability and objectivity included documentation of clinical stage, PSA level at diagnosis, surgical outcomes (rates of death, wound infection/bacteraemia and positive surgical margin), traditional measures of quality-of-care (10- and 15- year clinical and/or biochemical disease free survival) and patient assessed post-treatment function using a validated survey instrument.
Original languageEnglish
Article number109
Number of pages6
JournalJournal of Epidemiology and Preventive Medicine
Issue number2
Publication statusPublished - Oct 2015

Cite this