Centralisation of specialist cancer surgery services in two areas of England: the RESPECT-21 mixed-methods evaluation

Abstract

Background

Centralising specialist cancer surgical services is an example of major system change. High-volume centres are recommended to improve specialist cancer surgery care and outcomes. Objective

Our aim was to use a mixed-methods approach to evaluate the centralisation of specialist surgery for prostate, bladder, renal and oesophago-gastric cancers in two areas of England [i.e. London Cancer (London, UK), which covers north-central London, north-east London and west Essex, and Greater Manchester Cancer (Manchester, UK), which covers Greater Manchester]. Design

Stakeholder preferences for centralising specialist cancer surgery were analysed using a discrete choice experiment, surveying cancer patients (n = 206), health-care professionals (n = 111) and the general public (n = 127). Quantitative analysis of impact on care, outcomes and cost-effectiveness used a controlled before-and-after design. Qualitative analysis of implementation and outcomes of change used a multisite case study design, analysing documents (n = 873), interviews (n = 212) and non-participant observations (n = 182). To understand how lessons apply in other contexts, we conducted an online workshop with stakeholders from a range of settings. A theory-based framework was used to synthesise these approaches. Results

Stakeholder preferences – patients, health-care professionals and the public had similar preferences, prioritising reduced risk of complications and death, and better access to specialist teams. Travel time was considered least important. Quantitative analysis (impact of change) – only London Cancer’s centralisations happened soon enough for analysis. These changes were associated with fewer surgeons doing more operations and reduced length of stay [prostate –0.44 (95% confidence interval –0.55 to –0.34) days; bladder –0.563 (95% confidence interval –4.30 to –0.83) days; renal –1.20 (95% confidence interval –1.57 to –0.82) days]. The centralisation meant that renal patients had an increased probability of receiving non-invasive surgery (0.05, 95% confidence interval 0.02 to 0.08). We found no evidence of impact on mortality or re-admissions, possibly because risk was already low pre-centralisation. London Cancer’s prostate, oesophago-gastric and bladder centralisations had medium probabilities (79%, 62% and 49%, respectively) of being cost-effective, and centralising renal services was not cost-effective (12% probability), at the £30,000/quality-adjusted life-year threshold. Qualitative analysis, implementation and outcomes – London Cancer’s provider-led network overcame local resistance by distributing leadership throughout the system. Important facilitators included consistent clinical leadership and transparent governance processes. Greater Manchester Cancer’s change leaders learned from history to deliver the oesophago-gastric centralisation. Greater Manchester Cancer’s urology centralisations were not implemented because of local concerns about the service model and local clinician disengagement. London Cancer’s network continued to develop post implementation. Consistent clinical leadership helped to build shared priorities and collaboration. Information technology difficulties had implications for interorganisational communication and how reliably data follow the patient. London Cancer’s bidding processes and hierarchical service model meant that staff reported feelings of loss and a perceived ‘us and them’ culture. Workshop – our findings resonated with workshop attendees, highlighting issues about change leadership, stakeholder collaboration and implications for future change and evaluation.