Using benchmarking to assess aspects of specialisation in healthcare

Background Nordic health systems face challenges like an aging population, rising care demands, and high technology costs, making resource prioritization crucial. Efficiency, cost-effectiveness, and value for money are key for sustained performance, but market imperfections, such as information asymmetry and lack of price transparency, hinder affordable quality services. Benchmarking helps identify strengths, share best practices, and highlight areas for improvement, especially when market signals are lacking. Economies of scope can offer cost advantages from combining related services, but their benefits in healthcare are complex and context dependent.

Aim This thesis makes a twofold contribution: methodological and empirical. It identifies and suggest solutions for data comparability and methodological challenges in international benchmarking of efficiency and productivity, helping future researchers to avoid potential pitfalls. Empirically, it tests these solutions in studies that applies benchmarking to understand the effects of specialisation on costs, outputs and efficiency within Nordic healthcare system. The specific objectives of the studies are:

• To target and provide recommendations for overcoming data-related and methodological challenges inherent in cross-country benchmarking of hospital efficiency and productivity (Study I).
• To assess the cost-efficiency of highly diversified Nordic university hospitals by including measures of patient care, teaching, and research (Study II).
• To analyse cost-efficiency of specialised versus diversified acute-care hospitals in the Nordics and the potential existence of economies of scope (Study III).
• To evaluate the cost-efficiency and productivity changes in Swedish cancer care during a period of government-mandated collaboration and specialisation (Study IV).

Material and methods Study I presents solutions to data comparability and methodological challenges in international benchmarking, based on analyses of Nordic national patient administration databases from the Nordic Hospital Comparison Study Group. Study II uses the pooled Nordic patient database (2002-2004) from the Nordic Hospital Comparison Study Group, covering costs for patient care, teaching, research, outpatient visits, case-mix weighted admissions and discharges, full-time interns and residents, and bibliometric outcomes from Sweden, Finland, Norway, and Denmark. Study III uses an updated version of the database (2008-2009), including patient care costs, case-mix weighted outpatient visits, and inpatient admissions and discharges, all based on a common grouper with consistent DRG logic across the four countries. Study IV uses patient care costs and DRG points for malignant cancer (ICD-10 codes C00- C99) from the Swedish Association of Local Authorities and Regions' cost per patient database (2012-2022).

In Study I, a descriptive method is used. In Study II, III, and IV the bootstrapped nonparametric Data Envelopment Analysis is applied to estimate bias-corrected cost-efficiency scores. In Studies II and IV, a second-stage regression analysis is performed on the efficiency scores using an ordinary least squares model. In Study III scope convexity is defined as the ratio of cost efficiency relative to specialised and diversified hospital frontiers, using a data-driven approach to differentiate between specialised and diversified hospitals based on surgical vs. medical care and outpatient vs. inpatient care. Sensitivity analysis includes variations in production technology assumptions. In Study IV productivity changes using the Malmquist Productivity Index are also calculated.

Results In all Nordic analyses, Finnish hospitals consistently rank highest in average bias-corrected cost efficiency with the least variation around the mean, followed by Norwegian hospitals and Danish hospitals, while Swedish hospitals rank the lowest. Addressing data comparability and methodological challenges reduces variance and make results more robust, but the ranking across countries remains unchanged (Study I).

The results of Study II shows that Finland achieves the highest average cost- efficiency score in the patient care production model, with scores ranging from 0.90 to 0.95, regardless of the production technology assumption. However, when including costs and outcomes related to medical teaching and clinical research are included, the statistically significant differences in average efficiency between countries disappear. The findings also reveal that the optimal hospital size for delivering patient care differs from the ideal size for delivering teaching and research outcomes and confirms that a high case-mix weight negatively co-varies with efficiency.

The results of Study III confirm that the country specific efficiency rankings remain consistent with previous studies. A greater specialisation of surgical activity tends to generate higher average efficiency than a more diversified mix of surgical and medical activities. At the same time, bias-corrected scope convexity measures indicate economies of scope at the top- performing hospitals in Sweden (1.18), Finland (1.06) and Denmark (1.04). The impact of greater specialisation in outpatient care compared to inpatient care on hospital efficiency is more ambiguous.

The productivity analysis in Study IV shows a significant decline in both cancer care delivery (24.1 percentage points), and total care delivery (27.6 percentage points) during the study period, despite the introduction of multiple National Cancer Strategies promoting collaboration and specialisation. The bias-corrected average efficiency for cancer care (0.68) is lower than for total care (0.72), with oncology departments at university hospitals presenting the lowest scores. In contrast, a higher degree of specialisation, reflected in a greater proportion of oncology activities compared to other services, is positively linked to efficiency. Results also indicate regional differences in efficiency. Additionally, a weak positive association is found between efficiency and regional cancer mortality.

Conclusions This thesis addresses challenges in data comparability and methodology in international benchmarking using the Nordic countries' national patient administration databases, given their similar welfare models. It further explores these solutions through empirical studies that apply benchmarking to evaluate the effects of specialisation, contributing to the limited literature on the application of scope economics theory in healthcare. A better understanding of economies of scope, especially in hospital care, can inform healthcare configuration decisions. Policymakers should balance promoting specialisation for efficiency and innovation with supporting diversification when it improves patient outcomes. Future research should focus on defining outputs by clinical specialties, not care settings, and include episodes of care in efficiency analyses.

List of scientific papers

I. Medin E, Häkkinen U, Linna M, Anthun KS, Kittelsen SA, Rehnberg C; EuroHOPE Study Group. International hospital productivity comparison: experiences from the Nordic countries. Health Policy 2013 Sep; 112(1-2):80-7. https://doi.org/10.1016/j.healthpol.2013.02.004

II. Medin E, Anthun KS, Häkkinen U, Kittelsen SA, Linna M, Magnussen J, Olsen K, Rehnberg C. Cost efficiency of university hospitals in the Nordic countries: a cross-country analysis. Eur J Health Econ. 2011 Dec; 12(6):509-19. https://doi.org/10.1007/s10198-010-0263-1

III. Medin E, Janlöv N, Anthun KS, Kittelsen SA, Häkkinen U, Rehnberg C. Economies of scope in Nordic acute hospitals. [Manuscript]

IV. Medin E, Rehnberg C, Janlov N. Cancer care efficiency and productivity in Sweden. [Manuscript]