Multi-domain contracting of the non-physician primary care workforce and per-domain service output: a nationwide cross-sectional analysis of Polish National Health Fund data, 2022–2023

Abstract Background Combining community nursing, midwifery, and school medicine within a single provider organisation is widely advocated as a route to more comprehensive primary care, yet whether multi-domain provision is accompanied by differences in per-domain service output remains untested at population scale. Economic theories predict opposing effects: economies of scope favour joint provision, while specialisation may yield higher per-unit throughput. Methods A nationwide cross-sectional analysis was performed using administrative data of the Polish National Health Fund (NFZ) covering all contracts for community nursing (N), midwifery (M), and school medicine (S) in 2022 and 2023. The unit of observation was the unique provider-year ( n = 16,574; 8,302 in 2022 and 8,272 in 2023) from 8,511 providers across all 16 NFZ regional branches. Providers were classified into seven mutually exclusive service-domain combinations. Two primary outcomes were specified ex ante: nurse services per 1,000 nurse declarations and midwife services per 1,000 midwife declarations. Adjusted associations were estimated by negative binomial regression (NB2, dispersion estimated by maximum likelihood) with log(declarations) as offset, fixed effects for regional branch and year, mono-domain providers as reference categories, and standard errors clustered at the provider level. Results The nursing model included 11,753 provider-years (6,097 providers) and the midwifery model 8,333 provider-years (4,345 providers). Compared with N-only providers, nurse services per declaration were lower in N + M providers (adjusted IRR 0.77; 95% CI 0.71–0.85; p < 0.001) and in N​ + M + S ​providers (IRR 0.89; 95% CI 0.82–0.98; p = 0.016), whereas N + S providers did not differ (IRR 1.00; 95% CI 0.90–1.12). The pattern was stronger for midwifery: compared with M-only providers, N + M providers delivered 61% fewer midwife services per declaration (IRR 0.39; 95% CI 0.30–0.52; p < 0.001) and N + M + S providers 59% fewer (IRR 0.41; 95% CI 0.31–0.54; p < 0.001); the small M + S class did not differ significantly (IRR 1.28; 95% CI 0.56–2.94). Estimates were closely similar in a balanced panel of providers observed in both years. In analyses stratified by declaration count, the negative associations were concentrated among small and medium providers and attenuated towards the null in the largest tertile. Conclusions Multi-domain contracting, particularly the combination of community nursing with midwifery, is associated with substantially lower per-domain service output, a pattern consistent with specialisation effects. The data describe associations between contracted service configuration and domain-specific service rates, not substitution or specialisation as observed processes, and provider scale contributes to the pattern. The trade-off between organisational breadth and per-domain throughput deserves explicit consideration in benchmarking and payment design.