The impact of Community Health Centers on inappropriate use of emergency services
Lippi Bruni M., Ugolini C., Verzulli R., Leucci A.C., 2023 – Health Economics
The European health policy framework identifies community and primary care as strategic areas and underlines that strengthening territorial care through a coordinated approach can improve the efficiency, effectiveness, and responsiveness of healthcare systems. Although a common definition for integrated community care centers remains elusive – they are called Medical Homes, Patient-centered Medical Homes, Community Health Centers, etc. -, they share distinctive features including team-based care, integrated information technology systems, tools supporting clinical decisions such as population-based registries and the Chronic Care Model.
First developed in the United States to manage complex patients, similar initiatives are nowadays implemented worldwide. Examples can be found, in particular, in Canada and in several European countries. In the initial phase, the introduction of Community Health Centers (CHCs) was mainly motivated by growing concerns with the accelerating rate of healthcare spending, shifting demand away from acute hospital care in favor of less complex settings. Later, the focus moved from cost containment to quality improvement in territorial healthcare.
The Italian policy debate promoting the establishment of Community Health Centers (“Case della Salute”) starts in the years 2000s’. CHCs are seen as an investment to improve the effectiveness of treatments, especially for chronic patients, and to relieve congestion at the Emergency Departments (EDs). In the light of that, Regional Governments have launched large-scale programs, but the dissemination of the model has been gradual, depending on local infrastructures and funding. More recently, the PNRR has identified the development of CHCs as a national priority, promoting their transformation in “Case della Comunità”, thought of as public premises fostering a strict integration between health and social care. The new health policy design based on the OneHealth perspective aims at delivering integrated services where not only health but also social needs and community well-being are addressed.
Despite the prominence achieved in many institutional contexts, the empirical evidence of the impact of such programs on patients’ outcomes is surprisingly scant and, in most cases, it remains basically descriptive.
In a recent paper, “The impact of Community Health Centers on inappropriate use of emergency services”, published in Health Economics, Matteo Lippi Bruni, Cristina Ugolini, Rossella Verzulli and Anna Caterina Leucci aim at filling this gap by assessing the causal impact of General Practitioners’ (GP) participation into CHCs in the Italian NHS. The objective of the study is to test whether patients whose GP operates in a CHC display better outcomes compared to patients whose GP does not. The study focuses on Diabetes Mellitus type II and measures outcomes as inappropriate emergency admissions for minor conditions that should be treated in primary care. As diabetes treatment requires multiple processes and resources, these patients are well suited for studying the consequences of the new coordinated care system, because assistance relies strongly on GP’s controls and CHCs multidisciplinary teams are deemed to improve diabetes management.
The study uses an unbalanced panel of individual-level data covering the diabetic population of the Italian region Emilia-Romagna over years 2010–16 and employs a variety of Difference-in-Differences (DiD) estimators. At each point in time, patients whose GP operates in a CHC are assigned to the treatment group, while the control group consists of the patients registered with physicians working in traditional practices. The identification of the effect of the CHC model is challenged by the fact that GPs’ participation occurs on a voluntary basis. The authors employ multiple strategies to tackle such potential endogeneity bias. They use two-way fixed effects linear regression models as baseline, and evaluate the robustness of their estimates by using alternative DiD methods that take advantage of the staggered implementation of the program, also allowing for treatment effects heterogeneity.
The main findings suggest that CHC status significantly reduces the probability of inappropriate ED admissions compared with the traditional primary care model, although the magnitude of the effect is modest. The result holds after accounting for GPs’ self-selection into the program, controlling for patients’ and GPs’ characteristics, and allowing for treatment effect heterogeneity. The effect is driven by daytime visits on working days, whereas the finding is not confirmed for night and weekend admissions when CHCs are closed. This evidence reinforces the credibility of a causal link between CHC accessibility and a reduction in inappropriate utilisation of hospital emergency services. As the findings cannot be straightforwardly extended to other population groups, future work should include patients with other chronic conditions; with clinical data sources becoming available, it should be possible to examine a wider spectrum of the expected benefits of the policy to be weighed with the expected the costs.
The policy implications of the analysis suggest that investing in CHCs – especially those providing a wide array of integrated services – yields benefits in terms of higher appropriateness in the utilisation of different care settings. Nevertheless, even if the direction is in line with policymakers’ expectations, the magnitude of the effect remains small, thus suggesting a limited relief for ED overcrowding for diabetics. The development of primary care-based emergency services during nights and weekends could strengthen the effect further, provided that the prevailing mechanism in reducing inappropriate ED visits for diabetic patients is expanded access to high quality, integrated community care services, currently available through CHCs only at daytime in working days. With the CHC-model development process in its infancy, these achievements can be further enhanced in the future.