Leader: Giovanni Fattore (BOCCONI); Other collaborator(s): Simone Ghislandi (BOCCONI), Irene Torrini (BOCCONI)
Task 3 will perform an elaboration of a comprehensive framework of different models of community care services in Italy (social and healthcare) and develop a data platform integrating individual level data on the use of hospital and other health service (hospital discharge records SDO, administrative dataset from Lombardy Region) with other data collecting information on the interventions provided to elderly patients by social and primary care providers in the community. .
In particular, the activities carried out for this task involve the construction of two datasets. The first, which is at an advanced stage, brings together national data at the Local Health Authority (LHA) level from Hospital Discharge Records (SDOs) and Municipalities’ balance sheets in order to analyze the relationship between local public spending on territorial social services and the number, length of stay and cost of avoidable hospitalizations (AHs), used as a proxy for the quality of primary care. AHs in fact are hospitalizations required for all those conditions (Ambulatory Care Sensitive Conditions, ACSC) for which it is found that timely and effective care at the territorial level can prevent the use of inpatient hospital services. This dataset is then enriched by other context variables derived from national public sources, such as ISTAT, Internal Revenue Service, National Strategy for Inner Areas.
The second dataset comes from the Region of Lombardy registers and was received on June, 2024. These collect individual-level information on all over-65 residents with chronic conditions related to inpatient and outpatient care, pharmaceuticals, and residential, home-based, and preventive care services. For each patient, various information related to the General Practitioner (GP) such as age, gender, years of experience, and health district of membership is also available. When information at the municipality level will be added to the dataset, it will make it possible to analyze the relationship between health service utilization and GP characteristics and how such a relationship depends on territorial characteristics.
The Task aims at providing a unique example on how to supplement the databases typically used by policy makers with more specific details on the relevant supply-side aspects affecting the provision of health and community care services to the elderly. Moreover, a single data platform that collects information on the use of health services by the elderly and the current delivery of community services is crucial for analyzing whether, and how, the needs of the elderly population are being met.
The working group of Task 3 has begun construction of the first dataset. In particular, individual-level information from hospital discharge records is being merged with municipality-level information such as: socio-demographic and economic characteristics of the municipality, characteristics of territorial and hospital health service supply, geographic characteristics, social spending.
The working group of Task 3 completed the construction of the first dataset and has started work on the second.
The construction of the second dataset is ongoing.
The construction of the second dataset is ongoing.
During the reference period, we focused on preparing the dataset, particularly on data cleaning and verifying the quality of the available information. We standardized the variables and corrected any errors or inconsistencies.
Brief description of the activities and of the intermediate results
During the reference period, we focused on preparing the dataset, with particular attention to data cleaning activities and verifying the quality of the available information. We standardized the variables and corrected any errors or inconsistencies. Subsequently, we conducted a preliminary descriptive analysis aimed at understanding the distribution of the main variables and the correlations among them. Finally, we performed regressions on various outcomes, including as covariates both physician-related and patient-related characteristics.
Therefore, during this period, we initiated a new study aimed at evaluating the impact of primary care on post-discharge outcomes for elderly patients hospitalized for heart failure in Lombardy. We focused on constructing a clean and consistent analytical dataset using administrative healthcare data from 2010 to 2023, including hospital discharge records, demographic information, clinical data, and outpatient service use.
After a thorough data cleaning phase, we conducted a descriptive analysis to explore the distribution of primary variables and understand initial patterns, especially in the use of ambulatory services and readmission rates. To address potential endogeneity in the relationship between primary care use and patient outcomes, we implemented a quasi-experimental design using an instrumental variable (IV) approach.
We constructed an instrument capturing supply-side variation in the territorial distribution of general practitioners (GPs) at the ASST level, adjusted for population age structure and standard GP capacity ratios. The instrument showed a strong and significant association with primary care utilization, confirming its relevance and supporting its validity.
We then estimated regressions using this instrument to identify the causal impact of primary care access on 30-day all-cause hospital readmissions. The results suggest that better access to primary care significantly reduces the likelihood of early readmission, underscoring the importance of effective primary care as a gatekeeping mechanism, especially for elderly patients with chronic conditions like heart failure.
Main policy, industrial and scientific implications
In Italy, General Practitioners (GPs) play a central role as gatekeepers to secondary care, coordinating patient care pathways and contributing to the containment of national healthcare system costs. Their ability to promptly detect patients’ needs and ensure continuity of care is crucial in preventing avoidable hospital admissions, particularly those related to Ambulatory Care Sensitive Conditions (ACSCs). ACSCs are a key indicator of the accessibility, quality, and effectiveness of primary care, especially among older adults. However, empirical evidence on the direct impact of GPs on such hospital outcomes remains limited and inconclusive.
The first study investigates how GP characteristics influence avoidable hospitalization outcomes: risk of planned or emergency admissions, length of stay, readmission rates, and in-hospital mortality. For the second paper we know that heart failure is one of the leading causes of hospitalization among individuals over 65 and is associated with high rates of readmission and post-discharge mortality. Managing these patients poses significant clinical challenges due to the frequent presence of comorbidities and complex care needs. In this context, primary care services—including GPs and outpatient care—are increasingly recognized as essential in improving outcomes after hospital discharge. Strong and accessible primary care can support better disease management, timely follow-up, and reduce avoidable readmissions, thereby improving survival and quality of life. This study focuses on Lombardy, Italy’s most populous and economically significant region, where the healthcare system is organized into Aziende Socio Sanitarie Territoriali (ASSTs), local health authorities responsible for delivering integrated health and social care services.
The construction of the first dataset is completed.