Development and user-oriented visualization of health care composite indicators at district level: territorial health profiles.

Background: after the recent reform of territorial healthcare, districts have been designated within the Lombardy Regional Health System to coordinate and deliver territorial care. This entails the need of readily available information to measure the quality of provided healthcare, identify critical areas for improvement, monitor the balance between demand and supply of healthcare services.

Objectives: to present the development of a dynamic evaluation system of processes and outcomes resulting from the integration of territorial and hospital care, based on a set of composite indicators, called 'health profiles', and their visualization and release through a dedicated web platform. These summary measures aim to capture the complexity of a specific clinical area or population and easily convey it to health managers.

Methods: the definition of a reproducible process for the construction of composite indicators, having defined a theoretical framework, maps the potential indicators of the profile of interest in a matrix made up of health needs and healthcare quality dimensions, and selects them on the basis of desirable properties and statistical metrics. Single indicators are normalized in the range [0,1], weighted according to the value of their quality dimension and to their reliability (measured as intraclass correlation coefficient), and aggregated via a geometric mean. The result is the value of the health profile, a percentage ranging from 0 to 100.

Results: the 'cancer profile', related to the domain of cancer prevention and treatment, was developed as a case study to illustrate the methods and potential application of these composite indicators at the district level. The initial set of 37 candidate indicators investigated: adherence to organized screening; timeliness of first visits; efficacy and safety of treatment; and end-of-life care. Applying the methods described, 28 indicators were selected and used to compute the cancer profile for different territorial units. Four main user-oriented infographics were developed to convey the health profile, its individual indicators, and their variation over time and across territorial units. The cancer profile was calculated and graphically visualized for the years 2015-2022.

Conclusions: a method has been defined and implemented to build synthetic territorial indicators, called health profiles, that can be applied to other clinical areas, such as chronic conditions. The release of a dedicated web platform for the effective communication of the profiles is an important tool to support evidence-based public health decisions.