We’ve heard the expression “Hindsight is 20/20,” but that doesn’t mean we have to move forward blindly. Artificial intelligence (AI) and predictive analytics are increasingly helping us better anticipate the future and, from a population health perspective, are informing tools to aid in disease prevention and locate areas that need healthcare resources.
Think about the advantages of identifying areas with a high risk of behavioral health issues, mental health issues, or chronic diseases. Potential resources, time, and capital could be saved by having a preventive plan that will help direct these same funds more efficiently or allocate them for other severe issues. This helps not only health systems and organizations in dealing with population management but also companies in planning their strategic solutions for products. For example, identifying mental and behavioral health issues is as important in healthcare facility planning as incorporating space for patient surges and chronic diseases.
We are constantly scrambling for quick, inexpensive, and reliable data, whether it be deidentified demographic or clinical data. The real test for any analyst lies in being able to identify valuable and authentic data sources; most use creative methods of combining or scrapping data to compensate for the lack of complete datasets. While we learn the skills of using such innovative methods, we keep pondering the advantages of having access to clinical or behavioral health data that could ideally bolster the country’s health considerably, especially in current times.
With the current COVID-19 pandemic and realization of the usefulness of such analytics, the world is heading toward leveraging existing data to identify high-risk areas. Successful population health management solutions leverage a variety of data. However, the challenge most organizations face is finding updated data at a granular (zip code or block) level.
One dataset that is increasingly being used to improve health outcomes and achieve health equity is social determinants of health (SDOH). SDOH encompass five underlying factors: economic stability, education, social and community context, health and healthcare, and neighborhood and built environment. These five factors are highly interrelated; a lack of education could result in economic uncertainty in the future, whereas living in a neighborhood with socioeconomic and environmental barriers could result in higher rates of chronic diseases and reduced access to healthcare, which could in turn affect the ability to attend school. The root factors described in SDOH gives birth to multiple types of physical, mental, and even behavioral health disorders/issues. Hence, the best approach would be to utilize existing available data and try to identify at least one key aspect in these five categories to find the geographic areas that would most benefit from more community and healthcare investments. Thus, using data like county health ranking and zip code-level data, we were able to create a dashboard to help identify specific high-risk zones.
We created a dataset by combining multiple key variables and health indicators around demographics, quality of life, health behaviors, inpatient and outpatient care utilization, and mental health provider supply. Because behavioral health conditions affect populations across demographic and socioeconomic groups, varying weights were applied across the indicators to assign counties and zip codes a behavioral health score. This behavioral health score helps pinpoint the areas with potential issues. Similar dashboards and approaches could be used in the future for identifying chronic diseases, mental health issues, or any future epidemics.
