Predictive Analytics in Healthcare

Since 2020, as we dealt with COVID-19, it was clear that the global population, but the Mexican people in particular, were also victims of a second pandemic: chronic non-transmissible diseases, with obesity and overweight considered the main threats to public health. This phenomenon is the result of public and private health systems that were focused on the treatment of an active condition, not on the prevention of the particular triggers that gave rise to specific diseases, even though the concept of preventive medicine dates back to 1880 and that specific programs for preventive medicine in the public healthcare system in Mexico were designed in 2001.

This relatively recent preventive approach in Mexico is clearly ineffective. It consists of gathering population data on paper-based documents by unprepared and heterogeneous staff with limited access to users/patients. After inputting the collected information in paper databases or digital tools — text processors and Excel-like documents — the government officials try to establish interventional measures, such as anticancer programs or educational initiatives, to improve nutritional status and encourage physical activity among the general population. With this set of databases, institutions practice some sort of statistical analysis and deliver projections with known results: almost 70 percent of people in Mexico struggle with overweight and obesity, cardiovascular diseases are the leading cause of mortality after COVID-19 and diabetes mellitus affects one in every five citizens in the country, of whom some may not even know they are carriers of the disease. In summary, inaccurate data is used to build a retrospective analysis and generate prospective and potentially harmful epidemiological models.

The advent of digital technology, particularly in the last 15 years, has demonstrated that a new method for data capture and analysis is fundamental for organizations to thrive. The impact of this approach will make the implementation of public policy in healthcare and other industries more efficient. It will also have an effect in the private sector as well. This new point of view in data analysis and capture is not hindsight anymore; the new digital systems are looking for real-time data that could offer a priori models of how the tendencies and patterns of the analyzed datasets will probably occur. The expected result is to have accurate predictions of events to allow for planning and structuring of comparative and accountable strategies that should be updated, with sophisticated statistical analysis and support.  This is what predictive analytics is all about.

Who has not been surprised when opening the Amazon app to see immediate access to a book that you have always wanted even though you were not looking for it at the time?  This is a simple example of how a group of search inquiries, keywords, user characteristics and consumption patterns of individuals with similar archetypes can lead to a non-anticipated purchase. It is not a coincidence; it is the result of complex analytical systems and computer-based algorithms integrated in multilayer neural networks that support the processing of millions of bits of information.

Predictive analytics is more evident in some industries than others, but nowadays, it is present in almost all activities in our daily lives. It is clear that in e-commerce, it is more often used for normal citizens, but it is also needed in logistics, communications and even weather prediction.

Within this context, a big opportunity for predictive analytics in healthcare becomes evident.  There is a profound need for real-time data with analytics through cloud computing and machine learning that could allow us to identify patterns and tendencies that support the decisions and fulfillment of strategic health-related KPIs. With this integration of analytical tools and datasets, we could see a great impact on operational and financial efficiency in the public healthcare sector, replacing discretionary decisions with more accountable public systems. On an individual level, we could have the capacity to solve the complex balance between health and sickness considering all the variables, from genetic to environmental, and impacting how healthcare is provided at the personal level. In integrating medical records into other clinical variables, healthcare organizations could anticipate adverse events and make proper arrangements to avoid them. Predictive tools could collect and integrate data from lifestyles, symptoms, therapies and individual responses to develop new models of healthcare providers that could impact the supply chain, stock, human resources, infrastructure and administration to improve efficiency and user experience for both patients and collaborators. Last, but no least, is the possibility of personalized health and precision medicine.

There is more than one opportunity for predictive analytics to improve the way healthcare systems could impact people’s lives. The result of this approach is not only in the administrative aspect but directly on how an individual patient will have access to a complete, affordable, fair and efficient healthcare system. We really hope that public policy will be updated by integrating predictive analytics as a regular tool for gathering data and proper information analysis for a better comprehension of the complexity of a large and diverse population like ours.