By Dan Trott, Healthcare Strategist, Healthcare & Life Sciences, Dell EMC
Twitter: @DellHealth
Twitter: @DantrottDell
From identifying risk factors for cancer to infectious disease monitoring
Data culled from increasing numbers of shared databases of electronic health records (EHR), clinical research pools, geographical positioning systems, connected devices and more has given rise to using advanced analytics for improving population health management (PHM).
Advanced analysis of population health data is helping evaluate procedures already in place, identify policy and process improvements, and establish outcome criteria based on proven correlations by using sophisticated data analysis tools, including predictive analytics and machine-learning algorithms. Such analysis promises benefit at many levels of care, from the patient perspective, to that of providers, to larger populations over expansive geographies or disease states. It can garner insights from what would otherwise be “loud data,” or indecipherable information, with no easily identifiable correlations or connections. Opportunities abound in terms of where the advanced analysis of population health data will take us.
Identification of early-onset and risk factors for cancer
According to the American Society of Clinical Oncology, incidence of cancer will rise from 1.6 million in 2015 to 2.3 million in 2030—a 35 percent increase in 15 years. While many risk factors for cancer are well-known, population health analytics offer a deeper, broader knowledge base to help drive prevention, as well as earlier diagnosis and more individualized treatment.
For example, the Neuroblastoma and Medulloblastoma Translational Research Consortium (NMTRC), a consortium of more than 25 universities and children’s hospitals as well as the Translational Genomics Research Institute (TGen), set the goal of increasing remission rates in their patients. By using advanced analytics with support from Dell and Intel, NMTRC created a model for treating children with cancer. They were able to use aggregate data on patient genomes to increase the speed and depth of insight into treatment options and outcomes. Thus far, they have been able to stop the progression of cancer in 60 percent of patients and achieved remission in some patient populations that were formerly deemed incurable.
Monitoring and treatment of infectious disease outbreaks
Advanced analytics has major implications for PHM in monitoring and treatment of infectious disease outbreaks. Analysis of aggregated, up-to-the-minute information from multiple sources regionally or globally, including that from the World Health Organization (WHO) or the Centers for Disease Control (CDC), can incorporate risk factors, population density and mobility, and environmental factors, as well as account for missing data, to assess the rate of infection. For example, in Uganda where the prevalence of typhoid is high, they use the Ugandan Ministry of Health databases to address typhoid outbreaks. PHM officials can track an outbreak as it happens in multiple areas around the country.
More recently, public and governmental organizations have been tracking the spread of the Zika virus using data collated and reported by groups like the CDC’s ArboNET. Currently, the virus has affected Africa, Asia and the Americas, with more than 2,700 cases reported in the U.S. alone. The CDC is using data analytics to track the spread, train healthcare providers on how to identify the virus, and alert healthcare organizations and the public of progression. Data is also being shared with researchers to understand the disease better with the hopes of being able to develop a vaccine to stop the spread entirely.
Improved chronic care management for diabetes
The US spent $245 billion on diabetes in 2013. Problems still facing practitioners today include failure in patient understanding and compliance as well as insufficient follow-up care. Advanced analytics can help overcome such obstacles by identifying subtypes of type 2 diabetes that would require more individualized and also more effective treatment.
One study used a database built at Mount Sanai Medical Center to examine patients with type 2 diabetes, based on EHR data from more than 10,000 patients. The researchers were able to identify three subtypes of type 2 diabetes—each with a varied spectrum of signs and symptoms and rooted in different genetic traits as well as molecular pathways. Such analyses can inform assessment, diagnosis and treatment options for patients with type 2 diabetes and improve compliance because treatment is more tailored and, therefore, more effective.
Advanced analytics in addressing diabetes care management can result in cost savings, in additions to better care. Health Net Connect (HNC) initiated a population diabetic management program to improve clinical outcomes and healthcare savings for diabetes. They captured vitals and blood work from study participants over a 6-month period to measure the impact that routine teleconferencing and patient monitoring had on outcome. Patients in the program showed a significant decrease in key biomarkers, including 9.5% lower HB A1C and 35% decrease in LDL. To put that into perspective, for every 1% drop in HB A1C they estimate an $8,600 annual savings, and for every 1% decrease in LDL there is a 1% decrease in coronary heart disease, which costs on average a million dollars over a lifetime.
Where do we go from here
These examples represent just a small fraction of how advanced analytics are impacting population health and management, but they indicate a positive trend. As technology fuels the engine of analytics, and we become more adept at interpolating and understanding the enormous pools of data at our disposal, the promise to improve health on a grand scale becomes ever more within reach.