DURHAM, N.C., Oct. 27, 2016 /PRNewswire/ -- Forecast Health, an innovator in predictive analytics that improve patient and population health, today announced that a Society of Actuaries (SOA) study revealed that, compared to other vendors, the company's predictive modeling approach was at least 9% more accurate at predicting the top 1% of costs, and at least 25% more accurate at predicting healthcare costs overall.
"We are pleased with the results of the Society of Actuaries study," said Norm Storwick, FSA, MAAA, Forecast's Chief Actuary. "More accurate predictive models enable health plans, post-acute care managers, and provider organizations to identify more high-risk patients and intervene to prevent adverse events."
To ensure a fair comparison, the SOA assessed vendors on the same claims database of 1 million people. In cases where vendors provided multiple models, the best performing one was considered.
Forecast outperformed its industry peers in predicting the people who would be in the top 1% of costs and, as measured by the mean absolute error (MAE), also outperformed in predicting costs for all people. Using the R-squared measure, the results did not reflect Forecast's superior performance. However, the study concluded that the R-squared measure is susceptible to the influence of outliers, and that the MAE methodology is a more robust way to measure models.
The SOA study compared models using claims data only. In real world applications, Forecast Health integrates claims, clinical and person-level social determinants of health data. This achieves even higher levels of predictive accuracy than what was reported by the SOA, while also helping to identify which high-risk patients are impactable, and what factors are driving their risk.
"As an actuary, it was important to me to have 3rd party validation of our models by the SOA," said Storwick. "For our clients, greater accuracy will result in better patient outcomes and a meaningful competitive advantage."
About Forecast Health
Forecast Health's next generation analytics help health plans, provider organizations and post-acute care managers improve patient and population health by identifying "impactable" patients who are at high risk of adverse events, and providing patient-specific guidance using social determinants of health data to help mitigate that risk. To learn more, visit forecasthealth.com and follow us on Twitter or LinkedIn.
SOURCE Forecast Health