A PRACTICAL MODEL FOR EARLY IDENTIFICATION OF PROSPECTIVE HIGH NEED HIGH COST PATIENTS

Abstract

ABSTRACTBackgroundHigh Need-High Cost (HNHC) patients are those who experience poor health outcomes and high health care costs. Early identification may improve outcomes and lower costs.AimDevelopment of a model using retrospective data to identify patients at risk for becoming HNHC patients, in order to efficaciously plan interventions.MethodsData from a large Israeli Health Maintenance Organization (HMO) that includes 488,615 clients above the age of 21 were examined. Multivariate linear regression models were developed using 2012–2016 health expenditure as a dependent variable.ResultsThe number of yearly purchases of medications for chronic disorders, yearly outpatient visits, yearly emergency department and hospital admissions and the last measured HgA1c level were highly predictive of increased expenditure over a five-year period. Each of these indicators has a different coefficient of influence.ConclusionsWe developed a predictive model, based on easily obtained data from electronic medical records that enabled us to identify a population at risk for becoming HNHC in the next five years, a time window allows for intervention. Further research is needed to evaluate whether this is an early enough stage to implement pro-active intervention in the primary care setting.Trial registrationretrospectively registered.HIGHLIGHTSIn this study, we developed a numerical point system calculator, to indicate a risk score for health deterioration within 5 years of patients, by using numerical indicators existing in standard EMR data.The indicators introduced into this calculated risk can guide healthcare providers to the needed areas of intervention. The display of indicators also promotes optimization of care management and continuity of care.This risk score is expected to focus the attention of primary care teams on the population that will benefit most from it, as well as to evaluate the effectiveness of specific interventions.