The case‐mix of chronic illness hospitalization rates in a managed care population: implications for health management programmes

Abstract

If an HM programme was to be implemented in this population, the typical identification methods currently used by the industry would have resulted in most hospitalized patients either being initially classified as low risk or going undetected. Improving identification and stratification methods will allow HM programmes to better tailor interventions to impact hospitalization rates for the chronically ill.

Improving participant selection in disease management programmes: insights gained from propensity score stratification

“…Rather than targeting individuals with high costs, it may be better to target individuals with high potential savings. Data presented here and elsewhere [27] clearly suggest that targeting high‐cost individuals allows the programme to get a ‘free ride’ on the regression to the mean effect – those treated will show a natural decrease in admissions, while non‐treated individuals in the lower strata will have higher acute utilization. This may help make the case for a treatment effect, but it does not help the non‐participants who may truly benefit from the intervention.…”

“…Using this approach, the outcome measure can be described as the net change in admissions rate of participants over non-participants. The DID strategy ensures that any unobserved variables that remain constant over time, and are correlated with the participation decision and the outcome variable, will not bias the estimated effect [27].…”