A government agency wanted to evaluate the efficacy of a potential policy change for handling of long term disability leave payments to injured federal employees. In some cases an injured worker’s physician fails to provide appropriate medical justification for the worker’s continued absence from work on disability leave. In these cases, adjudicators will require a second opinion (SECOP) from another physician regarding the worker’s level of capability to return to work. Elder Research was tasked to evaluate the effectiveness of these second opinions to encourage employees to go back to work.
To inform decisions about which claimants were receiving second opinions specifically for the purpose of evaluating continued disability leave, Elder Research fused multiple noisy data sources in to a large DB2 database. To enhance future analysis of SECOPs we recommended they expand data logging. Survival analysis in R was then used to create a baseline for how long workers were paid compensation before returning to work, dying, or leaving the compensation rolls for other reasons.
Next we examined the typical behaviors of claims adjudicators regarding scheduling second opinions and concluded that the legislation requiring SECOPs did not significantly affect the rate of second opinions, but did affect the timing of SECOPs relative to the lifetime of the case. To evaluate the effectiveness of the procedures and when in the case lifetime a second opinion would be more effective, we compared lengths-of-stay on disability between claimants who did and did not receive second opinions.
A framework for future analysis was proposed where cases are targeted based on their prior propensity to return to work and the estimated benefit of a second opinion. The framework also included future causal analysis on the effectiveness of second opinions by examining the survival rates of claims assigned to different offices or claims adjudicators. The framework would utilize SECOP administration rates to determine the effect of SECOPs while controlling for intrinsic properties of a case which affect propensity to return to work.