Social and Medical Risk Factors in Value-Based Purchasing Results

A common criticism of quality improvement and value-based programs is whether they adequately adjust for patient medical condition and social determinants of health.  Research published in JAMA examines how these factors may impact provider performance under the Medicare value-based purchasing program.   (JAMA Article)    The Medicare value-based purchasing program invents or penalizes physicians for their performance on a variety of process, outcome, patient satisfaction and cost measures.  The authors attempted to ascertain the extent to which the medical risk or social environment characteristics of a large physician practice’s patient base might impact their VBP scores and reimbursement.  The practices included in the study had 100 or more doctors and were not in the shared savings or other similar demonstrations.  A practice was considered to have a high medical risk patient base if the hierarchical condition category risk scores were at or above the 75th percentile and those at a similar percentile for patients who were dually eligible were considered to have high social risk.  Quality composite scores and cost composite scores under the value-based purchasing program were the main outcome variables.  Individual components of those composite scores were also analysed.  Practices were divided into four types–low medical and social risk, high medical and social risk, low medical, high social risk, and high medical, low social risk.

Of 900 practices, 547 were low risk, 122 were high risk on both categories, 102 were high social risk only and 128 were high medical risk only.  The low risk practices performed much better on the composite quality score than the other categories of practices, with high social risk only practices bringing up the rear.  The performance on the cost composite score was different, here the high social risk only group had the lowest cost, followed by the low risk group. High medical risk only and high risk on both groups had higher costs.  Risk-adjusted readmission rates were lowest in the low-risk group and highest in the double high risk practices.  Per capita annual costs were $9506 at low risk practices and $11,692 at the double high risk practices and $13,683 in the high medical risk only practices.  In simulations of penalties or rewards, only 4% of low risk practices would be penalized; 10% of high social risk would be, 13% of double high risk and 18% of high medical risk ones.  The results could reflect some systemic issues at the practices or they could reflect patient characteristics that are beyond the control of physicians and should be taken into account in program design.