Inequality P4P Quality

Should you risk adjust for social factors?

If you want to implement value-based care, you need to define what a ‘good’ outcome is. A good outcome may be different for different individuals. For instance, if you look at mortality rates for a given disease, it makes sense that individuals age 80 and above will have higher mortality rates than those 20-29 year old with the identical condition. If you are measuring physician or hospital quality, you will want to control–or “risk-adjust”–for these differences in patient characteristics. Traditionally, risk adjustment controls for patient demographics and medical conditions. However, should you control for social factors such as poverty or education? There is a lot of debate:

The current policy of the National Quality Forum (NQF) allows adjustment for social risk.5 However, the Office of the Assistant Secretary for Planning and Evaluation (ASPE) at the Department of Health and Human Services recently released a report to Congress that recommended against adjusting for social risk factors for the process and outcome measures used in quality reporting and value-based purchasing programs

A paper by Nerenz et al. (2021) in Health Affairs aims to answer this question. While the authors do end up advocating for risk-adjustment for social factors, of more interest is the logic and assumptions needed to justify (or not justify) the inclusion of these factors. In essence, controlling for social factors means that the expected quality standard or outcomes are lower for some groups than others. This makes sense for demographics and medical conditions (e.g., older people and people with more comorbid conditions are likely to have worse outcomes on average), but should poorer individuals outcomes be held to a lower standard?

The authors argue that the answer to this question depends on the causal pathway you believe is operating. Consider the case that physicians decide to provide lower quality care for poor people compared to rich individuals. In that case, risk adjustment would not be appropriate. It could also be the case that low-quality providers are sorted to poor individuals and high-quality providers are sorted to rich individuals. Again, here, risk adjustment would not be appropriate since low-quality providers should not have a lower standard simply because people are poor.

On the other hand, disparities in quality of care provided (i.e., the process underlying the care) may not be associated with health disparities. For instance, let’s assume that physicians provide the same quality care to all individuals. However, assume when surgical follow-up requires numerous physician visits, poorer individuals are less likely to have consistent child care or easy access to transportation and thus outcomes are worse for reasons independent of the care the physician provides. Or, assume that when physicians may prescribe a medication, poorer individuals are less likely to be adherent to the prescribed medication because they are not able to afford the copayments. In both these examples, outcomes would be worse due to factors outside the physician’s control. In both these cases, one should risk-adjust for social factors.

In some cases, physicians may be able to overcome the social factors. For instance, they could prescribe generic rather than branded pharmaceuticals. Yet in other cases, overcoming the social factors would be too burdensome or expensive for many providers. For instance, providing free child care and transportation to all patients would be very expensive for practices. In this latter case, one should risk-adjust for social factors since it is unreasonable for practices to be able to shoulder this additional cost burden for individuals with social factors.

In short, payers and policymakers need to decide what is a bigger concern: risk adjusting and setting different standards of outcomes based on social factors or not risk adjusting and having physicians try to avoid patients with adverse social factors.


Leave a Reply

Your email address will not be published. Required fields are marked *