Tackling Quality—It's Never a Level Playing Field

Since 2008, health care in the United States has been shifting from a volume-driven, fee-for-service system to value-based reimbursement. Value in health care is defined as health outcomes achieved per dollar spent.¹ By changing the basis of provider payments, the goal has been to improve both individual and population health, while also limiting cost increases. This move to value-based care has been driven by legislation and programs managed by the Center for Medicare and Medicaid Services (CMS). The CMS exists at the vanguard of reimbursement policies, being the largest health care payer in the United States, with Medicare and Medicaid accounting for 37% of health care spending in 2017.² Private insurers often adopt efforts led by the CMS; for example, UnitedHealth Group, Aetna, and Anthem distributed nearly half of their reimbursements via value-based arrangements in 2017.³

Because of this shift to value-based care, there are increasing calls for health outcomes data to serve as end points for quality improvement processes. A work group of health policy experts and patient representatives was recently assembled by the American Academy of Neurology Institute (AANI) to formulate outcome measures that would (1) be broadly applicable across neurologic diseases and (2) function as meaningful tools to drive quality improvement in an individual provider's clinical practice.⁴ After substantial iterative discussion informed by a 21-day public comment period, 3 outcome measures were ultimately approved: Patient Communication Experiences for Patients with Neurologic Conditions; a Quality of Life Outcome for Patients with Neurologic Conditions; and EMG Utilization for Isolated Lower Back Pain. It is important to emphasize that these outcome measures reflect expert consensus that they are targets with the potential for added value in neurologic care. Whether the metrics lead to meaningful quality improvement will require continued study and consideration for risk adjustment to determine their value, and it will take time to understand how these proposed outcome metrics could fit into provider payment programs. The original CMS value-based programs illustrated the necessity of risk adjustment for fair use. Comparing patient health outcomes, and by extension the health care practices caring for these patients, without consideration of the presence or absence of all variables affecting the outcome metric of interest may lead to unjust impacts on reimbursement. The CMS value-based programs originated from a Department of Health and Human Services (HHS) proposal to shift medical reimbursement along a spectrum from straightforward fee for service with no link to quality to population-based payments (essentially outcome measures dependent on quality care). By 2015, the goal was to have 90% of Medicare fee-for-service payments in categories 2–4:

• Category 1: fee for service with no link of payment to quality.

• Category 2: fee for service with a link of payment to quality.

• Category 3: alternative payment models built on fee-for-service architecture.

• Category 4: population-based payment.

Correspondingly, the CMS developed value-based programs to guide US health care delivery away from category 1. The original 5 were the End-Stage Renal Disease Quality Incentive Program (ESRD QIP), the Hospital Value-Based Purchasing Program (VBP), the Hospital Readmission Reduction Program (HRRP), the Value Modifier Program (VM), and the Hospital Acquired Conditions Reduction Program (HAC).⁶

While a program rather than a measure, the underlying goal of the HRRP was to improve communication, similar to the AANI's Patient Communication Experiences for Patients' outcome metric. By penalizing hospitals with high readmission rates with a 3% reduction in payments, institutions would be incentivized to improve internal process measures to limit avoidable readmissions.^7,8 However, the HRRP did not consider the impact of socioeconomic status or race/ethnicity, and when studied, these 2 factors alone affected payments for 80% of acute care hospitals and accounted for 10–40% of the HRRP payment penalty. In effect, safety net and rural hospitals taking care of sicker or poorer patients were inordinately penalized.⁹ Sensitivity analysis testing only for socioeconomic status noted a similar overall effect.⁹ Consequently, recognition of this influential variable on payment penalties under HRRP resulted in the passage of legislation to separate hospitals into peer groups on the basis of socioeconomics.⁸

The HRRP experience must serve as a reminder that external variables can affect outcome measures and that contrasts in population outcomes on the basis of socioeconomic measures can be striking. Sico et al.⁴ emphasized the necessity of review following implementation of their measures to ensure that appropriate risk adjustment strategies are implemented. It is easy to imagine that caring for groups disadvantaged by limited caregiver support, access to electronic communication, or cognitive functioning due to coexisting health conditions would adversely affect scores on patient communication measures. With the outcome measure of “Quality of Life Outcome for Patients with Neurologic Conditions,” considerations for risk adjustment could include factors associated with the conditions itself (e.g., severity and stage of a given disease), sociodemographic characteristics (e.g., age, education, and homelessness), and provider type.⁸ Collecting high quality, timely and actionable survey data efficiently and with minimum burden on providers is the ideal. Payers and providers should partner together, likely with informatics/electronic medical records colleagues, to explore how to make this ideal a reality while having data attributed to a specific provider rather than an entire practice. As discussed in this AANI article, an approach to mitigate these biases could be to focus on provider self-improvement. To do this, the outcome of interest would be a provider's relative improvement in outcome metrics over time as opposed to a comparison to other clinicians, particularly in the initial years of use. There is precedent: VBP payments are adjusted based on whether a hospital scores well relative to other hospitals or as compared to their own prior performance.¹⁰

The authors have provided worthwhile measures to investigate quality. Hopefully, these outcome measures will prove to be meaningful tools to improve health outcomes in a value-based environment, but at this time, they have not been validated or tested for reliability. Therefore, they are not currently suitable for inclusion in a quality payment program. Although these measures might not be ready for reimbursement programs, neurology providers should take the lead in implementing outcomes measures, which may in turn lead to better risk adjustment development in the future as data are collated from usage.

Appendix Authors

• Received June 3, 2021.
• Accepted December 29, 2021.

• © 2022 American Academy of Neurology