Algorithm used to cut ER overutilization for kids penalizes Black, Hispanic patients: study

Payers aiming to cut healthcare costs by using a diagnosis claims-based algorithm to determine whether pediatric patients need to be taken to emergency departments risk billing Black and Hispanic patients more for the provided care than white patients, according to a study in JAMA Network Open.

Researchers with numerous children’s hospitals across the country participated in the study, spurred partly by the fact that in an effort to cut costs, Medicaid has focused on emergency departments, billing families for pediatric ED visits based on whether that care was “emergent” or not.

“Black and Hispanic children were significantly more likely to have their ED visits classified as nonemergent,” the study said. “Insurers using these algorithms for reducing reimbursement for certain ED visits may inadvertently contribute to relatively lower payment for ED clinicians caring for Black and Hispanic children.”

The risk of having to pay more out-of-pocket for ED visits may force families to delay or cancel needed care, a trend that’s been seen when the algorithm has been used by private health insurance companies, the study said.

The study investigated an approach used by some state regulators and health plans to reduce ED costs by using retrospective (after the visit) algorithms to process claims. Researchers used a simulation model to pinpoint equity implications.

Alon Peltz, M.D., the study’s corresponding author, told Fierce Healthcare in an email, that “we were alarmed to find that a common health care data algorithm appears to more often identify ED visits for Black and Latino children as low-acuity or non-emergent as compared to visits for white children. This is particularly concerning since algorithms are increasingly being used to measure the quality of care and to determine provider payment.”

Alon Peltz Harvard Medical School
Alon Peltz, M.D. (Harvard Medical School )

Though the differences across racial and ethnic groups who sought nonemergent care were relatively small, facilities serving a higher number of Medicaid beneficiaries might find themselves underfunded by this payment policy, according to the study.

The researchers reviewed over 8 million Medicaid claims filed between Jan. 1, 2016, and Dec. 31, 2019, and used an algorithm developed by the Virginia Department of Medical Assistance. The algorithm allows payers to reduce bills for ED visits down to the lowest level for those determined to be nonemergencies. The algorithm is based on 800 ICD-10-CM codes identified as nonemergent.

Researchers found that 47.7% of the visits were identified as possible nonemergencies and therefore allowed for a reduction in reimbursement. More nonemergent ED visits were tallied for Black (50.3%) and Hispanic (49%) children aged between 4 and 12 years than white children (45.3%).

“Modeling the impact of the reimbursement reductions across the cohort resulted in expected per-visit reimbursement that was 6% lower for visits by Black children and 3% lower for visits by Hispanic children relative to visits by white children,” the study said.

The authors stressed that the study isn’t meant to be an evaluation of Virginia’s ED payment policies and noted that each state uses the data collected by the algorithm in a different way. Researchers also understand the need to cut down on nonemergent visits to emergency rooms but urge that policymakers proceed with caution.

“While diagnosis-based claims algorithms for classifying ED visits can be efficient tools for monitoring utilization patterns, their lack of universally accepted conventions, discordant results across administrative and clinical definitions, and potential racial and ethnic bias should serve to raise concern regarding their use in informing reimbursement policy for children,” the study states.

Peltz said that “beyond this study, other concerns in the clinical community that have been raised about using ED data algorithms include (a) whether algorithms are sufficiently accurate to identify if a visit was low acuity/non-emergent, (b) appropriateness of using retrospective coverage denials/payment reductions to try to discourage ED visits and why ED providers would be penalized, and (c) more broadly whether programs to reduce ED payment will materially incentivize the necessary changes in population health we seek to improve primary care system capacity and promote more proactive clinical and social care interventions; or just reduce reimbursement.”