Health systems rely on commercial prediction algorithms to identify and help patients with complex health needs. We show that a widely used algorithm, typical of this industry-wide approach and affecting millions of patients, exhibits significant racial bias: At a given risk score, Black patients are considerably sicker than White patients, as evidenced by signs of uncontrolled illnesses. Remedying this disparity would increase the percentage of Black patients receiving additional help from 17.7 to 46.5%. The bias arises because the algorithm predicts health care costs rather than illness, but unequal access to care means that we spend less money caring for Black patients than for White patients. Thus, despite health care cost appearing to be an effective proxy for health by some measures of predictive accuracy, large racial biases arise. We suggest that the choice of convenient, seemingly effective proxies for ground truth can be an important source of algorithmic bias in many contexts.

/pdf/dissecting-racial-bias-in-an-algorithm-used-to-manage-the-3w9n49km45.pdf

Dissecting racial bias in an algorithm used to manage the health of populations

Medication adherence usually refers to whether patients take their medications as prescribed (eg, twice daily), as well as whether they continue to take a prescribed medication. Medication nonadherence is a growing concern to clinicians, healthcare systems, and other stakeholders (eg, payers) because of mounting evidence that it is prevalent and associated with adverse outcomes and higher costs of care. To date, measurement of patient medication adherence and use of interventions to improve adherence are rare in routine clinical practice. The goals of the present report are to address (1) different methods of measuring adherence, (2) the prevalence of medication nonadherence, (3) the association between nonadherence and outcomes, (4) the reasons for nonadherence, and finally, (5) interventions to improve medication adherence.

Medication Adherence: Its Importance in Cardiovascular Outcomes

America's health care system has become too complex and costly to continue business as usual. Best Care at Lower Cost explains that inefficiencies, an overwhelming amount of data, and other economic and quality barriers hinder progress in improving health and threaten the nation's economic stability and global competitiveness. According to this report, the knowledge and tools exist to put the health system on the right course to achieve continuous improvement and better quality care at a lower cost.The costs of the system's current inefficiency underscore the urgent need for a systemwide transformation. About 30 percent of health spending in 2009--roughly $750 billion--was wasted on unnecessary services, excessive administrative costs, fraud, and other problems. Moreover, inefficiencies cause needless suffering. By one estimate, roughly 75,000 deaths might have been averted in 2005 if every state had delivered care at the quality level of the best performing state. This report states that the way health care providers currently train, practice, and learn new information cannot keep pace with the flood of research discoveries and technological advances.About 75 million Americans have more than one chronic condition, requiring coordination among multiple specialists and therapies, which can increase the potential for miscommunication, misdiagnosis, potentially conflicting interventions, and dangerous drug interactions. Best Care at Lower Cost emphasizes that a better use of data is a critical element of a continuously improving health system, such as mobile technologies and electronic health records that offer significant potential to capture and share health data better. In order for this to occur, the National Coordinator for Health Information Technology, IT developers, and standard-setting organizations should ensure that these systems are robust and interoperable. Clinicians and care organizations should fully adopt these technologies, and patients should be encouraged to use tools, such as personal health information portals, to actively engage in their care.This book is a call to action that will guide health care providers; administrators; caregivers; policy makers; health professionals; federal, state, and local government agencies; private and public health organizations; and educational institutions.

Best Care at Lower Cost: The Path to Continuously Learning Health Care in America

In the United States, approximately 14 million people have had cancer and more than 1.6 million new cases are diagnosed each year. However, more than a decade after the Institute of Medicine (IOM) first studied the quality of cancer care, the barriers to achieving excellent care for all cancer patients remain daunting. Care often is not patient-centered, many patients do not receive palliative care to manage their symptoms and side effects from treatment, and decisions about care often are not based on the latest scientific evidence. The cost of cancer care also is rising faster than many sectors of medicine--having increased to $125 billion in 2010 from $72 billion in 2004--and is projected to reach $173 billion by 2020. Rising costs are making cancer care less affordable for patients and their families and are creating disparities in patients' access to high-quality cancer care. There also are growing shortages of health professionals skilled in providing cancer care, and the number of adults age 65 and older--the group most susceptible to cancer--is expected to double by 2030, contributing to a 45 percent increase in the number of people developing cancer. The current care delivery system is poorly prepared to address the care needs of this population, which are complex due to altered physiology, functional and cognitive impairment, multiple coexisting diseases, increased side effects from treatment, and greater need for social support. Delivering High-Quality Cancer Care: Charting a New Course for a System in Crisis presents a conceptual framework for improving the quality of cancer care. This study proposes improvements to six interconnected components of care: (1) engaged patients; (2) an adequately staffed, trained, and coordinated workforce; (3) evidence-based care; (4) learning health care information technology (IT); (5) translation of evidence into clinical practice, quality measurement and performance improvement; and (6) accessible and affordable care. This report recommends changes across the board in these areas to improve the quality of care. Delivering High-Quality Cancer Care: Charting a New Course for a System in Crisis provides information for cancer care teams, patients and their families, researchers, quality metrics developers, and payers, as well as HHS, other federal agencies, and industry to reevaluate their current roles and responsibilities in cancer care and work together to develop a higher quality care delivery system. By working toward this shared goal, the cancer care community can improve the quality of life and outcomes for people facing a cancer diagnosis.

/pdf/delivering-high-quality-cancer-care-charting-a-new-course-2u1fq5ml3k.pdf

Delivering High-Quality Cancer Care: Charting a New Course for a System in Crisis

ContextPrescription drugs are instrumental to managing and preventing chronic disease. Recent changes in US prescription drug cost sharing could affect access to them.ObjectiveTo synthesize published evidence on the associations among cost-sharing features of prescription drug benefits and use of prescription drugs, use of nonpharmaceutical services, and health outcomes.Data SourcesWe searched PubMed for studies published in English between 1985 and 2006.Study Selection and Data ExtractionAmong 923 articles found in the search, we identified 132 articles examining the associations between prescription drug plan cost-containment measures, including co-payments, tiering, or coinsurance (n = 65), pharmacy benefit caps or monthly prescription limits (n = 11), formulary restrictions (n = 41), and reference pricing (n = 16), and salient outcomes, including pharmacy utilization and spending, medical care utilization and spending, and health outcomes.ResultsIncreased cost sharing is associated with lower rates of drug treatment, worse adherence among existing users, and more frequent discontinuation of therapy. For each 10% increase in cost sharing, prescription drug spending decreases by 2% to 6%, depending on class of drug and condition of the patient. The reduction in use associated with a benefit cap, which limits either the coverage amount or the number of covered prescriptions, is consistent with other cost-sharing features. For some chronic conditions, higher cost sharing is associated with increased use of medical services, at least for patients with congestive heart failure, lipid disorders, diabetes, and schizophrenia. While low-income groups may be more sensitive to increased cost sharing, there is little evidence to support this contention.ConclusionsPharmacy benefit design represents an important public health tool for improving patient treatment and adherence. While increased cost sharing is highly correlated with reductions in pharmacy use, the long-term consequences of benefit changes on health are still uncertain.

/pdf/prescription-drug-cost-sharing-associations-with-medication-3gkob0unzd.pdf

Prescription drug cost sharing: associations with medication and medical utilization and spending and health.

Background: Each year, thousands of older adults develop delirium, a serious, preventable condition. At present, there is no well-validated method to identify patients with delirium when using Medicare claims data or other large datasets. We developed and assessed the performance of classification algorithms based on longitudinal Medicare administrative data that included International Classification of Diseases, 10th Edition diagnostic codes. Methods: Using a linked electronic health record (EHR)-Medicare claims dataset, 2 neurologists and 2 psychiatrists performed a standardized review of EHR records between 2016 and 2018 for a stratified random sample of 1002 patients among 40,690 eligible subjects. Reviewers adjudicated delirium status (reference standard) during this 3-year window using a structured protocol. We calculated the probability that each patient had delirium as a function of classification algorithms based on longitudinal Medicare claims data. We compared the performance of various algorithms against the reference standard, computing calibration-in-the-large, calibration slope, and the area-under-receiver-operating-curve using 10-fold cross-validation (CV). Results: Beneficiaries had a mean age of 75 years, were predominately female (59%), and non-Hispanic Whites (93%); a review of the EHR indicated that 6% of patients had delirium during the 3 years. Although several classification algorithms performed well, a relatively simple model containing counts of delirium-related diagnoses combined with patient age, dementia status, and receipt of antipsychotic medications had the best overall performance [CV– calibration-in-the-large <0.001, CV-slope 0.94, and CV–area under the receiver operating characteristic curve (0.88 95% confidence interval: 0.84–0.91)]. Conclusions: A delirium classification model using Medicare administrative data and International Classification of Diseases, 10th Edition diagnosis codes can identify beneficiaries with delirium in large datasets.

Identifying Medicare Beneficiaries With Delirium

Many older Americans do not receive needed care for mental health and substance use disorders (MHSUD), and there are substantial racial and ethnic disparities in receipt of this care across the lifespan. Medicare introduced cost-sharing parity for outpatient MHSUD care during the period 2010-14, reducing beneficiaries' out-of-pocket share of MHSUD spending from 50 percent to 20 percent. Among traditional Medicare beneficiaries ages sixty-five and older, we examined changes in MHSUD use and spending during the period 2008-18 for low-income beneficiaries with the cost-sharing reduction versus a control group of beneficiaries with free care throughout the study period among Black, Hispanic, Asian, and American Indian/Alaska Native versus White beneficiaries. Among older Medicare beneficiaries, overall use of MHSUD services increased during this period. For White beneficiaries, MHSUD cost-sharing parity was associated with an increased likelihood of having specialty MHSUD visits and medication use and a reduced likelihood of having unmonitored MHSUD medication use and MHSUD emergency department visits and hospitalizations. However, cost-sharing parity was associated with smaller or no gains in MHSUD services use for racial and ethnic minority beneficiaries compared with White beneficiaries, thus widening racial and ethnic disparities in MHSUD care.

Coverage Parity And Racial And Ethnic Disparities In Mental Health And Substance Use Care Among Medicare Beneficiaries.

P522. Cardiovascular Risk in Patients With Severe Mental Illness (SMI) Enrolled in the Fixed Dose Intervention Trial of New England Enhancing Survival in SMI (FITNESS)

OBJECTIVES
Commercial accountable care organization (ACO) contracts attempt to mitigate spending growth, but past evaluations have been limited to continuously enrolled ACO members in health maintenance organization (HMO) plans, excluding many members. The objective of this study was to examine the magnitude of turnover and leakage within a commercial ACO.


STUDY DESIGN
A historical cohort study using detailed information from multiple commercial ACO contracts within a large health care system between 2015 and 2019.


METHODS
Individuals insured through 1 of the 3 largest commercial ACO contracts during the study period, 2015-2019, were included. We examined patterns of entry and exit and the characteristics that predicted remaining in the ACO compared with leaving the ACO. We also examined predictors of the amount of care delivered in the ACO compared with outside the ACO.


RESULTS
Among the 453,573 commercially insured individuals in the ACO, approximately half left the ACO within the initial 24 months after entry. Approximately one-third of spending was for care occurring outside the ACO. Patients who remained in the ACO differed from those who left earlier, including being older, having a non-HMO plan, having lower predicted spending at entry, and having more medical spending for care performed within the ACO during the initial quarter of membership.


CONCLUSIONS
Both turnover and leakage hamper the ability of ACOs to manage spending. Modifications that address potentially intrinsic vs avoidable sources of population turnover and increase patient incentives for care within vs outside of ACOs could help address medical spending growth within commercial ACO programs.

Population turnover and leakage in commercial ACOs.

BACKGROUND Physicians often receive lower payments for dual-eligible Medicare-Medicaid beneficiaries versus nondual Medicare beneficiaries because of state reimbursement caps. The Affordable Care Act (ACA) primary care fee bump temporarily eliminated this differential in 2013-2014. OBJECTIVE To examine how dual payment policy impacts primary care physicians' (PCP) acceptance of duals. RESEARCH DESIGN We assessed differences in the likelihood that PCPs had dual caseloads of ≥10% or 20% in states with lower versus full dual reimbursement using linear probability models adjusted for physician and area-level traits. Using a triple-difference approach, we examined changes in dual caseloads for PCPs versus a control group of specialists in states with fee bumps versus no change during years postbump versus prebump. SUBJECTS PCPs and specialists (cardiologists, orthopedic surgeons, general surgeons) that billed fee-for-service Medicare. MEASURES State dual payment policies and physicians' dual caseloads as a percentage of their Medicare patients. RESULTS In 2012, 81% of PCPs had dual caseloads of ≥10% and this was less likely among PCPs in states with lower versus full dual reimbursement (eg, difference=-4.52 percentage points; 95% confidence interval, -6.80 to -2.25). The proportion of PCPs with dual caseloads of ≥10% or 20% decreased significantly between 2012 and 2017 and the fee bump was not consistently associated with increases in dual caseloads. CONCLUSIONS Pre-ACA, PCPs' participation in the dual program appeared to be lower in states with lower reimbursement for duals. Despite the ACA fee bump, dual caseloads declined over time, raising concerns of worsening access to care.

Mary Price

Papers

Identifying Medicare Beneficiaries With Delirium

Coverage Parity And Racial And Ethnic Disparities In Mental Health And Substance Use Care Among Medicare Beneficiaries.

P522. Cardiovascular Risk in Patients With Severe Mental Illness (SMI) Enrolled in the Fixed Dose Intervention Trial of New England Enhancing Survival in SMI (FITNESS)

Population turnover and leakage in commercial ACOs.

Payment Discrepancies and Access to Primary Care Physicians for Dual-eligible Medicare-Medicaid Beneficiaries.