Upcoding: Evidence from Medicare on Squishy Risk Adjustment

Utilization and Outcomes in U.S. ICU Hospitalizations

OBJECTIVES

Despite its importance, detailed national estimates of ICU utilization and outcomes remain lacking. We aimed to characterize trends in ICU utilization and outcomes over a recent 12-year period in the United States.

DESIGN/SETTING

In this longitudinal study, we examined hospitalizations involving ICU care ("ICU hospitalizations") alongside hospitalizations not involving ICU care ("non-ICU hospitalizations") among traditional Medicare beneficiaries using 100% Medicare part A claims data and commercial claims data for the under 65 adult population from 2008 to 2019.

PATIENTS/INTERVENTIONS

There were 18,313,637 ICU hospitalizations and 78,501,532 non-ICU hospitalizations in Medicare, and 1,989,222 ICU hospitalizations and 16,732,960 non-ICU hospitalizations in the commercially insured population.

MEASUREMENTS AND MAIN RESULTS

From 2008 to 2019, about 20% of Medicare hospitalizations and 10% of commercial hospitalizations involved ICU care. Among these ICU hospitalizations, length of stay and ICU length of stay decreased on average. Mortality and hospital readmissions on average also decreased, and they decreased more among ICU hospitalizations than among non-ICU hospitalizations, for both Medicare and commercially insured patients. Both Medicare and commercial populations experienced a growth in shorter ICU hospitalizations (between 2 and 7 d in length), which were characterized by shorter ICU stays and lower mortality. Among these short hospitalizations in the Medicare population, for common clinical diagnoses cared for in both ICU and non-ICU settings, patients were increasingly triaged into an ICU during the study period, despite being younger and having shorter hospital stays.

CONCLUSIONS

ICUs are used in a sizeable share of hospitalizations. From 2008 to 2019, ICU length of stay and mortality have declined, while short ICU hospitalizations have increased. In particular, for clinical conditions often managed both within and outside of an ICU, shorter ICU hospitalizations involving younger patients have increased. Our findings motivate opportunities to better understand ICU utilization and to improve the value of ICU care for patients and payers.

Growing divergence between Medicare Advantage plan bids and payments to plans

As the Medicare Advantage (MA) program grows in enrollment and costs, there has been increasing concern that federal payments to MA plans exceed necessary levels. Estimates suggest that, in 2023, MA plans were paid up to 6% more per enrollee than would have been spent had that beneficiary instead enrolled in traditional Medicare (TM). We evaluated the factors driving this overpayment, characterizing trends in MA benchmarks, bids, and total payments from pre-Affordable Care Act (pre-ACA) levels through 2023. We found that, despite an overall decrease in risk-adjusted bids relative to average risk-adjusted TM enrollee costs, total payments to plans have modestly increased since 2015. Decomposing these trends into various factors in the MA payment formula, we found that divergent trends in benchmarks and bids are, in part, due to the increasing influence of payment adjustments, such as quartile spending adjustments, quality bonus payments, and risk adjustment. Our results suggest that current payment rules have contributed to overpayments and policy reform may be necessary.

A framework for ex-ante evaluation of the potential effects of risk equalization and risk sharing in health insurance markets with regulated competition

Many health insurance markets are organized by principles of regulated competition. Regulators of these markets typically apply risk equalization (aka risk adjustment) and risk sharing to mitigate risk selection. Risk equalization and risk sharing can have various positive and negative effects on efficiency and fairness. This paper provides a comprehensive framework for ex-ante evaluation of these effects. In a first step, we distinguish 22 potential effects. In a second step, we summarize and discuss quantitative measures used for evaluating risk equalization and risk sharing schemes in academic research. To underline the relevance of our work, we compare our framework with an existing framework that was previously used in the Dutch regulated health insurance market. We conclude that this framework is incomplete and uses inappropriate measures. To avoid suboptimal policy choices, we recommend policymakers (1) to consider the entire spectrum of potential effects and (2) to select their measures carefully.

Medicare-Covered Services Near the End of Life in Medicare Advantage vs Traditional Medicare

Importance

Financial incentives in Medicare Advantage (MA), the managed care alternative to traditional Medicare (TM), were designed to reduce overutilization. For patients near the end of life (EOL), MA incentives may reduce potentially burdensome care and encourage hospice but could also restrict access to costly but necessary services.

Objective

To compare receipt of potentially burdensome treatments and transfers and potentially necessary postacute services in the last 6 months of life in individuals with MA vs TM.

Design, Setting, and Participants

A retrospective analysis of Medicare claims data among older Medicare beneficiaries who died between 2016 and 2018. The study included Medicare decedents aged 66 years or older covered by TM (n = 659 135) or MA (n = 360 430). All decedents and the subset of decedents with 1 or more emergent hospitalizations with a life-limiting condition (cancer, dementia, end-stage organ failure) that would likely qualify for hospice care were included.

Exposure

MA enrollment.

Main Outcomes

Receipt of potentially burdensome hospitalizations and treatments; receipt of postdischarge home and facility care.

Results

The study included 659 135 TM enrollees (mean [SD] age at death, 83.3 [9.0] years, 54% female, 15.1% non-White, 55% with 1 or more life-limiting condition) and 360 430 MA enrollees (mean [SD] age at death 82.5 [8.7] years, 53% female, 19.3% non-White, 49% with 1 or more life-limiting condition). After regression adjustment, MA enrollees were less likely to receive potentially burdensome treatments (-1.6 percentage points (pp); 95% CI, -2.1 to -1.1) and less likely to die in a hospital (-3.3 pp; 95% CI, -4.0 to -2.7) compared with TM. However, when hospitalized, MA enrollees were more likely to die in the hospital (adjusted difference, 1.3 pp; 95% CI, 1.1-1.5) and less likely to be transferred to rehabilitative or skilled nursing facilities (-5.2 pp; 95% CI, -5.7 to -4.6). Higher rates of home health and home hospice among those discharged home offset half of the decline in facility use. Results were unchanged in the life-limiting conditions sample.

Conclusions

MA enrollment was associated with lower rates of potentially burdensome and facility-based care near the EOL. Greater use of home-based care may improve quality of care but may also leave patients without adequate assistance after hospitalization.

In-Home Health Risk Assessments And Chart Reviews Contribute To Coding Intensity In Medicare Advantage

There is widespread agreement that taxpayers pay more when Medicare beneficiaries are enrolled in Medicare Advantage (MA) plans than if those beneficiaries were enrolled in traditional Medicare. MA plans are paid on the basis of submitted diagnoses and thus have a clear incentive to encourage providers to find and report as many diagnoses for their enrollees as possible. Two mechanisms that MA plans use to identify diagnoses that are not available for beneficiaries in traditional Medicare are in-home health risk assessments and chart reviews. Using MA encounter data for 2015-20, I isolated the impact of these two types of encounters on the risk scores used for payments to MA plans during 2016-21. I found that encounter-based risk scores for MA enrollees were higher by 0.091 points, or 7.4 percent, in 2021 when in-home health risk assessments and chart reviews were included than they would have been without the use of these tools.

Trends in Emergency Department Exam Medicare Reimbursements Between 2010 and 2018

BACKGROUND

As the Medicare population continues to grow, financial pressure is placed upon hospitals, physicians, and other providers as the payer mix has an increasing proportion of Medicare patients.

OBJECTIVE

The purpose of this study was to further the understanding of reimbursement trends surrounding the five levels of emergency department (ED) examinations (CPT codes 99281-99285) from 2010 to 2018 and determine how they have changed with respect to each procedure.

METHODS

CPT codes were filtered into the 2010 and 2018 Physician/Supplier Procedure Summaries from the Centers for Medicare and Medicaid Services' website to gather data on emergency physician submissions and Medicare denials and payments.

RESULTS

In 2010, 15,669,196 ED examinations were submitted to Medicare for $7,628,693,382 while in 2018, 16,432,184 ED examinations were submitted for $14,522,456,383. Despite an increase of $397/submission made by emergency physicians, Medicare paid 20.5% of the submitted charges in 2010 for ED examinations and 11.9% in 2018. The denial rate in 2018 was highest for level I ED examinations (11.3%), and the lowest for level V examinations (5.1%). The utilization of level V ED examinations increased 22.3% from 2010 to 2018, while the utilization of the others decreased. Of the five levels of ED examinations, only the level I examination did not exhibit a decrease.

CONCLUSIONS

From 2010 to 2018, emergency physicians charged a higher amount for ED examinations, yet Medicare reimbursement accounted for a smaller proportion of these charges, resulting in less payment per submission for the four most common levels. Downward trends in Medicare reimbursement may place financial burdens that could potentially hamper healthcare outcomes.

Medicare Advantage enrollment and outcomes of post-acute nursing home care among patients with dementia

Enrollment in Medicare Advantage (MA) has been rapidly growing. We examined whether MA enrollment affects the outcomes of post-acute nursing home care among patients with Alzheimer's disease and related dementias (ADRD). We exploited year-to-year changes in MA penetration rates within counties from 2012 through 2019. After adjusting for patient-level characteristics and county fixed effects, we found that MA enrollment was not associated with days spent at home, nursing home days, likelihood of becoming a long-stay resident, hospital days, hospital readmission, or 1-year mortality. There was a modest increase in successful discharge to the community by 0.73 percentage points (relative increase of 2.4%) associated with a 10-percentage-point increase in MA enrollment. The results are consistent among racial/ethnic subgroups and dual-eligible patients. These findings suggest an imperative need to monitor and improve quality of managed care among enrollees with ADRD.

Comparing the outcomes and costs of cardiac monitoring with implantable loop recorders and mobile cardiac outpatient telemetry following stroke using real-world evidence

Aim: Patients with ischemic stroke (IS) commonly undergo monitoring to identify atrial fibrillation with mobile cardiac outpatient telemetry (MCOT) or implantable loop recorders (ILRs). The authors compared readmission, healthcare cost and survival in patients monitored post-stroke with either MCOT or ILR. Materials & methods: The authors used claims data from Optum's de-identified Clinformatics® Data Mart Database to identify patients with IS hospitalized from January 2017 to December 2020 who were prescribed ambulatory cardiac monitoring via MCOT or ILR. They compared the costs associated with the initial inpatient visit as well as the rate and causes of readmission, survival and healthcare costs over the following 18 months. Datasets were balanced using patient baseline and hospitalization characteristics. Multivariable generalized linear gamma regression was used for cost comparisons. Cox proportional hazard regression was used for survival and readmission analysis. Sub-cohorts were analyzed based on the severity of the index IS. Results: In 2244 patients, readmissions were significantly lower in the MCOT monitored group (30.2%) compared with the ILR group (35.4%) (hazard ratio [HR] 1.23; 95% CI: 1.04-1.46). Average cost over 18 months starting with the index IS was $27,429 (USD) lower in the MCOT group (95% CI: $22,353-$32,633). Survival difference bordered on statistical significance and trended to lower mortality in MCOT (8.9%) versus ILR (11.3%) (HR 1.30; 95% CI: 1:00-1.69), led by significance in patients with complications or comorbidities with the index event (MCOT 7.5%, ILR 11.5%; HR 1.62; 95% CI: 1.11-2.36). Conclusion: The use of MCOT versus ILR as the primary monitor following IS was associated with significant decreases in readmission, lower costs for the initial IS and total care over the next 18 months, significantly lower mortality for patients with complications and comorbidities at the index stroke, and a trend toward improved survival across all patients.

Lower comorbidity scores and severity levels in Veterans Health Administration hospitals: a cross-sectional study

BACKGROUND

Previous studies found that documentation of comorbidities differed when Veterans received care within versus outside Veterans Health Administration (VHA). Changes to medical center funding, increased attention to performance reporting, and expansion of Clinical Documentation Improvement programs, however, may have caused coding in VHA to change.

METHODS

Using repeated cross-sectional data, we compared Elixhauser-van Walraven scores and Medicare Severity Diagnosis Related Group (DRG) severity levels for Veterans' admissions across settings and payers over time, utilizing a linkage of VHA and all-payer discharge data for 2012-2017 in seven US states. To minimize selection bias, we analyzed records for Veterans admitted to both VHA and non-VHA hospitals in the same year. Using generalized linear models, we adjusted for patient and hospital characteristics.

RESULTS

Following adjustment, VHA admissions consistently had the lowest predicted mean comorbidity scores (4.44 (95% CI 4.34-4.55)) and lowest probability of using the most severe DRG (22.1% (95% CI 21.4%-22.8%)). In contrast, Medicare-covered admissions had the highest predicted mean comorbidity score (5.71 (95% CI 5.56-5.85)) and highest probability of using the top DRG (35.3% (95% CI 34.2%-36.4%)).

CONCLUSIONS

More effective strategies may be needed to improve VHA documentation, and current risk-adjusted comparisons should account for differences in coding intensity.

A Novel Machine Learning Algorithm for Creating Risk-Adjusted Payment Formulas

Importance

Models predicting health care spending and other outcomes from administrative records are widely used to manage and pay for health care, despite well-documented deficiencies. New methods are needed that can incorporate more than 70 000 diagnoses without creating undesirable coding incentives.

Objective

To develop a machine learning (ML) algorithm, building on Diagnostic Item (DXI) categories and Diagnostic Cost Group (DCG) methods, that automates development of clinically credible and transparent predictive models for policymakers and clinicians.

Design, Setting, and Participants

DXIs were organized into disease hierarchies and assigned an Appropriateness to Include (ATI) score to reflect vagueness and gameability concerns. A novel automated DCG algorithm iteratively assigned DXIs in 1 or more disease hierarchies to DCGs, identifying sets of DXIs with the largest regression coefficient as dominant; presence of a previously identified dominating DXI removed lower-ranked ones before the next iteration. The Merative MarketScan Commercial Claims and Encounters Database for commercial health insurance enrollees 64 years and younger was used. Data from January 2016 through December 2018 were randomly split 90% to 10% for model development and validation, respectively. Deidentified claims and enrollment data were delivered by Merative the following November in each calendar year and analyzed from November 2020 to January 2024.

Main Outcome and Measures

Concurrent top-coded total health care cost. Model performance was assessed using validation sample weighted least-squares regression, mean absolute errors, and mean errors for rare and common diagnoses.

Results

This study included 35 245 586 commercial health insurance enrollees 64 years and younger (65 901 460 person-years) and relied on 19 clinicians who provided reviews in the base model. The algorithm implemented 218 clinician-specified hierarchies compared with the US Department of Health and Human Services (HHS) hierarchical condition category (HCC) model's 64 hierarchies. The base model that dropped vague and gameable DXIs reduced the number of parameters by 80% (1624 of 3150), achieved an R2 of 0.535, and kept mean predicted spending within 12% ($3843 of $31 313) of actual spending for the 3% of people with rare diseases. In contrast, the HHS HCC model had an R2 of 0.428 and underpaid this group by 33% ($10 354 of $31 313).

Conclusions and Relevance

In this study, by automating DXI clustering within clinically specified hierarchies, this algorithm built clinically interpretable risk models in large datasets while addressing diagnostic vagueness and gameability concerns.

Comparison of prior authorization across insurers: cross sectional evidence from Medicare Advantage

OBJECTIVE

To measure and compare the scope of US insurers' policies for prior authorization (PA), a process by which insurers assess the necessity of planned medical care, and to quantify differences in PA across insurers, physician specialties, and clinical service categories.

DESIGN

Cross sectional analysis.

SETTING

PA policies for five insurers serving most of the beneficiaries covered by privately administered Medicare Advantage in the US, 2021, as applied to utilization patterns observed in Medicare Part B.

PARTICIPANTS

30 540 086 beneficiaries in traditional Medicare Part B.

MAIN OUTCOME MEASURES

Proportions of government administered traditional Medicare Part B spending and utilization that would have required PA according to Medicare Advantage insurer rules.

RESULTS

The insurers required PA for 944 to 2971 of the 14 130 clinical services (median 1899; weighted mean 1429) constituting 17% to 33% of Part B spending (median 28%; weighted mean 23%) and 9% to 41% of Part B utilization (median 22%; weighted mean 18%). 40% of spending ($57bn; £45bn; €53bn) and 48% of service utilization would have required PA by at least one insurer; 12% of spending and 6% of utilization would have required PA by all insurers. 93% of Part B medication spending, or 74% of medication use, would have required PA by at least one Medicare Advantage insurer. For all Medicare Advantage insurers, hematology and oncology drugs represented the largest proportion of PA spending (range 27-34%; median 33%; weighted mean 30%). PA rates varied widely across specialties.

CONCLUSION

PA policies varied substantially across private insurers in the US. Despite limited consensus, all insurers required PA extensively, particularly for physician administered medications. These findings indicate substantial differences in coverage policies between government administered and privately administered Medicare. The results may inform ongoing efforts to focus PA more effectively on low value services and reduce administrative burdens for clinicians and patients.

Medicare Advantage Is Associated with Lower Utilization of Total Joint Arthroplasty

BACKGROUND

Medicare Advantage (MA) insurers use managed care techniques to review the utilization of medical services and control costs. It is unclear if MA enrollees have a lower utilization of elective surgical procedures such as inpatient hip and knee total joint arthroplasty (TJA), which have traditionally been covered by traditional Medicare (TM) without restrictions.

METHODS

We conducted a cross-sectional study using a 20% sample of 2018 TM claims and MA encounter records for 5,300,188 TM enrollees and 1,970,032 MA enrollees who were 65 to 85 years of age. We calculated unadjusted and adjusted differences (controlling for beneficiary and market characteristics) in the incidence of TJA for MA compared with TM, and by MA plan type. Finally, we calculated differences in the time to contact with an orthopaedic surgeon and time to the surgical procedure among enrollees with an osteoarthritis diagnosis.

RESULTS

After controlling for observable characteristics, there was a 15.6% lower incidence of TJA in MA enrollees compared with TM enrollees (p < 0.001). Compared with TM enrollees, health maintenance organization (HMO) enrollees were 28.1% less likely to undergo TJA, controlling for observable characteristics (p < 0.001). From the initial diagnosis, the time to contact with an orthopaedic surgeon and the time to the surgical procedure were also lower among TM enrollees compared with MA enrollees. At 2 years after an osteoarthritis diagnosis, 10.4% of TM enrollees, 7.9% of preferred provider organization (PPO) enrollees, and 5.7% of HMO enrollees had undergone inpatient TJA.

CONCLUSIONS

MA coverage was associated with a lower utilization of elective, inpatient hip and knee TJA. MA was also associated with a longer time to orthopaedic surgeon evaluation and surgical procedure.

LEVEL OF EVIDENCE

Prognostic Level III . See Instructions for Authors for a complete description of levels of evidence.

Postacute Care for Medicare Advantage Enrollees Who Switched to Traditional Medicare Compared With Those Who Remained in Medicare Advantage

Importance

Medicare Advantage (MA) plans receive capitated per enrollee payments that create financial incentives to provide care more efficiently than traditional Medicare (TM); however, incentives could be associated with MA plans reducing use of beneficial services. Postacute care can improve functional status, but it is costly, and thus may be provided differently to Medicare beneficiaries by MA plans compared with TM.

Objective

To estimate the association of MA compared with TM enrollment with postacute care use and postdischarge outcomes.

Design, Setting, and Participants

This was a cohort study using Medicare data on 4613 hospitalizations among retired Ohio state employees and 2 comparison groups in 2015 and 2016. The study investigated the association of a policy change with use of postacute care and outcomes. The policy changed state retiree health benefits in Ohio from a mandatory MA plan to subsidies for either supplemental TM coverage or an MA plan. After policy implementation, approximately 75% of retired Ohio state employees switched to TM. Hospitalizations for 3 high-volume conditions that usually require postacute rehabilitation were assessed. Data from the Medicare Provider Analysis and Review files were used to identify all hospitalizations in short-term acute care hospitals. Difference-in-difference regressions were used to estimate changes for retired Ohio state employees compared with other 2015 MA enrollees in Ohio and with Kentucky public retirees who were continuously offered a mandatory MA plan. Data analyses were performed from September 1, 2019, to November 30, 2023.

Exposures

Enrollment in Ohio state retiree health benefits in 2015, after which most members shifted to TM.

Main Outcomes and Measures

Received care in an inpatient rehabilitation facility, skilled nursing facility, or home health, or any postacute care; the occurrence of any hospital readmission; the number of days in the community during the 30 days after hospital discharge; and mortality.

Results

The study sample included 2373 hospitalizations for Ohio public retirees, 1651 hospitalizations for other Humana MA enrollees in Ohio, and 589 hospitalizations for public retirees in Kentucky. After the 2016 policy implementation, the percentage of hospitalizations covered by MA decreased by 70.1 (95% CI, -74.2 to -65.9) percentage points (pp), inpatient rehabilitation facility admissions increased by 9.7 (95% CI, 4.7 to 14.7) pp, use of only home health or skilled nursing facility care fell by 8.6 (95% CI, -14.6 to -2.6) pp, and days in the community fell by 1.6 (95% CI, -2.9 to -0.3) days for Ohio public retirees compared with other Humana MA enrollees in Ohio. There was no change in 30-day mortality or hospital readmissions; similar results were found by comparisons using Kentucky public retirees as a control group.

Conclusions and Relevance

The findings of this cohort study indicate that after a change in retiree health benefits, most Ohio public retirees shifted from MA to TM and received more intensive postacute care with no significant change in measured short-term postdischarge outcomes. Future work should consider additional measures of postacute functional status over a longer follow-up period.

Is there an advantage? Considerations for researchers studying the effects of the type of Medicare coverage

OBJECTIVE

To describe common methodological problems that arise in comparisons of Medicare Advantage (MA) and Traditional Medicare (TM) and within-MA studies and provide suggestions of how researchers can address these issues.

STUDY SETTING

Published research evaluating Medicare coverage options in the United States.

STUDY DESIGN

We considered key conceptual challenges and promising solutions that have been used thus far and suggest additional directions.

DATA COLLECTION

Not available.

PRINCIPAL FINDINGS

Many existing studies of MA versus TM include significant limitations, such as failing to account for unobserved confounders driving both beneficiary coverage choice and health outcomes once enrolled, not accounting for variation in benefit generosity, provider networks, or plan design across MA plans, and/or having been conducted at a time when MA enrollment was less than a third of all Medicare beneficiaries. We provide a review of methods that can help researchers to overcome these weaknesses and suggest additional methods and data sources that may aid future research.

CONCLUSIONS

The MA program is becoming an essential part of the US healthcare system. By accounting for non-random movement into and out of MA and studying the heterogeneity of beneficiary experience across plan and market characteristics, researchers can provide the high-quality evidence necessary for policymakers to design the program and reform TM in ways that maximize beneficiary outcomes.

Population-Based Trends in Complexity of Hospital Inpatients

Importance

Clinical experience suggests that hospital inpatients have become more complex over time, but few studies have evaluated this impression.

Objective

To assess whether there has been an increase in measures of hospital inpatient complexity over a 15-year period.

Design, Setting and Participants

This cohort study used population-based administrative health data from nonelective hospitalizations from April 1, 2002, to January 31, 2017, to describe trends in the complexity of inpatients in British Columbia, Canada. Hospitalizations were included for individuals 18 years and older and for which the most responsible diagnosis did not correspond to pregnancy, childbirth, the puerperal period, or the perinatal period. Data analysis was performed from July to November 2023.

Exposure

The passage of time (15-year study interval).

Main Outcomes and Measures

Measures of complexity included patient characteristics at the time of admission (eg, advanced age, multimorbidity, polypharmacy, recent hospitalization), features of the index hospitalization (eg, admission via the emergency department, multiple acute medical problems, use of intensive care, prolonged length of stay, in-hospital adverse events, in-hospital death), and 30-day outcomes after hospital discharge (eg, unplanned readmission, all-cause mortality). Logistic regression was used to estimate the relative change in each measure of complexity over the entire 15-year study interval.

Results

The final study cohort included 3 367 463 nonelective acute care hospital admissions occurring among 1 272 444 unique individuals (median [IQR] age, 66 [48-79] years; 49.1% female and 50.8% male individuals). Relative to the beginning of the study interval, inpatients at the end of the study interval were more likely to have been admitted via the emergency department (odds ratio [OR], 2.74; 95% CI, 2.71-2.77), to have multimorbidity (OR, 1.50; 95% CI, 1.47-1.53) and polypharmacy (OR, 1.82; 95% CI, 1.78-1.85) at presentation, to receive treatment for 5 or more acute medical issues (OR, 2.06; 95% CI, 2.02-2.09), and to experience an in-hospital adverse event (OR, 1.20; 95% CI, 1.19-1.22). The likelihood of an intensive care unit stay and of in-hospital death declined over the study interval (OR, 0.96; 95% CI, 0.95-0.97, and OR, 0.81; 95% CI, 0.80-0.83, respectively), but the risks of unplanned readmission and death in the 30 days after discharge increased (OR, 1.14; 95% CI, 1.12-1.16, and OR, 1.28; 95% CI, 1.25-1.31, respectively).

Conclusions and Relevance

By most measures, hospital inpatients have become more complex over time. Health system planning should account for these trends.

Managed competition in the United States: How well is it promoting equity and efficiency?

Managed competition frameworks aim to control healthcare costs and promote access to high-quality health insurance and services through a combination of public policies and market forces. In the United States, managed competition delivery systems are varied and diffused across a patchwork of divided markets and populations. This, coupled with extremely high national health spending per capita, makes a more unified managed competition strategy an appealing alternative to a currently struggling healthcare system. We examine the relative effectiveness of three existing programmes in the U.S. that each rely upon some principles of managed competition: health insurance exchanges instituted by the Affordable Care Act, Medicaid managed care organisations, and Medicare Advantage plans. Although each programme leverages some competitive features, each faces significant hurdles as a candidate for expansion. We highlight these challenges with a survey of academic health economists, and find that provider and insurer consolidation, highly segmented markets, and failing to incentivise competitive efficiencies all dampen the success of existing programmes. Although managed competition for all is a potentially desirable framework for future health reform in the U.S., successful expansion relies on addressing fundamental issues revealed by imperfect existing programmes.

Association Between the Medicare Advantage Quartile Adjustment System and Plan Behavior and Enrollment

Importance

Medicare Advantage (MA) has grown in popularity, but critics believe that insurers are overpaid, partially due to the quartile adjustment system that determines plan benchmarks. However, elimination of the quartile adjustments may be associated with less generous benefits and fewer plan offerings, which could slow MA enrollment growth.

Objective

To examine whether the quartile adjustment system is associated with differences in county-level benefits, insurer offerings, and MA enrollment.

Design, Setting, and Participants

The quartile adjustments create discontinuous jumps in county-level base payments based on historical traditional Medicare spending. Data from January 2017 to December 2021 and a regression discontinuity design were used to examine changes in insurer behavior and MA enrollment between quartiles. The analytic sample included 1557 county observations.

Main Outcomes and Measures

Study outcomes included monthly premiums, the share of plans charging premiums, primary care copayments, the share of plans using rebates to reduce Part B premiums, supplemental benefits, plan and contract availability, and MA enrollment.

Results

Discontinuities were found in the quartile adjustments and benchmarks. A 1-percentage point (pp) increase in the quartile adjustment was associated with a $6.36 increase in monthly benchmarks (95% CI, 5.10-7.62), a $0.51 decrease in monthly premiums (95% CI, -0.96 to -0.07), and a 0.68 pp decrease in the share of plans charging premiums (95% CI, -1.25 to -0.10). Significant changes were not found in primary care copayments (-$0.04; 95% CI, -0.17 to 0.09), the share of plans using rebates to reduce Part B premiums (-0.17 pp; 95% CI, -0.34 to 0.01), supplemental benefits (eg, preventive dental coverage; 0.17 pp; 95% CI, -0.25 to 0.0), the number of plans (1.06; 95% CI, -3.44 to 5.57) or contracts (0.31; 95% CI, -0.18 to 0.81), or the MA enrollment rate (0.16 pp; 95% CI, -0.61 to 0.94).

Conclusions and Relevance

The study results suggest that MA plans are not very sensitive to modest changes in payment rates. Modifications to the quartile adjustment system may generate savings without substantially affecting MA beneficiaries.

Upcoding in medicare: where does it matter most?

Upcoding in Medicare has been a topic of interest to economists and policy makers for nearly 40 years. While upcoding is generally understood as "billing for services at higher level of complexity than the service actually pro- vided or documented," it has a wide range of definitions within the literature. This is largely because the financial incentives across programs and aspects under the coding control of billing specialists and providers are different, and have evolved substantially over time, as has the published literature. Arguably, the primary importance of analyzing upcoding in different parts of Medicare is to inform policy makers on the magnitude of the process and to suggest approaches to mitigate the level of upcoding. Financial estimates for upcoding in traditional Medicare (Medicare Parts A and B), are highly variable, in part reflecting differences in methodology for each of the services covered. To resolve this variability, we used summaries of audit data from the Comprehensive Error Rate Testing program for the period 2010-2019. This program uses the same methodology across all forms of service in Medicare Parts A and B, allowing direct comparisons of upcoding magnitude. On average, upcoding for hospitalization under Part A represents $656 million annually (or 0.53% of total Part A annual expenditures) during our sample period, while up- coding for physician services under Part B is $2.38 billion annually (or 2.43% of Part B annual expenditures). These numbers compare to the recent consistent estimates from multiple different entities putting upcoding in Medicare Part C at $10-15 billion annually (or approximately 2.8-4.2% of Part C annual expenditures). Upcoding for hospitalization under Medicare Part A is small, relative to overall upcoding expenditures.

The impact of dual VA-Medicare use on a data-driven clinical management tool for older Veterans with multimorbidity

BACKGROUND

Healthcare systems are increasingly turning to data-driven approaches, such as clustering techniques, to inform interventions for medically complex older adults. However, patients seeking care in multiple healthcare systems may have missing diagnoses across systems, leading to misclassification of resulting groups. We evaluated the impact of multi-system use on the accuracy and composition of multimorbidity groups among older adults in the Veterans Health Administration (VA).

METHODS

Eligible patients were VA primary care users aged ≥65 years and in the top decile of predicted 1-year hospitalization risk in 2018 (n = 558,864). Diagnoses of 26 chronic conditions were coded using a 24-month lookback period and input into latent class analysis (LCA) models. In a random 10% sample (n = 56,008), we compared the resulting model fit, class profiles, and patient assignments from models using only VA system data versus VA with Medicare data.

RESULTS

LCA identified six patient comorbidity groups using VA system data. We labeled groups based on diagnoses with higher within-group prevalence relative to the average: Substance Use Disorders (7% of patients), Mental Health (15%), Heart Disease (22%), Diabetes (16%), Tumor (14%), and High Complexity (10%). VA with Medicare data showed improved model fit and assigned more patients with high accuracy. Over 70% of patients assigned to the Substance, Mental Health, High Complexity, and Tumor groups using VA data were assigned to the same group in VA with Medicare data. However, 41.9% of the Heart Disease group and 14.7% of the Diabetes group were reassigned to a new group characterized by multiple cardiometabolic conditions.

CONCLUSIONS

The addition of Medicare data to VA data for older high-risk adults improved clustering model accuracy and altered the clinical profiles of groups. Accessing or accounting for multi-system data is key to the success of interventions based on empiric grouping in populations with dual-system use.

Risk Adjustment in Health Insurance Markets: Do Not Overlook the "Real" Healthy

OBJECTIVES

The goals of this paper are (1) to identify groups of healthy people and (2) to quantify the extent to which the Dutch risk adjustment (RA) model overpays insurers for these groups.

BACKGROUND

There have been strong signals that insurers in the Dutch regulated health insurance market engage in actions to attract healthy people. A potential explanation for this behavior is that the Dutch RA model overpays insurers for healthy people.

METHODS

We identify healthy groups using 3 types of ex-ante information (ie, information available before the start of the health insurance contract): administrative data on prior spending for specific health care services (N = 17 m), diagnoses from electronic patient records (N = 1.3 m), and health survey data (N = 457 k). In a second step, we calculate the under/overpayment for these groups under the Dutch RA model (version: 2021).

RESULTS

We distinguish eight groups of healthy people using various "identifiers." Although the Dutch RA model substantially reduces the predictable profits that insurers face for these groups, significant profits remain. The mean per person overpayment ranges from 38 euros (people with hospital spending below the third quartile in each of 3 prior years) to 167 euros (those without any medical condition according to their electronic patient record).

CONCLUSIONS

The Dutch RA model does not eliminate the profitability of healthy groups. The identifiers used for flagging these groups, however, seem inappropriate for serving as risk adjuster variables. An alternative way of exploiting these identifiers and eliminating the profitability of healthy groups is to estimate RA models with "constrained regression."

Legislative and Regulatory Options for Improving Medicare Advantage

The Medicare Advantage program was created to expand beneficiary choice and to reduce spending through capitated payment to private insurers. However, many stakeholders now argue that Medicare Advantage is failing to deliver on its promise to reduce spending. Three problematic design features in Medicare Advantage payment policy have received particular scrutiny: (1) how baseline payments to insurers are determined, (2) how variation in patient risk affects insurer payment, and (3) how payments to insurers are adjusted for quality performance. The authors analyze the statute underlying these three design features and explore legislative and regulatory strategies for improving Medicare Advantage. They conclude that regulatory approaches for improving risk adjustment and for recouping overpayments from risk-score gaming have the highest potential impact and are the most feasible improvement measures to implement.

Health System Change in the Wake of COVID-19

This JAMA Forum discusses resiliency, telehealth, the health care labor force, and public health in the context of the health system changes occurring since the start of the COVID-19 pandemic.

Critical Bias in Critical Care Devices

Critical care data contain information about the most physiologically fragile patients in the hospital, who require a significant level of monitoring. However, medical devices used for patient monitoring suffer from measurement biases that have been largely underreported. This article explores sources of bias in commonly used clinical devices, including pulse oximeters, thermometers, and sphygmomanometers. Further, it provides a framework for mitigating these biases and key principles to achieve more equitable health care delivery.

Medicare advantage and dialysis facility choice

OBJECTIVE

To compare the characteristics of dialysis facilities used by traditional Medicare (TM) and Medicare advantage (MA) enrollees with end-stage kidney disease (ESKD).

DATA SOURCES

We used 20% TM claims and 100% MA encounter data from 2018 and publicly available data from the Centers for Medicare and Medicaid Services.

STUDY DESIGN

We compared the characteristics of the dialysis facilities treating TM and MA patients in the same ZIP code, adjusting for patient characteristics. The outcome variables were facility ownership, distance to the facility, and several measures of facility quality.

DATA COLLECTION/EXTRACTION

We identified point prevalent dialysis patients as of July 15, 2018.

PRINCIPAL FINDINGS

Compared to TM patients in the same ZIP code, MA patients were 1.84 percentage points more likely to be treated at facilities owned by the largest two dialysis organizations and 1.85 percentage points less likely to be treated at an independently owned facility. MA patients went to further and lower quality facilities than TM patients in the same ZIP code. However, these differences in facility quality were modest. For example, while the mean dialysis facility mortality rate was 21.85, the difference in mortality rates at facilities treating MA and TM patients in the same ZIP code was 0.67 deaths per 100 patient-years. Similarly, MA patients went to facilities that were, on average, 0.15 miles further than TM patients in the same ZIP code.

CONCLUSION

MA enrollees with ESKD were more likely than TM enrollees in the same ZIP code to use the dialysis facilities owned by the two largest chains, travel further for care, and receive care at lower quality facilities. While the magnitude of differences in facility distance and quality was modest, the direction of these results underscores the importance of monitoring dialysis network adequacy as ESKD MA enrollment continues to grow.

Home-Based Medical Care Use In Medicare Advantage And Traditional Medicare In 2018

The use of home-based medical care differed in Medicare Advantage and traditional Medicare in 2018. Having exactly one such visit was thirty-one times as likely for Medicare Advantage beneficiaries (18.6 percent) as for traditional Medicare beneficiaries (0.6 percent), likely reflecting incentives in the Medicare Advantage program to code all accurate diagnoses. Multiple home-based medical care visits were less likely in Medicare Advantage than in traditional Medicare (1.6 percent versus 2.1 percent of beneficiaries, respectively).

Favorable Selection In Medicare Advantage Is Linked To Inflated Benchmarks And Billions In Overpayments To Plans

Increases in Medicare Advantage (MA) enrollment, coupled with concerns about overpayment to plans, have prompted calls for change. Benchmark setting in MA, which determines plan payment, has received relatively little attention as an avenue for reform. In this study we used national data from the period 2010-20 to examine the relationships among unobserved favorable selection, benchmark setting, and payments to plans in MA. We found that unobserved favorable selection in MA led to underpayment to counties with lower MA penetration and overpayment to counties with higher MA penetration. Because the distribution of MA beneficiaries has shifted over time toward counties that were overpaid, we estimate that plans were overpaid by an average of $9.3 billion per year between 2017 and 2020. Changes to risk adjustment in benchmark setting could likely mitigate the impact of favorable selection in MA.

Incidence of moral hazards among health care providers in the implementation of social health insurance toward universal health coverage: evidence from rural province hospitals in Indonesia

Objective

To identify the incidence of moral hazards among health care providers and its determinant factors in the implementation of national health insurance in Indonesia.

Methods

Data were derived from 360 inpatient medical records from six types C public and private hospitals in an Indonesian rural province. These data were accumulated from inpatient medical records from four major disciplines: medicine, surgery, obstetrics and gynecology, and pediatrics. The dependent variable was provider moral hazards, which included indicators of up-coding, readmission, and unnecessary admission. The independent variables are Physicians' characteristics (age, gender, and specialization), coders' characteristics (age, gender, education level, number of training, and length of service), and patients' characteristics (age, birth weight, length of stay, the discharge status, and the severity of patient's illness). We use logistic regression to investigate the determinants of moral hazard.

Results

We found that the incidences of possible unnecessary admissions, up-coding, and readmissions were 17.8%, 11.9%, and 2.8%, respectively. Senior physicians, medical specialists, coders with shorter lengths of service, and patients with longer lengths of stay had a significant relationship with the incidence of moral hazard.

Conclusion

Unnecessary admission is the most common form of a provider's moral hazard. The characteristics of physicians and coders significantly contribute to the incidence of moral hazard. Hospitals should implement reward and punishment systems for doctors and coders in order to control moral hazards among the providers.

Price Increases Versus Upcoding As Drivers Of Emergency Department Spending Increases, 2012-19

Recent studies document a substantial increase in emergency department (ED) spending in the past decade, even though the number of ED visits per capita has remained relatively stable. Price increases and upcoding are sometimes cited as possible explanations, but their relative impacts are not known. We analyzed Blue Cross Blue Shield claims for patients of all ages who received care in EDs in five states in 2012 and 2019. We used estimates from spending regressions and regressions explaining coding intensity to decompose changes in spending between 2012 and 2019 into components attributable to price increases, changes in patient characteristics or treatment intensity, and upcoding. Prices accounted for at least half of the increase in ED spending per visit for four of the five states we examined. Increases in spending attributable to upcoding were notable but generally not as large. Future research should explore the associations between local market conditions, such as consolidation and ownership type, and both price increases and upcoding.

Impact of surgical complications on hospital costs and revenues: retrospective database study of Medicare claims

Aim: To compare the length of stay, hospital costs and hospital revenues for Medicare patients with and without a subset of potentially preventable postoperative complications after major noncardiac surgery. Materials & methods: Retrospective data analysis using the Medicare Standard Analytical Files, Limited Data Set, 5% inpatient claims files for years 2016-2020. Results: In 74,103 claims selected for analysis, 71,467 claims had no complications and 2636 had one or more complications of interest. Claims with complications had significantly longer length of hospital stay (12.41 vs 3.95 days, p < 0.01), increased payments to the provider ($34,664 vs $16,641, p < 0.01) and substantially higher estimates of provider cost ($39,357 vs $16,158, p < 0.01) compared with claims without complications. This results on average in a negative difference between payments and costs for patients with complications compared with a positive difference for claims without complications (-$4693 vs $483, p < 0.01). Results were consistent across three different cost estimation methods used in the study. Conclusion: Compared with patients without postoperative complications, patients developing complications stay longer in the hospital and incur increased costs that outpace the increase in received payments. Complications are therefore costly to providers and payers, may negatively impact hospital profitability, and decrease the quality of life of patients. Quality initiatives aimed at reducing complications can be immensely valuable for both improving patient outcomes and hospital finances.

Use of High-Risk Medications Among Older Adults Enrolled in Medicare Advantage Plans vs Traditional Medicare

Importance

Limiting the use of high-risk medications (HRMs) among older adults is a national priority to provide a high quality of care for older beneficiaries of both Medicare Advantage and traditional fee-for-service Medicare Part D plans.

Objective

To evaluate the differences in the rate of HRM prescription fills among beneficiaries of traditional Medicare vs Medicare Advantage Part D plans and to examine the extent to which these differences change over time and the patient-level factors associated with higher rates of HRMs.

Design, Setting, and Participants

This cohort study used a 20% sample of Medicare Part D data on filled drug prescriptions from 2013 to 2017 and a 40% sample from 2018. The sample comprised Medicare beneficiaries aged 66 years or older who were enrolled in Medicare Advantage or traditional Medicare Part D plans. Data were analyzed between April 1, 2022, and April 15, 2023.

Main Outcomes and Measures

The primary outcome was the number of unique HRMs prescribed to older Medicare beneficiaries per 1000 beneficiaries. Linear regression models were used to model the primary outcome, adjusting for patient characteristics and county characteristics and including hospital referral region fixed effects.

Results

The sample included 5 595 361 unique Medicare Advantage beneficiaries who were propensity score-matched on a year-by-year basis to 6 578 126 unique traditional Medicare beneficiaries between 2013 and 2018, resulting in 13 704 348 matched pairs of beneficiary-years. The traditional Medicare vs Medicare Advantage cohorts were similar in age (mean [SD] age, 75.65 [7.53] years vs 75.60 [7.38] years), proportion of males (8 127 261 [59.3%] vs 8 137 834 [59.4%]; standardized mean difference [SMD] = 0.002), and predominant race and ethnicity (77.1% vs 77.4% non-Hispanic White; SMD = 0.05). On average in 2013, Medicare Advantage beneficiaries filled 135.1 (95% CI, 128.4-142.6) unique HRMs per 1000 beneficiaries compared with 165.6 (95% CI, 158.1-172.3) HRMs per 1000 beneficiaries for traditional Medicare. In 2018, the rate of HRMs had decreased to 41.5 (95% CI, 38.2-44.2) HRMs per 1000 beneficiaries in Medicare Advantage and to 56.9 (95% CI, 54.1-60.1) HRMs per 1000 beneficiaries in traditional Medicare. Across the study period, Medicare Advantage beneficiaries received 24.3 (95% CI, 20.2-28.3) fewer HRMs per 1000 beneficiaries per year compared with traditional Medicare beneficiaries. Female, American Indian or Alaska Native, and White populations were more likely to receive HRMs than other groups.

Conclusion and Relevance

Results of this study showed that HRM rates were consistently lower among Medicare Advantage than traditional Medicare beneficiaries. Higher use of HRMs among female, American Indian or Alaska Native, and White populations is a concerning disparity that requires further attention.

Differences between integrated and non-integrated plans in Medicare Advantage

OBJECTIVE

To understand differences in financial performance, quality performance, supplemental benefits provision, and enrollee composition between integrated and non-integrated plans in the Medicare Advantage (MA) program.

DATA SOURCES

We used data from the Center for Medicare and Medicaid Services for 2015-2017. We included 156 integrated MA plans (31 unique contracts) and 2096 non-integrated MA plans (392 unique contracts).

STUDY DESIGN

We estimated linear probably models for financial performance, quality performance, supplemental benefits provision, and enrollee composition with state fixed effects and contract random effects. We adjusted for county-level market structure-related factors, cost-related factors, and demand-related factors. Our primary independent variable was an indicator of plan-provider integration.

PRINCIPAL FINDINGS

Integrated MA plans were associated with $19.4 (95% CI: 9.2, 29.7) and $16.6 (95% CI: 10.3, 22.9) higher Part C and Part D monthly premiums, but were associated with higher star quality ratings. There were no significant differences in revenues and plan payments per enrollee between integrated and non-integrated MA plans. Integrated MA plans were associated with $40.5 (95% CI: -54.0, -26.9) lower non-claims costs than non-integrated MA plans. There was limited evidence that integrated MA plans provided more generous supplemental benefits than non-integrated MA plans. Enrollment rates in integrated MA plans were particularly low among socially marginalized groups (3.4 [95% CI: -5.9, -1.0], 4.7 [95% CI: -8.5, -0.9], and 4.4 [95% CI: -6.4, -2.4] percentage points lower among non-Hispanic Black, Medicare-Medicaid dual eligible, and the disabled).

CONCLUSIONS

Our findings suggest that integrated MA plans may achieve higher efficiency and quality, but these benefits may not be experienced by all beneficiaries due to disparities in enrollment. As these models continue to spread, it is critical to develop policies to ensure that MA enrollees have equal access to integrated plans.

Differences In Use Of Services And Quality Of Care In Medicare Advantage And Traditional Medicare, 2010 And 2017

Medicare Advantage (MA) enrollment growth could make it difficult for MA plans to maintain their track record of limiting discretionary utilization while delivering higher-quality care than traditional Medicare. We compared quality and utilization measures in Medicare Advantage and traditional Medicare in 2010 and 2017. Clinical quality performance was higher in MA health maintenance organizations (HMOs) and preferred provider organizations (PPOs) than in traditional Medicare for almost all measures in both years. MA HMOs outperformed traditional Medicare on all measures in 2017. MA HMOs' performance on nearly all seven patient-reported quality measures improved, and MA HMOs outperformed traditional Medicare on five of those measures in 2017. MA PPOs performed the same as or better than traditional Medicare on all but one patient-reported quality measure in 2010 and 2017. The number of emergency department visits was 30 percent lower, the number of elective hip and knee replacements was approximately 10 percent lower, and the number of back surgeries was almost 30 percent lower in MA HMOs than in traditional Medicare in 2017. Utilization trends were similar in MA PPOs, but differences from traditional Medicare were narrower. Despite increased enrollment, overall utilization remains lower in Medicare Advantage than in traditional Medicare, whereas quality performance is the same or higher.

How Long-Term Care Quality Measures Address Alzheimer's Disease and Related Dementias in European Countries

OBJECTIVES

Alzheimer's disease and related dementias (ADRD) are prevalent conditions in long-term care homes (LTCHs) with most LTCH residents living with ADRD in many countries. Despite the prevalence of ADRD in LTCHs, a recent examination of LTCH quality measurement programs in 4 countries revealed few LTCH quality measures addressed ADRD, most commonly as a risk adjuster. We sought to better understand how quality measurement programs address ADRD internationally.

DESIGN

International comparative analysis.

SETTING AND PARTICIPANTS

We examined LTCH quality measures in 4 European countries-Germany, Switzerland, Belgium, and the Netherlands.

METHODS

The specifications to calculate each measure were assessed to determine whether the measure was calculated without assessing for ADRD, included only residents with ADRD, excluded residents with ADRD, or was risk adjusted for the presence of ADRD among the LTCH residents.

RESULTS

A total of 143 measures were examined across 4 quality measurement programs. Thirty-seven percent of the measures explicitly address ADRD. The programs addressed ADRD in starkly different ways. In Germany, most (13 of 15) measures addressed ADRD, and did so as an exclusion or inclusion criterion, and in Switzerland all the measures addressed ADRD through risk adjustment. In Flanders, Belgium, all measures were calculated without assessing for ADRD. In the Netherlands, one-third of the measures explicitly addressed ADRD by restricting the measure to psychogeriatric units.

CONCLUSIONS AND IMPLICATIONS

Although limited to examining measures from LTCH quality measurement programs in 4 European countries, this study adds evidence that ADRD tends not to be addressed by LTCH quality measures, but when ADRD is addressed, it tends to be through inclusion or exclusion criteria. LTCH regulators, policymakers, and providers can use this information to assess options for addressing ADRD in quality measurement programs. Future research is needed to assess how standard indicators of ADRD care quality differ across quality measurement programs.

Evaluation of Potentially Avoidable Acute Care Utilization Among Patients Insured by Medicare Advantage vs Traditional Medicare

Importance

Medicare Advantage plans have strong incentives to reduce potentially wasteful health care, including costly acute care visits for ambulatory care-sensitive conditions (ACSCs). However, it remains unknown whether Medicare Advantage plans lower acute care use compared with traditional Medicare, or if it shifts patients from hospitalization to observation stays and emergency department (ED) direct discharges.

Objective

To determine whether Medicare Advantage is associated with differential utilization of hospitalizations, observations, and ED direct discharges for ACSCs compared with traditional Medicare.

Design, Setting, and Participants

Cross-sectional study of US Medicare Advantage vs traditional Medicare beneficiaries from January 1 to December 31, 2018. Poisson regression models were used to compare risk-adjusted rates of Medicare Advantage vs traditional Medicare, controlling for patient demographic characteristics and clinical risk and including county fixed-effects. Data were analyzed between April 2021 and November 2022.

Main Outcomes and Measures

Hospitalizations, observation stays, and ED direct discharges for ACSCs.

Results

The study sample comprised 2 665 340 Medicare Advantage patients (mean [SD] age, 72.7 [9.8] years; 1 504 519 [56.4%] women; 1 859 067 [69.7%] White individuals) and 7 981 547 traditional Medicare patients (mean [SD] age, 71.2 [11.8] years; 4 232 201 [53.0%] women; 6 176 239 [77.4%] White individuals). Medicare Advantage patients had lower risk of hospitalization for ACSCs compared with traditional Medicare patients (relative risk [RR], 0.94; 95% CI, 0.93-0.95), primarily owing to fewer hospitalizations for acute conditions (eg, pneumonia). Medicare Advantage patients had a higher risk of ED direct discharges (RR, 1.44; 95% CI, 1.43-1.45) and observation stays (RR, 2.38; 95% CI, 2.34-2.41) for ACSCs vs traditional Medicare patients. Overall, Medicare Advantage patients were at higher risk of needing care for an ACSC (hospitalization, ED direct discharge, or observation stay) than traditional Medicare patients (RR, 1.30; 95% CI, 1.30-1.31). Within the Medicare Advantage population, patients in health maintenance organizations (HMOs) were at lower risk of ACSC-related hospitalization compared with patients in its preferred provider organizations (RR, 0.96; 95% CI, 0.95-0.98); however, those in the HMOs had a higher risk of ED direct discharge (RR, 1.08; 95% CI, 1.07-1.09) and observation stay (overall RR, 1.10; 95% CI, 1.02-1.12).

Conclusions and Relevance

The findings of this cross-sectional study of Medicare Advantage and traditional Medicare patients with ACSCs indicate that apparent gains in lowering rates of potentially avoidable acute care have been associated with shifting inpatient care to settings such as ED direct discharges and observation stays.

Association of Medicare Advantage vs Traditional Medicare With 30-Day Mortality Among Patients With Acute Myocardial Infarction

IMPORTANCE

Medicare Advantage health plans covered 37% of beneficiaries in 2018, and coverage increased to 48% in 2022. Whether Medicare Advantage plans provide similar care for patients presenting with specific clinical conditions is unknown.

OBJECTIVE

To compare 30-day mortality and treatment for Medicare Advantage and traditional Medicare patients presenting with acute myocardial infarction (MI) from 2009 to 2018.

DESIGN, SETTING, AND PARTICIPANTS

Retrospective cohort study that included 557 309 participants with ST-segment elevation [acute] MI (STEMI) and 1 670 193 with non-ST-segment elevation [acute] MI (NSTEMI) presenting to US hospitals from 2009-2018 (date of final follow up, December 31, 2019).

EXPOSURES

Enrollment in Medicare Advantage vs traditional Medicare.

MAIN OUTCOMES AND MEASURES

The primary outcome was adjusted 30-day mortality. Secondary outcomes included age- and sex-adjusted rates of procedure use (catheterization, revascularization), postdischarge medication prescriptions and adherence, and measures of health system performance (intensive care unit [ICU] admission and 30-day readmissions).

RESULTS

The study included a total of 2 227 502 participants, and the mean age in 2018 ranged from 76.9 years (Medicare Advantage STEMI) to 79.3 years (traditional Medicare NSTEMI), with similar proportions of female patients in Medicare Advantage and traditional Medicare (41.4% vs 41.9% for STEMI in 2018). Enrollment in Medicare Advantage vs traditional Medicare was associated with significantly lower adjusted 30-day mortality rates in 2009 (19.1% vs 20.6% for STEMI; difference, -1.5 percentage points [95% CI, -2.2 to -0.7] and 12.0% vs 12.5% for NSTEMI; difference, -0.5 percentage points [95% CI, -0.9% to -0.1%]). By 2018, mortality had declined in all groups, and there were no longer statically significant differences between Medicare Advantage (17.7%) and traditional Medicare (17.8%) for STEMI (difference, 0.0 percentage points [95% CI, -0.7 to 0.6]) or between Medicare Advantage (10.9%) and traditional Medicare (11.1%) for NSTEMI (difference, -0.2 percentage points [95% CI, -0.4 to 0.1]). By 2018, there was no statistically significant difference in standardized 90-day revascularization rates between Medicare Advantage and traditional Medicare. Rates of guideline-recommended medication prescriptions were significantly higher in Medicare Advantage (91.7%) vs traditional Medicare patients (89.0%) who received a statin prescription (difference, 2.7 percentage points [95% CI, 1.2 to 4.2] for 2018 STEMI). Medicare Advantage patients were significantly less likely to be admitted to an ICU than traditional Medicare patients (for 2018 STEMI, 50.3% vs 51.2%; difference, -0.9 percentage points [95% CI, -1.8 to 0.0]) and significantly more likely to be discharged to home rather than to a postacute facility (for 2018 STEMI, 71.5% vs 70.2%; difference, 1.3 percentage points [95% CI, 0.5 to 2.1]). Adjusted 30-day readmission rates were consistently lower in Medicare Advantage than in traditional Medicare (for 2009 STEMI, 13.8% vs 15.2%; difference, -1.3 percentage points [95% CI, -2.0 to -0.6]; and for 2018 STEMI, 11.2% vs 11.9%; difference, 0.6 percentage points [95% CI, -1.5 to 0.0]).

CONCLUSIONS AND RELEVANCE

Among Medicare beneficiaries with acute MI, enrollment in Medicare Advantage, compared with traditional Medicare, was significantly associated with modestly lower rates of 30-day mortality in 2009, and the difference was no longer statistically significant by 2018. These findings, considered with other outcomes, may provide insight into differences in treatment and outcomes by Medicare insurance type.

Administrative Data Use in National Registry Efforts: Blessing or Curse?

"Big data" refers to a growing field of large database research. Administrative data, a subset of big data, includes information from insurance claims, electronic medical records, and registries that can be useful for investigating novel research questions. While its use provides salient advantages, potential researchers relying on big data would benefit from knowing about how these databases are coded, common errors they may encounter, and how to best use large data to address various research questions. In the first section of this paper, Dr. Nicholas A. Bedard addresses the four major pitfalls to avoid with diagnosis and procedure codes in administrative data. In the next section, Dr. Jeffrey N. Katz et al. focus on the strengths and limitations of administrative data, suggesting methods to mitigate these limitations. Lastly, Dr. Elena Losina et al. review the uses and misuses of large databases for cost-effectiveness research, detailing methods for careful economic evaluations.

Comparison of Health Care Utilization by Medicare Advantage and Traditional Medicare Beneficiaries With Complex Care Needs

Importance

Medicare beneficiaries with co-occurring chronic conditions and complex care needs experience high rates of acute care utilization and poor outcomes. These patterns are well described among traditional Medicare (TM) beneficiaries, but less is known about outcomes among Medicare Advantage (MA) beneficiaries. Compared with TM, MA plans have additional levers to potentially address beneficiary needs, such as network design, care management, supplemental benefits, and value-based contracting.

Objective

To compare health care utilization for MA and TM beneficiaries with complex care needs.

Design, Setting, and Participants

This cross-sectional study analyzed beneficiaries enrolled in MA and TM using claims data from a large, national MA insurer and a random 5% sample of TM beneficiaries. Beneficiaries were segmented into the following cohorts: frail elderly, major complex chronic, and minor complex chronic. Regression models estimated the association between MA enrollment and health care utilization in 2018, using inverse probability of treatment weighting to balance the MA and TM cohorts on observable characteristics. The study period was January 1, 2017, through December 31, 2018. All analyses were conducted from December 2020 to August 2022.

Exposures

Enrollment in MA vs TM.

Main Outcomes and Measures

Hospital stays (inpatient admissions and observation stays), emergency department (ED) visits, and 30-day readmissions.

Results

Among a study population of 1 844 326 Medicare beneficiaries (mean [SD] age, 75.6 [7.1] years; 1 021 479 [55.4%] women; 1 524 458 [82.7%] White; 223 377 [12.1%] with Medicare-Medicaid dual eligibility), 1 177 896 (63.9%) were enrolled in MA and 666 430 (36.1%) in TM. Beneficiary distribution across cohorts was as follows: frail elderly, 116 047 with MA (10.0% of the MA sample) and 104 036 with TM (15.6% of the TM sample); major complex chronic, 320 954 (27.2%) and 158 811 (23.8%), respectively; and minor complex chronic, 740 895 (62.9%) and 403 583 (60.6%), respectively. Beneficiaries enrolled in MA had lower rates of hospital stays, ED visits, and 30-day readmissions. The largest relative differences were observed for hospital stays, which ranged from -9.3% (95% CI, -10.9% to -7.7%) for the frail elderly cohort to -11.9% (95% CI, -13.2% to -10.7%) for the major complex chronic cohort.

Conclusions and Relevance

In this cross-sectional study of Medicare beneficiaries with complex care needs, those enrolled in MA had lower rates of hospital stays, ED visits, and 30-day readmissions than similar beneficiaries enrolled in TM, suggesting that managed care activities in MA may influence the nature and quality of care provided to these beneficiaries.

Association of Insurance Mix and Diagnostic Coding Practices in New York State Hospitals

Question

What is the association between a hospital’s share of privately insured patients and diagnostic coding practices?

Findings

Analyses of 1 614 630 New York Medicaid admissions indicate that for a given patient with multiple hospitalizations, the number of diagnoses on a claim was positively associated with the hospital’s share of privately insured patients. Patients discharged from hospitals in the bottom quartile by share of privately insured patients were coded with 1 to 2 additional diagnoses when they were subsequently discharged from hospitals in the top quartile.

Meaning

Payer mix is associated with diagnostic coding patterns, suggesting that payment policy may be influencing investments in administrative infrastructure.

Importance

Given higher reimbursement rates, hospitals primarily serving privately insured patients may invest more in intensive coding than hospitals serving publicly insured patients. This may lead these hospitals to code more diagnoses for all patients.

Objective

To estimate whether, for the same Medicaid enrollee with multiple hospitalizations, a hospital’s share of privately insured patients is associated with the number of diagnoses on claims.

Design, Setting, and Participants

This cross-sectional study used patient-level fixed effects regression models on inpatient Medicaid claims from Medicaid enrollees with at least 2 admissions in at least 2 different hospitals in New York State between 2010 and 2017. Analyses were conducted from 2019 to 2021.

Exposures

The annual share of privately insured patients at the admitting hospital.

Main Outcomes and Measures

Number of diagnostic codes per admission. Probability of diagnoses being from a list of conditions shown to be intensely coded in response to payment incentives.

Results

This analysis included 1 614 630 hospitalizations for Medicaid-insured patients (mean [SD] age, 48.2 [20.1] years; 829 684 [51.4%] women and 784 946 [48.6%] men). Overall, 74 998 were Asian (4.6%), 462 259 Black (28.6%), 375 591 Hispanic (23.3%), 486 313 White (30.1%), 128 896 unknown (8.0%), and 86 573 other (5.4%). When the same patient was seen in a hospital with a higher share of privately insured patients, more diagnoses were recorded (0.03 diagnoses per percentage point [pp] increase in share of privately insured; 95% CI, 0.02-0.05; P < .001). Patients discharged from hospitals in the bottom quartile of privately insured patient share received 1.37 more diagnoses when they were subsequently discharged from hospitals in the top quartile, relative to patients whose admissions were both in the bottom quartile (95% CI, 1.21-1.53; P < .001). Those going from hospitals in the top quartile to the bottom had 1.67 fewer diagnoses (95% CI, −1.84 to −1.50; P < .001). Diagnoses in hospitals with a higher private payer share were more likely to be for conditions sensitive to payment incentives (0.08 pp increase for each pp increase in private share; 95% CI, 0.06-0.10; P < .001). These findings were replicated in 2016 to 2017 data.

Conclusions and Relevance

In this cross-sectional study of Medicaid enrollees, admission to a hospital with a higher private payer share was associated with more diagnoses on Medicaid claims. This suggests payment policy may drive differential investments in infrastructure to document diagnoses. This may create a feedback loop that exacerbates resource inequity.

Association of Private Equity Acquisition of Physician Practices With Changes in Health Care Spending and Utilization

Question

Findings

Meaning

Importance

Objective

Design, Settings, and Participants

Exposures

Main Outcomes and Measures

Results

Conclusions and Relevance

Association of Quality of Care With Where Veterans Choose to Get Knee Replacement Surgery

Importance

Recent legislation expanded veterans' access to Veterans Health Administration (VA)-purchased care. Quality should be considered when choosing where to get total knee arthroplasty (TKA), but currently available quality metrics provide little guidance.

Objective

To determine whether an association exists between the proportion of TKAs performed (vs purchased) at each VA facility and the quality of care provided (as measured by short-term complication rates).

Design, Setting, and Participants

This 3-year cohort study used VA and community care data (fiscal year 2017 to fiscal year 2019) from the VA's Corporate Data Warehouse. Complications were defined following the Centers for Medicare and Medicaid Services' methodology. The setting included 140 VA health care facilities performing or purchasing TKAs. Participants included veterans who had 43 371 primary TKA procedures that were either VA-performed or VA-purchased during the study period.

Exposures

Of the 43 371 primary TKA procedures, 18 964 (43.7%) were VA-purchased.

Main Outcomes and Measures

The primary outcome was risk-standardized short-term complication rates of VA-performed or VA-purchased TKAs. The association between the proportion of TKAs performed at each VA facility and quality of VA-performed and VA-purchased care was examined using a regression model. Subgroups were also identified for facilities that had complication rates above or below the overall mean complication rate and for facilities that performed more or less than half of the facility's TKAs.

Results

Among the study sample's 41 775 veterans who underwent 43 371 TKAs, 38 725 (89.3%) were male, 6406 (14.8%) were Black, 33 211 (76.6%) were White, and 1367 (3.2%) had other race or ethnicity (including American Indian or Alaska Native, Asian, and Native Hawaiian or other Pacific Islander); mean (SD) age was 66.9 (8.5) years. VA-performed and VA-purchased TKAs had a mean (SD) raw overall short-term complication rate of 2.97% (0.08%). There was no association between the proportion of TKAs performed in VA facilities and risk-standardized complication rates for VA-performed TKAs, and no association for VA-purchased TKAs.

Conclusions and Relevance

In this cohort study, surgical quality did not have an association with where veterans had TKA, possibly because meaningful comparative data are lacking. Reporting local and community risk-standardized complication rates may inform veterans' decisions and improve care. Combining these data with the proportion of TKAs performed at each site could facilitate administrative decisions on where resources should be allocated to improve care.

Modeling Coding Intensity of Procedures in a U.S. Population-Based Hip/Knee Arthroplasty Inpatient Cohort Adjusting for Patient- and Facility-Level Characteristics

Variations in procedure coding intensity, defined as excess coding of procedures versus industry (instead of clinical) standards, can result in differentials in quality of care for patients and have additional implications for facilities and payors. The literature regarding coding intensity of procedures is limited, with a need for risk-adjusted methods that help identify over- and under-coding using commonly available data, such as administrative claims. Risk-adjusted metrics are needed for quality control and enhancement. We propose a two-step approach to risk adjustment, using a zero-inflated Poisson model, applied to a hip-knee arthroplasty cohort discharged during 2019 (n = 313,477) for patient-level risk adjustment, and a potential additional layer for adjustment based on facility-level characteristics, when desired. A 21.41% reduction in root-mean-square error was achieved upon risk adjustment for patient-level factors alone. Furthermore, we identified facilities that over- and under-code versus industry coding expectations, adjusting for both patient-level and facility-level factors. Excess coding intensity was found to vary across multiple levels: (1) geographically across U.S. Census regional divisions; (2) temporally with marked seasonal components; (3) by facility, with some facilities largely departing from industry standards, even after adjusting for both patient- and facility-level characteristics. Our proposed method is simple to implement, generalizable, it can be used across cohorts with different sets of information available, and it is not limited by the accessibility and sparsity of electronic health records. By identifying potential over- and under-coding of procedures, quality control personnel can explore and assess internal needs for enhancements in their health delivery services and monitor subsequent quality improvements.

Hospital-physician integration and risk-coding intensity

Hospital-physician integration has surged in recent years. Integration may allow hospitals to share resources and management practices with their integrated physicians that increase the reported diagnostic severity of their patients. Greater diagnostic severity will increase practices' payment under risk-based arrangements. We offer the first analysis of whether hospital-physician integration affects providers' coding of patient severity. Using a two-way fixed effects model, an event study, and a stacked difference-in-differences analysis of 5 million patient-year observations from 2010 to 2015, we find that the integration of a patient's primary care doctor is associated with a robust 2%-4% increase in coded severity, the risk-score equivalent of aging a physician's patients by 4-8 months. This effect was not driven by physicians treating different patients nor by physicians seeing patients more often. Our evidence is consistent with the hypothesis that hospitals share organizational resources with acquired physician practices to increase the measured clinical severity of patients. Increases in the intensity of coding will improve vertically-integrated practices' performance in alternative payment models and pay-for-performance programs while raising overall health care spending.

Prescribing of low- versus high-cost Part B drugs in Medicare Advantage and traditional Medicare

OBJECTIVE

Examine whether Medicare Advantage (MA) coverage is associated with more efficient prescribing of Part B drugs than traditional Medicare (TM) coverage.

DATA SOURCES

Twenty percent sample of 2016 outpatient and carrier TM claims and MA encounter records and Master Beneficiary Summary File data.

STUDY DESIGN

We analyzed whether MA enrollees compared to TM enrollees more often received the low-cost Part B drug in four clinical scenarios where multiple similarly effective drugs exist: (1) anti-VEGF agents to treat macular degeneration, (2) bone resorption inhibitors for osteoporosis, (3) bone resorption inhibitors for malignant neoplasms, and (4) intravenous iron for iron deficiency anemia. We then estimated differences in spending if TM prescribing aligned with MA prescribing. Finally, using linear probability models, we examined whether differences in MA and TM prescribing patterns were attributable to differences in the hospitals and clinician practices who treat MA and TM enrollees or differences in how these hospitals and clinician practices treat their MA versus TM patients.

DATA COLLECTION/EXTRACTION METHODS

Not applicable.

PRINCIPAL FINDINGS

In all cases, a larger share of MA enrollees received the low-cost drug compared to TM enrollees, ranging from 8 percentage points higher for anemia to 16 percentage points higher for macular degeneration in the unadjusted analysis. Results were similar in regression analyses controlling for enrollee characteristics and market factors (5-13 percentage points). If TM prescribing matched MA prescribing, we estimated savings ranging from 6% to 20% of TM spending for each scenario. Differences in prescribing patterns were driven both by MA enrollees receiving treatment at more efficient hospitals and clinician practices and hospitals and clinician practices more often prescribing low-cost drugs to their MA patients.

CONCLUSIONS

Our findings show MA enrollees were more likely than TM enrollees to receive low-cost Part B drugs in four clinical scenarios where multiple similarly or equally effective treatment options exist.

How to deal with persistently low/high spenders in health plan payment systems?

Health insurance markets with community-rated premiums typically include risk adjustment (RA) to mitigate selection problems. Over the past decades, RA systems have evolved from simple demographic models to sophisticated morbidity-based models. Even the most sophisticated models, however, tend to overcompensate people with persistently low spending and undercompensate those with persistently high spending. This paper compares three methods that exploit spending-level persistence for improving health plan payment systems: (1) implementation of spending-based risk adjustors, (2) implementation of high-risk pooling for people with multiple-year high spending, and (3) indirect use of spending persistence via constrained regression. Based on incentive measures for risk selection and cost control, we conclude that a combination of the last two options can substantially outperform the first, which is currently used in the health plan payment system in the Netherlands.