Medicare Advantage vs. Traditional Medicare: What Matched Quality Measures Show

Medicare policy debates keep returning to one question: Does Medicare Advantage (MA) deliver better quality than traditional Medicare after accounting for who enrolls in each program?

On Wednesday, Sept. 23, Christie Teigland, Ph.D., Vice President of Research Science and Advanced Analytics at Inovalon, presented new evidence on that question during a pre-conference webinar for the National Committee for Quality Assurance (NCQA) Health Innovation Summit 2026. Sari Siegel, Ph.D., Senior Vice President of Research at AHIP, and Fredric Blavin, Ph.D., Associate Vice President of Research and Policy Development at the Blue Cross Blue Shield Association, joined her on the panel.

Teigland’s team matched MA and traditional Medicare members on health and social risk to make the comparison like for like. Rather than narrowing, the quality gaps widened: traditional Medicare’s 30-day readmission rate went from 73% higher than MAs to 126% higher after matching.

For health plan quality leaders, the takeaway reaches beyond the MA debate. A plan-wide rate reflects thousands of members with very different clinical and social circumstances. Breaking results out by member group shows where your programs are already succeeding and where targeted support will make the biggest difference.

For life sciences teams working in health economics and outcomes research (HEOR) and real-world evidence (RWE), this study demonstrates the methodology and level of rigor required to answer complex comparative questions, the kind that require years of pre-index history, neighborhood-level social risk data, and well-matched cohorts.

Medicare Advantage enrollees start out sicker and more socially vulnerable

Most comparisons of MA and traditional Medicare begin measurement at the point of enrollment. Teigland’s team drew on each person’s health history from commercial or Medicaid coverage in the three years before Medicare, which established a baseline before either program recorded a single diagnosis.

“If we don’t account for that, it becomes really hard to tell the differences in the populations versus the differences in how well their care is managed,” Siegel said.

Among people entering MA, 13.6% already had three or more chronic conditions, compared with 8.7% of those entering traditional Medicare. MA enrollees also had higher rates of every top condition the team examined, from diabetes with complications to chronic obstructive pulmonary disease and congestive heart failure.

Blavin then describes the social risk profile; 42% of MA enrollees came to Medicare from Medicaid, twice the rate among traditional Medicare enrollees. MA enrollees were also 1.6 times as likely to live in lower-income neighborhoods and less likely to own a home or a vehicle.

He also pointed to a gap in how the programs are compared. Medicare’s risk adjustment model captures clinical complexity reasonably well, Blavin said, but it does not account for social risk, which can independently drive differences in utilization and quality. For plans that serve many former Medicaid members or members in lower-income neighborhoods, that gap has real consequences.

How Inovalon researchers compared similar Medicare populations

First, the study used a long look-back period. Every participant had three years of continuous commercial or Medicaid coverage before Medicare, followed by two years in the same MA plan or in traditional Medicare, with medical and pharmacy benefits throughout.

Second, the team linked social drivers of health data at the nine-digit ZIP code level, which represents roughly 10 to 20 households. At that level, factors such as income, education, English proficiency, vehicle access, and living alone are far more predictive of outcomes than they are across larger geographic areas.

Third, the team matched the two groups. It applied exact matching on age and sex, then propensity matching on chronic conditions and social risk, so the MA and traditional Medicare groups were comparable before any outcomes were measured. The final matched cohort included 45,530 beneficiaries, with data spanning 2017 through 2023.

Matching changed Medicare Advantage and traditional Medicare quality results

Teigland presented readmissions first, using NCQA’s 30-day all-cause readmissions measure. MA had a lower rate even before matching: 10.8%, compared with 18.7% in traditional Medicare. After the researchers compared beneficiaries with more similar characteristics, the MA rate was 9.1%, while the traditional Medicare rate was 20.6%. The contrast in readmission rates was therefore even sharper in the matched comparison.

Preventable hospitalizations followed a similar pattern. Before matching, the traditional Medicare rate was 1.2 times the MA rate. After matching, it was 1.7 times the MA rate.

“These are not small differences,” Blavin said. Readmissions and preventable hospitalizations are not abstract quality measures: they reflect whether beneficiaries return to the hospital after discharge or experience hospitalizations that may have been avoidable.

The preventive-care findings, however, were not uniformly in MA’s favor. MA had a clear advantage in colorectal cancer screening, with enrollees screened at roughly twice the rate of traditional Medicare beneficiaries. After matching, that gap grew even larger. Traditional Medicare, meanwhile, had higher rates of annual flu vaccination and diabetic eye exams, but those advantages narrowed once the groups were matched.

Together, the results show why comparing similar populations matters. Matching can make some differences more visible and reveal that others are less pronounced than when they first appear. Teigland cautioned that when the populations being compared differ substantially, unadjusted rates can obscure meaningful differences in care and outcomes.

Medicare Advantage vs. traditional Medicare quality results before and after matching

Measure	MA, before matching	Traditional Medicare, before matching	MA, after matching	Traditional Medicare, after matching
30-day all-cause readmissions (lower is better)	10.8%	18.7%	9.1%	20.6%
Preventable hospitalizations (lower is better)	14.4%	17.1%	9.6%	16.4%
Colorectal cancer screening	57.4%	29.4%	56.8%	26.4%
Annual flu vaccine	49.7%	56.3%	49.0%	53.2%
Diabetes care: eye exam	66.1%	76.7%	65.0%	72.7%

Source: Inovalon research presented at the NCQA Health Innovation Summit 2026 pre-conference webinar, data years 2017 through 2023.

What matched quality comparisons mean for your plan

Teigland, who has co-chaired the National Quality Forum’s Scientific Methods Panel for years, knows the debate over risk adjustment in quality measurement well. But her guidance to plans is practical: rather than relying only on an overall rate, examine how results differ across member populations.

Most quality measures are not risk-adjusted for member characteristics, she said, and that is not necessarily a problem. The more useful approach is often to stratify the results and look separately at how different groups perform.

She recommended stratifying results by age and sex, as well as by the number of chronic conditions or a severity measure such as the Charlson Comorbidity Index. Where data is available, plans should also examine results by social drivers of health, including income, education, and language proficiency.

Strong overall results among healthier or more affluent members can often mask poorer outcomes and unmet needs in other groups.

“You really need to account for differences in the populations to find those true quality gaps that might be buried in an average rate somewhere,” Teigland said. The same principle applies beyond Medicare, she added, including when comparing commercial and Medicaid populations, provider groups, and health systems.

Siegel encouraged plans to treat the findings “less as a scorecard and more as a prompt.” Before drawing conclusions from any quality comparison, including comparisons used to assess their own performance, she said plans should ask four questions:

1. Which populations are included in the numerator and denominator?
2. How similar are the populations being compared?
3. What do the measure specifications account for?
4. What do they leave out?

Blavin attributed MA’s results to the structure of the program. Because MA plans bear the financial risk of readmission, he said, they have an incentive to invest in care coordination, chronic disease management, medication management, and supplemental benefits such as transportation and meals. He cautioned that Medicaid cuts and lower MA payments could put pressure on those benefits, with the greatest effect on members facing the most social risk.

As budgets tighten, plans need a clear view of where needs are the most pertinent, and which interventions are most likely to make a difference. That means looking beyond the plan-wide rate to identify variation in outcomes, care gaps, and performance across member populations.

Inovalon’s Converged Quality™ helps plans measure, report on, and improve quality, so your team can move beyond a single plan-wide rate and identify the populations, care gaps, and performance opportunities that demand attention.

The platform processes clinical quality metrics for more than 209 million lives annually, while national and regional benchmarking and population-outcomes analytics help teams put their performance in context and target improvement efforts with greater precision. With 26 consecutive years of NCQA HEDIS® Measure Certification and an NCQA-validated engine for digital quality measures, Converged Quality gives plans a proven foundation for today’s quality reporting requirements.

Medicare quality comparisons can teach HEOR and RWE teams

Teams working in health economics and outcomes research, real-world evidence, and comparative effectiveness research face the same fundamental challenge: patients who receive a therapy, enroll in a health plan, or seek care in a particular setting are rarely a random sample. Without accounting for those underlying differences, an unadjusted comparison can reflect who made the choice, not the effect of the choice itself. That is patient selection bias, and matching is one of the main tools researchers use to address it.

This matched cohort study offers a useful model. A three-year look-back can capture diagnoses and utilization history that a 12-month baseline may miss. Neighborhood-level social risk data adds context that clinical records alone may not capture. Together, those inputs, combined with exact and propensity score matching, can help researchers construct more comparable study populations and produce a more defensible assessment of the exposure under study.

Data access mattered as much as study design. Much of the existing research comparing Medicare Advantage and traditional Medicare has relied on beneficiaries who switch programs, Teigland noted, because MA encounter data at scale has historically been difficult to access. This study instead followed people from commercial or Medicaid coverage into their first Medicare coverage decision at age 65, using encounter data from MA plans of varying sizes across the country.

Inovalon’s Outcomes and Evidence Solutions team gives life sciences teams the real-world data foundation and research expertise to design more credible real-world evidence studies. With longitudinal patient-level data, linked context across the patient journey, and experienced research scientists, your team can move beyond surface-level comparisons and generate evidence designed to stand up to scrutiny.

Watch the full webinar

The complete session is available on demand, including the full quality results, the panel discussion, and the audience’s Q&A.

Watch the Recording

 

Frequently asked questions about Medicare Advantage vs. traditional Medicare quality

How did the study compare Medicare Advantage and traditional Medicare fairly?

It measured each person’s health from commercial or Medicaid coverage in the three years before Medicare, linked social drivers of health data at the nine-digit ZIP code level, and used exact and propensity matching on age, sex, chronic conditions, and social risk. The final matched cohort included 45,530 beneficiaries.

Who performed the study?

Inovalon’s research science team, led by Christie Teigland, Ph.D., Vice President of Research Science and Advanced Analytics, conducted the study in collaboration with researchers from AHIP and the Blue Cross Blue Shield Association. Teigland has served as principal investigator on NCQA measure development projects for all-cause 30-day readmissions and potentially avoidable hospitalizations, two of the outcomes at the center of the study and has co-chaired the National Quality Forum’s Scientific Methods Panel. To learn more about her and her team’s publications and work, visit our Meet the Experts page.

Did Medicare Advantage outperform traditional Medicare on quality?

On several measures, yes. In this study, MA had lower rates of 30-day readmissions and preventable hospitalizations and a higher rate of colorectal cancer screening, and each gap widened after matching. Traditional Medicare performed better on annual flu vaccination and diabetic eye exams, though both gaps narrowed after matching. Because the study was observational, the findings show how outcomes differed between comparable populations; they do not on their own prove that program enrollment caused the differences.

Why does a pre-Medicare baseline matter?

Diagnoses recorded before enrollment are not affected by differences in how the two programs document conditions. A three-year look-back also captures conditions that are not coded every year, which a 12-month baseline can miss.

Why should health plans stratify quality measures?

A plan-wide average can mask meaningful differences in outcomes across member populations. Stratifying results by age, sex, chronic-condition burden, severity, and available social drivers of health helps your plan see where care gaps and improvement opportunities are concentrated.

How should health plans stratify quality measures?

Teigland recommended starting with age group and sex, then adding the number of chronic conditions or a severity measure such as the Charlson Comorbidity Index, along with social drivers of health such as income, education, and language proficiency.

What makes a real-world comparison more defensible?

A more defensible real-world comparison starts with a baseline period long enough to capture each patient’s history, adds relevant clinical and social risk context, and matches study populations on observable characteristics before measuring any outcomes. Those steps help separate differences in who receives a treatment or enrolls in a program from the effect of the treatment or program itself.

Sources

All external sources are hyperlinked in text. Study findings and quotes come from the webinar. 

In order of appearance: Inovalon research and panelist remarks from the NCQA Health Innovation Summit 2026 pre-conference webinar, Sept. 23, 2026 (study design, findings, and quotes); Business Wire, July 23, 2026 (Converged Quality customer base, lives processed, and NCQA certification). Internal: Christie Teigland expert page, Converged Quality product page. 

HEDIS® is a registered trademark of the National Committee for Quality Assurance (NCQA). 

---

Inovalon and design® and Inovalon® are trademarks of Inovalon, Inc.