News & Insights

White paper: A pilot of a patient-centered primary care survey measure in a Medicare Advantage setting

Abstract

The rapid aging of the Medicare population requires a health system adapted to their unique needs.  The Age Friendly Health System framework is a rapidly growing response to that need.  In this system, patient goal attainment, known as “What Matters” in the framework, is a critical pillar.  Measurement of this concept remains undefined and is the subject of active measure development.  This report presents evaluation of a pilot of a patient-centered primary care survey measure.  This measure performed well on member response, cost of administration, uniqueness of measured factors, and actionable insights gained.

Introduction

This year, 2026, the first baby boomers are turning 80.  The number of older old and oldest old Americans is set to transform the demands on Medicare.  It is within this context that a national consensus is beginning to develop on the need for a health system responsive to the needs of the aging.  The Age Friendly Health System framework[1] is emerging as a leading contender for defining how payers and clinicians should organize care for the oldest old.  This report addresses the question of how one measures quality in the context of this framework, specifically, how one measures goal concordant care.  Is the healthcare system facilitating attainment of the patient’s goals?

Traditional measurement of quality in CMS has utilized HEDIS screening and chronic condition management goals.  The challenge of that framework is that patients become excluded from the measures as they age.  Breast cancer screening and colonoscopies no longer are recommended after 75.  Measurement of comprehensive diabetes care and its successor measures likewise cease to be applied to patients after 75.  And Blood pressure management recommendations cease to be applied after age 85.  Medicare beneficiaries age out of the current CMS quality measurement system long before many of them die, leaving Medicare without a standard of care for the oldest old.

The Age Friendly Healthcare framework is organized around 4 M’s, which are What Matters, Medication, Mentation and Mobility[1].  The subject of this paper is the measurement of What Matters.  The concept of “What Matters” to a patient is a challenge for measure development.  Patient centered outcomes measures have received much attention and thoughtful development work, and the Age Friendly “What Matters” concept of quality fits squarely in this tradition.  This report contrasts a provider documented measure sponsored and piloted by NCQA with a survey measure piloted by a Medicare Advantage Organization

The National Commission for Quality Assurance has been piloting goal attainment measures in which patients and providers collaboratively establish goals, set scaled targets, and monitor and measure progress towards the goals[2].  These measures may present challenges such as the time required to document progress, the shifting nature of goals in complex older populations, and the difficulty of abstracting data for consolidation of results at scale.   Independence Blue Cross, a Medicare Advantage Organization in Southeast Pennsylvania covering 120,000 Medicare Advantage lives, explored piloting the NCQA measure in its care management units.  NCQA is piloting the measures more broadly and results at a wider range of health plans may offer different results.

This paper presents findings from a pilot test of a patient level survey measure evaluating the physician role in goal attainment.  The findings suggest that the measure is a viable means of evaluating physician facilitation of goal attainment for Medicare beneficiaries.  Members respond to it at high rates, it captures unique factors of the patient experience relative to CAHPS questions, it is amenable to multivariate statistical analysis and yields significant relationships with predictors that offer insights into drivers of the patient experience.  The measure is relatively low cost to implement as it is structured as an add on to the CAHPS survey.

Quality measures for aging populations are desirable elements in a Stars program.  They can create incentives to develop models of care responsive to the unique needs of an aging population.  Criteria for a useful measure would include respondent receptivity, low cost of data collection, ease of data ingestion, low burden on providers, and high potential to yield insights on plan operations that impact the measured concept.  The patient-centered primary care measure evaluated here met these criteria.  It was appended to a CAHPS survey instrument, was ingested through the normal data capture method, was transparent to providers, and by linking survey responses to individual members it yielded rich insights on plan activities that supported goal attainment.

Method

The measure evaluated here is an adaptation of a measure taken from the Person-Centered Primary Care measurement set described by Elz, et al.[1] and accessed through the Larry A Green Center[2].  We adapted the survey question, “Over time, my personal doctor helps me meet my goals” to a CAHPS framework to more closely match the format of the CAHPS survey in which it was embedded.   The original format of the question asked the respondent to rate their experience on a scale of 1 to 10.  The potential responses were Never, Sometimes, Usually, Always. The survey was fielded in late summer 2025 to 35,500 Medicare Advantage members in Southeast Pennsylvania sampled using simple random sampling.  The survey yielded 8,273 unique responses, an overall survey response rate of 23%.  This measure was added to a preexisting biennial survey conducted for research into drivers of CAHPS scores.

We present the overall results of the survey measure alongside analyses of the measure compared to key CAHPS outcomes and composite measures. Additionally, we examine differences in response by age group using logistic regression to estimate the probability of a top-box (“Always”) response.

Results

Table 1 presents the distribution of valid responses along with selected summary statistics for the new measure. The measure had a high rate of completion, receiving 7,815 (94.5%) valid responses and 458 (5.5%) respondents with no or missing response. We find that 56% of respondents had a top-box, or “Always”, response, 26% responding “Usually”, 12% for “Sometimes”, and lastly 6% of respondents responded “Never” to the measure.

Additionally, 82% of total respondents indicated that their doctor helps them meet their goals either “Usually” or “Always.”

We examined the Spearman correlation between our measure, “Over time, my personal doctor helps me meet my goals” and 4 current CAHPS composites or measures: Overall rating of the health plan, Overall rating of all healthcare, Getting needed care, and Care coordination, and the results are summarized in Figure 2. We find that our measure of goal concordance is very weakly correlated with core CAHPS measures, with correlation coefficients ranging from 0.24 (Health Plan Rating) to 0.39 (Care Coordination Composite). This notion is reinforced by the finding that Cronbach’s alpha remains largely unchanged (Δ = -0.01) when the goals measure is removed from the set of CAHPS outcomes. This suggests that the patient reported measure of goal concordance is independent of other selected CAHPS measures and may be capturing unique factors of the patient experience not otherwise reflected in the current CAHPS survey instrument.

Lastly, to assess any differences in responses by age group we estimate the likelihood of a member giving a response of “Always” conditional on which age group the member belongs to. We find that those belonging to age groups 65 – 74 and 75 – 84 were  +17.7% (p <0.001) and +14.0% (p <0.001), respectively, more likely to respond “Always” than those aged 85+.

Discussion and Recommendations

This study explores the use of an adaptation of a Patient Centered Primary Care Measure survey element in the context of a CAHPS survey.  The goal was to evaluate a measure concept for the “What Matters” pillar of the Age Friendly Health System framework.  This pilot demonstrated that respondents will answer the question at rates similar to other CAHPS items, that it captures unique elements of the health care experience relative to other CAHPS measures, and that the cost of administration and data capture are minimal.  This survey measure appears to meet basic criteria for validity.

The responses revealed a statistically significant relationship between patients’ perception of physicians’ helping them attain their goals and age.  Patients over 85 reported lower scores than those under 85; age and goal concordance are inversely correlated.

In the context of a measure for an Age Friendly System of care this is a meaningful finding.  The central thesis of this inquiry is that current measurements used in Medicare such as comprehensive diabetes care and its successor metrics and screenings for cancer miss the aging population and new measures are needed.  The Age Friendly Health System has established What Matters as a central pillar of geriatric care.  The finding about the inverse relationship between age and physician facilitated goal attainment is concerning because at the older ages, treatment modalities for different conditions come into conflict, tradeoffs ensue, and patient wishes are the only true guide to optimized care.

The finding that physicians somehow are less effective at assisting members in goal attainment just when that function becomes most needed signals the amount of work to be done in the healthcare system.  This kind of transformation requires quality measurement, feedback, and incentives.

One critique frequently aimed at survey measures is that their subjectivity impedes their effectiveness at identifying actionable insights.  In fact, this survey measure was able to identify plan level operational factors that influence patient perceptions of physician ability to facilitate goal attainment.

The survey responses were individually identifiable and were linked to over 100 variables from health plan data, with ~50 across multiple domains added to the previously described logistic regression.  The domains included medical utilization, medical cost, pharmacy cost and utilization, health status and conditions, administrative interactions with the health plan, provider characteristics and patient demographics.  One example of a statistically significant covariate was first call resolution failure.  That is, if the member was unable to reach resolution of a concern in the first call, their experience of physician assistance in goal attainment declines.

The individually identifiable survey measure permits the health plan to investigate first call experiences of specific members and identify drivers of failures and craft interventions.  This is not true of a measure recorded in patient medical records that measures one specific element of the patient/physician intervention.  It is not clear how a health plan would impact one specific goal of the kind the NCQA pilot is studying, such as walking to the mail box.  A health plan can influence the overall experience of provider and patient through benefits availability, medical necessity determinations, and call center effectiveness.

Conclusion

This paper presents the results of a pilot study of a Person-Centered Primary Care Measure survey element.  The context for the study is the ongoing quest to develop a measure of “What Matters” for the Age Friendly Healthcare movement that could be applied to Medicare Advantage payers.  The survey measure attained a high response rate, was inexpensive to administer and analyze, capture unique elements of patient experience not captured in other CAHPS measures, and offered system level insights that could be used to improve goal concordant care for Medicare Advantage plan members.  These are attractive features of a quality measure for Medicare Advantage.  The patient survey format avoids imposing burdens on providers, the challenges of evolving patient goals if conditions change, and the costs of medical record data collection.  The inclusion of this measure in the stars measure set could create incentives for plans to support the Age Friendly Health System framework through effective interventions at scale.

Authors
Eric Cahow, Ph.D., Vice President of Government Markets Performance, Independence Blue Cross
Eric Wilkinson, Ph.D., Senior Research Scientist, Independence Blue Cross

Appendix 1.  Spearman Correlation Coefficients with selected CAHPS measures

Table 1: Distribution and Summary Statistics for Goals Measure

 

Figure 1: Correlation Matrix for Concordant Care and CAHPS Core Measures

 

[1] Author affiliations:  Independence Blue Cross, Philadelphia, PA.  Eric.Cahow@IBX.com

 

[2] Fulmer, et al.  2018. The Age-Friendly Health System Imperative.  JAGS.  66:22–24, 2018

 

[3] Fulmer, et al.  2022.  Age Friendly Health Systems:  A Guide to Using the 4Ms while Caring for Older Adults.  Institute for HealthCare Improvement.

 

[4] Mohanty, Sarita and Terry Fulmer.  2022.  “Person-Driven Outcome Measures Help Achieve Equitable Outcomes” in Health Affairs Forefront.  June 13, 2022.

 

[5] Etz, Rebecca S, et al. 2019.  A New Comprehensive Measure of High-Value Aspects of Primary Care. The Annals of Family Medicine May 2019, 17 (3) 221-230

 

[6] Larry S Green Center.  https://www.green-center.org/pcpcm