Personal physician access by preferred language among Medicare Advantage and Medicare Fee-for-Service older adults

INTRODUCTION

A personal physician is an ongoing medical advisor who assumes primary responsibility for a patient's care¹ and is vital for promoting health and improving the patient's quality of care received. The continuity of care from having a personal physician fosters patient–provider relationships, resulting in better communication, trust, and satisfaction.^{2, 3} It promotes access to preventive care,⁴ improves health outcomes,⁵ and lowers healthcare costs.

Using 2012 Medicare Consumer Assessment of Healthcare Providers and Systems (MCAHPS) survey data, Martsolf and colleagues⁶ found that Medicare Advantage (MA) enrollees were more likely to report having a personal physician than those in Medicare Fee-for-Service (FFS). That study also revealed that Spanish language survey respondents were less likely than other sociodemographic groups to have a personal physician. However, the analysis did not include Puerto Rico (PR) residents, a major Spanish-speaking group living in a US jurisdiction.

Participation in MA, particularly among Hispanic and Black individuals with Medicare, has grown rapidly in the past decade.⁷ It is unknown whether this increased MA enrollment has affected access to a personal physician for groups who were less likely to have one a decade ago.

We therefore used data from the 2022 MCAHPS survey to update our understanding of patterns in having a personal physician among Medicare enrollees. Furthermore, we evaluated whether MA enrollment among Spanish-responding older adults, including those in PR, is associated with greater likelihood of having a personal physician than FFS.

METHODS

Data were drawn from the 2022 MCAHPS surveys (both MA and FFS). We included 292,700 respondents who were 65 years of age or older, responded in English or Spanish, and answered the personal physician item on the survey (98%; “A personal doctor is the one you would see if you need a check-up, want advice about a health problem, or get sick or hurt. Do you have a personal doctor? Yes/No”). We compared respondents on this binary measure of whether they had a personal physician. Respondents were classified into four language/location categories based on survey language completion, a respondent's predicted language preference,⁸ and location of primary residence: English-responding, low probability Spanish preferring; English-responding, high probability Spanish-preferring; Spanish-responding people within PR; and mainland US Spanish-responding people. A linear regression model predicted self-report of having no physician from MA versus FFS coverage, language/location categories, and the interaction of coverage and language/location to test for differences by coverage and language/location.

RAND's Human Subjects Protection Committee approved this research.

RESULTS

Table 1 shows the weighted percentages of language/location groups within FFS (n = 70,871) and MA (n = 221,829). In MA, 6.1% of responses were in Spanish and 4.3% were in English among those with a high probability of Spanish preference. In FFS, 1.9% of responses were in Spanish and 2.2% were in English among those with a high probability of Spanish preference. Figure 1 shows the proportion of respondents without a personal physician by language/location within MA and FFS. The overall percentage of people without a personal physician was lower in MA (4.3%) than FFS (6.0%). The magnitude of this difference was notably larger for Spanish-responding people both in the mainland US (6.8% MA vs. 16.5% FFS) and within PR (4.7% MA vs. 12.9% FFS), and somewhat larger for English-responding people with a high probability for Spanish preference (7.3% MA vs. 11.7% FFS) relative to English-responding people with a low probability for Spanish preference (4.1% MA vs. 5.7% FFS), with p < 0.001 for the language/location by coverage interactions. This indicates that MA enrollment is more strongly associated with having a personal physician among Spanish-responding people (both in PR and the mainland) and English-responding people who have a high probability of Spanish preference than among English-responding people with a low probability of Spanish preference.

TABLE 1. Sample sizes and weighted percentages, by language/location and coverage.

Coverage type	Language/location	N	Weighted %
FFS	English-responding, low probability Spanish-preferring	68,689	95.9
	English-responding, high probability Spanish-preferring	981	2.2
	Spanish-responding, mainland	473	1.5
	Spanish-responding, Puerto Rico	728	0.4
MA	English-responding, low probability Spanish-preferring	203,380	89.7
	English-responding, high probability Spanish-preferring	7849	4.3
	Spanish-responding, mainland	7745	3.9
	Spanish-responding, Puerto Rico	2855	2.2

FIGURE 1

Proportion of respondents reporting not having a personal physician and 95% confidence intervals, by language/location stratified by probability of Spanish preference and coverage.

DISCUSSION

The disparity in having a personal physician for Spanish-responding people with Medicare has decreased since 2012⁶ but remains large in FFS. MA participation may have connected enrollees to personal physicians through several mechanisms. For example, MA plans generally require the selection of a primary care provider (PCP), and some assign a PCP to participants. Moreover, the positive interactions of MA and Spanish language suggest that MA plans are particularly successful in linking Spanish-preferring people to personal physicians. Further research is necessary to understand how these differences affect disparities in care for Spanish-preferring people with Medicare, assess whether activities that MA plans use to connect enrollees to providers are applicable to other populations and settings, and monitor the association between having a personal physician and patient experience.

AUTHOR CONTRIBUTIONS

Malcolm Williams contributed to the analysis and interpretation of data and drafting of the article. Marc N. Elliott contributed to the conception and design, acquisition of data, analysis and interpretation of data, and revising the manuscript for important intellectual content. Katrin Hambarsoomian contributed to the analysis and interpretation of data, drafting the article, and revising the manuscript for important intellectual content. Steven C. Martino contributed to the analysis and interpretation of data, and revising the manuscript for important intellectual content. Amelia Haviland contributed to the analysis and interpretation of data, and revising the manuscript for important intellectual content. Rob Weech-Maldonado contributed to the analysis and interpretation of data, and revising the manuscript for important intellectual content. Aditi Mallick contributed to the analysis and interpretation of data, and revising the manuscript for important intellectual content. Sarah Gaillot contributed to the acquisition of data, analysis and interpretation of data, and revising the manuscript for important intellectual content. Sarah Johaningsmeir contributed to the analysis and interpretation of data, and revising the manuscript for important intellectual content. Nate Orr contributed to analysis and interpretation of data, and revising the manuscript for important intellectual content. Debra Saliba contributed to the analysis and interpretation of data, and revising the manuscript for important intellectual content.

ACKNOWLEDGMENTS

The authors would like to thank Katherine Osby for assistance in preparing the manuscript.

FUNDING INFORMATION

This study was funded by the Centers for Medicare & Medicaid Services (contract/task order: GS-10F-0275P/75FCMC20F0101).

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

SPONSOR'S ROLE

Aditi Mallick, Sarah Gaillot, and Sarah Johaningsmeir are employees of the sponsoring agency, the Centers for Medicare & Medicaid Services.