Assessing eligibility of Medicare beneficiaries with chronic conditions for the ACCESS model

Abstract

Introduction

The Advancing Chronic Care with Effective Scalable Solutions (ACCESS) model is a new outcome-aligned payment model intended to support technology-enabled chronic disease management among Traditional Medicare (TM) beneficiaries. However, the national population potentially eligible for ACCESS remains poorly characterized, particularly those with multiple chronic conditions.

Methods

Using nationally representative survey data from the 2023 Medical Expenditure Panel Survey, we described the health characteristics and health care utilization patterns of TM beneficiaries potentially eligible for the ACCESS model.

Results

A total of 56.3% of beneficiaries (estimated 9.1 million) were potentially eligible to participate in the model, 24% of beneficiaries (estimated 2.2 million) had multiple chronic conditions, and beneficiary telehealth use for cardio-kidney-metabolic and musculoskeletal conditions was limited.

Conclusion

The sizable proportion of TM beneficiaries potentially eligible for more than one ACCESS track suggests a deliberate approach to care coordination to support technology-enabled chronic disease management in older adults.

Introduction

Older adults insured by Medicare experience a high burden of chronic diseases that frequently co-occur, including cardio-kidney-metabolic (CKM), musculoskeletal (MSK), and behavioral health (BH) conditions.^1-3 Management of these conditions often requires longitudinal care delivered across primary care and outpatient settings, which may be enhanced by digital health technologies such as telehealth.⁴ However, fee-for-service payment structures in Traditional Medicare (TM) have historically provided limited incentives for technology-enabled care.⁵

The Advancing Chronic Care with Effective Scalable Solutions (ACCESS) model is a new voluntary outcome-aligned payment model intended to support technology-enabled chronic disease management among TM beneficiaries.⁶ The ACCESS model will reimburse both digital health companies providing technology-enabled care to beneficiaries (ACCESS organizations) and beneficiaries’ primary care or referring physician responsible for coordinating care. The Centers for Medicare and Medicaid Services (CMS) began implementing the ACCESS model in July 2026, with the model scheduled to run for 10 years. However, the national population potentially eligible for ACCESS remains poorly characterized, particularly those with multiple chronic conditions.

As CMS continues implementing the ACCESS model, understanding the size, health characteristics, and health care utilization patterns of the potentially eligible population will be important for guiding implementation and evaluating the model's reach and impact. We therefore analyzed nationally representative survey data to characterize TM beneficiaries potentially eligible for the ACCESS model.

Data and methods

Data source

We used 2023 Medical Expenditure Panel Survey (MEPS) data, a national survey of community-dwelling persons in the United States that catalogs healthcare utilization and expenditures. We identified individuals with outpatient visits linked to ACCESS-specific qualifying chronic conditions by combining the following MEPS files: Full-Year Consolidated File, Office-Based Visits File, Outpatient Visits File, and Medical Conditions File.^7,8

We included individuals with ≥1 outpatient visit and continuous TM coverage during 2023 to identify individuals who had stable TM coverage who were actively receiving care.⁹ Outpatient visit diagnoses were classified using Clinical Classifications Software Refined (CCSR) codes, which groups International Classification of Diseases 10th Revision (ICD-10) codes into clinically meaningful categories,¹⁰ to determine potential eligibility for the 4 ACCESS clinical tracks that cover common chronic conditions: early cardio-kidney-metabolic (eCKM; hypertension, dyslipidemia, obesity, prediabetes), CKM (diabetes mellitus, chronic kidney disease, atherosclerotic cardiovascular disease), MSK (chronic musculoskeletal pain), and BH (depression, anxiety). Individuals were classified as track-eligible based on track-specific inclusion and exclusion criteria,⁶ with full methodology detailed in the Supplementary material.

We also examined telehealth, a common modality of digital health.¹¹ For each outpatient visit, respondents reported if the visit was in person or via telehealth. Overall telehealth use was defined as having ≥1 telehealth outpatient visits for any medical condition. ACCESS track-eligible telehealth use was defined as having ≥1 telehealth outpatient visit associated with qualifying medical conditions for the specified ACCESS track.

Statistical analysis

Using MEPS survey weights to account for the complex survey design and nonresponse, we generated national estimates of TM beneficiaries potentially eligible for ACCESS tracks as well as demographic characteristics and healthcare utilization of individuals across ACCESS tracks. MEPS does not contain a CCSR code for prediabetes, resulting in potentially ambiguous eCKM eligibility status for some individuals. We classified these individuals as potentially eCKM-eligible in our primary analysis and as potentially eCKM-ineligible in a sensitivity analysis.

Analyses were conducted using Stata v19.5 (StataCorp) and independently verified using R v.4.5.1. The University of Utah institutional review board exempted this study from review and informed consent, as we used deidentified, publicly available data.

Results

Our study included 1272 respondents who were classified as TM beneficiaries with ≥1 outpatient visits in 2023, corresponding to an estimated 16.2 million (95% confidence interval [CI]; 14.9–17.6M) beneficiaries nationally. Overall, 56.3% (95% CI; 53.0%-59.5%) were potentially eligible for ≥1 ACCESS track, corresponding to an estimated 9.1 million (95% CI; 8.2–10.0M) beneficiaries (Figure 1). Among beneficiaries potentially eligible for any ACCESS track, 24.0% (95% CI, 20.2%-28.3%) qualified for ≥2 ACCESS tracks.

Figure 1.

Estimated ACCESS track eligibility and overlap in traditional medicare beneficiaries, main analysis. Source: Authors’ analyses of data from the 2023 Medical Expenditures Panel Survey. Abbreviations: ACCESS, Advancing Chronic Care with Effective Scalable Solutions; BH, behavioral health; CKM, cardiovascular-kidney-metabolic; eCKM, early cardiovascular-kidney-metabolic; MSK, musculoskeletal. Estimates use sample weights to account for the complex design of the Medical Expenditure Panel Survey and are shown as weighted national estimates and weighted percentages of traditional medicare beneficiaries in our study population (16.2M total, 9.1M ACCESS-eligible, and 7.1M ACCESS-ineligible). For overlap analyses, CKM and eCKM were combined into a single CKM domain consistent with the ACCESS model, where an individual is either classified as eCKM or CKM but not both. Counts are displayed in millions.

Among ACCESS tracks, CKM had the largest number of potentially eligible beneficiaries (3.9 million [95% CI; 3.3–4.4M]), followed by MSK (3.8 million [95% CI; 3.2–4.3M]), eCKM (2.7 million [95% CI; 2.2–3.2M]), and BH (1.2 million [95% CI; 0.9–1.5M]) (Table 1).

Table 1.

Characteristics of Traditional Medicare beneficiaries potentially eligible for ACCESS tracks.

	ACCESS Trackb
Characteristic	eCKM			CKM			MSK			BH
Weighted estimate of eligible beneficiaries, No. (95% CI)	2 670 924 (2 167 331-3 174 518)			3 863 214 (3 331 478-4 394 951)			3 769 644 (3 191 265-4 348 022)			1 225 586 (923 344-1 527 828)
	Unweighted, no	Weighted %/mean	95% CI	Unweighted, no	Weighted %/mean	95% CI	Unweighted, no.	Weighted %/mean	95% CI	Unweighted, no	Weighted %/mean	95% CI
Age, mean years	204	73	72-74	301	73	72-75	299	73	72-74	86	68	65-71
<65 y	10	4.9%	2.5-9.3	26	8.4%	5.5-12.5	19	8.1%	4.4-14.5	22	29.1%	18.9-41.9
65 y and older	194	95.1%	90.7-97.5	275	91.6%	87.5-94.5	280	91.9%	85.5-95.6	64	70.9%	58.1-81.1
Sex
Men	81	40.6%	33.3-48.3	152	49.9%	43.7-56.0	136	45.3%	39.4-51.3	31	37.9%	27.0-50.3
Women	123	59.4%	51.7-66.7	149	50.1%	44.0-56.3	163	54.7%	48.7-60.6	55	62.1%	49.7-73.0
Race and Ethnicity
Hispanic	11	8.0%	4.3-14.4	15	3.6%	1.7-7.4	9a	2.9%	1.3-6.2	2a	2.0%	0.5-8.0
Non-Hispanic White	171	82.2%	74.9-87.6	246	83.0%	77.1-87.6	260	87.0%	81.2-91.1	76	89.8%	80.8-94.9
Non-Hispanic Black	16	6.9%	3.8-12.3	24	7.8%	4.8-12.6	15	4.5%	2.2-8.7	6a	4.3%	1.6-11.2
Non-Hispanic Asian	3a	2.1%	0.6-7.0	8a	2.6%	1.1-6.0	4a	2.3%	0.7-7.4	1a	2.6%	0.4-16.6
Non-Hispanic other or multiracial	3a	0.9%	0.2-3.2	8a	2.9%	1.4-6.1	11a	3.4%	1.7-7.0	1a	1.3%	0.2-8.6
Education level
High school or lower	82	43.9%	36.5-51.5	160	52.9%	46.0-59.6	116	40.0%	33.4-46.9	37	39.4%	30.1-49.6
Bachelor's degree or higher	122	56.1%	48.5-63.5	139	47.1%	40.4-54.0	182	60.0%	53.1-66.6	49	60.6%	50.4-69.9
Income
<100% of FPL	16	5.9%	3.3-10.4	31	8.0%	4.9-12.7	20	4.9%	2.8-8.6	11	8.2%	3.7-17.2
100 to <200% of FPL	25	12.0%	7.0-19.8	48	14.6%	10.2-20.5	31	10.1%	6.5-15.4	16	16.0%	9.8-25.2
200 to <400% of FPL	55	29.2%	21.5-38.4	88	27.4%	21.6-34.1	90	28.9%	23.3-35.2	15	18.5%	10.2-31.3
≥400% of FPL	108	52.9%	44.2-61.4	134	50.0%	43.2-56.8	158	56.1%	49.6-62.3	44	57.2%	44.5-69.1
Self-reported health status
Good, very good, or excellent	176	86.3%	79.7-91.0	210	70.0%	63.9-75.5	248	79.8%	73.1-85.2	58	62.0%	49.2-73.3
Fair or poor	26	13.7%	9.0-20.3	88	30.0%	24.5-36.1	51	20.2%	14.8-26.9	28	38.0%	26.7-50.8
Outpatient visits, mean	204	19	16-22	301	24	21-28	299	29	25-33	86	37	24-49
Number of chronic medications,c mean	180	4.6	(4.1-5.2)	269	6.6	(6.0-7.1)	265	4.8	(4.3-5.2)	76	6.2	(5.1-7.3)
Number of ED visits
0	163	79.7%	72.3-85.5	211	69.4%	64.2-74.1	238	79.7%	73.8-84.5	59	71.2%	59.8-80.4
1	30	14.3%	9.7-20.7	63	21.4%	16.6-27.2	45	14.6%	10.5-19.9	19	20.9%	13.2-31.6
>1	11	6.0%	3.1-11.4	27	9.2%	5.9-14.2	16	5.7%	3.1-10.4	8a	7.9%	3.3-17.7
Number of hospitalizations
0	178	86.7%	80.0-91.4	233	77.0%	71.7-81.6	245	80.9%	75.6-85.3	66	79.2%	67.2-87.6
1	19	9.2%	5.4-15.3	53	16.9%	12.8-22.1	43	14.3%	10.6-19.1	14	14.1%	7.3-25.4
>1	7a	4.1%	1.9-8.4	15	6.1%	3.4-10.5	11	4.8%	2.5-9.1	6a	6.7%	2.9-15.0
Overall telehealth use
No	154	78.0%	69.9-84.4	252	80.8%	74.2-86.1	238	79.5%	73.4-84.4	40	43.5%	32.5-55.2
Yes	50	22.0%	15.6-30.1	49	19.2%	13.9-25.8	61	20.5%	15.6-26.6	46	56.5%	44.8-67.5
ACCESS track-eligible telehealth use
No	189	94.1%	89.1-96.9	287	96.0%	92.7-97.8	296	98.6%	95.6-99.6	49	54.9%	42.5-66.8
Yes	15	5.9%	3.1-10.9	14	4.0%	2.2-7.3	3a	1.4%	0.5-4.4	37	45.1%	33.2-57.5

Source: Authors’ analyses of data from the 2023 Medical Expenditures Panel Survey.

Estimates use sample weights, and the 95% CIs account for the complex design of the Medical Expenditure Panel Survey. Respondents with missing data were not included in the generation of national estimates.

Abbreviations: ACCESS, Advancing Chronic Care with Effective Scalable Solutions; BH, behavioral health; CI, confidence interval; CKM, cardiovascular-kidney-metabolic; ED, emergency department; eCKM, early cardiovascular-kidney-metabolic; FPL, federal poverty level; MSK, musculoskeletal.

^aEstimates may be unreliable due to unweighted sample <10.

^bACCESS track columns are not mutually exclusive, and beneficiaries may qualify for multiple tracks. However, per ACCESS eligibility guidelines, individuals cannot qualify for both eCKM and CKM. In this table, eCKM and CKM are presented separately to illustrate differences in beneficiary characteristics and health care utilization across early vs established CKM populations.

^cChronic medications (≥30-day supply) are counted only among patients with ≥1 prescription fill with a nonmissing days-supply value. Patients who had medications, but no recordable days’ supply were excluded from this row (unweighted N): 167 overall in the cohort, and by track—eCKM 24, CKM 32, MSK 34, BH 10. All other rows use the full track denominator.

Across all tracks, overall telehealth use exceeded ACCESS track-eligible telehealth use. This difference was greatest among potential MSK-eligible beneficiaries, where 20.5% (95% CI, 15.6%-26.6%) of individuals had a telehealth visit for any medical condition while 1.4% (95% CI, 0.5%-4.4%) had a telehealth visit for an MSK track-specific qualifying medical condition. This difference was smallest among potential BH-eligible beneficiaries (56.5% [95% CI, 44.8%-67.5%] vs 45.1% [95% CI, 33.2%-57.5%]).

In the eCKM sensitivity analysis assuming nonprediabetes status, an estimated 1.9 million (95% CI, 1.5–2.2M) beneficiaries remained potentially eligible for eCKM (Supplementary material).

Discussion

More than half of TM beneficiaries in our study were potentially eligible for at least 1 ACCESS track, with approximately 1 in 4 of those beneficiaries qualifying for multiple tracks. These findings highlight the substantial burden of chronic disease and multimorbidity among older adults insured by TM, corresponding to millions of beneficiaries nationwide.

CMS has estimated that more than two-thirds of TM beneficiaries would be eligible for the ACCESS model,⁶ compared with 56.3% in our study. This difference may reflect our restriction to TM beneficiaries with at least 1 outpatient visit in 2023, as well as our use of survey-reported medical conditions to determine ACCESS eligibility. This may underestimate the number of potentially eligible individuals who would qualify for ACCESS but did not have an outpatient visit in 2023.

Beyond estimating the size of the potentially eligible population, our study found that approximately one-quarter of ACCESS-eligible beneficiaries potentially qualified for multiple tracks. Because ACCESS Organizations will provide care alongside patients’ existing primary care clinicians, the high prevalence of potential multitrack eligibility underscores the importance of effective care coordination to minimize fragmented care and mitigate additional strain on overburdened primary care teams.^12-16 Experience among Veterans receiving care through both the Veterans Health Administration (VHA) and VHA-purchased community care illustrates how expanding access across multiple providers can introduce additional care coordination challenges.^17,18 As CMS has announced future expansion of ACCESS to cover additional chronic conditions such as chronic obstructive pulmonary disease, heart failure, and substance use disorders, the need for effective care coordination will be even more important.¹⁹

Furthermore, our study provides new insights into how potentially eligible TM beneficiaries use telehealth. ACCESS track-eligible telehealth use was less common than overall telehealth use across all ACCESS tracks. While many TM beneficiaries have previously used telehealth, telehealth use for non-BH conditions was limited, highlighting both opportunity and potential challenges for technology-enabled CKM and MSK disease management among older adults who may not regularly use digital health technologies for chronic disease management.²⁰ A recent national survey of US adults aged 50-80 years found greater use of digital health technologies among privately insured individuals (40.2% [95% CI, 37.3%-43.1%]) than among those covered only by Medicare (33.6% [95% CI, 30.7%-36.3%]), highlighting the potential role of payment structures in facilitating technology-enabled care.¹¹

This study has several limitations. First, we used MEPS CCSR coding for medical conditions to assign ACCESS track eligibility, which are self-reported and thus subject to potential recall bias and misclassification.²¹ Although ACCESS track eligibility is based on ICD-10 codes, MEPS only provides ICD-10 base codes, which lack the granularity of ICD-10 extended codes that CMS will ultimately use. However, our objective was to characterize the population potentially eligible for the model using clinically meaningful disease groupings rather than replicate administrative claims-based eligibility using ICD-10 codes designed for billing. Thus, our estimates represent approximations of potential ACCESS eligibility which may be compared against claims-based analyses. Second, not all ACCESS inclusion and exclusion criteria could be operationalized. For example, MEPS lacks a CCSR code for prediabetes, which limits eCKM inclusion eligibility classification. Our sensitivity analysis focusing on individuals potentially eCKM-eligible allowed us to estimate the upper and lower bounds of eCKM eligibility status; however, this large difference (an estimated 0.8 million individuals) should be contextualized with future studies that use claims-based data. Third, while we were able to use MEPS CCSR codes for the inclusion criteria of all tracks, there were fewer corresponding CCSR codes for exclusion criteria (Supplementary material); thus, our estimates likely overestimate the number of potentially eligible individuals. Fourth, our study used MEPS data from 2023, which do not collect data for other digital health technologies besides telehealth use. Thus, our findings do not fully reflect current or future patterns of ACCESS eligibility or other technology-enabled care (eg, remote physiologic monitoring or mobile health applications). Fifth, some subgroups had low number of respondents (eg, Hispanic individuals) resulting in potentially unreliable estimates that should be interpreted with caution. Future studies should prioritize adequate representation of smaller subgroups to identify heterogeneity that may be obscured in population-level estimates.

Conclusion

In summary, our findings have important implications for the implementation of the ACCESS model among patients, clinicians, and ACCESS Organizations. The sizable proportion of TM beneficiaries potentially eligible for more than one ACCESS track suggests a deliberate approach to care coordination to support technology-enabled chronic disease management in older adults. As the implementation of ACCESS continues, our findings provide important context for ongoing evaluations of whether the model can fulfill its potential to incentivize high-value, technology-enabled care for TM beneficiaries with chronic conditions.

Author contributions

T.T.L. had full access to all data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: T.T.L. and J.A.J. Acquisition, analysis, or interpretation of data: T.T.L. and J.A.J. Drafting of the manuscript: T.T.L., J.A.J., H.S., K.K., and M.B.C. Critical revision of the manuscript for important intellectual content: T.T.L., J.A.J., H.S., K.K., and M.B.C. Statistical analysis: T.T.L. and J.A.J. Obtained funding: T.T.L. Administrative, technical, or material support: M.B.C. Study supervision: M.B.C.

Supplementary material

Supplementary material is available at Health Affairs Scholar online.

Funding

There was no direct grant funding for this work.

Role of the funder: No funding organization was involved in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data availability

The data underlying this study are publicly available at https://meps.ahrq.gov/mepsweb/.