Cognitive impairment associated with greater care intensity during home health care

Abstract

BACKGROUND:

In Medicare-funded home health care (HHC), 1 in 3 patients has cognitive impairment (CI), but little is known about the care intensity they receive in this setting. Recent HHC reimbursement changes fail to adjust for patient CI, potentially creating a financial disincentive to caring for these individuals.

METHODS:

This cohort study included a nationally representative sample of 1,214 Medicare HHC patients between 2011-2016. Multivariable logistic and negative binomial regressions modelled the relationship between patient CI and care intensity—measured as the number and type of visits received during HHC and likelihood of receiving multiple successive HHC episodes.

RESULTS:

Patients with CI had 45% (p<0.05) greater odds of receiving multiple successive HHC episodes and received an additional 2.82 total (p<0.001), 1.39 nursing (p=0.003), 0.72 physical therapy (p=0.03), and 0.60 occupational therapy visits (p=0.01) during the index HHC episode.

DISCUSSION:

Recent HHC reimbursement changes do not reflect the more intensive care needs of patients with CI.

1. INTRODUCTION

Older adults with cognitive impairment—ranging from mild cognitive impairment¹ to Alzheimer’s disease and related dementias (ADRD)— are heavy users of health care with high comorbidity burdens and uniquely challenging care needs.^2-8 Industry leaders and policymakers have turned their attention toward development of targeted interventions to improve the quality and continuity of home- and community-based care for those with cognitive impairment. ^9-11 At the same time, payment policies in the United States do not consistently account for the greater care needs of older adults with cognitive impairment. In this context, there is growing interest in better understanding the service utilization patterns of these individuals and how payment policies may impact these patterns.^9-11

Medicare-funded home health care (HHC), a benefit offering skilled services delivered in the patient’s home environment, is an important source of support for community-living older adults with cognitive impairment. Medicare beneficiaries with cognitive impairment are more likely to access HHC than those without cognitive impairment,^{3, 12} and 1 in 3 Medicare home health patients has an ADRD diagnosis.¹³ However, to our knowledge, no prior work examines differences in care intensity during HHC based on patient cognitive impairment. During a Medicare home health episode, the patient may receive varying numbers of visits from home health staff, and visits may include a mix of skilled nursing, skilled therapy, personal care aide, and other service types. The type and intensity of visits received reflect the patients’ care needs, drives variable costs, and is a major determinant of profitability for home health agencies.¹⁴

Understanding care intensity among HHC recipients with cognitive impairment is particularly relevant in light of recent payment policy changes. Revisions to the Medicare home health payment system, under the Patient-Driven Groupings Model (PDGM), have the potential to negatively impact care for patients with cognitive impairment. PDGM is intended to more closely tie reimbursement to patient characteristics—particularly, diagnoses and comorbidities associated with higher care intensity. However, patient cognitive impairment is largely ignored, affecting reimbursement only if the patient has diagnosed dementia and a pressure ulcer or other specific skin condition.¹⁵ Thus, to the extent that cognitive impairment prompts higher intensity care, PDGM may negatively affect home health agencies caring for a higher volume of cognitively impaired patients, ultimately threatening access to care and/or impeding providers’ ability to deliver high-quality care to this vulnerable subpopulation. Information regarding care intensity by Medicare beneficiaries with cognitive impairment is necessary to predict whether PDGM is likely to lead to such unintended consequences.

The Centers for Medicare and Medicaid Services previously commissioned an analysis to inform home health case-mix methodology revisions which indicated that patient cognitive impairment was associated with reduced care intensity during HHC.¹⁶ However, this analysis did not adjust for any potential confounding patient or provider characteristics, nor employ econometric techniques to minimize endogeneity bias.¹⁶ Additionally, its findings diverge from existing literature which has found greater care intensity across numerous care settings outside of HHC,^{2-8, 17} as well as greater likelihood of HHC entry,^{3, 12} among those with cognitive impairment, suggesting a need for further examination of the relationship between cognitive impairment and care intensity during HHC. Drawing on a nationally representative sample of community-living Medicare beneficiaries who received HHC between 2011-2016, this study provides the first information regarding the relationship between patient cognitive impairment and care intensity during HHC. Findings offer a better understanding of the resources needed to care for this vulnerable subpopulation, potentially guiding future interventions to support aging in place in the context of cognitive decline, and may inform future revisions to the Medicare home health payment system.

2. METHODS

2.1. Data

To capture a rich array of quantitative data relating to the patient’s social context, health status, and experiences during home health care we link four data sources at the patient level: 1) the National Health and Aging Trends Study (NHATS), an annual, nationally representative survey of Medicare beneficiaries aged 65 and older with rich information regarding participants’ health status, functional impairment, and social supports.¹⁸ 2) The Outcome and Assessment Information Set (OASIS), a standard patient assessment completed during Medicare home health with information regarding patients’ health and functional status, including cognitive impairment. 3) Medicare claims, including the number and type of home health visits billed by the provider during an episode and the number of episodes in a spell of care. 4) The Medicare Provider of Services file (POS), a listing of Medicare-certified providers with information on home health agency characteristics. For each individual, we identified the first—“index” —HHC episode during the study timeframe (2011-2016) from OASIS and Medicare claims and linked these data to the immediately preceding NHATS survey for this individual and the POS data from the year of the episode for the home health agency that provided their care.

Our analytic sample included 1,214 NHATS respondents who entered the NHATS survey in 2011 and received HHC between 2011-2016, reflecting a nationally representative sample of Medicare beneficiaries (weighted n=5,856,333) in the United States. We excluded individuals who did not receive HHC during this period, as well as those living in congregate settings (e.g. assisted living) at the time of NHATS interview, due to the availability of supports that may affect HHC care intensity. We limited our analyses to the index HHC episode during the study timeframe; therefore, each individual appears in the dataset once.

2.2. Measures

Cognitive Impairment

We measure patients’ identified cognitive impairment during home health via OASIS item M1700, which requires the home health clinician to rate patient cognitive function on a 5-item scale. This approach to measuring cognitive impairment provides several benefits. First, billing/claims data routinely underestimate the presence of cognitive impairment in older adults.^{19, 20} Second, it is more precise to use a measure of cognitive impairment collected during HHC as deteriorations in cognition often precede home health utilization as a result of hospitalization²¹ or as a precipitating factor for home health referral.²² Third, as this item is already routinely collected during Medicare-funded HHC, it could easily be incorporated into potential revisions to the Medicare home health payment system.

Prior research on the psychometric properties of OASIS item M1700 finds sufficient “gold-standard” validity²³ in identifying cognitive impairment (compared to the Short Portable Mental Status Questionnaire) and strong inter-rater reliability.²⁴ We categorized those rated “0” on the 5-item scale—which corresponds to being “alert/oriented, able to focus and shift attention, comprehends and recalls task directions independently”—as having no identified cognitive impairment. A small proportion of our sample (13.3%) was distributed across the response categories indicating greatest cognitive impairment (“2”, “3”, and “4”). Therefore, we categorized those rated “1”, “2”, “3”, or “4” on the 5-item scale—ratings which range from “requires prompting” to “totally dependent”—as having identified cognitive impairment. (Please see Appendix A for greater detail regarding this OASIS item.)

Home Health Care Intensity

We measured care intensity by whether the patient received multiple successive home health episodes and by the number and type of visits provided during the index episode. We measured receipt of multiple successive episodes via Medicare home health claims data. Patients were categorized as having incurred a single episode of care if they received no additional home health episodes within 60 days of the end of the index episode and as having incurred multiple episodes otherwise.

From claims, we derived counts of nursing, physical therapy, occupational therapy, aide, and total home health visits incurred during the index spell of HHC. We identified the type of visit using Healthcare Common Procedure Coding System codes.²⁵ Nursing visits include visits from a Registered Nurse (RN) or Licensed Practical Nurse (LPN) to provide direct care, evaluate the plan of care, or observe/assess patient’s condition. Physical therapy visits include visits from a Physical Therapist (PT) to provide direct care or develop a program of therapy. Occupational therapy visits include visits from an Occupational Therapist (OT) to provide direct care or develop a program of therapy. Personal care aide visits include visits from a Home Health Aide. We examined visits received across three timeframes: 30, 60, and 120 days following the home health start of care date. These timeframes represent the new payment period for home health reimbursement (30 days), the payment period in place during our study timeframe of 2011-2016 (60 days), and the current average stay in HHC (120 days), as the average patient receives 2 successive 60-day episodes of care.²⁶

Older Adult and Home Health Agency Characteristics

From NHATS, we measured older adults’ sociodemographic characteristics (age, sex, race, Medicaid-enrollment). From OASIS, we measured older adults’ post-acute status (receiving inpatient care in the 14 days preceding HHC), clinical severity (overall clinical severity, elevated hospitalization risk, presence of surgical wound), functional impairment, and receipt of caregiver assistance during the HHC episode. Overall clinical severity and functional impairment were determined from Health Insurance Prospective Payment System codes²⁷ that identify home health patients as having low, moderate, or significant clinical severity and little or no, moderate, or significant functional impairment. Elevated hospitalization risk is determined by clinician responses to an item querying whether the patient displays “signs or symptoms” that suggest a heightened risk for hospitalization (e.g. polypharmacy, exhaustion, history of falls).²⁸

From the POS files, we measure whether the home health agency was nonprofit, the number of full-time equivalent employees, accreditation status, and whether the provider operated in a rural or urban area. Sources of provider accreditation included The Joint Commission, the Accreditation Commission for Health Care, and the Community Health Accreditation Partner.

2.3. Statistical Analysis

Propensity Score Weighting

We used propensity score weighting to account for a number of underlying patient characteristics that could confound the relationship between identified cognitive impairment and HHC intensity. Propensity score weighting yields treatment and comparison groups that are balanced in the distribution of observed characteristics, minimizing the threat of endogeneity bias.²⁹ We used logistic regression to model individuals’ probability of having identified cognitive impairment during the HHC episode (the “treatment”) as a function of a range of characteristics observed prior to the home health episode.

We used weighting by the odds to adjust for each individual’s propensity score. Treated individuals received a weight of 1 and untreated individuals received a weight of p/(1-p), where p represents their propensity score.³⁰ Following propensity score weighting, we observed satisfactory covariate balance between our treatment and comparison groups (standardized differences in means of observed covariates between treatment and comparison groups were all less than 15%).²⁹ We then created composite weights by multiplying each individual’s propensity score weight by their NHATS survey weight—an approach described and validated in prior research.^30-33 These composite weights adjust for both probability of treatment (propensity score) and complex survey design and were used in our final analyses. (See Appendix B for greater detail regarding propensity score weighting.)

Regression Models

We used multivariable, weighted logistic regression to model the odds of receiving multiple successive HHC episodes as a function of identified cognitive impairment. We used multivariable, weighted negative binomial regression to model the expected additional number of visits associated with identified cognitive impairment, by visit type. In descriptive analyses, we observed over-dispersion for all visit types, suggesting a need for negative binomial models; high frequency of participants who received zero physical therapy, occupational therapy, or aide visits suggested a need for zero-inflated models for these visit types.

All models adjusted for the following older adult characteristics measured prior to the home health episode: older adults’ age, sex, race, and Medicaid-enrollment. We adjusted for the following characteristics measured during the home health episode: older adults’ post-acute status, clinical severity, functional impairment, elevated hospitalization risk, presence of surgical wound, and receipt of family caregiver assistance. Finally, models adjusted for home health agency nonprofit status, number of full-time equivalent employees, accreditation status, and operating in a rural or urban area. All models were weighted using composite weights to adjust for both propensity score and complex survey design. All analyses were performed using Stata 14 (StataCorp LLC, College Station, TX).

3. RESULTS

Among 1,214 (weighted n=5,856,333) community-living Medicare beneficiaries receiving HHC between 2011-2016, home health clinicians identified 43.9% as having cognitive impairment (Table 1). Nearly 1 in 3 (29.2%) experienced multiple successive home health episodes. Patients received an average of 10.5 total visits within 30 days of start of care, 14.6 visits within 60 days, and 18.8 total visits within 120 days. (See Appendix C for averages by visit type). The average age was 81.6 years, over half (62.5%) were female, and 1 in 4 (22.2%) were Medicaid-enrolled. Over half (64%) were receiving post-acute care, 1 in 3 (32.9%) had high clinical severity, and 1 in 4 (24.3%) had high functional impairment.

Table 1. Characteristics of Community-Living Medicare Beneficiaries Receiving Home Health 2011-2016 by Identified Cognitive Impairment (Before and After Propensity Score Weighting).

(unweighted n=1,214; weighted n=5,856,333)^*

				Full sample (n=1,214)	Identified cognitive impairment (n=602, 43.9%)	No identified cognitive impairment (n=859, 56.1%)
						Before propensity score weighting		After propensity score weighting
n (wtd %) or mean ± SE							p- value†		p- value†
Receives multiple successive home health episodes				341 (29.2)	196 (31.8)	145 (20.5)	<0.001	145 (26.8)	0.17
Number of total visits received:
	Days since start of care	[	30	10.5 (0.29)	11.3 ± 0.41	9.7 ± 0.27	<0.001	9.7 ± 0.36	0.003
			60	14.6 (0.43)	15.8 ± 0.54	12.7 ± 0.41	<0.001	13.5 ± 0.58	0.002
			120	18.8 (0.65)	20.2 ± 0.83	15.7 ± 0.64	<0.001	17.4 ± 0.79	0.006
Older adult characteristics measured before home health episode
Age				81.6 ± 0.37	81.5 ± 0.45	78.4 ± 0.31	<0.001	81.7 ± 0.49	0.72
Female				748 (62.5)	383 (61.6)	365 (42.0)	0.30	365 (63.3)	0.64
	Race	[	White	815 (76.3)	384 (76.8)	431 (83.3)	0.02	431 (75.8)	0.76
			Non-white	399 (23.7)	218 (23.2)	181 (16.6)		181 (24.2)
Medicaid enrolled				249 (22.2)	148 (20.7)	101 (13.4)	<0.001	101 (23.7)	0.39
Older adult characteristics measured during home health episode
Post-acute				792 (64.0)	350 (62.5)	442 (77.8)	<0.001	442 (65.4)	0.48
	Clinical severity	[	Low	330 (25.7)	140 (21.7)	190 (30.3)	0.02	190 (29.6)	0.09
			Moderate	485 (41.4)	253 (43.0)	232 (39.8)		232 (39.9)
			High	399 (32.9)	209 (35.3)	190 (29.9)		190 (30.5)
	Functional impairment	[	Low/none	190 (15.1)	57 (11.6)	133 (21.7)	<0.001	133 (18.5)	0.04
			Moderate	762 (60.6)	384 (62.6)	378 (62.6)		378 (58.7)
			High	262 (24.3)	161 (25.9)	101 (15.7)		101 (22.8)
Elevated hospitalization risk				959 (74.2)	482 (73.7)	477 (68.0)	0.13	477 (74.8)	0.78
Wound				253 (18.4)	69 (13.5)	184 (35.5)	<0.001	184 (23.1)	0.002
Receives family caregiver assistance				1,123 (91.7)	558 (91.9)	565 (89.4)	0.28	565 (91.5)	0.87
Home health agency characteristics
Non-profit				455 (42.9)	226 (44.6)	229 (47.7)	0.40	229 (41.3)	0.42
Number of FTEs‡				86.4 ± 14.6	80.4 ± 11.1	113.1±28.8	0.13	92.5± 13.6	0.44
Accredited§				247 (23.3)	120 (22.3)	127 (23.0)	0.83	127 (24.2)	0.56
Rural provider				193 (18.0)	92 (16.9)	101 (18.5)	0.68	101 (19.0)	0.60

^*Data drawn from linked National Health and Aging Trends Study (NHATS), Outcomes and Assessment Information Set (OASIS), Medicare claims, and Medicare Provider of Services files for 1,214 Medicare beneficiaries receiving home health care between 2011-2016.

^†Results of weighted Pearson and Wald tests for differences between groups.

^‡Full time equivalent employees

^§Accreditation from The Joint Commission, Accreditation Commission for Health Care, or Community Health Accreditation Partner.

In propensity score adjusted multivariable models, home health patients with identified cognitive impairment had 45% greater odds of receiving multiple successive home health episodes (Adjusted Odds Ratio: 1.45; 95% CI: 1.01-2.09), after adjusting for relevant older adult and home health agency characteristics (Table 2).

Table 2. Odds of Receiving Multiple Successive Home Health Episodes, by Patient’s Identified Cognitive Impairment.

(unweighted n=1,214; weighted n=5,856,333)^*

				Adjusted Odds Ratio (95% CI)
Identified cognitive impairment				1.45 (1.01-2.09)
Characteristics measured before home health episode
Age				1.00 (0.98-1.02)
Female				1.14 (0.78-1.68)
	Race	[	White	REF
			Non-white	0.97 (0.58-1.62)
Medicaid enrolled				1.18 (0.67-2.09)
Characteristics measured during home health episode
Post-acute				0.39 (0.24-0.63)
	Clinical severity	[	Low	REF
			Moderate	1.41 (0.93-2.14)
			High	2.16 (1.35-3.45)
	Functional impairment		Low/none	REF
			Moderate	1.03 (0.60-1.75)
			High	1.19 (0.69-2.05)
Elevated hospitalization risk				1.35 (0.84-2.18)
Wound				1.04 (0.60-1.80)
Receives family caregiver assistance				0.79 (0.39-1.61)
Home health agency characteristics
Non-profit				0.30 (0.18-0.49)
Number of FTEs†				0.99 (0.99-1.00)
Accredited‡				1.13 (0.72-1.77)
Rural provider				1.88 (1.28-2.74)

^*Data drawn from linked National Health and Aging Trends Study (NHATS), Outcomes and Assessment Information Set (OASIS), Medicare claims, and Medicare Provider of Services files for 1,214 Medicare beneficiaries receiving home health care between 2011-2016.

^†Full time equivalent employees

^‡Accreditation from The Joint Commission, Accreditation Commission for Health Care, or Community Health Accreditation Partner.

In propensity score adjusted multivariable models, home health patients with identified cognitive impairment received a significantly greater number of visits and this difference in care intensity increased over the course of the home health episode. Within 30 days of home health start of care, patients with identified cognitive impairment received an additional 1.72 total visits (95% CI: 0.60-2.83; p=0.003) and 1.04 nursing visits (95% CI: 0.40-1.67; p=0.002) compared to those without identified cognitive impairment (Figure 1). Within 60 days of start of care, cognitively impaired patients received an additional 2.82 total visits (95% CI: 1.32-4.31; p<0.001), 1.39 nursing visits (95% CI: 0.49-2.29; p=0.003), 0.72 physical therapy visits (95% CI: 0.06-1.39; p=0.03), and 0.60 occupational therapy visits (95% CI: 0.15-1.05; p=0.01). Within 120 days of home health episode start of care, cognitively impaired patients received an additional 3.44 total visits (95% CI: 1.69-5.20; p<0.001), 1.68 nursing visits (95% CI: 0.60-2.77; p=0.003), 0.94 physical therapy visits (95% CI: 0.10-1.78; p=0.03), 0.67 occupational therapy visits (95% CI: 0.11-1.24; p=0.02), and 0.88 personal care visits (95% CI: 0.09-1.67; p=0.03).

Figure 1. Additional Number of Home Health Visits Received by Patients with Identified Cognitive Impairment, Holding Covariates at their Means*.

*Data drawn from linked National Health and Aging Trends Study (NHATS), Outcomes and Assessment Information Set (OASIS), Medicare claims, and Medicare Provider of Services files for 1,214 Medicare beneficiaries receiving home health care between 2011-2016.

4. STRENGTHS AND LIMITATIONS

This analysis has multiple notable strengths and limitations. Access to the linked NHATS and OASIS allows us to control for a rich array of covariates both before and during the home health episode and to develop propensity score weights to minimize potential endogeneity bias. While we adjust for multiple measures of health status during the home health episode, we do not examine specific comorbidities that may contribute to care intensity, such as diabetes or heart disease. The unweighted analytic sample size is small compared to the overall Medicare home health patient population. However, with available design variables and survey weights we are able to produce nationally representative estimates. We measure visits exclusively during the first 120 days following the start of HHC; thus, results may not reflect visit patterns across a longer spell of HHC. As claims data are not available for Medicare Advantage enrollees, our findings are limited to Medicare fee-for-service enrollees. We measure cognitive impairment using an item currently collected during routine home health patient assessments, maximizing the relevance of our findings to future payment policy revisions.

5. DISCUSSION

Identified cognitive impairment was present in nearly half (43.9%) of Medicare home health patients from 2011-2016 and was associated with greater care intensity, even after adjusting for underlying patient and home health agency characteristics. Patients with cognitive impairment had 45% greater odds of receiving multiple successive home health episodes and received an additional 2.8 total visits (95% CI: 1.32-4.31) during the index 60-day episode. Recent changes to the Medicare home health payment system do not account for patient cognitive impairment and reduce reimbursement for later episodes in a sustained spell of HHC. Our results raise the possibility that these policies may create a financial disincentive to serving beneficiaries with cognitive impairment—potentially reducing their access to HHC. Additionally, our findings confirm the recognized need for supportive home-based services for community-dwelling older adults with cognitive impairment.^{10, 34}

Compared to those without cognitive impairment, cognitively impaired Medicare beneficiaries access HHC at higher rates³⁵ and using different referral pathways.²² During home health they require greater family caregiver support³⁶ and a greater number of visits (particularly skilled nursing visits). These differences are especially notable given that cognitively impaired patients face a higher risk of readmission during HHC³⁷ and greater visit intensity is linked to reduced likelihood of readmission and other unplanned facility admission for this population.³⁸ Thus, there may be potential benefits to developing targeted HHC models for patients with cognitive impairment that consider services available within HHC and the unique needs of patients with cognitive impairment. While existing research has relied on a range of strategies for determining patient cognitive impairment,^{22, 35-38} findings from this work point to the potential feasibility of using an existing item in standardized home health patient assessments—OASIS M1700, requiring home health clinicians to rate patient cognitive function—to identify those who are likely to require more intensive care and may benefit from these targeted care models. Future research should further explore the psychometric properties of this measure and compare it to other modes of assessing cognitive impairment and/or ADRD during HHC.

Home health providers are highly responsive to payment system revisions.^39-42 This has previously led to reduced access for beneficiaries with greater disability, frailty, and social vulnerability^{39, 42, 43} as provider behavior altered to align with new financial incentives. Recent changes in Medicare home health reimbursement appear designed to shift the benefit towards a model of care that may be misaligned with the unique needs of cognitively impaired patients. Three recent changes to the Medicare home health payment system under the Patient-Driven Groupings Model (PDGM)⁴⁴ have the potential to reduce HHC access for beneficiaries with cognitive impairment.

First, episodes not immediately preceded by a hospitalization receive lower reimbursement under PDGM. These include episodes prompted by a community referral—which prior research finds are more common among older adults with cognitive impairment^{22, 35}—as well as later episodes during a sustained period of home health utilization. In this study, we find that home health patients with cognitive impairment have significantly higher odds of incurring multiple successive episodes. Reduced reimbursement for these episodes may create a financial incentive for providers to discharge patients before clinicians feel they are ready or may reduce access to HHC for those with cognitive impairment as providers incur heavier costs of care for these patients without commensurate reimbursement. Other chronic conditions may also increase the likelihood of incurring multiple episodes; however, cognitive impairment deserves specific attention as it is prevalent among home health patients¹³ and community-living older adults with cognitive impairment are at heightened risk for unmet care needs.⁴⁵

Second, PDGM shortens the payment period for a HHC episode from 60 days to 30.⁴⁴ Care provided after day 30 is considered to be part of a “later” episode and is subject to lower levels of reimbursement. Patients with dementia are more likely to receive visits in days 31-60 of a HHC episode³⁵ and in this analysis we observed that the difference in care intensity between those with vs without cognitive impairment becomes more pronounced over the course of a home health episode. Future analyses are warranted which consider more granular visit patterns; for example, charting the number of visits received, by day, across an episode of care and comparing utilization patterns by cognitive impairment.

Third, PDGM determines reimbursement by patient clinical grouping and comorbidity category, neither of which are affected by the presence of identified cognitive impairment or diagnosed dementia (with the exception of a comorbidity category which includes patients who have both a dementia diagnosis and a pressure ulcer or other specific skin condition).¹⁵ The present study finds that cognitive impairment is associated with receiving a greater number of visits across disciplines during home health. As visits are the major determinant of variable costs and profits for home health providers,¹⁴ failure to adjust reimbursement based on patient cognitive impairment may financially penalize providers who serve a larger proportion of patients with cognitive impairment.

This study offer a contrasting image to previous work suggesting that home health patients with cognitive impairment have lower resource use than patients without such impairment.¹⁶ Our results echo findings from settings other than HHC that indicate greater care intensity for those with cognitive impairment^2-8 and support recent work by Ankuda et al (2020) that raises the possibility of reduced access to HHC for those with cognitive impairment under PDGM.³⁵ Findings from the present study indicate that CMS should carefully monitor the potential unintended consequences of PDGM for beneficiaries with cognitive impairment and/or diagnosed dementia, particularly in regards to their access to, and outcomes during, HHC. Ensuring access to HHC among community-living older adults with cognitive impairment is especially important given that this population often faces unmet care needs⁴⁶ and there are significant barriers to accessing home-based care via Medicaid⁴⁷ or private pay.⁴⁸ Future inclusion of patient cognitive impairment as a determinant of reimbursement under PDGM, perhaps as a stand-alone comorbidity category, merits further attention and review.

6. CONCLUSION

During Medicare-funded home health care (HHC), patients with identified cognitive impairment were more likely to receive multiple successive episodes and received a greater number of skilled nursing, physical therapy, and occupational therapy visits. Findings raise concerns that recent changes to Medicare home health reimbursement may disincentivize providers from caring for those with cognitive impairment, potentially limiting their access to HHC. Future research analyzing shifts in HHC utilization patterns among those with cognitive impairment following implementation of these payment system revisions will be crucial to ensuring that community-based health care payment and delivery systems can adequately support this vulnerable population of older adults.