Impact of Managed Long-Term Services and Supports on Nursing Home Use

Abstract

Background:

States have increased their use of managed care organizations to cover long-term services and supports (LTSS) to rebalance use from nursing homes to home and community-based settings. We study the impact of managed long-term services and supports (MLTSS) programs on long-stay Medicaid nursing home visits.

Methods:

We used the Minimum Data Set (MDS) 3.0 and Master Beneficiary Summary Files (MBSF) from 2011 to 2021 to identify 3,685,771 older adults aged 65 above with long-stay nursing home visits. Using a staggered difference-in-differences (DiD) regression design, we compared individuals in states that implemented MLTSS programs with individuals in states that did not implement MLTSS.

Results:

Although overall Medicaid nursing home enrollment for older adults decreased from 2011 to 2021, states implementing MLTSS experienced a 5.83% decrease in their share of older adults residing in long-stay nursing homes relative to states that did not implement MLTSS. We observed no significant changes in the share of residents with low care needs due to MLTSS.

Conclusion:

The findings of this study suggest MLTSS programs can potentially fulfill states’ objectives of rebalancing Medicaid LTSS use from nursing homes to home and community-based settings, but further design changes may be necessary to target low-care nursing home residents.

Introduction

Managing the delivery of long-term services and supports (LTSS) for older adults is an important policy concern. Nearly 25 million individuals over the age of 65 already necessitate some form of LTSS. ¹ In 2020, state Medicaid programs spent roughly one-third (or $200 billion) of their budgets on LTSS. ² Older adults generally prefer receiving LTSS care in their home or community, as opposed to an institutional setting.³ However, since the inception of the Medicaid program, states have been required to provide institutional services (i.e., nursing home care), while home and community-based services (HCBS) benefits are optional. Since then, Medicaid LTSS policy has focused on initiatives to rebalance LTSS spending away from institutional care and towards HCBS.^4,5

One such initiative is the use of managed long-term services and supports (MLTSS) programs, where state Medicaid programs pay managed care organizations (MCOs) a per beneficiary, per month capitated amount to deliver LTSS to beneficiaries.⁶ There has been a sizable increase in the use of MLTSS over the past decade and a half.⁴ By 2021, 24 states operated MLTSS programs covering roughly two million beneficiaries. Depending on the structure of the capitation rate, MLTSS plans may have financial incentives to provide HCBS to certain LTSS beneficiaries, or to move them to an institutional setting. In several states, MCOs receive a “blended rate,” which does not depend on whether a beneficiary lives in the community or an institutional setting, while in other states, the level of the capitation rate does.⁶ States can also “carve out” institutional services from the capitated payment system, such that the MCO forgoes responsibility of a plan member if they move into an institutional setting.^6,7 In addition to impacting the rebalancing of LTSS beneficiaries between home or community and institutional settings, the capitation structure may influence the type of LTSS-beneficiary in each setting. For example, for nursing home residents with low care needs, financial incentives may exist to transfer these individuals from the nursing home to the lower cost option of home or community-based care, because the amount of care required is low.

Recent studies analyzing MLTSS have found mixed evidence on whether MLTSS programs rebalance LTSS expenditures and reduce institutional care use.^5,8–11 In this study, we measure the impact of MLTSS programs on long-stay Medicaid nursing home residents, focusing on overall use and changes in the share of individuals with low care needs.

Methods

We estimated a difference-in-differences (DiD) regression specification to measure the impact of a state implementing a MLTSS program on the state’s share of older adults aged 65 and above residing in a nursing home and the share of low care needs residents. We used the Minimum Data Set (MDS) 3.0 and Master Beneficiary Summary Files (MBSF) from 2011 to 2021. Using MACPAC reports, we identified the year of MLTSS implementation across states.⁷ We defined low care needs as those individuals identified as not needing any physical assistance with the following activities of daily living: bed mobility, toileting, eating, and transferring, and not classified as either “special rehab” or “clinically complex”.^10,12 Our sample included only older adults aged 65 and above who were dually eligible. Because Medicare covers nursing home stays of up to 100 days, we subset our sample to individuals with length of stays longer than 100 days to identify nursing home stays covered by Medicaid. To account for “Medicaid spend down,” we also included individuals who transitioned into dual eligible status over the course of their nursing home stay. ⁷Because states implemented MLTSS programs in different years, we used the Callaway and Sant’Anna estimator.^,13 All regression estimates were adjusted for state-level demographic variables such as race, ethnicity, sex, and participation in the Balancing Incentives Program, which provided federal funding to improve LTSS rebalancing. In sensitivity analyses, we add restrictions on the possible values of post-treatment differences in trends given the “pre-trend”.

Results

Our sample included on average 976,883 long-stay Medicaid nursing home residents per year. Nursing home residents were similar in terms of demographic characteristics, however MLTSS states had larger Hispanic populations and a higher share of their nursing home population in urban areas (Table 1). We also found that overall, long-stay Medicaid nursing home enrollment for older adults decreased over time, and this decrease accelerated during the COVID-19 pandemic (Supplementary Figure S1). Although enrollment decreased by 5.41 percentage points from 2011 to 2019, it further decreased by 3.68 percentage points from 2019 to 2021 alone.

Table 1.

Descriptive Statistics by Managed Long-Term Services and Supports (MLTSS) Status at baseline

	MLTSS States		Non MLTSS States
	Mean	S.D.	Mean	S.D.
Age	83.38	1.01	82.83	1.20
Race/Ethnicity
White	76.14%	17.04%	84.49%	11.23%
Black	11.13%	9.00%	8.76%	10.72%
Asian	4.05%	14.17%	0.75%	1.10%
Other races	1.53%	2.95%	2.17%	5.16%
Hispanic	5.44%	7.91%	1.70%	2.50%
Female	72.55%	2.45%	72.44%	2.66%
Married	19.15%	2.30%	18.77%	2.17%
Urban	73.37%	17.33%	56.86%	21.13%
N (states)	24		26

From our regression estimates, we found that implementing MLTSS was associated with a total reduction of -0.15 percentage points in the share of older adults residing in a long-stay nursing home within a state (95% CI: −0.297 percentage points, −0.002 percentage points) or a 5.8% decrease (Table 2). The event study plot for changes in this share is displayed in Figure 1. In addition, we do not find any significant changes in the share of residents with low care needs due to MLTSS (Supplementary Figure S2). The results of our assessment of pre-existing trends of our outcomes are provided in Supplementary Table S1.

Table 2.

Difference-in-Differences Estimates for Nursing Home Enrollment and Share of Residents with Low Care Needs

	Mean	ATT, percentage points (95% C.I.)
Share of Older Adults Residing in Long Stay Nursing Homes, %	2.58%	−0.15** (−0.297, −0.002)
Low Care Needs, %	10.20%	−1.37 (−3.06, 0.33)
No Need of Bed Mobility Assistance, %	20.29%	−1.11 (−3.37, 1.16)
No Need of Toileting Need Assistance, %	12.76%	−1.55 (−3.37, 0.26)
No Need of Transfer Need Assistance, %	17.05%	−1.35 (−3.08, 0.37)
No Need of Eating Need Assistance, %	60.88%	−1.86 (−4.08, 0.37)
N (state-years)	468

^**p<0.05

Notes: Models are adjusted for race, ethnicity, sex, and state participation in the Balancing Incentives Program, all of which are aggregated at the state-year level.

“Always treated” states are not included. Robust standard errors are clustered at the state level.

Figure 1.

Event study plot for the share of older adults residing in long-stay nursing homes.

Notes: Regression estimates include the same control variables and state and year fixed effects, as listed in Table 2. Estimates obtained from an event study regression where the event time is equal to 0 in the first year of MLTSS implementation in a state. 95% confidence bars are based on standard errors clustered at the state level.

Discussion

This study found that implementing MLTSS programs was associated with changes in the share of older adults residing in long-stay nursing homes but no corresponding changes in the share of residents with low care needs. To our knowledge, this study is unique in examining the impact of MLTSS on nursing homes at a national level; nevertheless, the results are consistent with a three-state study which found no association between MLTSS and the share of low care needs residents,¹⁰ and a New York-state study which found a decrease in nursing home use for dual eligible individuals with dementia.⁹

MLTSS was associated with a slight decrease in the share of older adults residing in long-stay nursing homes, however the lack of reduction in the share of residents with low care needs warrants further discussion. Although this could partly be attributed to the facility-reported measurement of low care needs in the MDS, the relationship between MLTSS and low-care nursing home use could hinge on the structure of the capitation rates paid to the plans. If states offer a “blended rate” to MCOs, which does not depend on whether a beneficiary lives in the community or an institutional setting, then HCBS use is expected to increase, because institutional care is more expensive than HCBS. However, because capitation rates are recalculated annually, a higher share of HCBS users would result in a lower blended rate. Therefore, this scenario could potentially create a perverse incentive for MCOs to keep low care beneficiaries in nursing homes to maintain higher capitation rates.⁶ When states “carve out” institutional services from the capitated payment system, the MCO forgoes responsibility of a plan member if the individual moves into an institutional setting. In this situation, a financial incentive exists for the plan to move their high care needs beneficiaries to institutional settings, as opposed to providing them with HCBS.

In addition, existing programs, such as the Balancing Incentive Program (BIP) and Money Follows the Person (MFP), which provided federal funding to improve LTSS rebalancing, have operated in states with and without MLTSS and have been found to increase HCBS use and transition individuals from institutional settings to the community.^14,15 Thus, there may not be any additional impact of MLTSS in reducing the share of low care nursing home residents, beyond existing federal programs, although we do account for the presence of BIP in a state.¹⁰

The lack of change in outcomes other than nursing home resident share also brings into question the overall value of MLTSS programs, and whether it is able to offer LTSS beneficiaries with less expensive options of care. ^14,15Nevertheless, future work is needed to determine if MLTSS works better for certain subpopulations, whether the size of a state’s MLTSS program impacts the level of reduction of nursing home use, or whether there are existing specific capitation payment structures that are more successful in helping meet state objectives around MLTSS.

Limitations

This study had some limitations. First, we cannot identify the specific MLTSS program in which an individual is enrolled. Next, multiple plans may exist within the same counties within a state; however, we are unable to account for MLTSS plan heterogeneity within a given state. Finally, we are unable to identify the MLTSS plan’s enrollment policies (e.g., mandatory, passive, voluntary) and capitation payment structure.

Conclusion

The findings of this study suggest, when states implement a MLTSS program, they experience a slight decrease in Medicaid nursing home use. However, how a state structures its program’s capitation payment may impact whether MLTSS programs can fulfill states’ objectives of rebalancing LTSS use from nursing homes to home and community-based settings for low care needs residents.

Supplementary Material

Supplemental

The Online Supplement provides additional results and sensitivity analyses.

Figure S1. Long-Stay Nursing Home Enrollment among Dual Eligible Older Adults, from 2011 to 2021

Figure S2. Event study plot for share of Nursing Home Residents with Low Care Needs

Table S1. Sensitivity Analysis of violations to pre-trends assumption

Key Points Box

• State adoption of managed long-term services and supports (MLTSS) programs was associated with a slight decrease in a state’s share of older adults residing in long-stay Medicaid nursing homes.

• There was no corresponding change in the share of low care needs (i.e., not needing any physical assistance with bed mobility, toileting, eating, and transfers) residents

• The structure of the program’s capitation payment may play a role in keeping individuals with low care needs out of nursing homes

Why does this paper matter?

• While states have increased their use of managed care organizations to cover long-term services and supports (LTSS), how a state structures its program’s capitation payment may impact whether MLTSS programs can fulfill states’ objectives of rebalancing LTSS use from nursing homes to home and community-based settings for low care needs residents.

Data Sharing:

The dataset from this study is held securely at Harvard Medical School. The patient-level dataset cannot be made publicly available.