Evolution of the home health care market: The expansion and quality performance of multi‐agency chains

Abstract

Objective

To examine the growth and evolution of the home health agency (HHA) market and to compare quality performance across HHA ownership categories.

Data Source

Agency characteristics were extracted from Medicare cost reports and Provider of Services file. Quality of care and patient characteristics were extracted from Quality of Patient Care Star Ratings and HHA Public Use File.

Study Design

Agency‐ and state‐level analyses were conducted to describe HHA market trends. Patient characteristics and quality measures were compared across ownership categories of interest.

Data Collection/Extraction Methods

All Medicare‐certified HHAs in operation, 2005‐2018.

Principal Findings

Over the study period, the HHA sector grew substantially, increasing from 7899 to 10 818 agencies, and chain‐owned HHAs doubled in number from 903 (11.4% of all agencies) to 1841 (17.0%). In 2018, across agency types, for‐profit nonchain agencies were the largest category both in the number of agencies (67.8%) and the number of Medicare enrollees served (40.7%). Additionally, for‐profit nonchain agencies grew most in total number, from 4293 (54.3%) to 7337 (67.8%), while for‐profit chain agencies grew most in the number of Medicare enrollees served, from 439 998 (12.9%) to 1 082 385 (28.3%). Regarding patient composition, for‐profit nonchain agencies served the highest proportion of dual eligible beneficiaries (42.2%) and African‐Americans (27.9%) among all agency types. Regarding quality performance, a higher star rating is significantly (P < .01) associated with chain agency status. Moreover, chain HHAs performed better on self‐reported process measures, and risk‐adjusted self‐reported outcome measures; however, they performed worse on risk‐adjusted claims‐based outcome measures. These results were similar across for‐profit and nonprofit chain agencies.

Conclusion

Chains play a growing role in the home health sector. Substantial differences in geographic distribution, patient composition, and quality performance were observed between chain‐ and nonchain HHAs. Examining the growth and performance of multi‐agency chains can help inform quality reporting and monitoring, assess payment adequacy, and facilitate greater transparency and accountability within the HHA marketplace.

What This Study Adds.

• Prior studies have examined the home health market and performance of home health agencies by ownership as well as the relationship between ownership type and chain status.

• However, the home health market and home health agency performance have never been examined at the intersection of chain and profit status, which can provide important organizational and managerial insights on the HHA sector.

• For‐profit nonchain agencies served the highest proportion of dually eligible beneficiaries and African‐Americans in their patient composition in 2017.

• Yet, for‐profit chain agencies demonstrated the highest growth rate in terms of the number of Medicare enrollees served during the past two decades.

• Overall, chain HHAs performed better on summary star ratings, self‐reported process, and outcome measures, but they performed worse on claims‐based outcome measures.

1. INTRODUCTION

The home health care sector has expanded dramatically over the past two decades. From 2002 to 2017, the total number of home health episodes increased by 62 percent. ¹ In 2017, about 3.4 million Medicare beneficiaries received home health care services at a cost to Medicare of $17.7 billion. ¹ Today, Medicare‐covered home health care provides a wide range of services to homebound individuals such as skilled nursing, physical therapy, and occupational therapy to assist with recovery following an injury or illness. ²

As health care payers strive to encourage greater value, policy makers and researchers have targeted home health care as both a source of wasteful spending and a lower‐cost substitute for institutional postacute care. ³ Many alternative payment models have targeted home health utilization in the past decade, such as the Bundled Payment for Care Initiative and Medicare Shared Savings Program. ⁴ The goals of these models were to reduce costs or improve quality while encouraging closer cooperation between home health agencies (HHAs) and other providers in the health system. To date, evaluations have suggested home health is a source of savings under these models but there have not been big changes in terms of care integration across HHAs and health care systems.

Reflecting the substantial orientation to serving Medicare beneficiaries, Medicare policy and payment reforms have played large roles in shaping the HHA industry. For example, before 2000, the home health industry faced substantial policy changes that influenced market composition. Prior to 1980, Congress did not allow for‐profit ownership in the home health industry. ⁵ Between 1990 and 1997, home health spending and utilization rose so drastically that the Balanced Budget Act (BBA) of 1997 was passed to restrict the reimbursements given to home health agencies through the Interim Payment System (IPS). ⁶ This resulted in about 20 percent of Medicare‐certified home health agencies closing. ⁵ , ⁷ , ⁸ , ⁹ Three years later, the IPS was replaced with a BBA‐mandated prospective payment system (PPS), with home health agencies garnering significant profits in the subsequent years. ¹⁰ According to the Medicare Payment Advisory Commission (MedPAC), the majority of agencies opened since the implementation of the PPS have been for‐profit. ¹

Given the growing importance of home health in health care reform efforts and the large impact of policies and payments on evolution of the HHA industry, it is essential to understand the organizational structure of the industry and monitor how the sector is evolving in response to payment models. Ultimately, policy makers need to understand whether emerging ownership structures are associated with better HHA performance.

Studies of the skilled nursing facility and hospice sectors have shown that the market can change over time and this change can often be correlated with provider performance. Within the skilled nursing facility sector, there has been a great deal of churning in terms of chain ownership and skilled nursing facilities. Grabowski et al found that nursing homes that were bought or sold by chains had more deficiency citations, and therefore lower quality. ¹¹ Research on the hospice industry suggested an expansion in multi‐agency chains over time. Stevenson et al tracked the evolution of chains within the hospice sector and found that between 2000 and 2011, the number of for‐profit hospice chains rapidly increased. ¹² Interestingly, chain and nonchain hospice agencies had similar patient populations and service use, with greater differences observed between hospice agencies by profit status. ¹³ Eliason and colleagues studied the postacquisition behavior changes of dialysis facilities and found that acquired facilities converged to the behaviors of their new parent companies. ¹⁴

Research on the role of chains within the home health industry is more limited. A cross‐sectional 2014 study found that, on average, nonprofit agencies served a higher number of patients than for‐profit agencies, whereas for‐profit agencies had higher costs per patient. The same study also found that for‐profit agencies performed worse on overall quality indicators compared to nonprofits, especially in the South, where for‐profits dominated. ⁵ There was evidence that about 32 percent of HHAs are part of a chain ¹⁵ and that for‐profits are less likely to be associated with a chain. ⁸ However, we are not aware of prior research that tracks the emergence of for‐profit chain HHAs and evaluates their patient characteristics and quality performance.

To address this research gap, we analyzed the overall home health market, focusing on ownership and chain status, over the period 2005‐2018, during which substantial policy changes were happening. We also examined the change in the geographic distribution of chain‐owned agencies between 2005 and 2018. Additionally, we compared the characteristics of HHAs and their patients across all agency types, focusing on HHA chain status and ownership type. Finally, we analyzed the quality of care across all agency types. The term “agency type” is further used to describe HHAs by all combinations of HHA chain status and ownership type. Our study addresses the transparency, safety, and accountability of the multibillion‐dollar home health industry and how it intersects with other parts of the entire health system.

2. METHODS

2.1. Data sources

The primary analyses for this study relied on data from 2005 to 2018 Medicare Cost Reports, a public data source maintained by the Centers for Medicare and Medicaid Services (CMS). All Medicare‐certified providers, including skilled nursing facilities, hospices, hospitals, and HHAs, are required to submit a cost report annually. Freestanding home health agency information was extracted from Medicare Home Health Agency Cost Reports; information regarding hospital‐based HHAs and skilled nursing facility‐based HHAs was extracted from Medicare hospital and skilled nursing facilities cost reports.

To describe the HHA market, we obtained agency characteristics from Medicare cost reports, including geographic location, ownership type, chain status (identified by either chain name or chain code), and the number of enrollees served by Medicare‐certified HHAs from 2005 to 2018. Because the Medicare cost report data did not provide reliable and complete information about the chain or parent organization, these data were extracted through an iterative process. This iterative process involved thorough cross‐checking of the chain name and the chain code in the cost reports across years to clarify any inconsistencies that may have arisen due to mismatches, abbreviations, or typographical errors in the HHA chain names and chain status. Inconsistencies were resolved, if applicable, through Internet searches of relevant, publicly available information. Similar chain coding processes have been previously implemented in the hospice and skilled nursing home sectors. ¹⁶ , ¹⁷ , ¹⁸

About 10% of the HHAs lacked ownership information in their cost reports starting from 2010. Consequently, we completed ownership information by incorporating the ownership information from Provider of Services data, which were collected through the CMS Regional Offices and contains a quarterly record of provider characteristics for Medicare‐approved hospitals and other types of health care facilities. Additionally, some of the geographic locations for HHAs in the cost reports were listed as their parent organization locations, rather than as their exact locations. For these agencies, we identified the agency locations through the Provider of Services data.

To analyze HHA patient characteristics across agency types (for‐profit chain‐owned, for‐profit nonchain, nonprofit chain‐owned, nonprofit nonchain), we obtained data on patient characteristics at each HHA from the Home Health Agency Utilization and Payment Public Use File (PUF) for the year 2017, which is the most recent year available. The PUFs, maintained by CMS, provided information on utilization, costs, payments, as well as patient demographics and chronic conditions for each home health agency. The patient demographic and chronic condition data were evaluated for each home health agency, including average enrollee age, percent of male enrollees, the average number of Medicare‐Medicaid dual beneficiary enrollees, average hierarchical condition category (HCC) score, percent of enrollees with Alzheimer's disease, and the percent of enrollees with chronic kidney disease. More specifically, HCC score represents patients’ risk adjustment factor score, derived from a risk adjustment model originally designed to estimate future health care costs for the patients.

Additionally, we obtained data on the agency‐level and national‐level quality of care data for each HHA from Home Health Compare for the year 2018 and calculated the average and standard deviation of ten HHA quality measures for each agency type. Home Health Compare is publicly reported by CMS to facilitate consumers' choice regarding home health care providers. It provides HHA‐level information on process and outcome quality measures, a summary measure of the quality of patient care star ratings, patient survey results, and a summary measure of patient survey star ratings.

The quality measures we examined were derived by CMS from the Outcome and Assessment Information Set (OASIS), Medicare claims data, and Consumer Assessment of Healthcare Providers and Systems (CAHPS) survey data. OASIS is an assessment tool developed to collect and report performance data on HHAs. This tool captured 24 process and outcome quality metrics and was completed by managers representing individual home health agencies. Agencies receive star ratings on a 1 to 5 scale, with 5 representing the highest level of quality of care and 3.5 representing the national average level. The summary rating on the quality of patient care was generated by CMS through a combination of eight separate process and outcome measures. The patient survey results were derived from CAHPS, a set of surveys that asked patients to report on their health care experiences. The summary rating of patient experience was generated by CMS through a combination of four topic‐specific measures from CAHPS.

For our analyses, we focused on ten representative measures, including two summary measures, three process measures, five outcome measures. The two summary measures were star ratings for quality of patient care and HHCAHPS survey summary. The three process measures, which were self‐reported from OASIS, were the percent of home health team members that began their patients’ care in a timely manner, the percent of home health team members that addressed physician‐recommended medication issues in a timely manner, and the percent of home health team members that taught patients or their family caregivers about their prescription drugs. Process measures were not risk‐adjusted for variation in patient characteristics, as risk adjustment is not considered to be necessary for those measures. The five outcome measures were all risk‐adjusted for patient characteristics and were derived from two sources: (a) self‐reported outcomes from OASIS, and (b) claim‐based outcomes from Medicare claims. The three self‐reported outcome measures were the percent of patients who got better at walking or moving around, the percent of patients who got better at bathing, and how often patients receiving home health care needed urgent, unplanned care in the ER without being admitted. The two claims‐based outcome measures were how often home health patients had to be admitted to the hospital and Medicare spending per episode of care relative to the national median level of Medicare spending per episode.

2.2. Analyses

We conducted several analyses to investigate the home health market. We first examined overall trends in the number of HHAs and the number of Medicare enrollees served by HHAs by ownership type and chain status from 2005 to 2018. When provided partial year entries for a given agency‐year, we re‐weighted partial year observations to make annual estimations. We excluded agencies with missing entries.

We also compared HHAs by agency types for the years 2005 and 2018, focusing on agency characteristics and chain size, geographic reach, and agency affiliation (freestanding, hospital‐based, skilled nursing facility (SNF)‐based) including, but not limited to, the average number of Medicare enrollees per agency, the average number of Medicare enrollees per chain, the average number of states in which a chain operates, total number of Medicare enrollees served in each agency type, and the percent of Medicare HHA enrollees served. We calculated the percent of Medicare HHA enrollees served by dividing the total number of Medicare enrollees in HHAs of each agency type by the total number of Medicare enrollees served by HHAs for that year.

Additionally, we analyzed the percent of Medicare enrollees served by chain‐owned HHAs in each state for the years 2005 and 2018. We assigned the state of each HHA using its registered location in Medicare cost reports. We calculated the percent by dividing the total number of Medicare enrollees in chain‐owned HHAs in each state by the total number of Medicare enrollees served by HHAs in that state for that year.

We compared patient characteristics across different HHA agency types in 2017. For each patient characteristic measure, we calculated the mean value for a given HHA category by averaging the values of a patient characteristic for all HHAs within that category. The agencies with missing information for each patient characteristic were excluded from the calculation for that patient characteristic.

Finally, we compared the quality of care for each HHA agency type. For each quality measure, we calculated the average values for each HHA category by averaging across all HHAs in that category. Additionally, we examined the proportion of underperforming HHAs of which quality measures were in the national bottom decile for each agency type. We calculated the proportion of underperforming HHAs by dividing the number of HHAs in the national bottom decile by the total number of HHAs in this agency type. The agencies with missing information for each quality measure were excluded from the calculation for that measure. Moreover, multiple linear regression models were then performed to evaluate the association between quality of care star rating and different HHA types, by controlling for the number of Medicare enrollees in HHAs to account for the agency size and enrollee health insurance status.

3. RESULTS

3.1. Evolution of home health market

The home health sector expanded from 7899 agencies in 2005 to 12 079 in 2013 and then slowly declined to 10 818 in 2018. Exhibiting a similar trend, the total number of Medicare enrollees receiving home health services increased from approximately 3.4 million in 2005 to 4.1 million in 2016, then slowly decreased to 3.8 million in 2018 (see Figure 1).

FIGURE 1.

Home health market evolution over the period of 2005‐2018. Panel A. Number of home health agencies, by ownership type and chain status, 2005‐2018. Panel B. Number of home health agencies medicare enrollees, by ownership type and chain status, 2005‐2018. Source: Medicare Cost reports, 2005‐2018 [Color figure can be viewed at wileyonlinelibrary.com]

Over the 2005‐2018 period, chain‐owned HHAs doubled in number (from 903 to 1841 agencies), and the number of nonchain HHAs increased from 6162 to 8615. Among different types of HHAs, for‐profit nonchain agencies had the highest growth rate in the number of HHAs. Similarly, for‐profit chain‐owned agencies demonstrated a moderate increase in the number of HHAs. Conversely, the number of nonprofit nonchain and government‐owned HHAs declined.

Despite a high agency growth rate in the for‐profit nonchain agency category, for‐profit chain agencies demonstrated the highest growth rate in terms of the number of Medicare enrollees served. Medicare enrollees for this agency category increased from 12.9% (0.4 million) in 2005 to 28.3% (1.0 million) of the Medicare enrollee market in 2018. Medicare enrollees served by chain‐owned HHAs increased from approximately 0.5 million to 1.2 million, representing a change in Medicare market share from 15.6% to 32.5%. On the other hand, Medicare enrollees served by nonchain HHAs decreased from 2.7 million to 2.5 million (or from 78.7% to 65.5% of the Medicare market).

3.2. Characteristics of home health agency chains

Between 2005 and 2018, home health agency chains grew bigger, on average, with the largest for‐profit chains in the market playing an especially big role in more recent years. The average number of agencies per chain almost doubled for both for‐profit chains and nonprofit chains. (see Table 1) On average however, by 2018 for‐profit chains had a much larger number of agencies per chain (19.84 agencies per chain) compared to nonprofit chains (4.73 agencies per chain).

TABLE 1.

Characteristics of Home Health Agencies by ownership type and chain status, for the years 2005 and 2018 ^a

Agency characteristic	2005				2018
	For‐profit		Not‐profit		For‐profit		Not‐profit
	Chain	Nonchain	Chain	Nonchain	Chain	Nonchain	Chain	Nonchain
Size
Number of agencies (% of all agencies)	804	4293	99	1869	1647	7337	194	1278
	(10.2%)	(54.3%)	(1.3%)	(23.7%)	(15.2%)	(67.8%)	(1.8%)	(11.8%)
Total number of Medicare enrollees (% of all Medicare enrollees)	439 998	1 310 544	90 214	1 361 955	1 082 385	1 556 026	162 104	946 739
	(12.9%)	(38.6%)	(2.7%)	(40.1%)	(28.3%)	(40.7%)	(4.2%)	(24.8%)
Average number of Medicare enrollees per agency	608	360	930	785	752	295	921	851
Total number of all enrollees (% of all enrollees)	713 860	1 884 624	175 471	2 597 005	1 726 951	3 163 242	416 670	2 102 627
	(12.5%)	(33%)	(3.1%)	(45.5%)	(22.9%)	(41.9%)	(5.5%)	(27.9%)
Average number of all enrollees per agency	986	523	1809	1568	1205	607	2367	2030
Affiliation
Freestanding	795	4109	82	766	1644	7264	177	668
Hospital‐based	9	148	17	1022	3	52	16	551
Skilled nursing facility‐based	0	36	0	81	0	21	1	59

Chain Characteristic	For‐profit chain	Non‐profit chain	For‐profit chain	Non‐profit chain
Size
Average number of agencies per chain	10.86	2.83	19.84	4.73
Average number of Medicare enrollees per chain	6376	2653	13 877	4381
Median number of Medicare enrollees per chain	1229	1293	1998	3345
Average number of all enrollees per chain	10 498	5161	22 140	11 574
Median number of all enrollees per chain	2147	2079	3759	6289
Geographic reach
Average number of states in which chains operate	3.8	1.4	5.4	2
Number of chains in 3 or more states	24	4	39	7
Number of agencies in top 5 chains	461	48	847	82
Number of agencies in top 10 chains	586	69	1118	129
Number of Medicare enrollees in top 5 chains	261 691	47 029	674 869	64 394
Number of Medicare enrollees in top 10 chains	333 413	66 584	817 779	100 092
Number of all enrollees in top 5 chains	451 240	95 056	1 048 407	198 754
Number of all enrollees in top 10 chains	554 760	131 111	1 290 532	290 056

^aGovernment‐owned agencies are not included in this table. The percentages of all reported categories in the table did not add up to 100% due to the fact that government‐owned agencies were not reported in the table.

Source: Authors’ analysis of data from Medicare cost reports, 2005‐2018.


In terms of enrollment, Medicare enrollees and total enrollees per chain more than doubled between 2005 and 2018. At agency level, in both 2005 and 2018, for‐profit chain‐owned HHAs had considerably larger number of Medicare enrollees and total enrollees than for‐profit nonchain HHAs, on average, whereas the nonprofit chain‐owned and nonprofit nonchain HHAs were more similar.

The largest for‐profit chains have grown to play a prominent role in the home health sector. The total number of Medicare enrollees in the top 5 for‐profit chains increased from around 262 000 to 675 000, while Medicare enrollees in the top 5 nonprofit chains only grew from around 47 000 to 64 000. Similarly, for‐profit home health agency chains have grown to serve enrollees across a larger number of states, with for‐profit home health chains now serving enrollees in 5.4 states, on average, compared to only 2.0 states for nonprofit chains in 2018.

Finally, the majority of the HHAs were free‐standing agencies. From 2005 to 2018, the number of free‐standing HHAs in each agency type largely increased except for nonprofit nonchain category. Conversely, number of hospital‐based HHAs and skilled nursing facility‐based HHAs both decreased.

3.3. Geographic distribution of home health chains

The number of Medicare enrollees served by chain agencies shifted substantially on a state level from 2005 to 2018 (see Figure 2). In 2018, only one state (Montana) did not have enrollees served by any home health chains, compared to seven in 2005. The shift in enrollees served by chains was most prominent in the Southeast. The five states with the highest percentage of Medicare enrollees who received home health services from chain providers in 2018 were South Carolina (80.5%), Alabama (80.0%), Delaware (75.9%), Tennessee (70.4%), and West Virginia (68.5%).

FIGURE 2.

Percent of Medicare enrollees served by chain home health agencies, by state, for the years 2005 and 2018. Panel A. Percent of Medicare enrollees in chain home health agencies, by state, in 2005. Panel B. Percent of Medicare enrollees in chain home health agencies, by state, in 2018. Source: Authors’ analysis of data from Medicare cost reports, 2005‐2018

The change in the proportion of Medicare enrollees varies from state to state, but overall, the proportion increased in almost all states, with the only two exceptions being Michigan and Montana. The absolute value ranges from a decrease of 3.9% (Montana, 3.9% to 0) to an increase of 59.4% (West Virginia, 9.1% to 68.5%), with an average increase of 19.8%. The five states with the highest increase in the proportion of Medicare enrollees between 2005 and 2018 were as follows: Hawaii (49.7%), Oregon (52.8%), Georgia (53.1%), South Carolina (54.6%), and West Virginia (59.4%).

3.4. Patient characteristics of home health agencies

Several patient characteristics were distinct across agency types in 2017 (see Table 2). The biggest differences between HHA types were between for‐profit and nonprofit agencies, regardless of chain status. Among chain and nonchain agencies, for‐profit HHAs generally had enrollees with longer stays, a lower proportion of Medicare Advantage enrollment, a lower proportion of Medicare enrollees living in rural areas, and a much higher proportion of African‐American enrollees. Patient health conditions were similar across home health agencies by profit status. Comparing chain to nonchain agencies, there were some differences between chain for‐profit and nonchain for‐profit agencies, as the nonchain for‐profit agencies had more enrollees with longer enrollment periods, more dual eligible beneficiaries, fewer living in rural areas, and more African‐American enrollees.

TABLE 2.

Patients characteristics of Home Health Agencies by ownership type and chain status, for the year 2017

	For‐profit		Not‐profit		Government	National level
	Chain	Nonchain	Chain	Nonchain
Patients with service days
Percent of beneficiaries with 7 or fewer service days	3.4%	1.7%	5.0%	4.7%	3.4%	2.6%
Percent of beneficiaries with 30 or fewer service days	23.5%	22.2%	38.0%	40.4%	39.2%	26.7%
Percent of Beneficiaries with 60 or more Service Days	37.6%	49.2%	23.2%	22.2%	27.0%	42.4%
Percent of beneficiaries with 180 or more service days	12.1%	23.4%	5.5%	5.4%	8.8%	18.3%
Patients insurance status
Percent Medicare advantage beneficiaries	4.6%	6.9%	8.7%	7.5%	5.7%	6.5%
Percent Dual Beneficiaries	21.5%	42.2%	21.3%	20.9%	21.8%	34.5%
Percent Medicare beneficiaries in a rural ZIP	17.6%	8.2%	24.4%	29.0%	60.1%	14.7%
Patients demographics
Average age	77.3	76.0	77.4	77.8	77.9	76.6
Percent male beneficiaries	38.2%	38.6%	39.7%	39.2%	38.5%	38.6%
Percent White beneficiaries	80.7%	64.0%	85.7%	86.9%	90.1%	72.0%
Percent African‐American beneficiaries	14.9%	27.9%	6.9%	9.0%	7.2%	21.8%
Patients health conditions
Average HCC score	2.30	2.46	2.24	2.19	2.06	2.38
Percent of beneficiaries with Alzheimer's	38.4%	40.1%	32.0%	31.9%	30.7%	38.1%
Percent of beneficiaries with cancer	13.9%	12.8%	15.9%	16.3%	15.3%	13.8%
Percent of beneficiaries with chronic kidney disease	55.5%	56.5%	56.2%	53.9%	52.1%	55.8%
Percent of Beneficiaries with COPD	34.1%	33.6%	34.8%	33.2%	34.7%	33.7%
Percent of beneficiaries with depression	42.5%	41.8%	39.5%	38.6%	36.1%	41.2%
Percent of beneficiaries with diabetes	45.0%	51.9%	42.3%	42.2%	42.4%	48.8%
Percent of beneficiaries with stroke	12.9%	12.8%	12.0%	12.1%	11.6%	12.7%

HCC = Hierarchical Condition Category; COPD = Chronic obstructive pulmonary disease.

Source: Authors’ analysis of data from Medicare cost reports, 2017; Home Health Agency Public Use File, 2017

3.5. Quality of care of home health agencies

Quality of care was evaluated by analyzing star ratings and process and outcome measures for different HHA types. For‐profit and nonprofit chain‐owned HHAs had the highest average star ratings (3.84 and 3.48, respectively) compared to for‐profit and nonprofit nonchain HHAs (3.17 and 3.20, respectively), and government‐owned HHAs (2.99; see Table 3). The quality star ratings were significantly associated with chain status (P < .01), when adjusting for the number of Medicare enrollees in HHAs.

TABLE 3.

Quality performance of Home Health Agencies by ownership type and chain status, for the year 2018

	For‐profit		Not‐profit		Government	National level
	Chain	Nonchain	Chain	Nonchain
	Mean (% d )	Mean (%)	Mean (%)	Mean (%)	Mean (%)
Summary measures
Quality of patient care star rating (1 to 5)	3.84 (4.2%)	3.17 (19.9%)	3.48 (5.4%)	3.20 (9.3%)	2.99 (21.4%)	3.5
HHCAHPS Survey Summary Star Rating	3.64 (9.1%)	3.36 (19.2%)	3.60 (4.1%)	3.75 (5.5%)	4.26 (0.9%)	3.64
Process measures
How often the home health team began their patients' care in a timely manner (Percentage) a	95.71 (2.4%)	91.89 (12.6%)	93.65 (6.5%)	93.20 (7.5%)	92.68 (9.0%)	94.3
How often the home health team taught patients (or their family caregivers) about their drugs (Percentage) a	98.29 (4.1%)	96.16 (12.2%)	98.56 (2.7%)	97.18 (7.9%)	96.02 (13.8%)	98.2
How often physician‐recommended actions to address medication issues were completely timely (Percentage) a	91.11 (8.5%)	88.25 (11.6%)	95.02 (2.2%)	91.53 (5.8%)	89.43 (8.2%)	92.3
Outcome measures
How often patients got better at walking or moving around (percentage) a	78.44 (1.7%)	70.94 (14.0%)	75.35 (1.1%)	72.51 (3.8%)	69.03 (10.0%)	75.6
How often patients got better at bathing (percentage) a	81.35 (1.5%)	72.09 (14.4%)	78.15 (0.0%)	73.84 (3.4%)	71.21 (4.8%)	77.9
How often home health patients had to be admitted to the hospital b , c	16.3 (2.9%)	15.4 (13.9%)	15.9 (4.9%)	16.6 (4.9%)	16.6 (9.0%)	15.8
Medicare spending per episode of care, relative to the national median level of Medicare spending per episode b	1.03 (1.5%)	0.96 (15.1%)	0.97 (2.7%)	0.96 (6.2%)	0.96 (9.0%)	1.00
Potentially avoidable events measure
How often patients receiving home health care needed urgent, unplanned care in the ER without being admitted (percentage) a	13.87 (2.1%)	12.25 (14.6%)	13.69 (3.9%)	13.57 (4.2%)	15.08 (3.0%)	13

^aThese outcome measures were self‐reported by HHAs based on Outcome and Assessment Information Set (OASIS) data.

^bThese outcome measures were based on Medicare claims data.

^cThis measure was calculated based on data of 2017, as it was not available in 2018.

^dThe proportion of HHAs with performance in the national bottom decile for each agency type.

Source: Authors’ analysis of data from Medicare cost reports, 2017‐2018; Home Health Compare, 2017‐2018.


For the three self‐reported process measures, chain‐owned agencies (for‐profit and nonprofit) on average performed better compared to all other agency types. Two self‐reported outcome measures (the percent of patients who got better at walking or moving around, the percent of patients who got better at bathing) told a similar story: chain agencies out‐performed nonchain agencies. Finally, chain agencies (both for‐profit and nonprofit) had a greater share of patients who needed urgent, unplanned care in the emergency room, compared to nonchain agencies (both for‐profit and nonprofit). Government‐owned HHAs performed the worst on this measure.

Conversely, nonchain HHAs, on average, performed better on claims‐based outcome measures compared to chain‐owned HHAs. Chain‐owned HHAs (both for‐profit and nonprofit) had a higher percentage of patients who had to be admitted to the hospital, compared to for‐profit nonchain HHAs after risk adjustment. However, nonprofit nonchain agencies and government‐owned agencies performed the worst on this measure. For‐profit chain‐owned HHAs had the highest risk‐adjusted Medicare spending per eposide of care, 3% above the national median level. In contrast, the spending of all other HHA categories was all 3%‐4% below the national median level.

Additionally, the proportion of HHAs with performance in the national bottom decile was also examined for each agency type. Nonchain HHAs, on average, had higher proportion of underperforming agencies compared to chain‐owned HHAs.

4. DISCUSSION

The HHA sector grew substantially over the last two decades, with chain‐owned HHAs largely driving this growth. Among agency types, for‐profit nonchain agencies were most distinct, serving a higher proportion of enrollees with longer lengths of stay, who were dually eligible, and who were African‐American. Overall, chain HHAs performed better on many self‐reported process and outcome measures of quality but not on claims‐based outcome measures. We consider the implications of these results for regulatory oversight, quality measurement and reporting, and payment reform.

4.1. Regulatory oversight and fraud monitoring

As with other postacute care services, legislators and regulatory agencies have long been concerned about over‐utilization of home health services, inadequate care, and questionable billing. In 2011, CMS designated new HHAs as high‐risk providers, citing their record of fraud, waste, and abuse. ¹⁹ CMS recently issued guidance that new home health providers (defined as providers that newly enroll in the program or those that have predefined changes of ownership) are subject to being placed in a provisional period of enhanced oversight to address these issues. ²⁰ Based on our analyses, chain‐owned HHAs largely drove the growth of the HHA sector during the past two decades. Moreover, we observed that chain‐owned HHAs increasingly dominated the HHA market in various states, especially in the Southeast. Given the increasing growth of chain‐owned HHAs, tracking ownership of HHAs and auditing HHAs on a chain‐level could be a valuable strategy to ensure fiscal compliance. As agency‐level fraud often goes unnoticed, chain‐level oversight could more easily detect fraud and potentially increase compliance with the law downstream. Additionally, systematic evaluation by chain status can also inform functions such as investigation of complaints, compliance monitoring, and licensure renewals. ¹²

4.2. Quality measurement and reporting

Our study found that for‐profit nonchain HHAs performed relatively worse on self‐reported process and outcome measures but better on claims‐based outcome measures. One explanation for this difference is that multi‐agency chains can develop standardized care practices and more easily share resources. Standardization of protocols and practices makes it easier for specific tasks to be completed. For example, a study by Kamimura et al found that corporate standardization was associated with fewer deficiencies in chain‐owned nursing homes. ²¹

In contrast, nonchain HHAs demonstrated better performance on the risk‐adjusted Medicare claims‐based outcomes. A possible explanation for this difference is that risk adjustment has a substantial impact on HHA outcomes, in particular as the nonchain for‐profit HHAs provide services to a population that is more likely to be dual eligible, African‐American, and have a long length of stay. Another explanation is that these measures are less susceptible to upcoding on the part of providers.

Additionally, the proportions of HHAs with performance in the national bottom decile were higher for nonchain agencies than those for chain‐owned agencies. One explanation for this is that chains could better maintain floors of quality, compared to nonchain HHAs.

Monitoring and reporting quality of care is critical to ensure transparency and accountability within the home health market. A series of quality measures have been reported on Home Health Compare since 2003 to facilitate informed decision making for providers to refer patients and for enrollees to choose an HHA for themselves. In the home health industry, it is common for individual agencies to have a different name from their parent chain; therefore, chain‐level quality information could provide enrollees with more clarity about the culture and organizational structure of an HHA.

Our study provides evidence that notable differences existed in quality performance and patient outcomes between chain and nonchain agencies, in the aggregate. In future research, it will be crucial to investigate the root causes leading to this difference. The causes could be underlying quality differences or systematic reporting practice differences between chain and nonchain agencies. Moreover, as the behaviors of newly acquired health care providers could be largely impacted by the new parent companies, quality reporting and monitoring at chain‐level could provide information on quality performance at an organizational level, beyond just an agency basis. ¹⁴ , ¹⁶ Additionally, chain‐level quality performance could be considered when approving business activities such as mergers and acquisitions.

In addition to public reporting and transparency efforts, developing a chain‐level measurement focus seems especially relevant as integrated provider networks seek to build contractual relationships across geographic markets. Health systems, accountable care organizations, and managed care plans look to home health as a central component of their efforts, especially since postacute care historically has been a source of such variation in overall spending. Building provider relationships with home health companies, as opposed to individual agencies, offers organizations a greater ability to achieve efficiencies in care delivery through standardized procedures and practices.

4.3. Payment reform

This study confirmed the finding by MedPAC that the majority of agencies opened since the implementation of the PPS have been for‐profit. ¹ Additionally, in 2019, MedPAC recommended a 5% reduction in HHA payment in reaction to longstanding double‐digit profit margins in the home health industry. In this study, we found a growing number of Medicare beneficiaries have enrolled in chain HHAs over time. Therefore, it is necessary to examine further the patient selection and referral practice by chain status, besides just profit status, to ensure payment reforms do not have heterogenous impacts by ownership type.

4.4. Limitations

Our findings were limited in several ways. First, even with our detailed chain coding process, the data we obtained from the Medicare Cost Reports might still have missing information on certain chains, out‐of‐date chain information due to acquisitions or mergers, or incorrect coding of an agency as part of a chain organization. Second, quality measures analyzed in the study might not represent the complete picture of quality performance for different agency types, as we missed 3%‐35% of agencies for various quality measures (Appendix Table S1). The majority of the missing values in quality measures were due to the small sample size of the residents living in the HHAs, and thus, the quality measures were suppressed in those cases. Finally, the OASIS‐based quality outcomes are self‐reported and may be susceptible to differential reporting by agency type.

5. CONCLUSION

Chains play a growing role in the home health sector, as they have in other postacute care sectors and hospice care. Substantial differences in geographic distribution, patient composition, and quality performance are observed between chain and nonchain HHAs. Further research should seek to understand better the mechanisms that might contribute to these differences. By examining the growth and performance of multi‐agency HHA chains, our analyses can help inform efforts to monitor and assure the quality of care, assess payment adequacy and options for reform, and facilitate greater transparency and accountability within the HHA marketplace.

Supporting information

Author Matrix

Appendix S1

Geng F, Mansouri S, Stevenson DG, Grabowski DC. Evolution of the home health care market: The expansion and quality performance of multi‐agency chains. Health Serv Res. 2020;55:1073–1084. 10.1111/1475-6773.13597