| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
Correspondence: Address correspondence to Dr. Beth Han, 1 Choke Cherry Road, Room 7-1010, Rockville, MD 20857. E-mail: beth.han{at}samhsa.hhs.gov
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Home care • Length of service in home care • Home health agency ownership • Medicare home health patients • Medicaid home health patients • Private health insurance patients in home care
Before the 1997 Balanced Budget Act (BBA), home health agencies received Medicare payment under a cost-based reimbursement system, thus providing more services among Medicare patients meant higher revenues to home health agencies (Levine et al., 2003; Liu, Long, & Dowling, 2003; McCall, Komisar, Petersons, & Moore, 2001; McCall, Petersons, Moore, & Korb, 2003; Murtaugh, McCall, Moore, & Meadow, 2003). Home health agencies might seek to take advantage of opportunities to bill third parties for unnecessary visits and to keep cases open longer than necessary. Researchers have described for-profit home health agencies as being motivated by economic profitability and profit maximization, whereas they have viewed not-for-profit home health agencies as having more altruistic motives than for-profit agencies (Carpenter, 1986; Estes & Wood, 1986; Schmid, 1993; Waldo, Levit, & Lazenby, 1986; Williams, 1994). Therefore, for-profit agencies may be more likely to engage in this revenue-management behavior than not-for-profit agencies. It is not surprising that Medicare patients received more visits from for-profit home health agencies than from not-for-profit agencies prior to the passage of the BBA (U.S. General Accounting Office, 1996, 1997; U.S. Office of the Inspector General, 1997; Williams, 1994). It appears to make sense that Murkofsky, Phillips, McCarthy, Davis, and Hamel (2003) found that Medicare patients receiving care from for-profit home health agencies had longer lengths of service than those receiving care from not-for-profit agencies before the passage of the BBA. Han, Remsburg, Lubitz, and Goulding (2004) reported that before the enactment of the BBA, Medicare patients from for-profit home health agencies were less likely to be discharged within 60 days than their counterparts from not-for-profit agencies. The U.S. General Accounting Office (1996) also reported bivariate results indicating that Medicare patients from for-profit agencies had longer lengths of service than their counterparts from not-for-profit agencies in selected regions. However, according to Williams (1994), among Medicare home care patients in 1984, length of service was not associated with agency ownership, although the total number of visits, total charges, and the intensity of service were higher among those served by for-profit home health agencies. Before the Medicare interim payment system (IPS), for-profit and not-for-profit home health agencies were equally required to recertify the eligibility of each patient receiving the Medicare home care benefit every 60 days (Brega et al., 2002; Han et al., 2004; Han & Remsburg, 2005; Murkofsky et al., 2003; Williams, 1994). Williams explained that based on the isomorphism theory (Estes & Swan, 1994), nonprofit and for-profit home health agencies should tend to behave similarly under the clearly structured payment system.
The BBA mandated the use of the IPS to limit costs while a prospective payment system for home health care was being developed. The IPS, which was in effect between October 1997 and October 2000, set an aggregate cap on how much a Medicare home health agency would be reimbursed and gave home health agencies a strong incentive to reduce the number of visits to Medicare beneficiaries (McCall, Korb, Petersons, & Moore, 2002). The prospective payment system replaced the IPS after October 2000. If for-profit home health agencies are motivated by economic profitability and profit maximization, and not-for-profit home health agencies have greater focus on altruistic behavior than for-profit agencies (Carpenter, 1986; Estes & Wood, 1986; Schmid, 1993; Waldo et al., 1986; Williams, 1994), then it seems reasonable that in 1999 and 2000, Medicare patients in eight states received shorter lengths of service from for-profit home health agencies than from not-for-profit agencies (Brega et al., 2002). Moreover, under IPS, the number of visits fell more sharply in for-profit home health agencies than not-for-profit home health agencies (FitzGerald et al., 2006; McCall et al., 2003). However, Murkofsky and colleagues (2003) found that Medicare patients receiving care from for-profit home health agencies had longer lengths of service in 1998 than their counterparts receiving care from not-for-profit agencies, leading them to suggest that for-profit agencies may provide poor-quality home care, resulting in longer lengths of service than would be the case for patients receiving care from not-for-profit agencies. Two empirical studies have examined differences in the quality of care delivered by for-profit and not-for-profit home health agencies (Ellenbecker, 1995; Jette, Smith, & McDermott, 1996). Ellenbecker found that not-for-profit home health agencies had superior performance on quality, whereas Jette and colleagues determined that home health agency ownership was not associated with the risk of having a care quality deficiency.
Hospital-based agencies are more likely to provide postacute home health services (Williams, 1994), thus patients from hospital-based agencies may have shorter lengths of service than patients from non-hospital-based agencies. Moreover, hospital-based agencies are more likely to be not for profit (Williams, 1994). Therefore, in order to truly understand the relationship between agency ownership and length of service, it is critical to adjust for the effect of hospital-based agencies on length of service. The study by Williams clearly controlled for whether agencies were hospital based or not and found no relationship between agency ownership and length of service. The investigations by Murkofsky and colleagues (2003) and Han and associates (2004) did not consider the possible impact of hospital-based agencies on length of service in home care. Because for-profit and not-for-profit home health agencies were equally required to recertify the eligibility of each patient receiving the Medicare home care benefit every 60 days (Han et al., 2004; Murkosfky et al., 2003; Williams, 1994) before and during the Medicare IPS, the isomorphism theory (Estes & Swan, 1994) would predict similar lengths of stay by ownership type during this period, as Williams found in the pre-IPS policy environment.
Furthermore, little is known about how nonprofit and for-profit home health agencies behaved with respect to length of service among Medicaid patients in the 1990s. None of the existing studies examined the association between agency ownership and length of service in home care among Medicaid patients. The Medicaid home health benefit covers more services and has less restrictive eligibility criteria than does the Medicare home health benefit. For example, although some Medicaid programs mimicked the eligibility for the Medicare home health benefit, Medicaid beneficiaries do not always have to be homebound. Many states do not require part-time or intermittent care and do not limit the amount of care delivered (Harrington, Summers, & Wellin, 2000), which may allow agencies to offer more extended home health services under Medicaid.
Home care benefits also vary among private health insurance plans (Center for Medicare Education, 2004). Periodic recertification for the eligibility of private insurance patients receiving home care is not often required. There are no consistent criteria to address the amount of care provided for home care patients with private health insurance. Little is known about the association between agency ownership and length of service in home care among patients with private health insurance.
It is important to know how nonprofit and for-profit home health agencies respond to changes in payment methods with respect to length of service and the number of visits over time among patients under different payment systems. The Centers for Medicare & Medicaid Services (CMS) made major cost-control Medicare policy changes through the IPS and added controls on fraud and abuse that occurred at about the same time. During the same period, CMS also sought to redefine the Medicare home health benefit as postacute care only and move it away from the more long-term-care model that it had become prior to the policy changes. Therefore, it is reasonable and critical to examine whether these new policies, which were heavily aimed toward reducing the length of service in Medicare home health care, differentially affected length of service in for-profit and not-for-profit home health agencies. In other words, the study can help researchers understand whether and how for-profit and not-for-profit agencies, with different financial motives, responded differentially to the Medicare policy changes. If there were differences, after accounting for patient mix and other factors, then the controls on length of service were probably insufficient to overcome the fact that for-profit agencies tend to change their services, patient mix, and length of service to maintain their profit margins. If there were no differences after controls, it is likely that one or more of the following held: (a) CMS managed to maintain sufficient control over length of service during the policy changes to limit for-profit/not-for-profit length of service differentials, (b) for-profit agencies managed to maintain their margins without changing length of service (e.g., by reducing the total number of visits or limiting the number of more costly registered nurse visits), or (c) not-for-profit agencies were concerned about costs and revenues in much the same way as for-profit agencies.
Using National Home and Hospice Care Survey (NHHCS) data, this study investigated the association between agency ownership and length of service in home care among patients with different payment sources—including Medicare, private health insurance, or Medicaid—during the period from 1992 through 2000. This study could not examine the relationship between agency ownership and the number of home care visits over time because only the 1992 NHHCS collected information on the number of home care visits per episode.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Each NHHCS was conducted from September to December of the survey year. The sampling frames for current patients included lists of patients enrolled in home care at the time of the surveys. The sampling frames for discharged patients included home care patients discharged during the 12 months prior to the day the survey was completed in each sampled agency. NHHCS data were obtained by personally interviewing an agency staff member who was familiar with the medical record of the sampled patient and who referred to patient medical and other records (Han & Remsburg, 2003; Haupt & Jones, 1999).
Differences in Patient Characteristics by Agency Ownership
Each current patient sample provides only a snapshot view of patient characteristics at the time of each survey data collection. In contrast, each discharged patient sample provides more stable estimates of patient characteristics during the 12 months prior to the day the survey was completed. Thus, we used discharged patient samples of all existing NHHCSs to examine patient characteristics such as age at admission (< 18, 18–49, 50–64, 65–74, 75–84, or 
85 years old), gender (male or female), race, reasons for home care discharge (hospitalized, placed in nursing home, transferred to other inpatient or residential care, recovered, stabilized, died, moved out of the area or changed agencies, family/friends resumed care, or services no longer needed), expected primary and secondary payment sources, number of activities of daily living (e.g., bathing, dressing, eating, transferring in or out of beds or chairs, walking, and using the toilet room) and number of instrumental activities of daily living (e.g., doing light housework, managing money, shopping for groceries or clothes, using the telephone, preparing meals, and taking medications) assisted by home health agency staff in the past 30 days/since admission, referral source (physician, hospital, self/family, or other), diagnoses at admission (cancer, congestive heart failure, fracture, chronic obstructive pulmonary disease, diabetes mellitus, and pneumonia) that may have related to chronic home care need, metropolitan statistical area and non-metropolitan statistical area as defined by the Office of Management and Budget, and region (Northeast, Midwest, West, or South). This study focused on for-profit and not-for-profit status of home health care agency ownership and did not examine characteristics of home care patients in government-owned agencies and in agencies with other miscellaneous ownership types due to their limited sample sizes.
The 1992–1996 NHHCSs allowed for only single-race reporting, whereas the 1998 and 2000 NHHCSs allowed for multiple-race reporting. To analyze trends by race, we coded race for the 1998 and 2000 NHHCSs as non-Hispanic Black, non-Hispanic White, or other. The non-Hispanic White category included only patients reported as non-Hispanic White with no other race. The non-Hispanic Black category included patients reported as Black alone or with other race(s). We included all other race groups and Hispanics (regardless of race) in the other category. To fully capture multiple payment sources of home care patients, we created a payment source variable based on the patient's expected primary and secondary payment sources for the episode of care (private health insurance, Medicare, Medicare plus private health insurance, Medicare plus Medicaid, Medicaid, or other miscellaneous payment sources including Veterans Administration coverage).
Estimating Length of Service in Home Care by Agency Ownership and Patient Payment Source
Using only current patients overestimates length of service, and using only discharged patients underestimates length of service (Cox, 1967). The SAS callable SUDAAN (Research Triangle Institute, 2004) software cannot adjust for this well-known statistical phenomenon, length-biased sampling. Thus, we examined both current patients and discharged patients simultaneously by applying the Kaplan-Meier method.
For each survey, we calculated the number of days a patient was enrolled in home care (for discharged patients, by subtracting admission date from discharge date; for current patients, by subtracting admission date from survey interview date). Moreover, we created variables for discharged patient status (1 = discharged patient, 2 = current patient) and survey year (1 = year 1992, 2 = year 1994, 3 = year 1996, 4 = year 1998, 5 = year 2000).
We created a variable indicating censoring status (censored observations or events) as follows: (a) all current patients who were in home care at the time of the surveys as censored observations; (b) discharged patients who moved out of the area or changed agencies as censored observations; and (c) all other discharged patients (hospitalized, placed in nursing home, transferred to other inpatient or residential care, recovered, stabilized, died, family/friends resumed care, or services no longer needed) as events. The Kaplan–Meier method estimated the median length of service among patients in for-profit and not-for-profit agencies in 1992, 1994, 1996, 1998, and 2000. Furthermore, we applied the Kaplan–Meier method to estimate the median length of service among patients with private health insurance, Medicare, or Medicaid in these years.
Associations Between Agency Ownership, Patient Payment Source, and Length of Service
We developed multivariate Cox proportional hazards models to determine the association between agency ownership and length of service in home care. Prior to model building, we completed collinearity diagnostics among independent variables to reduce multicollinearity. During model building, we used –2 log-likelihood statistics and Martingale residual methods to determine how well the models fit the data. We included an interaction term in the multivariate model if it was significant. Due to the complex sampling design, we used SAS callable SUDAAN (Research Triangle Institute, 2004) to conduct all analyses.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
Associations Between Agency Ownership, Patient Payment Source, and Length of Service
Table 2 presents results of the pooled model on the association between agency ownership and length of service in home care after controlling for covariates. The table reports adjusted hazard ratios with 95% confidence intervals. An adjusted hazard ratio less than 1.0 was associated with a lower risk of being discharged from home care and was related to a longer length of service. In the pooled model, we found an interaction effect between payment source and year (p <.0001) on length of service in home care. Our pooled model also identified main effects of other covariates (e.g., year, payment source, region, metropolitan statistical area, hospital-based agency). Patients with Medicare were more likely to have longer lengths of service than patients with private health insurance (p <.001) and to have shorter lengths of service than patients with Medicaid (p <.01). We found no statistically significant interaction between ownership and year (p =.3041), indicating that ownership was not associated with length of service in home care (for the pooled model) from 1992 through 2000. To reveal these complex relationships, we conducted further stratification analyses by payment source.
|
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
We found that during the IPS, the unadjusted median length of service in home care decreased among patients from both for-profit agencies and not-for-profit agencies. Moreover, unadjusted median length of service among patients in for-profit agencies decreased more than that among their counterparts in not-for-profit agencies, indicating that for-profit agencies may have reacted to the home health IPS more than not-for-profit agencies. However, after we adjusted for patient case-mix differences and agency characteristics, our results indicated that length of service was not associated with agency ownership from 1992 through 2000 among patients with any type of the examined payment sources. In addition to confirming the results by Willams (1994) and the isomorphism theory mentioned by Estes and Swan (1994) and Williams, our results indicate that regardless of the policy changes in the home care arena, for-profit and nonprofit home health agencies behaved similarly with regard to length of service in the 1990s among patients under a clearly structured payment system (e.g., Medicare) and among patients under inconsistently structured systems (e.g., Medicaid or private health insurance). Thus, our results do not support the notion that for-profit home health agencies act differently from not-for-profit home health agencies where length of service is concerned.
Before and during the Medicare IPS, for-profit and not-for-profit home health agencies were equally required to recertify the eligibility of each patient receiving the Medicare home care benefit every 60 days (Brega et al., 2002; Han et al., 2004; Murkofsky et al., 2003; Williams, 1994). Moreover, the eligibility criteria for the Medicare home health care benefit are clearly defined, such as under a physician's care, confined to the home, and in need of skilled nursing care; or in need of physical, speech, or occupational therapy services, medical social services, or home health aide services on a part-time or intermittent basis. Therefore, for-profit and not-for-profit agencies tended to operate similarly in providing length of care for Medicare patients, which is consistent with the isomorphism theory that nonprofit and for-profit home health agencies tend to behave similarly in length of service under clearly structured systems (Estes & Swan, 1994; Williams, 1994).
However, the eligibility criteria for home care patients with Medicaid or private health insurance are not always consistently defined. More importantly, there are no consistent and specific criteria to address the amount of care provided for home care patients with private health insurance or Medicaid, because different states (Kitchener, Ng, Miller, & Harrington, 2004) and private insurance plans have different rules and requirements. Periodic recertification is not always required for the eligibility of each home care patient with Medicaid or private health insurance. We did not find an association between agency ownership and length in service among patients with Medicaid or private insurance. Therefore, for-profit and not-for-profit home health agencies behaved quite similarly in length of service in the 1990s even for patients under inconsistently structured payment systems.
This study has several limitations that we should mention. We were unable to investigate the relationships among length of service in home care, quality of home care, and home health agency ownership because the NHHCS lacks information on quality of care by home health agencies. Moreover, we were unable to examine the association between the number of home care visits per episode and length of service in home care between 1992 and 2000 because only the 1992 NHHCS collected information on the number of home care visits per episode. Furthermore, we could not examine the availability of informal support in the home and the unmet needs of home care patients because of the limitations of the NHHCS questionnaires.
Regardless of these limitations, our findings clearly demonstrate that regardless of the policy changes in the home care arena in the 1990s, for-profit and not-for-profit home health agencies behaved similarly regarding length of service for patients receiving home health care under Medicare, Medicaid, and private insurance. It is not too surprising that Medicare length of service decreased similarly within the relatively well-controlled Medicare system. However, our study does provide new information regarding Medicaid and private insurance. The findings suggest that although state-to-state and company-to-company variation existed, any over-time change in length of service between for-profit and not-for-profit agencies was similar under controls for relevant patient and other characteristics. Importantly, with sufficient controls on relevant patient characteristics and other factors, and despite a substantial decline in length of service, our results show that patients in for-profit and nonprofit home health agencies did not experience differential changes in length of service. We do not have the information to fully address whether (a) CMS managed to maintain sufficient control over length of service during the policy changes to limit for-profit/not-for-profit length of service differentials, (b) for-profit agencies maintained their margins without changing length of service, or (c) not-for-profit agencies were concerned about costs and revenue in much similar ways as for-profit providers. However, the results of this study will be informative for future studies on these issues. Future studies are needed to examine the relationships among agency ownership status, length of service, and the number of home care visits per episode, as well as quality of care among home care patients with specific payment sources. Researchers should investigate how for-profit agencies are able to make profits with the reductions in length of service. For example, research is needed to examine whether the reductions in Medicare length of service simply shifted the costs to private insurance and Medicaid. It is also possible that for-profit agencies undertook fewer visits or other actions in order to maintain profits, or that the profits were large before the IPS and continued to be adequate. These pieces of information can be critical for policy makers to fully understand how for-profit and not-for-profit agencies behave before they eventually decide whether policy changes are needed to limit the continuous growth of for-profit agencies.
|
Footnotes |
|---|
1 U.S. Department of Health and Human Services, Washington, DC.
2 George Mason University, Fairfax, VA.
Decision Editor: Linda S. Noelker, PhD
Received for publication October 25, 2006. Accepted for publication April 11, 2007.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
