Misalignment Between Medicare Policies and Depression Care in Home Health Care: Home Health Provider Perspectives

The homebound condition requires that patients “have a condition due to an illness or injury that restricts their ability to leave their place of residence” except with some kind of aid or assistance (16). Medicare also requires that patients have a need for skilled care (by nursing or therapy professionals) on an intermittent basis and that is “reasonable and necessary.” When elaborating on the “reasonable and necessary” principles, the Medicare Benefit Policy Manual emphasizes the need to prevent future complications or acute episodes and to ensure medical safety.

Medicare’s prospective payment system (PPS), first implemented in 2000, specifies a 60-day episode rate with applicable adjustments. Case-mix adjustment is based on home health diagnoses, functional severity, and home health service utilization (17). Only home visits received for physical, speech-language, and occupational therapies are factored into the service utilization adjustment. The PPS provides strong incentives to home health agencies (HHAs) to contain costs within the per-episode payment and, in particular, to limit the number and duration of nursing visits in a given episode.

A related condition of Medicare participation requires that HHAs conduct the initial assessment in a timely manner (within 48 hours of referral, within 48 hours of the patient’s return home, or on the physician-ordered start-of-care [SOC] date) (17). This requirement, coupled with the fact that only new admissions (but not ongoing care) generate additional revenue to HHAs under the PPS, leads HHAs to assign higher priorities to SOC visits than to follow-up visits.

The Centers for Medicare and Medicaid Services (CMS) requires every HHA to conduct standardized assessments of patients using the Outcome and Assessment Information Set (OASIS). The OASIS is more comprehensive at the SOC and less so at discharge (to home), transfer (to an inpatient setting), or resumption of care (after a hospital admission). The current OASIS (OASIS-C) at SOC contains 79 items (18), making the first home visit a long and involved session. The SOC OASIS must be completed within five days of SOC and, after agency review for errors and updates, submitted to the state within 30 days (17). HHAs are regularly monitored for submission timeliness.

The OASIS-C encourages the systematic assessment of depressive symptoms at SOC with a two-item depression screen known as the Patient Health Questionnaire–2 (PHQ-2) (19–21). In addition, the discharge OASIS queries whether an intervention was conducted for patients who screened positive at SOC and whose physicians ordered interventions.

HHAs are subject to many other reporting and documentation regulations by CMS, partly as a result of mounting concerns regarding inefficient or inappropriate use of Medicare resources and rapidly growing expenditures (9). Software vendors have responded by supplying HHAs with electronic clinical and billing systems. HHAs are typically small, independent, and resource poor. They rely heavily on vendor products and rarely have home-grown systems. Clinicians routinely use laptops or tablets in the field that are synchronized or linked to a central system (22). Home health managers rely on the system to supervise frontline clinicians and manage cases.

Our multidisciplinary research team comprised complementary expertise in mental health policy (YB), evidence-based depression care intervention in home health care (MLB), qualitative research methods (JER), and health care organizational behavior (AAE).

We sought administrators and nurses from five HHAs located in five states in the Northeast, South, and Midwest regions of the United States. These agencies recently participated in a randomized trial led by one of the coauthors to test the Depression CAREPATH intervention (11,12). Based on our experience with the trial, we selected three types of key informants to provide complementary perspectives: home health nurses who worked on a CAREPATH intervention team and home health nurses who worked on a usual care team, nurse supervisors of CAREPATH intervention teams, and medical or clinical directors involved in their agency’s initial decision to participate in CAREPATH or who played a leadership role in facilitating participation. A liaison at each agency identified potential informants, who were then contacted by the study team by e-mail or phone to recruit them and schedule interviews. After the study was described to informants, written informed consent was obtained. The study received approval from the Institutional Review Board at Weill Cornell Medical College.

We designed a semistructured interview guide to assess perceptions of depression and perceived roles in and responsibilities for depression care, current practice of caring for depression either by themselves (nurses) or at their agency (administrators), and perspectives regarding facilitators of and barriers to depression care quality improvement at four levels: system (including policy), organization (HHA), provider (home health and non–home health), and patient. Open-ended questions allowed informants to elaborate on issues they found most salient. Three interview guides covered the same domains of inquiry but were worded differently to account for informants’ roles. The guides were pilot-tested with a home health clinical director, a nursing team supervisor, and a nurse outside the CAREPATH agencies and iteratively revised throughout the interviews. [The final interview guides are included in an online data supplement to this article.]

The first author conducted telephone interviews with informants between February 2012 and November 2012. Notes taken during and immediately after the interview were incorporated into data analysis (23).

Audio-recorded interviews were transcribed into Microsoft Word documents by a third-party service and analyzed with grounded theory methodology (24), an established qualitative analytic approach by which researchers generate overall themes from transcripts by iteratively reading and reviewing them, attributing labels (codes) to salient text, and documenting recurring topics that emerge.

Following accepted qualitative analytic methods (25–28), two investigators (YB and AAE) independently coded each transcript and met to discuss the definitions of codes and negotiate consensus. For further analytical rigor, a third investigator (JER) independently coded selected transcripts and participated in coding meetings to provide an outside voice and help negotiate consensus. The process was documented throughout and led to the development of a codebook and themes (the final codebook is available from the authors on request).

Data were organized and managed with the qualitative analysis software, Nvivo, version 9. The software facilitated text and code consolidation into larger themes intended to reflect how current Medicare policies are aligned (or misaligned) with depression care quality improvement in home health care.

Informants expressed concerns that Medicare eligibility requirements for patients to remain homebound and to demonstrate a “skilled need” were at odds with evidence-based depression care. Informants expressed frustration that they were not able to follow up with patients to track progress once patients were no longer homebound or their “skilled needs” were considered met and therefore ought to be discharged. This was a prominent theme voluntarily endorsed by more than half of informants in all three roles. One supervisor stated, “You [home health nurse] know that you’ve called the doctor, you’ve expressed to the doctor how you felt, but at this point—it’s not a skilled need to keep the patient open. . . . I’ve felt a lot of times that I’ve abandoned the patient, in a way. There’s no skilled need.”

Informants believed that the eligibility requirements were especially restrictive for antidepressant management because it takes four to eight weeks before a psychotropic medication can be determined effective. This concern intensified and became more difficult to reconcile among supervisors and nurses in the CAREPATH intervention, because they received training on antidepressant management as part of the CAREPATH protocol and practiced it in routine care.

Nevertheless, several administrators and nurses recognized the fiscal reality of Medicare and believed that Medicare could not afford to have patients stay longer in home health care because “the problem [need for depression and other chronic condition management] is so pervasive.” These informants reflected on the short-term, transitional nature of home health care and believed that their role was to get patients “on track” and to leave them to community-based providers for long-term management: “It would just feel better if there was good mental health support to turn [patients] over to . . . after we leave,” one informant said.

The lack of explicit recognition of nursing time and quality of care in the home health PPS provides misaligned incentives for chronic condition management in general and depression care in particular. Several agency directors described myriad challenges providing high-quality, evidence-based mental health services in ways that are financially sustainable. As one said, “The payment system is the same whether you focus on completing tasks . . . or if you spend this additional time [on depression care management]. . . . Nursing salary is very high, so yeah, we’re doing great care . . . but it’s destroying our budget.”

Two subthemes emerged regarding how the PPS incentives “trickled down” to affect how home health nurses manage their caseloads and compromise their ability to care for depressed patients. First, the productivity expectation (in the range of six or seven home visits per day) pressures nurses to limit time in patients’ homes. Informants acknowledged that they would like to do much more care coordination and patient and caregiver education, but they did not have time. Although one agency allowed nurses to schedule extended visits for patients with complex conditions, interviews with nurses revealed that extending visits was not always feasible. One noted, “You don’t always have the luxury of doing those extended visits.”

Second, incentives embedded in PPS coupled with the regulation regarding timely assessment at SOC led HHAs to assign higher priority to new admissions. A common solution to a scheduling conflict between a SOC and a follow-up visit was to have other nurses “pick up” the follow-up visit. However, several informants expressed concern that doing so might disrupt the clinician-patient relationship and continuity of care. One nurse noted that patients with depression or other mental health needs may be adversely affected by the inconsistency in care providers: “I try to make sure that I’m the person that sees [patients with mood issues] . . . because they don’t like a lot of change.”

Several administrators applauded the inclusion of the two-item depression screen in the current OASIS, believing that it raised awareness of depression among home health providers. However, informants cited two ways in which the screening tool as currently incorporated limits its potential to assist nurses with depression care management. First, the SOC visit may not be the best time to screen for depression; it should not be the sole screening attempt, and the results should not solely determine a patient’s need for depression care. Informants noted that the SOC OASIS assessment typically takes two hours and is too involved to allow adequate assessment of depression. In addition, patients are not always willing to discuss mental health issues at the first visit. Several informants mentioned that they usually readminister the screen in a week or two to confirm or update the initial assessment, which is consistent with evidence-based practice and the CAREPATH protocol. However, because the OASIS assessment becomes “locked” to the nurse after five days (in compliance with the timely completion requirement), nurses who readministered the screen did not usually revise the SOC assessment—even though nurse supervisors or other administrators are able to “unlock” it to correct errors or make updates. Thus the adequacy and accuracy of screening results recorded in SOC OASIS are of concern.

Second, the lack of follow-up assessment for depression in OASIS represents missed opportunities for continued evaluation and ongoing management. Currently, depression assessment is not included in the OASIS at resumption of care (after hospitalization), recertification (for another 60-day episode), transfer (to an inpatient setting), or discharge. An agency director put it this way, “I think it [OASIS] needs to go a whole lot further because, as you know, it just asks, ‘Was an assessment done?’ But it doesn’t ask, ‘What was the result? What was the follow-up?’ ”

Underdevelopment of clinical decision support for depression care management in vendor-supplied electronic health records (EHRs) for home health care and the high cost of customization constitute an important barrier to quality improvement. Several informants affirmed that having structured prompts or protocols built into the system would help with nurses’ adoption of evidence-based depression care. They pointed out that before the CAREPATH study customized their EHRs to support the intervention, functionalities for documenting depression care were weaker than for those documenting other major medical conditions, such as congestive heart failure and diabetes. In particular, informants noticed that once nurses documented depression, the system they worked with offered little guidance (such as through prompts or drop-down menus) on what to do next.