Late-Life Depression in Home Healthcare

Abstract

Major depression is disproportionately common among elderly adults receiving home healthcare and is characterized by greater medical illness, functional impairment, and pain. Depression is persistent in this population and is associated with numerous poor outcomes such as increased risk of hospitalization, injury-producing falls, and higher health care costs. Despite the need for mental health care in these patients, significant barriers unique to the home healthcare setting contribute to under-detection and under-treatment of depression. Intervention models target the home healthcare nurse as liaison between patients and physicians, and instruct in the identification and management of depression for their patients. Successful implementation requires interventions that ‘fit’ how home healthcare is organized and practiced, and long distance implementation strategies are required to increase the reach of these interventions.

Introduction

Depression is the third leading contributor to the overall burden of disease in the US, reflecting both its pervasiveness and consequences. Beyond personal suffering, depression in late life increases the risk of caregiver burden, medical illness, disability, social isolation, institutionalization, suicide and non-suicide mortality.[1] In this paper, we summarize published research on the epidemiology, course, treatment, and interventions associated with late-life depression in the home healthcare setting. We highlight examples from our own work as well as future directions for the field.

The 12-month prevalence of geriatric major depression is approximately 3% in community-dwelling samples[2, 3] and 6–8% in older primary care patients.[4] Rates of depression are even twice as high, if not higher, among geriatric medical inpatients, home healthcare patients, residents of assisted living facilities, home-delivered meal clients, and other populations of older adults who are characterized by the medical, functional, and social disadvantages that are associated with the risk and outcomes of depression. Rates of minor and subthreshold depression are also high in these geriatric populations and have been shown to be clinically significant in terms of significant in terms of persistence and negative effects on health and functioning.[5–9]

Our interest in home healthcare grew from epidemiologic evidence of disproportionately high rates of depression in homebound older adults.[10, 11] These findings were consistent with the population’s significant medical burden, disability and social isolation, conditions that are both risk factors and outcomes of depression.[1, 7, 12, 13] From a services delivery perspective, homebound status not only increases the need for mental health care but also poses significant barriers to meeting this need. We turned to home healthcare as the sector providing care to the homebound and found little research on depression.

Prevalence and correlates of depression in home healthcare

Home health (HH) agencies provide subacute care to over 3.5 million Medicare beneficiaries annually, mostly aged 65 and older.[14] Home healthcare is less costly[15] and generally preferred by patients over hospitals or nursing homes.[16, 17] Fee-for-service Medicare pays HH agencies by 60-day payment episodes using a case-mix adjusted standard rate. Few (2%) patients are referred to HH agencies with a primary psychiatric diagnosis; our research has thus focused on depression that co-occurs among patients referred for a medical or surgical (med/surg) condition. [15] Home healthcare patients receive an average of 22 nursing visits per episode of care. Many also receive home-based physical therapy or other services.

In the first study on rates and correlates of major depressive disorder in this population, we collected longitudinal data (N=539) in older adults (age ≥ 65) newly admitted to a Medicare certified HH agency in Westchester County, New York.[18] We interviewed patients at baseline (start of care), one, six, and twelve months after initiating home healthcare, collected additional data on depression from family members, reviewed medical records, and interviewed nurses.

To examine rates of depression, we determined DSM-IV diagnoses by best estimate conference where the Structured Clinical Interview for DSM-IV (SCID) for both patients and informants, medical record, and medication history were reviewed by the supervising psychologist, geriatric psychiatrist, geriatrician, and principle investigator. Using this best-estimate procedure, the one-month prevalence of major depression in this sample was 13.5% and that of minor depression was 10.8%.

Ell and colleagues estimated a comparable, albeit somewhat lower, 8.5% rate of probable major depression using the self report Patient Health Questionnaire (PHQ-9).[19, 20] Besides providing- an estimate of depression, this study demonstrated the feasibility of integrating the PHQ-9 into routine home healthcare practice and lay the groundwork for the addition of the PHQ-2 when Medicare revised its mandatory assessments in 2010.[21, 22]

Among patients who met DSM-IV criteria for major depression in our first home healthcare study, most (71%) were experiencing their first episode of major depression. The average depressive episode lasted over two months. Major depression was significantly associated with medical severity (OR=1.13 per point in the Charlson Comorbidity Index), disability in instrumental activities (OR=1.25 per IADL), reported pain (OR=1.82), and history of depression (OR=4.32).[18] Major depression was not, however, associated with sociodemographic factors such as gender, age, or marital status that are commonly associated with depression in community, primary care, and groups of older adults who do not have the level of medical severity, disability or pain that both characterize home healthcare patients and strongly contribute to depression risk.

Although the prevalence of major depression, specific symptom groups, and overall symptom severity did not differ between African-American and Caucasian patients in our research sample,[23] racial disparities are evident in the identification and treatment of depression in real world practice. In a recent national sample of US elderly patients receiving home healthcare, Caucasians were significantly more likely than African-Americans to have a formal depression diagnosis noted in their home healthcare chart (OR=4.06, 95% CI=1.66–9.93).[24] Future research is needed to determine whether this finding reflects differences in depression recognition, documentation, or transfer of clinical information across health care settings.

Medical severity and disability also contributed to the incidence of depression among home healthcare patients who were not depressed during the first weeks of home healthcare. Among patients without depression, 10% reported major or minor depression at one year follow-up.[25] Factors predicting depression incidence were similar to correlates observed at baseline, including worse self-rated health, somatic depressive symptoms, greater number of ADL limitations, and greater decline in ADL.

The prevalence of suicide ideation was also disproportionally high in home healthcare patients; in the Westchester patient sample, 10.6% of patients reported passive and 1.2% reported active suicidal ideation.[26] Higher depression severity, greater medical comorbidity, and lower subjective social support were independently associated with the presence of any level of suicidal ideation.[27] In particular, satisfaction with one’s relationships and feeling useful to family and friends were significantly associated with a lower likelihood of suicidal ideation. Despite the predominantly mild levels of ideation, we consider the home healthcare population to nevertheless be at relatively “high risk” for suicide based on their combination of several risk factors as identified in the literature, including depression, medical comorbidity, disability, and low perceived social support. Among patients without suicide ideation at baseline, 5.4% reported suicidal ideation at their one year follow-up interview.[26]

Course and outcomes of depression

Longitudinal evidence on the persistence of depression among our sample suggests that major depression has clinical significance in this patient population.[28] Among patients with baseline major depression, 42% continued to meet criteria at one month, 27% achieved partial remission, and 31% were in full remission. Baseline factors that predicted full remission at one month included fewer IADL limitations, experiencing a “great deal” of pain, and absence of a recent stressful life event. Similarly, suicidal ideation persisted for 36.7% of those with such ideation at baseline.

Evidence that depression increases home healthcare patients’ risk for other adverse outcomes underscores the clinical meaningfulness of depression in these patients and the overall costs involved in their care. Longitudinal analyses documented that depressed patients compared to other home healthcare patients, controlling for other risk factors, had higher risk of hospitalization, injury-producing falls, and higher health care costs.[29–32] For example, the mean time to hospitalization in our research sample was 8.4 for depressed patients versus 19.5 days for non-depressed patients after start of care.[32]. Hospitalization risk was significantly higher for depressed patients during the first few weeks. In an analysis of administrative data on all geriatric patients admitted to one HH agency over two years, patients with adverse falls were over twice as likely to be been identified as depressed at the start-of-care than matched nonfallers.[29]

Depression recognition

Our group investigated nurses’ recognition of depression in their home healthcare patients in two ways. First, we interviewed the nurses who provided care to patients in our study sample and compared their assessment to that of the SCID. The nurses correctly identified depression among 44 of the 97 SCID-depressed patients (sensitivity of 45.4) and 230 of the 306 patients who were not depressed (specificity of 75.2).[33] Nurses’ reports of moderate or severe depression agreed poorly with our SCID diagnosis of major depression (Kappa=.19) and with the presence of gateway symptoms of depressed mood or anhedonia (kappa=.25) identified in our best estimate consensus conference. Nurses who had more geriatric nursing experience, however, were more likely to correctly identify depression.

A second analyses compared nurses’ identification of depressed mood or anhedonia (i.e., the gateways symptoms for a DSM-IV depression diagnosis) on mandated start of care assessments (OASIS). OASIS ratings also agreed poorly with research-assessed gateway symptoms (kappa=.38) (kappa=.27) and with SCID major depression.[34] The home healthcare nurses accurately documented the presence of depression in 13 of 35 cases (sensitivity=37.1%), and agreed on the absence of depression in 175 of 185 cases (specificity=94.6%; negative predictive value=88.8%).

These data suggested that home healthcare nurses had difficulty making accurate assessments of depression among older home care patients. From interviews of nurses, we also learned that many nurses reported similar misconceptions as much of the general public regarding the symptoms (e.g., social isolation) and causes of depression (e.g., depression is an inevitable consequence of aging). These studies identified inaccurate assessment of depression as a significant barrier to treatment in this elderly homebound population.

Treatment

In our group’s representatives sample of home healthcare patients, only 21.9% of patients with major depression were receiving antidepressant treatment at start of home healthcare, and none was receiving psychotherapy.[18] Moreover, over 40% of these treated patients received inadequate treatment due to under-dosing compared to evidence-based guidelines or to patient non-adherence. While rates of identification and treatment of depression in home healthcare patients have risen sharply in the past decade,[24, 35, 36] most patients with clinically significant depressive symptoms are still either not being treated or are being treated inadequately. Reasons for inadequate treatment range from patients’ refusal or nonadherence to physicians’ not following evidence-based guidelines in prescribing or in adjusting treatment in response to persistent symptoms.[18, 37–39].

Examination of patient preferences for depression treatment may shed light on the reluctance of patients to initiate and adhere to recommended treatments. In analyses of a recent research sample of home healthcare patients, we found that 47% preferred an active treatment such as antidepressant medication or psychotherapy as their first choice.[40] Over half, however, preferred nonactive or complementary approaches such as religious or spiritual activities, exercise, or “doing nothing”. Current antidepressant use, previous psychotherapy experience, white and Hispanic versus African-American patients, greater IADL impairment, and less personal stigma about depression were independently associated with preference for an active treatment.

The data presented above on prevalence, course, outcomes, recognition and treatment of depression in home healthcare are summarized in Table 1.

TABLE 1.

KEY OBSERVATIONAL STUDIES OF DEPRESSION IN HOME HEALTHCARE AND FINDINGS

Study	Sample	Prevalence	Correlates (Higher rates)	Course and Outcomes	Recognition	Depression Treatment
Weill Cornell: Westchester HH Agency	Representative sample of new Medicare medical/surgical patients Age ≥65, N=539	SCID Diagnosis:[18] MDD: 13.5% Minor: 10.8% Suicide Ideation: [26] Active: 1.2% Passive: 10.6%	Medical Comorbidity [18] Disability[18] Less Social Support[27] No difference race/ethnicity, gender or age [18]	Course of MDD at one month: [28] MDD: 42% Partial Remission: 27% Full Remission: 31% Days to Hospitalization:[32] Depressed 8.4 days Not Depressed: 19.5 Use of Agency Services (by MDD):[30] Skilled nurse: no difference Medical social work: increased Therapist: no difference HH Aide: increased	Compared to SCID Nurse Reports: [33] Sensitivity: 45.4% Specificity: 75.2% OASIS: [34] Sensitivity: 37.45% Specificity: 94.6%	Among Patients with MDD[18]: antidepressants: 21.9% psychotherapy: 0.0%
U. Southern California: 3 HH Agency Trial	Newly enrolled patients aged ≥65 Medicare or Medi-Cal recipients (N=9,178)	PHQ9: [19] probable MDD: 8.5% mild depression: 1.6%	Females [19] Latinos Not married Living w/others Poorer health		Compared to PHQ 10–14 OASIS:[19] Sensitivity=55% Compared to PHQ 15+ OASIS:[19] Sensitivity=64%	Among patients with high PHQ9 scores:[19] antidepressants: 36%
National Homecare and Hospice Survey	Cross-sectional analysis of national survey data of HH Age ≥65, N= 2,137	Chart Diagnosis:[35, 36] ICD9 codes: 6.4%	Younger Age [35] White Race [24]			Antidepressant Use(by chart diagnosis):[24, 36] Depression: 73.2% No Depression: 31.6% Higher in whites in both depressed and nondepressed [24]
Weill Cornell: Westchester HH Agency Administrative Data Study	All patients (age ≥65) over 2 years with Serious Falls and Matched Controls (N=908)	OASIS (M0590)		Depressed increased risk of falling (OR=1.9)[29, 31]

Abbreviations:

HH: Home Healthcare

SCID: Structured Clinical Interview for Axis I DSM-IV Disorders (SCID)[93]

MDD: Major Depressive Disorder

PHQ9: Patient Health Questionnaire [20]

ICD9: International Classification of Disease, 9^th Edition

OASIS: Medicare’s Outcome Assessment Information Set

Partnership approach to intervention development

Studies documenting the problems of under-recognition and under-treatment have led to initiatives aimed at improving depression assessment, referral, and treatment. These projects confronted the kinds of challenges typically faced when trying to implement new practices in routine care as well as some unique to home healthcare.[37, 38, 40, 41] Working in partnership with HH agencies proved to be a successful approach to the development of depression management in home healthcare. Advantages to this approach are increased acceptability, feasibility, and effectiveness of the intervention efforts.[42–44]

Using this approach, our group worked with three regional HH agencies to identify factors that would facilitate better depression assessment. At that time, the Medicare OASIS-B did not include any standard assessment for depression. After working with the agencies, we decided against the inclusion of a standardized scale into the routine nursing assessment because it increased the already time consuming and unpopular burden of formal assessments. Further, there was little evidence that the addition of rating scales to routine home healthcare practice, or to clinical practice generally, would improve patient care or outcomes.[45–47] Rather, the intervention trained nurses in more accurate use of existing required assessments, specifically those in the OASIS that target the evaluation of depression. Training nurses to improve the existing depression assessment also alleviated the nurses’ concerns that assessments would be cumbersome and incongruent with other aspects of routine care.

TRIAD: Training in the Assessment of Depression

Weill Cornell’s intervention, TRIAD: Training in the Assessment of Depression, operationalized the gateway depression items (depressed mood and anhedonia) in the OASIS to be consistent with DSM-IV criteria with appropriate questions and, when indicated, clinical probes.[48] The goal of this intervention was for nurses to accurately identify patients with persistent depressed mood or anhedonia, symptoms warranting further evaluation by the primary care provider or through a mental health referral. In an effort to improve the sensitivity of detection to depression symptoms, the training called attention to factors that frequently cause difficulties the course of depression in HH patients, such as medical illness, disability, and pain, Nurses were also trained to ask follow-up depression assessment with clinically relevant questions about duration and persistence of symptoms. Nurses were also encouraged to observe and assess behavioral or nonverbal cues such as crying, eye contact, and affect.

Training procedures used proven techniques for effectiveness in continuing education. One such technique was the development of videos depicting nurse interviews with patients showing different scenarios and complicating conditions that may be encountering while using the OASIS for depression assessment. Other techniques incorporated into the intervention were tool kits, role play, and didactic instruction.[49] The training totaled 4.5 hours divided into two sessions, 1 month apart, as well as two e-mail boosters. By giving enough time for nurses to practice their skills with patients between sessions, the booster session could respond to the nurses’ “real world” experience rather than hypothetical situations. Trainings were set up in a way at the agency’s request to minimize the time that nurses spent “out of the field” in an effort not to disrupt clinical care.

A nurse randomized trial conducted in three HH agencies compared TRIAD to a minimal intervention consisting of a training video only and to a control condition.[37] The study compared nurses’ assessments with assessments conducted by research associates. Based upon review of patient records, the full TRIAD intervention improved the likelihood that a patient with persistent depressed mood or anhedonia would be successfully referred for a mental health evaluation. Specifically, 50.0% of depressed TRIAD patients received a referral, versus 18.5% of depressed patients in the minimal and 21.4% of the depressed control patients (p=.047). Furthermore, the intervention did not promote “false positives” in mental health referrals. In analyses of patients with a depression diagnosis, patients who were referred had clinical outcomes, specifically reduction in depressive symptoms.

Because the group of nurses who were provided a simple video did not have a better referral rate than the control group, the TRIAD results suggest that changing nurse practice involves more than providing a script or standardized assessment. An important component was helping nurses to understand the importance of depression to their patients’ care, giving them an opportunity to practice their assessments, and building upon their existing clinical skills. As these factors are relevant to the use of standardized assessments, the same strategy was used when developing the training curriculum and resources for appropriate use of the PHQ-2 (and PHQ-9) when it was added to revised OASIS-C.[50, 51]

Primary care and home-based intervention models

Improving depression assessment is only a first essential building block to quality depression care for elderly depressed home healthcare patients. A more comprehensive approach would integrate broader depression care management functions into home healthcare. An international evidence-base demonstrates that depression can be treated and managed in community-dwelling, homebound older adults.[52] Much of this research comes from outside the United States. These studies have demonstrated the effectiveness of home-based psychiatric services and/or integrating psychiatric care into routine home healthcare; however, the relevance of these specific models to the US is limited by significant differences in how home healthcare is organized and financed in different countries.[53–55] In the US, other studies have demonstrated the effectiveness of home-based depression care but were not conducted in the context of home healthcare. For example, the PATCH intervention conducted in senior housing combined case finding with home visits from university-based psychiatric nurses.[56] The PEARLS intervention demonstrated improved outcomes for older adults with minor depression or dysthymia who were referred by senior centers.[57]

One barrier to integrating depression care into home healthcare is that, unlike depression assessment which has been required of home healthcare nurses, depression care management has not been an expected part of nursing care for medical or surgical patients. Home healthcare nurses may be reluctant to add a new clinical responsibility to their workload. This challenge was similar to that originally faced by depression care management interventions designed for primary care. Two successful primary care models form the evidence bases used to develop depression care management interventions for home healthcare.

The Hartford Foundation-funded IMPACT and the NIMH-funded PROSPECT randomized controlled trials similarly investigated the effectiveness of primary care-based multi-faceted interventions for late-life depression.[58, 59] Both interventions are based on the Collaborative Depression Care model,[60] an evidenced-based approach shown effective in improving quality of care and clinical outcomes. Its cornerstone is managing depression as a chronic – rather than acute – illness. Thus patients benefit from not only active treatments (e.g., pharmacological and/or psychotherapy) but also ongoing care, e.g., monitoring symptoms and adherence, and teaching patients self-management skills. Primary care clinicians are supported both by a “depression care manager” (e.g., nurse or social worker) who provides much of the direct care and by access to mental health specialists for consultation as needed.

The Collaborative Depression Care model has been the subject of many meta-analyses, cost effectiveness, quality improvement and implementation initiatives.[47, 60–67] In IMPACT and PROSPECT, primary care physicians provided depressed older patients guideline-based treatment with depression care management. And in both trials, patients in the Intervention arms achieved better clinical outcomes than similar patients provided usual care alone. There is also evidence that depressed care management is associated with reduced mortality rates compared to usual care for depressed patients.[68]

In home healthcare, Ell and colleagues implemented a randomized intervention, modeled on IMPACT, that included nurse training in depression assessment, designated depression care specialists, and the option of home-based PST.[69] Their experience confirmed the feasibility and potential effectiveness of conducting depression interventions in home healthcare. However, as the authors acknowledged, barriers to implementing their model confirmed the need for further adaptation of the primary care model of depression care management to fit the needs of home healthcare patients and the organization and practice of home healthcare The following section describes our group’s adaptation of the IMPACT and PROSPECT to home healthcare. As with TRIAD, the Depression CAREPATH (CARE for PATients at Home) was developed using the evidence-based practice partnership model.[44]

Depression care management in home healthcare; the Depression CAREPATH

The Depression CAREPATH was designed for medical/surgical home healthcare patients who suffer clinically significant depressive symptoms.[70, 71] The underlying premise in developing the Depression CAREPATH was that successful implementation requires an intervention that ‘fits’ how home healthcare is organized and practiced. Our group worked with partnership with HH agencies to adapt the IMPACT and PROSPECT models of depression care management (DCM) from primary care to home healthcare. We found that the basic model fits naturally with HH which already uses a team approach to develop and follow a patient’s Medicare-mandated “Care Plan”.[72] The patient’s physician authorizes the Care Plan and is responsible for treatment decisions. The HH nurse supports the physician by providing in-home patient care and consulting with the physician and experts as clinically indicated both during care and at discharge.

In the Depression CAREPATH, the basic clinical functions of DCM remained the same as primary care, but whereas primary care assigns many of these functions to a single depression care manager, the Depression CAREPATH integrated these functions into the routine care practice of all medical/surgical nurses. This decision to train all nurses in DCM reflected two major considerations: 1. Cost: HH agencies are reimbursed based on a prospective payment system for fee-for-service Medicare patients. Additional home visits by a new provider would likely increase patient costs relative to reimbursement; 2. Skills: Nurses commonly manage chronic diseases regardless of the formal reason for home care (e.g., diabetes management with patients receiving wound care). Managing depression is fundamentally comparable to managing other chronic diseases, making nurses clinically prepared to practice DCM and fitting DCM easily into routine care.[73, 74]

The CAREPATH DCM protocol was designed to fit within a routine visit and used weekly or, for patients seen less frequently, at each visit. Each DCM visit involves (in home healthcare terminology):

1. Assessment of depression severity: Nurses are expected to assess patients who screen positive on the OASIS PHQ-2 with the full PHQ-9 to determine the severity of depression.[20] Although the PHQ-9 can be used as a self-report paper-and-pencil screen, nurses are expected to ask patients the PHQ-9 questions and encouraged to use visual and verbal clues when relevant when reviewing response with patients. Nurses continue to assess depression with the PHQ-9 weekly to monitor clinical course; lack of improvement or worsening symptoms trigger additional case coordination.

2. Case Coordination involves consultation with the patient’s physician about possible initiation or changes in medication and/or psychotherapy treatment, as well as possible consultation or referral to a psychiatric nurse (if available), social worker, or outside mental health specialist.[37, 75–77]. Case Coordination includes ensuring that depression care continues after discharge.[72]

3. Medication Management includes monitoring side effects and adherence for patients taking antidepressants.[78] Problems of treatment initiation and adherence are exacerbated in home healthcare as: 1. accessing psychotherapy is difficult for homebound older adults making it easy to skip appointments, and 2. home healthcare patients typically take many prescription medications; antidepressant medication are often viewed as the least important medication, contributing to the likelihood that patients will skip doses.

4. Education: As with other conditions, nurses educate patients about depression and its treatment, working to dispel myths and stigma.[79–81] One goal of patient education is to increase patient involvement in decision-making. Such involvement, while generally shown to be beneficial,[81–83] has particular relevance for depression as it may directly affect the helplessness and hopelessness associated with depression. Although research in this area is relatively new, especially with older adults or home healthcare patients, we do know that treatment[82] and in participating in decision-making[83] and, when randomly assigned to treatments that matched preferences had higher treatment initiation and adherence over 3 months.[84] The Depression CAREPATH intervention trains nurse to educate patients about depression and its treatments, to help patients clarify their own treatment preferences, and to use that information when coordinating care.

5. Goal Setting: While not psychotherapists, nurses typically assist patients in goal setting and activation.[85–87] For DCM, nurses help patients set and review weekly goals for self care, pleasurable activities, and social contact.[73, 88] Nurses are encouraged to involve family caregivers in developing strategies to support patients’ goals.

Family caregivers are a potential resource to all home healthcare interventions. A promising area for future research is skill training for family members to help them support DCM and determining the impact such training can have on patient health outcomes. A survey of the family caregivers of older home healthcare patients found that family members already provided a variety of relevant caregiver tasks but were also interested in receiving training in improve and extend those skills.[89] This interest in training was independent of the nature and extent of tasks they currently provided to their family member. Black caregivers expressed greater overall interest in receiving training than did white caregivers, as did younger caregivers compared to same-generation caregivers.

The depression intervention research, described above. is summarized din Table 2.

TABLE 2.

KEY INTERVENTION STUDIES OF DEPRESSION IN HOME HEALTHCARE AND FINDINGS

Study	Design	Sample	Depression Criteria	Intervention	Outcome
Weill Cornell TRIAD RCT [37]	Nurses randomized within each of 3 HH agencies in one county	Newly enrolled aged ≥65 Medicare medical/surgical patients (N=256)of nurse randomized sample	SCID Gateway symptoms (mood and anhedonia)	TRIAD (Training in assessment of Depression) of routine visiting nurses	Outcomes by end of first episode of care Completed referral of depressed patients: Intervention: 50% minimal intervention: 18% controls: 21%
U. Southern California Depression Care Management RCT [69]	Patients randomized within each of three HH programs in Los Angeles region	Depressed (PHQ9), newly enrolled aged ≥65 Medicare or Medi-Cal recipients (N=311)	PHQ9	Depression Intervention (Depression Care Management by routine visiting nurses plus Problem Solving Therapy)	Outcomes at 12 months Symptom Improvement: Enhanced Usual Care: 78% Intervention: 79% (OR=1.12) Response (50% reduction in PHQ9): Enhanced Usual Care: 36% Intervention: 44% (OR=1.39)
Weill Cornell CAREPATH Demonstration Trial [71]	Nurses randomized within each of four HH agencies across New York State	Depressed (PHQ9) Newly enrolled aged ≥65 Medicare medical/surgical patients (N=200)of nurse randomized sample	PHQ9	Depression CAREPATH (Care for Patients at Home): Depression Care Management by routine visiting nurses	Outcomes (Intervention only) by end of first episode of care: Full Remission: 50% Partial Remission: 23%

Abbreviations:

RCT: Randomized Control Trial

HH: Home healthcare

SCID: Structured Clinical Interview for Axis I DSM-IV Disorders (SCID)[93]

OR: Odds Ratio

PHQ9: Patient Health Questionnaire [20]

Implementation strategies

To be effective, a protocol needs to be used. The Depression CAREPATH was developed with this in mind to ensure that the intervention could be implemented and sustained in real world home healthcare practice. In addition to training home healthcare nurses in depression care management skills, the intervention also includes guidelines and resources to help HH agencies develop the infrastructure needed to support the use of the DCM protocol in routine care.[71] As part of infrastructure development, HH agencies are helped to tailor the guidelines to fit their own policies and local resources, including: 1. Case Coordination Guideline that designates whom to contact (e.g., patients’ physician, specialist) for referral or consultation; 2. Suicide Risk Protocol that operationalizes responses to levels of risk; 3. Mental Health Resources Directory for communities that serve their patients; 4. Supervision Strategies, including consultation on how to oversee the nurses’ use of the protocol; and 5. Benchmark Reports for use in supervision and quality improvement.

The dissemination and effective implementation of best practices commonly entail face-to-face involvement by the implementation staff.[90–92] Such intensity, however, may limit their availability, affordability or acceptability, especially to agencies that are small, geographically dispersed, decentralized, or resource poor. Most of the >10,000 Medicare certified HH agencies in the US are freestanding, small, and without resources to support advanced quality improvement. This situation leads to the unequal distribution of evidence-based care.

In response to these concerns, our group has developed a long distance implementation strategy designed to reach HH agencies regardless of size, affiliation, or location. It uses a web-based platform to support the implementation of Weill Cornell’s Depression CAREPATH Intervention by HH agencies. The implementation strategy employs e-learning modules, webinars, email/telephone consultation, toolkits, and social networking technology for long distance delivery of four implementation activities: 1. Infrastructure development, 2. Training in the Depression CAREPATH protocol, 3. Supervision and performance feedback on nurses’ use of the Depression CAREPATH protocol, and 4. Social learning among HH agencies that use the Depression CAREPATH. Future research will evaluate the effectiveness of using such long distance strategies to support the implementation of evidence-based practices.

Conclusion: Lessons learned and future perspective

Major depression affects one in seven elderly home healthcare patients, is mostly undetected or poorly treated, and leads to worse functional and health outcomes and higher healthcare costs. All older home healthcare patients, regardless of gender, race, ethnicity or age, share this high risk of depression. Over the past decade, an increasing number of older adults, with or without a depression diagnosis, start home healthcare already taking antidepressants. Nonetheless, most patients with clinically significant depressive symptoms are either still not being treated or are being treated inadequately. Thus interventions designed to improve depression detection, care and outcomes in home healthcare are still needed.

The challenge is to design interventions that are both effective in helping patients and sustainable in the organization. Our group’s strategy has been to work in partnership with key stakeholders to adapt the depression evidence based practice to fit naturally into routine home healthcare practice. At the same time, we help HH agencies modify their infrastructure so they can support these new practices. Effective and sustainable interventions have the potential for wide dissemination.

Home healthcare serve patients regardless of socioeconomic status and in locations ranging from the inner city to the most rural and isolated regions. Thus home healthcare is well positioned to reach all patients in need of evidence-based depression care. Effective implementation strategies with extended reach are needed to support the dissemination and implementation of evidence-based depression care by the large number of HH agencies who serve the nation’s homebound older adults.

Executive summary.

Depression in home healthcare

The prevalence of major depression among older adults newly admitted to home healthcare is 13.5%, and that of minor depression is 10.8%.

Depression is persistent in this population and further increases the risk of hospitalization, injury-producing falls, and higher health care costs.

Home healthcare nurses have difficulty making accurate assessments of depression, and rates of treatment are low.

Depression care management in home healthcare

An intervention aimed at increasing nurses’ ability to identify key symptoms of depression improved the likelihood that depressed patients would be successfully referred for a mental health evaluation.

Depression care management interventions are being developed and tested in home healthcare and focus on ongoing monitoring of depression symptoms, case coordination with other health care professionals, medication management, patient and family education, and patient goal setting.

Implementation strategies

Effective implementation of depression care management skills in home healthcare necessitates helping agencies develop infrastructure support. Long distance implementation strategies are being developed in order to reach HH agencies regardless of size, affiliation, or location.

Strategies include infrastructure development within HH agencies, web-based training modules, ongoing supervision and automated feedback on nurses’ use of care management interventions, and social learning among HH agencies.