Improving Care Coordination Between Veterans Health Administration Primary Care Teams and Community Home Health Aide Providers: A Qualitative Study

Design, Setting, and Participants

VHA's Homemaker/Home Health Aide benefit serves 136,000 Veterans with 10.5 million visits annually (Office of Audits and Evaluations, 2020), aiming to honor Veterans’ preferences to remain independent as long as possible. Veterans are generally eligible if they need assistance with 3+ activities of daily living (e.g., bathing, eating), or have significant cognitive impairment, live alone, or were recently discharged from a nursing facility. Eligibility criteria are set nationally, but can be tailored locally to meet individual medical centers’ needs. Veterans are referred from inpatient or outpatient settings, and typically receive 9–15 hours of care per week (Veterans Health Administration, 2006).

From September to December 2020, we interviewed VHA primary care teams, agency administrators, and home health aides as part of a quality improvement project to improve coordination between home health and primary care at a large urban Veterans Affairs medical center. The medical center (VAMC) provides clinic and home-based primary care services to over 14,000 veterans annually, half of whom are over the age of 65. The VAMC employs 20 full-time equivalent primary care providers. Approximately 600 veterans receive VHA-paid home health aide services from approximately 8 to 10 contracted community home health agencies.

Because our project examined coordination across the VHA and non-VHA caregiving team, we recruited participants from each group. We included primary care team members in different roles (physicians, social workers, and nurses) as they may have different interactions with and knowledge of aides’ roles. Agencies worked with multiple VAMCs and interviews reflected their experience across medical centers.

Data Collection

We recruited through purposeful and snowball sampling. Our team’s two physicians (KSB and MA) and social worker (NSK) identified primary care team members who regularly worked with older patients receiving home health aide services. The project lead and coordinator (EF and TR), both VAMC employees, recruited team members by phone and email, aiming for a balanced sample across roles (physician, nursing, social work). We requested interviews with administrators at the seven home health agencies serving the highest volume of VAMC patients. Administrators were then asked to refer 2–3 aides who provided care to veterans within the past year. Aides gave their employers permission to be contacted, and received a US$25 incentive.

We (EF and TR) conducted 30–45 minute semi-structured phone interviews. Interviews were recorded and transcribed verbatim. Two participants preferred not to be recorded; in these cases, the interviewer took detailed notes which were reviewed and verified by a second interviewer on the call. One aide interview was conducted in Spanish by a bilingual team member, translated by a professional transcriptionist, and reviewed for accuracy by the interviewer.

Questions were adapted for each group but focused broadly on: (1) perceptions of aides’ roles and tasks; (2) changes in aides’ roles and patient care during COVID-19; (3) communication between aides, agencies, and VAMC primary care teams; and (4) recommendations to improve coordination and delivery of care between VHA and contracted agencies in the future. We developed the interview guide collaboratively based on the researchers’ experience in geriatrics, primary care, and home health and emerging COVID-19 research. We piloted the guide with two VAMC primary care team members and an outside expert in qualitative interviewing with home health agencies and aides.

The site’s Research and Development committee determined this to be non-research quality improvement. Nonetheless, we were consistent with ethical standards of research, including obtaining consent and maintaining participant confidentiality.

Data Analysis

We conducted qualitative thematic analysis using a combined inductive and deductive approach (Fereday & Muir-Cochrane, 2006). First, three coders (EF, KMJ, and EG) independently reviewed three interviews (aide, agency and primary care team). We noted a priori topic areas (e.g., communication, information sharing) and emerging concepts (e.g., managing gaps in care) and developed an initial codebook. We applied the codebook to these interviews, discussing and refining it with additional team members (KSB, MA) until no new codes emerged. We then coded the remaining interviews independently. We maintained rigor by reviewing the accuracy of code definitions and their application in team meetings, comparing them against our analytic memos, and recoding segments as necessary. Primary care, aide, and administrator interviews were analyzed separately, then compared and contrasted to highlight emergent themes and divergent perspectives (Kendall et al., 2010). We recruited subjects until reaching thematic saturation (Hennink et al., 2017). Data were analyzed using NVivo 12. We adhered to the Consolidated Criteria for Reporting Qualitative Research (COREQ) (Supplemental File 1).

Establishing Home Health Services is Complex and Requires Many Individuals and Systems to Work Together

Participants across groups noted that coordinating home care services across the medical center, agency and patients’ homes was a multi-step process, involving physicians, nurses, social workers, service coordinators, agencies, and aides. The referral process began with the primary care provider. “If the patient needs an HHA, I would put the order for HHA, and then the social worker contacts the patient,” explained one physician. “Then, the nurses do their evaluation. And then, we will see the patient again in six months, usually.” (MD 3) Once the order was signed, coordinating services fell to the VHA home health coordinator, VHA social workers, and agency administrators. This included identifying a contracted agency to take the case, authorizing services, managing complaints from veterans or agencies, and at times moving veterans between agencies.

Both social workers and agencies noted described the point from referral to initiation of services as the “bumpiest,” with backlogs on both the VHA and agency sides due to administrative delays, staffing challenges and aide workforce shortages. “Once we get [the veteran] admitted and onboard, things get better…it’s that initially getting them on, and getting them the services, and getting the list of cases that [the medical center] needs filled,” noted one administrator (Agency 5). Another administrator shared that VHA caseloads appeared higher than the Medicaid managed care plans the agency worked with, which could delay the start of services, saying “you’ve got one [VHA] person overseeing hundreds of patients, of veterans. So…that’s a time-consuming process.” (Agency 4)

Sharing information between VHA and agencies could also be challenging. One administrator described the medical center's new secure online referral process as difficult to use. Prior to this change, VHA coordinators “would call us and say, send me over the list of who needs placement. We would send them an updated list of assessments…there was just a much more fluid communication.” (Agency 5) However, even between VHA staff timely updates were not always available. One social worker often referred veterans discharged from acute care to non-VHA home health aide services (for instance, through Medicare) because it was “faster, more efficient” whereas within VHA, a referral entered in a patient’s medical record “goes wherever it goes...you don’t have a sense of what’s going on unless somebody [the veteran or family] calls [to follow up].” She explained that “after months of waiting, you lose that patient,” recalling one referral that was canceled after the Veteran had waited for his initial aide visit for over a month and no longer needed the service (Social worker 3).

Communication between Medical Teams and Agencies is Often Reactive, Rather than Proactive

In general, primary care teams reported that once a referral was made, VHA was largely hands-off unless a problem arose. Physicians in particular reported that after signing an order for services, they had little knowledge of whether an aide was placed or what they were doing in the home. “I have like, a [patient] visit every 6 months, so if something happens in between, I have no idea,” noted one physician. “But we don’t monitor. I don’t know who’s monitoring the patient with the home health aide. I have no idea.” (MD 2) Another physician shared that while he reviewed the agency’s plan of care and other documentation, as long as the veteran’s needs were being met he “wouldn’t necessarily [seek] out” more information. (MD 3)

Nurses and social workers also noted that since VHA contracted out home health services it did not have oversight, and agencies primarily managed services on their own unless there was a clinical or service issue. “Honestly, once we put in the aide, whatever goes on or issues or complaints, it’s between the patient and the agency,” explained one nurse. “If there’s something more complex, the [VHA] social worker…gets involved.” (RN 2) In one case, an RN recalled reaching out to the agency directly to request a change in aides after a patient’s long-time aide died and the replacement was not “a good match.” Agency administrators similarly noted that “if there is a change in a patient’s condition, it’s something clinically or we feel that the veteran is not safe in their home...then we’re definitely going to involve the [VHA] team.” (Agency 1) However, this level of involvement could vary by site, depending on the medical center’s preference.

One case where communication appeared more proactive was during hospital discharges, particularly during the COVID-19 pandemic. Within VHA, these services are coordinated by the inpatient medical team rather than primary care. One administrator reported VHA staff were “very helpful and responsive” in sharing information about “who has COVID, when they were diagnosed, what their status is so we could then place the aide in quickly” compared to patients coming from community hospitals where “we couldn’t get proper information” (Agency 1). She described working closely with a VHA social worker to help one disabled veteran with COVID safely transition from hospital to home. “He had no family, no caregivers, [and was] an amputee and we really agreed [he needed an aide]…I requested that the VA provide a mask to the patient, then it would be easier for us to place the [aide]. They were able to right away [source masks]…we also gave the [aide] additional enhanced PPE.” However, for long-term care and community-admitted veterans, this type of coordination and information sharing was rare.

No Formal Communications Processes between the Primary Care Team and Agency

Both agencies and providers reported that there was no “formal process” or “consistent method” of communication between the agency and medical team, and contact was often ad hoc. Agency administrators shared that it could be difficult to “get hold of a live person” or find the right VHA staff member to call about patient issues, with one noting “it is difficult for us to make those connections as an outside person.” (Agency 5)

In the absence of formal communication channels, agencies often reached out to VHA contacts they knew. Several agencies reported contacting the local VHA home health services coordinator first. As an administrator shared, “we don’t really know who [the patient’s] doctor is” (Agency 1); and one social worker agreed that agency contact with the medical team could be “very tricky” because the agency would need to know the primary care physician’s name to reach them. (Social Worker 2) Another administrator shared that “if I can’t reach anybody else I usually go to the social workers and they reach out to the patient’s social worker or to the doctor or the team…they have to kind of navigate it for us.” (Agency 1) Veterans themselves were also an important line of communication; as one physician explained, veterans “have been with us for quite a while…they know who to contact, when to contact” if there was an issue. (MD 1)

Veterans enrolled in home-based primary care, VHA’s team-based home visiting program, were a notable exception. As one administrator noted, “we pretty much have a direct line [to the HBPC team] or they have a direct line to us if they have any problems in the veteran’s home.” (Agency 1) A nurse agreed that “[the agency staff] all have access to our numbers…whether it’s the nutritionist, therapist, case manager. But mostly, they call us, the nurse.” (RN 2)

Both agencies and providers, particularly nurses, felt more frequent and standard communication channels could be helpful for patient care, particularly during critical transitions or incidents. “It’s nice to know maybe [the veteran is] going through sundowning, or they are an Alzheimer’s patient,” shared one administrator. “Then we can educate the [aide] as needed.” He also noted it could be helpful to share contextual information with VHA if a veteran changed agencies to avoid the patient becoming “a ping-pong ball” between providers (Agency 4). One nurse in particular noted more proactive communication could alert the medical team to changes in the patient’s health. “It would be nice if [issues] were communicated to us, because then we know, hey, maybe something’s going on in the home. Why isn’t he opening the door for his aide? Maybe we should call 911…or if they are abusive to the aide, maybe it’s not just their behavior. Maybe it’s something mentally [or] physically that’s going on that we can help with.” (RN 9) However, when asked about how communication might become more formalized, participants were wary of adding new responsibilities onto already demanding schedules, asking “well, who’s going to do that?” (MD 3) or “I’m not looking to take that on.” (Social Worker 3)

Aides are an Important Source of Patient Information

Most participants across groups agreed that aides were an important source of patient information. One nurse called aides “the number one point of communication,” particularly if the patient was not able or willing to verbalize a need. “Sometimes [the veteran] may not call us, no matter how many times you tell them if they need anything just call me…but the aide steps in, and she will call and say, ‘I noticed this, this, and that’. And so that prompts another visit or just communication with the providers.” (RN 2) Another nurse shared that for some patients, “I will make the [home] visit when the aide is there because I know the aide knows where things are, I know that I can direct the aide or I know I have to say, ‘well, listen, here is a list of his appointments.’” (RN 1)

Similarly, social workers shared that aides could offer useful information about a veteran’s health status and social risks. One social worker explained that she would ask about needs that might indicate larger issues, asking the aide “Are you making sure he eats? How is he with his family? Are they close together? Do they call? Do they visit?” (Social Worker 1)

Aides working for veterans enrolled in the home-based primary care program reported more frequent contact with VHA nurses, often during home visits. One aide shared that her patient’s nurse was “helpful,” saying “if my patient is doing something that I know he didn’t do last week, [the nurse is] like oh, this is what happens with the dementia patients.” (Aide 1) Another noted that her patient’s nurse would often speak to her separately so she could share information freely. “I go by the door, and [the nurse] always ask me is [every] thing okay with him, do you see anything? And I always tell her the truth, if I see anything, he’s sad or something, I definitely give a heads up.” (Aide 6) In some cases, aides might proactively contact the medical team; for instance, one doctor noted an aide might “step in” and call the VHA’s 24-hour nursing hotline to report if a patient was dehydrated, not eating, or had cognitive changes. (MD 1)

Most agencies required aides to report concerns directly to their supervisor, even if they had spoken with a family member or VHA team member. “Our policy is that they come to us directly, to the nurse [supervisor] first, and then the nurse would communicate” to the VHA team if necessary, explained one administrator. (Agency 5) Certain issues, such as falls, were “must report,” while others relied on the aide’s judgment. In one case, an aide alerted her supervisor when she suspected her patient had sleep apnea, explaining “he would go for like 30 seconds without taking air… so I brought that to my supervisor’s attention.” (Aide 7)

Aides Reported Patient Information, but Rarely Received it

While aides were important channels of communication between their agencies, patients and VHA, they rarely knew the outcome of their reporting. As one aide explained, “if there is an issue, I call the [agency], and talk to the coordinator, and they talk to the nurse, and they talk to the doctor,” but that was often where communication ended. (Aide 4) Aides reported that they most frequently received updates from veterans themselves or family members. One aide specifically noted that she made an effort to “get close to the family” because “I’m not getting much input with the agency.” (Aide 8)

Lack of reporting back may be due to privacy concerns. Several administrators and social workers expressed concern that sharing patient information with aides could violate veterans’ privacy and that there was a “fine line” (RN 3). Nurses also felt sharing information was challenging because aides were not listed as contacts or providers in patient records. “The aide can call and give [information] but we can’t give the aide any medical information about the patient,” said one RN. “Most of the time we have to call whoever’s listed in their chart…if the home health aide happens [to answer, the veteran] might say okay. But we won’t ask any specific medical information because they’re not listed.” (RN 2)