At the onset of the COVID-19 pandemic, there was a rapid increase in the use of telehealth services at the US Department of Veterans Affairs (VA), which was accelerated by state and local policies mandating stay-at-home orders and restricting non-urgent in-person appointments. Even though the VA was an early adopter of telehealth in the late 1990's, the vast majority of VA outpatient care continued to be face-to-face visits through February 2020. Since VA telehealth programs vary by clinic, this proposed study examined how telehealth services were used at three types of outpatient clinics, primary care, cardiology, and home-based primary care (HBPC), within the VA Greater Los Angeles Healthcare System (GLA). Primary care is a gateway to all other care in the VA, and Veterans rely on it for the management of both acute and chronic conditions; cardiology manages a highly acute and medically vulnerable population; and HBPC has both a highly vulnerable population and a unique framework for supporting patients in their homes.
Aim1. To illustrate the use of telemedicine for each of the three clinics. a) Using the VA administrative and clinical encounter data from the VA Corporate Data Warehouse (CDW), we examined the rate and utilization patterns of synchronous tele-visits (e.g., telephone, VA Video Connect, and other VA-approved video conferencing platforms) 12-months before onset of COVID-19 (to create a baseline) and compared to 12-months after the onset of COVID-19 at each clinic. b) We identified the patient characteristics of telehealth users at each clinic by examining socio-demographic data (age, gender, race/ethnicity, marital status, health insurance coverage), and health risk scores (Nosos).
Aim2. To evaluate barriers and facilitators to achieving rapid implementation of telemedicine delivery. We conducted individual interviews with 34 key stakeholders and informants, including healthcare providers, hospital administrators and staff at the three clinics. Interviews queried respondents about facility preparedness policies and procedures with regards to telehealth, what types of telehealth resources were made available and what types of telehealth services were actually used, how telehealth services were tracked and coded, types of support received to transition to telehealth, ability to maintain continuity of care, and facilitators and barriers to implementing the telehealth response during the COVID-19 pandemic.
For this study, a parallel mixed methods approach was used where quantitative data management/analyses, as well as qualitative data collection/analyses were conducted simultaneously. For the quantitative portion, VA administrative and clinical data from the VA CDW were used. Outpatient visits were identified as either telehealth (telephone or video) or non-telehealth in-person encounters. For the analysis, total number of telehealth outpatient visits 12-months before and 12-months after the onset of COVID were first identified for each clinic (PC, cardiology, and HBPC). The monthly utilization of all telehealth outpatient services was calculated by dividing the total number of telehealth visits by the total number of outpatient visits for each clinic. The distribution of the percentage of monthly telehealth visits 12-months before and 12-months after the onset of COVID-19 was examined for each clinic. The distribution of the percentage of monthly telephone vs. VA Video Connect (VVC) visits were also examined 12-months before and 12-months after the onset of COVID-19. Additional multivariate analyses were conducted for each clinic cohort. Individual-level interrupted time series (ITS) analysis through segmented logistic regression on repeated monthly observations on telehealth use over 24-months (March 1, 2019 thru March 1, 2021), was used. ITS was divided into four segments: 1) pre-COVID, 2) onset of COVID (stay-at-home orders), 3) lifting of stay-at-home orders, and 4) start of the 2020-2021 flu season. Moreover, as outpatient records are nested in patients as well as providers, the analysis included patient- and provider-level clustering and adjusted for socio-demographic (age, gender, race/ethnicity, marital status, health insurance), as well as health risk factors (Nosos). The statistical significance level was set at p<0.05. All analyses were conducted in Stata (v.15) on the VINCI secure platform.
For the qualitative portion, semi-structured 30-minute telephone interviews were conducted with 34 GLA staff members who were involved in providing or supporting telehealth services within PC, cardiology, and HBPC during the COVID-19 pandemic. Respondents included: 18 clinical providers (physicians, nurse practitioners, RN care managers, and clinical fellows), 8 ancillary providers (social workers, psychologists, dieticians, pharmacists, and occupational therapists), 5 nurse managers, and 3 Health Administration Service (HAS) leaders. All interviews were audio recorded and transcribed. The study team utilized a rapid analysis approach to analyze the interview transcripts and prepare the dissemination of findings. Transcripts were divided and individually summarized by study team members. Summaries were created using a deductive approach, where data was grouped by domains based on the interview guide. Each team member conducted a randomized, secondary review of five to six summaries and discussed discrepancies with the team to ensure consistency in the data being recorded. Then the summaries were consolidated into three high-level summary documents, one for each clinic, to identify key points and commonly occurring themes across all interviews.
This study was approved by the VA GLA Institutional Review Board.
Quantitative Results (Aim 1)
The findings indicate that before the onset of COVID-19, for all three clinics, telehealth (telephone and video) use varied between 15% and 27% per month (PC: 16%-20%; cardiology: 7%-16%; HBPC: 16%-27%), the majority were telephone encounters. At the onset of COVID (during March 2020), however, telehealth use increased substantially: PC 44%, cardiology 45%, and HBPC 41%. After the onset of COVID-19, use of telehealth services continued to increase and reached its peak for PC at 80% in April 2020. For cardiology and HBPC, the peak was 70% and 79%, respectively, in May 2020. Starting in June 2020, use of telehealth services for all three clinics started to decline slightly, but never reached pre-COVID-19 levels during the 12-month post onset of COVID-19.
According to ITS analyses, at the onset of COVID-19, the odds of telehealth use were 9.2 (PC), 11.97 (cardiology), 3.52 (HBPC) times more likely compared to the end of the pre-COVID-19 time series. On the other hand, with the lifting of stay-at-home orders, the odds of telehealth use reduced to 10% (PC), 27% (cardiology), 20% (HBPC) of what telehealth use was when stay-at-home orders were in place. However, following the start of the 2020 flu season, the odds of telehealth use increased by a factor of 1.2 (PC), 1.05 (cardiology), 1.03 (HBPC), on monthly basis. For all 3 types of VA care (PC, cardiology, and HBPC), there were similar patterns of telehealth use during the 12-months after the onset of COVID-19.
In terms of social demographic factors, for PC, compared to non-Hispanic Whites, African Americans were more likely (OR:1.10, 95% CI: 1.05-1.14), whereas Hispanics were less likely (OR:0.93, 95% CI: 0.89-0.97) to use telehealth for PC visits. Married compared to not currently married (OR:1.08, 95% CI: 1.05-1.12), and Veterans with other health insurance coverage (OR:1.09, 95% CI:1.05-1.13) were more likely to use telehealth for PC visits. For cardiology, Whites compared to all others (OR:1.38, 95% CI: 1.23-1.54), married compared to not currently married (OR:1.25, 95% CI: 1.11-1.40), and Veterans with other health insurance coverage (OR:1.19, 95% CI:1.06-1.35), were more likely to use telehealth for cardiology visits. In terms of health factors, Veterans who had higher health risk factors, were less likely (OR:0.95, 95% CI: 0.93-0.97) to use cardiology telehealth visits. For HBPC, Whites compared to all others (OR:1.51 CI: 1.23-1.85) were more likely to use telehealth for HBPC visits. It should be noted that telehealth use was predominantly driven by use of telephone versus video-conferencing.
Qualitative Results (Aim 2)
The interviews with 34 stakeholders attempted to evaluate barriers and facilitators to achieving rapid implementation of telemedicine delivery. Three main themes emerged from the interviews regarding the transition to telehealth services: 1) telehealth expansion; 2) telehealth scheduling; and 3) telehealth modalities.
Telehealth expansion: All study respondents indicated that the rapid transition to telehealth was driven by the dual declarations of California's Stay Home Order and the VAMC suspending all non-urgent procedures. Leadership from each clinic met independently to discuss strategies for the transition. All non-essential appointments were converted to tele-visits. Staff in all three clinics began taking on additional roles, with many acting as champions to facilitate the switch to telehealth.
Telehealth scheduling: The successful transition to telehealth appointments was largely dependent on the level of communication between the scheduling clerk and the provider in each clinic. Each clinic had its own scheduling infrastructure, which in turn significantly impacted the way the clinic's providers perceived the transition to telehealth. Respondents in PC and HBPC reported limited scheduling challenges and confusion, due to PC providers having a close relationship with the scheduling clerks and HBPC providers scheduling their own patients. However, clerks in cardiology are not closely integrated into the clinics, so almost all respondents described scheduling as a key barrier to smooth telehealth adoption.
Telehealth modalities: Supporting our quantitative findings, respondents across all clinics and service roles described a heavy reliance on telephone as the main modality of choice in the initial transition period to telehealth. Providers reported that most patients preferred using the telephone, as many did not have the equipment necessary for a video visit or else found the technology to difficult to navigate. Further, the VVC platform was described as confusing for both patients and providers, and it would often take 10-15 minutes of a 30-minute appointment slot to explain to a patient how to get onto a VVC video link. Providers started seeing more patents in-person starting in June 2020, when the VAMC reauthorized non-urgent procedures and stay-at-home orders were lifted. However, most respondents cited the desire to continue virtual care in some form even after the pandemic is under control.
The movement to integrate telehealth into clinical practice has been growing for several years, but there have been significant barriers to widespread adoption. The COVID-19 pandemic, however, forced rapid expansion of telehealth services. Exploring the adoption of telehealth within a single VA medical center has provided the opportunity to understand the varied barriers and facilitators of different clinics and care providers. An individual VA medical center is an umbrella for a multitude of clinics and service groups, each with distinct needs and priorities.
- For all 3 types of VA care (PC, cardiology, and HBPC), there were similar patterns of telehealth use during the 12-months after the onset of COVID-19.
- There are some variabilities in use of the telehealth among different racial/ethnic groups and health risk factor levels.
- Veterans at GLA used telephone more than VVC for telehealth visits during COVID-19.
- The successful transition to telehealth appointments was largely dependent on the level of communication between the scheduling clerk and the provider in each clinic. Each clinic had its own scheduling infrastructure. This impacted how quickly each clinic transitioned to virtual care.
Our findings highlight the flexibility and creativity of VA clinical staff and leadership in rapidly responding to a massive disruption in healthcare after the onset of COVID-19. This underscores the need to understand individual clinic processes and workflows, in order to provide appropriate resources for each clinic.
- Midboe AM, Javier SJ, Salsbury SA, Katsovich L, Burgess DJ, King HA, Taylor SL, Martino S, Mayer JM, Wallace RB, Der-Martirosian C, Kerns RD. Impact of COVID-19 pandemic on nonpharmacological pain management trials in military and veteran healthcare settings: an evaluation informed by implementation science. Translational behavioral medicine. 2023 Aug 11; 13(8):601-611.
- Leung LB, Yoo C, Chu K, O'Shea A, Jackson NJ, Heyworth L, Der-Martirosian C. Rates of Primary Care and Integrated Mental Health Telemedicine Visits Between Rural and Urban Veterans Affairs Beneficiaries Before and After the Onset of the COVID-19 Pandemic. JAMA Network Open. 2023 Mar 1; 6(3):e231864.
- Bossarte RM, Ross EL, Liu H, Turner B, Bryant C, Zainal NH, Puac-Polanco V, Ziobrowski HN, Cui R, Cipriani A, Furukawa TA, Leung LB, Joormann J, Nierenberg AA, Oslin DW, Pigeon WR, Post EP, Zaslavsky AM, Zubizarreta JR, Luedtke A, Kennedy CJ, Kessler RC. Development of a model to predict combined antidepressant medication and psychotherapy treatment response for depression among veterans. Journal of affective disorders. 2023 Apr 1; 326:111-119.
- Cannedy S, Leung L, Wyte-Lake T, Balut MD, Dobalian A, Heyworth L, Paige NM, Der-Martirosian C. Primary Care Team Perspectives on the Suitability of Telehealth Modality (Phone vs Video) at the Veterans Health Administration. Journal of primary care & community health. 2023 Jan 1; 14:21501319231172897.
- Balut MD, Wyte-Lake T, Steers WN, Chu K, Dobalian A, Ziaeian B, Heyworth L, Der-Martirosian C. Expansion of telemedicine during COVID-19 at a VA specialty clinic. Healthcare (Amsterdam, Netherlands). 2022 Mar 1; 10(1):100599.
- Der-Martirosian C, Wyte-Lake T, Balut M, Chu K, Heyworth L, Leung L, Ziaeian B, Tubbesing S, Mullur R, Dobalian A. Implementation of Telehealth Services at the US Department of Veterans Affairs During the COVID-19 Pandemic: Mixed Methods Study. JMIR formative research. 2021 Sep 23; 5(9):e29429.
TRL - Applied/Translational
Care Management Tools, Technology Development
None at this time.