One semistructured interview guide for veteran patients and one for clinicians (Multimedia Appendix 1) was developed by the investigator team using the updated Consolidated Framework for Implementation Research [14]. The veteran interview guide was developed based on a prior veteran survey about RM [15] and revised with input from the Rural Colorado Veteran Research Engagement board. The clinician interview guide was developed through an iterative process with input solicited from practicing VHA cardiologists and the incorporation of concepts from new HRS recommendations [2].

Both interview guides sought to understand barriers and facilitators to an “RM-first strategy,” defined as in-person CIED clinic visits only if clinically prompted among patients engaged in RM. Patients were informed that similar data were obtained through RM as in-person visits; they may need in-person visits for abnormalities identified on remote transmissions; they could still contact their device clinic; and their other visits, such as with primary care, would continue. Patients were asked about the travel burden to VHA, how their care may have changed during the COVID-19 pandemic, and any concerns about reducing routine in-person CIED clinic visits. Device clinicians were asked about the benefits and barriers to this new care model, and how this may change their practice flow. A 23-item Qualtrics survey was also administered to gather professional and demographic data as well as preinterview information about clinician impressions of RM-first care (Multimedia Appendix 1). Specifically, this survey asked clinicians how often they conducted routine evaluations for patients with CIEDs, stratified by adherent and nonadherent patients, and what clinicians did when patients did not want to schedule routine in-person CIED checks or missed an in-person CIED check. This survey also asked clinicians about the anticipated benefits and concerns of an RM-first strategy, how effective that it would be concerning cardiovascular outcomes, and if such a strategy would help their clinic.

Of note, partway through the clinician interview process, the draft 2023 HRS expert consensus was released [2], introducing an “alert-based care” model, similar to RM-first care. Therefore, the interview guide was then adapted to solicit feedback about this recommendation. For the veteran interviews, a question was added about the veteran’s view of the new recommendations.

This was a quality improvement project conducted in partnership with the VHA Measurement Science Quality Enhancement Research Initiative and the VHA National Cardiac Device Surveillance Program.

Veterans were eligible for interview inclusion if they had a CIED, were completely adherent to RM in the past 400 days (which means that they had sent a remote transmission covering this timeframe), [12] and lived in a rural area. Introductory letters were sent to 100 randomly selected veterans meeting these criteria (since these participants did not know the project team), 91 of whom were then contacted at least once via a telephone connection to Microsoft Teams. The letter described the study background and objectives as well as topics that would be covered by a named VHA staff member (SM). Up to 3 contact attempts were made, with a message left for each unanswered attempt.

A purposive sample of VHA CIED clinic-focused clinicians who had been interviewed for a prior project about best practices to support RM adherence were contacted for interview [16]. An introductory email described this study’s background, objectives, and potential changes that may result from findings as well as information about the project team and funding source. Snowball sampling was then used, asking these clinicians to recommend colleagues at their device clinic. Finally, purposive sampling was used to contact clinicians caring for a high proportion of veterans living in rural areas to more adequately represent rural clinician perspectives.

Informed consent was obtained before recording all interviews, which were conducted on and recorded using Microsoft Teams. Between November 2022 and February 2023, a total of 40 veterans were interviewed by coauthor SM (BS, male, qualitative researcher), with each of these 40 individual interviews lasting 5‐15 minutes in length and some attended by coauthors TLR (MPH, male, public health researcher) and SSD (MD, MHS, male, cardiologist). Between November 2022 and February 2023, a total of 22 clinician interviews between 30‐60 minutes were conducted by TLR, with some attended by SSD. Field notes were taken during both sets of interviews to summarize key points and supplemented with transcribed interview recordings to ensure accuracy. There were no repeat interviews.

Reflexive thematic analysis [17,18] of interview field notes and transcripts was used to elucidate veteran and clinician views about RM-first care.

First, authors AK (MD, female, cardiology fellow) and TLR familiarized themselves with the dataset by reading the field notes and transcripts, making notes about the overall findings within both sets of interviews (veteran and clinician) and reflecting on their experiences in the direct care of patients with CIEDs (AK) and research and quality improvement efforts for care of patients with CIEDs (TLR). Next, the authors generated separate codebooks based on the domains of the distinct interview guides. For veteran interviews, AK and TLR independently coded 6 distinct interview notes, which involved generating additional codes identified inductively, for the goal of reflexivity. These 2 authors then met and reviewed each other’s codes, sought SSD’s input, and resolved any differences by consensus, creating 1 final codebook. AK then coded the remaining interviews and developed candidate themes, supporting each theme based on coded data and direct quotations. AK’s candidate themes were intentionally broad. TLR and SSD reviewed these themes with AK against the coded data, leading to refining and then naming these themes. Finally, AK wrote the analytic narrative and supported these themes with quotations directly from the veteran interviews to describe veteran perspectives. Coauthor SSD provided iterative feedback on several versions of the analytic narrative to improve clarity and increase confirmability.

For clinician interviews, AK and TLR first independently coded 3 distinct interview notes, which involved generating additional codes identified inductively. These 2 authors then reviewed each other’s codes and resolved any differences by consensus. AK then coded the remaining interviews. The authors used the same process as described above for thematic generation, refinement, and naming. AK wrote the analytic narrative, which is presented in the Results section of this paper, and supported these themes with quotations directly from the interviews. We conducted both clinician and patient interviews until reaching thematic saturation on two criteria, (1) no new concepts were identified in iterative analysis interviews (code frequency counts) and (2) there was consistent repetition among interviewee responses without any new information being added to existing codes (code meaning) [19,20]. The number of interviews that we conducted with both our population of veterans and Veterans Affairs (VA) clinicians exceeded the number (n=17) found in recent empiric studies [20].

Atlas.ti 23 (ATLAS.ti Scientific Software Development GmbH), a qualitative analysis software, was used to organize and apply analytic codes.

This work was conducted as a quality improvement project and not human subjects research. Per the Department of Veterans Affairs Office of Research & Development Program Guide: 1200.21, “VHA (Veterans Health Administration) Operations Activities That May Constitute Research,” data were collected as part of a quality improvement study to assess and improve the quality of RM care for veterans with CIEDs and did not require institutional review board approval. Veteran and clinician participants were informed at study enrollment that responses would be anonymized, and verbal consent to recording was acquired before each interview. No compensation was provided. Study data were deidentified and stored in a secure, encrypted VA database.

Of 22 clinicians interviewed, 20 (87%) participated in the survey, 14 (64%) of which were fully complete. Of the 20 respondents, 6 were MD/DOs, 7 were advanced practice providers (APPs), 6 were registered nurses (RNs), and 1 was a medical instrument technician (Table 3). Ten self-identified as female and 6 self-identified as non-White. Almost half of the respondents had been working at their current VHA cardiology clinic for >10 years. All clinicians were focused on CIED-related care and were not serving as patients’ primary cardiology clinician.

The most commonly reported scheduling frequency for routine in-person ICD and pacemaker evaluations was every 12 (range 4‐12) months, used by 72% (n=13) and 83% (n=15) of clinicians, respectively (Table 4).

Seven (39%) clinicians reported using an RM-first strategy for some patients. Sixteen (89%) thought this strategy would improve veteran convenience by reducing appointments and travel time. Six (33%) expected it would enable more care for other patients with heart rhythm disorders.

However, 12 (63%) clinicians were concerned about a reduction in the quality of veteran care and 10 (53%) about veteran-perceived abandonment. Fifteen (83%) respondents were confident that an RM-first strategy was as effective as RM with in-office visits regarding cardiovascular outcomes, while 3 (17%) were not. Seven (39%) expected an RM-first strategy would benefit their clinic, 7 (39%) were undecided, and 4 (22%) thought it would not.

Clinicians thought that patients who were the best candidates for RM-first care were those without cardiac resynchronization therapy (CRT) devices who were adherent to RM, clinically stable and noncomplex, not device-dependent, not having frequent arrhythmias, good communicators, and facile with technology. One APP explained,

Many clinicians expected the decision about appropriateness for an RM-first strategy would initially be determined by the patient’s clinician, as an APP explained,

When asked how an RM-first care model should be evaluated for success, most clinicians thought patient satisfaction should be a key indicator, along with patient RM adherence. As an APP said,

Respondents also thought it would be important to ensure there was no increase in adverse outcomes or rhythm issues not being identified promptly.

Respondents also discussed potential time savings with an RM-first approach. As an RN said,

Many interviewees also noted that monitoring for missed RM transmissions would be central for a new RM-first care model, but most already had a process in place for doing so. One APP explained,

The 2023 HRS expert consensus statement introduced “alert-based remote monitoring,” defined as “a combination of continuous connectivity with clinic visits that are prompted only by the detection of actionable events,” [2] which provides the basis for the RM-first care model that we discussed with veterans and clinicians. Both expressed interest in this model of CIED care and cited the benefit of reducing patient travel burden and enabling clinical bandwidth to care for other patients. However, patients sometimes preferred in-person evaluations (generally for non-CIED related medical reasons and the patient-clinician relationship), and some expressed concerns regarding technological issues with RM. Given the VHA’s central RM infrastructure that reviews all remote transmissions, VHA is well-positioned to implement and study this care model, which could inform other health systems and clinicians about the context of implementing RM-first care. Indeed, most clinicians expected that RM-first would ultimately become the standard of care for CIED management.

There is often substantial lag in implementing research and consensus recommendations into clinical practice, including inertia in initiating new care models [21,22]. Reasons for such inertia include overestimation of existing care as well as lack of practice organization to achieve therapeutic goals [22]. Providing patient and clinician education and support when implementing an RM-first care model will be important to overcome inertia, leverage facilitators, and surmount barriers.

Some patients worried about the quality of RM. To address this, patient-centered RM education should be provided before transitioning to RM-first care and emphasize to patients that any actionable findings on RM will prompt appropriate clinical actions, sometimes including in-person evaluations. Additionally, for patients to qualify for this care strategy, they need to be consistently and continuously connected to RM so clinically actionable events can be identified promptly. Thus, patients should be educated about ensuring RM connectivity and troubleshooting strategies based on their specific transmitter. Patients and clinicians also raised concerns regarding the loss of the in-person relationship and the inability to perform in-person assessment, such as a physical examination. To address this, device clinicians should ensure that patients have regular follow-ups with their general cardiologist or electrophysiologist (as appropriate) or at least routine primary care, and that the device clinic is not their primary source of cardiology care.

Clinicians also noted a potential increased risk of patients being lost to follow-up. Clinics must have a method of tracking patients outside of in-person visits and ensuring RM adherence [16]. Patients who become disconnected from RM will require in-person evaluation. Finally, patients and clinicians raised concerns about technical comfort with troubleshooting home monitors and RM adherence, which requires a high workload [23]. To alleviate this burden, postcard reminders that recommend patients contact their CIED manufacturer for assistance have been shown to increase RM adherence, without burdening clinicians [24]. Additionally, sending informational text messages to recently disconnected patients can improve RM adherence [25].

Although there are several barriers to be addressed, the RM-first care model has the potential to provide many improvements for patients and clinicians. With the growing potential of digital health technology in cardiovascular medicine [26], the lessons from our study have broad applicability but it will be critical to ensure that an RM-first care model, as with any virtual care modality, is implemented equitably [27,28]. Reduced patient travel burden is particularly important for patients who live in rural locations. From a reimbursement perspective, while VHA is a single-payer, other health care payers would need to adopt novel reimbursement strategies for RM that facilitate sustainable and cost-effective CIED follow-up care [2,29,30]. Finally, a reduction in unnecessary device-related clinic visits will allow clinicians to see other patients with heart rhythm disorders and reduce wait times, which may result in higher-value care, particularly given the shortage of cardiovascular health professionals [31]. An RM-only model has been successfully implemented at a large clinic in Italy since the COVID-19 pandemic and was associated with time savings for clinicians and patients with no increase in adverse clinical outcomes [32]. Further, although not currently available, if remote reprogramming is demonstrated to be safe and feasible to implement, it could further reduce the need for in-person visits and could improve patient perceptions around an RM-first care model.

Our study should be considered in the context of its limitations. First, although we studied a single health system with specific patient population demographics (more often rural, predominantly White, and predominantly male) and clinicians providing care in an integrated health care delivery system, the Veterans Affairs National Cardiac Device Surveillance Program (VANCDSP) centrally monitors more than 64,000 veterans with CIEDs, making VHA well-positioned to implement and evaluate RM-first care. Future studies should evaluate other patient populations, which would help to assess the transferability of our findings. Second, although this was a national study, our results represent a limited number of both patient and clinician perspectives. However, qualitative methods intentionally provide granular data from smaller numbers of participants, patients were randomly selected, and our methodology provided detailed information on perspectives from clinicians across the United States. Third, interviews were conducted while new HRS consensus was released in draft form [2], so questions were modified partway through the interview process, and the ideas being introduced were new; patients and clinicians may feel differently when they have had more time to assimilate the recommendations. We did not inform patients about the additional safety offered by consistent and continuous RM connectivity. Fourth, we did not interview patients who were new or nonadherent to RM. Fifth, we did not have participant validation of our findings. Sixth, this study’s team represented an institution (VANCDSP) with some influence on both patient care and clinical support. While it was not apparent in the review of interview recordings or transcripts, this power dynamic may have incentivized veteran patients and clinicians to speak more favorably of the VANCDSP or caused interviewees to present their care or their patient’s existing care in a more favorable light. Finally, this study represents patient and clinician expectations of RM-first care, instead of their views based on experience; as RM-first is implemented in the future, patient and clinician perceptions on barriers and facilitators to this care model should be evaluated.

Both patients and CIED clinicians experienced in RM within the VHA were open to an RM-first care model that reduces in-person visits but conveyed barriers about solely relying on RM and possible diminution of the patient-clinician relationship. Implementation of new RM recommendations will require attention to these perceptions and prioritization of patient-centered approaches.