Provider Communication and Telepresence Enhance Veteran Satisfaction with Telestroke Consultations

Abstract

Background and Purpose:

Telestroke has been demonstrated to be a cost-effective means to expand access to care and improve outcomes in stroke; however, information on patient perceptions of this system of care delivery are limited. This study seeks to examine patient feedback of a national telestroke system within the Veterans Health Administration.

Methods:

Patients who received a telestroke consultation were eligible for a phone interview two weeks later, including questions about technology quality, telepresence, and telestroke provider communication. Satisfaction scores ranged from 1-7, (higher = more satisfied), and for analyses were dichotomized as 6-7 indicating high satisfaction vs. < 6. Patient variables including stroke severity (measured by the NIH Stroke Scale) were obtained from study records. Generalized estimating equation models were used to determine what factors were associated with patient satisfaction.

Results:

Over 18 months, 186 interviews were completed and 142 (76%) reported high satisfaction with telestroke. Patients with more severe stroke were less likely to recall the consultation. Factors significantly associated with patient satisfaction were higher ratings of the technology (p < 0.0001), telepresence (p < 0.0001), provider communication ratings (p < 0.0001), and overall VA satisfaction (p = 0.02). In the multivariate model, telepresence (OR 3.10, 95% CI 1.81-5.31) and provider ratings (OR 2.37, 95% CI 1.20-4.68) were independently associated with satisfaction. Veterans that were satisfied were more likely to recommend the technology (p < 0.0001).

Conclusion:

Provider qualities, including telepresence and provider ratings, were associated with overall Veteran satisfaction with the telestroke consultation. Technology quality may be necessary but not sufficient to impact patient experience. Training providers to improve telepresence could improve patient experience with telestroke consultation.

Introduction

Telestroke has been demonstrated to be a cost-effective means to expand access to acute care and improve outcomes in patients with ischemic stroke.^1,2 Originally described in the late 1990’s, telemedicine for acute stroke care uses video teleconferencing to provide timely access to vascular neurologic expertise and facilitate rapid treatment decisions.³ Furthermore, telestroke has the potential to reduce geographic and racial disparities in access to acute stroke care.^4,5 Numerous regional telestroke systems have been implemented around the United States with recommendations for incorporating telestroke into stroke systems of care now appearing in national policy statements.^6,7,8 In 2016, the Veterans Health Administration (VHA) implemented the first nationwide telestroke program. As of June 2020, 40 VHA hospitals are participating in this program including hospitals serving both rural and urban areas of the United States.

Evaluating patient and hospital level outcomes in telemedicine poses unique challenges. In 2016, the American Heart Association (AHA) published recommendations on measuring quality and outcomes in telestroke.⁹ Commonly tracked metrics include process metrics (time to consult, treatment, and transfer), accuracy of diagnosis, thrombolytic utilization, and patient safety outcomes.^10,11,12 In a review of current literature on telestroke quality, the majority of publications focus on these process or patient outcome metrics.

Less is known about the patient’s experience of a telestroke consultation. In general, telehealth for many conditions is associated with improved overall patient satisfaction with their healthcare.¹³ Systematic reviews suggest patients may both embrace and resist change, where patient age is an important mediator with older patients generally rating telehealth programs less favorably than younger patients.^13,14,15 While some of the early studies on telestroke did report general patient satisfaction,¹⁶ very few studies have undertaken an in depth analysis of the patient experience with a telestroke provider and health care delivery system.¹⁷ Finally, although some studies have examined telehealth satisfaction among Veterans for PTSD care, there is an overall paucity of data unique to this patient population.¹⁸ This study seeks to examine the patient perspective of the acute telestroke consultation within the context of a nationally implemented telestroke system, and to identify patient, provider, and hospital-level factors associated with patient satisfaction.

Methods

This project was conducted as part of the Department of Veterans Affairs National Telestroke Program (NTSP) Evaluation, which was approved by VHA national and local offices as an operational quality improvement project (non-research). As such, local IRB approval was not required and written informed consent was not obtained from individual participants (although verbal assent was obtained prior to the structured telephone interview). The data that support the findings of this study are available from the corresponding author upon reasonable request.

Patient Sample

Patients included in this analysis were Veterans receiving a NTSP telestroke consultation between the initial program implementation in September 2016 through March 2018. All patients receiving a consultation that included a video component were eligible to be called two weeks following the consultation for a telephone interview. We did not attempt to call Veterans who died, were placed on palliative care or hospice during the hospitalization, remained hospitalized four weeks after stroke, or had no identifiable telephone number in hospital records. All other Veterans had three telephone calls attempted, beginning at two weeks post-consultation. If the Veteran was not available or did not recall the consultation, a proxy who was present during the consultation was interviewed.

Structured Interview

The telephone interview included 18 questions with standardized measures of consultation technical quality, telestroke provider communication, telepresence (how much the interaction felt like the physician was in the room), overall satisfaction with telestroke care, and if the patient would recommend the service to another veteran (Supplemental File). Telestroke satisfaction was determined by taking the mean of two questions related to satisfaction. The final score then ranged from 1-7, with higher scores indicating greater satisfaction.¹⁹ Technology was a mean of five questions and provider communication was a mean of six questions all of which were scored on a 5-point Likert scale (higher scores = better). Telepresence was scored by taking the mean of two items, one which asked about the degree to which it felt like the physician was in the room and the other about the degree to which it felt like a face-to-face visit; these items were scored on a 1-7 interval scale (higher = greater sense of telepresence). Open-ended responses to a question about why the Veteran would or would not recommend telestroke to another Veteran were also recorded and categorized. Satisfaction with Veterans Affairs (VA) care in general was measured by taking the mean of 4 items which were scored on a 5-point Likert scale (higher scores indicate more overall satisfaction).

Other Variables

Patient demographics including age, stroke severity (measured by the NIH stroke scale), and telestroke provider diagnosis (stroke or TIA vs. other) were obtained from evaluation program data. Rurality was based on the Veteran zip code at the time of stroke, and classified according to VA definitions of rural and urban. Anonymous codes were assigned to each telestroke provider to facilitate provider-level analysis.

Statistical Analysis

The primary outcome of patient-reported satisfaction was measured with a mean of two Likert items ranging from 1 to 7 and had large ceiling effects since a high proportion of respondents were fully satisfied. Therefore, for analysis, the outcome was dichotomized as fully satisfied (mean score ≥ 6) vs less than fully satisfied (mean score < 6). Generalized estimating equations (GEE) methodology with a compound-symmetry covariance structure were further used to determine odds of increases in patient satisfaction with each outcome (OR calculated based on single point increases in mean scores for each outcome). The secondary outcome was the patient recall of the telestroke consultation, this was recorded with a binary Yes or No outcome. Bivariate logistic regression models were used to determine what factors were associated with patient satisfaction with telestroke and with patient recall of the consultation. Covariates of interest included patient age, urban/rural residence, stroke severity, whether the patient was diagnosed with a stroke during the visit, previous experience with telehealth, technology ratings, provider communication, telepresence, and VA Satisfaction.

Spline models were used to assess the linearity assumption of continuous variables with the log-odds of being fully satisfied. The final multivariable model was performed by selecting all variables with a bivariate p-value ≤ 0.25, and then using the same generalized estimating equation methodology as was used for the bivariate analysis with clustering by facility.²⁰ Odds ratios and associated 95% confidence intervals are reported. All statistical analyses were conducted using SAS V9.4 software (SAS Institute Inc., Cary, NC).

Results

Over the 18-month period, 776 telestroke consultations were completed, of whom 208 provided complete responses for the satisfaction ratings on the follow-up telephone questionnaire (Figure 1). Among these patients, 76 had a telestroke diagnosis of stroke and 29 were diagnosed with a TIA. Among those who were reached but did not complete an interview, 65 did not recall the consultation (Figure 1).

Figure 1:

Patient CONSORT Diagram

Of the 208 Veterans who completed the satisfaction questions, 186 had complete covariate information thus were used for all analysis. Of these, 142 (76.3%) were fully satisfied with the telestroke consultation (Table 1). Average age of the patients was 67.1 (13.1), 75.7% of the patients classified themselves as White, and more than 90% were male. There was no association between being fully satisfied and patient demographics (age, urban/rural residence), stroke severity, or prior experience with a telehealth service of any type. Patient rating of provider communication and telepresence, technology, and satisfaction with VA care were significantly associated with being fully satisfied with the telestroke consultation.

Table 1.

Satisfaction with Telestroke Consultation

		Fully Satisfied N =142 (76.3%)	Less than fully satisfied N =44 (23.7%)	Odds Ratio (95% CI)	p-value
Age	Mean (SD)	67.0 (13.2)	67.5 (12.5)	0.99 (0.97, 1.03)	0.822
Rurality	Rural	53 (37.3%)	20 (45.5%)	0.73 (0.39, 1.34)	0.307
	Urban	89 (62.7%)	24 (54.5%)	Ref
NIHSS	Mean (SD)	3.2 (4.5)	3.1 (4.5)	1.01 (0.94, 1.08)	0.877
Diagnosis of Stroke	Stroke/TIA	74 (52.1%)	18 (40.9%)	1.55 (0.76, 3.17)	0.231
	Other	68 (47.9%)	26 (59.1%)	Ref
Previous Telehealth Experience	Yes	43 (30.3%)	17 (38.6%)	0.69 (0.22, 2.19)	0.532
	No	99 (69.7%)	27 (61.4%)	Ref
Technology Rating (Range 1-5)	Mean (SD)	4.7 (0.5)	4.0 (0.9)	3.48 (1.92, 6.31)	<0.001
Provider Communication (Range1-5)	Mean (SD)	4.7 (0.6)	3.7 (0.9)	5.57 (2.67, 11.64)	<0.001
Telepresence (Range 1-7)	Mean (SD)	6.5 (0.8)	4.5 (1.7)	4.05 (2.46, 6.67)	<0.001
VA Satisfaction (Range 1-5)	Mean (SD)	4.4 (0.9)	3.9 (1.0)	1.65 (1.14, 2.40)	0.008

From the bivariate models (Table 1), variables of whether the patient was diagnosed with a stroke during the visit, technology ratings, provider communication, telepresence, and VA satisfaction were selected to be included in the multivariable logistic model since p-values were less than 0.25. From the generalized estimating equation models (Table 2), the odds of being fully satisfied with the consultation were higher in encounters with a higher telepresence rating (OR 3.15, 95% CI 1.83, 5.42) and a higher provider communication rating (OR 2.37, 95% CI 1.19-4.72, Table 2). However, technology rating, stroke diagnosis, and general satisfaction with VA care were not significantly associated with overall satisfaction. These results suggest that telepresence and communication play a larger role in overall consultation satisfaction.

Table 2:

Multivariable Logistic regression model of Telestroke satisfaction

	Odds Ratio (95% CI)	95% Confidence Interval	p-value
Diagnosis of Stroke	1.52	(0.45,5.17)	0.498
Technology Rating	1.48	(0.78,2.82)	0.227
Provider Communication	2.37	(1.19,4.72)	0.014
Telepresence	3.15	(1.83,5.42)	<0.001
VA Satisfaction	1.21	(0.84, 1.74)	0.299

Among the 142 patients who indicated that they were fully satisfied with the telestroke service, 97% indicated that they would recommend this service to another Veteran who was experiencing stroke symptoms. Interestingly, among the 44 patients who indicated that they were less than fully satisfied with the consultation, 69% would still recommend the service.

Finally, we evaluated 153 open-ended responses about why a Veteran would or would not recommend telestroke to another Veteran. Comments were generally focused on either the provider or the technology. Approximately 74% of comments were identified as generally positive, 17% as generally negative, and 9% as other. Comments illustrating each of these categories are shown in Table 3. The largest group of comments (45%) were positive comments about the technology, reflecting on the ease and speed of access to a specialist. Of the negative comments, most related to a general dislike or distrust of the technology.

Table 3.

Example open-ended responses about recommending Telestroke

Positive Technology (N = 69, 45%)29%)	Positive Provider (N = 45, 29%)	Negative Technology (N = 20, 13%)	Negative Provider (N = 6, 4%)
“It’s easier to be connected directly to a specialist than to have to wait for a specialist to arrive on site…it was like Star Trek!”	“As soon as the Telestroke doctor came on the screen he was very clear and immediately got my confidence…fantastic about letting me know what was going on.”	“It’s not the same on the computer…it didn’t feel personal.”	“They would take the screen out in the hallway and speak to the [other] doctors, and that made me feel uncomfortable.”
“It was done so quickly; they brought the stand in and it was instant. She did an exam on him and she could have been anywhere in the world—it was just so quick!”	“She [Telestroke doctor] was very calm, cool, collected. Her treatment felt specialized, and that was very comforting.”	“There is no way a doctor in [another state] could diagnose me with a brain bleed.”	“I didn’t have any input…he [Telestroke doctor] was talking to the other doctor.”

Factors associated with Veterans’ recall of the telestroke consultation are shown in Table 4. Recall of the consultation was associated with lower stroke severity (mean NIHSS 2.8 vs. 5.7, p = 0.03). Patients with a stroke diagnosis were nearly three times more likely (OR 2.88, 95% CI 1.09-7.60, p=0.033) to recall the consult compared with those with a non-stroke diagnosis (mimic).

Table 4.

Recall of Telestroke Consultation

		Remember Consult N = 147 (90.2%)	Do Not Remember Consult N = 16 (9.8%)	Odds Ratio (95% CI)	p-value
Age	Mean (SD)	66.6 (12.9)	67.8 (6.9)	0.99 (0.96, 1.03)	0.738
NIHSS	Mean (SD)	2.8 (3.9)	5.7 (7.8)	0.90 (0.82, 0.99)	0.027
Diagnosis of Stroke	Stroke/TIA	79 (53.7%)	5 (31.3%)	2.21 (0.75, 6.53)	0.153
	Other	68 (46.3%)	11 (68.8%)	Ref

Discussion

In this sampling of a nationwide telestroke service, we found that patient ratings of provider communication and telepresence (how much the interaction felt like the physician was in the room) were the strongest predictors of overall satisfaction. Importantly, these factors are to some extent modifiable through provider training and technological factors, and they ultimately may influence whether a patient would recommend this service to another patient. To our knowledge, this is the largest study of patient-reported telestroke experiences focusing on factors related to satisfaction with their consultation.

Over the past 20 years, remote provision of acute stroke care by telemedicine has dramatically changed emergency care delivery for cerebrovascular diseases with increasing emphasis being placed on ensuring that these services are providing high quality and positively affecting patient outcomes.⁹ Early work focused on ensuring technical quality of the consultation including remote assessment of the NIHSS^{21, 22} and rapid access and interpretation of CT scans.²³ Subsequently, recommendations evolved to standardized data collection to examine process times and outcomes.²⁴ From numerous robust telestroke datasets, it is now established that telestroke systems have the ability to achieve treatment rates and time based metrics in similar to in-person care at hub centers.^{11, 25-27} In fact, telestroke may be associated with higher utilization of IV alteplase, although this may also reflect the improvements in stroke education and processes at spoke sites that implement telestroke services.²⁴ Finally, several reports have demonstrated the safety of care by telestroke as well as long term impacts on functional outcomes and mortality.^{12, 28-30}

While the large body of literature on telestroke quality speaks to the enormous benefits of this care delivery model, it is relatively lacking in data from the patient perspective. Understanding factors that influence patient satisfaction with this type of service is critical for its ongoing acceptance and success. The existing literature suggests that telestroke systems are generally well received by patients and felt to enhance care delivery.^{16, 31} Recently, a study of 29 patients conducted semi-structured interviews of patients and caregivers examined experiences of a telestroke system. Although general acceptance of the consultation was noted, several issues regarding information exchange and interpersonal skills were highlighted.¹⁷ Our study adds to this literature by providing determinants of satisfaction from a large national sample. Similar to the previous report, telestroke provider communication skills with the patient as well as local care team providers were highly valued. An important pitfall of inadvertently excluding the patient from care discussions was noted and offers a potential target for telestroke improvement strategies and is part of the broader concept of telepresence. Our study also adds to the literature by identifying factors related with ability to recall the conversation. As expected, having less impairments (lower NIHSS) was associated with recall among patients we were able to contact and those with a final diagnosis of stroke were nearly 3 times more likely to recall the interaction.

The idea of telepresence dates back for nearly three decades and was initially described in the human factors literature.³² Early experience with telepresence in medicine was defined from robotics and surgical care.^33-35 Eye contact and nonverbal signals can enhance the perception of sharing space with a remote site.³² In our sample, we assessed telepresence by asking to what extent the telestroke consultation felt like a face-to-face experience and to what extent did the patient feel that they were engaged with the telestroke physician. Telepresence is clearly highly associated with overall satisfaction. Skills that may impact telepresence may not be intuitive and require education and training for providers in telestroke and other telemedicine fields. The telestroke provider frequently has divided attention between the patient, family members, local care providers, robotic camera controls, electronic medical records, and neuroimaging interfaces. In this context, it is easy to lose focus on the patient which may be perceived as disengagement. Subjective comments in this study reflect this and serve as a reminder that the patient remains at the center of the consultation.

Finally, several components of our interview focused on technological aspects of the consultation including audio and visual clarity and ease of interfacing with the on screen physician. Only 30% of our patients had prior experience with any form of telemedicine although this did not seem to influence satisfaction. Given that there are many platforms and interfaces available for telemedicine services, it is important to know the technological limitations of each and have easily understood processes in place to address technological problems. Although we found that video and audio clarity is certainly important in effective communication, our results suggest that adequate technology is not sufficient to define a successful consultation from the patient’s perspective. Ultimately, communication and provider skills in telepresence play a larger role.

This study has several strengths including being the only prospective sampling of patient perspectives from a national telestroke program, including patients who were ultimately not diagnosed with a cerebrovascular disease. The sample is limited to those patients who were able to provide responses which excluded patients with more severe strokes, ongoing hospitalization, or those with deficits limiting participation (although proxy/caregiver responses were collected when possible). Response rate, although not terribly out of proportion to other survey based studies, is low at 25% which raises the possibility of non-response bias. In addition to collecting standardized questions, Veterans were able to provide open ended responses about their care to identify factors influencing patient perceptions that were not collected otherwise. There are also limitations to this study. Given the variability in telestroke systems based on system structure and technology interfaces, our findings may not be generalizable to all telestroke programs. This study also was conducted among a Veteran population with a unique relationship to their healthcare system, including healthcare deliver perceptions, and perhaps increased access to telehealth technology which has been employed widely in the VA for more than a decade.

Notwithstanding these limitations, this study was able to demonstrate several determinants of patient satisfaction in a telestroke consultation. Provider communication and telepresence are key to patient satisfaction with teleconsultation; provider training in these areas will likely lead to improved patient perceptions of telestroke. As telestroke and telehealth systems of care expand, we must not only measure clinical outcomes and processes of care associated with telehealth but also must keep in mind patient perceptions of the program. Having fully satisfied patients will help ensure continued expansion and success of these programs which can efficiently increase access to highly specialized care for patients nationwide.

Non-standard Abbreviations and Acronyms

VHA

Veterans Health Administration

NTSP

National Telestroke Program

VA

Veterans Affairs