Abstract
Objective
To determine whether synchronous video-based telemedicine visits with specialists are feasible and to evaluate clinical effectiveness, patient perceptions, and other benefits of telemedicine visits for follow-up migraine care in a tertiary headache center.
Design
A one-year, randomized clinical trial.
Results
Fifty patients were screened and 45 entered the study (43 women, two men). Out of 96 scheduled visits, 89 were successfully conducted using telemedicine. Eighteen patients (out of 22) in the telemedicine cohort and 12 patients (out of 23) in the in-office cohort completed the study. In this small study, clinical outcomes, namely improvement in MIDAS, number of headache days, and average severity at 12 months for participants in the telemedicine group, were not different from those in the in-office group. Convenience was rated higher and visit times were shorter in the telemedicine group.
Conclusions
In this cohort of patients with severe migraine-related disability, telemedicine was a feasible mode of treatment and an effective alternative to in-office visits for follow-up migraine care. Physician productivity could be higher with telemedicine, and patients may get better access because of its convenience.
Trial Registration
This study is listed on ClinicalTrials.gov (NCT01706003).
Background and introduction
Despite the large disease burden of migraine, there are currently fewer than 600 headache medicine specialists in the US and Canada with United Council for Neurologic Subspecialties (UCNS) certification, and they are largely concentrated in limited geographic areas (1,2). Individuals with migraine face similar challenges regarding access to care in other countries too (3). Increasing the pool of qualified providers is difficult, as headache medicine fellowship programs are largely unfilled (4). Thus, there is a growing need for alternative means of providing increased access to care for patients with migraine, particularly for those patients with severe migraine-related disability.
This subpopulation generally requires visits at relatively frequent intervals in order to obtain optimal care. They thus face a time commitment that far exceeds the duration of the office visits and costs pertaining to travel, child and dependent care, and time missed from work or school. They may also need an accompanying person in cases of long-distance travel or an inability to drive during a migraine attack. Missed visits occurring because of illness or inclement weather may be difficult to reschedule as well, leading to lapses in care.
Telemedicine offers a way to increase access to care by providing clinical expertise to patients in diverse locations and circumstances. Migraine care lends itself well to this technology, as there is generally no need for a physical examination after the initial office visit. Telemedicine allows for face-to-face interaction so the provider can review the interim history and medications taken, assess treatment response, evaluate the need for additional testing, discuss pharmacologic and non-pharmacologic treatment options, educate, and prescribe a course of action. A typical follow-up visit for migraine patients includes at least 20 to 30 minutes with a provider.
To our knowledge, long-term follow-up care for migraine patients using synchronous two-way video-based telemedicine visits with specialists and patients logging in from convenient locations has not been assessed systematically. Therefore, over a one-year period, we evaluated such care in a headache medicine program at a tertiary referral center using telemedicine for migraine patients with severe headache disability. We hypothesized that follow-up care using telemedicine would be feasible and lead to patient outcomes and patient satisfaction similar to those of in-office visits.
Methods
Enrollment and study procedures
Inclusion and exclusion criteria.
After we obtained informed consent, participants were randomized by a research coordinator using an on-line randomization tool to receive their follow-up care either by telemedicine or through in-office visits (1:1). All the study participants and the investigators were blind to the randomization process. All participants completed a Migraine Disability Assessment Scale (MIDAS) and an allodynia questionnaire during their initial visit. Follow-up visits were scheduled at 4–6 weeks and 3, 6, 9, and 12 months. All visits (including the initial visit) were conducted by a single headache medicine specialist (DIF) using a similar format, covering interim headache and medical history (including emergency department/urgent care visits and hospitalizations), review of current medications, response to treatment including adverse reactions, descriptions of headaches, allergies, blood pressure, and weight.
Participants randomized to telemedicine had their blood pressure and weight measured within 5 days of each telemedicine session at a convenient location. A study coordinator scheduled and confirmed the visits. We attempted to reschedule all missed visits in order to have a full complement of visits for evaluation. The length of each visit was recorded. Participants were offered reimbursements for their copayments, travel expenses (mileage) and parking fees so there was no difference in visit costs between participants.
Participants were assigned a unique identifying code to maintain confidentiality for on-line data capture (BR). All the online surveys used this unique code to capture patient data and feedback. In-office visits took place in the Department of Neurology outpatient clinic at UT Southwestern. Telemedicine visits were conducted using Zoom (Zoom Video Communications, San Jose, CA) software and were scheduled on-line using the identification code. The telemedicine visits were synchronous two-way audio- and video-based, and participants chose locations convenient to them. Participant-initiated office communications, other medical conditions, extra visits, procedures, and hospitalizations were accessed from the medical record.
Questionnaires
All participants completed an initial questionnaire that gathered demographic data, such as age, distance from the office, travel time and total time required for office visits, activities missed for study visits, and costs associated with study visits (e.g. child care), and inquired about their perceptions of telemedicine. Prior to each study visit, enrollees completed an on-line questionnaire (SurveyMonkey, San Mateo, CA) regarding their interim progress, status, medication tolerance, and MIDAS score. After each follow-up visit, all participants completed short questionnaires that were appropriate to the assigned study group. We asked each participant to complete an on-line questionnaire (incorporating a modified version of the Group Health Association of America’s Consumer Satisfaction Scale) after the 12-month visit, which inquired about their overall experience. After the final visit, one of us (BR) also interviewed willing enrollees by telephone.
Outcome measures and data analysis
Excel 2016 and R (version 3.5.2) were used to perform the statistical analysis. The primary outcome measure was the percentage of visits using telemedicine completed as scheduled; the sample size was chosen to provide evidence for this proof of concept (5). Secondary outcome measures included clinical outcomes related to headache disability (MIDAS, number of headache days in the previous 3 months, and severity on a scale of 0–10 with 10 = pain that is as bad as it can be), visit time, and perceptions of telemedicine. The average change from baseline to 12 months for the in-office group was compared with the average change for those in the telemedicine group (difference in average difference). The McNemar χ2 test was used to examine differences in the proportion of participants reporting moderate to severe disability (MIDAS ≥ 11) between the baseline and 12 months (6). Logistic regression was used to test whether membership in the in-office group was associated with experiencing a 50% or greater improvement in clinical outcomes. The logistic regression was also controlled for baseline values. Two sample t-tests were used to estimate whether participants in different groups differed in their survey responses.
Results
Patient characteristics at baseline.
Women are much more likely to seek care for migraine and headaches than men in the US. It may also be a practice/referral bias (17,18).
+Within the preceding 3 months.
Routine appointments and study adherence
There were 200 possible appointments, based on five visits each for the 40 enrolled participants who had at least one visit after the baseline (90 in-office visits and 110 telemedicine visits). Of these, 173 visits were scheduled (77 in-office and 96 telemedicine), and the others were lost to follow-up due to administrative and coordination issues or lack of participant response. Five appointments were cancelled by the patient (four in-office, one telemedicine), three were cancelled by the provider (one in-office, two telemedicine), and nine were no-show appointments (five in-office, four telemedicine). Overall, 156 visits (67 in-office and 89 telemedicine) were completed successfully. Thus, 92.7% of telemedicine visits scheduled were completed successfully, in contrast to 87% of in-office visits. Of the 18 participants randomized to the in-office group, 12 completed the 12-month visit. In the telemedicine group, 22 participants entered the study and 18 completed the 12-month visit.
Technical difficulties with the telemedicine system (two audio, one video) occurred during three visits for two participants. In each case, the visit was conducted by telephone instead.
Clinical course
Feasibility and clinical outcomes (95% CI).
The percentage of participants reporting moderate to severe disability decreased from the baseline to 12 months in both groups (91.6% to 66.6% for the in-office group and 83.3% to 55.6% for the telemedicine group), but based on the McNemar χ2 test the change was not significant (p-value = 0.37 and 0.13 respectively). Based on our logistic regressions, telemedicine patients were not associated with a lesser likelihood of experiencing a 50% or greater reduction in MIDAS (odds ratio: 1.25, 95% CI 0.29–5.54, p = 0.77), number of headache days (odds ratio: 1.12, 95% CI 0.25–5.09, p = 0.88), or severity (odds ratio: 0.29, 95% CI 0.05–1.35, p = 0.12). Our conclusions did not change even after we controlled for the baseline scores in the regression.
Time distribution for visits
The mean total time spent (including travel) was 194 minutes for the in-office group (n = 46 visits) and 37 minutes for the telemedicine group (n = 64 visits). The difference in average time spent is statistically significant (p < 0.001). Individuals in the in-office group spent more face-to-face time with the physician on average than those in the telemedicine group (34 minutes vs. 25 minutes, p = 0.01). However, patient perception regarding visit times were similar (see Figure 1).
Patient perception about length of visits.
Convenience of visits
Based on post-visit responses at 6 weeks, 6 months, and 12 months, participants using telemedicine (n = 40) rated the convenience of visits as excellent (80%) or very good (10%), in contrast to those in the in-office group (n = 29; excellent: 24%, very good: 31%, see Figure 2). Using a 5-point Likert scale, the average score for the convenience of telemedicine visits was 4.7 and the average score for in-office visits was 3.5 (p < 0.001).
Patient perception about convenience of visits.
Time missed from work
Participants in the in-office group (n = 46 visits) were more likely to take considerably greater time off from work to complete their visits. Four patients needed to take a day or more to travel and seek care. The maximum amount of time needed for telemedicine visits (n = 63) was only 2 hours.
Other resource utilization: Emergency visits, procedure visits, telephone and patient portal communications
One person in the in-office group reported seeing a provider urgently during the study, and three participants in the telemedicine group reported an emergency scenario; two of those saw a practitioner in the office. Six patients were hospitalized (nine admissions total) for headache during their study participation; four were assigned to the telemedicine group, and two were in the in-office group. Two participants in the in-office group scheduled two non-procedure appointments each with the nurse practitioner (one of which was likely a scheduling error), and one telemedicine participant had an office visit with the nurse practitioner during the study. Six enrollees in the in-office group and seven in the telemedicine group returned to the office for onabotulinumtoxinA injections, accounting for 17 additional visits in the in-office group and 16 in the telemedicine group. Three participants had additional visits for nerve blocks or sphenopalatine ganglion blocks, including two in the in-office group (11 visits) and one in the telemedicine group (one visit).
Twelve participants in the in-office group and 17 in the telemedicine group used the patient portal for communication with the office. There were 96 patient portal communications in the in-office group (range 1–25 per person) and 300 in the telemedicine group (range 1–61 per person). Based on a two-sample t-test assuming unequal variances, the average number of communications was not significantly different between the two groups (p-value = 0.08). Six participants in the in-office group and five in the telemedicine group did not use the patient portal.
Change in diagnosis
Three participants with an initial diagnosis of chronic migraine were diagnosed with spontaneous intracranial hypotension during their study participation (one in-office, two telemedicine). Re-analysis of clinical outcomes after excluding these individuals did not change our conclusions.
Discussion
A case report successfully utilized telemedicine to facilitate communication between a headache medicine specialist and a patient presenting for care at a branch hospital (7). In contrast to that case report, patients participating in our telemedicine visits were not located in a clinical setting. We show that telemedicine (from a remote patient location) is a feasible mode of treatment for follow-up care of migraine, as demonstrated by 89 visits (out of 96 scheduled) successfully completed. An early study of internet-based self-help behavioral interventions for chronic headache showed improvement in participants randomly assigned to a 6-week active intervention compared to those placed on a symptom-monitoring waitlist (8). There was a significant reduction of mean headache activity, the number of weekly headache days, and mean weekly headache peak intensity in the CBT group compared to the waitlist controls. A later study added weekly telephone contact by a therapist, and the authors reported improvement in headache disability, depression, maladaptive coping strategies, and perceived stress measures (9). In our small study, both groups received intervention through a specialist, and the improvements in migraine-related disability (MIDAS), number of headache days, and average severity were similar across the groups.
The feasibility of telemedicine for diagnosing and treating non-acute headaches and medication-overuse headaches has been evaluated in northern Norway (10,11). In contrast to our study, those telemedicine participants were not located remotely from the neurologist (they were just in another office at the same university outpatient clinic). The primary outcome of the study was the incidence of a secondary headache within 12 months of the initial consultation visit. The main secondary outcome was a change in headache intensity using a 10-point visual analogue scale (VAS). At 3 and 12 months, there was no difference between HIT-6 or VAS scores between the two groups. There was an overall decrease in the HIT-6 score, from 64–65 to 60–61, which was maintained at month 12, and a decrease from 7 to 5.0–5.5 in the VAS score; neither change was statistically significant from baseline.
We found that in-office visits took significantly longer than telemedicine visits. An in-office visit requires additional steps prior to seeing the provider, including the check-in process, rooming, obtaining vital signs, and updating the medication list and allergies, which are performed by non-physician personnel. However, the face-to-face time with the physician was also significantly longer during office visits. While our finding is similar to that of the randomized study by Müller et al. (11), the patient location was fixed in their case whereas in our case the locations for telemedicine visits could be quite different. Since telemedicine visits may not involve much travel, the shorter consultation time may reflect an inherent difference in live interactions, the patient’s desire to use all of their allotted appointment time (30 minutes for a follow-up visit), or the physician’s subconscious effort to meet perceived patient expectations in general, particularly for those traveling long distances.
Throughout the study, it was readily apparent to the physician that telemedicine visits were much shorter than those conducted in the office. This increased efficiency for telemedicine visits corresponds to a similar conclusion reached through analytical modeling by Rajan et al. (12). This could increase productivity for physicians who are already a scarce resource and hence provide access to more patients.
Innovations in treatment can founder on patient resistance to change, so we gathered information about participants’ perceptions during the study. We found a much higher rate of patient satisfaction with visits conducted using telemedicine than Müller et al. (13) found. Despite the shorter consultation time for the telemedicine group, participants in both groups were equally satisfied with the effectiveness of care and with the provider. Participants also rated the time spent with the specialist as satisfactory, hinting that patients did not feel rushed or that the quality was compromised. Convenience may have offset shorter appointments, too, as several participants in the telemedicine group were able to complete their visits from their workplace and also during business-related travel.
Selection of participants’ open-ended responses.
Previous studies of internet-based interventions and telemedicine for treating various neurological (including headache) conditions support their utility in consultation and psychological therapies (5,7–15). Our study differs from previous studies in several respects. First, we used synchronous telemedicine visits with both audio and video for long-term migraine follow-up care. Participants assigned to visits by telemedicine communicated with the specialist from a remote location of their choice. We followed our patients using visits according to the randomized assignment throughout the duration of the study. Patients were randomized to receive five follow-up visits conducted by the same headache medicine specialist at the same intervals using face-to-face telemedicine technology. The initial outpatient consultation was performed in person (as was required by Texas state law) and included a neurological examination. Regardless of legal requirements, we favor this strategy when possible as the initial examination is important to exclude a secondary cause, and not all aspects of the exam (e.g. ophthalmoscopy) can be readily performed remotely. Our cohort included migraine patients with severe disability, three of whom were ultimately diagnosed with spontaneous intracranial hypotension. No other published study to our knowledge has investigated migraine management using modern telemedicine or has assessed long-term outcomes of telemedicine visits.
Limitations of our study include the number of uncompleted visits and the resulting lack of complete data. There may be bias in our cohort, as patients who were unwilling to be randomized to either treatment were excluded; however, only five patients declined participation during the recruitment phase. The quality of the telemedicine platform was inconsistent, and a few visits were ultimately performed by telephone because of audio or video failure. While participants in the study indicated willingness to pay for telemedicine, our university practice consists almost exclusively of individuals with commercial or national governmental insurance who may be less likely to consider the out-of-pocket expense to be a financial hardship than individuals in other circumstances (16).
Despite these limitations, our study provides evidence that telemedicine is a viable method of conducting follow-up visits for migraine patients with severe migraine-related disability, with similar outcomes at one year compared with in-office visits. Physician productivity was significantly higher, and participants randomized to telemedicine were highly satisfied with their care and the convenience afforded by the technology. Our results support the use of this modality and coverage by third-party payers, as telemedicine offers the opportunity for migraine patients who do not have ready access to a headache medicine specialist to receive care that is comparable to in-office treatment.
Clinical implications
Telemedicine is a feasible mode of treatment and an effective alternative to in-office visits for follow-up migraine care. Physician productivity could be higher with telemedicine. Patients may get better access because of its convenience.
Footnotes
Acknowledgements
The authors are grateful to Jennifer Fehmel, research coordinator, and Nancy Alexander-Moreno, RN, for their valuable contributions to the administration of this study.
Data sharing statement
The study protocol, statistical analysis plan, informed consent form, and de-identified data collected during the trial will be available by request to the corresponding author (DIF) for 36 months following article publication.
Declaration of conflicting interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: DIF serves on advisory boards for Alder Biopharmaceuticals, Allergan, Amgen, Biohaven Pharmaceuticals, electroCore, Eli Lilly, Promius, Supernus, Teva, Theranica and Zosano. She has received speaking honoraria for Allergan, Amgen, Autonomic Technologics Inc., electroCore, Revance, Supernus, Teva and Theranica. She has grant support from Eli Lilly, Autonomic Technologies Inc., Zosano, and Merck. She is a contributing author to Neurology Reviews and MedLink Neurology. BR reports no disclosures. AS reports no disclosures.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by Merck, Inc. The sponsor did not contribute to the concept, study design, collection of data, data analysis, conclusions, or writing of the manuscript.