This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Aging, is properly cited. The complete bibliographic information, a link to the original publication on http://aging.jmir.org, as well as this copyright and license information must be included.
The debilitating effects of recurrent stroke among aging patients have urged researchers to explore medication adherence among these patients. Video narratives built upon Health Belief Model (HBM) constructs have displayed potential impact on medication adherence, adding an advantage to patient education efforts. However, its effect on medication understanding and use self-efficacy have not been tested.
The researchers believed that culturally sensitive video narratives, which catered to a specific niche, would reveal a personalized impact on medication adherence. Therefore, this study aimed to develop and validate video narratives for this purpose.
This study adapted the Delphi method to develop a consensus on the video scripts’ contents based on learning outcomes and HBM constructs. The panel of experts comprised 8 members representing professional stroke disease experts and experienced poststroke patients in Malaysia. The Delphi method involved 3 rounds of discussions. Once the consensus among members was achieved, the researchers drafted the initial scripts in English, which were then back translated to the Malay language. A total of 10 bilingual patients, within the study’s inclusion criteria, screened the scripts for comprehension. Subsequently, a neurologist and poststroke patient narrated the scripts in both languages as they were filmed, to add to the realism of the narratives. Then, the video narratives underwent a few cycles of editing after some feedback on video engagement by the bilingual patients. Few statistical analyses were applied to confirm the validity and reliability of the video narratives.
Initially, the researchers proposed 8 learning outcomes and 9 questions based on HBM constructs for the video scripts’ content. However, following Delphi rounds 1 to 3, a few statements were omitted and rephrased. The Kendall coefficient of concordance, W, was about 0.7 (
The Delphi method was proven to be helpful in conducting discussions systematically and providing precise content for the development of video narratives, whereas the Video Engagement Scale was an appropriate measurement of video realism and emotions, which the researchers believed could positively impact medication understanding and use self-efficacy among patients with stroke. A feasibility and acceptability study in an actual stroke care center is needed.
Australian New Zealand Clinical Trials Registry ACTRN12618000174280; https://www.anzctr.org.au /Trial/Registration/TrialReview.aspx?id=373554&isReview=true
Medication nonadherence is prevalent at large especially in major chronic diseases, despite patient education and advanced knowledge and methods [
In educational strategy, 75% of information is engaged visually and about 13% of it is engaged using our hearing senses [
The researchers proposed a patient education intervention at an outpatient stroke clinic as it may be a perfect venue for focused recurrent stroke education because it provides access to a common variation of people who are at high risk for recurrent stroke. In a quest for cost-effectiveness, the researchers utilized the prolonged waiting time in the clinic as an opportunity to deliver the intervention adjunct to the current medication therapy adherence clinic’s (MTAC) effort that may benefit the patients with stroke. Time spent in the waiting area is a potential period for patients to gain knowledge and confidence in managing their medication [
The researchers believed that video narratives shown simultaneously with patient education modules are expected to have a positive impact on self-efficacy. Consequently, if the video is incorporated with theoretical behavioral constructs, it could induce self-reflection and simulation by a role model. In addition, if the video is repetitively seen, a
This study hypothesized that providing video narratives incorporated with theoretical behavioral constructs adjunct to the existing MTAC’s patient education effort, informational brochures, counseling, and medication review would result in better stroke awareness, medication understanding, and use self-efficacy toward improved adherence. This study was the intervention development and validation phase of a randomized controlled trial (universal trial number: U1111-1201-3955) [
The Delphi method originated from RAND Corporation studies from the 1950s and aimed to develop a reliable technique to obtain consensus from experts. Since then, many researchers have applied this organized method for expert problem-solving issues. They have also developed systematic guidelines of the process and analysis of the Delphi Method [
The researchers in this study applied a Delphi method to obtain anonymous consensus on learning outcomes, Health Belief Model (HBM) constructs, and content of video scripts which took place from October 2017 to December 2017 among experts experienced in stroke patient education. The consensus procedure incorporated 3 rounds of questionnaires via email to finalize expert panelists’ viewpoints.
The process started with literature findings on the local need for stroke survivors. Most patients’ crucial need encompassed feelings of being independent to have a good quality of life, reducing the severity and preventing recurrent stroke [
Fundamentally, the design of the content was based on the most widely used framework, HBM [
The core of the Delphi method was the selection of a knowledgeable and experienced expert panel of members within the specific need of content development [
The expert panel team of 8 comprised 2 neurologists, 2 pharmacists, 2 medical educationists, and 2 patients who had experienced a stroke. The neurologists were selected based on their 10 to 12 years of professional experience of diagnosing and prescribing medications to patients with stroke. The pharmacists were also selected based on their 10 to 12 years of professional experience of reviewing and dispensing prescribed medications to patients with stroke at the hospital and community level. Whereas, the medical educationists, who were also knowledgeable in developing curricular pedagogy, contributed to the suitability of learning outcomes for stroke according to local context and sensitivity. Finally, the patients with stroke for about 5 years had experiences and an awareness of the need for emotional support to enhance self-efficacy.
There is no specific sample size recommendation for the Delphi method in this area of study as different disciplines and purpose of discussion often result in dissimilar response rates and time [
A fruitful discussion with the panel of experts led to the video narrative script development. The researchers developed the scripts in English and translated them into the Malay language with the help of a professional bilingual translator. Then, back translation was performed by another bilingual researcher who was not exposed to the initial scripts to verify the similarity of meanings. Both scripts (a neurologist’s and a patient’s version) addressed a brief summary of (1) the debilitating impact of stroke; (2) related risk factors of recurrent stroke, its prevention strategy, and benefit; (3) belief in self-confidence; and (4) real-life cues of successful recovery regardless of the severity of stroke. The Flesch-Kincaid reading level for the narrative scripts scored an average grade level of 6 [
The researchers believed that it was ideal and realistic to have actual actors (ie, neurologist and a patient who had experienced a stroke) to narrate the scripts. Meanwhile, the video was taken at the Arts and Social Sciences School, Monash University, Malaysia, with the help of a technical officer. They narrated each video script, both in English and Malay language within 2 min, and the manner of speech was according to communication principles [
The researchers drafted the initial narrative script content guide from literature findings, which comprised 8 learning outcomes and 9 HBM-related questions linked to individual perceptions, cues to action, the likelihood of action, and self-efficacy. The panel of experts was given options (ie, yes: to agree to accept or no: do not agree to accept) and an open-ended question to add any other relevant information to the list or justify any redundancy. Hence, this round helped to establish the initial content and construct development of the list, clarification of meaning, and rephrasing or merging of a redundant statement. They were given 3 weeks’ time to respond to the Delphi method coordinator.
It was accepted that, approximately, an 80% agreement from the panel (ie, 6 or 7 out of 8 experts) for response frequencies for each learning outcome and HBM question was to be accepted or omitted. This percentage cut off was an appropriate reference point to attain content and construct validity [
The researchers repeated the same procedures and timeline as the previous discussion except that the panel of experts was asked to rank the level of relevance using a 7-point Likert scale (ie, 1: not at all relevant and 7: extremely relevant). They were asked to justify their choice of rank if it was 4 points and lower. Kendall W coefficient of concordance was used to measure the nonparametric rankings [
The coordinator received comments and feedback to rephrase a few statements to illustrate appropriate meanings. The coordinator asked the expert panelists if they were willing to continue the rounds until the W value rises and all agreed. Hence, a final edition of learning outcomes and HBM questions were resent via an online survey questionnaire for the Delphi method round 3 discussion.
Round 2 discussion and analysis produced a summary of responses and clarification from the panel of experts, which gave an overall picture of final scoring and the current level of consensus of the
A purposeful sample of 10 bilingual patients with stroke (within the inclusion and exclusion criteria of the trial) were requested to provide written feedback on the comprehension of the English and Malay video narratives scripts. The informed and consented patients were asked to reply either via email or via a prepaid postal service. Their responses contributed to face and content validity. They also viewed the video narratives in both languages and responded to the Video Engagement Scale (VES) that was presented to them face-to-face during their follow-up clinic visit. Test-retest was not appropriate as these patients were exposed to patient education materials, which could affect their follow-up responses. We expected occurrences of revision in every round of iteration. Therefore test-retest was not applicable to the Delphi method.
To the researchers’ knowledge, there were no fixed guidelines to validate a video narrative for patient education; however, there has been a link between the construct of engagement and persuasive communication [
Approvals for this development and validation study have been obtained from the Malaysian Medical Research and Ethics Committee (NMRR ID-15-851-24737) and the Monash University Human Research Ethics Committee (ID 9640) whereas the MyStrokeStory trial was registered with the Australian New Zealand Clinical Trials Registry (ACTRN12618000174280; universal trial number U1111-1201-3955).
The researchers made no addition to the initial draft of the learning outcomes and HBM questions before the Delphi method round 1. We omitted statements that were redundant, had less than 80% agreement (ie, What is a stroke? How serious is having a stroke?), or were rephrased (ie, How common is a stroke? to Who is at high risk of stroke?). Whereas, few other statements or questions had only a minor correction. Therefore, 8 learning outcomes and 9 HBM questions were edited to 6 statements with 6 questions each for the Delphi method round 2.
In round 2, the W value was below 0.7. The mean ranking for learning outcomes and HBM questions also varied (ie, 2 experts were asked to justify their low score for learning outcomes and HBM construct questions 1 and 2).
However, in round 3, the list of learning outcomes and HBM questions was finalized (
The researchers received positive feedback on the scripts (ie, good script, short and meaningful, and direct points), but there were not many comments on the structure of sentences or usage of words. Therefore, the researchers concluded that the scripts were suitable to the local context; hence, the narrative scripts were finally confirmed.
The finalized video narrative scripts’ learning outcomes and questions parallel with the Health Belief Model constructs.
Health Belief Model constructs | Learning outcomes | Questions |
Individual perception: Perceived susceptibility; Perceived severity | 1. To be able to recognize and understand stroke cause, symptoms, and effects |
1. What happens to you when you have a stroke? |
Likelihood of action: Perceived benefit; Perceived barrier | 3. To understand lifestyle risk factors of stroke |
3. How do you prevent another stroke? |
Self-efficacy | 5. To understand and acquire skills of medication understanding and use self-efficacy after a stroke | 5. How do you ensure your medication works for you? |
Final analysis of the Delphi method (n=8).
Raters | 10 items, meana,b,c |
Member 1 | 4.4 |
Member 2 | 6.2 |
Member 3 | 4.6 |
Member 4 | 4.4 |
Member 5 | 5.5 |
Member 6 | 3.4 |
Member 7 | 5.4 |
Member 8 | 5.4 |
aCronbach alpha: .908.
bIntraclass correlation coefficient (95% CI): 0.733 (0.384-0.919).
c
The Video Engagement Scale scores (n=10).
Raters | 15 items, meana,b,c |
Patient 1 | 5.3333 |
Patient 2 | 5.6000 |
Patient 3 | 5.7333 |
Patient 4 | 5.9333 |
Patient 5 | 6.4000 |
Patient 6 | 6.5333 |
Patient 7 | 6.7333 |
Patient 8 | 6.8667 |
Patient 9 | 6.8667 |
Patient 10 | 6.6667 |
aCronbach alpha: .925.
bIntraclass correlation coefficient (95% CI): 0.797 (0.572-0.921).
c
The VES scores were above average, which exhibited a good link with perceived realism (
This study explicitly developed and validated video narratives to be used as intervention materials in a randomized controlled trial [
There were some apparent limitations in this video narrative development. Face-to-face discussion was unable to be carried out in the Delphi method rounds owing to the lack of interval time and slow responses from the expert panel despite constant reminders. Hence, the Delphi method discussion ended in round 3 whereby force agreement would have occurred. The researchers were also aware that face validity and video engagement responses lacked the required number of participation from poststroke patients because of specific inclusion and exclusion criteria via purposive sampling method. Therefore, the video narratives’ validation and study aim were skewed toward particular samples only, and hence, results could not be generalized to the whole population of patients with stroke. In addition, responses from nonbilingual patients were also not assessed owing to the delay during the purposive sampling period and having the VES available in the English version only.
Nevertheless, the Delphi method proved to be a versatile and helpful technique in conducting discussions systematically and reaching a consensus unanimously, eliciting precise ideas, and providing rich, in-depth data in defining an intervention strategy. In addition, the video narrative development processes were found to be useful as a guideline for other behavioral studies, which use video as their intervention, samples with chronic illness, and study sites other than health care centers.
The researchers believed that
Health Belief Model
intraclass correlation coefficient
medication therapy adherence clinic
Video Engagement Scale
The authors would like to acknowledge the Jeffrey Cheah School of Medicine and Health Sciences, Monash University, Malaysia, for their financial, material, and other facilities support. The authors also wish to acknowledge the contributions of Saw Pui San, Rohit Verma, Allah Bukhsh, and Amutha Selvaraj as members of the panel experts. We wish to extend our thanks to the staffs and patients from the Neurology Clinic, Hospital Kuala Lumpur who had helped in the video development. The authors are also grateful to the Director General of Health, Malaysia, for his permission to publish this paper.
None declared.