A pilot investigation of a mobile phone application and progressive reminder system to improve adherence to daily prevention treatment in adolescents and young adults with migraine

Abstract

Background

Effective management of migraine requires adherence to treatment recommendations; however, adolescents with migraine take their daily medications only 75% of the time. Low-cost adherence-focused interventions using technology may improve adherence, but have not been investigated.

Methods

Thirty-five adolescents and young adults (13–21 years) with migraine participated in an AB-design pilot study to assess the use of a mobile phone adherence-promotion application (“app”) and progressive reminder system. Adherence was calculated using electronic monitoring during the baseline period and medication adherence intervention.

Results

Relative to baseline, adherence significantly improved during the first month of the intervention. Specifically, improvements existed for older participants with lower baseline adherence. Self-reported app-based adherence rates were significantly lower than electronically monitored adherence rates. Participants rated the intervention as acceptable and easy to use.

Conclusions

“Apps” have the potential to improve medication adherence and are a promising intervention for adolescents and young adults with low adherence. Involving parents in the intervention is also helpful. Providers should assess barriers to adherence and use of technology-based interventions, encourage parents to incorporate behavioral incentives, and provide referrals for more intensive interventions to improve long-term outcomes. Further, tracking adherence in an app may result in an underestimation of adherence. Future full-scale studies should be conducted to examine adherence promotion app interventions.

Keywords

Pediatric adherence intervention mHealth technology mobile phone headache

Introduction

Adherence to prescribed treatment recommendations for adolescents and young adults (AYAs) with migraine has the potential to decrease headache frequency, severity, and disability. A 2014 review of adherence in individuals with headache revealed that adherence rates ranged from 25 to 94% for adults (1), while data on rates of medication nonadherence in patients with headache under the age of 18 were nonexistent in the literature (1). More recent studies have demonstrated rates of medication adherence for adolescents with headache to range from 64% for daily self-report (2) to 75% in studies using electronic monitors (2) and 90% in randomized clinical trials (3).

Few studies have examined interventions aimed at improving treatment adherence for individuals with migraine. The only study examining an adherence-promotion intervention for adults with migraine found higher adherence rates in participants who received education on migraine symptoms, diagnosis, treatment, and prevention than a group of participants who did not receive the education (4). For adolescents, one study recommended that it may be preferential to prescribe a once-daily medication for migraine prevention, rather than one prescribed twice daily, in order to increase medication adherence (2). No studies have specifically examined the effects of a behavioral intervention aimed at improving adherence to headache treatment in AYAs. A recent meta-analysis examining psychological interventions to promote treatment adherence in pediatric and adult chronic health conditions demonstrated that adherence interventions are effective and that behavioral and/or multi-component interventions are particularly potent in improving adherence among youth (5). In addition, mobile health (i.e. mHealth) interventions are effective in eliciting meaningful improvements in pediatric health behavior and associated pediatric health outcomes (6).

Given the rates of nonadherence for adolescents and adults with headache, the prevalence and chronicity of this condition, and the complexity of headache treatment, the primary goal of this study was to test the feasibility, acceptability, and preliminary efficacy of a mobile application-based behavioral treatment aimed at improving adherence for adolescents with migraine. As 91% of AYAs have access to a mobile device (7), delivering adherence-promotion interventions to adolescents through a mobile platform may also improve user engagement and satisfaction. The specific adherence-promotion intervention examined in this study combined several evidence-based behavioral components, including daily medication reminders (8), medication monitoring (9,10), and caregiver involvement. Specifically, the objectives of the current study were to examine: a) the impact of a mobile application and progressive reminder system (PRS) on adherence to daily preventive medication in AYAs with migraine, b) predictors of change in adherence, c) the relationship between electronically monitored adherence with self-reported medication adherence through a mobile application and phone calls, and d) acceptability (i.e. user satisfaction and usability) of the mobile phone application and PRS.

Methods

Participants

This pilot study was conducted at a pediatric headache clinic in the USA between March and July 2016. Participants were recruited during routine outpatient visits. Inclusion criteria included: a) Participant age 13–21 years; b) diagnosis of migraine according to current International Classification of Headache Disorders criteria (ICHD-3 beta) (11) for migraine with or without aura and chronic migraine; c) current prescription for migraine prevention medication in pill format; d) mobile wireless internet via a data plan or wi-fi access; and e) English fluency for participant and caregiver. Participants were excluded from the study if they met any of the following criteria: Continuous migraine (no pain-free periods for an entire month; n = 0), developmental delay or impairment (n = 1), or serious mental illness (n = 4). Patients' medical charts were screened for exclusionary criteria, with additional patients eliminated after physician input.

Study design and procedures

The study followed a single-center AB (baseline and intervention) design to assess adherence using the mobile phone application (“app”) and PRS. The study was approved by the hospital's institutional research ethics board. Adult ( ≥ 18 years old) and adolescent participants' caregivers provided informed consent. Adolescents provided assent.

Data collection for participants occurred in two phases throughout a 16-week period. The first 8-week phase served as an initial “run-in” period, during which only electronic monitoring adherence data were collected to establish a baseline adherence rate. At the conclusion of the run-in, participants were oriented to the app and PRS by study staff. During Phase 2 (8 weeks), participants utilized the app and PRS and electronic medication monitors.

Intervention: The intervention was developed based on the Health Belief Model's principle of cued actions (12), which suggests that external cues, including reminders and prompts, may be needed to increase adherence behaviors. Indeed, mobile reminders have been used to promote successful adherence to behavior change in multiple pediatric health populations (13,14), although none have focused on AYAs with headache. The PRS component of the intervention was also informed by Family Systems Theory (15).

Participants downloaded the MedaCheck app (www.medacheck.com) on personal mobile devices and entered their medication dose schedules into the app. Participants without smartphones were provided with an iPod Touch. Figure 1 displays the visual and audio notifications provided to prompt participants to take medication and record medication administration into the app. If an adolescent did not endorse taking the medication dose (i.e. ignored the prompt or selected “skip”), the PRS was activated one hour later and the participant received a reminder phone call from a call center. The call center was staffed by customer service representatives who were not specifically trained in behavioral change techniques. In the case that the phone call was not answered or the adolescent denied taking medication, a phone call was immediately placed to the caregiver to inform them of a missed dose and request parental assistance to ensure that the dose was taken. Participants and caregivers had the opportunity during the reminder phone calls to report that the participant had successfully taken medication; this information was uploaded into the system.

Figure 1.

Screenshots from MedaCheck phone application.

Measurement

Demographics and medical characteristics: Background information was collected from the caregiver using a demographics form. Self-reported family income was collected as a proxy of socioeconomic status. Disease information such as headache type and prescribed medication were obtained through medical chart review. Headaches were classified according to current criteria (ICHD-3 beta) (11).

Pediatric migraine disability: Pediatric Migraine Disability Assessment (16) (PedMIDAS) is a six-item instrument that assesses the impact of headaches on functioning in school, home, and social environments over the past three months. A total score is created by summing responses on all six items. PedMIDAS uses a grading scale of none (I; 0–10), mild (II; 11–30), moderate (III; 31–50) and severe (IV; > 50) and has shown excellent internal consistency and test-retest reliability (16). The total score was used in this study.

Electronically monitored adherence: The primary measure of efficacy was medication adherence to preventive medication using Medication Event Monitoring Systems Trackcap6® (MEMS; Aardex Group, Sion, Switzerland; www.medamigo.com). This system includes a pill bottle and cap with a microprocessor that records each instance in which the cap is removed. This technology has been successfully employed in adolescents and young adults with migraine (2). Participants were informed that the purpose of the MEMS cap was to track medication adherence by recording bottle openings. They were also provided with a pill bottle label to record medication refills in which medication was not simultaneously taken. Data were downloaded upon return of the MEMS bottle. Adjustments were made to account for incidental openings reported on the label. Daily adherence was calculated as the number of cap openings divided by the expected number of doses prescribed. Values were multiplied by 100 to create percentages. Weekly adherence rates were calculated by averaging daily adherence during the week. Separate adherence rates were calculated for the last half of the run-in period, the first four weeks of the intervention, the last four weeks of the intervention, and the entire 8-week intervention.

Self-reported adherence: Self-report of adherence during the intervention was measured through the MedaCheck app and PRS system. Participants were prompted to answer one-item assessing medication adherence through the app, which was uploaded to a secure and HIPPA-compliant cloud server, and immediately available to study personnel through a web-based portal. If participants or caregivers received a reminder phone call, they were also prompted to report whether medication had been taken. Self-reported app-based adherence rates were computed using the data from participants' responses to the mobile prompt for taking medication; self-reported PRS adherence was computed by combining adherence reports entered into the app and recorded by call center. These values were also divided by the number of expected daily doses and multiplied by 100. Participants who did not respond to the app or PRS prompts were considered nonadherent.

Intervention feasibility and usability: Participant and caregiver quantitative and qualitative ratings for the feasibility and usability of the app and PRS were assessed by a brief health technology questionnaire administered at the conclusion of the study. All items were anchored by a 5-point (0-4) Likert-type scale.

Statistical analysis

The intention of this feasibility pilot study was to obtain effect size data to power a larger randomized clinical trial. A power analysis was conducted according to the primary repeated measures analysis. Given power of .80 with α = .05 and f = .25 (i.e. medium effect size), a total sample of 28 was needed. To account for estimated 30% study attrition, we aimed for a sample size of 40.

Initial analysis of the 8-week run-in phase revealed evidence of reactivity, suggesting significant within-subject changes (i.e. improved adherence) in the absence of an intervention (13). This can occur when participants are not fully blind to the target of the intervention. Recommendations for the length of the run-in phase tend to be approximately 4–8 weeks (17,18), with many adherence studies utilizing a 4-week run-in (19,20). To minimize the likelihood of bias, only the last 4 weeks of the run-in phase were included in pre-post intervention analyses.

Frequencies and descriptive statistics were run to assess intervention improvements and describe implementation of the PRS. To determine the efficacy of the intervention compared to the run-in, a repeated measures generalized linear model with planned contrasts was analyzed (Objective 1). The within-subjects variable (i.e. time) had three levels: 4-week run-in, intervention weeks1–4, and intervention weeks 5–8. The intervention phase was divided into two 4-week intervals to better understand the efficacy of the intervention over time. To determine whether the intervention was more efficacious for participants who were more or less adherent, a two-level between-subjects variable (i.e. baseline adherence group) was included in the analysis. Adherence group was determined by median split (i.e. 88% or greater MEMS adherence was classified as more adherent). Post-hoc paired t-tests were planned should the analysis reveal a significant time-by-adherence group interaction effect. Eta-squared values were calculated as an indicator of effect size (small = .01, medium = .06, large = .14) (21).

Linear regression examined predictors of changes in MEMS adherence over time (Objective 2). Paired t-tests were used to compare adherence rates by method (Objective 3). Cohen's d was calculated (with dependent observations corrected) as a measure of effect size, with d = .20, .50, and .80 indicating small, medium, and large effects, respectively (21,22). Finally, frequencies and descriptive statistics were run to quantify participant usability and feasibility ratings (Objective 4). All analyses were run using SPSS Statistics version 23.0. Statistical significance was determined by p < .05.

Results

Demographic and baseline characteristics

Of the 55 patients eligible to participate, 40 patients and caregivers provided consent to participate in the study; 15 patients declined to participate (e.g. “already have a successful method for tracking adherence”, “family is too busy”). Five participants were withdrawn from the study due to inability to make further contact (n = 4) and stopping preventative headache medication during the run-in period (n = 1). The total sample of 35 participants was mostly female (77%) and white (89%), with a mean age of 15.37 years (see Table 1). Three participants subsequently weaned or discontinued their medication during the intervention (i.e. one in week 12 and two in week 16); one of these participants was not included in analyses assessing the last half of the intervention due to incomplete data.

Table 1.

Participant characteristics.

	Total sample (n = 35)		iOS sample (n = 29)		Android sample (n = 6)
Characteristic	M	SD	M	SD	M	SD	t
Child age (years)	15.4	1.9	15.1	1.6	16.8	2.7	−2.14*
PedMIDAS Total Score	22.4	28.3	25.8	30.3	7.7	8.1	1.44
	n	%	n	%	n	%	Χ ²
Child sex
Male	8	22.9	6	20.7	2	33.3
Female	27	77.1	23	79.3	4	66.7	.45
Child race
White	31	88.5	27	93.1	4	66.7
Black	2	5.7	1	3.4	1	16.7
Hispanic	1	2.9	1	3.4	0	0.0
Asian	1	2.9	0	0.0	1	16.7	3.43^†
Child headache diagnosis
Intractable migraine without aura	23	65.7	19	65.5	4	66.7
Intractable migraine with aura	5	14.2	3	10.3	2	33.3
Intractable migraine with and w/o aura	3	8.5	3	10.3	0	0.0
Refractory migraine without aura	4	11.4	4	13.8	0	0.0	3.29
Child anti-headache drug therapy
Amitriptyline	11	31.4	8	27.6	2	33.3
Valproic acid	3	8.5	2	6.9	1	16.7
Levetiracetam	4	11.4	3	10.3	1	16.7
Topiramate	17	48.5	16	55.2	2	33.3	1.25
Mobile device
Android	6	17	–	–	–	–
iPhone	24	69	–	–	–	–
iPod Touch	5	14	–	–	–	–
Family income
0–25,000	2	5.7	1	3.4	1	6.7
25,001–50,000	5	14.3	4	13.8	1	16.7
50,001–75,000	2	5.7	1	3.4	2	33.3
75,001–100,000	4	11.4	4	13.8	0	0.0
100,001–125,000	6	17.1	5	17.2	1	16.7
125,001–150,000	4	11.4	4	13.8	0	0.0
150,000+	7	19.9	7	26.9	0	0.0
Not reported	5	14.3	3	10.3	1	16.7	2.92^‡

Note: * indicates p < .05. ^†race analyzed as white vs. non-white. ^‡income analyzed as ≤ $100,000 vs. > ≤ $100,000.

Unexpectedly, there was a technical issue with the Android devices (n = 6) and those data were excluded from the PRS analyses. There were no significant between-group differences at baseline between Android and iOS users on gender (χ²(1) = .45, p = .50); race (χ²(1) = 3.43, p = .06); income (χ²(1) = 2.92, p = .09); headache type (χ²(3) = 3.29, p = .35); medication type (χ²(3) = 1.25, p = .74); dosing schedule (χ²(1) = .08, p = .78), or PedMIDAS total score (t(30) = 1.44, p = .16); however, Android users were more likely to be older (t(30) = −2.14, p = .041. Android users were removed from analyses pertaining to PRS usage, combined app and PRS adherence rates, and usability.

Baseline adherence: Adherence assessed using MEMS was relatively high in this sample of adolescents during the 4-week run-in, with a mean of 81.9% and median of 87.0%. At the 25th and 75th percentiles, participant adherence rates were 76.7% and 94.4%, respectively.

Activation of PRS: For the iOS users (n = 29), adolescents received a medication reminder call due to not endorsing a scheduled dose for 33% (SD = 22.23) of their required doses. Caregivers were contacted for 15% (SD = 12.31) of their child's medication dosages. In other words, when a participant did not enter his or her medication in the app, the escalation to a caregiver call occurred approximately 46% of the time.

Primary analyses

Examining effect of time across 4-week intervals: A repeated measures generalized linear model with planned contrasts for the total sample revealed a significant large main effect of time, F(2,64) = 7.17, p = .002, η² = .17 (Figure 2). Planned contrasts demonstrated a significant increase in MEMS adherence from the run-in period (M = 81.52) to the first four weeks of the intervention (M = 87.45), F(1,32) = 5.42, p = .026. However, there was no significant difference between the run-in period and the last four weeks of the intervention (M = 78.69), F(1,32) = 1.38, p = .25.

Figure 2.

Adherence percentages during run-in and intervention for adolescents with “optimal” and “less than optimal”.

The time X baseline adherence (i.e. low vs. high) interaction revealed a medium-to-large significant effect (F(2,64) = 4.43, p = .016, η² = .10), suggesting patterns in MEMS adherence pre- and post-intervention differed depending on participants' baseline adherence rates. To investigate the interaction, post-hoc paired t-tests were conducted. Results demonstrated large improvements in adherence from the run-in (M = 70.68) to the first four weeks of the intervention (M = 83.01), (t(17) = −3.05, p = .007, Cohen's d = −.73), for participants with lower baseline adherence; again, no significant differences between the run-in and the last four weeks (M = 70.96) of the intervention were found (Cohen's d = −.02). In contrast, MEMS adherence for participants with high baseline adherence did not increase from run-in (M = 93.72) to the first half of the intervention (M = 92.43), (t(15) = .60, p = .56, Cohen's d = .17). However, a significant, large decrease in adherence rates was observed when the last half of the intervention (M = 87.39) was compared to the run-in (t(15) = 2.40, p = .030, Cohen's d = .75).

Secondary analyses

Predictors of change in adherence from run-in to post-intervention: Regression results revealed older participants were more likely to demonstrate improved MEMS adherence from run-in to post-intervention implementation (B = 2.30, β = .40, t = 2.15, p = .041). Neither PedMIDAS Total score, medication dosing regimen, adolescent calls per dose, nor caregiver calls per dose significantly predicted pre-post intervention changes in MEMS adherence.

Comparing adherence rates by method (iOS users only; Table 2): App-only adherence rates (M = 60.42) for the duration of the eight-week intervention period were significantly lower than MEMS adherence rates (M = 84.07; t(25) = −6.00, p < .001, Cohen's d = −1.19) (Figure 3). The PRS adherence rate (M = 78.71) was not significantly different from the MEMS rates (t(25) = 1.49, p = .15, Cohen's d = .29); however, PRS adherence was higher than app-only reports of adherence throughout the duration of the intervention (t(25) = −6.10, p < .001, Cohen's d = −1.14 (Figure 3).

Table 2.

Adherence rates by assessment method.

	MEMS			App-only report			PRS (phone + app) report
	N	M	SD	N	M	SD	N	M	SD
Total sample
Run-in	35	81.94	16.46	–	–	–	–	–	–
Post-intervention	35	84.00	14.35	29^a	60.41	22.89	35	75.86	23.80
Low adherence (<88%)
Run-in	18	70.68	15.78	–	–	–	–	–	–
Post-intervention	18	76.25	15.58	14^a	52.07	22.24	18	75.23	20.81
High adherence (≥88%)
Run-in	17	93.85	4.18	–	–	–	–	–	–
Post-intervention	17	92.21	6.35	15^a	68.19	21.31	17	76.52	27.26

Note: ^aAndroid users were not included in analyses.

Figure 3.

Adherence percentages during the intervention by report type.

Acceptability and usability

Over 80% of the participants gave the app high ratings for ease and convenience of use, app satisfaction, and appearance (Table 3). Over half of adolescent participants found the phone calls to themselves or to their caregiver to be helpful; while 70–74% of the caregivers found the phone calls to be helpful. The amount of phone calls received by each participant varied depending on how often they tracked medication adherence using the app.

Table 3.

Adolescent and parent's perception of usability and acceptability.

	M	High rating (%)	SD	Range
Adolescent (n = 28)
How easy was it for you to understand how to use the app?	3.79	96	.50	2–4
How much did you feel you could trust the features of the app?	3.46	86	.84	1–4
How satisfied were you with the app?	3.36	86	.83	1–4
How much did you like the way the app looked?	3.30	85	.72	2–4
How much did the app keep your interest and attention?	3.02	79	.87	1–4
How convenient was the app to use?	3.34	86	.86	1–4
Has tracking your medication intake been helpful?	3.43	86	1.03	0–4
How good were the app reminders at helping you remember to take your meds on time?	3.00	82	.82	1–4
How useful did you find the phone call reminder to your parent?	2.57	57	1.43	0–4
How useful did you find the phone call feature?	2.57	61	1.50	0–4
How good of a fit was the progressive reminder system for your family?	3.18	82	.98	1–4
If headache difficulties return, how likely would you be to come back to the app to track your med adherence?	3.00	71	1.01	1–4
Caregiver (n = 24)
How helpful were the push notifications for helping your child take his/her medication?	2.98	75	1.07	1–4
How helpful were the phone calls for helping your child to take his/her medication?	2.76	74	1.58	0–4
How helpful were the parent phone calls for helping your child to take his/her medication?	2.93	70	1.25	0–4

Note: Android users not included; 0–4 scale for all items; High rating indicates a score of “mostly” or “very helpful.”

Discussion

The current pilot study suggests that an adherence promotion mobile application and PRS are feasible, acceptable, and a promising intervention for AYAs with headache. Nearly three-quarters of the patients approached to participate were interested in participating in the adherence intervention, suggesting that AYAs with headache and their families may be interested in technology-based strategies for improving medication adherence. Although technical difficulties arose with participants who were not iOS users, the majority of iOS participants and their caregivers expressed satisfaction with the mobile app and found it easy to use.

With regards to preliminary efficacy, results demonstrated that adherence to daily preventive medications may initially be improved for adolescents and young adults (AYAs) with migraine using a low-cost technology-based adherence-promotion intervention. Nearly half of the participants demonstrated some level of improvement in adherence as a result of the app, with one adolescent demonstrating a 36% increase in adherence during the initial treatment period. Results also highlight the importance of providing adherence interventions for patients with less than optimal adherence, as participants with lower adherence appeared to benefit most. Further, large improvements in adherence for the participants with low baseline adherence occurred within the initial month of the app intervention. Results provide preliminary support for the use of adherence interventions for patients with less than optimal adherence, as participants with lower baseline adherence improved most from pre- to post-intervention Thus, clinicians and health care providers may wish to consider recommending medication reminder apps when quick improvements are desired. However, it should be noted that regression to the mean for those patients with lower baseline adherence cannot be ruled out and that the effects of the intervention did not last the full duration of the study (i.e. pre-post results yielded minimal effects). Nevertheless, it is a straight-forward solution that medical providers can quickly provide to adolescents with headache to boost adherence until additional strategies or solutions are deemed necessary.

To sustain engagement in this intervention modality, providers may consider checking in with participants to discuss barriers, challenges, and concerns with both the app and medication administration. Understanding these barriers will allow medical providers to recommend the use of specific behavioral interventions to improve engagement and the effects of the intervention over time. While speculative, it is possible that an increase in preventive medication adherence may reduce headache severity, frequency, and associated disability. However, it must be acknowledged that a recent study found no significant differences in the reduction of headache days experienced by children regardless of whether they received amitriptyline, topiramate, or placebo (23). While a reexamination of preventive medication for pediatric migraine may be necessary, topiramate remains an FDA-approved treatment for episodic migraine in adolescents. Adherence to medications (and pill-taking in general) continues to be imperative for current disease management practice; therefore, efforts to improve such adherence are worthy of continued investigation.

Although the cause of the improvements in adherence cannot be determined, it is possible that the PRS utilized in this study was responsible in part for these improvements. Specifically, adolescents required a phone call for one-third of their doses and the PRS was escalated to a caregiver phone call approximately half of the time following the adolescent call. This system was specifically selected for patients in this developmental stage due to the desire to have increased independence from their caregivers. The PRS gives adolescent patients the opportunity to be independent in taking their medication, but also allows for parental oversight in circumstances where the adolescent does not follow through. Our results suggest that phone calls, which provide both additional reminders and accountability, are important for improving adherence. Consistent with studies demonstrating increased effectiveness of adherence apps with caregiver involvement (6), health care providers should continue to work with families to include caregivers in both adherence assessment and intervention, either as a primary member of the adherence plan or through peripheral oversight for missed doses.

Although the app and PRS demonstrated improvement in adherence for AYAs with less than optimal adherence, there was a subset of participants for whom use of the app did not improve adherence. These adolescents may require more intensive multi-component interventions including a combination of behavioral, organizational, educational, health care system, technology, and social support components. Indeed, multi-component adherence-promotion interventions produce higher mean adherence effects than solely educational/instructional interventions (24). These multi-component interventions could be recommended to those adolescents who experience an initial increase in adherence with the use of an app but then begin to decline once the app becomes less novel.

In addition to the potential benefits of the app and PRS intervention to AYAs with low adherence, this intervention may be more beneficial to older AYAs. It is possible that older participants are developmentally ready to take more responsibility for their health care needs (25). In addition, they may use their phones with greater frequency (e.g. due to fewer parental or school restrictions, social norms of phone use) than younger AYAs. When recommending use of apps and PRS interventions, providers are encouraged to consider the patient's age and developmental readiness.

We were surprised to find that self-reported adherence rates in the application were an underrepresentation of electronically monitored adherence. This is contrary to most studies of adherence, which have shown self-reported ratings of adherence to be higher than rates measured by electronically monitoring (26,27); however, one study of adolescents with headache revealed a similar result (2). Given our findings, health care providers working with AYAs with migraine should exercise caution when assessing adherence based on self-reported values generated by apps alone because these adherence rates may be lower than actual adherence. It is possible that AYAs use the reminder from the application to take their medication but do not take the time to manually indicate within the app that they have taken it.

Results of the study should be considered within the context of several limitations. First, this was a pilot study with a sample size that limited our ability to detect significant findings. Future studies should examine the use of promising medication adherence applications in randomized controlled trials with larger samples. Second, the study sample was predominantly white and affluent, which limits the generalizability of this study. Third, although Android users were omitted from analyses examining the PRS, we acknowledge that technical difficulties encountered by this group may have influenced the results and reduce the ability to generalize findings to AYAs who do not have iOS devices. Finally, this study was conducted at a single site; future studies would benefit from testing this medication adherence app at multiple locations.

Clinical implications

Results of this study suggest several important clinical implications. First, adherence is less than optimal for most AYAs with migraine, indicating the importance of assessing adherence at medical appointments. Second, this study suggests that use of a medication adherence app and PRS, especially for older AYAs or patients with lower adherence, may have promise for improving short-term adherence. Third, it is important for providers to consider that some patients will need more intensive intervention than can be provided by an app, and recommendations such as a multi-component interventions including educational, organizational, and behavioral components should be considered. Fourth, involving caregivers in adherence interventions or establishing caregiver accountability for missed medication may improve adherence in adolescents with migraine. Finally, health care providers should be aware that adherence rates that rely on self-report information entered into a mobile app may actually be an underestimation of adherence.

Clinical implications

Adherence to treatment recommendations, including pill-taking recommendations for prevention medications, is less than optimal for most adolescents and young adults with migraine and it is important for health care providers to assess adherence.

Medication adherence applications and PRS, especially for adolescents and young adults with lower adherence and older age, may be promising strategies for improving adherence in the short term.

Some patients will need more intensive intervention than can be provided in an application and adherence intervention recommendations such as a multi-component interventions including educations, organizations, and behavioral components should be considered.

Involving caregivers in medication adherence procedures and interventions may improve adherence in adolescents with migraine.

Healthcare providers should be aware that adherence rates that rely on self-report information entered into a mobile app may be an underestimation of the adolescent's adherence.

Footnotes

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: Jeffrey Shepard, is CEO of MedaCheck, LLC, which programmed the adherence mobile application used in this study. Dr. Shepard did not have a role in data management or analysis in this study. Other authors have nothing to disclose.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by a training grant from the National Institutes of Health (T32HD068223) to the primary and corresponding author (RRR).

References

Ramsey

Ryan

Hershey

et al.

Treatment adherence in patients with headache: A systematic review. Headache 2014; 54: 795–816.

Kroon Van Diest

Ramsey

Aylward

et al.

Adherence to biobehavioral recommendations in pediatric migraine as measured by electronic monitoring: The Adherence in Migraine (AIM) Study. Headache 2016; 56: 1137–1146.

Kroon Van Diest

Ramsey

Kashikar-Zuck

et al.

Treatment adherence in child and adolescent chronic migraine patients: Results from the Cognitive Behavioral Therapy and Amitriptyline Trial. Clin J Pain 2017; 33: 892–898.

Rothrock

Parada

Sims

et al.

The impact of intensive patient education on clinical outcome in a clinic-based migraine population. Headache 2006; 46: 726–731.

Pai

McGrady

. Systematic review and meta-analysis of psychological interventions to promote treatment adherence in children, adolescents, and young adults with chronic illness. J Pediatr Psychol 2014; 39: 918–931.

Fedele

Cushing

Fritz

et al.

Mobile health interventions for improving health outcomes in youth: A meta-analysis. JAMA Pediatrics 2017; 171: 461–469.

Lenhart A. Teen, social media and technology overview 2015. Washington, DC: Pew Research Center, 2015.

Dayer

Heldenbrand

Anderson

et al.

Smartphone medication adherence apps: Potential benefits to patients and providers. J Am Pharm Assoc 2013; 53: 172–181.

Spaulding

Devine

Duncan

et al.

Electronic monitoring and feedback to improve adherence in pediatric asthma. J Pediatr Psychol 2012; 37: 64–74.

10.

Chan AH, Stewart AW, Harrison J, et al. The effect of an electronic monitoring device with audiovisual reminder function on adherence to inhaled corticosteroids and school attendance in children with asthma: a randomised controlled trial. Lancet Respir Med 2015; 3: 210–219.

11.

Headache Classification Committee of the International Headache Society. The International Classification of Headache Disorders, 3rd edition (beta version). Cephalalgia 2013; 33: 629–808.

12.

Strecher

Rosenstock

The health belief model. In: Glanz

Lewis

Rimer

(eds). Health behaviour and health education: Theory, research, and practice, 2nd edn. San Francisco: Jossey-Bass, 1997, pp. 41–59.

13.

Garofalo

Kuhns

Hotton

et al.

A randomized controlled trial of personalized text message reminders to promote medication adherence among HIV-positive adolescents and young adults. AIDS Behav 2016; 20: 1049–1059.

14.

Hommel

. Using technology to assess and promote adherence to medical regimens in pediatric chronic illness. J Pediatr 2014; 164: 922–927.

15.

Bowen

. Family therapy in clinical practice, Lanham, MD: Jason Aronson, 1993.

16.

Hershey

Powers

Vockell

et al.

PedMIDAS: Development of a questionnaire to assess disability of migraines in children. Neurology 2001; 57: 2034–2039.

17.

Cook

Schmiege

McClean

et al.

Practical and analytic issues in the electronic assessment of adherence. Western J Nurs Res 2012; 34: 598–620.

18.

Williams

Amico

Bova

et al.

A proposal for quality standards for measuring medication adherence in research. AIDS Behav 2013; 17: 284–297.

19.

Rapoff

Belmont

Lindsley

et al.

Prevention of nonadherence to nonsteroidal anti-inflammatory medications for newly diagnosed patients with juvenile rheumatoid arthritis. Health Psychol 2002; 21: 620–623.

20.

Modi

Guilfoyle

Rausch

. Preliminary feasibility, acceptability, and efficacy of an innovative adherence intervention for children with newly diagnosed epilepsy. J Pediatr Psychol 2013; 38: 605–616.

21.

Cohen

. Statistical power analysis for the behavioral sciences, 2nd edn. Hillsdale, NJ: Lawrence Erlbaum Associates, 1988.

22.

Cohen

. A power primer. Psychol Bulletin 1992; 112: 155–159.

23.

Powers

Coffey

Chamberlin

et al.

Trial of amitriptyline, topiramate, and placebo for pediatric migraine. New Eng J Med 2017; 376: 115–124.

24.

Kahana

Drotar

Frazier

. Meta-analysis of psychological interventions to promote adherence to treatment in pediatric chronic health conditions. J Pediatr Psychol 2008; 33: 590–611.

25.

Annunziato

Parkar

Dugan

et al.

Brief report: Deficits in health care management skills among adolescent and young adult liver transplant recipients transitioning to adult care settings. J Pediatr Psychol 2011; 36: 155–159.

26.

Pai

ALH

Gray

et al.

Development and reliability of a correction factor for family-reported medication adherence: Pediatric inflammatory bowel disease as an exemplar. J Pediatr Psychol 2013; 38: 893–901.

27.

Modi

Guilfoyle

Morita

et al.

Development and reliability of a correction factor for parent-reported adherence to pediatric antiepileptic drug therapy. Epilepsia 2011; 52: 370–376.