The Utility and Acceptability of a Self-Help Smartphone Application for Eating Disorder Behaviors

Abstract

Objective

Smartphone applications (apps) for eating disorders are a promising approach to assist individuals who do not receive traditional treatment. This study examines usage characteristics, perceptions regarding the acceptability of a new self-help intervention developed for users with eating disorders, and associations between attitudes and use patterns.

Methods

189 individuals pilot-tested a personalized app-based program, and 133 completed the required components of the pilot-test over an 8-day period. Of these, 64 individuals (49%) completed an exit survey pertaining to acceptability.

Results

Seventy percent of those who pilot-tested the app-based program completed the required components, i.e. a baseline review and then a 1-week assessment. Body mass index was associated with the total number of recorded meal logs. Study participants rated the app as highly suitable and acceptable, providing evidence of the feasibility and appropriateness of the program.

Conclusion

The app-based program demonstrated feasibility of deploying the app across user groups and high acceptability.

Introduction and Aims

Eating disorders contribute annually to the global burden of disease (Neumark-Sztainer et al., 2006), affecting approximately 3.5–6.5% of females (Allen, Byrne, Oddy, & Crosby, 2013; Hoek & van Hoeken, 2003; Hudson, Hiripi, Pope Jr, & Kessler, 2007) and 3–3.5% of males (Allen et al., 2013; Raevuori, Keski-Rahkonen, & Hoek, 2014), who face significant barriers to treatment such as stigma, access, and affordability (Kazdin, Fitzsimmons-Craft, & Wilfley, 2017). Smartphone applications (apps) offer the possibility to engage individuals across the care continuum and reach those who are not in treatment (Luxton, McCann, Bush, Mishkind, & Reger, 2011, Tregarthen, Lock, & Darcy, 2015). While most apps are offered as fixed interventions, interventions that are tailored to the needs of individuals address heterogeneity in the eating disorder population, leading to more acceptable and effective care, and better clinical outcomes (Collins, Murphy, & Bierman, 2004, Darcy & Sadeh-Sharvit, 2017; Juarascio, Manasse, Goldstein, Forman, & Butryn, 2015). With such advantages in mind, we developed an unguided self-help program with tailored content for identified user clusters (Sadeh-Sharvit et al., submitted) to complement an existing freely available app.

This paper concerns a pilot study whose aim was to assess the feasibility and acceptability of an adaptive app based program among a sample of people with eating disorders. The data reported in this paper were collected during a phase of pilot-testing, following the development of a new adaptive intervention. The second phase of the study uses a randomized controlled trial to test the efficacy of this new version, and is currently in progress. We hypothesized that the app would be feasible and acceptable to participants on average. An exploratory aim was to assess whether app usage patterns among pilot-testers were associated with body mass index (BMI) or eating disorder symptoms such as frequency of binge-eating.

Method

Procedures and sample

The design and procedures used in the current study were approved by our Institutional Review Board. Users of Recovery Record, a smartphone app for eating disorders, who were not linked with a treating clinician, were eligible to participate in this pilot study. Upon registration, 200 new users were issued an in-app invitation to pilot test a new adaptive version of the app for an 8-day period. The newly developed program provides content to complement the standard app, according to baseline individual characteristics and individualized feedback on progress over time. The standard freely available app provides users with meals and symptom self-monitoring in an evidence-based cognitive behavioral treatment (CBT) format and has been described previously (Tregarthen, Lock, & Darcy, 2015).

One hundred eighty-nine users accepted the invitation and provided informed consent. Among these users, 70 percent (n=133 users, “sample 1”) completed the baseline review (“review 1”) and a one-week individualized progress review after approximately 7 days (“review 2”). In addition, after completion of the pilot-testing phase, participants were sent an invitation to participate in a survey evaluating the acceptability of the adaptive version. Of 200 users invited, 64 (32 %) respondents consented (“sample 2”). Acceptability data were completely anonymized and were not linked to personal information provided elsewhere on the app.

Measures

Feasibility

Feasibility was measured by retention rates during the course of the 8-day piloting period (recruitment; enrollment and completion rate).

App utilization

Usage data on the aggregate number of logs (total, and those pertaining to meals, behaviors, thoughts, feelings), binge/purge episodes, goals selected and coping skills used were collected via the smartphone application server pertaining to app usage characteristics.

Acceptability of the intervention – measuring how well they liked the program

Data were collected on users’ perceptions of the app’s acceptability in improving behaviors related to eating disorders after completion of the pilot-testing phase. Users were asked to respond to a survey of 11 items that assessed 1) suitability of the app; 2) predicted efficacy; 3) satisfaction, on a 10-point scale with higher numbers indicating greater acceptability. Sample items include: “how suitable do you think this program is for your problem?”, “how successful do you think the app will be at helping you?”. Users were also asked about their satisfaction with various elements of the app, including: the “look and feel of the program”, 7-day review, suggested goals, and suggested skills. Lastly, users were prompted to rank the app features (i.e., goal setting, coping skills, setting values, successes, identifying obstacles, self-esteem pie chart, meditations, audio message) that were perceived as most helpful to the user on a 8-point scale. Open-ended feedback was collected as well.

Intervention

A multidisciplinary team comprised of experts in eating disorders, information technologies, and user experience design, conceptualized the most urgent needs of each user group and the empirically supported strategies effective for reducing eating disorder symptoms for said prototypes. The adaptive version of the intervention provided users with symptom-specific content in addition to the standard features in the app (e.g. goal setting, learning and practicing coping skills to augment symptom alleviation).

Statistical Analysis

Descriptive statistics such as means and standard deviations as well as frequencies were reported for continuous and categorical variables as appropriate. Chi-squared tests, logistic regression, and Poisson regression models were used to assess the associations between app usage and eating disorder symptoms along with user characteristics.

Covariates

We explored associations between user demographics i.e. age and BMI, and user preferences i.e. frequency of goal reminders, number of coping skills selected, number of goals selected, number of goal reviews completed, and app usage (i.e. logs completed).

For Section 1 of the results, we present the utilization characteristics for the users who accepted the invitation to pilot test the intervention and who completed two consecutive reviews (n=133). For Section 2 of the results, we present survey results from the 64 individuals who accepted the invitation to complete the acceptability survey. All statistical analyses were performed using SAS v9.4 (SAS Institute) and R v3.3.

Results

Primary Aim

Feasibility

Users were 96% female and had a mean age of 23 years. Table 1 presents a summary of demographic and usage characteristics of sample 1, meal log completion, coping skills and goals selection in the first review, coping skills usage (as indicated in review 2), and perceived helpfulness and utility of those skills and goals.

Table 1.

Characteristics of pilot-test users

	Overall N = 133
Gender‡
Female	96%
Male	4%
Median Age (mean ± sd)	23 (24 ±14)
Body Mass Index*	25 (29 ±19)
Log count in 7 days from review 1	21 (21 ±15)
Meal Log count in 7 days from review 1	20 (20 ±14)
Frequency of goal reminders**
Once a day	73%
Every two days	8%
On third day	19%
Feeling about the next 7-days†
Excited	19%
Good	29%
Okay	37%
Not great	15%
Coping skills chosen in review 1††
1	17%
2	30%
3+	53%
Unique goals selected from review 1
1	42%
2	24%
3+	33%
Goal reviews completed midweek
0	21%
1	28%
2	14%
3+	38%
Coping skills used from review 1†
0	23%
1	22%
2	28%
3+	27%

^‡missing observations, m = 11;

^*m = 45;

^**m = 2,

^†m = 4,

^††m=3;

^***m = 6

Respondent perceptions of how they felt about the next review in 7 days were positive on average. Most users (83%, N=108) chose to practice two or more coping skills at the first review. According to review 2, 77% (N = 98) of the users indicated they had used at least one coping skill in the past week. Of these 98 users, 79% had found at least one of the coping skills to have “helped a little” and 55% found at least one coping skill to be “very helpful”. In regard to goals set by users at the beginning of the program, approximately half of the users (51.1%, N = 68) noted that they had at least one goal that they either made good progress on or achieved the goal.

Exploratory findings of associations between personal characteristics and logging behavior

Higher BMI was associated with fewer total recorded logs, adjusting for reminder frequency (β for 2^nd, 3^rd and 4^th BMI quartiles = −0.4, −0.3, −0.3, p<0.01). Lower BMI quartiles were also associated with the total number of restriction logs recorded: 68% of the first BMI quartile had submitted at least 1 restriction log in the period after submitting the first review, compared to 45%, 50%, and 5% from the 2^nd, 3^rd and 4^th quartiles respectively (chi-squared = 19.5, p-value from Chi-squared test =0.001).

A selection of daily reminders for the selected goals was associated with an increased number of recorded logs on average(31 logs on average when reminders were daily, 27.7 logs on average when reminders were every two or three days).

Pilot test exit survey: Acceptability and Suitability of the Mobile App Intervention

A total of 64 app users (i.e., sample 2) completed the acceptability survey. On average, users believed that the app was suitable (mean = 8.5, sd = 1.6) and predicted that it would be successful at helping them (mean = 7.9, sd =1.8). Table 2a summarizes the satisfaction ratings of the 64 users who completed the survey, and includes a full list of features ranked by perceived utility.

Table 2a.

Satisfaction ratings and Program Feature Utility

Satisfaction ratings	Satisfied to Very Satisfied
With the program so far	94% (n=60)
With 7-day review	83% (n=52)
With the suggested goals	91% (n=58)
With the suggested skills	84% (n=54)
With the look and feel of the program	91% (n=58)
Felt that the amount of content on the app
Was appropriate	80% (n=51)
Too much content	14% (n=9)
Too little content	6% (n=4)
Likely to recommend the app to someone with an ED	94% (n=60)
Program Feature	Average Rank (sd)
Goal setting	2.6 (1.8)
Coping skills	3.4 (1.9)
Setting your values	4.0 (2.1)
Your successes	4.1 (1.6)
Identifying obstacles	4.1 (2.0)
Self-esteem pie chart	5.3 (2.2)
Meditations	5.9 (1.8)
Audio messages	6.5 (2.1)

^*ratings for Program Features were measured on a 8-point scale

Qualitative narrative responses

Open-ended qualitative responses about the helpfulness of the review revealed both strengths and opportunities provided by the app intervention. Common recurring strengths were access (e.g. “Being able to get some sort of structured support. My family does not know about my ED, so I don’t get any formal treatment.”), the built-in prompt to select weekly goals and the focus on successes, graphed progress (e.g. “The ability to see what I’m doing as far as successes and improvement”), and the encouragement given through features such as aspiration images and social support. Among areas for improvement offered, was a greater consideration of user burden related to ongoing recording of their data over time, and more instructions and directions given to users.

Discussion

This study piloted an unguided self-help program among a set of mobile app users and found that the adaptive app is feasible, acceptable and is perceived as helpful by users with eating disorder symptoms. Moreover, users were selective in their utilization patterns, and did not use the features they had not found to be of help.

Strengths of the study include the balanced sample of users of a range of BMIs that are reflective of the population living with EDs, and the fact that users selected to use the app spontaneously. Limitations include the comparatively low response rate for the exit survey, the reliance on self-report data. Our study was also a pilot by design, and was not a comparative study linked to clinical outcomes. A randomized controlled trial to test the efficacy of this new version is currently in progress. Despite these limitations, our findings indicate that an adaptive unguided self-help app can be used in at least a subpopulation of users.

For app interventions to be maximally effective, they must be tailored to the needs of their users and be perceived as relevant and useful. Intervention development, pilot testing, and seeking perspectives from the users themselves enabled us to demonstrate the acceptability of an app intervention that is likely to be used with people living with eating disorders. While data gathered related to developing, refining, and evaluating self-help apps for eating disorders may increase treatment options (Fairburn & Patel, 2017), further research is still needed for evidence-based apps with linkage to clinical outcomes.