A mixed-method systematic review of text-based telehealth interventions in eating disorder management Text-basedtelehealthinEDmanagement

Purpose – Eatingdisorders(EDs)isamajorhealthconditionaffecting9%oftheglobalpopulationand10%of thosewithEDslosttheirlivesasaresult.Text-basedtelehealthinterventions(TTIs)seemtoprovidealow-costandconvenienttreatmentoption;however,theevidenceisscarce.ThisstudyaimedtosynthesizeevidencerelatingtotheuseofTTIsforthemanagementofEDs. Design/methodology/approach – Five databases were searched published between January 2020 and May 2019. The authors used keywords relating to telehealth and EDs. The authors used Joanna Briggs Institute ’ s (JBI ’ s) critical appraisal instrument to assess the methodology quality of included studies. Findings – Fifteenstudieswereincludedinthismix-methodsystematicreviewandassessedformethodology quality. Email, web-based texting, text-messaging and online chat room were used as mode for deliver healthcareforpatientswithEDs.Inthetreatmentphase,allstudies(ninestudies; n 5 860participants)showed effectiveness (for RCTs) and usefulness (for non-RCT studies). In the aftercare phase (six studies; n 5 364 participants), the results regarding the effectiveness of TTIs were mixed. Two studies showed effectiveness whilst four studies did not find statistically significant change of ED outcomes. Research limitations/implications – The qualities of these studies varied; firstly, 66% ( n 5 10) of the studies were non-randomized studies (e.g. single-arm trial, case report) with small samples. Moreover, one- fourth ( n 5 4) of the studies did not use validated instruments or indicate the instrument. Also, half ( n 5 7) of the studies used TTIs as adjunct to face-to-face treatment or bigger online treatment, it is hard to make conclusionthat the changeswere due to TTIs ’ effect. In addition,follow-up rate is notsatisfactory,thus results should be interpreted cautiously. Practical implications – TTIs seem to be promising for management of EDs, particularly in the treatment phase. This provides an important treatment option for health practitioners and people with EDs as an alternative or in adjunct with face-to-face services. Originality/value – This is the first review to synthesis the use of TTIs for ED management.

Previous research examined the use of text-based telemedicine techniques for providing healthcare services to patients with EDs. However, to our knowledge, no review has attempted to appraise the evidence for the utility of TTIs, techniques and technologies used and the impact of TTIs on patient outcomes. In order to expand knowledge of these topics, we conducted a review of qualitative and quantitative studies to synthesize the evidence on the use of TTIs.

Search procedures
We conducted a literature search using the keywords related to EDs and text-based telemedicine in five electronic databases, including PsycINFO, PubMed, Embase, Cochrane Library and ProQuest, with the assistance of an experienced library consultant (JH). We used this combination of search terms: (telemedicine OR mHealth OR "mobile health" OR "text message" OR text-messaging OR text-messages OR text-based OR email OR e-mail OR chatroom OR "chat room" OR forum OR "discussion board" OR "message board") AND ("eating disorder" OR "eating disorders" OR bulimi* OR anorexi* OR binge eat* OR EDNOS) (Appendix 1). Additional hand searches with reference lists were also performed. Reviewed papers included the research papers published between January 2000 and May 2021.
The inclusion and exclusion criteria for reviewed papers were formulated using the following PICO statement: (1) Population: participants have ED-related symptoms or are formally diagnosed with EDs, including anorexia nervosa (AN), BN, binge-eating disorder (BED) and EDNOS. Thus, preventive programs were excluded; (2) Intervention: therapy related communication between therapist and patient used text-based telemedicine techniques. Interventions that were delivered by videoconferencing and phone were excluded. Studies examining self-help interventions were excluded; (3) Comparison: randomized controlled studies assessing the effectiveness of text based telemedicine interventions in EDs were included. Since evidence relating to text-based telemedicine interventions on EDs were relatively limited, single arm trials and qualitative studies were also included. (4) Outcomes: studies reporting ED-related outcomes related to text-based telemedicine techniques were included. We screened the selected articles against inclusion and exclusion criteria according to the PRISMA checklist [42].

Data extraction
In addition to basic demographic details of the research articles (ex: author, year, country and study design), two independent researchers (XZ&SE) collected and tabulated information about the type of TTI, disease area, type of therapy and outcomes.

Assessment of methodological quality
Qualitative studies were assessed by using Joanna Briggs Institute's (JBI) critical appraisal instrument for qualitative research [43]. Quantitative studies were assessed by using the JBI's critical appraisal instrument for RCTs and non-randomized experimental studies [44,45]. Two reviewers (XZ and SE) conducted the assessment independently. Discrepancies were resolved by the involvement of a third reviewer (MB). For randomized controlled trials, studies meeting at least ten out of thirteen criteria were categorized into "good" quality, seven Text-based telehealth in ED management to nine were categorized into "moderate" quality, less than seven were deemed "poor" quality. For quasi-experimental studies, studies meeting at least seven out of nine criteria were categorized into "good" quality, five to seven criteria were "moderate" quality, less than five were "poor". For qualitative studies, studies meeting eight out of ten criteria were deemed "good" quality, five to seven were "moderate" quality, and less than five were categorized into "poor" quality.

Results
The initial search resulted in 653 articles. After duplicate records were removed, 341 articles remained for review. After screening titles and abstracts, we excluded 270 articles ( Figure 1).

TTIs during treatment
During treatment period, text-based telehealth techniques included email, web-based texting, text-messaging and online chat rooms. The psychological theories underpinning these interventions were cognitive behavioral techniques (CBT) and motivational interviewing techniques (MI). CBT used in TTIs included keeping a food diary, psychoeducation, eating regular meals, as well as modifying negative automatic thoughts, emotional regulation and maladaptive behaviors were used in TTIs (Table 2). Researchers also sent motivational messages via messaging apps. All nine studies (a total of 860 participants) showed effectiveness (for RCTs) and usefulness (for quasi-experimental studies and qualitative studies) of TTIs among individuals with EDs. Robinson and Serfaty [46] tested the effectiveness of using email to reduce ED-related symptoms among patients with BN, BED and EDNOS. This study found email-delivered intervention was effective both at the end of the intervention and at 3-month follow up. Wagner et al. [47] found that an online asynchronous chat on a webpage with patients with BED was effective in reducing binge-eating episodes, depression level, as well was promoting recovery. Zerwas et al. [48] conducted an intervention in which patients with BN received psychotherapy from therapists. This study found that the text-based telemedicine intervention was equivalent to face-to-face psychotherapies in increasing abstinence rate at follow up. The intervention also alleviated binge eating and depression and improving quality of life at both end of treatment and follow-up. Robinson and Serfaty [49] tested the effectiveness of email therapy on participants with BN, BED and EDNOS. Results indicated that email therapy was effective in reducing bulimic and relevant depressive symptoms. ter Huurne et al. [50] tested the effectiveness of a web-based program in which participants with BN and related EDNOS communicated with therapists via text asynchronously. Results indicated that the intervention was effective in reducing BN-relevant symptoms and improving body satisfaction, quality of life, and general mental and physical health. The improvements sustained after 6 months. Shingleton et al. [51] tested effectiveness of text-messaging in reducing AN and BN. This study found that the intervention was effective in improving motivation among normal weight individuals. Yager [52,53]  JHR useful tool for communication in combination with face-to-face sessions and improved recovery among this population. Moessner and Bauer [54] reported that email as a stand-alone intervention facilitated access to routine care for underserved individuals.

TTIs during aftercare stage
Six studies (a total of 364 participants) used TTIs to provide aftercare for participants to prevent relapse. During aftercare, TTIs were used mainly for monitoring patients. In our review, we found that phones were mainly used text-messaging (in four studies) and a combination of online chat and email (in two studies). Researchers used TTIs to monitor the change of ED-related symptoms, thoughts and emotions as well as to provide feedback. When necessary, TTIs were also used to deliver online counseling (see Table 2). The results regarding the effectiveness of TTIs were mixed. Three studies [55][56][57] used text-messaging to monitor changes in ED-related symptoms and provide tailored feedback. Their results indicated a reduction in ED-related symptoms after intervention. In contrast, a similar intervention [58] did not find statistically significant changes. Two studies [59,60] used an online chat group and email to provide aftercare support for ED patients. This study found that this intervention was acceptable and feasible, but not effective.

Methodological quality of included studies
For randomized controlled trials (Table 3), we found that the quality of evidence for three studies were "good" and two were "moderate." For quasi-experimental studies (Table 4), quality assessment showed that five studies were categorized into "good" quality, while two were "moderate." Because six out of seven quasi-experimental studies were single-arm trial, none of them had control group. For qualitative studies (Table 5), all three studies were of "good" quality.

Discussion
We found evidence suggesting that telemedicine is effective in treating Eds [20,26]. For example, video-based modality (i.e. videoconferencing) produced an equivalent outcome to face-to-face conditions [17,27]. Online CBT as well as Internet-based self-help programs Treatment Aftercare (1) Deliver cognitive behavioral therapies for participants (2) Monitor symptoms relating to eating disorders (3) Deliver motivational interviewing (MI) (4) Send motivational messages to participants (5) Participants report eating disorderrelated symptoms (6) Clients do "homework" and send it back to the researcher (7) Provide feedback to participants about their progress (8) Set goal and/or plan (9) Psycho-education (10) Deliver self-control techniques (11) Deliver exposure techniques (1) Monitor eating disorder symptoms (2) report positive or negative life events, thoughts, feelings and their progress (or lack of progress) (3) Automatic personalized feedback to educate the recipients about positive and negative changes in their status. The feedback aims to motivate participants to solve their own problems (4) Send feedback about the participants' progress (5) Online counseling Table 2.

Examples of Online
Text-based Interventions (OTIs) used in the treatment phase and aftercare phase for eating disorders in mixed methods systematic review Text-based telehealth in ED management reduced ED-related symptoms more effectively than the waiting list control group [61,62]. Our current review found the evidence that TTIs have positive effects on managing EDs. Thus, this analysis adds new evidence showing effectiveness of telemedicine for Eds [63].
Among studies testing the effectiveness and usefulness in treating BN, BED and EDNOS, all studies showed that TTIs were effective in reducing bulimic symptoms and severity. We found that TTIs which were underpinned by CBT and were particularly effective in interventions. This finding is consistent with previous evidence on the effectiveness of CBT on EDs [64]. Another technique used in the included studies was motivational interviewing (MI). It was found that text-messaging underpinned by MI techniques had effect on motivation to change rather than on eating behaviors [51].
In the aftercare stage, previous studies have shown that outpatient care prevents relapse after acute care. In addition, mobile and internet-based interventions were promising and may help maintain treatment gains [65]. Our current study showed that text-based interventions were helpful during the aftercare phase because they enabled monitoring of patients' symptoms in a less disruptive way for patients. When the text-based feedback is tailored to the reported symptoms, patients may want to maintain behavior changes. However, some participants provided feedback that TTIs may not be able to provide the empathetic feedback which human therapists are able to. Therefore, future studies may explore ways to improve the patients' experience.
Current studies also found various features, as well as advantages and disadvantages of different modalities of TTIs. Interventions delivered via text messages were preprogrammed in reviewed studies. These interventions usually served a single purpose such as symptom monitoring [57]. Others focused on global health and increasing motivation to change eating behaviors [51]. Previous studies also reported that the purpose of the text messaging was to provide support and assist in self-monitoring. They distributed supportive messages and self-monitoring procedures in mental health interventions [66]. Therefore, text messaging  Clear definition of "cause" and "effect" Text-based telehealth in ED management In terms of the feasibility and acceptability of TTIs, the current literature generally shows that retention rate and adherence rate were high and participant feedback was positive [26,[66][67][68]. However, some participants and therapists in reviewed studies preferred to face-toface consultations. In contrast to texting interventions, email or online chat (chat group, webpage-based chat) interventions involved a human therapist on the other end. Therefore, chat interventions were able to conduct more complex and complete therapeutic process. Also, the responses were more personalized and appropriate. These modalities gave participants opportunities to discuss their situations and concerns [47,50]. But this type of intervention was less frequent than text-message interventions.
There are some limitations in our review that are worth noting. The quality of these studies varied. First, 66% (n 5 10) of the studies were non-randomized studies (e.g. single-arm trial, case report) with small samples. Moreover, one-fourth (n 5 4) of the studies did not use validated instruments or indicate the instrument. Also, half (n 5 7) of the studies used TTIs as an adjunct to face-to-face treatment or bigger online treatment. For these studies, it is hard to conclude that the changes in outcome were due to TTIs, those results therefore should be interpreted cautiously.

Conclusion
Overall, our review found that TTIs generally had positive results on the management of EDs. TTIs including email, text-messaging, webpage and online chat room, are viable ways to enhance or deliver multiple levels of treatment for EDs. They help reduce ED-related symptoms, comorbid depression and anxiety, improve body satisfaction and contact between clients and therapists. However, due variability in quality of the reviewed studies, our conclusions should be interpreted with some caution. In the future, more rigorously designed studies with larger samples are required to evaluate the effectiveness of TTIs for EDs.