Real-world patient outcomes for telehealth-delivered, remote eating disorder treatment: a scoping review

Abstract

Background

Only 30% of individuals with eating disorders receive specialized treatment. While preliminary evidence suggests that telehealth-delivered, remote eating disorder treatment may offer improved accessibility with similar effectiveness to in-person treatment, research on these services remains limited, particularly regarding the communities that are disproportionately affected by barriers to standard care. This scoping review sought to map the existing research on real-world patient outcomes in remote eating disorder treatment, identify knowledge gaps, and prioritize areas for future studies.

Methods

This review followed the Joanna Briggs Institute methodology for scoping reviews. It comprises observational evaluations of telehealth-delivered, remote eating disorder treatment conducted in routine clinical settings. An electronic database search was performed in PsycINFO, PubMed, and ProQuest Dissertations & Theses Global in August 2024 and updated in September 2025.

Results

Following the search and screening process, 27 articles, comprising six case reports and 21 cohort/case series designs, were deemed eligible for inclusion. Remote treatments evaluated differed across level of care, therapeutic modalities, provider types, dosage, and adjunctive technologies used. Just under half of the studies compared outcomes from remote and in-person treatment, while the remainder examined remote treatment alone. Articles were published between 2011 and 2025 and, when excluding case reports, nearly 60% evaluated programs that rapidly transitioned to remote delivery due to COVID-19. While demographic reporting was limited and inconsistent, available information indicated that participants ranged from three to 75 years old and were predominantly White, cisgender women/females diagnosed with anorexia nervosa. Though preliminary, findings tentatively suggest that remote eating disorder treatment can yield improvements across core outcome domains, largely comparable to in-person settings. Less is known about how outcomes may differ across demographic groups.

Conclusions

Overall, this body of literature remains small and characterized by limitations and inconsistencies, including differences in the treatment services evaluated as well as disparities in study design, methodology, and reporting. Utilization of remote treatment by historically excluded groups remains low, calling for further reflection around its accessibility for target communities. Additional studies with more rigorous, intentional designs are needed. The field would also benefit from standardization in relation to data collection and reporting to allow for better synthesis of findings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40337-025-01441-5.

Plain language summary

Remote eating disorder treatment (i.e., telehealth) may help improve access to care, especially for groups like racial and ethnic minorities who often face additional barriers, such as stigma. Research on patient outcomes in remote eating disorder services delivered in real-world clinical settings is limited, especially in relation to these historically underrepresented groups. This scoping review mapped the existing research to identify gaps and prioritize directions for future studies. Twenty-seven articles from 2011 to 2025 were included in the review. Many studies evaluated programs that quickly switched to remote care because of COVID-19. Overall, the treatment services evaluated were quite varied, studies had limitations related to design and methodology, and there were inconsistencies in how things were reported, making it difficult to combine findings and draw conclusions. Tentatively, results suggest that remote eating disorder treatment can be effective, however this conclusion should be interpreted with caution given the inconsistencies and limitations identified, including a lack of diversity in study participants which limits generalizability. Additional high-quality research is needed to confirm these findings. More consistency in what data are collected and how data are reported would allow for better interpretation of results across studies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40337-025-01441-5.

Background

Eating disorders (EDs) are severe, life-threatening illnesses that confer significant mental and physical impairment [1–3]. Despite advances, effective treatment of EDs is complicated by two critical gaps [4]. First, although controlled studies may identify efficacious treatments, a disconnect remains regarding what is routinely offered in clinical practice (i.e., the research-practice gap). Furthermore, even when effective treatment is available in real-world settings, a treatment gap persists, with many individuals unable to access this care. These challenges are evident in reported treatment rates, with only about 30% of individuals with EDs receiving any formal specialized care [5]. This limited access is in part attributed to the myriad barriers individuals with EDs face when seeking treatment, including financial, geographic, identification, sociocultural, and treatment quality challenges [6]. Given these barriers, identifying treatment approaches utilized in real-world ED treatment settings that have the potential to expand access is essential.

Telehealth-delivered, remote ED treatment (i.e., care delivered by clinicians via videoconferencing) has emerged as one potential strategy to improve access [4] by addressing some common barriers to in-person treatment, such as geographic disparities in services, time constraints, and stigma [6, 7]. Indeed, reported benefits have included reduced travel burden and increased flexibility and convenience [8–10]. While remote interventions are not without limitations, such as privacy concerns and technological challenges [11], preliminary studies in real-world settings suggest that remote ED treatment yields meaningful improvements across symptom domains [12, 13], such as ED, depressive, and anxiety symptoms, with changes similar to those seen in in-person care [14–16]. Such research was spurred by the onset of the COVID-19 pandemic but continues to emerge as the body of evidence develops alongside the ongoing adoption of remote services in clinical practice. Further investigation into patient outcomes in remote ED treatment is still needed, particularly for long-term applicability.

Importantly, barriers to ED treatment disproportionately affect certain groups [6], as some communities encounter a greater number of barriers overall and are also more likely to face distinct types of barriers. For instance, gender and sexual minorities report significantly more barriers to treatment than their cisgender and heterosexual peers [6], with qualitative research highlighting barriers among this population such as affordability and availability of services, stereotypes that undermine access to ED identification and treatment, and cultural responsiveness of care [17]. In addition, individuals who have a household income of ≤$70,000 report significantly more obstacles to care than their higher income counterparts, a finding unsurprising given that financial challenges are the most frequently cited barrier to ED treatment [6]. Moreover, some groups are more likely to experience specific kinds of barriers (e.g., identification, sociocultural) due to systemic misperceptions and biases around EDs. For example, clinician bias and gendered, racialized norms around help-seeking are obstacles to care commonly faced by men and racial and ethnic minorities [18–20]. Additionally, while there is limited research on EDs in older adults, it seems there is a lack of ED identification among this demographic due to stigma and stereotypes as well as difficulties seeking care due to social norms involving familial and work responsibilities [21]. The disproportionate amount and nature of treatment barriers faced by these groups have resulted in disparities in ED treatment access [22–24], with many of these communities being historically underrepresented in ED treatment services. Thus, implementing strategies to expand access to care for these populations is critical. However, despite speculation that remote ED treatment may help bridge access to care for these groups (e.g., by reducing financial burden associated with travel, expanding the availability and reach of culturally responsive care, or increasing flexibility and convenience to accommodate work or parenting), it remains unclear if remote treatment has achieved this goal. Furthermore, little is known regarding if and how remote ED treatment outcomes may differ across demographic groups, thus the effectiveness of remote treatment for these target groups is yet to be established.

In sum, while telehealth-delivered, remote ED treatment holds promise for narrowing the treatment gap and improving access to care, particularly for groups disproportionately affected by barriers to traditional treatment, the evidence base on these services remains limited. As societal perceptions of remote care shift away from its role as a pandemic-driven stopgap towards its potential as a sustainable and accessible care model, it is important to examine the breadth and focus of the real-world treatment literature and clarify the necessary directions for future investigation to address the research-practice gap. Accordingly, the purpose of this scoping review is to map the existing observational research on patient outcomes in remote ED treatment, identify knowledge gaps, and prioritize areas for future studies. The review will summarize how patient outcomes in these treatments have been evaluated thus far, as well as provide insights into who is utilizing these services in the real world, what the literature reveals about preliminary patient outcomes, and how these outcomes might differ across demographic groups. A scoping review methodology was deemed most appropriate considering the broad objectives and emerging nature of this field of research. A preliminary search of MEDLINE, the Cochrane Database of Systematic Reviews, and JBI Evidence Synthesis was conducted and no current or underway systematic or scoping reviews on the topic were identified.

Review questions

The following research question was used to guide this review: How have patient outcomes for telehealth-delivered, remote ED treatment been evaluated in the existing observational literature, and what do these studies suggest regarding patient characteristics and treatment outcomes? For clarity, this overarching question was separated into four subquestions:

1. How have patient outcomes for telehealth-delivered, remote ED treatment been evaluated in the existing observational literature?

2. Who is utilizing telehealth-delivered, remote ED treatment in the real world?

3. What does the existing literature suggest regarding preliminary patient outcomes for telehealth-delivered, remote ED treatment?

4. What are the outcomes across different demographic groups?

Inclusion criteria

Types of participants

The population of interest comprises individuals of all ages who meet diagnostic criteria and are engaged in treatment for an ED.

Concept

The concept evaluated is patient outcomes in telehealth-delivered, remote ED treatment, defined as treatment in which patients and providers are not in the same location but rather care is delivered by clinicians via videoconferencing to patients in their home or another environment outside of a treatment facility. Treatments delivered across all levels of care and encompassing various treatment components (e.g., individual therapy, group therapy) were included, provided they met these criteria. Research evaluating digital therapeutics or other asynchronous technology-based interventions as well as remote interventions delivered by non-credentialed providers (e.g., coaching) was excluded. Additionally, patient outcomes were defined using published consensus guidelines for EDs [25–27] and included the following domains: physical health (e.g., weight restoration, medical indicators), ED symptoms, co-occurring mental health conditions, and quality of life (QOL)/functioning. Papers which did not include any of these outcomes, instead exclusively focusing on concepts such as patient experiences or satisfaction, were excluded.

Context

The context of interest includes remote treatment settings in any geographic location. Research conducted in in-person treatment settings or in hybrid programs, where participants received both remote and in-person care, was excluded unless the study design allowed for a meaningful comparison or differentiation of outcomes between the two formats of care.

Types of sources of evidence

This review considered peer-reviewed articles and dissertations that feature empirical studies. Conference abstracts, book chapters, and non-peer reviewed articles were excluded, as well as review articles, meta-analyses, commentaries, and editorials because they are not primary sources presenting original data and/or generally lack sufficient methodological detail as well as a formal review process. The review examined observational treatment studies only (i.e., in which treatment was delivered in routine clinical settings, without active participant recruitment or experimental manipulation), including quantitative, qualitative, and mixed methods designs. Experimental studies were excluded as the review aimed to examine patients and outcomes in real-world settings rather than under controlled or simulated conditions. Additionally, non-treatment studies were excluded because they do not address the review’s objective to understand how patient outcomes of remote ED treatments have been evaluated.

Methods

This scoping review followed the Joanna Briggs Institute (JBI) methodology for scoping reviews [28, 29] as well as the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR) checklist [30]. The review’s objectives, inclusion criteria, and methods were specified in advance and documented in a protocol that was not pre-registered.

Search strategy

A three-step literature search was conducted to identify relevant studies. First, preliminary words and phrases were searched in the PsycINFO Thesaurus and the Medical Subject Headings (MeSH) vocabulary to identify relevant terms. These terms were used to perform a limited database search in PsycINFO and PubMed to screen titles, abstracts, and keywords for any additional relevant terms. Next, the identified terms were used to develop a full search strategy for three electronic databases (i.e., PsycINFO, PubMed, ProQuest Dissertations & Theses Global), selected to ensure coverage of psychological and medical-oriented studies and to encompass published journal articles as well as theses/dissertations. Lastly, reference lists of studies selected for the review were manually searched for additional relevant articles. The initial search was conducted in August 2024 and later updated in September 2025. Any literature published in English (due to feasibility constraints) up until that time was included (i.e., no restrictions on date of publication were applied). For full details regarding both the initial and updated search strategies, please see Supplementary Material 1.

Source of evidence screening and selection

Following the search, all identified citations were uploaded into the systematic review software, Rayyan [31]. Duplicates were removed and the remaining articles were retained for a two-stage screening process conducted by three authors (HW, CBS, RL) after the initial search and two authors (HW, CBS) after the updated search (also due to feasibility constraints). Following a pilot test, reviewers first independently screened the articles’ titles and abstracts, using the eligibility criteria (see Supplementary Material 2) to determine their suitability for further review. Potentially relevant sources were then retrieved in full and imported into Rayyan. The full-text manuscripts of these citations were then independently screened by the same authors to determine their eligibility for inclusion. Throughout screening, all discrepancies were resolved via reviewer discussion until consensus was achieved.

Data extraction

Data were extracted from papers included in the scoping review by two independent authors (HW and either RL or CBS) using a data extraction tool (See Supplementary Material 3) developed by the reviewers and informed by consensus guidelines for EDs [25–27]. Data extracted included the following categories: (1) Publication Info and Methodology (e.g., study design, country of origin), (2) Treatment Characteristics (e.g., level of care, therapeutic modality), (3) Sample Characteristics (e.g., sample size, demographic and diagnostic information), and (4) Outcomes (e.g., domains assessed; measures utilized; baseline, discharge, and follow-up scores). The extraction form was piloted on three sources to establish consistency across reviewers and ensure that all relevant results were extracted. All discrepancies were resolved through discussion and/or with the assistance of the third reviewer. In line with standard scoping review practice and methodology [28, 29], a formal critical appraisal of included studies was not performed as the purpose was to map the breadth of existing evidence rather than assess study quality.

Analysis and presentation of results

Extracted data were collated, summarized, and reported in tabular and narrative forms.

Results

Search results

Results from the initial and updated searches have been summed and are reported in total below. 859 articles published from 1972 to 2025 were identified in the database searches. No additional articles were identified via the backward reference searching. See Fig. 1 for the PRISMA flow diagram of study selection [32]. Following the removal of duplicates, a total of 675 articles were retained. In the initial title and abstract screening stage, 628 articles were excluded as not relevant. Upon full-text examination of the remaining 47 articles, an additional 18 were excluded for the following reasons: not an observational study (n = 8), not clinician-delivered treatment via videoconferencing (n = 6), not published in English (n = 1), not an empirical study (n = 1), not a clinical ED sample or subsample (n = 1), and no inclusion of recommended ED outcomes (n = 1) (see Supplementary Material 4 for the list of sources excluded following full-text review with primary reasons for exclusion). After screening concluded, two additional articles were excluded during data extraction as the authors noted that demographic and clinical details of the participants had been altered due to privacy concerns; thus, the accuracy of the information presented could not be confirmed to address our research question [33, 34]. Consequently, a total of 27 articles were deemed eligible to be included in the present review [12, 14–16, 35–57].

Fig. 1.

PRISMA Flow Diagram. *Articles were excluded following full-text screening during data extraction when reviewers noticed that the authors had altered participants’ demographic and clinical details to maintain their privacy, thus the accuracy of the data needed to answer the research question could not be confirmed

Inclusion of sources of evidence

As per the inclusion criteria, all 27 studies were empirical, observational evaluations that examined patient outcomes of telehealth-delivered, remote ED treatment and were available in English.

Review findings

How have patient outcomes for telehealth-delivered, remote ED treatment been evaluated in the existing observational literature?

Of the 27 included studies, 26 were peer reviewed articles and one was a doctoral dissertation [57]. Table 1 provides an overview of key characteristics related to the treatments evaluated in each of these studies, which varied across articles. Just over half examined outpatient care (n = 15) and the remainder examined day treatment (n = 12), such as partial hospitalization (PHP) and intensive outpatient programs (IOP). The frequency and intensity of outpatient treatment ranged from biweekly to five sessions per week, while day treatment programs ranged from two to six hours per day, three to seven days per week. Most day treatment programs (n = 10) and two thirds of outpatient treatments (n = 10) reported multidisciplinary care teams, typically including a psychotherapist (n = 18) and a dietitian (n = 17), with some articles also mentioning a psychiatric provider (n = 9), nurse (n = 4), or other medical provider (n = 10), and/or additional clinical support staff (e.g., peer mentors, care partners) (n = 10). The remaining treatments were delivered solely by psychotherapists (n = 6). All treatments consisted of individual, family, and/or group therapy, with some also including nutrition therapy (n = 17), psychiatric consultation (n = 13), sessions with clinical support staff (n = 10), and meal support (n = 8). Thirteen studies reported using adjunctive technologies, including online portals or platforms (n = 8; e.g., for text messaging, progress monitoring), telephone consultations to supplement and/or replace some videoconferencing sessions (n = 2), email to reinforce session material and give feedback on food diaries (n = 1), mobile applications with features like meal logs, self-guided content, and chat functions (n = 4), and remote monitoring devices (e.g., blood pressure cuff, thermometer, scale) to track vitals and weights (n = 4). As for therapeutic modalities, approximately half of the programs (n = 13) offered cognitive behavioral therapies (CBTs), including manualized protocols as well as treatment informed by all waves of CBTs (e.g., dialectical behavior therapy, acceptance and commitment therapy), while roughly 40% of studies (n = 11) evaluated family-based treatment (FBT) or FBT-informed approaches. Finally, several treatment programs were eclectic or integrative (n = 6), predominantly combining CBTs, psychoeducation, and experiential modalities, and one case report examined outpatient schema therapy.

Table 1.

Remote treatment characteristics

Citation	Level of care	Type(s) of providers	Treatment component(s)	Therapeutic modalities	Frequency/intensity of treatment	Adjunctive technology
Baker 2024	OP	Multidisciplinary T, D, P, OMP, CSSa	IT, GT, FT, NT, PC, SSa	FBT or FBT-informed	~ 5 50 min sessions/wka	Online portal/platforma
Blalock 2025	DT (IOP)	Multidisciplinary T, D	IT, GT, FT, NT, MS	CBTs or CBTs-informed	5 sessions (10.5 h)/wk	Online portal/platform; mobile application; remote medical monitoring devices
Cai 2025	OP	Multidisciplinary T, D, OMP, CSS	IT, GT, FT, NT, PC, SSa	FBT or FBT-informed	~ 5 50 min sessions/wka	Online portal/platforma
Carr 2022	DT (unspecified)	NR	IT, GT, MS	Eclectic/Integrative	NR	NR
Drury 2025	OP	Multidisciplinary T, D, P	FT, NT, PC	FBT or FBT-informed	NR	NR
Franklin 2025	OP	Multidisciplinary T, D, N or OMP	IT, GT, FT, NT	FBT or FBT-informed; CBTs or CBTs-informed	~ 4 h/wk	NR
Hellner 2025	OP	Multidisciplinary T, D, OMP, CSS	IT, GT, FT, NT, PC, SSa	FBT or FBT-informed; CBTs or CBTs-informed	~ 5 50 min sessions/wka	Online portal/platforma
Levinson 2021	DT (IOP)	Multidisciplinary T, D, OMP	IT, GT, NT, MS, PC	CBTs or CBTs-informed	3 h/day, 5 days/wkb	NR
Melisse 2024	OP	Multidisciplinary T, D	IT, NT	CBTs or CBTs-informed	Twice weekly to biweekly IT sessions	NR
Murray 2022	OP	T	IT	CBTs or CBTs-informed	2 sessions/wk	NR
Ortiz 2023	DT (PHP/IOP)	Multidisciplinary (unspecified)	IT, GT, FT, NT, MS	CBTs or CBTs-informed	PHP: 6 h/day, 5 days/wk IOP: 3 h/day, 5 days/wk	NR
Penwell 2024	DT (PHP)	Multidisciplinary T, D, P	IT, GT, FT, NT, MS, PC	CBTs or CBTs-informed	5 h/day, 5 days/wk	NR
Pereira 2023	OP	Multidisciplinary T, D, P, OMP	FT	FBT or FBT-informed	NR	NR
Perry 2024	OP	Multidisciplinary T, D, OMP, CSS	IT, GT, FT, NT, PC, SSa	FBT or FBT-informed	~ 5 50 min sessions/wka	Online portal/platforma
Plumley 2021	DT (PHP)	Multidisciplinary T, OMP, CSS	IT, GT, FT, MS, SS	CBTs or CBTs-informed	5 h/day, 4 days/wk	Telephone consultations
Price 2025	OP	T	IT	CBTs or CBTs-informed	NR	Online portal/platform
Raykos 2021	OP	T	IT, FT	FBT or FBT-informed; CBTs or CBTs-informed	NR	NR
Shepherd 2023	DT (PHP/IOP)	Multidisciplinary T, D, P, N, CSS	IT, GT, FT, NT, MS, PC, SS	Eclectic/Integrative	PHP: 6 + h/day, 5 days/wk IOP: 3 + h/day, 3–5 days/wk	Mobile application; remote medical monitoring devices
Shepherd 2024	DT (PHP/IOP)	Multidisciplinary T, D, P, N, CSS	IT, GT, FT, NT, PC, SS	Eclectic/Integrative	PHP: 6–7 days/wk, avg 26 h/wk IOP: 4–7 days/wk, avg 13 h/wk	Mobile application; remote medical monitoring devices
Shepherd 2025	DT (PHP/IOP)	Multidisciplinary T, D, P, N, CSSa	IT, GT, FT, NT, PC, SSa	Eclectic/Integrative	PHP: 5 + h/day, 5–7 days/wk IOP: 3 + h/day, 3–5 days/wk	Mobile application; remote medical monitoring devices
Simpson 2011	OP	T	IT	Schema Therapy	2 sessions/wk for 4 wks, 1/wk thereafter	Email; telephone consultations
Steiger 2022	OP	T	IT, GT	CBTs or CBTs-informed	Weekly GT, bimonthly IT	NR
Steinberg 2023a	OP	Multidisciplinary T, D, P, OMP, CSSa	IT, GT, FT, NT, PC, SSa	FBT or FBT-informed	~ 5 50 min sessions/wka	Online portal/platforma
Steinberg 2023b	OP	Multidisciplinary T, D, P, OMP, CSS	IT, GT, FT, NT, PC, SS	FBT or FBT-informed	~ 5 50 min sessions/wk	Online portal/platform
Thaler 2024	DT (IOP)	T	GT, MS	CBTs or CBTs-informed	2–3 h/day, 4 days/wk	NR
Van Huysse 2023	DT (IOP)	Multidisciplinary (unspecified)	GT, FT, PC	FBT or FBT-informed; Eclectic/Integrative	~ 13 h across 4 days/wk	NR
Wilkes 2019	DT (IOP)	Multidisciplinary T, D, P	IT, GT, FT, NT	Eclectic/Integrative	11 + hr/wk	NR

avg, average; CBTs, cognitive behavioral therapies; CSS, clinical support staff; D, dietitian; DT, day treatment; FBT, family-based treatment; FT, family therapy; GT, group therapy; h, hours; IOP, intensive outpatient program; IT, individual therapy; MS, meal support; N, nurse; NR, not reported; NT, nutrition therapy; OMP, other medical provider; OP, outpatient; P, psychiatric provider; PC, psychiatric consultation; PHP, partial hospitalization program; SS, sessions with clinical support staff; T, therapist; wk, week

^aAll components not reported in manuscript, but included in another publication (i.e., Steinberg 2023b, Shepherd 2023, or Shepherd 2024) that readers are directed to

^bAll components not reported in manuscript, but included in supplementary materials

As for study design and methodology, the review comprises 20 cohort studies, six case reports, and one article that has elements of both a cohort and case series design (see Table 2 for a summary of study designs and methodologies). Importantly, since the latter article [48] reported cohort outcomes without distinguishing remote and in-person treatment formats, this review only reports on the case series findings. Just over half of the studies examined remote treatment only (n = 16), while the remainder compared remote and in-person outcomes (n = 11). Most comparison studies seemed to have mutually exclusive groups (i.e., remote vs. in-person), however two studies included participants who received both formats, with one classifying patients by predominant format [38] and the other utilizing a within-subjects design [48]. Studies were published between 2011 and 2025, with all but two being published between 2021 and 2025, and the majority of participants receiving remote treatment between 2020 and 2021. Relatedly, just under half of the studies (n = 12) evaluated programs that rapidly transitioned to remote delivery due to the onset of the COVID-19 pandemic and most were conducted in the United States (n = 18).

Table 2.

Study design and methodology

Citation	Country of origin	Study design	In-person comparison	Covid-19 transition	Data collection timeline
Baker 2024	USA	Cohort	No		Sep 1 2020–Nov 1 2022
Blalock 2025	USA	Cohort	Yes	✓	In-person: Mid 2021–Mid 2022 Remote: Jan 2021–Mid 2022
Cai 2025	USA	Cohort	No		Sep 2020 - May 2023
Carr 2022	UK	Cohort	Yesa	✓	In-person: Feb 2 2018–Jan 3 2020 Remote: Sep 17 2019–May 19 2021
Drury 2025	USA	Cohort	Yes	✓	In-person: Jan 1 2014–Mar 16 2020 Remote: Mar 17 2020–Sep 30 2023
Franklin 2025	USA	Cohort	No		Patients referred to treatment between Sep 25 2021–Mar 30 2023
Hellner 2025	USA	Cohort	No		Aug 14 2023–Jul 31 2024
Levinson 2021	USA	Cohort	Yes	✓	In-person: Mar 8 2018–Mar 15 2020 Remote: Mar 16 2020–Jan 10 2021
Melisse 2024	NED	Case report	No		Presented for treatment in Jan 2024
Murray 2022	USA	Case report	No		NR
Ortiz 2023	USA	Cohort	Yes	✓	In-person: Jan 2019–Mar 2020 Remote: Mar 2020–Sep 2021
Penwell 2024	USA	Cohort	Yes	✓	In-person: Oct 2019–Feb 27 2020 Remote: Mar 23 2020–Aug 2020
Pereira 2023	CAN	Cohort	Yes	✓	In-person: 2017–2018 Remote: 2020–2021
Perry 2024	USA	Cohort	No		Patients seeking treatment between Sep 2020–May 2023
Plumley 2021	UK	Cohort	No	✓	NR
Price 2025	UK	Case report	No		Feb 2023–Oct 2023
Raykos 2021	AUS	Cohort; Case series	Yesb	✓	Remote sessions began Mar 23–Apr 22 2020
Shepherd 2023	USA	Case report	No		NR
Shepherd 2024	USA	Case report	No		NR
Shepherd 2025	USA	Cohort	No		2021–2023
Simpson 2011	UK	Case report	No		NR
Steiger 2022	CAN	Cohort	Yes	✓	In-person: Sep 2017–Feb 2020 Remote: Apr 2020–May 2021
Steinberg 2023a	USA	Cohort	No		Sep 1 2020–Jun 1 2022
Steinberg 2023b	USA	Cohort	No		Sep 1 2020–Aug 31 2021
Thaler 2024	CAN	Cohort	Yes	✓	In-person: Jan 2017–Dec 2019 Remote: Mar 2020–Dec 2022
Van Huysse 2023	USA	Cohort	Yes	✓	Patients admitted to treatment between: In-person: Oct 2018–2019 Remote: Oct 2020–Oct 2021
Wilkes 2019	USA	Cohort	No		Patients inquired about treatment between Jan 2016–Mar 2017

AUS, Australia; CAN, Canada; NED, Netherlands; NR, not reported; UK, United Kingdom; USA, United States

^aPatients were classified into remote and in-person groups based on how the majority of their treatment was delivered, so some patients in the remote group received some in-person treatment and vice versa

^bRemote and in-person outcomes were compared within-subjects rather than between-subjects

With regard to study samples, aside from case reports, selection was based on varied criteria. Five studies appeared to include all patients who admitted during their study timeframe [15, 35, 46, 48, 54], while the remaining 16 studies utilized additional eligibility criteria, including requiring research consent (n = 6) [12, 16, 36, 44, 56, 57], as well as exclusion based on diagnosis (n = 8) [14, 16, 37, 38, 40, 41, 44, 45], completion of outcome measures or weights (n = 7) [12, 14, 16, 36–38, 45], weight/need for weight restoration (n = 5) [16, 37, 44, 45, 53], length of stay (n = 5) [12, 14, 37, 40, 45], age (n = 4) [16, 53, 55, 57], attendance/engagement (n = 3) [16, 40, 55], readmission status (n = 2) [12, 39], level of care (n = 1) [51], hybrid treatment (n = 1) [39], and/or randomized controlled trial participation (n = 1) [39]. In addition to sample criteria, some treatment programs had admission criteria beyond standard considerations for determining an appropriate level of care [58, 59], including limitations based on age (n = 11) [14, 35, 38–40, 44, 46, 51, 53, 54, 56] and diagnosis (n = 3) [39, 40, 46]. All criteria considered, the cohort and case series studies had remote sample sizes ranging from nine to 1,235, though one third (n = 7) of remote samples comprised less than 50 participants and more than 60% (n = 13) comprised less than 100.

While there was considerable consistency in the assessment and operationalization of outcome domains, some differences were also observed. For physical health, all but two cohort studies reported on weight restoration, as well as five case reports where it was not relevant to the client’s treatment goals. Weight restoration was most commonly examined via changes in body mass index (BMI; n = 11). One cohort study and four case studies reported on other medical indicators, two of which documented concerns at admission and tracked outcomes throughout treatment, including changes in endocrine, gastrointestinal, and/or cardiac indicators. The majority of studies also reported on ED symptoms (n = 22). Three patient-reported outcome measures (PROMs) were utilized to assess ED symptoms, with the majority using the Eating Disorder Examination-Questionnaire (EDE-Q; n = 16). Most studies also assessed at least one co-occurring mental health condition throughout treatment (n = 18), including depressive symptoms (n = 17), anxiety symptoms (n = 15), and/or suicidal ideation (n = 2). While six different PROMs were used to assess depressive symptoms and five used to assess anxiety symptoms across studies, the most commonly used measures were the Patient Health Questionnaire-9 (PHQ-9; n = 11) and the Generalized Anxiety Disorder-7 (GAD-7; n = 9). Finally, only eight studies reported on QOL/functioning. Four PROMs were used across studies, with the Clinical Impairment Assessment (CIA; n = 3) and the Eating Disorder Quality of Life Instrument (EDQOL; n = 3) being the most frequently used. While this review was not restricted to quantitative research, no relevant qualitative findings were identified.

Who is utilizing telehealth-delivered, remote ED treatment in the real world?

A high-level summary of commonly reported demographic and clinical characteristics is provided in Table 3. Regarding demographics, age, sex/gender, and race/ethnicity were the only variables reported on by at least half of included articles. Participants in remote samples ranged from three to 75 years old with the mean age falling between 14 and 36 across all studies. Nineteen studies specified age groups, with 63% of those indicating a majority or entirely adult sample. As for sex/gender, the majority of participants were female/cisgender women, with exceptions limited to avoidant/restrictive food intake disorder (ARFID)-only samples: one male case report and two cohort/case series studies with less than 75% females and/or cisgender women. Among the cohort and case series studies, six reported only the percentage of participants in the majority sex/gender category (i.e., female, cisgender woman), without providing data on other groups. The remaining 15 that provided a full breakdown showed that males/cisgender men were most often the next most prevalent category (up to 40%), with other gender identities less represented (i.e., ≤ 6%). Importantly, only three studies [15, 39, 56] reported sex at birth and gender identity as distinct constructs, with the remaining articles reporting on either sex or gender, with several using these terms interchangeably (e.g., describing gender as male and female) and/or not clearly distinguishing between the two variables (e.g., combining male, female, and non-binary into one variable). Similarly, race and ethnicity were frequently combined, with only four out of the 21 studies that reported on them describing both as distinct variables [39, 54, 56, 57]. Where reported, all case study participants identified as White and all but three cohort/case series studies had remote samples consisting of at least 70% White participants. Six of the cohort/case series studies described the percentage of White participants only. The remaining twelve shared percentages for all reported race/ethnicity categories, with non-White representation largely coming from individuals who identified as Asian or Pacific Islander (1.7–18.2%), Black or African American (1.3–5.9%), or multiracial (3–11.8.8%), with minimal representation of Middle Eastern or Northern African and/or Native American or Indigenous individuals (i.e., ≤ 2%). Additionally, where reported, Hispanic or Latinx individuals constituted 0–60.3% of study samples.

Table 3.

Sample characteristics in remote treatment

Citation	Sample size	Age	Age group	Sex and/or gender	Race	ED diagnosis	Duration of treatment
Case reports
Melisse 2024	1	40s	Adult	Woman	NR	BED, OSFED (Night eating syndrome)	20 wks
Murray 2022	1	16	Adolescent	Male	NR	ARFID	21 sessions
Price 2025	1	26	Adult	Female	White	ARFID	23 sessions over 9 mos
Shepherd 2023	1	57	Adult	Female	White	BED	186 days (PHP: 15 wks, IOP: 11.5 wks)
Shepherd 2024	1	46	Adult	Woman	White	AN-BP	150 days (PHP: 104, IOP: 46)
Simpson 2011	1	39	Adult	Woman	NR	OSFED/EDNOS	7 sessions over 11 wks

		M (SD)	% by age group	% Female and/or % Cisgender Woman	% White	% by ED diagnosis	M (SD)
Cohort studies and case series
Baker 2024	1,235	Mode: 15–17	84.8% youth/adolescents 15.1% adults	78.9 CW	76	AN: 77.1% (AN-R: 60.2%, AN-BP: 15.6%, AN-unspecified: 1.3%) BN: 2.4% BED: 1.7% ARFID: 15.2% OSFED/EDNOS: 0.80% UFED: 2.8%	NR
Blalock 2025	231	28.03 (11.09)	6.5% adolescents 93.5% adults	84.8 F	75.6	AN: 23.8% (AN-R: 16.9%, AN-BP: 6.9%) BN: 6.5% BED: 19.5% ARFID: 3% OSFED/EDNOS: 47.2%	58.84 (26.69) days
Cai 2025	233	15.3 (2.7)	NR	85.8 CW	66.5	AN: 100%	46.8 (18.9) wks
Carr 2022	13	26.85 (6.94)	100% adults	92.3 F	92.3	AN: 100% (AN-R: 84.6%, AN-BP: 15.4%)	30.85 (13.97) wks
Drury 2025	77	15.19 (1.66)	100% adolescents	85.7 F 84.4 CW	57.1	AN: 58.4% (AN-R: 53.2%, AN-BP: 5.2%) ARFID: 2.6% OSFED/EDNOS: 39%	7.33 (6.42) mos
Franklin 2025	68	14.5 (1.6)	100% adolescents	83.8 F	27.9a	AN: 100% (AN-R: 72.1%, AN-BP: 27.9%)	NR
Hellner 2025	783	NRb	68% youth/adolescents 32% adults	53 CW	73	ARFID: 100%	Median: 35 wks
Levinson 2021	33	24.52 (9.27)	NR	90.91 F 87.88 CW	90.91	AN: 42.4% BN: 12.12% BED: 12.12% ARFID: 3.03% OSFED/EDNOS: 30.03%,	11.07 (6.3) wks
Ortiz 2023	46	19.38 (3.08)	NR	93.47 F	84.78	AN: 54.34% BN: 8.7% OSFED/EDNOS: 36.96%	7.34 (2.85) wks
Penwell 2024	70	23.1 (10.4)	35.7% youth/adolescents 64.3% adults	98.6 CW	82.9	AN: 47.2% (AN-R: 32.9%, AN-BP: 14.3%) BN: 11.4% BED: 10% ARFID: 1.4% OSFED/EDNOS: 30%	35 (15) days
Pereira 2023	10	14.27 (1.89)	100% youth/adolescents	100 F	NR	AN: 90% OSFED/EDNOS: 10%	NR
Perry 2024	130	14.3 (4.1)	NR	57.7 CW	70	ARFID: 100%	NR
Plumley 2021	9	30.33 (13.93)	NR	88.89 F	77.78	AN: 100% (AN-R: 88.88%, AN-BP: 11.11%)	NR
Raykos 2021	25	24.2 (7.62)	NR	93 F	70	AN: 48% BN: 20% OSFED/EDNOS: 28% UFED: 4%	1–12 sessions
Shepherd 2025	116	27.53 (12.24)c	31.9% adolescents 68.1% adultsc	81.9 CW	84.5	AN: 50.8% (AN-R: 37.9%, AN-BP: 12.9%) BN: 8.6% BED: 3.4% ARFID: 9.5% OSFED/EDNOS: 27.5%c	20.83 (11.46) wksc
Steiger 2022	76	28.41 (10.3)	100% adults	85.5 F	NR	AN: 42.1% BN: 27.6% ARFID: 2.6% OSFED/EDNOS: 27.6%	8.04 (2.0) wks
Steinberg 2023a	609	15.6 (2.29)	NR	85 CW	78	AN: 83.1% (AN-R: 66.2%, AN-BP: 16.3%) ARFID: 12.2% BN, BED, and OSFED: remaining 5%	Median: 23 (IQR: 14, 39) wks
Steinberg 2023b	210	16.1 (2.9)	NR	83 CW	71	AN: 80% (AN-R: 63%, AN-BP: 15%, AN-unspecified: 2%) BN: 1% BED: 2% ARFID: 14% OSFED/EDNOS: 2%	42.84 (0.18) sessions
Thaler 2024	54	30.50 (8.53)	100% adults	91 Fd	NR	AN: 54.7% (AN-R: 22.6%, AN-BP: 32.1%) BN: 15.1% ARFID: 1.9% OSFED/EDNOS: 28.3%	13.90 (2.05) wks
Van Huysse 2023	53	15.28 (2.26)	100% youth/adolescents	93 F 89 CW	94	AN: 67.9% (AN-R: 58.5%, AN-BP: 9.4%) BN: 1.9% ARFID: 1.9% OSFED/EDNOS: 28.3%	50.83 (13.83) days
Wilkes 2019	34	35.65 (13.73)	100% adults	100 F	82.4	AN: 14.7% BN: 35.3% BED: 8.8% OSFED/EDNOS: 17.6%	10.97 (6) wks

AN, anorexia nervosa; AN-BP, anorexia nervosa binge eating/purging type; AN-R, anorexia nervosa restricting type; ARFID, avoidant/restrictive food intake disorder; BED, binge eating disorder; BN, bulimia nervosa; CW, cisgender woman; EDNOS, eating disorders not otherwise specified; F, female; mos, months; OSFED, other specified feeding or eating disorder; NR, not reported; UFED, unspecified feeding or eating disorder; wks, weeks

^aNote that this represents the proportion of Non-Hispanic White patients in the study sample. The sample also included a high proportion of Hispanic patients, some of whom may also identify as White

^bReported for age subgroups, but not across the full sample

^cFull details not reported in manuscript, but included in supplementary materials

^dSex/gender information provided for the full sample, but not broken down by treatment condition (i.e., remote and in-person)

Beyond age, sex/gender, and race/ethnicity, demographic reporting was limited. Three studies reported on sexual orientation; one cohort study found that of the 52% who reported sexual orientation, approximately 45% identified as heterosexual [15], and two case study participants also identified as heterosexual [42, 50]. Only one article reported on religion or spiritual identity, with this case study participant identifying as Jewish [50]. Additionally, one cohort study described the educational status across their sample, most of whom had at least some higher education experience (94.1%) [57]. Two case reports also mentioned education level, with one participant holding advanced degrees [50] and the other having left school at the age of 16 [52]. Relatedly, three cohort/case series studies described their sample’s employment status, with most participants being either employed or students [46, 48, 57]. Additionally, two case study participants were employed [42, 47], one unemployed [50], and one retired [49]. One cohort study [46] and three case reports [47, 50, 52] also described their participants’ living situations, with nearly 90% of cohort participants and all case study participants living with their partner or family. Lastly, one cohort study reported financial status as follows: <$50,000 (23.5%), $50,001-$80,000 (35.3%), $80,001-$100,000 (14.7%), $100,001-$150,000 (11.8%), >$150,000 (14.7%) [57]. Meanwhile, two case reports noted that their participants had the necessary resources to participate in treatment and/or did not experience any financial barriers to care [49, 50].

Among clinical characteristics, ED diagnoses, comorbid diagnoses, and prior ED treatment history were most frequently reported. All articles described their sample’s ED diagnoses. Among cohort/case series, 62% (n = 13) were predominantly or exclusively anorexia nervosa (AN), with the restricting subtype (AN-R) being more prevalent than the binge/purge subtype (AN-BP) in all but one study. Many samples did have representation across other ED diagnoses, including bulimia nervosa (BN; 1–35.3.3%), binge eating disorder (BED; 1.7–19.5%), ARFID (1.4–100%), other specified feeding or eating disorder/eating disorder not otherwise specified (OSFED/EDNOS; 0.80–47.2%), and unspecified feeding or eating disorder (UFED; 2.8–4.8%). Furthermore, while one case report focused on AN [50], the remaining case reports focused on non-AN diagnoses, including ARFID [43, 47], BED [42, 49], and OSFED/EDNOS [42, 52]. Additionally, comorbid diagnoses were reported in 15 articles. Eight cohort studies reported the prevalence of common comorbidities in their remote [12, 36, 39, 40, 48, 51, 54] or full (i.e., in-person and remote) samples [15], with anxiety (20.5–82.8%) and mood disorders (22.1–68.1%) being most common. Two studies [12, 36] reported the mean number of comorbid diagnoses in their remote sample (1.23–2.23) while four [39, 46, 48, 51] reported the percentage diagnosed with any comorbidity (51.9–100%). A final cohort study stated that cisgender boys/men and transgender and gender expansive (TGE) patients were more likely to have comorbid attention-deficit/hyperactivity disorder than cisgender girls/women [35]. In addition, five of the six case studies mentioned comorbid diagnoses: two had none [43, 50], one had autism traits [47], one had suicidal ideation with major depressive disorder and post-traumatic stress disorder (PTSD) [49], and one had insomnia, a depressive disorder, and remitted PTSD [42]. Furthermore, 13 studies described their remote sample’s prior ED treatment history. Across four cohort studies, 33.33–84.8% of their samples had accessed any prior ED treatment [35, 39, 46, 54]. Five articles detailed proportions of their samples that had previously attended specific levels of ED care [35, 36, 39, 54, 56]. For instance, 24.7–70% had received prior outpatient treatment. One additional study reported that 41.4% of their sample had stepped down from a higher level of care [12]. Meanwhile, four case study participants [43, 47, 49, 52] had not accessed any prior ED treatment while one had one prior IOP stay [50] and one had tried a variety of prior outpatient treatments [42].

Some articles also reported on prior ED diagnosis, duration of the ED, age of ED onset, payment status, and duration of treatment, as recommended by consensus guidelines [25–27]. Prior ED diagnosis was mentioned in two studies, with both case reports suggesting undiagnosed BN in young adulthood [42, 50]. Two cohort/case series studies [46, 48] reported mean or median duration of illness (3–5.89 years) and four case studies [47, 49, 50, 52] reported age of onset (5–15 years). Regarding payment status, three cohort studies conveyed that 48.5–100% of their sample utilized health insurance [36, 40, 54], with one also specifying that 31% of their sample self-paid and 3% obtained care through other means [54]. Additionally, three case report participants utilized health insurance to obtain treatment [42, 49, 50], one being supported by employer reimbursement [42] and another self-paying for part of their stay [50]. Lastly, 22 studies reported duration of treatment. At the outpatient level, individual and mean/median lengths of stay ranged from 1 to 42.84 remote sessions and 8.04 to 46.8 weeks. For participants enrolled in day treatment programs, individual and mean durations of treatment ranged from 5 to 30.85 weeks.

In addition to describing the demographic and clinical characteristics, many comparison studies ran statistical analyses to determine if there were significant differences between remote and in-person samples in these variables, including age (n = 9) [12, 14, 16, 36, 38, 39, 44, 55, 56], ED diagnoses (n = 8) [12, 14, 16, 36, 38, 39, 55, 56], sex/gender (n = 7) [12, 14, 16, 36, 38, 39, 56], race/ethnicity (n = 6) [12, 14, 36, 38, 39, 56], ED treatment history (n = 3) [12, 39, 56], and comorbid diagnoses (n = 3) [12, 36, 39], as well as duration of treatment (n = 9) [12, 14–16, 36, 38, 39, 55, 56]. Across these variables, few significant differences were identified. Patients in one remote sample were significantly younger than their in-person counterparts [14], whereas patients in another remote sample were significantly older [36]. Additionally, one study noted their remote and in-person samples displayed significantly different proportions of ED diagnoses [36], while another remote sample comprised fewer patients diagnosed with ARFID than the comparative in-person sample [16]. Regarding treatment history, one study noted that more patients in their remote sample had stepped down from a higher level of care [12]. Another reported that a higher proportion of their remote patients had previously accessed any higher level of care, but a lower proportion had accessed prior outpatient treatment [56]. As for comorbid diagnoses, one study found that patients in their remote sample had a greater mean number of diagnoses than their in-person patients [36]. Finally, one study noted longer duration of treatment for their in-person sample [16] and two for their remote samples [55, 56]. No significant differences were found with respect to sex/gender or race/ethnicity.

What does the existing literature suggest regarding preliminary patient outcomes for telehealth-delivered, remote ED treatment?

In line with the purpose of a scoping review, preliminary findings from included studies are presented here without definitive conclusions about effectiveness as no formal risk of bias assessment and advanced data synthesis was conducted. Patients’ discharge statuses along with the core outcome domains outlined above (i.e., physical health, ED symptoms, co-occurring mental health conditions, QOL/functioning) will be discussed. Where reported, key details (e.g., measures utilized, admission and discharge values, effect sizes) are summarized in Table 4 (weight restoration and ED symptom outcomes) and Table 5 (co-occurring symptom and QOL outcomes). Of note, physical health indicators aside from weight restoration and co-occurring suicidal ideation are excluded from these tables as so few articles reported on these variables.

Table 4.

Weight restoration and ED symptom outcomes in remote treatment

Citation	Weight restoration						Global ED symptoms
	Sample	Sample size	Measure	Baseline	Discharge	Effect size	Measure	Baseline	Discharge	Effect size
Case reports
Melisse 2024a	NA	NA	NA	NA	NA	NA	EDE; EDE-Q	2.9; 3.6	1.4; 3.3	NR
Murray 2022a	NA	NA	NA	NA	NA	NA	PARDI-AR-Q ARFID Severity	3.5	1.5	NR
Price 2025a	NA	NA	NA	NA	NA	NA	Self-report	NR	NR	NR
Shepherd 2023a	NA	NA	NA	NA	NA	NA	EDE-Q	3.88	1.44	NR
Shepherd 2024	NA	1	Weight (lbs); BMI; %IBW	90; 15.94; 78.26%	112.66; 19.96; 97.97%	NR	EDE-Q	3.55	1.24	NR
Simpson 2011a	NA	NA	NA	NA	NA	NA	EDE-Q	4.49	1.8	NR

				M (SD)	M (SD)			M (SD)	M (SD)
Cohort studies and case series
Baker 2024	WR plan	NR	%EBWb	84% [credible interval: 83%, 84%]	NR	NR	EDE-QS	NRc	NR	NR
Blalock 2025	NR	NR	NR	NR	NR	NR	EDE-Q	3.42 (1.33)	1.92 (1.23)	L
Cai 2025	All patientsd	233	Weight (lbs); %EBWb	106.4 (18.0); 82.5% (8.5%)	NR	NR	EDE-QS	NR	NR	NR
Carr 2022	All patientse	13	BMI	17.15 (2.32)	17.64 (2.86)	NR	EDE-Q	4.62 (0.99)	3.66 (1.20)	NR
Drury 2025	All patients	75	%EBWb	88.15% (9.14%)	NR	NR	NR	NR	NR	NR
Franklin 2025	NR	NR	NR	NR	NR	NR	EDE-Q	NR	NR	NR
Hellner 2025	WR plan	413	%EBWb	NRc	NRc	NR	PARDI-AR-Q (abbreviated)	NRc	NRc	NR
Levinson 2021	All patients/AN patients	33/14	BMI	All patients: 25.24 (10.83)	All patients: 26.26 (10.39)	All patients: L/AN patients: M	EDE-Q	3.56 (1.42)	2.56 (1.14)	L
Ortiz 2023	WR plan	20	BMI	18.61 (1.42)	20.09 (1.50)	M	EDE-Q	3.17 (1.64)	1.69 (1.28)	M
Penwell 2024	AN patients	33	Weight; BMI	NR	NR	NR	EDE-Q	3.33 (1.40)	2.49 (1.39)	M
Pereira 2023	All patients	10	%EBWb	86.54% (2.62%)	NR	NR	NR	NR	NR	NR
Perry 2024	All patientsd	130	Weight (lbs); %EBW	92 (25.6); 84.6% (7.4%)	NR	NR	NR	NR	NR	NR
Plumley 2021	All patientse	9	BMI	15.38 (1.82)	18.36 (1.71)	L	EDE-Q	4.78 (1.02)	3.26 (1.47)	L
Raykos 2021	AN patients	12	BMI	NR	NR	NR	EDE-Q	NR	NR	NR
Shepherd 2025	AN patients	57	BMI; IBW	22.00 (6.25); NRc	NRc	L	EDE-Q	NRc	NRc	L
Steiger 2022	BMI cutoff	24	BMI	19.62 (3.3)f	NR	NR	EDE-Q	3.90 (1.4)	3.38 (1.4)	M
Steinberg 2023a	All patientsd	609	Weight (lbs); %EBWb	108.2 (21.5); NR	NR	NR	EDE-QS	15.6 [credible interval: 8, 22]	NR	NR
Steinberg 2023b	All patients/WR plan	210/163	Weight (lbs); %EBWb	108.7 (27.3); NR	NR	NR	EDE-QS	11.08 [credible interval: 6.93, 15.37]	NR	NR
Thaler 2024	BMI cutoff	29	BMI	17.18 (1.70)	18.48 (2.54)	L	EDE-Q	4.35 (0.98)	3.34 (1.51)	M
Van Huysse 2023	WR plan	41	%EBWb	84% (7%)	94% (7%)	NR	NR	NR	NR	NR
Wilkes 2019	All patients	34	BMI	26 (10.44)	26.50 (10.18)	S	EDE-Q	4.44 (1.03)	3.18 (1.45)	L

AN, anorexia nervosa; BMI, body mass index; EBW, estimated body weight; ED, eating disorder; EDE-Q, Eating Disorder Examination-Questionnaire; EDE-QS, Eating Disorder Examination-Questionnaire Short; IBW, ideal body weight; L, large; lbs, pounds; M, medium; NA, not applicable; NR, not reported; PARDI-AR-Q, Pica, ARFID, Rumination Disorder Interview–Questionnaire; S, small; WR, weight restoration

^aWeight restoration was not a treatment goal for the participants in these case studies

^bThough language varied across studies (e.g., target weight, estimated body weight), EBW is used to encompass any study in which an individualized estimated target weight was calculated after assessment by dietetic and/or medical staff, for instance based on age, growth history, median BMI, etc

^cReported for sub-groups of the sample (e.g., based on gender identity, age), but not across the full remote sample

^dAll patients in the sample were included, but the sample comprised patients who required weight restoration only

^eAll patients in the sample were included, but the sample comprised patients diagnosed with AN only

^fBMI was reported for the full remote sample, but not for those with BMI < 18 or < 20, which are the subgroups on which their weight restoration analyses were run

Table 5.

Co-occurring symptom and QOL outcomes in remote treatment

Citation	Depressive symptoms				Anxiety symptoms				QOL/functioning
	Measure	Baseline	Discharge	Effect size	Measure	Baseline	Discharge	Effect size	Measure	Baseline	Discharge	Effect size
Case reports
Melisse 2024	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR
Murray 2022	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR
Price 2025	PHQ-9	NR	NR	NR	GAD-7	NR	NR	NR	CIA	NR	NR	NR
Shepherd 2023	PHQ-9	19	5	NR	STAI	39 State, 61 Trait	21 State, 33 Trait	NR	EDQOL	1.64	0.32	NR
Shepherd 2024	PHQ-9	12	2	NR	STAI	54 State, 56 Trait	49 State, 44 Trait	NR	EDQOL	1.2	0.16	NR
Simpson 2011	SCL-90 Depression	1.69	0.46	NR	NR	NR	NR	NR	CORE- OM	1.29	0.18	NR

		M (SD)	M (SD)			M (SD)	M (SD)			M (SD)	M (SD)
Cohort studies and case series
Baker 2024	PHQ-9	NRa	NR	NR	GAD-7	NRa	NR	NR	NR	NR	NR	NR
Blalock 2025	PHQ-9	12.98 (5.87)	7.69 (5.33)	L	NR	NR	NR	NR	NR	NR	NR	NR
Cai 2025	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR
Carr 2022	NR	NR	NR	NR	NR	NR	NR	NR	WSAS	27.62 (8.40)	19.38 (9.54)	NR
Drury 2025	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR
Franklin 2025	PHQ-9	NR	NR	NR	GAD-7	NR	NR	NR	NR	NR	NR	NR
Hellner 2025	PHQ-8	NRa	NRa	NR	GAD-7	NRa	NRa	NR	NR	NR	NR	NR
Levinson 2021	BDI-II	26.16 (12.62)	20.13 (11.80)	M	NR	NR	NR	NR	NR	NR	NR	NR
Ortiz 2023	NR	NR	NR	NR	GAD-7	11.70 (6.21)	7.89 (4.81)	< S	CIA	27.83 (13.94)	16.35 (9.76)	< S
Penwell 2024	CES-D	28.26 (12.07)	24.24 (13.28)	S	OASIS	10.17 (3.83)	9.09 (4.00)	S	NR	NR	NR	NR
Pereira 2023	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR
Perry 2024	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR
Plumley 2021	PHQ-9	20.78 (5.14)	16.22 (7.12)	M	GAD-7	15.56 (5.57)	13.67 (4.77)	S	NR	NR	NR	NR
Raykos 2021	PROMIS-Depression	NR	NR	NR	PROMIS- Anxiety	NR	NR	NR	CIA	NR	NR	NR
Shepherd 2025	PHQ-9	NRa	NRa	L	STAI-Trait	NRa	NRa	L	EDQOL	NRa	NRa	L
Steiger 2022	PHQ-9	16.22 (7.4)	15.96 (6.5)	NR	GAD-7	14.64 (5.4)	14.18 (5.9)	NR	NR	NR	NR	NR
Steinberg 2023a	PHQ-9	12.9 [credible interval: 5, 20]	NR	NR	GAD-7	10.6 [credible interval: 5, 17]	NR	NR	NR	NR	NR	NR
Steinberg 2023b	PHQ-9	8.38 [credible interval: 5.43, 11.25]	NR	NR	GAD-7	10.28 [credible interval: 6.59, 13.77]	NR	NR	NR	NR	NR	NR
Thaler 2024	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR
Van Huysse 2023	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR
Wilkes 2019	BDI-II	24.53 (10.20)	17.96 (12.84)	S	BAI	26.68 (10.11)	21.93 (11.78)	S	NR	NR	NR	NR

BAI, Beck Anxiety Inventory; BDI-II, Beck Depression Inventory-II; CES-D, Centre for Epidemiologic Studies Depression Scale; CIA, Clinical Impairment Assessment; CORE-OM, Clinical Outcomes in Routine Evaluation Outcome Measure; EDQOL, Eating Disorder Quality of Life Instrument; GAD-7, Generalized Anxiety Disorder-7 Item; L, large; M, medium; NR, not reported; OASIS, Overall Anxiety Severity and Impairment Scale; PHQ-9, Patient Health Questionnaire-9; PROMIS, Patient-Reported Outcomes Measurement Information System; QOL, quality of life; S, small; SCL-90, Symptom Checklist-90; STAI, State-Trait Anxiety Inventory; WSAS, Work and Social Adjustment Scale

^aReported for sub-groups of the sample (e.g., based on gender identity, age), but not across the full remote sample

First, eight cohort studies described the discharge status of their participants. One article [12] did not provide specific percentages of discharge types; however, two reported the proportion of their patients who completed treatment [40, 51], two split their sample into two discharge categories (e.g., premature or not premature) [38, 46], and two provided a more detailed breakdown of discharge status [54, 57]. Across articles reporting frequencies, the percentage of patients that discharged treatment complete ranged from 4% in a sample where the majority of patients remained in treatment [54] to 69.8% [51]. Three articles also noted that they did not find significant differences between remote and in-person samples in discharge status [12, 38, 39].

Regarding physical health, all but one cohort study examining weight restoration saw improvements throughout remote treatment. Eight articles documented statistically significant increases [14–16, 37, 45, 51, 55, 56] while four reported improvements in weight restoration based on raw values and effect sizes only [12, 44, 46, 57]. Three studies reported increases in weight based on Bayesian statistics [35, 41, 53] and one appeared to combine frequentist and Bayesian approaches, reporting posterior summaries (e.g., credible intervals) alongside p-values that suggested significant improvement [54]. Reported effect sizes ranged from small [57] to medium [14, 15] to large [15, 46, 51, 55]. With respect to the rate of weight restoration, one article reported an average of 1.09 pounds gained per week [54]. Five others reported achievement of 95% of target weight, with outcomes ranging from 50% achieved by 20 weeks [37, 45] to 80% by 16 weeks [54], 56% by discharge [39], and 70% by 6 months post-treatment [44]. In contrast, one study found significant pre- to post-treatment weight increases for in-person patients’ but not remote [38]. However, this study saw no difference between remote and in-person treatment groups in mean change in BMI [38]. Similarly, five studies found no time by condition interaction [14–16, 55, 56] and two found no significant differences between remote and in-person groups in weight gained, percent of estimated target weight, and/or proportion that reached 95% of their target weight [39, 44]. Additionally, one study using visual inspection to assess patients’ trajectories across in-person and remote care found no evidence that the switch to a remote setting was associated with negative outcomes for weight restoration [48]. Conversely, one study did find that treatment type significantly predicted discharge BMI, with remote treatment associated with lesser change [12]. Among case reports, weight restoration was only relevant to one patient’s treatment goals [50]; this patient with AN-BP was able to successfully weight restore in remote treatment, with an average rate of weight gain of 1.06 pounds per week. Finally, with respect to physical health indicators aside from weight restoration, this case report and one other saw improvements across endocrine, gastrointestinal, and cardiac indicators where relevant [49, 50]. One cohort study reported the proportion of patients experiencing amenorrhea at admission (28.6%), but did not follow up on this at end of treatment [39].

With respect to ED symptoms, 15 cohort studies saw improvements throughout remote treatment, eight of which were statistically significant [14–16, 36, 38, 40, 51, 55], though one only saw significant improvements for treatment completers [40]. Three additional articles reported improvements in ED symptoms based solely on mean scores and effect sizes [12, 46, 57]. Three studies also noted decreases in ED symptoms based on Bayesian statistics [35, 41, 53] and another based on what appeared to again be a combination of Bayesian and frequentist statistics [54]. Reported effect sizes ranged from medium [12, 14, 16, 55] to large [15, 36, 46, 51, 57]. Additionally, one study included 6-month follow-up scores, which indicated a continued decline in average ED symptoms beyond discharge [12]. When comparing remote and in-person treatment, no differences in change scores from admission to discharge [36, 38] or time by condition interactions [14–16, 55] were identified, indicating that symptoms decreased similarly regardless of treatment type. Similarly, one study found that treatment type did not predict discharge ED symptoms except for binge eating, with remote treatment predicting lower binge eating frequency at discharge [12]. An additional study using visual inspection of patients’ trajectories again revealed no evidence that the switch to remote care was associated with negative outcomes [48]. All six case reports also reported improvements in ED symptoms. For ARFID cases, one self-reported increased meal frequency and variety of foods/drinks consumed throughout treatment [47] while the other showed 45–62% decreases in scores on the Food Neophobia Scale (FNS), Nine Item ARFID Screen (NIAS), and Pica, ARFID Rumination Interview-Questionnaire (PARDI-AR-Q), that were sustained at 2-month follow-up [43]. Additionally, two case reports saw reductions in global ED symptoms from admission to discharge as well as at follow-up (1- or 12-month) [49, 50], one demonstrating a clinically significant change from admission to discharge [49]. Another patient was abstinent from binge eating post-intervention and demonstrated improvements in ED symptoms based on both the EDE and EDE-Q, though their EDE-Q scores remained above the clinical cutoff [42]. The final case participant experienced a 77% decrease in severity of eating pathology from admission to discharge and continued improvement in symptoms by 1-month follow-up [52].

Regarding co-occurring mental health conditions, three studies found that patients in their remote cohort experienced statistically significant improvements from admission to discharge in depressive symptoms [15, 36, 51] and anxiety symptoms [14, 40, 51], though one study only saw improvements in anxiety symptoms for treatment completers [40]. An additional three articles described improvements in both domains based on average scores and effect sizes alone [12, 46, 57]. Post-treatment, one study included 6-month follow-up scores, which illustrated continued improvement in average depressive and anxiety symptoms [12]. Using Bayesian statistics, another four articles conveyed improvements in depressive and anxiety symptoms throughout treatment [35, 41, 53, 54], one of which again appeared to combine this approach with frequentist statistics [54]. Lastly, two studies also found reductions in patients’ suicidal ideation scores as per item nine on the PHQ-9 [35, 36]. Reported effect sizes for changes from admission to discharge ranged from small [12, 57] to medium [15, 46] to large [36, 51] for depressive symptoms and from less than small [14] to small [12, 46, 57] to large [51] for anxiety symptoms. One small effect size was reported for suicidality [36]. In contrast, two studies found no significant effect of time, indicating that depressive [16, 40] and/or anxiety symptoms [16] did not significantly improve during remote treatment. In comparing outcomes across remote and in-person samples, one study found that remote patients exhibited significantly greater improvements in depressive symptoms and suicidality as compared to in-person patients [36]. Meanwhile, two studies found that there was no time by condition interaction for depressive [15, 16] and anxiety symptoms [14, 16], suggesting symptoms changed similarly regardless of treatment type. One additional study reported that treatment type did not predict either depressive or anxiety symptoms at discharge [12]. Meanwhile, one study using visual inspection of patients’ trajectories revealed no adverse effect of transitioning to remote care for either symptom domain [48]. Finally, four of the six case reports evaluated changes in depressive symptoms and three in anxiety symptoms. For depressive symptoms, one patient remained in the subclinical range from start to end of treatment [47], one experienced a 73% reduction from admission to discharge [52], and two showed clinically significant change from admission to discharge and maintenance by follow-up (1- or 12-month) [49, 50]. For anxiety, one patient experienced a meaningful change in general anxiety, with scores no longer in the clinical range at discharge [47], and two demonstrated clinically significant change from admission to discharge in state and/or trait anxiety and maintenance by follow-up (1- or 12-month) [49, 50].

Among studies examining patients’ QOL/functioning, three cohort studies reported significant improvements throughout remote treatment [14, 38, 51]; two effect sizes were reported, one less than small [14] and one large [51]. No differences in mean change between remote and in-person groups [38] nor a time by condition interaction [14] was identified, suggesting the level of impairment decreased similarly regardless of treatment type. Another study saw no indication that the switch to remote treatment was associated with negative outcomes for their patients’ functioning [48]. Furthermore, four of the six case reports reported on this outcome. Two found decreases from admission to discharge and from discharge to follow-up (1- or 12-month) in total impact on QOL [49, 50], one showed an 86% reduction in global distress from admission to discharge [52], and one reported that a meaningful change was observed, with scores no longer in the clinical range at discharge [47].

What are the outcomes across different demographic groups?

There were only five studies that addressed how these preliminary patient outcomes of telehealth-delivered, remote ED treatment might differ across demographic status, four of which found no differences across groups in remote care. One study compared patient characteristics and treatment outcomes by gender identity among children, adolescents, and young adults receiving remote, enhanced outpatient FBT [35]. Despite differences in clinical presentation, cisgender girls/women, cisgender boys/men, and TGE patients improved throughout treatment on all outcomes (i.e., weight restoration; ED, depressive, and anxiety symptoms; suicidal ideation) at similar rates. Similarly, another study conducted in the same treatment program found no significant differences in ED symptom reduction among children and adolescents with respect to age, race, ethnicity, or gender [54]. In addition, another study identified no significant time by age interactions across adolescents, young adults, and mature adults, indicating that treatment response in their intentionally-remote PHP/IOP was similar regardless of age across all outcome domains (i.e., weight restoration; ED and depressive symptoms; trait anxiety; quality of life) [51]. Lastly, one study noted a moderating effect of age on the relationship between treatment format and treatment completion, with older age predicting lower odds of completing treatment among patients who received in-person FBT, but not for those who received remote FBT [39]. Age, however, did not moderate the impact of treatment format on weight restoration. In contrast, one study comparing outcomes in patients (ages 9–23) who attended in-person versus remote care in an integrative but largely FBT-informed higher-level-of-care ED treatment program found that age was a significant predictor of follow-up %EBW, with older patients experiencing lower %EBW at 6-months post-treatment [56].

Discussion

Main findings

This scoping review sought to summarize how patient outcomes for telehealth-delivered, remote ED treatment have been evaluated in the extant observational research in order to map the existing evidence, identify current knowledge gaps, and propose directions for future studies. This work is especially relevant given the aforementioned research-practice and treatment gaps, which reflect both limited understanding of how treatments perform in real-world settings and persistent barriers that prevent many individuals with EDs from accessing care. Despite growing interest in remote care models as a means of addressing these challenges, this body of literature remains relatively small with only 27 studies identified, six of which examined outcomes for a single patient only. Additionally, limitations and inconsistencies were identified across these studies, including differences in the treatment services themselves as well as disparities in study design, methodology, and reporting.

Variability in the treatment services evaluated by studies in this review allowed for preliminary findings to be assessed across the breadth of treatments offered in real world settings, however it created substantial challenges when comparing outcomes across studies, given the drastic differences in the amount and nature of services being delivered. Included studies evaluated remote ED treatment across a range of levels of care (i.e., slightly more than half were outpatient, with the remainder being day treatment) and therapeutic modalities (i.e., CBTs or CBTs-informed, FBT or FBT-informed, eclectic/integrative, schema therapy). The frequency, intensity, and duration of this treatment varied considerably across studies, from a single remote session [48] to 26.5 weeks of multi-hour care [49]. In addition, although many programs employed multidisciplinary care teams, the composition of these teams varied. Consistent with practice guidelines [58, 59], all teams included a therapist and all but one included a dietitian. However, some also included additional various types of medical staff, with their specific professions not always made clear. Relatedly, while all treatments consisted of individual, family, and/or group therapy and most offered nutrition therapy, less than half included psychiatric consultation or meal support, despite recommendations from practice guidelines to include these services, particularly at higher levels of care [58, 59]. Beyond the providers and interventions outlined by these guidelines, select programs also offered sessions with clinical support staff (e.g., peer mentors, care partners) to supplement their standard care teams and services. Lastly, despite the adoption of remote services, the use of other adjunctive technologies remains limited, with less than half of the articles reporting use of these tools in their treatment and those that did reporting relatively narrow functionality (e.g., email or telephone communications, messaging and progress monitoring in an online portal or app). Moving forward, treatments that are designed for remote environments (i.e., intentionally-remote) may have a greater capacity to leverage more innovative uses of technology, such as the remote medical monitoring devices described by Shepherd et al. [49–51] and Blalock et al. [36] used to obtain vitals and closed weights to improve data integrity in the remote setting. However, more work is needed to evaluate the impact of these and other emerging technologies (e.g., artificial intelligence), as well as the numerous other treatment attributes that varied across studies, on patient outcomes.

In addition to these differences in the treatment services evaluated, limitations and inconsistencies in study design and methodology were also noted. To begin, with regard to study location and time, studies were conducted exclusively in Western nations. Unfortunately, two articles conducted in Israel were excluded during data extraction due to altered demographic and clinical details [33, 34], though they could have provided valuable information about the effectiveness of remote ED treatment in a different geographic and cultural context. Additionally, nearly all included studies were published between 2021 and 2025 and, excluding case reports, nearly 60% of studies evaluated programs that were rapidly transitioned to remote delivery due to the COVID-19 pandemic, limiting long-term applicability.

Accordingly, it appeared that many studies were arranged without extensive planning or preparation in an effort to quickly evaluate the transition to remote care during the pandemic, resulting in limitations and lack of clarity related to study design and methodology. Similar methodological constraints were documented in a scoping review of the impacts of COVID-19 on EDs [60] as well as in a systematic review of broader COVID-19 clinical research [61]. These discrepancies and ambiguity were largely related to the studies’ samples. For example, it was rarely specified and frequently unclear if a study was conducted prospectively or retrospectively and therefore whether all patients enrolled in treatment were included in the study or only those who consented for research purposes. Furthermore, while all study designs had to allow for a meaningful comparison or differentiation of remote and in-person treatment outcomes in order to meet the review inclusion criteria, some studies’ distinctions were less clear, with participants having received both formats of care to some degree (e.g., examining within-subjects changes in treatment trajectory following a transition from in-person to remote care or comparing separate samples categorized by predominant mode of treatment even if some individuals received both formats). In addition, there were varying exclusion criteria enacted by different studies and treatment programs (e.g., exclusions based on diagnosis, age, weight/need for weight restoration, completion of outcome measures, level of care, length of stay, attendance/engagement, readmission status, etc.). Perhaps as a result, most samples were quite small. Taken together, the ambiguity and variability surrounding study samples raise questions about the generalizability of these studies’ findings and complicate cross-study comparisons.

There was also considerable variability among studies regarding what outcome domains were assessed and how they were operationalized. While most studies evaluated weight restoration and ED symptoms, over 30% did not examine any co-occurring mental health conditions and 70% did not assess QOL/functioning. Furthermore, studies examining comorbid symptoms focused on depressive symptoms, anxiety symptoms, and/or suicidal ideation, omitting other recommended domains, such as trauma or obsessive compulsive symptoms [26]. Additionally, only five studies considered physical health outcomes aside from weight restoration, all but one being case reports. Given published guidelines to measure these indicators [25, 26], further clarification may be needed as to how such outcomes should be summarized and reported in an aggregated manner. Of the outcomes examined, there were also inconsistencies in how they were operationalized. ED symptoms were most consistent, with all studies utilizing the EDE-Q, EDE-Q Short, or the PARDI-AR-Q for ARFID. However, measures used for other outcome domains were more varied; for example, across the 17 studies assessing depressive symptoms, six different measures were utilized. Published recommendations on ED outcomes assessment, including specific measures, were released during the same time period that most studies in this review were published and thus may not have been taken into consideration by these articles [25–27]. In the future, researchers should attend to these recommendations to enable greater consistency across studies.

In addition to summarizing how remote ED treatment has been evaluated thus far in the literature, this review also sought to examine who is utilizing these services in the real-world. While healthcare utilization can be used as a proxy for access, offering insights into remote ED treatment’s accessibility for historically excluded groups, it is influenced by other factors (e.g., individual preferences) [62] and should be interpreted with caution. Furthermore, given the numerous inclusion and exclusion criteria implemented to define study samples, the presented sample characteristics may not portray a representative picture of remote ED patient populations. These challenges are then further compounded by the inconsistencies in what demographic and clinical characteristics were reported by the included articles as well as their varied method of reporting this information. Several studies described the majority demographic group only (e.g., only reporting the percent identifying as White or female), omitting details about specific populations. Additionally, as noted above, terms such as sex and gender were frequently conflated, making cross-study comparisons difficult. Moreover, variables aside from age, sex/gender, race/ethnicity, and diagnoses were seldom reported. While demographic reporting in ED research is often centered on the majority group, incomplete, or entirely absent [63], these inconsistencies in reporting make it even more challenging to draw conclusions around service utilization for several of the groups most in need of improved access to care. The omission of variables such as sexual orientation and socioeconomic status are particularly concerning, considering evidence that sexual minorities experience elevated and distinct ED symptoms [64, 65] and that food insecurity and low income may be associated with increased disordered eating behaviors [66, 67]. Furthermore, in summarizing the demographic and clinical information that was reported, findings revealed that remote samples were not particularly diverse, nor representative of the population struggling with EDs. For instance, research shows that racial and ethnic minorities [68, 69] and TGE individuals [70, 71] may be at comparable if not higher ED risk relative to White, cisgender individuals, while prevalence for males may be nearly as high as in females [72, 73]. Yet, study samples were overwhelmingly composed of White, females/cisgender women, as is typical in ED research [63, 74]. In addition, there were no significant differences between remote and in-person samples in sex/gender nor race/ethnicity, with minimal differences in other demographic and clinical variables either, aside from a few exceptions in age [14, 36], ED and comorbid diagnoses [16, 36], treatment history [12, 56], and duration of treatment [16, 55, 56].

Collectively, the issues with representativeness among remote treatment samples and their minimal differences from in-person treatment samples may indicate that telehealth-delivered, remote ED treatment has not yet fulfilled its promise of improved access. Similar concerns have been expressed in relation to the accessibility and equity of digital mental health care more broadly [75–77]. These findings encourage further reflection around dimensions of accessibility of remote ED treatment, including approachability (e.g., patient outreach and recruitment), acceptability (e.g., inclusivity, stigma, cultural relevance), availability (e.g., hours of operation, scheduling), affordability (e.g., direct costs, insurance options), and appropriateness (e.g., quality and effectiveness of care) [62]. Given that many studies in this review evaluated programs that shifted to remote delivery during the COVID-19 transition, these samples might reflect patients who would have otherwise been obtaining in-person care. It is possible that programs that are intentionally designed for remote delivery and thus implement outreach and programming with enhanced access in mind may be able to reach a different set of patients. For instance, one qualitative exploration of remote ED treatment noted that specialized programming (e.g., gender- or diagnosis-specific groups) may help to foster patients’ sense of belonging, which may improve the acceptability of care, particularly for those who hold identities historically underrepresented in ED treatment [62, 78]. To further assess this, additional research evaluating the accessibility and effectiveness of intentionally-remote ED treatment programs is still needed.

Lastly, this scoping review also sought to summarize preliminary patient outcomes of telehealth-delivered, remote ED treatment and how those outcomes may differ by demographic status. Preliminary findings tentatively suggest that remote ED treatment can yield improvements across key outcome domains, with small to large effects, that appear to be largely comparable to those achieved in in-person settings. All studies examining ED symptoms and QOL/functioning reported improvements throughout remote care. For physical health, all but one study [38] saw improvements in weight restoration throughout remote treatment and in the two case studies that monitored other physical health indicators, both showed improvements. Similarly, for co-occurring mental health conditions, all but two studies [16, 40] saw improvements in this domain. Across all outcome domains, only one study saw worse outcomes in remote versus in-person care, with patients experiencing lesser change in BMI throughout remote treatment [12], perhaps warranting additional focus on weight restoration in remote settings in future work. Of note, the analyses conducted to examine outcomes and the methods of reporting vastly differed across the included studies. For example, many studies did not report average admission and discharge values or effect sizes, while others only presented raw values or effect sizes due to small samples. These additional inconsistencies further challenge cross-study comparisons and may impede efforts to synthesize findings in future systematic reviews or meta-analyses. In addition, only five of the 27 included studies examined how patient outcomes differed across demographic groups. Though not atypical in ED treatment research [63], this could in part stem from such analyses being underpowered given the small sample sizes and poor demographic representation outlined previously. Of these five articles, four concluded there were no differences in remote treatment outcomes by demographic status [35, 39, 51, 54], while the fifth noted that older patients were at a lower percent of their target weight by 6-month follow-up [56]. More research is needed to confirm these findings; particular attention should be paid to long-term outcomes, including across age groups. Until then, considering the lack of sample diversity and limited studies comparing outcomes across demographic groups, the outcomes described above should be interpreted with caution as there has not yet been sufficient evaluation to ensure remote treatments are effective across the diverse population afflicted by EDs.

Strengths and limitations

To our knowledge, this is the first review to map the available research on patient outcomes of real-world remote ED treatment, providing valuable guidance for future research in this burgeoning field. The review followed a rigorous methodological procedure, as informed by the JBI guidelines [28, 29], including thorough search, selection, and extraction processes. However, several limitations should be considered. As with any review, it is possible that relevant sources of information may have been omitted. For instance, articles published in languages other than English were not included due to feasibility. Moreover, the review’s rigor could have been strengthened by pre-registering or publishing the review protocol. Finally, while a formal critical appraisal of included studies was not conducted based on scoping review guidelines [28, 29], several limitations of the included articles have been outlined above, including the design and methodological weaknesses related to the abrupt COVID-19 transition to remote care, varying sample criteria, small sample sizes, poor demographic diversity and representation, differing outcome domains and operationalizations of those domains, and inconsistencies in terms of statistical analyses and reporting.

Future directions

Due to the limitations outlined above, it is difficult to draw definitive practical implications from this review regarding the remote treatment of EDs. Included studies largely indicate that remote ED treatment can yield improvements in core outcome domains across levels of care. However, the aforementioned differences in the treatment services evaluated alongside the methodological constraints of included studies call into question the tenability and generalizability of these results, especially given the inconsistent sample criteria and lack of sample diversity. To better understand the effectiveness of remote ED treatment, additional studies with more rigorous, intentional designs are needed. Additionally, further evaluation of intentionally-remote treatment programs is crucial, as their accessibility and effectiveness may differ from services that abruptly shifted to the remote setting without preparation. Research leveraging insurance claims databases to obtain more representative information regarding the utilization and accessibility of remote ED treatment would also be useful. The field would further benefit from more standardized data collection and reporting, including consistent use of appropriate measurement tools, clear definitions of assessed variables (e.g., biological sex vs. gender identity), uniform response options (e.g., sexual orientations, treatment history), and comprehensive reporting (e.g., demographic breakdowns beyond majority groups, average scores on self-report measures), some of which is addressed in the consensus guidelines informing this review [25–27]. Additionally, exploring long term outcomes would also be valuable as only seven studies in this review evaluated outcomes beyond discharge, four of which were case reports. If more diverse populations are reached, moderation analyses could also provide meaningful insight into how outcomes may differ according to demographic status as well as diagnosis. Finally, considering the differences in the treatment services evaluated, mediation analyses could offer useful information regarding which aspects of remote treatments are most contributing to symptom improvements. In sum, based on the methodological issues identified in this scoping review and the heterogeneous nature of the existing literature, a systematic review on this topic is not recommended at this time, though may be warranted soon if the recommended methodological changes are implemented. While scoping reviews are intentionally broad in focus [29], future systematic reviews should either be more narrow in scope or ensure the body of literature is large enough to segment out the research (e.g., by level of care or therapeutic modality) to account for the many differences in real-world remote ED treatments.

Conclusions

To our knowledge, the current review is the first to summarize patient outcomes of telehealth-delivered, remote ED treatment conducted in routine clinical settings. Overall, this body of literature remains small and characterized by limitations and inconsistencies, including differences in the treatment services themselves as well as disparities in study design, methodology, and reporting. Additionally, utilization of remote ED treatment by historically excluded groups remains low, calling for further reflection around its accessibility for target communities. Nevertheless, findings tentatively suggest that telehealth-delivered, remote ED treatment can be effective across key outcome domains, yielding outcomes that appear largely comparable to in-person services. However, this conclusion should be interpreted with caution given differences in treatments, study limitations, and poor demographic representation. Additional high-quality studies evaluating intentionally-remote treatment programs are needed, along with studies employing larger samples and examining long-term outcomes. The field would also benefit from standardization in relation to data collection and reporting to allow for better synthesis of findings. Ultimately, addressing critical research-practice and treatment gaps through rigorous real-world evaluation and expansion of care to underserved populations will be essential to realizing the full potential of telehealth-delivered, remote ED treatment.

Abbreviations

AN

Anorexia nervosa

AN-BP

Anorexia nervosa-binge/purge subtype

AN-R

Anorexia nervosa-restricting subtype

ARFID

Avoidant/restrictive food intake disorder

BED

Binge eating disorder

BMI

Body mass index

BN

Bulimia nervosa

CBTs

Cognitive behavioral therapies

CIA

Clinical Impairment Assessment

EBW

Expected body weight

ED

Eating disorder

EDE-Q

Eating Disorder Examination-Questionnaire

EDNOS

Eating disorder not otherwise specified

EDQOL

Eating Disorder Quality of Life Instrument

FBT

Family-based treatment

FNS

Food Neophobia Scale

GAD-7

Generalized Anxiety Disorder-7

IBW

Ideal body weight

IOP

Intensive outpatient program

JBI

Joanna Briggs Institute

MeSH

Medical Subject Headings

NIAS

Nine Item ARFID Screen

OSFED

Other specified feeding or eating disorder

PARDI-AR-Q

Pica, ARFID, Rumination Disorder Interview-Questionnaire

PHP

Partial hospitalization program

PHQ-9

Patient Health Questionnaire-9

PRISMA-ScR

Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews

PROM

Patient-reported outcome measure

PTSD

Post-traumatic stress disorder

QOL

Quality of life

TGE

Transgender and gender expansive

UFED

Unspecified feeding or eating disorder

Author contributions

HW and CBS conceptualized the project with input from RL. HW managed and coordinated the project. WOP was responsible for funding and resource acquisition and provided general supervision. HW, CBS, and RL conducted the search and screening process. HW and RL completed data extraction with consultation from CBS. HW wrote and prepared the initial manuscript. All authors read and edited the manuscript and approved the final version.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Data availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

HW and RL are employees of Within Health Provider Services FL PLLC. CBS is an employee of Within Health Provider Services FL PLLC and is a consultant to and holds stock options issued by Within Health Group, Inc., an affiliate of Within Health Provider Services FL PLLC. WOP is the sole owner and President of Within Health Provider Services FL PLLC and is a co-founder, co-owner, and the Chief Executive Officer of Within Health Group, Inc.