Patient Engagement

Specific details of engagement activities were not delineated in the protocol. Further details about participant selection and engagement activities are described in Patient Engagement Methods below.

7

The protocol did not address the specific patient engagement activities that would be conducted, therefore further details are supplied in Patient Engagement Methods below.

Clinical Effectiveness and Clinical Harms

Rather than having 2 reviewers conduct the clinical review (i.e., data extraction, critical appraisal, data analysis), a single reviewer was responsible for the clinical review thus altering the study design from a systematic review to a rapid review, except for study selection which involved 2 reviewers agreeing on their decisions to include or exclude each study screened. With this change, the literature search methods were also streamlined, updating the database searches monthly (initial search conducted on May 24, 2023 and last alert completed on August 24, 2023) but not the grey literature search (conducted once from May 25 to June 5, 2023).

13,14, 15, 19 to 22

The study design and approaches to data extraction, critical appraisal, and data analysis was modified due to feasibility and resourcing constraints.

No attempt was made to quantitatively synthesize the data from the findings via meta-analyses.

21

The data from the findings was deemed too heterogenous to appropriately pool and provide a quantitative synthesis.

Rather than posting a list of studies selected for inclusion on the CADTH website for broad feedback, the list was sent to a group of select external stakeholders for targeted feedback.

19, 32

The targeted feedback approach was used due to feasibility and resourcing constraints during the data selection phase.

Outcome-level risk of bias assessment for the critical appraisal was not done. Instead, an overall assessment of study risk of bias from the domain level was used to inform the critical appraisal of included studies.

20

This change is in line with the approach used in CADTH’s rapid reviews, which was used to guide the clinical review.

Health Economics: Health care resource implications

Rather than consulting with program administrators and clinical experts, CADTH identified the health care resources needed for implementing and running an early intervention program for eating disorders through a review of the literature. This included a grey literature search for existing programs in Canada and review of their descriptions, as well as a review of relevant articles that were identified via the clinical and economic reviews for descriptions of the components of the interventions assessed within those studies.

17

The approach to identifying the resources needed to implement or run an early intervention program for eating disorders was modified due to feasibility concerns and to avoid potential delays to obtaining the information.

Social and Ethical Dimensions

This section of the project was removed.

25 to 32

This change was due to resourcing constraints.

Richards et al., (2023)¹

Pre-post cohort study

UK

Academic Health Science Network National Programme; Health Foundation; NIHR Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King's College London; NIHR Senior Investigator Award; NHS Innovation Accelerator Fellowship

Inclusion criteria: Participants aged 16 to 25 with an ED diagnosis of < 3 years duration

Participant characteristics:

Sample size:

• FREED-4-All cohort, n = 2473

• FREED-Up cohort, n = 278

Age, mean (SD):

• FREED-4-All = 19.87 (2.29)

• FREED-Up = 20.19 (2.39)

Gender: NR

ED Diagnosis, % (n):

• FREED-4-All (n = 1779)^a

  • AN = 46% (819)

  • BN = 25% (450)

  • BED = 4% (67)

  • ARFID = 1% (22)

  • OSFED = 24% (421)

• FREED-Up (n = 278)

  • AN = 35% (96)

  • BN = 27% (75)

  • BED = 1% (3)

  • ARFID = 0% (0)

  • OSFED = 37% (104)

DUED, mean (SD):

• FREED-4-All = 14.86 (9.73)

• FREED-Up = 17.85 (10.38)

Intervention:

FREED service model

• FREED-4-All cohort represents the most recent cohort of FREED participants.

• FREED-Up cohort represents a past cohort of FREED participants included in a multi-site study.

Comparator: NA (single-arm pre-post analysis on FREED-4-All and FREED-Up cohorts)

• Adherence to wait time targets

• ED symptomology (EDE-Q)

• Binge eating, vomiting, laxative episodes (behavioural items from EDE-Q)

• Change in BMI

• Psychological distress (CORE-10/OM)

• FREED-4-All cohort: changes between pre-treatment and post-treatment (over unspecified duration)

• FREED-Up cohort: changes between baseline to 3-, 6-, and 12-month follow-up

Austin et al., (2022)²

Retrospective cohort study

UK

Health Foundation

Participant data was extracted from the FREED-Up study (see Flynn et al., [2020] for inclusion/exclusion criteria and participant characteristics)

Intervention:

FREED service model^a

Comparator: TAU cohort^b

• ED symptomology (EDE-Q)

• Psychological distress (CORE-10)

• Psychological impairment due to ED (CIA)

• Change in mood (DASS-21)

• Functional impairment due to ED (WSAS)

• Perception of emotion for caregiver or partner (LEE)

• Function and wellbeing (PSYCHLOPS)

• Change in BMI

• Baseline to 3-, 6-, and 12-month follow-up

Radunz et al., (2021)³

Single-arm pre-post cohort study

Australia

Funding: NR

Inclusion criteria:

Participants aged 16 to 25 with ED symptoms for < 3 years who accessed treatment in one of two clinics servicing South Australia (n = 96)

Participant characteristics:

Age, M (SD); min, max = 19.3 (2.39); 16, 26

Gender (female), % = 92%

Intervention:

Early intervention services for ED in “emerge-ED” program which provides tailored treatment (e.g., CBT) to service users within pre-specified wait time targets

Comparator: NA (single-arm pre-post intervention analysis)

• ED cognitions and behaviours (ED-15)

• ED symptomology (EDE-Q)

• Psychosocial impairment (CIA)

• Depression, anxiety and stress (DASS-21)

• Change in BMI

Baseline to end of treatment (approximately 6 months in duration)

Richards et al., (2021)⁴

Pre-Post cohort study

UK

Shine and Scaling Up Improvement Award from the Health Foundation (GIFTS 7294/CRM 1216); PhD studentship from the Health Foundation; King’s College London International

Postgraduate Research Scholarships; NHS Innovation Accelerator Fellowship; NIHR Biomedical Research Centre for Mental Health, South London and Maudsley NHS Foundation Trust and Institute of Psychiatry, Psychology and Neuroscience at King’s College London; NIHR Senior Investigator Award

Participant data was extracted from the FREED-Up study (see Flynn et al., [2020] for inclusion/exclusion criteria)

Participant characteristics from FREED cohort included in the FREED-Up study (n analyzed = 259)^c:

Age, M (SD) = 20.19 (2.34)

Gender, female:male = 241:18

Ethnicity, n (%):

• White = 170 (66%)

• Asian = 25 (10%)

• Black = 10 (45%)

• Mixed = 19 (7%)

• Other/unknown = 35 (14%)

Baseline EDE-Q score, M (SD) = 4.06 (1.23)

Intervention:

FREED service model^a

Comparator: TAU cohort^b

Program fidelity (adherence to wait times)

NA

Flynn et al., (2020)⁵

Pre-Post cohort study

UK

Health Foundation; Scaling Up Improvement Award

Inclusion criteria:

• FREED cohort (from which FREED-Up cohort [i.e., past FREED participants included in a multi-site study] was derived): participants aged 16 to 25 who had a primary diagnosis of ED and < 3 years duration of illness

• TAU cohort: participants aged 16 to 25 with an ED illness duration of < 3 years who accessed ED services approximately 1.5 to 2 years before the implementation of FREED

Exclusion criteria:

Participants in need of immediate in-participant admission, a primary comorbid physical or mental disorder, severe intellectual disability, and insufficient English language to complete study procedures

Participant characteristics:

Sample size

• FREED-Up cohort (n = 278)

• TAU cohort (n = 224)

Age, M (SD):

• FREED-Up = 20.19 (2.39)

• TAU = 20.28 (2.43)

Sex, female:male:

• FREED-Up = 259:19

• TAU = 216:8

ED diagnosis, n (%):

• FREED-UP

  • AN = 117 (42.1%)

  • BN = 71 (25.9%)

  • BED = 3 (1.1%)

  • OSFED = 86 (30.9%)

• TAU

  • AN = 116 (51.8%)

  • BN = 59 (26.3%)

  • BED = 6 (2.7%)

  • OSFED = 44 (19.6%)

Ethnicity, n (%):

• FREED-Up

  • White = 181 (65.1%)

  • Asian = 27 (9.7%)

  • Black = 11 (4.0%)

  • Mixed = 20 (7.2%)

  • Unknown = 39 (14.1%)

• TAU

  • White = 174 (77.7%)

  • Asian = 21 (9.4%)

  • Black = 5 (2.2%)

  • Mixed = 7 (3.1%)

  • Unknown = 17 (7.6%)

Intervention:

FREED service model^a

Comparator: TAU cohort^b

• ED onset, duration, frequency, and severity (DUSC, DUED)

• Wait times (weeks)

• Treatment uptake

NA

Fukutomi et al., (2019)⁶

Pre-Post cohort study

UK

NIHR; The Health Foundation

Participant data was extracted from the FREED pilot study (see McClelland et al., [2018] for inclusion/exclusion criteria) but only included participants diagnosed with AN

Participant characteristics:

Sample size:

• FREED-AN cohort (n = 22)

• TAU-AN cohort (n = 35)

Age (combined), M = 20.4

Intervention:

FREED service model^a

Comparator: TAU cohort^b

• 24-month service utilization

• Last measured BMI

Baseline to 24-month follow-up

McClelland et al., (2018)⁷

Pre-Post cohort study

UK

NIHR Health Foundation

FREED cohort (from which FREED pilot cohort [i.e., past FREED participants included in a multi-site study] was derived):

Inclusion criteria:

• FREED cohort = participants aged 18 to 25 with a primary ED diagnosis and < 3 year duration of illness

• TAU cohort = participants aged 18 to 25 with an ED illness duration of < 3 years who accessed ED services 2 years before the implementation of FREED

Exclusion criteria:

Participants in need of immediate in-participant admission, a primary comorbid physical or mental disorder, inability to participant for 12-month duration of study, and insufficient English language to complete study procedures

Participant characteristics:

Sample size:

• FREED cohort (n = 56)

• TAU cohort (n = 86)

Age at referral, M (SD):

• FREED = 20.4 (2.24)

• TAU = 20.4 (2.0)

Age of illness onset, M (SD):

• FREED = 19.3 (2.6)

• TAU = 19.3 (2.1)

Gender, female, n (%):

• FREED = 54 (96%)

• TAU = 85 (98%)

Diagnosis, n (%):

• FREED:

  • AN = 22 (35%)

  • BN = 18 (32%)

  • BED = 1 (2%)

  • OSFED = 15 (27%)

• TAU:

  • AN = 35 (40%)

  • BN = 24 (28%)

  • BED = 4 (5%)

  • OSFED = 23 (27%)

Intervention:

FREED service model^a

Comparator: TAU cohort^b

• Wait times (weeks)

• Treatment uptake

• Change in BMI

• ED symptomology (EDE-Q)

• Change in mood (DASS-21)

• Psychological impairment due to ED (CIA)

• Perception of emotion for caregiver or partner (LEE)

• Psychological distress (CORE-10)

• Work and social adjustment impairment

Baseline to 3-, 6-, and 12-month follow-up

Brown et al., (2016)⁸

Pre-Post cohort study

UK

Shine award from the Health Foundation); NIHR Biomedical Research Centre for Mental Health, SLaM and Institute of Psychiatry, Psychology and Neuroscience, King’s College London

Inclusion criteria:

• FREED cohort = participants aged 18 to 25 with a primary ED diagnosis and < 3 year duration of illness

• TAU cohort = participants aged 18 to 25 with an ED illness duration of < 3 years who accessed ED services 2 years before the implementation of FREED

Exclusion criteria:

Participants in need of immediate in-participant admission, a primary comorbid physical or mental disorder, severe learning disability

Participant characteristics:

Sample size:

• FREED cohort (n = 51)

• TAU cohort (n = 89)

Age, M (SD):

• FREED = 20.64 (2.52)

• TAU = 20.47 (1.99)

Gender, female, %:

• FREED = 49:2

• TAU = 87:2

Diagnosis, n (%):

• FREED

  • AN = 20 (39.2%)

  • BN = 17 (33.3%)

  • OSFED = 14 (27.5%)

• TAU

  • AN = 33 (37.9%)

  • BN = 25 (28.1%)

  • BED = 4 (4.5%)

  • OSFED = 25 (28.1%)

  • No ED = 2 (0.02%)

Intervention:

FREED service model^a

Comparator: TAU cohort^b

• DUSC

• DUED

• Wait times (weeks)

• Treatment uptake

NA

Godart et al., (2022)⁹

Long-term follow-up analysis of an RCT

France

Projet Hospitalier de Recherche Clinique (CRC-PHRC, 1997, AOM97133 APHP, French Ministry of Health), the Caisse Nationale d’Assurance Maladie des Travailleurs Salaries (CNAMTS), and the Fondation de France

See Godart et al., (2012) for inclusion/exclusion criteria and participant characteristics

Intervention:

Systematic family therapy in combination with a multidisciplinary outpatient care program

Comparator:

TAU multidisciplinary outpatient care program

• Change in BMI

• AN clinical functioning (GOAS)

• ED psychological and behavioural traits (EDI)

• Psychological distress and/or psychological status (SCL-90-R)

• Family adaptability and cohesion (FACES III)

Baseline to 6-, 12-, 18-, and 54-month follow-up

Herpertz-Dahlmann et al., (2021)¹⁰

Pre-post cohort study

Germany

Ministry of Labour, Health and Social Policies of the State of North-Rhine-Westphalia, Germany; Open access funding enabled and organized by Projekt DEAL

Inclusion criteria:

Participants between the ages of 12 and 18 with a diagnosis of AN (or atypical AN) during their first or second admission for AN with at least 1 carer

Exclusion criteria:

Anyone with organic brain disease or other severe psychiatric disorders, substance abuse, severe self-injurious behaviour, low intelligence, severe comorbid somatic disorder, inability to speak German, or planned residential treatment

Participant characteristics:

Sample size:

• Home treatment cohort (n = 22)

• Non-home treatment cohort (n = 10)

Age, M (SD); Min, Max:

• Home treatment = 15.06 (1.15); 13.17, 17.03

• Non-home treatment = 16.33 (1.13); 14.69, 17.90

Gender, female, n (%):

• Home treatment = 22 (100%)

• Non-home treatment = 10 (100%)

AN subtype diagnosis, n (%):

• Home treatment restrictive = 22 (100%)

• Non-home treatment restrictive = 10 (100%)

• Home treatment atypical AN = 3 (13.6%)

• Non-home treatment atypical AN = 1 (10%)

Duration of illness in weeks, M (SD); Min, Max:

• Home treatment = 50.82 (30.75); 3.57, 111.57

• Non-home treatment = 54.93 (30.77); 4.86, 100.14

Psychiatric comorbidities, n (%):

• Home treatment:

  • At least 1 comorbidity = 18 (81.8%)

  • Affective disorder = 17 (77.3%)

  • Anxiety disorder = 10 (45.5%)

  • OCD = 0

  • Other = 3 (13.6%)

• Non-home treatment:

  • At least 1 comorbidity = 9 (90%)

  • Affective disorder = 10 (100%)

  • Anxiety disorder = 6 (60%)

  • OCD = 5 (50%)

  • Other = 1 (10%)

Intervention:

Home-based treatment post inpatient treatment which included an individualized treatment plan and multidisciplinary methods of therapy delivery

Comparators:

Change in clinical outcome at the beginning of treatment to end of treatment; non-home-based treatment participants were used to compare for categorical variables

• Change in BMI

• ED-specific psychopathology (EDE; EDI)

• AN clinical functioning (MRAOS)

• Comorbid psychiatric disorder (Mini-International Neuropsychiatric Interview for Children and Adolescents)

• Depressive symptoms (BDI)

• Health-related quality of life (Kidscreen-27)

• Treatment satisfaction (ZUF-8 [CSQ-8])

Start of treatment to end of treatment and 1-year follow-up

Coelho et al., (2019)¹¹

Single-arm pre-post cohort study

Canada

British Columbia Mental Health and Substance Use Services

Inclusion criteria:

Participants with a duration of illness of < 3 years admitted to FBT outpatient ED program with a diagnosis of AN or OSFED

Participant characteristics (n = 62):

Age, M (SD); min, max = 14.6 (2.1); 9, 18

Gender (female), n (%) = 58 (93.5)

Diagnosis, n (%):

• AN restrictive subtype = 49 (79%)

• AN binge/purge subtype = 2 (3.2%)

• OSFED restrictive subtype = 10 (16.1%)

• OSFED purge subtype = 1 (1.6%)

Ethnicity, n (%):

• Caucasian = 28 (45.2%)

• Asian = 7 (11.2%)

• Mixed background = 1 (1.6%)

• Not available = 26 (41.9%)

Psychiatric comorbidities, n (%):

• MDD = 7 (11.3%)

• GAD = 8 (12.9%)

• SAD = 3 (4.8%)

• OCD = 2 (3.2%)

• Other anxiety disorder = 14 (22.6%)

Intervention:

Family-based therapy

Comparator:

NA (Pre-post intervention analysis)

• Change in BMI

• Treatment progression

Beginning of treatment to end of treatment (over unspecified duration)

Hurst et al., (2019)¹²

Single-arm prospective cohort study

Australia

Funding: none

Inclusion criteria:

Participants aged 12 to 17 diagnosed with AN with an illness duration of < 3 years and referred to a specialist outpatient child and adolescent ED service

Participant characteristics:

Age, M (SD)  = 14.9 (1.2)

Intervention

Family-based therapy in combination with cognitive behavioural therapy focusing on perfectionism

Comparator

NA (Pre-post intervention analysis)

• ED symptomology (EDI-3)

• ED psychopathology and behaviour (EDE-Q)

• Perfectionism (CAPS)

• Expected body weight

• Outcomes were measured at 4 phases: after FBT commencement [T1]; FBT phase 2 and CBT commencement [T2]; completion of CBT [T3]; and completion of FBT and CBT [T4] (all over unspecified duration)

Rosling et al., (2016)¹³

Single-arm pre-post cohort study

Sweden

Crown Princess Lovisa’s Fund for Child Health Care; the Gillbergska Foundation; the First of May Flower Annual Campaign; Professor Bror Gadelius Memorial Foundation; the Sven Jerring Foundation; and Uppsala University

Inclusion criteria:

Adolescent females aged 10 to 17.9 from Uppsala County who were referred for assessment to the Eating Disorder Unit

Relevant participant characteristics:

Sample size: AN cohort (n = 31)

Age, M (SD) = 15.1 (2.0)

DUED (months), M (SD); range = 9.1 (7.3); < 1 to 32

Intervention:

Outpatient family-based therapy program

Comparator:

NA (Pre-post intervention analysis)

• ED symptomology (EDI-C)

• Depressive symptoms (MADRS-S)

• AN clinical functioning (MRAOS)

Baseline to 1-year follow-up

Godard et al., (2012)¹⁴

RCT

France

Projet Hospitalier de Recherche Clinique (CRC- PHRC, 1997, AOM97133 AP-HP), French Ministry of Health

Inclusion criteria:

Female participants ages 13 to 21 with a diagnosis of AN and < 3 years duration of illness

Exclusion criteria:

Inability to speak French or understand interview questions, any metabolic pathology interfering with eating or digestion, any psychotic disorder

Participant characteristics:

Sample size:

• Family therapy cohort (n = 30)

• TAU cohort (n = 30)

Age of illness onset, M (SD):

• Family therapy cohort = 14.7 (1.7)

• TAU cohort = 15 (1.5)

Age at study inclusion, M (SD):

• Family therapy cohort = 16.4 (1.7)

• TAU cohort = 16.6 (1.7)

Duration of illness in months, M (SD):

• Family therapy cohort = 17.1 (8.3)

• TAU cohort = 16.1 (5.2)

Intervention:

Systematic family therapy in combination with a multidisciplinary outpatient care program

Comparator:

TAU multidisciplinary outpatient care program

• Change in BMI

• Menstrual status

• Contraceptive use

• Number of hospitalizations

• AN clinical functioning (GOAS)

• ED psychological and behavioural traits (EDI)

• Social adjustment (SAS)

Baseline to 6-, 12-, and 18-months follow-up

ED symptomology

EDE-Q

The EDE-Q is a 28-item self-report questionnaire designed to assess the range, frequency, and severity of behaviours associated with an ED.¹⁵ Users are assessed on 4 subscales including restraint, eating concern, shape concern, and weight concern. Each subscale is scored as an average between 0 and 6, with higher scores indicating greater frequency or severity of eating disorder psychopathology over the previous 28 days.¹⁶ An overall global score ranging from 0 and 6 is assigned by summing the four subscale scores and diving by the number of subscales (i.e., 4), with a higher score indicating more problematic eating outcomes.¹⁵

EDE-Q global score’s clinically significant cut-off in populations including people living with an ED diagnosis = ≥ 2.17 to 3.19^{17 to 19,a}

ED-15

The ED-15 is a 15-item self-report questionnaire to assess eating attitudes and behaviours.²⁰ The questionnaire scores the frequency of 10 attitudes over the preceding week using a 7-point Likert scale, ranging from 0 (not at all) to 6 (all the time).²¹ An overall attitudinal score between 0 and 6 is assigned using the mean of the scores on all 10 attitudinal items.²¹ The questionnaire also includes 5 questions related to the frequency of problematic eating behaviours in the previous week (i.e., binge eating, vomiting episodes, laxative misuse, eating restraint, and excessive exercise), scored as the number of times or the number of days each behaviour occurred.²¹

No information

EDI

The EDI is a standardized, 64-item, self-report questionnaire that assesses a broad range of behavioural and attitudinal characteristics associated with EDs.²² The EDI consists of 8 subscales measuring: drive for thinness, bulimia, body dissatisfaction, ineffectiveness, perfectionism, interpersonal distrust, interoceptive awareness, and maturity fears.²³ Each item is rated as occurring always, usually, often, sometimes, rarely, or never. Responses to each item are assigned a score between from 0 to 3.²³ Subscale scores are calculated by summing scores from each item within the subscale, with higher scores indicating increased frequency of cognitive and behavioural characteristics associated with EDs.²³

No information

DUED

DUED refers to the length of time (often reported in months or years) between when an individual developed an ED and when they first initiated evidence-based treatment.⁵

No information

DUSC

DUSC refers to the length of time (often reported in months or years) between when an individual developed an ED and the date of specialist clinical assessment.⁵

No information

BMI and menstrual outcomes

BMI score

BMI is a value derived from the mass and height of an individual. It is calculated by dividing a person’s weight in kilograms by their height in metres. The Canadian Guidelines for Body Weight Classification in Adults²⁴ assigns four categories of BMI ranges in adults:

• underweight (BMI less than 18.5 kg/m²)

• normal weight (BMI from 18.5 kg/m² to 24.9 kg/m²)

• overweight (BMI from 25 kg/m² to 29.9 kg/m²)

• obese (BMI 30 kg/m² and over)

In children, it is not feasible to categorize individuals into categories based on absolute BMI thresholds because most anthropometric measures vary by age and sex.²⁵

Not applicable

EBW

%EBW is a measure of an individual’s BMI relative to a typical person of their age. It is calculated by dividing an individual’s BMI by the median age-adjusted BMI and multiplying by 100.¹⁰ A value greater than 100% indicates the individual has a BMI higher than the median age-adjusted BMI, while values lower than 100% indicate the individual has a BMI lower than the median age-adjusted BMI.¹⁰

No information identified

MROAS

The MROAS is a guided interview that identifies clinical features central to the syndrome of anorexia nervosa.²⁶ It consists of 5 domains: including: food intake and nutritional status, menstrual state, mental state, psychosexual adjustment, and socioeconomic status.²⁶ A score from 0 to 12 is determined for each domain depending on the individual’s responses, with higher scores indicating better clinical status.^10,26 A final average score is determined by calculating the average score across the 5 domains.²⁶

No information identified

MROC

MROC is used to classify an individual’s outcome following treatment as good, intermediate, or poor. In Godart et al. (2022),⁹ categories were defined as:

• good (BMI equal to or higher than the 10th percentile and regular menstruation)

• intermediate (BMI greater than the 10th percentile but amenorrhea (i.e., the absence of menstruation for at least the past three months)

• poor (BMI less than 10th percentile or presence of bulimic symptoms)

No information identified

Psychological impact

CORE-10/OM

The CORE-OM is a 34-item self-report instrument that includes 4 subscales designed to assess subjective well-being, symptoms, function, and risk.²⁷ The frequency that each item has occurred over the previous week is scored on a 5-point Likert scale between 0 (not at all) and 4 (most or all the time). A total raw score can be calculated by summing the scores for each item (ranging from 0 to 136), and scores for each subscale can be calculated by adding the value assigned to each item within the domain.²⁸ The average response for an individual (ranging from 0 to 4) can be calculated by dividing the total raw score by the number of items (i.e., 34).²⁸ Higher scores indicate higher psychological distress.²⁷

The CORE-10 is shortened version of the CORE-OM that includes 10 items.²⁹ The frequency that each item has occurred over the previous week is scored on a 5-point Likert scale between 0 (not at all) and 4 (most or all the time).²⁹ Total scores range between 0 and 40 and are calculated using the sum of the scores for each item.²⁹ Higher CORE-10 scores indicate higher level of general psychological distress, with a total score of 11 or above being clinically significant.³⁰ Average scores (ranging from 0 to 4) can also be calculated by diving the total score by the number of items (i.e., 10).²⁹

CORE-OM clinically significant cut-off points for men (M) and women (W) for unspecified clinical and non-clinical populations:^b

• Mean item score = 1.19 (M); 1.29 (W)³¹

CIA

The CIA is a 16-item self-report questionnaire to assess severity of psychosocial impairment due to ED.^32,33 It includes 3 subscales: personal impairment (6 items), social impairment (5 items), and cognitive impairment (5 items). Each item is rated on a 4-point Likert scale (0 = not at all; 3 = a lot) that reflects how often the item has occurred in the past month.³² A global score (ranging from 0 to 48) is calculated by adding the values for each item, with higher values indicating higher levels of psychosocial impairment.³² A global score of 16 represents clinically significant impairment.³³

No information identified

DASS-21

The DASS-21 is a 21-item self-report scale designed to measure the negative emotional state of depression, anxiety, and stress.³⁴ It is the short form of the DASS-42, and consists of 3 subscales (depression, anxiety, stress) that each contain 7 items.³⁴ Items are scored on a scale of 0 (did not apply at all) to 3 (applied very much or most of the time) to indicate how much the statement applied to the individual over the past week.³⁴ Scores for each subscale ranging from 0 to 21 are calculated by summing the values for each relevant item, with higher scores indicating higher levels of depression, anxiety, or stress.³⁴ Subscale scores are multiplied by 2 to yield values that can be compared with the original DASS-42.³⁵ Total scores are calculated by summing the 3 subscale scores.³⁴

No information identified

LEE

The LEE scale is a 60-item self-administered questionnaire that measures the perception of expressed emotion in a person’s influential relationships.^2,7,36 It consists of 4 subscales that assess attitude toward illness, emotional response, intrusiveness, and tolerance and expectations.^2,36 Each subscale includes 15 items that are rated in true-false format, and the scale generates scores for each of the 4 subscales and an overall expressed emotion score.³⁷ Higher scores indicate greater perceived expressed emotion.^7,36

No information identified

PSYCHLOPS

PSYCHLOPS is a psychometric instrument that can be used as an outcome measure to assess participants perspectives on their psychological distress.³⁸ It consists of 3 domains: problems, functioning, and wellbeing.³⁸ Four questions included in the PSYCHLOPS are rated using a using a 6-point Likert scale, ranging from 0 to 5.³⁹ Total scores are generated by summing the value assigned to each of these questions and range from 0 to 20.³⁹ Higher scores indicate higher psychological difficulty.³⁹

No information identified

SCL-90-R

The SCL-90-R is a 90-item self-report questionnaire for measuring a range of psychological and psychiatric symptoms.^40,41 It assesses 9 primary symptom dimensions containing 6 to 13 items each, including somatization, obsessive-compulsive, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation, and psychoticism.⁴² Each item is scored on a 5-point scale (0 = not at all; 4 = extremely), with higher numbers indicting more intense symptoms within the past week.⁴² A score is determined for each of the 9 symptom scales by summing values from relevant items.⁴⁰ These 9 primary dimensions are then summed to provide 3 global indices of psychological distress: Global Severity Index, the Positive Symptom Distress Index, and the Positive Symptoms Total.⁴¹ A total score is assigned using the sum of all items.^40,42

SCL-90-R clinically significant cut-off points for population with generalized psychological conditions:^c

• Global severity index = 0.60^d; 1.20^43,e

BDI-II

The BDI-II is a 21-item self-report questionnaire that assesses depressive symptoms among the emotional, cognitive, motivational, and physiological domains of depression.⁴⁴ Published in 1996, it is a revised version of the BDI that that corresponds with the depression diagnostic criteria defined in DSM-IV.⁴⁵ Each item is answered on a 4-point Likert scale between 0 and 3, with higher scores indicating increasing symptom severity.⁴⁵ Total scores range between 0 and 63 and are calculated using the sum of the scores for each item.⁴⁵ Total scores can be used to classify the severity of depressive symptoms as minimal (0 to 13), mild (14 to 19), moderate (20 to 28), and severe (29 or greater).⁴⁶

No information identified

CAPS

The CAPS is a 22-item self-report questionnaire used to assess perfectionism in young people.¹² It includes 2 subscales that measure self-oriented perfectionism (12 items) and socially prescribed perfectionism (10 items). Each item is rated on a 5-point scale (0 = false—not at all true of me; 4 = very true of me).⁴⁷ Ratings are used to assigned scores for each item, which are then summed to generate subscale scores.⁴⁷ Self-oriented perfectionism and socially prescribed perfectionism subscale scores range between 12 and 60 and 10 and 50, respectively.⁴⁷ Higher scores indicate higher levels of perfectionism.⁴⁷

No information identified

Work and social adjustment

WSAS

The WSAS is a 5-item self-report scale of social functional impairment attributable to a specific problem or disorder (e.g., an individual’s ED).⁴⁸ Each item is evaluated on a scale ranging from 0 (no impairment at all) to 8 (very severe impairment).⁴⁸ A total score ranging from 0 to 40 is calculated by summing the value assigned to each item, with higher scoring indicating higher impairment.⁴⁸

No information identified

SAS

The SAS is a 54-item self-report scale used to assess social adjustment and role performance in the past 2 weeks across 6 domains: work and school, social and leisure, extended family, primary relationship, parental, and family unit.⁴⁹ Each item is assigned a score between 1 and 5, with higher scores indicating greater impairment in functioning.⁵⁰ An overall score can be calculated by summing the scores of all the items and dividing by the number of items.⁵⁰

No information identified

Health care utilization

ZUF-8

The ZUF-8 is an 8-item self-report questionnaire used to measure treatment satisfaction.¹⁰ Each item is rated on a 4-point Likert scale, which are coded from 1 to 4.¹⁰ Scores from each of the 8 items are summed to generate a total score that ranges from 8 to 32, with higher values indicating decreased treatment satisfaction.¹⁰

No information identified

Global functioning outcomes

Kidscreen-27

The Kidscreen-27 is a 27-item self-report questionnaire that assesses health-related quality of life across five domains: physical well-being (5 items), psychological wellbeing (7 items), parent relations and autonomy (7 items), social support and peers (4 items), and school environment (4 items).⁵¹ It can be applied to both children and caregivers.⁵¹ Each item is rated using 5 possible multiple-choice responses (e.g., not at all, slightly, moderately, very, extremely), which are assigned a score between 1 and 5.⁵² For each domain, a scoring algorithm is used to calculate T-scores with a mean of 50 and a standard deviation of 10.⁵¹ A total score ranging from 27 to 135 is calculated by summing the values from each item, with higher scores indicating higher health-related quality of life.⁵²

No information identified

GOAS

The GOAS evaluates the central clinical features of anorexia nervosa.

No information identified

Richards et al., (2023)¹

Serious ROB [?]

1.1 (PY) Confounding factors may impact the effect of intervention

1.2 (N) All participants received the same intervention so analysis was not based on follow-up time

1.4 (N) Authors did not report using appropriate analysis method to control for confounding

1.6 (N) Authors did not control for any post-intervention variables that could have been affected by intervention

Moderate ROB [+]

2.1 (N) Selection of participants was not based on participant characteristics observed after the start of the intervention

2.4 (Y) Intervention and follow-up was applied uniformly across participant groups (e.g., baseline data, 6-month and 12-month follow-up data)

Serious ROB [?]

3.1 (N) There was sufficient ambiguity in FREED-4-All participant and FREED-Up participant groups. It was unclear if there was any cross over between groups

3.2 (N) It is unclear when information used to define intervention groups was recorded

3.3 (N) Knowledge of intervention status would not have affected potential outcomes

Low ROB [+]

4.3 (NI) No co-interventions were included in the analysis

4.4 (PY) Implementation of intervention was likely successful for included participants

4.5 (PY) It is unlikely that participants would not adhere to FREED intervention regimen

4.6 (NA)

Serious ROB [+]

5.1 (N) There was a significant amount of missing data from both FREED-4-All and FREED-Up participants for follow-up measurements

5.2 (PN) It was not indicated that participants were excluded due to missing data

5.3 (PN) It was not indicated that participants were excluded due to missing data on other variables needed for the analysis

5.4 (PN) One intervention group had a much higher proportion of missing data at follow-up

5.5 (PN) It is not clear that the results are robust with the presence of missing data

Moderate ROB [?]

6.1 (PN) Outcome measures would not be influenced by knowledge of intervention group

6.2 (PY) Assessors were likely aware of which intervention group was being assessed

6.3 (Y) Similar methods of outcome assessments were used across intervention group

6.4 (PN) Errors in outcome measurements are likely not attributable to intervention received

Moderate ROB [?]

7.1 (PN) Outcomes assessed were not likely measured multiple times

7.2 (PY) For certain outcomes multiple analyses were done over different time point to assess change

7.3 (N) Different subgroups were not analyzed

Serious ROB [+]

Austin et al., (2022)²

Serious ROB [?]

1.13 (PY) Confounding factors (e.g., age, gender, location, illness type, treatment) may impact the effect of intervention

1.2 (N) All participants received the same intervention so analysis was not based on follow-up time

1.4. (N) Authors did not report using appropriate analysis method to control for confounding

1.6 (N) Authors did not control for any post-intervention variables that could have been affected by intervention

Moderate ROB [+]

2.1 (N) Selection of participants was not based on participant characteristics observed after the start of the intervention

2.4 (Y) Intervention and follow-up was applied uniformly across participant groups (e.g., baseline data, 6-month and 12-month follow-up data)

Low ROB [?]

3.1 (Y) Intervention groups were presented with clear detail (FREED cohort vs TAU cohort)

3.2 (Y) Information used to classify intervention groups were not likely to be confused due to one group being a historical cohort comparator

3.3 (N) Knowledge of the outcomes would not impact classification of intervention group

Low ROB [+]

4.3 (NI) No co-interventions were included in the analysis

4.4 (PY) Implementation of intervention was likely successful for included participants

4.5 (PY) It is unlikely that participants would not adhere to FREED intervention regimen

4.6 (NA)

Moderate ROB [?]

5.1 (Y) Outcome data was presented for nearly all participants from baseline to follow-up

5.2 (PN) It was not indicated that participants were excluded due to missing data

5.3 (PN) It was not indicated that participants were excluded due to missing data on other variables needed for the analysis

Moderate ROB [?]

6.1 (PN) Outcome measures would not be influenced by knowledge of intervention group

6.2 (PY) Assessors were likely aware of which intervention group was being assessed

6.3 (Y) Similar methods of outcome assessments were used across intervention group

6.4 (PN) Errors in outcome measurements are likely not attributable to intervention received

Moderate ROB [?]

7.1 (PN) Outcomes assessed were not likely measured multiple times

7.2 (PY) For certain outcomes multiple analyses were done over different time point to assess change

7.3 (PN) Different diagnostic subgroups were analyzed but it is unclear if this impacted reported effect estimates

Serious ROB [?]

Radunz et al., (2021)³

Serious ROB [?]

1.1 (PY) Confounding factors (e.g., age, gender, location, illness type, treatment) may impact the effect of intervention

1.2 (N) All participants received the same intervention so analysis was not based on follow-up time

1.4 (N) Authors did not report using appropriate analysis method to control for confounding

1.6 (N) Authors did not control for any post-intervention variables that could have been affected by intervention

Serious ROB [+]

2.2 (N) Selection of participants was not based on participant characteristics observed after the start of the intervention

6.3 (PY) It is likely that the start of intervention and data extracted at follow-up were at similar time points for participants

Serious ROB [+]

3.1 (Y) Only 1 intervention group was included in the analysis

3.2 (Y) Information used to define intervention group was likely recorded at the start of the intervention

3.3 (PN) Since only 1 intervention group was analyzed, it is unlikely that knowledge of outcomes would impact intervention group

Moderate ROB [+]

4.3 (NI) Co-interventions were not included in this analysis

4.4 (PY) It is likely that intervention implementation was successful for most participants

4.5 (Y) Participants likely adhered to intervention regimen

Serious ROB [+]

5.1 (N) There was a significant amount of missing data for follow-up outcome measurements

5.2 (PN) It was not indicated that participants were excluded due to missing data

5.3 (PN) It was not indicated that participants were excluded due to missing data on other variables needed for the analysis

5.4 (NA) Only 1 intervention group was analyzed

5.5 (PN) There is no indication that appropriate methods were used to account for missing data

Serious ROB [+]

6.1 (PN) Only 1 intervention group was included in the analysis which likely did not impact outcome measures

6.2 (Y) Assessors were aware of which intervention group was being assessed

6.3 (NI) Only 1 intervention group was included in the analysis

6.4 (PN) Errors in outcome measurements are likely not attributable to intervention received

Serious ROB [+]

7.1 (PY) Multiple outcome measurements were used for certain outcome domains

7.2 (PY) For certain outcomes multiple analyses were done over different time point to assess change

#7.3 (N) No subgroup analysis was complete

Serious ROB [+]

Richards et al., (2021)⁴

Serious ROB [?]

1.1 (PY) Confounding factors (e.g., age, gender, location, illness type, treatment) may impact the effect of intervention

1.2 (N) All participants received the same intervention so analysis was not based on follow-up time

1.4 (N) Authors did not report using appropriate analysis method to control for confounding

1.6 (N) Authors did not control for any post-intervention variables that could have been affected by intervention

Moderate ROB [?]

2.1 (N) Selection of participants was not based on participant characteristics observed after the start of the intervention

2.4 (PY) It is likely that intervention and follow-up analysis was done uniformly for most participants

Moderate ROB [?]

3.1 (Y) Intervention groups were presented with clear detail (FREED cohort vs TAU cohort)

3.2 (Y) Information used to classify intervention groups were not likely to be confused due to one group being a historical cohort comparator

3.3 (N) Knowledge of the outcomes would not impact classification of intervention group

Low ROB [?]

4.3 (NI) No co-interventions were included in the analysis

4.4 (PY) Implementation of intervention was likely successful for included participants

4.5 (PY) It is unlikely that participants would not adhere to FREED intervention regimen

4.6 (NA)

Moderate ROB [+]

5.1 (Y) Outcome data was available for nearly all participants

5.2 (PN) It was not indicated that participants were excluded due to missing data

5.3 (PN) It was not indicated that participants were excluded due to missing data on other variables needed for the analysis

Moderate ROB [?]

6.1 (PN) Outcome measures would not be influenced by knowledge of intervention group

6.2 (PY) Assessors were likely aware of which intervention group was being assessed

6.3 (Y) Similar methods of outcome assessments were used across intervention group

6.4 (PN) Errors in outcome measurements are likely not attributable to intervention received

Moderate ROB [?]

7.1 (PN) Outcomes assessed were not likely measured multiple times

7.2 (PY) For certain outcomes multiple analyses were done over different time point to assess change

7.3 (PN) Different diagnostic subgroups were analyzed but it is unclear if this impacted reported effect estimates

Serious ROB [?]

Flynn et al., (2020)⁵

Serious ROB [?]

1.1 (PY) Confounding factors (e.g., age, gender, location, illness type, treatment) may impact the effect of intervention

1.2 (N) All participants received the same intervention so analysis was not based on follow-up time

1.5 (N) Authors did attempt to minimize confounding factors, but no appropriate analysis method to control for confounding was used

1.6 (N) Authors did not control for any post-intervention variables that could have been affected by intervention

Moderate ROB [?]

2.1 (N) Selection of participants was not based on participant characteristics observed after the start of the intervention

2.4 (PY) It is likely that intervention and follow-up analysis was done uniformly for most participants

Low [?]

3.1 (Y) Intervention groups were presented with clear detail (FREED cohort vs TAU cohort)

3.2 (Y) Information used to classify intervention groups were not likely to be confused due to one group being a historical cohort comparator

3.3 (N) Knowledge of the outcomes would not impact classification of intervention group

Low ROB [+]

4.1 (NI) No co-interventions were included in the analysis

4.2 (PY) Implementation of intervention was likely successful for included participants

4.3 (PY) It is unlikely that participants would not adhere to FREED intervention regimen

Serious ROB [+]

5.1 (NI) No information was reported related to loss to follow-up or missing outcome data

5.2 (NI) No information was provided relating to how potential missing data was handled

5.3 (NI) No information was provided relating to how potential missing data was handled for variables needed for the analysis

Serious ROB [+]

6.1 (PN) Outcome measures would not be influenced by knowledge of intervention group

6.2 (PY) Assessors were likely aware of which intervention group was being assessed

6.3 (N) Different methods of outcome assessment were used between the FREED cohort and TAU cohort

6.4 (PN) Errors in outcome measurements are likely not attributable to intervention received

Moderate ROB [?]

7.1 (PN) Outcomes assessed were not likely measured multiple times

7.2 (PY) For certain outcomes multiple analyses were done over different time point to assess change

7.3 (N) No subgroup analysis was complete

Serious ROB [?]

Fukutomi et al., (2019)⁶

Serious ROB [?]

1.1 (PY) Confounding factors (e.g., age, gender, location, illness type, treatment) may impact the effect of intervention

1.2 (N) All participants received the same intervention so analysis was not based on follow-up time

1.4 (N) Authors did not report using appropriate analysis method to control for confounding

1.6 (N) Authors did not control for any post-intervention variables that could have been affected by intervention

Moderate ROB [?]

2.1 (N) Selection of participants was not based on participant characteristics observed after the start of the intervention

2.4 (PY) It is likely that intervention and follow-up analysis was done uniformly for most participants

Moderate [?]

3.1 (Y) Intervention groups were presented with clear detail (FREED cohort vs TAU cohort)

3.2 (PY) Analysis was done on historical cohort data so misclassification of intervention status is unlikely

3.3 (N) Knowledge of the outcomes would not impact classification of intervention group

Low ROB [?]

4.3 (NI) No co-interventions were included in the analysis

4.4 (PY) Implementation of intervention was likely successful for included participants

4.5 (PY) It is unlikely that participants would not adhere to FREED intervention regimen

Moderate ROB [?]

5.1 (PY) Due to use of retrospective cohort data, it is unlikely that significant amount of data was missing

5.2 (N) Analysis was complete to include participants with potential missing data on intervention status

5.3 (N) Analysis was complete to include participants with potential missing data on variables needed for analysis

Low ROB [?]

6.1 (PN) Outcome measures would not be influenced by knowledge of intervention group

6.2 (PY) Assessors were likely aware of which intervention group was being assessed

6.3 (Y) Similar methods of outcome assessments were used across intervention group

6.4 (PN) Errors in outcome measurements are likely not attributable to intervention received

Moderate ROB [?]

7.1 (PN) Outcomes assessed were not likely measured multiple times

7.2 (PY) For certain outcomes multiple analyses were done over different time point to assess change

7.3 (N) No subgroup analysis was complete

Serious ROB [?]

McClelland et al., (2018)⁷

Serious ROB [?]

1.1 (PY) Confounding factors (e.g., age, gender, location, illness type, treatment) may impact the effect of intervention

1.2 (N) All participants received the same intervention so analysis was not based on follow-up time

1.4 (N) Authors did not report using appropriate analysis method to control for confounding

1.6 (N) Authors did not control for any post-intervention variables that could have been affected by intervention

Moderate ROB [?]

2.1 (N) Selection of participants was not based on participant characteristics observed after the start of the intervention

2.4 (PY) It is likely that intervention and follow-up analysis was done uniformly for most participants

Low ROB [?]

3.1 (Y) Intervention groups were presented with clear detail (FREED cohort vs TAU cohort)

3.2 (Y) Information used to classify intervention groups were not likely to be confused due to one group being a historical cohort comparator

3.3 (N) Knowledge of the outcomes would not impact classification of intervention group

Low ROB [?]

4.3 (NI) No co-interventions were included in the analysis

4.4 (PY) Implementation of intervention was likely successful for included participants

4.5 (PY) It is unlikely that participants would not adhere to FREED intervention regimen

Serious ROB [+]

5.1 (N) There was a significant amount of missing data for follow-up outcome measurement for FREED cohort and audit cohort

5.2 (N) Analysis was complete to include participants with potential missing data on intervention status

5.3 (N) Analysis was complete to include participants with potential missing data on variables needed for analysis

5.4 Proportions of missing data to follow-up were larger for TAU cohort compared to FREED cohort

5.5 (PY) Appropriate statistical analysis using mixed models was used allowing for missing data to be included and may allow for a robust analysis

Serious ROB [+]

6.1 (PN) Outcome measures would not be influenced by knowledge of intervention group

6.2 (PY) Assessors were likely aware of which intervention group was being assessed

6.3 (N) Different methods of outcome assessment were used between the FREED cohort and audit cohort

6.4 (PN) Errors in outcome measurements are likely not attributable to intervention received

Moderate ROB [?]

7.1 (PN) Outcomes assessed were not likely measured multiple times

7.2 (PY) For certain outcomes multiple analyses were done over different time point to assess change

7.3 (N) No subgroup analysis was complete

Serious ROB [?]

Brown et al., (2016)⁸

Serious ROB [?]

1.1 (PY) Confounding factors (e.g., age, gender, location, illness type, treatment) may impact the effect of intervention

1.2 (N) All participants received the same intervention so analysis was not based on follow-up time

1.4 (N) Authors did not report using appropriate analysis method to control for confounding

1.6 (N) Authors did not control for any post-intervention variables that could have been affected by intervention

Moderate ROB [?]

2.1 (N) Selection of participants was not based on participant characteristics observed after the start of the intervention

2.4 (PY) It is likely that intervention and follow-up analysis was done uniformly for most participants

Low ROB [?]

3.1 (Y) Intervention groups were presented with clear detail (FREED cohort vs TAU cohort)

3.2 (Y) Information used to classify intervention groups were not likely to be confused due to one group being a historical cohort comparator

3.3 (N) Knowledge of the outcomes would not impact classification of intervention group

Low ROB [?]

4.3 (NI) No co-interventions were included in the analysis

4.4 (PY) Implementation of intervention was likely successful for included participants

4.5 (PY) It is unlikely that participants would not adhere to FREED intervention regimen

Serious ROB [?]

5.1 (NI) No information was reported related to loss to follow-up or missing outcome data

5.2 (NI) No information was provided relating to how potential missing data was handled

5.3 (NI) No information was provided relating to how potential missing data was handled for variables needed for the analysis

Serious ROB [+]

6.1 (PN) Outcome measures would not be influenced by knowledge of intervention group

6.2 (PY) Assessors were likely aware of which intervention group was being assessed

6.3 (N) Different methods of outcome assessment were used between the FREED cohort and audit cohort

6.4 (PN) Errors in outcome measurements are likely not attributable to intervention received

Moderate ROB [?]

7.1 (PN) Outcomes assessed were not likely measured multiple times

7.2 (PY) For certain outcomes multiple analyses were done over different time point to assess change

7.3 (N) No subgroup analysis was complete

Serious ROB [?]

Herpertz-Dahlmann et al., (2021)¹⁰

Serious ROB [+]

1.1 (PY) Confounding factors (e.g., age, gender, location, illness type, treatment) may impact the effect of intervention

1.2 (N) All participants received the same intervention so analysis was not based on follow-up time

1.4 (N) Authors did not report using appropriate analysis method to control for confounding

1.5 (N) Authors did not control for any post-intervention variables that could have been affected by intervention

Moderate ROB [+]

2.1 (N) Selection of participants was not based on participant characteristics observed after the start of the intervention

2.4 (PY) The intervention and follow-up was likely applied similarly for all participants included in the study

Serious ROB [+]

3.1 (Y) Only 1 intervention group was included in the analysis

3.2 (Y) Information used to define intervention group was likely recorded at the start of the intervention

3.3 (PN) Since only 1 intervention group was analyzed, it is unlikely that knowledge of outcomes would impact intervention group

Moderate ROB [+]

4.1 (PN) It is unlikely that there are significant deviations from intended intervention that would impact outcome assessment

Moderate ROB [+]

5.1.3 (Y) Outcome data and follow-up data was available for nearly all participants

5.2 (NI) No information was provided relating to how potential missing data was handled

5.3 (NI) No information was provided relating to how potential missing data was handled for variables needed for the analysis

Serious ROB [+]

6.1 (PN) Only 1 intervention group was included in the analysis which likely did not impact outcome measures

6.2 (Y) Assessors were aware of which intervention group was being assessed

6.3 (NI) Only 1 intervention group was included in the analysis

6.4 (PN) Errors in outcome measurements are likely not attributable to intervention received

Moderate ROB [+]

7.1 (PN) Outcomes assessed were not likely measured multiple times

7.2 (PY) For certain outcomes multiple analyses were done over different time point to assess change

7.3 (N) No subgroup analysis was complete

Serious ROB [+]

Coelho et al., (2019)¹¹

Serious ROB [+]

1.1 (PY) Confounding factors (e.g., age, gender, location, illness type, treatment) may impact the effect of intervention

1.2 (N) All participants received the same intervention so analysis was not based on follow-up time

1.4 (N) Authors did not report using appropriate analysis method to control for confounding

1.5 (N) Authors did not control for any post-intervention variables that could have been affected by intervention

Moderate ROB [+]

2.1 (N) Selection of participants was not based on participant characteristics observed after the start of the intervention

2.4 (PY) The intervention and follow-up was likely applied similarly for all participants included in the study

Serious ROB [+]

3.1 (Y) Only 1 intervention group was included in the analysis

3.2 (Y) Information used to define intervention group was likely recorded at the start of the intervention

3.3 (PN) Since only 1 intervention group was analyzed, it is unlikely that knowledge of outcomes would impact intervention group

Moderate ROB [+]

4.1 (PN) It is unlikely that there are significant deviations from intended intervention that would impact outcome assessment

Moderate ROB [+]

5.1 (Y) Outcome data and follow-up data was available for nearly all participants

5.2 (NI) No information was provided relating to how potential missing data was handled

5.3 (NI) No information was provided relating to how potential missing data was handled for variables needed for the analysis

Serious ROB [+]

6.1 (PN) Only 1 intervention group was included in the analysis which likely did not impact outcome measures

6.2 (Y) Assessors were aware of which intervention group was being assessed

6.3 (NI) Only 1 intervention group was included in the analysis

6.4 (PN) Errors in outcome measurements are likely not attributable to intervention received

Moderate ROB [+]

7.1 (PN) Outcomes assessed were not likely measured multiple times

7.2 (PY) For certain outcomes multiple analyses were at program admission and discharge

7.3 (N) No subgroup analysis was complete

Serious ROB [+]

Hurst et al., (2019)¹²

Serious ROB [+]

1.1 (PY) Confounding factors (e.g., age, gender, location, illness type, treatment) may impact the effect of intervention

1.2 (N) All participants received the same intervention so analysis was not based on follow-up time

1.4 (N) Authors did not report using appropriate analysis method to control for confounding

1.5 (N) Authors did not control for any post-intervention variables that could have been affected by intervention

Moderate ROB [+]

2.1 (N) Selection of participants was not based on participant characteristics observed after the start of the intervention

2.4 (PY) The intervention and follow-up was likely applied similarly for all participants included in the study

Serious ROB [+]

3.1 (Y) Only 1 intervention group was included in the analysis

3.2 (Y) Information used to define intervention group was likely recorded at the start of the intervention

3.3 (PN) Since only 1 intervention group was analyzed, it is unlikely that knowledge of outcomes would impact intervention group

Moderate ROB [+]

4.1 (PN) It is unlikely that there are significant deviations from intended intervention that would impact outcome assessment

Moderate ROB [+]

5.1 (Y) Outcome data and follow-up data was available for nearly all participants

5.2 (NI) No information was provided relating to how potential missing data was handled

5.3 (NI) No information was provided relating to how potential missing data was handled for variables needed for the analysis

Serious ROB [+]

6.1 (PN) Only 1 intervention group was included in the analysis which likely did not impact outcome measures

6.2 (Y) Assessors were aware of which intervention group was being assessed

6.3 (NI) Only 1 intervention group was included in the analysis

6.4 (PN) Errors in outcome measurements are likely not attributable to intervention received

Serious ROB [+]

7.1 (PY) For each outcome domain, multiple outcome measurements were included in the analysis

7.2 (PY) Multiple analysis of intervention-outcome results were included in the analysis

7.3 (N) No subgroup analysis was complete

Serious ROB [+]

Rosling et al., (2016)¹³

Serious ROB [+]

1.1 (PY) Confounding factors (e.g., age, gender, location, illness type, treatment) may impact the effect of intervention

1.2 (N) All participants received the same intervention so analysis was not based on follow-up time

1.4 (N) Authors did not report using appropriate analysis method to control for confounding

1.5 (N) Authors did not control for any post-intervention variables that could have been affected by intervention

Moderate ROB [+]

2.2 (N) Selection of participants was not based on participant characteristics observed after the start of the intervention

2.4 (PY) The intervention and follow-up was likely applied similarly for all participants included in the study

Serious ROB [+]

3.1 (Y) Only 1 intervention group was included in the analysis

3.2 (Y) Information used to define intervention group was likely recorded at the start of the intervention

3.3 (PN) Since only 1 intervention group was analyzed, it is unlikely that knowledge of outcomes would impact intervention group

Serious ROB [+]

4.3 (NI) No co-interventions were included in the analysis

4.4 (PY) Implementation of intervention was likely successful for included participants

4.5 (PY) Only 1 intervention was used and it is likely that participants adhered to intervention given

Serious ROB [+]

5.1 (N) There was a significant amount of participants that were lost to follow-up and had missing data

5.2 (NI) No information was provided relating to how potential missing data was handled

5.3 (NI) No information was provided relating to how potential missing data was handled for variables needed for the analysis

5.4 (NA) Only 1 intervention group was included in the analysis

5.5 (PY) Analysis of missing data was included and provides evidence that results of analysis were robust

Serious ROB [+]

6.1 (PN) Only 1 intervention group was included in the analysis which likely did not impact outcome measures

6.2 (Y) Assessors were aware of which intervention group was being assessed

6.3 (NI) Only 1 intervention group was included in the analysis

6.4 (PN) Errors in outcome measurements are likely not attributable to intervention received

Serious ROB [+]

7.1 (PN) Outcomes assessed were not likely measured multiple times

7.2 (PY) For certain outcomes multiple analyses were done over different time point to assess change

7.3 (N) No subgroup analysis was complete

Serious ROB [+]

Godart et al., (2022);⁹ Godart et al., (2012)¹⁴

Some concerns [?]

1.1 (Y) Randomization was done in block methods using an SPSS randomization program

1.2 (Y) Allocation of intervention group was sealed in an envelope so participants were unaware of intervention status

1.3 (N) There was minimal baseline differences between intervention groups post-randomization which suggests there was no issue with the randomization process

Some concerns [+]

2.1 (PY) Based on the type of intervention, it would be unlikely that the participants were unaware of the intervention they would be receiving (i.e., family therapy vs no family therapy)

2.2 (Y) Blinding program administrators would not be possible for the intervention included

2.3 (N) No changes or deviations to assigned intervention group was reported

2.6 (PY) The trial first used intention-to-treat analysis then per-protocol analysis for to estimate the effect of assignment to treatment

Some concerns [?]

2.6 (Y) Nearly all data was available for participants that were randomized at baseline and last available follow-up, any missing data was similar across intervention groups

Some concern [?]

4.1 (PN) Methods of outcome measures were verified and appropriate

4.2 (PN) Methods of outcome measurements were applied uniformly for each intervention group at comparable time points

4.3 (N) Assessors of outcomes measures were blinded to participant intervention status

High ROB [?]

5.1 (Y) A pre-specified analysis plan was used before outcome data was available for analysis

5.2 (PY) Outcome measurements were assessed at multiple time points using results from scales, which may impact selection of reported results

5.3 (PN) Measurements of results were likely only analyzed in one way but at multiple time points

High ROB [?]

EDE-Q score

Richards et al., (2023)¹

Results of all participants from the FREED-Up cohort (n analyzed = 278 at baseline [T1]; 182 at 6 months [T2]; 175 at 12 months [T3])

• EDE-Q scores, M (SD) at T1 vs T2; MD = 4.08 (1.21) vs 2.85 (1.57); 1.23 (P < 0.001)

• EDE-Q scores, M (SD) at T1 vs T3; MD = 4.08 (1.21) vs 2.31(1.55); 1.77 (P < 0.001)

• % (n) above EDE-Q clinical cut-off (> 2.8) at T1, T2, and T3 = 84% (233); 49% (89); 35% (61)

Results of all participants from the FREED-4-All cohort (n analyzed = 793 at baseline [T1]; 135 at post-treatment [T2])

• EDE-Q scores, M (SD) at T1 vs T2; MD = 4.06 (1.29) vs 2.04 (1.39); 2.02 (P < 0.001)

• % (n) above EDE-Q clinical cut-off (> 2.8) at T1 and T2 = 84% (633); 29% (39)

Austin et al., (2022)²

Results of all participants from the FREED cohort (n analyzed = 278 at baseline [T1]; 216 at 3 months [T2]; 182 at 6 months [T3]; 175 at 12 months [T4])

• EDE-Q score MD (95% CI) at T1 vs T2; SE (P value) = −0.92 (−1.07 to −0.78); 0.074 (P < 0.001)

• EDE-Q score MD (95% CI) at T2 vs T3; SE (P value) = −0.34 (−0.50 to −0.18); 0.080 (P < 0.001)

• EDE-Q score MD (95% CI) at T3 vs T4; SE (P value) = −0.49 (−0.66 to −0.32); 0.11 (P < 0.001)

• EDE-Q score MD (95% CI) at T1 vs T4; SE (P value) = −1.75 (−1.97 to −1.54); 0.11 (P < 0.001)

Radunz et al., (2021)³

Results of mean EDE-Q scores from baseline (n analyzed = 96) and end of treatment (n analyzed = 30)

• Baseline EDE-Q score, M (SD) = 4.25 (1.12)

• End of treatment EDE-Q score, M (SD) = 1.87 (1.12)

• Between group difference, d (95% CI) = 2.05 (1.43 to 2.68)

• P < 0.001

McClelland et al., (2018)⁷

Results of mean EDE-Q score from all participants from the FREED cohort (n analyzed = 53 at baseline [T1]; 37 at 3 months [T2]; 32 at 6 months [T3]; 25 at 12 months [T4])

• T1 score, M (SD) = 4.0 (1.3)

• T2 score, M (SD) = 3.2 (1.4)

• T3 score, M (SD) = 2.5 (1.4)

• T4 score, M (SD) = 2.2 (1.6)

Mean change in EDE-Q score from T1 (n analyzed = 53) to T2 (n analyzed = 37) of all participants from the FREED cohort

• M (95% CI) = −0.82 (−1.21 to −0.43)

• P = 0.001

Mean change in EDE-Q score from T1 (n analyzed = 53) to T3 (n analyzed = 32) of all participants from the FREED cohort

• M (95% CI) = −1.55 (−2.06 to −1.05)

• P = 0.001

Mean change in EDE-Q score from T1 (n analyzed = 53) to T4 (n analyzed = 25) of all participants from the FREED cohort

• M (95% CI) = −2.08 (−2.76 to −1.41)

• P = 0.001

Mean change in EDE-Q score from T3 (n analyzed = 32) to T4 (n analyzed = 25) of all participants from the FREED cohort

• M (95% CI) = −0.53 (−1.02 to −0.03)

• P = 0.03

ED cognition

Radunz et al., (2021)³

Change in linear trend for ED-15 outcome (ED cognition) across days since treatment commencement (0 to 70 days)

• Change (SE) = −0.022 (0.0022)

• P < 0.001

Change in quadratic trend for ED-15 outcome (ED cognition) across days since treatment commencement (0 to 70 days)

• Change (SE) = 0.00006 (0.00001)

• P < 0.001

Binge episodes

Richards et al., (2023)¹

Results of all participants from the FREED-Up cohort (n analyzed = 278 at baseline [T1]; 182 at 6 months [T2]; 175 at 12 months [T3])

• Binge episodes per month, M (SD) at T1 vs T2; MD = 6.41 (8.39) vs 3.70 (8.17); 2.71 (P < 0.001)

• Binge episodes per month, M (SD) at T1 vs T3; MD = 6.41 (8.39) vs 2.39 (4.60); 4.02 (P < 0.001)

Results of all participants from the FREED-4-All cohort (n analyzed = 820 at baseline [T1]; 151 at post-treatment [T2])

• Binge episodes per month, M (SD) at T1 vs T2; MD = 4.83 (10.17) vs 2.19 (4.84); 2.64 (P < 0.001)

Austin et al., (2022)²

Results of participants diagnosed with BN, BED, or OSFED from the FREED cohort (n analyzed = 125 at baseline [T1]; 76 at 12 months [T4])^a

• Binge episodes MD (95% CI) at T1 vs T2; SE (P value) = −5.53 (−7.28 to −3.79); 0.88 (P < 0.001)

• Binge episodes MD (95% CI) at T2 vs T3; SE (P value) = −0.19 (−1.72 to 2.10); 0.97 (P = 0.84)

• Binge episodes MD (95% CI) at T3 vs T4; SE (P value) = −2.56 (−4.58 to 0.55); 1.02 (P = 0.13)

• Binge episodes MD (95% CI) at T1 vs T4; SE (P value) = −8.29 (−10.09 to −6.48); 0.92 (P < 0.001)

Radunz et al., (2021)³

Change in linear trend for ED-15 outcome (binge eating) across days since treatment commencement (0 to 70 days)

• Change (SE) = −0.02 (0.0041)

• P < 0.001

Change in quadratic trend for ED-15 outcome (binge eating) across days since treatment commencement (0 to 70 days)

• Change (SE) = 0.00009 (0.00003)

• P < 0.001

Purging episodes

Richards et al., (2023)⁴

Results of all participants from the FREED-Up cohort (n analyzed = 278 at baseline [T1]; 182 at 6 months [T2]; 175 at 12 months [T3])

• Vomit episodes per month, M (SD) at T1 vs T2; MD = 6.97 (11.76) vs 3.27 (9.73); 3.70 (P < 0.001)

• Vomit episodes per month, M (SD) at T1 vs T3; MD = 6.97 (11.76) vs 2.39 (4.60); 4.79 (P < 0.001)

Results of all participants from the FREED-4-All cohort (n analyzed = 821 at baseline [T1]; 150 at post-treatment [T2])

• Vomit episodes per month, M (SD) at T1 vs T2; MD = 5.84 (15.07) vs 1.43 (3.98); 4.41 (P < 0.001)

Austin et al. (2022)²

Results of participants diagnosed with BN, BED, or OSFED from the FREED cohort (n analyzed = 98 at baseline [T1]; 56 at 12 months [T4])^a

• Vomiting episodes MD (95% CI) at T1 vs T2; SE (P value) = −6.51 (−8.42 to −4.61); 0.97 (P < 0.001)

• Vomiting episodes MD (95% CI) at T2 vs T3; SE (P value) = −0.76 (−2.84 to 1.31); 1.05 (P = 0.47)

• Vomiting episodes MD (95% CI) at T3 vs T4; SE (P value) = −2.86 (−5.14 to −0.58); 1.16 (P = 0.014)

• Vomiting episodes MD (95% CI) at T1 vs T4; SE (P value) = −10.13 (−13.23 to −7.03); 1.58 (P < 0.001)

Radunz et al., (2021)³

Change in linear trend for ED-15 outcome (vomiting) across days since treatment commencement (0 to 70 days)

• Change (SE) = −0.008 (0.0029)

• P = 0.008

Change in quadratic trend for ED-15 outcome (vomiting) across days since treatment commencement (0 to 70 days)

• Change (SE) = 0.00005 (0.00002)

• P = 0.02

Laxative use

Richards et al., (2023)¹

Results of all participants from the FREED-Up cohort (n analyzed = 278 at baseline [T1]; 182 at 6 months [T2]; 175 at 12 months [T3])

• Laxative episodes per month, M (SD) at T1 vs T2; MD = 2.03 (6.52) vs 1.13 (4.22); 0.90 (P < 0.05)

• Laxative episodes per month, M (SD) at T1 vs T3; MD = 2.03 (6.52) vs 0.55 (2.93); 1.48 (P < 0.001)

Results of all participants from the FREED-4-All cohort (n analyzed = 823 at baseline [T1]; 153 at post-treatment [T2])

• Laxative episodes per month, M (SD) at T1 vs T2; MD = 1.30 (5.71) vs 0.46 (2.83); 0.84 (not SS)

Austin et al., (2022)²

Results of participants diagnosed with BN, BED, or OSFED from the FREED cohort (n analyzed = 39 at baseline [T1]; 23 at 12 months [T4])^a

• Laxative use MD (95% CI) at T1 vs T2; SE (P value) = −5.66 (−8.50 to −2.82); 1.42 (P < 0.001)

• Laxative use MD (95% CI) at T2 vs T3; SE (P value) = −1.05 (−4.16 to 2.06); 1.56 (P = 0.5)

• Laxative use MD (95% CI) at T3 vs T4; SE (P value) = −2.55 (−5.80 to −0.70); 1.00 (P = 0.12)

• Laxative use MD (95% CI) at T1 vs T4; SE (P value) = −9.26 (−12.40 to −6.12); 1.56 (P < 0.001)

Radunz et al., (2021)³

Change in linear trend for ED-15 outcome (laxative use) across days since treatment commencement (0 to 70 days)

• Change (SE) = 0.004 (0.02)

• P = 0.86

Change in quadratic trend for ED-15 outcome (laxative use) across days since treatment commencement (0 to 70 days)

• Change (SE) = 0.000001 (0.00001)

• P = 0.92

Excessive exercise

Austin et al., (2022)²

Results of participants diagnosed with BN, BED, or OSFED from the FREED cohort (n analyzed = 112 at baseline [T1]; 62 at 12 months [T4])^a

• Excessive exercise MD (95% CI) at T1 vs T2; SE (P value) = −6.10 (−7.56 to −4.64); 0.74 (P < 0.001)

• Excessive exercise MD (95% CI) at T2 vs T3; SE (P value) = −2.22 (−3.82 to −0.62); 0.81 (P = 0.007)

• Excessive exercise MD (95% CI) at T3 vs T4; SE (P value) = −0.63 (−2.38 to 1.13); 0.89 (P = 0.48)

• Excessive exercise MD (95% CI) at T1 vs T4; SE (P value) = −8.95 (−11.04 to −6.86); 1.06 (P < 0.001)

Radunz et al., (2021)³

Change in linear trend for ED-15 outcome (driven exercise) across days since treatment commencement (0 to 70 days)

• Change (SE) = −0.013 (0.0042)

• P < 0.001

Change in quadratic trend for ED-15 outcome (driven exercise) across days since treatment commencement (0 to 70 days)

• Change (SE) = 0.00003 (0.00005)

• P = 0.03

Restrictive dieting

Radunz et al., (2021)³

Change in linear trend for ED-15 outcome (restrictive dieting) across days since treatment commencement (0 to 70 days)

• Change (SE) = −0.024 (0.0072)

• P < 0.001

Change in quadratic trend for ED-15 outcome (restrictive dieting) across days since treatment commencement (0 to 70 days)

• Change (SE) = 0.00007 (0.00005)

• P = 0.01

DUSC

Flynn et al., (2020)⁵

Results of DUSC (months) for participants diagnosed with AN, BN, BED, OSFED from the total FREED cohort (n analyzed = 278)

• AN, M (SD) = 16.50 (10.58)

• BN, M (SD) = 19.35 (10.34)

• BED, M (SD) = 17.67 (8.33)

• OSFED, M (SD) = 15.09 (9.85)

• Total, M (SD) = 16.82 (10.31)

Results of DUSC (months) for participants diagnosed with AN, BN, BED, OSFED under optimal conditions from FREED cohort (n analyzed = 157)

• AN, M (SD) = 13.29 (8.85)

• BN, M (SD) = 18.94 (10.69)

• BED, M (SD) = 17.67 (8.33)

• OSFED, M (SD) = 12.95 (8.35)

• Total, M (SD) = 15.11 (9.58)

Results of DUSC (months) for participants diagnosed with AN, BN, BED, OSFED from TAU cohort (n analyzed = 224)

• AN, M (SD) = 15.62 (10.67)

• BN, M (SD) = 19.81 (9.30)

• BED, M (SD) = 16.75 (10.87)

• OSFED, M (SD) = 16.38 (11.20)

• Total, M (SD) = 16.47 (10.41)

Between group comparison of total DUSC (months) from total FREED cohort (n = 278) vs TAU cohort (n = 224)

• P = 0.71

• 95% CI = −1.49 to 2.13

Between group comparison of total DUSC (months) from FREED cohort under optimal conditions (n = 157) vs TAU cohort (n = 224)

• P = 0.200

• 95% CI = −3.45 to 0.72

Brown et al., (2016)⁸

Mean DUSC (months) for all FREED cohort (n analyzed = 51), FREED cohort with minimal gatekeeping (n analyzed = 14), FREED cohort with complex gatekeeping (n analyzed = 37), and TAU cohort (n analyzed = 89)

• All FREED cohort, M (SD) = 15.67 (10.04)

• FREED cohort with minimal gatekeeping, M (SD) = 12.45 (9.14)

• FREED cohort with complex gatekeeping, M (SD) = 16.89 (10.21)

• TAU cohort, M (SD) = 16.16 (10.63)

DUED

Flynn et al., (2020)⁵

Results of DUED (months) for participants diagnosed with AN, BN, BED, OSFED from the total FREED cohort (n analyzed = 278)

• AN, M (SD) = 17.50 (10.62)

• BN, M (SD) = 20.26 (10.45)

• BED, M (SD) = 18.67 (8.33)

• OSFED, M (SD) = 16.30 (9.84)

• Total, M (SD) = 17.85 (10.38)

Results of DUED (months) for participants diagnosed with AN, BN, BED, OSFED under optimal conditions from FREED cohort (n analyzed = 157)

• AN, M (SD) = 14.02 (9.08)

• BN, M (SD) = 19.72 (10.76)

• BED, M (SD) = 18.67 (8.33)

• OSFED, M (SD) = 14.05 (8.37)

• Total, M (SD) = 15.95 (9.74)

Results of DUED (months) for participants diagnosed with AN, BN, BED, OSFED from TAU cohort (n analyzed = 224)

• AN, M (SD) = 18.57 (11.27)

• BN, M (SD) = 23.05 (9.35)

• BED, M (SD) = 18.00 (11.40)

• OSFED, M (SD) = 19.90 (12.64)

• Total, M (SD) = 19.98 (11.13)

Between group comparison of total DUED (months) from total FREED cohort (n = 278) vs TAU cohort (n = 224)

• P < 0.05

• 95% CI = −4.23 to −0.31

Between group comparison of total DUED (months) from FREED cohort under optimal conditions (n = 157) vs TAU cohort (n = 224)

• P < 0.001

• 95% CI = −6.04 to −1.68•

Brown et al., (2016)⁸

Mean DUED (months) for all FREED cohort (n analyzed = 51), FREED cohort with minimal gatekeeping (n analyzed = 14), FREED cohort with complex gatekeeping (n analyzed = 37), and TAU cohort (n analyzed = 65)

• All FREED cohort, M (SD) = 16.39 (10.08)

• FREED cohort with minimal gatekeeping, M (SD) = 13.04 (9.29)

• FREED cohort with complex gatekeeping, M (SD) = 17.66 (10.20)

• TAU cohort, M (SD) = 19.09 (11.67)

• P = 0.07 for FREED cohort with minimal gatekeeping vs TAU cohort

BMI score

Richards et al., (2023)¹

Results of all AN participants from the FREED-Up cohort (n analyzed = 96 at baseline [T1]; 76 at 6 months [T2]; 66 at 12 months [T3])

• BMI score, M (SD) at T1 vs T2; MD = 16.42 (1.19) vs 17.67 (1.77); −1.25 (P < 0.001)

• BMI score, M (SD) at T1 vs T3; MD = 16.42 (1.19) vs 18.43 (2.23); −2.01 (P < 0.001)

• % (n) of participants with AN above BMI threshold (> 18.5 kg/m²) at T1, T2, and T3 = 0% (0); 33% (25); 52% (34)

Results of all AN participants from the FREED-4-All cohort (n analyzed = 429 at baseline [T1]; 88 at post-treatment [T2])

• BMI score, M (SD) at T1 vs T2; MD = 17.41 (2.24) vs 19.08 (2.55); −1.67 (P < 0.001)

• % (n) of participants with AN above BMI threshold (> 18.5 kg/m²) at T1 and T2 = 22% (93); 59% (52)

Austin et al., (2022)²

Estimated mean BMI score (kg/m²) of AN participants for FREED cohort (n = 117) vs TAU cohort (n = 116)

• M (95% CI) = 18.65 (18.27 to 19.03) vs 17.33 (16.75 to 17.90)

• MD (95% CI) = 1.32 (0.63 to 2.02)

Estimated mean BMI points gained for AN participants at baseline (T1) to 12 months (T4) for FREED cohort vs TAU cohort^a

• M (95% CI) = 2.09 (1.66 to 2.53) vs 1.22 (0.59 to 1.86)

Proportion of participants who were weight recovered (BMI > 18.5 kg/m²) at each time point, n/N (%)

• FREED cohort vs TAU cohort at baseline = 5/117 (4.35%) vs 5/78 (6.4%)

• FREED cohort vs TAU cohort at 3 months = 18/105 (17.1%) vs 8/59 (13.6%)

• FREED cohort vs TAU cohort at 6 months; P value = 31/92 (33.7%) vs 8/55 (14.5%); P = 0.011

• FREED cohort vs TAU cohort at 12 months; P value = 42/79 (53.2%) vs 5/28 (17.9%); P < 0.001

Radunz et al., (2021)³

Results of mean BMI score (kg/m²) from baseline (n analyzed = 70) and end of treatment (n analyzed = 43)

• Baseline BMI score, M (SD) = 22.14 (0.85)

• End of treatment BMI score, M (SD) = 23.11 (0.85)

• Between group difference, d (95% CI) = −0.21 (−0.72 to 0.30)

• P < 0.001

Fukutomi et al., (2019)⁶

Results of mean BMI (kg/m²) at final time point (24-month follow-up) for FREED-AN cohort (n analyzed = 11) vs TAU-AN cohort (n analyzed = 8)

• FREED-AN, M (95% CI) = 19.2 (18.21 to 20.16)

• TAU-AN, M (95% CI) = 18.0 (16.90 to 19.15)

• MD (95% CI) = 1.1 (−0.44 to 2.66)

Mean BMI increase (kg/m²) from assessment to final time point (24-month follow-up) for FREED-AN cohort (n analyzed = 11) vs TAU-AN cohort (n analyzed = 8)

• FREED-AN, M (95% CI) = 2.7 (1.57 to 3.85)

• TAU-AN, M (95% CI) = 1.9 (0.75 to 3.14)

• P = 0.06

Proportion of participants who were weight recovered (BMI > 18.5 kg/m²) between 12- and 24-month follow-up for FREED-AN cohort and TAU-AN cohort

• FREED-AN, n/N (%) = 12/17 (71%)

• TAU-AN, n/N (%) = 2/9 (22%)

• P = 0.02

Proportion of participants who were weight recovered (BMI > 18.5 kg/m²) across all time points for FREED-AN cohort and TAU-AN cohort

• FREED-AN, n/N (%) = 13/22 (59%)

• TAU-AN, n/N (%) = 5/28 (21%)

• P = 0.003

McClelland et al., (2018)⁷

Results of mean BMI score (kg/m²) from all participants from the FREED cohort (n analyzed = 50 at baseline [T1]; 45 at 3 months [T2]; 35 at 6 months [T3]; 30 at 12 months [T4])

• T1 score, M (SD) = 19.8 (3.7)

• T2 score, M (SD) = 19.7 (3.3)

• T3 score, M (SD) = 19.9 (2.9)

• T4 score, M (SD) = 20.7 (3.2)

Mean change in BMI score (kg/m²) from T1 (n analyzed = 50) to T2 (n analyzed = 45) of all participants from the FREED cohort

• M (95% CI) = 0.16 (−0.40 to 0.71)

• P = 1.00

Mean change in BMI score (kg/m²) from T1 (n analyzed = 50) to T3 (n analyzed = 35) of all participants from the FREED cohort

• M (95% CI) = 0.69 (−0.02 to 1.41)

• P = 0.064

Mean change in BMI score (kg/m²) from T1 (n analyzed = 50) to T4 (n analyzed = 30) of all participants from the FREED cohort

• M (95% CI) = 1.20 (0.29 to 2.12)

• P = 0.004

Mean change in BMI score (kg/m²) from T3 (n analyzed = 35) to T4 (n analyzed = 30) of all participants from the FREED cohort

• M (95% CI) = 0.51 (−0.16 to 1.18)

• P = 0.229

Psychological distress

Richards et al., (2023)¹

Results of all participants from the FREED-Up cohort (n analyzed = 277 at baseline [T1]; 182 at 6 months [T2]; 175 at 12 months [T3])

• CORE-10/OM score, M (SD) at T1 vs T2; MD = 1.97 (0.75) vs 1.45 (0.74); 0.52 (P < 0.001)

• CORE-10/OM score, M (SD) at T1 vs T3; MD = 1.97 (0.75) vs 1.39 (0.85); 0.58 (P < 0.001)

Results of all participants from the FREED-4-All cohort (n analyzed = 577 at baseline [T1]; 76 at post-treatment [T2])

• CORE-10/OM score, M (SD) at T1 vs T2; MD = 1.93 (0.72) vs 1.42 (0.83); 0.51 (P < 0.001)

Austin et al., (2022)²

Results of all participants from the FREED cohort (n analyzed = 277 at baseline [T1]; 216 at 3 months [T2]; 182 at 6 months [T3]; 175 at 12 months [T4])

• CORE-10/OM score MD (95% CI) at T1 vs T2; SE (P value) = −2.59 (−3.42 to −1.77); 0.42 (P < 0.001)

• CORE-10/OM score MD (95% CI) at T2 vs T3; SE (P value) = −2.49 (−3.39 to −1.58); 0.46 (P < 0.001)

• CORE-10/OM score MD (95% CI) at T3 vs T4; SE (P value) = −0.94 (−1.8 to 0.02); 0.49 (P = 0.054)

• CORE-10/OM score MD (95% CI) at T1 vs T4; SE (P value) = −6.02 (−7.08 to −4.95); 0.54 (P < 0.001)

McClelland et al., (2018)⁷

Results of mean CORE-10 score from all participants from the FREED cohort (n analyzed = 53 at baseline [T1]; 37 at 3 months [T2]; 32 at 6 months [T3]; 25 at 12 months [T4])

• T1 score, M (SD) = 19.8 (8.2)

• T2 score, M (SD) = 16.1 (7.0)

• T3 score, M (SD) = 14.2 (7.8)

• T4 score, M (SD) = 15.4 (8.3)

Mean change in CORE-10 score from T1 (n analyzed = 53) to T2 (n analyzed = 37) of all participants from the FREED cohort

• M (95% CI) = −3.61 (−6.81 to −0.42)

• P = 0.019

Mean change in CORE-10 score from T1 (n analyzed = 53) to T3 (n analyzed = 32) of all participants from the FREED cohort

• M (95% CI) = −5.57 (−9.00 to −2.13)

• P = 0.001

Mean change in CORE-10 score from T1 (n analyzed = 53) to T4 (n analyzed = 25) of all participants from the FREED cohort

• M (95% CI) = −5.43 (−9.33 to −1.54)

• P = 0.002

Mean change in CORE-10 score from T3 (n analyzed = 32) to T4 (n analyzed = 25) of all participants from the FREED cohort

• M (95% CI) = −0.13 (−3.83 to −4.09)

• P = 1.00

Psychological impairment due to ED

Austin et al., (2022)²

Results of all participants from the FREED cohort (n analyzed = 276 at baseline [T1]; 214 at 3 months [T2]; 180 at 6 months [T3]; 173 at 12 months [T4])

• CIA score MD (95% CI) at T1 vs T2; SE (P value) = −5.35 (−6.59 to −3.90); 0.67 (P < 0.001)

• CIA score MD (95% CI) at T2 vs T3; SE (P value) = −3.85 (−5.31 to −2.38); 0.75 (P < 0.001)

• CIA score MD (95% CI) at T3 vs T4; SE (P value) = −4.26 (−5.82 to −2.69); 0.80 (P < 0.001)

• CIA score MD (95% CI) at T1 vs T4; SE (P value) = −13.35 (−15.31 to −11.38); 1.00 (P < 0.001)

Radunz et al., (2021)³

Results of mean CIA score from baseline (n analyzed = 96) and end of treatment (n analyzed = 30)

• Baseline CIA score, M (SD) = 35.23 (1.66)

• End of treatment CIA score, M (SD) = 14.53 (1.66)

• Between group difference, d (95% CI) = 2.32 (1.66 to 2.97)

• P < 0.001

McClelland et al., (2018)⁷

Results of mean CIA score from all participants from the FREED cohort (n analyzed = 52 at baseline [T1]; 32 at 3 months [T2]; 33 at 6 months [T3]; 26 at 12 months [T4])

• T1 score, M (SD) = 1.8 (0.62)

• T2 score, M (SD) = 1.67 (0.66)

• T3 score, M (SD) = 1.20 (0.69)

• T4 score, M (SD) = 1.0 (0.71)

Mean change in CIA score from T1 (n analyzed = 52) to T2 (n analyzed = 32) of all participants from the FREED cohort

• M (95% CI) = −0.18 (−0.47 to 0.10)

• P = 0.102

Mean change in CIA score from T1 (n analyzed = 52) to T3 (n analyzed = 33) of all participants from the FREED cohort

• M (95% CI) = −0.66 (−0.95 to −0.36)

• P = 0.001

Mean change in CIA score from T1 (n analyzed = 52) to T4 (n analyzed = 26) of all participants from the FREED cohort

• M (95% CI) = −0.98 (−1.33 to −0.63)

• P = 0.001

Mean change in CIA score from T3 (n analyzed = 33) to T4 (n analyzed = 26) of all participants from the FREED cohort

• M (95% CI) = −0.33 (−0.66 to 0.00)

• P = 0.053

Depression, anxiety, and stress

Austin et al., (2022)²

Results of all participants from the FREED cohort (n analyzed = 278 at baseline [T1]; 216 at 3 months [T2]; 182 at 6 months [T3]; 175 at 12 months [T4])

• DASS-21 score MD (95% CI) at T1 vs T2; SE (P value) = −5.06 (−6.54 to −3.57); 0.76 (P < 0.001)

• DASS-21 score MD (95% CI) at T2 vs T3; SE (P value) = −3.54 (−5.16 to −1.92); 0.83 (P < 0.001)

• DASS-21 score MD (95% CI) at T3 vs T4; SE (P value) = −3.10 (−4.82 to −1.38); 0.88 (P < 0.001)

• DASS-21 score MD (95% CI) at T1 vs T4; SE (P value) = −11.70 (−13.77 to −9.62); 1.05 (P < 0.001)

Radunz et al., (2021)³

Results of mean depression score measured by DASS-21 from baseline (n analyzed = 96) and end of treatment (n analyzed = 30)

• Baseline depression score, M (SD) = 1.94 (0.13)

• End of treatment depression score, M (SD) = 0.82 (0.13)

• Between group difference, d (95% CI) = 1.60 (1.02 to 2.18)

• P < 0.001

Results of mean anxiety score measured by DASS-21 from baseline (n analyzed = 96) and end of treatment (n analyzed = 30)

• Baseline anxiety score, M (SD) = 1.62 (0.15)

• End of treatment anxiety score, M (SD) = 0.90 (0.15)

• Between group difference, d (95% CI) = 0.89 (0.36 to 1.42)

• P < 0.001

Results of mean stress score measured by DASS-21 from baseline (n analyzed = 96) and end of treatment (n analyzed = 30)

• Baseline stress score, M (SD) = 1.94 (0.10)

• End of treatment stress score, M (SD) = 1.18 (0.10)

• Between group difference, d (95% CI) = 1.14 (0.85 to 1.98)

• P < 0.001

McClelland et al., (2018)⁷

Results of mean DASS-21 score from all participants from the FREED cohort (n analyzed = 51 at baseline [T1]; 37 at 3 months [T2]; 33 at 6 months [T3]; 26 at 12 months [T4])

• T1 score, M (SD) = 32.7 (13.7)

• T2 score, M (SD) = 24.3 (15.5)

• T3 score, M (SD) = 21.1 (14.6)

• T4 score, M (SD) = 23.0 (14.0)

Mean change in DASS-21 score from T1 (n analyzed = 51) to T2 (n analyzed = 37) of all participants from the FREED cohort

• M (95% CI) = −9.09 (−14.94 to −3.25)

• P = 0.001

Mean change in DASS-21 score from T1 (n analyzed = 51) to T3 (n analyzed = 33) of all participants from the FREED cohort

• M (95% CI) = −12.21 (−18.24 to −6.17)

• P = 0.001

Mean change in DASS-21 score from T1 (n analyzed = 51) to T4 (n analyzed = 26) of all participants from the FREED cohort

• M (95% CI) = −12.33 (−18.92 to −5.74)

• P = 0.001

Mean change in DASS-21 score from T3 (n analyzed = 33) to T4 (n analyzed = 26) of all participants from the FREED cohort

• M (95% CI) = −0.12 (−5.49 to −5.27)

• P = 1.00

Expressed emotion

Austin et al., (2022)²

Results of all participants from the FREED cohort (n analyzed = 278 at baseline [T1]; 216 at 3 months [T2]; 180 at 6 months [T3]; 175 at 12 months [T4])

• LEE score MD (95% CI) at T1 vs T2; SE (P value) = −2.38(−3.65 to −1.11); 0.65 (P < 0.001)

• LEE score MD (95% CI) at T2 vs T3; SE (P value) = −0.77 (−2.16 to 0.63); 0.71 (P = 0.28)

• LEE score MD (95% CI) at T3 vs T4; SE (P value) = −0.87 (−2.34 to 0.61); 0.75 (P = 0.25)

• LEE score MD (95% CI) at T1 vs T4; SE (P value) = −4.02 (−5.64 to −2.39); 0.82 (P < 0.001)

McClelland et al., (2018)⁷

Results of mean LEE score from all participants from the FREED cohort (n analyzed = 51 at baseline [T1]; 37 at 3 months [T2]; 31 at 6 months [T3]; 26 at 12 months [T4])

• T1 score, M (SD) = 17.3 (11.0)

• T2 score, M (SD) = 14.9 (9.9)

• T3 score, M (SD) = 12.0 (7.4)

• T4 score, M (SD) = 12.2 (12.3)

Mean change in LEE score from T1 (n analyzed = 51) to T2 (n analyzed = 37) of all participants from the FREED cohort

• M (95% CI) = −1.45 (−4.96 to −2.06)

• P = 1.00

Mean change in LEE score from T1 (n analyzed = 51) to T3 (n analyzed = 31) of all participants from the FREED cohort

• M (95% CI) = −3.52 (−7.35 to 0.32)

• P = 0.088

Mean change in LEE score from T1 (n analyzed = 51) to T4 (n analyzed = 26) of all participants from the FREED cohort

• M (95% CI) = −3.86 (−8.17 to −0.46)

• P = 0.102

Mean change in LEE score from T3 (n analyzed = 31) to T4 (n analyzed = 26) of all participants from the FREED cohort

• M (95% CI) = −0.34 (−4.66 to 3.98)

• P = 1.00

Function and wellbeing

Austin et al., (2022)²

Results of all participants from the FREED cohort (n analyzed = 275 at baseline [T1]; 216 at 3 months [T2]; 178 at 6 months [T3]; 175 at 12 months [T4])

• PSYCHLOPS score MD (95% CI) at T1 vs T2; SE (P value) = −3.79 (−4.35 to −3.24); 0.28 (P < 0.001)

• PSYCHLOPS score MD (95% CI) at T2 vs T3; SE (P value) = −1.42 (−22.03 to −0.81); 0.31 (P = 0.28)

• PSYCHLOPS score MD (95% CI) at T3 vs T4; SE (P value) = −1.71 (−2.35 to −1.07); 0.33 (P < 0.001)

• PSYCHLOPS score MD (95% CI) at T1 vs T4; SE (P value) = −6.92 (−7.67 to −6.17); 0.38 (P < 0.001)

Work and social adjustment

Austin et al., (2022)²

Results of all participants from the FREED cohort (n analyzed = 278 at baseline [T1]; 216 at 3 months [T2]; 182 at 6 months [T3]; 175 at 12 months [T4])

• WSAS score MD (95% CI) at T1 vs T2; SE (P value) = −3.14 (−4.19 to −2.09); 0.54 (P < 0.001)

• WSAS score MD (95% CI) at T2 vs T3; SE (P value) = −2.94 (−4.09 to −1.79); 0.58 (P < 0.001)

• WSAS score MD (95% CI) at T3 vs T4; SE (P value) = −2.07 (−3.29 to −0.86); 0.62 (P < 0.001)

• WSAS score MD (95% CI) at T1 vs T4; SE (P value) = −8.15 (−9.67 to −6.62); 0.77 (P < 0.001)

McClelland et al., (2018)⁷

Results of mean WSAS score from all participants from the FREED cohort (n analyzed = 51 at baseline [T1]; 36 at 3 months [T2]; 32 at 6 months [T3]; 26 at 12 months [T4])

• T1 score, M (SD) = 21.0 (9.7)

• T2 score, M (SD) = 18.1 (9.7)

• T3 score, M (SD) = 14.5 (10.5)

• T4 score, M (SD) = 11.8 (10.3)

Mean change in WSAS score from T1 (n analyzed = 51) to T2 (n analyzed = 36) of all participants from the FREED cohort

• M (95% CI) = −2.87 (−7.07 to 1.34)

• P = 0.354

Mean change in WSAS score from T1 (n analyzed = 51) to T3 (n analyzed = 32) of all participants from the FREED cohort

• M (95% CI) = −7.16 (−11.74 to −2.58)

• P = 0.001

Mean change in WSAS score from T1 (n analyzed = 51) to T4 (n analyzed = 26) of all participants from the FREED cohort

• M (95% CI) = −10.21 (−15.50 to −2.58)

• P = 0.001

Mean change in WSAS score from T3 (n analyzed = 32) to T4 (n analyzed = 26) of all participants from the FREED cohort

• M (95% CI) = −3.04 (−8.13 to 2.05)

• Z-score = −1.49

• P = 0.541

• SES = −0.31

Wait times

Richards et al., (2021)⁴

Proportion of all participants from FREED-Up cohort with an attempted engagement call ≤ 48 hours

• AN participants, n/N (%) = 93/101 (92%)

• BN/BED participants, n/N (%) = 53/59 (90%)

• OSFED, n/N (%) = 63/74 (85%)

• All participants, n/N (%) = 209/234 (89%)

• Between group comparison, P = 0.34

Proportion of participants diagnosed with optimal conditions from FREED-Up cohort with an attempted engagement call ≤ 48 hours

• AN participants, n/N (%) = 50/54 (93%)

• BN/BED participants, n/N (%) = 42/47 (89%)

• OSFED, n/N (%) = 36/42 (86%)

• All participants, n/N (%) = 128/143 (90%)

• Between group comparison, P = 0.90

Proportion of all participants from FREED-Up cohort that received an engagement call ≤ 48 hours

• AN participants, n/N (%) = 53/100 (53%)

• BN/BED participants, n/N (%) = 32/66 (49%)

• OSFED, n/N (%) = 36/75 (48%)

• All participants, n/N (%) = 121/241 (50%)

• Between group comparison, P = 0.76

Proportion of participants diagnosed with optimal conditions from FREED-Up cohort that received an engagement call ≤ 48 hours

• AN participants, n/N (%) = 26/55 (47%)

• BN/BED participants, n/N (%) = 24/50 (48%)

• OSFED, n/N (%) = 20/42 (48%)

• All participants, n/N (%) = 70/147 (48%)

• Between group comparison, P = 0.31

Proportion of all participants from FREED-Up cohort that were offered an assessment ≤ 2 weeks

• AN participants, n/N (%) = 54/104 (52%)

• BN/BED participants, n/N (%) = 36/63 (57%)

• OSFED, n/N (%) = 36/78 (46%)

• All participants, n/N (%) = 126/245 (51%)

• Between group comparison, P < 0.01

Proportion of participants diagnosed with optimal conditions from FREED-Up cohort that were offered an assessment ≤ 2 weeks

• AN participants, n/N (%) = 35/55 (64%)

• BN/BED participants, n/N (%) = 31/48 (65%)

• OSFED, n/N (%) = 20/42 (48%)

• All participants, n/N (%) = 86/145 (59%)

• Between group comparison, P < 0.01

Proportion of all participants from FREED-Up cohort that received an assessment ≤ 2 weeks or 4 weeks

• AN participants, n/N (%) = 50/104 (46%) or 78/109 (72%)

• BN/BED participants, n/N (%) = 30/69 (44%) or 49/69 (71%)

• OSFED, n/N (%) = 30/81 (37%) or 61/81 (75%)

• All participants, n/N (%) = 110/259 (43%) or 188/259 (73%)

• Between group comparison for assessment received ≤ 2 weeks, P = 0.47

• Comparison to TAU cohort^a for assessment received ≤ 2 weeks, P < 0.001

Proportion of participants diagnosed with optimal conditions from FREED-Up cohort that received an assessment ≤ 2 weeks or 4 weeks

• AN participants, n/N (%) = 30/55 (55%) or 45/55 (82%)

• BN/BED participants, n/N (%) = 28/55 (55%) or 43/51 (84%)

• OSFED, n/N (%) = 17/43 (40%) or 38/43 (88%)

• All participants, n/N (%) = 75/149 (50%) or 126/149 (85%)

• Between group comparison for assessment received ≤ 2 weeks, P < 0.01

Proportion of all participants from FREED-Up cohort that were offered treatment ≤ 4 weeks

• AN participants, n/N (%) = 40/100 (40%)

• BN/BED participants, n/N (%) = 20/63 (32%)

• OSFED, n/N (%) = 18/76 (24%)

• All participants, n/N (%) = 78/239 (33%)

• Between group comparison, P = 0.07

Proportion of participants diagnosed with optimal conditions from FREED-Up cohort that were offered treatment ≤ 4 weeks

• AN participants, n/N (%) = 23/52 (44%)

• BN/BED participants, n/N (%) = 17/46 (37%)

• OSFED, n/N (%) = 10/42 (24%)

• All participants, n/N (%) = 50/140 (36%)

• Between group comparison, P = 0.29

Proportion of all participants from FREED-Up cohort that received treatment ≤ 4 weeks or 8 weeks

• AN participants, n/N (%) = 28/108 (26%) or 64/108 (59%)

• BN/BED participants, n/N (%) = 15/69 (22%) or 41/69 (59%)

• OSFED, n/N (%) = 17/79 (22%) or 42/79 (53%)

• All participants, n/N (%) = 60/256 (23%) or 147/256 (57%)

• Between group comparison for assessment received ≤ 4 weeks, P = 0.72

• Comparison to TAU cohort^b for assessment received ≤ 2 weeks, P < 0.001

Proportion of participants diagnosed with optimal conditions from FREED-Up cohort that received treatment ≤ 4 weeks or 8 weeks

• AN participants, n/N (%) = 17/54 (32%) or 40/54 (74%)

• BN/BED participants, n/N (%) = 14/51 (28%) or 35/51 (69%)

• OSFED, n/N (%) = 10/41 (24%) or 26/41 (63%)

• All participants, n/N (%) = 41/146 (28%) or 101/146 (69%)

• Between group comparison for assessment received ≤ 2 weeks, P = 0.04

Flynn et al., (2020)⁵

Wait time to assessment (weeks) for participants diagnosed with AN, BN, BED, OSFED from the total FREED cohort (n analyzed = 278)

• AN, M (SD) = 3.27 (2.65)

• BN, M (SD) = 3.45 (3.10)

• BED, M (SD) = 3.10 (0.54)

• OSFED, M (SD) = 4.18 (5.42)

• Total, M (SD) = 3.58 (3.79)

Wait time to assessment (weeks) for participants diagnosed with AN, BN, BED, OSFED under optimal conditions from FREED cohort (n analyzed = 157)

• AN, M (SD) = 2.54 (1.70)

• BN, M (SD) = 2.40 (1.56)

• BED, M (SD) = 3.10 (0.54)

• OSFED, M (SD) = 2.70 (1.77)

• Total, M (SD) = 2.56 (1.64)

Wait time to assessment (weeks) for participants diagnosed with AN, BN, BED, OSFED from TAU cohort (n analyzed = 224)

• AN, M (SD) = 5.41 (5.64)

• BN, M (SD) = 6.59 (4.80)

• BED, M (SD) = 14.0 (2.13)

• OSFED, M (SD) = 11. 50 (19.71)

• Total, M (SD) = 6.72 (8.70)

Between group comparison of wait time to assessment (weeks) from total FREED cohort (n = 278) vs TAU cohort (n = 224)

• P < 0.001

• 95% CI = −4.28 to −2.00

Between group comparison of wait time to assessment (weeks) from FREED cohort under optimal conditions (n = 157) vs TAU cohort (n = 224)

• P < 0.001

• 95% CI = −5.54 to −2.78

Wait time to treatment (weeks) for participants diagnosed with AN, BN, BED, OSFED from the total FREED cohort (n analyzed = 278)

• AN, M (SD) = 7.41 (4.78)

• BN, M (SD) = 7.72 (5.35)

• BED, M (SD) = 7.24 (3.19)

• OSFED, M (SD) = 9.27 (3.19)

• Total, M (SD) = 8.06 (5.73)

Wait time to treatment (weeks) for participants diagnosed with AN, BN, BED, OSFED under optimal conditions from FREED cohort (n analyzed = 157)

• AN, M (SD) = 5.81 (2.82)

• BN, M (SD) = 6.12 (2.77)

• BED, M (SD) = 7.24 (3.19)

• OSFED, M (SD) = 7.31 (3.97)

• Total, M (SD) = 6.36 (3.21)

Wait time to treatment (weeks) for participants diagnosed with AN, BN, BED, OSFED from TAU cohort (n analyzed = 224)

• AN, M (SD) = 18.41 (15.36)

• BN, M (SD) = 21.34 (13.71)

• BED, M (SD) = 19.54 (3.01)

• OSFED, M (SD) = 26.80 (22.78)

• Total, M (SD) = 20.76 (16.60)

Between group comparison of wait time to treatment (weeks) from total FREED cohort (n = 278) vs TAU cohort (n = 224)

• P < 0.001

• 95% CI = −14.86 to −10.54

Between group comparison of wait time to treatment (weeks) from FREED cohort under optimal conditions (n = 157) vs TAU cohort (n = 224)

• P < 0.001

• 95% CI = −17.08 to −11.70

McClelland et al., (2018)⁷

Wait time median (days) from referral to assessment for FREED Cohort (n = 56) and TAU cohort (n = 86)

• FREED cohort, median (IQR) = 42.5 (23 to 66)

• TAU cohort, median (IQR) = 62 (41 to 98)

• RR (95% CI) = 0.74 (0.53 to 1.05)

• P = 0.084

Wait time median (days) from assessment to treatment for FREED Cohort (n = 56) and TAU cohort (n = 86)

• FREED cohort, median (IQR) = 20 (11 to 31)

• TAU cohort, median (IQR) = 34 (16 to 125)

• RR (95% CI) = 0.34 (0.23 to 0.49)

• P < 0.001

Brown et al., (2016)⁸

Mean wait time to assessment (weeks) for all FREED cohort (n analyzed = 51), FREED cohort with minimal gatekeeping (n analyzed = 14), FREED cohort with complex gatekeeping (n analyzed = 37), and TAU cohort (n analyzed = 89)

• All FREED cohort, M (SD) = 6.44 (5.38)

• FREED cohort with minimal gatekeeping, M (SD) = 3.67 (3.35)

• FREED cohort with complex gatekeeping, M (SD) = 7.48 (5.66)

• TAU cohort, M (SD) = 9.94 (5.87)

• P < 0.001 for all FREED cohort vs TAU cohort

• P < 0.001 for FREED cohort with minimal gatekeeping vs TAU cohort

• P < 0.05 for FREED cohort with complex gatekeeping vs TAU cohort

Mean wait time to treatment (weeks) for all FREED cohort (n analyzed = 51), FREED cohort with minimal gatekeeping (n analyzed = 14), FREED cohort with complex gatekeeping (n analyzed = 37), and TAU cohort (n analyzed = 65)

• All FREED cohort, M (SD) = 9.59 (5.78)

• FREED cohort with minimal gatekeeping, M (SD) = 6.25 (3.63)

• FREED cohort with complex gatekeeping, M (SD) = 10.86 (5.97)

• TAU cohort, M (SD) = 19.87 (15.11)

• P < 0.001 for all FREED cohort vs TAU cohort

• P < 0.001 for FREED cohort with minimal gatekeeping vs TAU cohort

• P < 0.001 for FREED cohort with complex gatekeeping vs TAU cohort

Mean wait time from assessment to treatment (weeks) for all FREED cohort (n analyzed = 51), FREED cohort with minimal gatekeeping (n analyzed = 14), FREED cohort with complex gatekeeping (n analyzed = 37), and TAU cohort (n analyzed = 65)

• All FREED cohort, M (SD) = 3.16 (2.19)

• FREED cohort with minimal gatekeeping, M (SD) = 2.58 (1.41)

• FREED cohort with complex gatekeeping, M (SD) = 3.38 (2.40)

• TAU cohort, M (SD) = 10.07 (11.70)

• P < 0.001 for all FREED cohort vs TAU cohort

• P < 0.001 for FREED cohort with minimal gatekeeping vs TAU cohort

• P < 0.001 for FREED cohort with complex gatekeeping vs TAU cohort

Service use

Austin et al., (2022)²

Proportion of treatment completion for FREED cohort (n = 270) vs TAU cohort (n = 157)

• FREED cohort, n (%) = 189 (70%)

• TAU cohort, n (%) = 103 (65.6%)

• P = 0.35

Number of treatment sessions attended by FREED cohort vs TAU cohort across 12-month follow-up period

• FREED cohort, M (SD) = 18.64 (12.64)

• TAU cohort, M (SD) = 16.67 (15.01)

• P = 0.16

Number of participants requiring addition intensive treatment for FREED cohort (n = 272) vs TAU cohort (n = 169) across 12-month follow-up period

• FREED cohort, n (%) = 18 (6.6%)

• TAU cohort, n (%) = 21 (12.4%)

• P = 0.037

Number of days in intensive treatment for FREED cohort vs TAU cohort across 12-month follow-up period

• FREED cohort, M days (SD) = 7.03 (34.55)

• TAU cohort, M days (SD) = 17.93 (58.39)

• P = 0.02

Flynn et al., (2020)⁵

Treatment uptake after assessment for total FREED cohort (n = 278) vs TAU cohort (n = 224)

• FREED cohort, n (%) = 272 (97.84%)

• TAU cohort, n (%) = 160 (71.43%)

• P < 0.01

Fukutomi et al., (2019)⁶

Mean number of treatment sessions attended at 24-month follow-up for FREED-AN cohort (n = 22) and TAU-AN cohort (n = 35)

• FREED-AN, M (SD) = 30.5 (17.0)

• TAU-AN, M (SD) = 20.5 (15.4)

Number of participants needing intensive treatment at 24-month follow-up for FREED-AN cohort vs TAU-AN cohort

• FREED-AN, n/N (%) = 5/22 (23%)

• TAU-AN, n/N (%) = 9/28 (32%)

• P = 0.54

McClelland et al., (2018)⁷

Number of participants that took up treatment after assessment for FREED cohort vs TAU cohort

• FREED cohort, n/N (%) = 56/56 (100%)

• TAU cohort, n/N (%) = 64/86 (74%)

• P < 0.001

Median number of sessions attended for FREED cohort (n = 56) vs TAU cohort (n = 64)

• FREED cohort, median (IQR) = 21.5 (9 to 29.5)

• TAU cohort, median (IQR) = 16 (8 to 24)

• RR (95% CI) = 1.16 (0.80 to 1.70)

Number of participants that completed treatment for FREED cohort and TAU cohort

• FREED cohort, n/N (%) = 40/56 (71%)

• TAU cohort, n/N (%) = 45/64 (71%)

Number of participants that required additional intensive treatment for FREED cohort vs TAU cohort

• FREED cohort, n/N (%) = 5/56 (8.9%)

• TAU cohort, n/N (%) = 9/64 (14.1%)

• P = 0.999

Brown et al., (2016)⁸

Number of participants that took up treatment after assessment for FREED cohort vs TAU cohort

• FREED cohort, n/N (%) = 51/51 (100%)

• TAU cohort, n/N (%) = 65/89 (73%)

• X² = 16.60

• P < 0.001

EDI

Godart et al., (2022)⁹

Between group comparison from all participants with AN from the FT-S with TAU cohort (n = 30) vs TAU (n = 30); 3 years after the end of treatment

• EDI total score, FT-S with TAU vs TAU; M (SD) = 47.2 (36.9) vs 48.3 (39.1); −0.2 (P = 0.860); absolute effect size (95%CI) = −1.1 (−21.1 to 18.87); relative effect size^a (95%CI) = −0.03 (−0.5 to 0.5)

Herpertz-Dahlmann et al., (2021)¹⁰

Results of EDI-2 global score of all participants with AN that received HoT (n analyzed = 22 at admission [T1]; 21 at the start of HoT [T2]; 21 at the end of HoT [T3]; 21 at 1-year follow-up [T4])

• M (SD) at T1 vs T4; M = 280.68 (53.21) vs 222.42 (52.23)

• T2 score, M (SD) = 261.71 (57.37)

• T3 score, M (SD) = 244.90 (52.91)

Hurst et al., (2019)¹²

Results of EDI-3 global score of all participants with AN that received FBT with CBT-P (n = 21); at FBT phase one commencement (T1); at FBT phase two and CBT-P commencement (T2); after completion of CBT-P (T3); after FBT with CBT-P completion (T4)

• M (SD) at T1 vs T2; M = 56.2 (17.6) vs 50.0 (21.8); 1.55 (d = 0.31)

• M (SD) at T1 vs T3; M = 56.2 (17.6) vs 41.7 (24.2); 3.18 (d = 0.69); P < 0.01

• M (SD) at T1 vs T4; M = 56.2 (17.6) vs 36.1 (26.5); 3.64 (d = 0.90); P < 0.01

EDE-Q

Herpertz-Dahlmann et al., (2021)¹⁰

Results of EDE global score of all participants with AN that received HoT (n analyzed = 22 at admission [T1]; 21 at the start of HoT [T2]; 21 at the end of HoT [T3]; 21 at 1-year follow-up [T4])

• M (SD) at T1 vs T4; M = 4.04 (1.05) vs 1.53 (1.15); P < 0.001

• T2 score, M (SD) = NR

• T3 score, M (SD) = 1.72 (1.01)

AN symptom remission

Hurst et al., (2019)¹²

Results of all participants with AN (n = 19) after the completion of FBT with CBT-P

• Full remission, n (%) = 11 (57%)

• Partial remission, n (%) = 8 (43%)

BMI score

Godart et al., (2022)⁹

Between group comparison of BMI score of all participants with AN from the FT-S with TAU cohort (n = 30) vs TAU (n = 30); 3 years after the end of treatment

• FT-S with TAU vs TAU; M (SD) = 18.61 (2.13) vs 17.91 (2.72); 1.2 (P = 0.268); absolute effect size (95%CI) = 0.706 (−0.56 to 1.97); relative effect size^a (95%CI) = 0.288 (−0.219 to 0.797)

Herpertz-Dahlmann et al., (2021)¹⁰

Results of BMI score of all participants with AN that received HoT (n analyzed = 22 at admission [T1]; 21 at the start of HoT [T2]; 21 at the end of HoT [T3]; 21 at 1-year follow-up [T4])

• M (SD) at T1 vs T4; M = 16.26 (1.15) vs 19.72 (1.32); P < 0.001

• T2 score, M (SD) = 18.35 (1.01)

• T3 score, M (SD) = 19.66 (1.03)

Rosling et al., (2016)¹³

Results from all participants with AN that received FB specialized out-patient service (n analyzed = 31 at baseline [T1]; 1 at 1-year follow-up [T2])

BMI score, M ± SD:

• T1, M = 15.1 ± 1.22

• T2, M = 14.1

BMI percentile

Godart et al., (2022)⁹

Between group comparison from all participants with AN from the FT-S with TAU cohort (n analyzed = 30) vs TAU (n analyzed = 30), 3 years after the end of treatment

• BMI ≥ 10th percentile, FT-S with TAU vs TAU; n (%) = 22/30 (73.3) vs 15/30 (50.0)^b; 3.4 (P = 0.063); absolute effect size (95%CI) = 23.3 (−1.6 to 44.3); relative effect size^a (95%CI) = 2.5 (0.9 to 8.1)

Herpertz-Dahlmann et al., (2021)¹⁰

Results of BMI percentile of all participants with AN that received HoT (n analyzed = 22 at admission [T1]; 21 at the start of HoT [T2]; 21 at the end of HoT [T3]; 21 at 1-year follow-up [T4])

• M (SD) at T1 vs T4; M = 3.61(4.36) vs 28.96 (14.98); P < 0.001

• T2 BMI percentile, M (SD) = 17.29 (10.56)

• T3 BMI percentile, M (SD) = 31.19 (10.17)

EBW

Herpertz-Dahlmann et al., (2021)¹⁰

Results of %EBW^c of all participants with AN that received HoT (n analyzed = 22 at admission [T1]; 21 at the start of HoT [T2]; 21 at the end of HoT [T3]; 21 at 1-year follow-up [T4])

• M (SD) at T1 vs T4; M = 77.99 (4.94) vs 92.52 (5.72); P < 0.001

• T2%EBW, M (SD) = 87.68 (4.40)

• T3%EBW, M (SD) = 93.28 (3.76)

Menstruation

Godart et al., (2022)⁹

Between group comparison from all participants with AN from the FT-S with TAU cohort (n = 30) vs TAU (n = 30); 3 years after the end of treatment

• Resumption of menstruation, FT-S with TAU vs TAU; n (%) = 22/30 (73.3) vs 15/30 (50.0)^b; 5.4 (P = 0.020) ; absolute effect size (95%CI) = 30 (4.8 to 50.4); relative effect size^a (95%CI) = 4.2 (1.2 to 10.2)

Herpertz-Dahlmann et al., (2021)¹⁰

Results of menstruation in the last 3 months of all participants with AN that received HoT (n analyzed = 22 at admission [T1]; 21 at the start of HoT [T2]; 21 at the end of HoT [T3]; 21 at 1-year follow-up [T4])

More than three regular cycles

• T1, n (%) = 1 (4.5)

• T2, n (%) = NR

• T3, n (%) = 8 (38.1)

• T4, n (%) = 7 (33.3)

Irregular

• T1, n (%) = 4 (18.2)

• T2, n (%) = NR

• T3, n (%) = 7 (33.3)

• T4, n (%) = 6 (28.6)

Amenorrhea

• T1, n (%) = 17 (77.3)

• T2, n (%) = NR

• T3, n (%) = 6 (28.6)

• T4, n (%) = 4 (19.0)

Oral contraceptive use

• T1, n (%) = 0 (0.0)

• T2, n (%) = NR

• T3, n (%) = 0 (0.0)

• T4, n (%) = 4 (19.0)

Rosling et al., (2016)¹³

Results of menstrual status of all participants with AN that received FB specialized out-patient service (n = 31 at baseline [T1]; 1 at 1-year follow-up [T2])

Pre-menarcheal

• T1, n = 10

• T2, n = 1

Secondary amenorrhea

• T1, n = 19

• T2, n = 0

Contraceptives

• T1, n = 2

• T2, n = 0

MROC/MROAS categories of general outcome based on BMI and menstrual function^d

Godart et al., (2022)⁹

Results of outcome categories^d of all participants with AN from the FT-S with TAU cohort (n analyzed = 30); 3 years after the end of treatment

• Good outcome category, n (%) = 17/30 (56.7)

• Intermediate outcome category, n (%) = 1/30 (3.3)

Results of outcome categories^c of all participants with AN from the TAU cohort (n analyzed = 29), 3 years after the end of treatment

• Good outcome category, n (%) = 7/29 (24.1)

• Intermediate outcome category, n = 2/29

Between group comparison of all participants with AN from the FT-S with TAU cohort (n analyzed = 30) vs TAU (n analyzed = 29); 3 years after the end of treatment

• Good/intermediate outcome category, FT-S with TAU vs TAU; n (%) = 18/30 (60.0) vs 9/29 (31.0); 5.0 (P = 0.026); absolute effect size (95%CI) = 28.9 (3.6 to 49.6); relative effect size^a (95%CI) = 3.8 (1.1 to 9.7)

Rosling et al., (2016)¹³

Results of outcome categories^d of all participants with AN from the FB specialized out-patient service cohort (n = 29); 1 at 1-year follow-up

• Good outcome category, n (%) = 13 (45)

Psychological distress

Godart et al., (2022)⁹

Results from all participants with AN from the FT-S with TAU cohort (n = 30) vs TAU (n = 30); 3 years after the end of treatment

• SCL-90-R/GSI score, FT-S with TAU vs TAU; M (SD) = 0.63 (0.64) vs 0.59 (0.63); −0.3 (df = 55, P = 0.807); absolute effect size (95%CI) = 0.04 (−0.29 to 0.38); relative effect size^a (95%CI) = −0.8 (−1.4 to −0.3)

• SCL-90-R/PST score, FT-S with TAU vs TAU; M (SD) = 29.8 (21.4) vs 29.1 (20.1); −0.1 (df = 55, P = 0.902); absolute effect size (95%CI) = 0.68 (−10.3 to 11.7); relative effect size^a (95%CI) = 0.03 (−0.5 to 0.5)

• SCL-90-R/PSDI score, FT-S with TAU vs TAU; M (SD) = 0.02 (0.01) vs 0.02 (0.01); −1.3 (df = 55, P = 0.362); absolute effect size (95%CI) = 0.001 (−0.002 to 0.005); relative effect size^a (95%CI) = 0.24 (−0.3 to 0.7)

Depression

Herpertz-Dahlmann et al., (2021)¹⁰

Results of BDI-II sum score of all participants with AN that received HoT (n analyzed = 22 at admission [T1]; 21 at the start of HoT [T2]; 21 at the end of HoT [T3]; 21 at 1-year follow-up [T4])

• M (SD) at T1 vs T4; M = 21.50 (11.25) vs 10.29 (9.71); P = 0.003

• T2 score, M (SD) = 14.95 (11.14)

• T4 score, M (SD) = 11.00 (9.70)

Perfectionism

Hurst et al., (2019)¹²

Results of all participants with AN that received FBT with CBT-P (n = 21); at FBT phase one commencement [T1]; at FBT phase two and CBT-P commencement [T2]; after completion of CBT-P [T3]; after FBT with CBT-P completion [T4]

EDI perfectionism score

• M (SD) at T1 vs T2; M = 14.3 (4.9) vs 14.0 (6.2); 0.29 (d = 0.05)

• M (SD) at T1 vs T3; M = 14.3 (4.9) vs 11.0 (6.0); 3.01 (d = 0.60); P < 0.01

• M (SD) at T1 vs T4; M = 14.3 (4.9) vs 10.2 (6.7); 3.02 (d = 0.70); P < 0.01

EDI overcontrol score

• M (SD) at T1 vs T2; M = 29.3 (11.1) vs 28.6 (12.7); 0.31 (d = 0.06)

• M (SD) at T1 vs T3; M = 29.3 (11.1) vs 23.7 (14.5); 2.20 (d = 0.43), P < 0.05

• M (SD) at T1 vs T4; M = 29.3 (11.1) vs 21.0 (16.0); 2.7 (d = 0.60), P < 0.05

CAPS self-oriented perfectionism score

• M (SD) at T1 vs T2; M = 47.9 (8.5) vs 46.3 (9.8); 0.99 (d = 0.17)

• M (SD) at T1 vs T3; M = 47.9 (8.5) vs 43.3 (11.4); 2.61 (d = 0.46); P < 0.05

• M (SD) at T1 vs T4; M = 47.9 (8.5) vs 40.1 (12.0); 3.3 (d = 0.76); P < 0.01

CAPS socially prescribed perfectionism score

• M (SD) at T1 vs T2; M = 28.0 (8.3) vs 29.7 (8.4); −1.06 (d = 0.20)

• M (SD) at T1 vs T3; M = 28.0 (8.3) vs 28.5 (9.5); −0.24 (d = 0.06)

• M (SD) at T1 vs T4; M = 28.0 (8.3) vs 26.0 (10.4); 0.82 (d = 0.21)

School attendance

Rosling et al., (2016)¹³

Results from all participants with AN (n = 31) that received FB specialized out-patient service; at 1-year follow-up

• Back to school on a full-time basis, n (%) = 27 (93%)

Social Adjustment

Godart et al., (2012)¹⁴

Between group comparison of SAS global score of all participants with AN from the FT with TAU cohort (n = 30) vs TAU (n = 30 at baseline [T1], 29 at 8 months of follow-up [T2])

• T1, FT with TAU vs TAU; M (SD) = 2.6 (0.6) vs 2.6 (0.6); −0.11 (P = 0.91)

• T2, FT with TAU vs TAU; M (SD) = 2.0 (0.8) vs 2.0 (0.8); −0.23 (P = 0.82); absolute effect size (95%CI) = 0; relative effect size^a (95%CI) = 0 (−0.29 to 0.29)

Service use

Godart et al., (2022)⁹

Results from all participants with AN from the FT-S with TAU cohort (n = 30) vs TAU (n = 30); 3 years after the end of treatment

• Psychiatric re-hospitalizations, FT-S with TAU vs TAU; n (%) = 13/30 (43.3) vs 18/30 (60)^a; 1.7 (P = 0.196); absolute effect size (95%CI) = 16.7 (8.2 to 38.8); relative effect size^b (95%CI) = 1.05 (0.18 to 1.4)

• Re-hospitalization for AN, FT-S with TAU vs TAU; n (%) = 11/30 (36.7) vs 15/30 (50)^a; 1.1 (P = 0.297); absolute effect size (95%CI) = 13.3 (−11.2 to 35.7); relative effect size^b (95%CI) = 0.6 (0.2 to 1.6)

Rosling et al., (2016)¹³

Results from all participants with AN that received FB specialized out-patient service (n = 29); at first year of treatment (T1); 1-year follow-up (T2)

Treated at EDU in day care some part of the year

• T1, n = 14

• T2, n = 0

Treated at EDU only in out-patient during the year

• T1, n = 15

• T2, n = 1

Coelho et al., (2019)¹¹

Results of all participants with AN or other specified/unspecified eating disorder that received FBT (n analyzed = 62)

• Number of days of FBT, Mdn (IQR) = 207 (21 to 1,556)

• Number of participants that completed FBT, n (%) = 25 (40.3)

• Number of participants that required continued ED treatment, n (%) = 25 (40.3)

• Number of participants that required additional intensive treatment, n (%) = 13 (21)

• Number of participants that required discontinuation of FBT, n (%) = 5 (8.1)

Treatment satisfaction

Herpertz-Dahlmann et al., (2021)¹⁰

Results of ZUF-8 score of all participants with AN that received HoT (n analyzed = 21 at the start of HoT [T2]; 21 at the end of HoT [T3])

• T2 score, M (SD) = 1.77 (0.39)

• T3 score, M (SD) = 1.64 (0.41)

Quality of life

Herpertz-Dahlmann et al., (2021)¹⁰

Results of Kidscreen-27 score of all participants with AN that received HoT (n analyzed = 22 at admission [T1]; 21 at the start of HoT [T2]; 21 at the end of HoT [T3]; 21 at 1-year follow-up [T4])

Physical well-being

• M (SD) at T1 vs T4; M = 30.04 (10.75) vs 47.82 (11.51); P < 0.001

• T2 score, M (SD) = NR

• T3 score, M (SD) = 44.27 (9.32)

Psychological well-being

• M (SD) at T1 vs T4; M = 29.05 (17.46) vs 44.67 (13.76); P = 0.010

• T2 score, M (SD) = NR

• T3 score, M (SD) = 40.16 (12.27)

Parent relations and autonomy

• M (SD) at T1 vs T4; M = 52.34 (7.85) vs 56.56 (8.75); P = 0.023

• T2 score, M (SD) = NR

• T3 score, M (SD) = 53.77 (7.82)

Social support and peers

• M (SD) at T1 vs T4; M = 41.95 (11.86) 51.54 (11.47); P = 0.008

• T2 score, M (SD) = NR

• T3 score, M (SD) = 46.19 (8.22)

School environment

• M (SD) at T1 vs T4; M = 50.14 (9.61) vs 56.48 (10.68); P = 0.078

• T2 score, M (SD) = NR

• T3 score, M (SD) = 54.60 (8.49)

General outcomes on socioeconomic status, food intake, menstrual state, mental state and psychosexual state

Godart et al., (2022)⁹

Results of GOAS Global Score of all participants with AN from the FT-S with TAU cohort (n = 30) vs TAU (n = 30), 3 years after the end of treatment

• FT-S with TAU vs TAU; M (SD) = 8.8 (2.8) vs 8.4 (2.4); 1.14 (P = 0.252); absolute effect size (95%CI) = 0.47 (−0.908 to 1.85); relative effect size^a (95%CI) = 0.177 (−0.33 to 0.689)

Herpertz-Dahlmann et al., (2021)¹⁰

Results of MROAS global score of all participants with AN that received HoT (n analyzed = 22 at admission [T1]; 21 at the start of HoT [T2]; 21 at the end of HoT [T3]; 21 at 1-year follow-up [T4])

• M (SD) at T1 vs T4; MD = 4.28 (1.39) vs 8.72 (1.60); P < 0.001

• T2 score, M (SD) = NR

• T3 score, M (SD) = 7.97 (1.67)