Table 1 Summary of the populations, design characteristics and main results of the selected studies
From: Systematic review of prospective studies assessing risk factors to predict anorexia nervosa onset
Authors | Year of pub | Country | Cohort size | Initial age | Follow-up period | Measuring risks factors tolls | ED Diagnoses tools | Results | Attrition bias | ||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Statistical method | Retained risk factors | Association with AN onset | Cut-off calculation | ||||||||||
Stice et al. [45] | 2017 | USA | 257 | 12–16 years | Follow up at 2 and 8 years | Ideal-Body Stereotype Scale, Thinness expectancies scale, 5-item scale about denial of the cost of pursuing thin ideals, Weight Control Scale, Dutch Restrained Eating Scale, Negative Affect Scale-Revised, Studies-Depression Scale, Beck Depression inventory, Functional impairment measured with 17 items from the Social Adjustment Scale-Self Report for Youth, Frequency of visits to mental health care institutions, BMI | EDDI, DSM-5 criteria for eating disorders | Univariate, multivariate Cox analysis |  | Subthreshold/Threshold AN | No | Not taken into account | |
HR | p value | ||||||||||||
Functional impairment Body Mass Index | 1.71 0.05 | .0047  < .0001 | |||||||||||
| Â | Threshold AN | ||||||||||||
HR | p-value | ||||||||||||
Negative affect Functional impairment Body Mass Index | 2.06 2.17 0.16 | .0144 .0292 .0201 | |||||||||||
Stice and Desjardins [46] | 2018 | USA | 1272 | Mean age 18 years | 3 years | 8-item Ideal-Body Stereotype Scale–Revised, 9-item Thinness Expectancy scale, 5 items scale about denial of cost of pursuing thin ideal, 9 items from the Satisfaction and Dissatisfaction with Body Parts Scale, 5-item Weight Control scale, Dutch Restrained Eating Scale, Negative Affect Scale Revised, Center for Epidemiologic Studies-Depression Scale, Beck Depression Inventory, 17 items from the Social Adjustment Scale-Self Report for Youth, Frequency of visits to mental health care, BMI | EDDI, DSM-5 criteria for eating disorders | Multivariate analysis, Classification tree | Low BMI | Participants in top 83% of BMI: Incidence = 0.3% | Part. in bottom 17% BMI: Incidence = 11.1% | No | Not taken into account |
Interaction of low BMI/high body dissatisfaction | Part. in bottom 17% BMI and bottom 89% of body dissatisfaction: Incidence = 9.0% | Part. in bottom 17% BMI and top 11% of body dissatisfaction: Incidence = 71.4% | |||||||||||
Overall classification rate = 0.982 Null tree classification rate = 0.979 F-index = 0.303 G-mean = 0.438 | |||||||||||||
Tyrka et al [50] | 2002 | USA | 1272 | Mean age 18Â years | 3Â years | SCID, EAT-26, SIQYA, Family Environment Scale | EDI, EAT 26, Structured Clinical Interview for DSM-III-R | Multivariate analysis, Logistic regression | Perfectionism BMI | T2 | T3 | No | Not taken into account |
Unstandardized B Coeff. = .746; -.319 p-value ≤ .05 | Unstandardized B Coeff = 1.314; -.817 p-value =  ≤ .05 | ||||||||||||