Any overlap between orthorexia nervosa and obsessive–compulsive disorder in Lebanese adults? Results of a cross-sectional study and validation of the 12-item and 4-item obsessive–compulsive inventory (OCI-12 and OCI-4)

Abstract Background Orthorexia Nervosa (ON), a compulsive obsession with vigorous eating, has increasingly caught researchers' attention. Although Orthorexia Nervosa has not been labeled an eating disorder, research about ON highlighted a strong link with anorexia nervosa or obsessive–compulsive disorder (OCD). Therefore, this study aimed to (1) validate the Arabic version of the Obsession-Compulsion Inventory (OCI-12 and OCI-4) and (2) check if there is an overlap between ON and OCD among a sample of Lebanese adults. Methods This cross-sectional study involved 487 Lebanese participants between July and August 2021. The Dusseldorf Orthorexia Scale (DOS) was used to assess ON; scores between 25 and 29 indicate probable ON, whereas scores ≥ 30 indicate ON tendencies. A confirmatory factor analysis (CFA) was carried out using SPSS AMOS v.24 on the OCI-12 and OCI-4 scales’ items. The root mean square error of approximation (RMSEA) statistic, the Tucker Lewis Index (TLI) and the comparative fit index (CFI) were used to evaluate the goodness-of-fit of the model. Results The CFA results indicated an excellent fit of the model: the Maximum Likelihood Chi-Square = 147.73 and Degrees of Freedom = 48, which gave a χ2/df = 3.08, TLI = 0.934, CFI = 0.952, and RMSEA = 0.065 [95% CI 0.054–0.078]. The fit indices of the one-factor structure of the OCI-4 were excellent as well: χ2/df = 6.15/2 = 3.08, TLI = 0.95, CFI = 0.98 and RMSEA = 0.065 [95% CI 0.007–0.127]. The Area Under the Curve was 0.600 [95% CI 0.524–0.674]. There was no cutoff value that showed good sensitivity or specificity at the same time. At the DOS cutoff of 25, sensitivity was 19.1%, whereas the specificity was 90.6%. The positive and negative predictive values (PPV and NPV) at this cutoff value were 24.4% and 88.7% respectively. At the DOS cutoff of 30, sensitivity was 8.8%, whereas the specificity was 94.3%. The PPV and NPV at this cutoff value were 10.6% and 92.5% respectively. The results showed that higher total OCD scores (Beta = 0.15) were significantly associated with more ON tendencies. Moreover, higher OCD washing scores (Beta = 0.52), physical activity index (Beta = 0.06), and Body Mass Index (Beta = 0.17) were significantly associated with more ON tendencies. Conclusion The present results suggest that ON, as measured by the DOS, shares more common features with disordered eating and cannot adequately predict the presence of OCD symptoms