The results indicated that 347 (79.2 %) subjects (out of 438) used sports supplements, among whom 69 (15.8 %) used protein powders, 96 (21.9%) used vitamins, 97 (22.1) used creatine, 64 (14.6 %) used amino acids, 21 (4.8%) used other supplements, and 91 (20.8%) used no supplements. Furthermore, 293 (66.9%) people used no drugs. Of the 145 (33.15) people who used drugs, 71 (16.2%) used steroid products, 55 (12.6%) used peptide hormones and growth factors, and 19 (4.3%) used other drugs.
Table 1 shows that there are statistically significant differences between supplement use and non-use in bodybuilders in those with muscle dysmorphia (t (436) = 13.48, P < .001) and social physique anxiety (t (358) = 3.69, P < .001) scores. Furthermore, the difference between drug use and non-use in those with muscle dysmorphia (t (396) = 8.05, P < .001) and social physique anxiety (t (181.8) = 8.12, P < .001) was statistically significant (
Table 1).
To examine the research hypothesis, a logistic regression was performed to ascertain the effects of muscle dysmorphia and social physique anxiety on the participants’ likely use of supplements. When both MDD and SPA enter The logistic regression model was statistically significant, χ2(1) = 150.07, P < 0.0001. The overall model explained 45.0% (Nagelkerke R2) of the variance in supplement use and correctly classified 88.4% of cases. As well as, result show that model was statistically significant, when MDD as predictor χ2(1) = 146.09, P < 0.0001. The model explained 44.0% (Nagelkerke R2) of the variance in supplement use and correctly classified 89.5% of cases.
Moreover, the overall logistic regression model was statistically significant for the effect of muscle dysmorphia and social physique anxiety on the probability of drug use in bodybuilders, χ
2(1) = 130.72, P < 0.001. The model explained 35.0% (Nagelkerke R
2) of the variance in drug use. Also, prediction model whit SPA was statistically significant χ
2 (1) = 80.37, P < 0.0001, and 23.0% (Nagelkerke R
2) of the variance in drugs was explain. In addition, SPA alone able to correctly classified 82.0% of cases that use drugs. Additionally, result show that regression model was statistically significant, when MDD as predictor χ
2(1) = 48.20, P < .0001. MDD explained 14.0% (Nagelkerke R
2) of the variance in drugs use and correctly classified 82.0% of cases (
Tables 2 and
3).