Analysis of the discriminant ability of shorter versions of the French ADAM questionnaire

Abstract

Objective: To investigate whether shorter versions of the ADAM test, a screening questionnaire for andropause, provide better diagnostic value than the original tool.

Methods: Five thousand and twenty-eight volunteer men aged 50-70 years attending a screening campaign for andropause, provided a fasting blood sample and completed the French ADAM test. Logistic regression analysis identified items that best predict andropause defined as serum free testosterone level below 70 ng/l. ROC curves assessed the diagnostic value of modified versions of the ADAM test, obtained by elimination of the less relevant predictors of andropause.

Results: Only four items of the ADAM questionnaire may account for the diagnosis of andropause. These items concerned loss of height, decrease in libido and in enjoyment of life and deterioration in work performance. Item 9 was borderline significant. The area under the ROC curve for the short versions varied slightly from 0.555 to 0.560. As expected, model 6 has a greater specificity (56.02%) than the original tool while the efficiency increased slightly (54.85%).

Conclusion: The modified versions of the ADAM test do not provide better diagnostic value than the original tool.