Bruno Falissard1, Mickael Lukasiewicz, Emmanuelle Corruble. 1. Faculté de médecine Paris-sud, Hôpital Paul Brousse, AP-HP, Département de santé publique, 12 Av. Paul Vaillant Couturier, 94804 Villejuif, France. falissard_b@wanadoo.fr
Abstract
BACKGROUND AND OBJECTIVE: The MDP75 (Minimal Difference Perceived 75% of the time) is the difference in the global score from a scale or a questionnaire which, 75% of the time, leads to the higher score being spontaneously attributed by patients or clinicians to the most severe situation. This may be considered as a "minimal clinically perceptible difference" because the proportion of "consistent" choices only exceeds the proportion that could be observed "by chance" by 50%. METHODS: The MDP75 can be estimated through a comparison of two groups of items (group A and group B) randomly selected from the instrument, using a sample of subjects. A logistic regression is then used that relates the choice of A or B to the difference: score(A) - score(B). RESULTS: Data were collected on a depression scale, the HDRS. Three clinicians classified 300 pairs of situations. The MDP75 obtained was 2.7, with a 95% confidence interval of [1.9, 4.1]. CONCLUSION: It can be noted that this value is consistent with expert opinion (3 or 4). Nevertheless, in view of the small number of clinicians involved in the data collection phase, further studies are required to confirm this result.
BACKGROUND AND OBJECTIVE: The MDP75 (Minimal Difference Perceived 75% of the time) is the difference in the global score from a scale or a questionnaire which, 75% of the time, leads to the higher score being spontaneously attributed by patients or clinicians to the most severe situation. This may be considered as a "minimal clinically perceptible difference" because the proportion of "consistent" choices only exceeds the proportion that could be observed "by chance" by 50%. METHODS: The MDP75 can be estimated through a comparison of two groups of items (group A and group B) randomly selected from the instrument, using a sample of subjects. A logistic regression is then used that relates the choice of A or B to the difference: score(A) - score(B). RESULTS: Data were collected on a depression scale, the HDRS. Three clinicians classified 300 pairs of situations. The MDP75 obtained was 2.7, with a 95% confidence interval of [1.9, 4.1]. CONCLUSION: It can be noted that this value is consistent with expert opinion (3 or 4). Nevertheless, in view of the small number of clinicians involved in the data collection phase, further studies are required to confirm this result.