Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Predicting disagreement between physicians and patients on depression response and remission Janet L. Cunningham a , Lisa Wernroth b , Lars von Knorring a , Lars Berglund b and Lisa Ekselius a Demographic, personality, and disease-related factors all contribute when patients disagree with physicians on the severity of subjective symptoms. This study aims to create a model, on the basis of patient factors at treatment initiation, for longitudinal prediction of disagreement on treatment response and remission in depressed patients. Four hundred patients with major depressive disorder were studied during a clinical drug trial. Repeated assessments with the Montgomery–A ˚ sberg Depression Rating Scale (MADRS) and the self-rating version (MADRS-S) were used to indicate response or remission. Factors at baseline and week 2 were tested for inclusion in a model for the prediction of discordance on remission and response between patients and physicians at week 8. The models were then tested, in the same population, at weeks 12, 16, and 24. Model AUCs ranged from 0.71 to 0.74 for week 8. The models that were validated at weeks 12, 16, and 24 indicated stability in the predictive value of the models. The risk for longitudinal disagreement in the evaluation of depression treatment response and remission in clinical practice and drug trials can be predicted using factors at study initiation and at week 2. Int Clin Psychopharmacol 00:000–000  c 2013 Wolters Kluwer Health | Lippincott Williams & Wilkins. International Clinical Psychopharmacology 2013, 00:000–000 Keywords: assessment, clinical trial, major depressive disorder, Montgomery–A ˚ sberg Depression Rating Scale, personality, self and observer rating, subjective symptoms a Department of Neuroscience, Psychiatry Unit and b Uppsala Clinical Research Centre, Uppsala University, Uppsala, Sweden Correspondence to Janet L. Cunningham, MD, PhD, Department of Neuroscience, Uppsala University, Entrance 15, 3rd floor, SE-751 85 Uppsala, Sweden Tel: + 46 18 611 5243; fax: + 46 18 515810; e-mail: Janet.cunningham@neuro.uu.se Janet L. Cunningham and Lisa Wernroth contributed equally to the writing of this article. Received 21 November 2012 Accepted 7 February 2013 Introduction Evaluation of subjective symptoms of depression in patients is a daily challenge for physicians. Many factors potentially influence the communication and consensus between the patient and the physician on the severity of subjective symptoms over time. Evaluation and following such symptoms is, however, essential to determine treatment response and remission. The Montgomery–A ˚ sberg Depression Rating Scale (MADRS) and the self-rating version (MADRS-S) are well-known scales for assessment of depression severity and are designed to follow subjective symptoms identi- fied as selective and sensitive to treatment response (Montgomery and Asberg, 1979; Svanborg and Asberg, 2001). For 9/10 items on MADRS physicians are reliant on information from the patient in evaluating the severity, and these nine items are mirrored in MADRS-S. We recently reported acceptable concordance between pa- tient and physician ratings and general trends in treatment response when using MADRS and MADRS-S during a clinical drug trial (Cunningham et al., 2011). The intraclass coefficient for the total score was the highest, 0.75, at week 8, and patients and physicians agreed in 82% of cases on both remission and response at this time point (Cunningham et al., 2011). The use of self- evaluations saves both time and money. However, an unexplained discordance of 18% between patients and physicians is still high and limits the use of self- evaluations as a primary endpoint in clinical trials or in clinical practice. Previous studies have shown that differences in self and observer raw depression rating scores correlate to personality traits and demographic factors. It has been known for decades that those with neurotic or none- ndogenous depression tend to rate their depression as more severe than those with endogenous depression (Prusoff et al., 1972; Paykel and Prusoff, 1973; Moller and Steinmeyer, 1990). The current depression diagnosis no longer includes this distinction. However, the presence of neurotic personality traits is also related to overestimation of the severity of depression (Enns et al., 2000; Duberstein and Heisel, 2007). Other personality traits such as openness to experience, obsessiveness, low extraversion, and novelty seeking influence the agreement between patient and physician ratings (Paykel and Prusoff, 1973; Enns et al., 2000; Duberstein and Heisel, 2007; Carter et al., 2010). Patients with comorbid personality disorders have higher scores on both MADRS and MADRS-S; however, correlations between the scales are lower (Mattila-Evenden et al., 1996). Low self-esteem, high levels of neuroticism, and dysfunctional attitudes are associated with inflated self-report scores. In a regression Original article 1 0268-1315  c 2013 Wolters Kluwer Health | Lippincott Williams & Wilkins DOI: 10.1097/YIC.0b013e32835ff2a8 untitled - 00004850-900000000-99812.pdf http://graphics.tx.ovid.com.ezproxy.its.uu.se/ovftpdfs/FPDD... 1 av 7 2013-03-19 13:27