The Study Of Major Depressive Disorder Clinical Sreening Scales Among General Hospital Inpatients

Background and Objective:Major depressive disorder (MDD) has become one of the common mental disorders in general hospital inpatients, which is a leading cause of disability and associated with lowering quality of health and increasing morbidity and mortality. Both of PHQ-2and PHQ-9are widely used for indentifying, diagnosing MDD and guiding MDD treatment in general hospital in many countries. Domestic scholars have evaluated validities of PHQ-9and PHQ-2for screening MDD in outpatients and community population, and the results showed that they are simple and easy to use and suitable for Chinese culture. HADS is one of the depression screening instruments and widely used in general hospitals in China, but it is lack of the ability of diagnosing and guiding treatment in MDD. However, few studies about PHQ-9or PHQ-2for MDD screening have been done in general hospital inpatients, and no studies have been done to compare differences among depression screening scales in general hospitals inpatients. In this study we evalutated the reliability and validity of PHQ-9, PHQ-2, HADS-D to make MDD diagnosis in general hospital inpatients, and compare accuracy of recognizing MDD between HADS-D, PHQ-2and PHQ-9each other.Materials and Methods:One thousand and two hundred inpatients18years or older with medical conditions were recuited from three tertiary hospitals and two secondary hospitals in Tianjin, China. All the subjects were evaluated with the Chinese version of Mini-International Neuropsychiatric Interview(MINI), which was considered as the gold standard of MDD diagnosis in this study. All patients completed PHQ-2, PHQ-9and HADS-D survey. Total69subjects were diagnosed as MDD by MINI based on the Diagnostic and Statistical Manual of Mental Disorders(DSM-IV). SPSS17.0statistical software packages was used to describe general informations, and to analyze the reliability and validity of PHQ-9, PHQ-2, and HADS-D. The index of reliability was the Crobah alpha coefficient. The major indices of validity was the receiver operating characteristic curve (ROC), the area under ROC (AUC) was the best diagnostic evaluation index, which represented the diagnostic accuracy of MDD. Based on ROC analysis, sensitivity(Se), specificity(Spe), positive predictive value(PPV), negative predictive value(NPV), positive likelihood ratio(LR+), negative likelihood ratio(LR-) was calculated at different cut-off scores. The MedCalc-11.4.0version was used to compare the differences of AUC between PHQ-2, PHQ-9and HADS-D for diagnosing MDD. a=0.05(bilateral level) was selected.Results:(1) PHQ-9Crobah alpha coefficient is0.783(after correction of0.809) with item-total correlation of0.428-0.609, PHQ-2Crobah alpha coefficient is0.719(after correction of0.720) with item-total correlation of0.562, HADS-D Crobah alpha coefficient is0.837(after correction of0.847) with item-total correlation of0.428-0.609.(2) MINI as the gold standard of MDD diagnosis, the AUC of PHQ-9and PHQ-2was0.952(95%CI0.922-0.981) and0.899(95%CI0.856-0.942) respectively. The optimal cut-off score of PHQ-9for screening MDD wa≥7, which indicated a sensitivity of0.92and a specificity of0.90with positive likelihood ratio of9.20. The optimal cut-off score of PHQ-2for screening MDD was≥2, which indicated a sensitivity of0.85and a specificity of0.86, with positive likelihood ratio of6.07. When using the PHQ-9office diagnostic code algorithms to make MDD diagnosis based on the DSM-IV MDD diagnosis standard, it had a sensitivity of0.38and a specificity of0.99with positive likelihood ratio of38.(3) There were no statistical differences in AUC between PHQ-2and HADS-D(z=0.892,P=0.372),and between PHQ-9and HADS-D(z=1.899,P=0.058), but it was an exception between PHQ-9and PHQ-2(z=2.929,P=0.003).(4) There was a dose titration relationship between the severity of MDD and function damage indices of inpatients (disability days during the past six months F=16.5, P≤0.001; health self-rating at present F=29.1, P<0.001; health self-rating at the worst F=35.4,p<0.001).Conclusion:(1) As a screening tool for MDD, PHQ-2and PHQ-9have good characteristics of psychological measurement, the optimal cut-off scores were≥2and≥7respectively;(2) As a clinical diagnostic tool for MDD, the specificity and the positive likelihood ratio of the PHQ-9office diagnostic code algorithms were very high respectively, but the sensitivity was problematic. Therefore, it can not be used for clinical diagnosis.(3) PHQ-2, PHQ-9and HADS-D are the best screening tool for patients with major depressive disorder in general hospital, but PHQ-9may be more favored.(4) The total score of PHQ-9was the effective tool to assess depression severity.