Systematic review of factor validity of psychiatric scales in longitudinal studies

Validity shows the degree of concurrence between the results received by an actual measuring and that of what an instrument is supposed to measure. There are three main types of validity: content validity, criterion-related validity and construct validity. Factor validity is a special approach to constructive validity, tested by statistical analysis called factor analysis. Hamilton rating scale for depression and Montgomery-Asberg depression scale are the most widely used psychiatric instruments. The aim was to carry out a systematic review of the literature on factor structure of psychiatric scales reported in different times during longitudinal studies. The units of analysis were published papers obtained by searching the two bibliographic databases: MEDLINE and PsycINFO. Factor validity of the HAMD scale was reported in 6 (0.09%) of 6590 studies which used this scale, whereas the factor validity of the MADRS scale was reported in 4 (0.2%) of 2051 studies which used the MADRS scale. The difference between the first and the last measuring in relation to values was not statistically significant: median of the number of extracted factors of the HAMD scale (p=0.371), median of total explained factor variance of the HAMD scale (p=0.250), median of variance explained by the first factor of the HAMD scale (p=0.125). Factor validity of the MADRS scale in repeated measuring also did not have statistically significant difference for the following values: median of the number of extracted factors of the MADRS scale (p=0.174), median of variance explained by the first factor of the MADRS scale (p=0.125). Coefficients of concurrent validity of the HAMD i MADRS depression scales show the trend of increase in longitudinal studies and their values are for about a third higher at the end of studies than in their beginning. Low frequency of reporting the data about reliability and validity of applied rating scales is the main problem in using the meta-analytical methods effectively to study changes in adequacy of measures in longitudinal studies.