Academic journal article European Journal of Psychotraumatology

Assessment of Depression in Veterans across Missions: A Validity Study Using Rasch Measurement Models

Academic journal article European Journal of Psychotraumatology

Assessment of Depression in Veterans across Missions: A Validity Study Using Rasch Measurement Models

Article excerpt


Substantial proportions of soldiers returning from deployment in war zones go on to develop mental health problems (2-17% across studies; Richardson, Frueh, & Acierno, 2010). Posttraumatic stress disorder (PTSD) is one of the most investigated mental health consequences of military deployment. However, approximately 40% of those evaluated after deployment are not diagnosed with PTSD but with other psychiatric disorders (Packnett, Gubata, Cowan, & Niebuhr, 2012; Piccirillo, Packnett, Boivin, & Cowan, 2015), most frequently depression or symptoms thereof (Stander, Thomsen, & HighfillMcRoy, 2014).

While depression after deployment often occurs comorbidly with PTSD (Stander et al., 2014; Van Hooff et al., 2014), it is also seen with no concurrent PTSD (O'Donnell, Creamer, & Pattison, 2004). Research results suggest post-deployment prevalence of depression with or without comorbid PTSD ranging from 2-15% (Hoge et al., 2004; Reijnen, Rademaker, Vermetten, & Geuze, 2015). This variance in prevalence across studies is hypothesized to partly reflect differences in sampling, assessment tools, time of assessment, attrition rates or other methodological issues such as unrecognized noninvariance of measurement (Ramchand et al., 2010). However, depression prevalence differences may also reflect an actual difference due to different combat exposure or specific mission characteristics. Disentangling methodological from actual differences is crucial in order to understand the toll on mental health that different wars and different missions cause among soldiers. However, studies that systematically compare rates of mental health problems across multiple cohorts using the same methodology while ensuring measurement invariance across these cohorts are rare (Richardson et al., 2010), leaving us with limited knowledge on how different deployments affect soldiers differently.

Since 1992, approximately 31,000 soldiers have been deployed to international missions with the Danish Military (Statistics from the Danish Veteran Centre's deployment database). From 1998 onwards, the Psychological Reactions following deployment to International Missions (PRIM) questionnaire on post-deployment mental health has been distributed to all deployed soldiers 7-8 months after home-coming (Andersen, 1998). This large amount of systematically collected data provides a unique opportunity to compare the psychological toll of deployment across different cohorts, wars and missions with different levels and types of combat exposure.

Included in the PRIM is a subscale consisting of 10 items aimed at measuring symptoms of depression (PRIM-Depression, see Table 1). The scale was constructed by Danish military psychologists based on other depression questionnaires and the literature on depression (Andersen, 1998). However, no formal validation of the scale has been conducted and, hence, the reliability and validity of the scale can be questioned. Valid assessment tools for deployment- and psychotrauma-related outcomes are crucial to obtain valid results (Frewen, Dean, & Lanius, 2012; Olff, 2015).

Validation of psychological measurement scales and instruments have historically been, and are still, mainly conducted within the classical test theory (CTT) framework. However, in recent decades, there has been a growing number of studies employing modern test theory, particularly confirmatory factor analysis (CFA) focusing on the dimensionality issue of such instruments and, to a lesser degree, item-response models (IRT) focusing on the issue of item fit (Bentley, Gallagher, Carl, & Barlow, 2014; Tsai et al., 2015). IRT provides a solid foundation for assessing the accuracy and invariance of a scale in measuring the trait it intends to measure (Van der Linden & Hambleton, 2013). A special instantiation of IRT is the family of Rasch models (RM; Fischer & Molenaar, 2012), with the RM for dichotomous items, also known as the one-parameter or 1PL model, being the simplest (Rasch, 1960). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.