U.S. Military Surveillance of Mental Disorders, 1998–2013

All articles published in the MSMR from January 1998 to December 2013 and listed in the master table of contents (www.afhsc.mil/Msmr/TableOfContents) were identified (N=993) and examined by two independent reviewers (RN and ECR) to identify feature articles for inclusion in the analysis. We categorized feature articles (N=329) by topic according to the category of disorder of the primary health outcome reported and resolved disagreements by consensus. Interrater reliability for the categorization of feature articles was assessed by Cohen’s kappa statistic. The relative proportion of articles by topic of health outcome was compared with the relative proportion of total ambulatory encounters by major category of disorder during successive four-year periods and overall by using two-sided binomial exact tests. [Specific methods for feature article identification and categorization, tabulation of ambulatory encounters, and details of additional statistical methods and power calculations are included in an online supplement to this report.]

Among the 329 articles, we initially agreed on all but 22 (κ=.91). Through consensus, 14 discrepancies were adjudicated in favor of the first reviewer (RN), and eight were adjudicated in favor of the second reviewer (ECR) (p=.28 by two-sided binomial exact test). After consensus, among the 329 feature articles over the 16-year period, 11% were deemed to focus on mental disorders, 4% on neurologic disorders, 29% on musculoskeletal disorders and injuries, and 39% on infectious, respiratory, digestive, or dermatologic disorders (Table 1). During successive four-year periods, the proportion of feature articles focusing on mental disorders rose from 2% in 1998–2001 to 15% in 2009–2013. The proportion of feature articles focusing on neurologic disorders varied between 0% and 8% during the 16-year period. The proportion focusing on musculoskeletal disorders and injuries decreased from a high of 42% in 1998–2001 to 25% in 2009–2013. In contrast, the proportion focusing on infectious, respiratory, digestive, or dermatologic disorders varied between 32% and 54% during the period. Articles focusing on all other categories of disorder (“other,” which are not discussed further) accounted for fewer than one in five articles overall and during each four-year period.

Excluding the “other” category, over the 16-year period overall the proportion of feature articles was found to vary significantly from the proportion of ambulatory encounters in the U.S. military health care system for all categories except for mental disorders. The overall proportions of articles focusing on neurologic disorders and musculoskeletal disorders and injuries were significantly lower than the overall proportion of ambulatory encounters for these categories (p<.001 and p<.01, respectively), and the overall proportion of articles focusing on infectious, respiratory, digestive, or dermatologic disorders was significantly higher (p<.001) than the overall proportion of ambulatory encounters in this category. During successive four-year periods beginning in 2002 and overall, the excess in the proportion of articles compared with ambulatory encounters by category was significantly greater than 5% only for infectious, respiratory, digestive, or dermatologic disorders. [A figure in the online supplement presents these findings.]

This analysis found evidence that over the 16 years from 1998 to 2013, MSMR feature articles focusing on mental disorders were published in numbers proportionate to the overall burden to the U.S. military health care system of mental health morbidity during the period. This proportionality reflects an early recognition by AFHSC leadership and the editors of the MSMR of the growing number of encounters for mental disorders in the U.S. military population associated with the Iraq and Afghanistan wars. In editorial commentary in the MSMR, this trend was first noted in the annual tabulation of ambulatory encounter data for 2006, published in early 2007, which identified and specifically commented on the “sharp” 18% increase in the number of visits for mental disorders over the previous two years—the largest by far across all categories of adverse health outcomes. This and subsequent annual tabulations confirmed a marked increase in the number of ambulatory encounters for mental disorders over successive prior five-year periods: 27% in 2007, 55% in 2008, 68% in 2009, and 120% in 2010 [see figure in the online supplement].

In November 2010, a special issue of the MSMR was published that contained feature articles focused exclusively on mental health. This special issue, and subsequent special issues published in February 2012 and July 2013, contributed 13 of the 25 feature articles (52%) focusing on mental disorders during the final four-year period of our study, contributing significantly to the findings of proportionality reported here.

Long-standing U.S. military policy has emphasized that “[h]ealth surveillance activities shall be prioritized based upon the greatest beneficial impact on commanders’ [force health protection] planning, response, and decision making” (9). Therefore, disorders that are prioritized for routine analysis may be those considered amenable to preventive interventions. This analysis found, perhaps not unexpectedly, that conditions that have historically been the focus of U.S. military preventive medicine efforts, in particular infectious, respiratory, digestive, and dermatologic disorders (10), received disproportionate prioritization for surveillance analysis during this 16-year period. The prevention of such disorders in the U.S. military has historically been amenable to relatively simple interventions, including vaccines and prophylactic medications, and to improved sanitation and hygiene, and such prevention renders additional forces immediately available to commanders, which aligns well with military operational priorities.

Prevention of mental disorders, in contrast, is a concept significantly less familiar to the U.S. military preventive medicine community. Effective preventive interventions for mental disorders may align poorly with the operational priorities of commanders during periods of heightened or prolonged military activity, limiting the translation of results of epidemiological analysis into effective policy. For example, strict adherence to deployment mental health standards, a critical measure to reduce exposure to combat among personnel with strong risk factors for subsequent severe mental health morbidity (11,12), would significantly limit the number available for repeated deployments.

For reasons that are unknown, although mental disorders were not found to be disproportionately represented overall, detailed epidemiological analyses of mental disorders in the U.S. military were regularly published in the MSMR only in recent years after the initial conclusion of operations in Iraq and the surge into Afghanistan and only after a significant burden of preventable mental health morbidity and mortality by suicide had accumulated among those with clearly identifiable risk factors (13). This suggests a need to prioritize surveillance relevant to the prevention of mental disorders and suicide—for example, by making the timely monitoring of adherence to deployment mental health standards (14) a routine component of military health public health surveillance activities.

This study had a number of limitations. We did not have access to the precise ICD-9-CM codes that constituted the annual MSMR tabulation categories, because these have been described only as approximating the global burden of disease (GBD) classifications (15). However, owing to the high degree of overlap between the GBD classifications and top-level ICD-9-CM major diagnostic categories, it is unlikely that any discordance would have resulted in a significant systematic miscategorization of feature articles.

Of greater concern is that the categorization of articles was a fundamentally subjective process. Although we agreed on 307 of 329 (93%) initial categorizations, a relatively large number of articles were discordantly categorized as “infectious, respiratory, digestive, or dermatologic.” Through consensus, seven of these (or nearly a third of the discordantly categorized articles) were recategorized as “other.” However, had consensus recategorization resulted in these articles retaining their original discordant category, our findings that this category was overrepresented would have been further reinforced.

Our results indicate that since 1998 the increased burden to the U.S. military health care system from mental health morbidity has been met with a proportional increase in the number of MSMR feature articles focusing on these disorders. However, the delayed focus on mental disorders shown in this study suggests a need to prioritize the routine surveillance of these conditions to more regularly inform military public mental health policy and to better promote effective and timely population-based preventive interventions.

The authors thank Michelle Carras, Ph.D., Tamar Mendelson, Ph.D., and Lisa Townsend, Ph.D., for helpful comments on the manuscript.