The post-deployment screening process

The CAF has adjusted its mental health system over the past 15 years in an effort to minimise the impact of mental health problems on its members.25 Subtle changes were made in the application of medical policy that made it easier for personnel who recover from mental disorders to remain in uniform, reducing a potential career-related barrier to care seeking. The CAF also introduced a resilience and mental health training programme in 2008 with a focus on educating members on mental illness awareness and stigma reduction. This programme was further integrated across the deployment cycle in 2009 and included instruction with an additional emphasis on prevention and psychological resilience. In-depth post-deployment mental health screening was introduced in 2002 and it became fully implemented in the CAF in 2004.

Currently, CAF policy requires all service members who deploy to a special duty area for a period of 60 or more days to complete PDS between 90 and 180 days following deployment return; those deploying to some locations in Europe and the USA are exempted. Additionally, members who are already receiving mental healthcare are to also complete the process but without the time constraint.

The PDS process begins with an orientation briefing of the screening’s purpose, confidentiality and the associated follow-up. The member then completes a questionnaire that contains questions on sociodemographic and military characteristics, deployment experiences and health problems that are assessed using standardised instruments: the SF-36 Health Survey, a general health status measure with an emphasis on functioning26; portions of the Patient Health Questionnaire assessing physical symptoms, depression, suicidality, panic disorder and generalised anxiety disorder27; the Patient Checklist for PTSD—Civilian Version (PCL-C)28; the 10-item Alcohol Use Disorder Identification Test29; a 30 item combat exposure scale30 and the Brief Traumatic Brain Injury Screen.31 These screening instruments were chosen to provide an indication of any concern in nine areas: (1) military-related post-traumatic stress disorder (PTSD) symptoms, (2) civilian-related PTSD symptoms, (3) depressive symptoms, (4) anxiety symptoms, (5) substance use, (6) post-concussive symptoms, (7) other physical health issues, (8) family/marital problems and (9) workplace conflict. The responses from these instruments are scored as indicated by the questionnaire guidelines and concerns are identified using either the recommended cut-offs or ones that were slightly adjusted for the CAF population. The responses and indicated concerns are then reviewed by a mental health professional who subsequently conducts a semi-structured interview. The interviewer then completes a ‘tracking form’ that captures a summary disposition of concerns about potential health problems in the above stated nine areas and an additional ‘other’ to be specified area, along with the interviewer’s follow-up care recommendations. Approximately 6 to 8 weeks after the interview, the appointment register is reviewed to assess whether recommended follow-up care was received. Those without such follow-up are contacted to determine whether further care is still needed.

Study population and sampling

The study cohort consisted of all CAF personnel (n=28 460) who had a deployment within the 01 January 2009 to 31 December 2014 time frame. This time frame was chosen partly based on the availability of electronic health records data to assist with sampling and also to ensure greater than 2 years of post-deployment follow-up among those not releasing from military service beforehand. A stratified random sample of 3004 individuals was identified from this cohort. Seven sampling strata were defined by administrative data that indicated increasing amounts of post-deployment mental health services use (MHSU) in order to guide over-sampling of those who would likely have a mental disorder: six strata categorised any speciality MHSU (0, 1 to 3, ≥4 psychiatrist or psychologist appointments) among those with and those without an indicated PDS-associated appointment and a seventh stratum included those with no identified MHSU in the administrative data. The study was powered to discern a delay to care difference of at least 50 days between screened and non-screened individuals with 85% power when employing a log-rank test. Sample size per stratum was determined using a Neyman optimal allocation approach.32 Medical records for 2997 individuals in the sample were reviewed and seven were inaccessible; 2598 individuals had a total of 2875 deployments that required PDS.

Data collection

Data on deployments were extracted from deployment tasking (extract date: 30 March 2016), deployment-related pay (extract date: 30 March 2016) and human resources (extract date: 01 August 2017) administrative databases. Clinical and other associated data were abstracted from paper and electronic medical records over the period of 06 February 2017 to 01 May 2018. PDS associated data was obtained from the medical records review and this was supplemented with electronic data from the PDS programme (extract date: 01 August 2012). Additional data on sociodemographic and military characteristics were obtained from human resources administrative data.

Outcome measures

The primary outcome assessed was compliant completion of PDS as indicated in the CAF PDS policy. Each deployment that required PDS was considered compliant only if the process was completed within a 90 to 180 days post-deployment time frame. Administrative data provided details on the sampled individuals’ deployments over the period of interest (01 January 2009 to 31 December 2014).

A completed PDS occurred only when both the questionnaire and subsequent interview with a mental health professional were completed and this was based on documentation in either the medical record or electronic PDS programme data. The interview date determined the date of PDS completion. Those who screened early (<90 days), late (>180 days) or in compliance (90 to 180 days), relative to the date of deployment return, were identified. Non-screeners were determined by the absence of documentation to indicate a completed screening (ie, a questionnaire and interview) in both the medical chart and electronic PDS programme data.

Covariates of interest

Previous research has identified a number of covariates to be associated with post-deployment mental health problems or delays to accessing care and these were considered for possibly influencing compliance with the PDS completion requirement.1 33 Arguably, individuals with these characteristics are the ones who would benefit most from PDS and as such, knowing whether these covariates are associated with compliance would be beneficial. First official language is among the sociodemographic covariates being considered; it has been found to be associated with better mental health statistics in some previous work4 and it may be a proxy for a base cluster as these service members tend to be concentrated in a subset of bases. The covariates of interest included:

1. Military and sociodemographic information: age (19 to 24, 25 to 34, 35 to 44 or 45 to 60 years), sex,34 first official language (English or French),4 marital status (married/common law, divorced/separated/widowed or single - never married),35 years of military service (<5, 5 to 9, 10 to 19 or ≥20 years), rank category (junior non-commissioned member (JNCM), senior non-commissioned member (SNCM) or officer),1 component (Regular or Reserve Force),36 service (Army, Air Force or Navy)37 and combat arms military trade/occupation.38

2. Deployment-related information: deployment location (Afghanistan or ‘other’), deployment return date (2009 or ≥2010) and deployment length (≤180 days, >180 days).

3. Mental health-related information: a past mental health problem documented in the medical chart and early MHSU interaction. In those screening early or compliant, an early MHSU interaction was defined as either a mental health diagnostic assessment or a MHSU appointment having been made on or before PDS; however, among those screening late or not at all, this early MHSU interaction would need to have occurred ≤180 days post-deployment (ie, before the upper limit for compliant screening). The creation of an early MHSU interaction covariate was an attempt to identify and control for service members who receive mental healthcare soon after deployment return as these individuals are to also complete PDS but without the 90 to 180 day post-deployment constraint.

All time-varying covariates were assessed relative to deployment return date with the exception of marital status which was assessed on the human resources administrative data extract date. Variable categorisations were based on the data’s distribution and previous work with this population.

Statistical analysis

The data were predominantly analysed using SAS for Windows, V.9.3 (SAS Institute Inc, North Carolina); marginal effect estimates were obtained using Stata for Windows, release 14 (StataCorp LP, College Station, Texas). We applied the sample design weights to determine descriptive and regression statistics and Taylor Series Linearisation39 was used to generate the associated SE estimates and 95% CIs. The primary unit of analysis was deployment. Since some individuals (ie, 266 of 2598 sampled individuals) had multiple deployments, individuals were handled as cluster variables in the design specification to allow SEs to be corrected for the repeated measure.40

List-wise deletion was used for missing values and this excluded less than 1% of the deployments; the deployment location covariate had a single missing value and all other covariates in the analyses were complete.

Descriptive statistics were generated and Wald X² tests assessed associations between each variable and PDS compliance. While compliance with the 90 to 180 day post-deployment completion time was the main outcome, late completers (>180 days), early completers (<90 days) and non-completers were examined descriptively. Kaplan-Meier curves were generated to assess the time-to-PDS completion overall and by covariates that were determined to be significant in regression modelling; zero time was deployment return date, event date was PDS completion date and non-completers were censored at the start date of a subsequent deployment, their military release date or chart review date, whichever was earlier. Follow-up time for censored individuals ranged from a 72 day minimum to a 3128 day maximum (mean: 1640 days; median: 1634 days; IQR: 2214–1052=1162 days). Follow-up time for PDS completers ranged from 1 day to a maximum of 2345 days (mean: 167 days; median: 142 days; IQR: 114–183=69 days).

Logistic regression modelling investigated the association of covariates with PDS compliance, completion 90 to 180 days post-deployment. A conservative variable reduction strategy was used; variables were sequentially dropped if their covariate-adjusted Wald X² p value was ≥0.25, starting with the variable having the largest p value.

We used a marginal standardisation approach to estimate the proportion of PDS-requiring deployments that, based on the final logistic regression model, would be expected to be compliant with screening had they had the characteristic of interest; this approach operates by statistically forcing the total population to have the characteristic of interest while other covariates retain their observed value.32 Expected marginal PDS compliance prevalence differences (MPD) were computed and a z-test assessed the MPD statistical significance with standard errors computed using the delta method.32

Patient and public involvement

CAF service members, patients and/or the public were not involved in developing the research question, the study design or in the conduct of the study. The findings from this study and the larger research project will be shared with CAF service members and other interested stakeholders through targeted conference venues, CAF community newsletters or communiques and other venues.

Study population characteristics

Table 1 summarises the demographic and military characteristics among unique individuals with deployments that required PDS during the study period (ie, 2009 to 2014). Individuals were predominantly English speaking, married, male, Regular Force members, in the JNCM rank category and Army service. On their first deployment return during the study period, a majority were less than 35 years of age (mean: 34.1 years, 95% CI 33.6 to 34.6) and just over half had been in military service for less than 10 years. Additionally, during the period of study 89.1% of individuals had only one deployment and 83.5% had an Afghanistan deployment.

Post-deployment screening completion and compliance

Figure 1 characterises the cumulative proportion of deployments with a PDS requirement that had one completed as post-deployment time increases (ie, Kaplan-Meier estimates); while the rate of completions tended to decrease after 180 days, the completions did not plateau until 390 to 450 days post-deployment. An estimated 67.3% (95% CI 65.0 to 69.6) of deployments that required a PDS had one completed; 43.3% (95% CI 40.6 to 46.0) were completed within the 90 to 180 day post-deployment period (policy compliant), 6.5% (95% CI 5.0 to 7.9) were completed early (ie <90 days post-deployment) and 17.6% (95% CI 15.4 to 19.7) were completed late (ie, >180 days post-deployment). Moreover, contrary to expectations, 16.8% of deployments that did not require PDS had one completed; these deployments and associated PDS were not assessed further.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Cumulative proportion of deployments occurring over the 2009 to 2014 period that required post-deployment screening and had one completed as post-deployment time increased. The compliant completion period was 90 to 180 days post-deployment. PDS, post-deployment screening.

Service members who receive mental healthcare soon after deployment return are to also complete PDS but without the 90 to 180 day post-deployment constraint. We identified individuals who had interactions with mental health services post-deployment but prior to screening or, for non-screeners and late completers, prior to the upper time cut-off for screening compliance (ie, 180 days post-deployment). In many instances, such individuals are compliant screeners as they do not have to screen within a certain time period if they were receiving care before this 180 day post-deployment period. Overall, 3.3% (95% CI 2.6 to 3.9) of deployments were from individuals with an early MHSU interaction. Breaking this down by PDS completion status, 2.6% of compliant completers, 0.6% of early completers, 4.0% of late completers and 4.2% of non-completers had an early MHSU interaction. If all non-completers with this early interaction ultimately complete a PDS, completed screenings would increase from 67.3% to 68.7% (95% CI 66.4 to 71.0). Similarly, if all those who were identified as non-compliant completers but had this early MHSU interaction were designated compliant screeners, the compliant screening estimate would increase from 43.3% to 45.4% (95%CI 42.7 to 48.1).

We further looked at MHSU interactions among both late completers and non-completers to identify the fraction that may have benefitted from earlier screening and 32.9% (95% CI 28.0 to 37.9) of late completers had a MHSU interaction; this was comprised of 28.9% (95% CI 24.0 to 33.8) with an interaction that was more than 180 days post-deployment and 4.0% that were earlier. In comparison, 20.5% (95% CI 17.7 to 23.2) of non-completers had a MHSU interaction; this was comprised of 16.2% (95% CI 13.7 to 18.8) with an interaction that was more than 180 days post-deployment and 4.2% that were earlier. Thus, 28.9% of late completers and 16.2% of non-completers might have had a shortened delay to care had they been compliant screeners.

The sociodemographic and military characteristics of individuals with deployments that required PDS are provided in table 2 by PDS completion status. The three non-compliant completion groups (ie, not completed, early completion and late completion) differed on most covariates. Notably, those not completing PDS were more likely to be older, in higher ranks, in non-combat arms occupations, Navy (less from the Army), have a non-Afghanistan deployment location and have shorter deployments (ie, ≤180 days) relative to both early and late completers. Other univariate differences were such that non-completers were more likely to have more years of military service, a more recent deployment return date (ie, ≥2010) and to be Regular Force members relative to late completers. Additionally, relative to early completers, non-completers were more likely to have an English first official language, a less recent deployment return date and an early MHSU interaction.

Additionally, Wald X² tests comparing covariates among the early, late and compliant completion groups indicated that they were largely similar. However, relative to the other two groups early completers were more likely to have a first official language that was French and a more recent deployment return date (ie, ≥2010), but less likely to have an early MHSU interaction. Late completers were more likely to have shorter deployments (ie, ≤180 days).

Wald X² tests comparing the compliant and non-compliant (overall) completion groups indicated that the non-compliant group was more likely to be older, in higher ranks, have more years in military service, Navy (less from the Army), in non-combat arms occupations, have a non-Afghanistan deployment location and have shorter deployments relative to the compliant group (table 2).

Logistic regression results

Table 3 presents the logistic regression analysis for PDS completion compliance. Five covariates were excluded in the final variable-reduced model: years of military service, deployment return date, service, age category and marital status. We observed that the odds of PDS completion compliance was significantly associated with a number of covariates. Compliance was higher for the JNCM rank category relative to officers (adjusted OR (AOR): 1.59; 95% CI 1.17 to 2.16), combat arms occupations (AOR: 1.44; 95% CI 1.11 to 1.86) and Afghanistan deployments (AOR: 2.37; 95% CI 1.62 to 3.47); it was lower for shorter deployments (ie, ≤180 days relative to >180 days) (AOR: 0.65; 95% CI 0.48 to 0.86) and among personnel with a past mental health problem (AOR: 0.65; 95% CI 0.42 to 1.00). Completion compliance was lower with marginal significance (0.05<p≤0.10) among those whose first official language was French (AOR: 0.79; 95% CI 0.61 to 1.03) and those with an early MHSU interaction (AOR: 0.63; 95% CI 0.39 to 1.05). Completion compliance differences were not significant for the sex and component covariates, or for SNCM ranks relative to officers.

Marginal PDS compliance prevalence difference estimates are presented in table 3. PDS completion compliance was 10.6% higher for JNCM ranks relative to officers, 8.4% higher for the combat arms occupation, 19.2% higher for Afghanistan deployments, 10.1% higher for longer deployments (>180 days) and 9.7% higher in those without a past mental health problem.

Figure 2 characterises the cumulative proportion of deployments with a PDS requirement that had one completed as post-deployment time increases (ie, Kaplan-Meier estimates) for statistically significant covariates identified by the logistic regression analysis. There are three time intervals of interest. The rate of early PDS completion appears to be higher for Afghanistan deployments and deployments greater than 180 days, relative to their counterpart covariate values. Differences for the compliant PDS completion period are reflective of what was noted in the logistic regression analysis. However, for Afghanistan deployment locations, combat arms occupations and non-commissioned member (NCM) ranks (both JNCM and SNCM) the rate of late PDS completions appears to continue at an elevated level relative to their counterpart covariate value. In contrast, the late completion rate appears to increase comparably for both deployment length covariate values (ie, ≤180 days, >180 days) before the cumulative proportion plateaus. The rate of late PDS completions for those with a past mental health problem appears to be more variable; initially elevated then comparable or lower relative to those without a past mental health problem. Moreover, it was observed that among individuals with each of these more-likely-to-screen characteristics, the cumulative proportion with a completed PDS approached 80% or less as time passed.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Cumulative proportion of deployments occurring over the 2009 to 2014 period that required PDS and had one completed as post-deployment time increased by specific covariates. All covariates had a statistically significant association with compliant PDS completion (ie, 90 to 180 days post-deployment) as determined by multiple logistic regression. PDS, post-deployment screening.

Key findings

The objective of the current study was to estimate PDS completion compliance, as measured by completion of screening 90 to 180 days post-deployment for deployments that required it, and to examine factors associated with this compliance. We found that 67.3% of deployments that required PDS had one completed and 43.3% were completed within the compliant post-deployment period. PDS compliance was higher with a lower rank category (10.6% for JNCM relative to officers), combat arms occupations (8.4%), Afghanistan deployments (19.2%), longer deployments (10.1%) and among those without a past mental health problem (9.7%). Additionally, screening compliance was higher with marginal significance among those whose first official language was English and those without an early MHSU interaction.

The rate of late PDS completions was observed to continue at an elevated level as time passed beyond the upper time limit for compliance (ie, >180 days post-deployment) among individuals with an Afghanistan deployment, a combat arms occupation and NCM ranks (both JNCM and SNCM) relative to their counterpart covariate values. This may reflect a delayed recognition of need for care, with PDS used as a means to achieve this, or it may simply be reflective of a late attempt to be compliant with PDS. However, we also observed that 28.9% of late completers and, additionally, 16.2% of non-completers had a MHSU interaction more than 180 days post-deployment. This suggests a delay to care in a sizeable fraction of late and non-screeners that may have possibly been averted had they screened sooner, especially if the MHSU was related to a prior deployment.

Comparison with other research

We were unable to locate much research on compliance with PDS in military personnel from other countries. In the US military, PDS consists of a post-deployment health assessment shortly (ie, approximately 30 days) after a deployment ends and a second health status reassessment 90 to 180 days post-deployment; these assessments use standardised screening questionnaires and a meeting with a healthcare provider.24 Some research has estimated compliance with the initial post-deployment health assessment screening to be 76.8% and this reducing to 33.5% for the reassessment screening, estimated among 289 355 US service members returning from deployments (ie, Operation Iraqi Freedom) ending in 2005.24 The researchers suggested that service members who were female, in the Army or Air Force, Reserve component and in non-combat occupations were more likely to complete a screening; however, this interpretation was based on univariate summary statistics without formal statistical testing. Our findings were a little different; 67.3% of required screenings were completed, and we did not observe statistically significant differences in compliant completions (ie, 90 to 180 days post-deployment) for sex, service or component covariates and we observed a reverse finding for combat arms occupations as they were 8.5% more compliant.

Limitations

The primary limitation of our study relates to it being retrospective. While we were able to identify records and dates for deployment information, PDS completion and other associated details, we were reliant on the information that was available. We were unable to identify specific individual reasons for non-compliance, particularly for those who didn’t complete a required PDS. Some individuals would have completed their PDS early or late, or not yet, because they already had interactions with mental health services and a completed screening among such individuals is an administrative requirement without a time constraint; we estimate that total completions and compliant completions would have only increased by approximately 1.4% and 2.1%, respectively, for this reason. Additionally, we determined PDS completion date based on the date of the summary ‘tracking form’, which is completed following the interview with a mental health professional at the end of screening. An unknown number of individuals may have started the screening process prior to the required 180 day post-deployment period but finished it after this time point, or started without continuing on to complete the interview; these individuals would have been classified as non-compliant late completers, or non-completers in those without an interview. Moreover, while screening in the CAF requires that personnel returning from requisite deployments complete PDS 90 to 180 days after deployment return, this study and assessment was not designed to determine whether this screening window is optimal with respect to desired outcomes for the CAF setting.

Interpretation

In the CAF, the responsibility for mental health is shared among a member’s command, clinicians and the member, with each having different but overlapping responsibilities that vary across the deployment cycle. Nevertheless, compliance with PDS is largely the responsibility of the member’s current commanding officer. We found that the compliance with the PDS requirement, as measured by completion timing, was less than ideal for deployments that required individuals to undergo screening. Both the fraction of individuals who completed a screening (67%) and those who completed it within the compliance time frame (43%) were lower than expected and this suggests that there is a need for improvement. The CAF does not currently have a formal process to report PDS completion statistics for the review of commanding officers. Our findings suggest that the implementation of such feedback mechanisms would be beneficial, especially if the lag from deployment return to compliance reporting was optimised. This would assist in clarifying where the barriers to compliance may occur (ie, at the service member or commanding officer level) and it would allow for adjustments that could potentially enhance compliance.

In order to better understand the influences on PDS completion compliance, we investigated a number of factors that had the potential to influence screening compliance and this indicated that selection processes may be occurring with respect to the characteristics of individuals who were more likely to screen. A number of factors that have previously been identified as having some association with developing post-deployment mental health problems, and are commonly known among military personnel, were observed to also have a strong influence on compliance with screening. Specifically, deployments to Afghanistan, longer deployments, lower ranks and combat occupations have each been found to be associated with a higher rate of post-deployment mental health problems1 2 41 42 and these characteristics were associated with higher compliance. However, even among individuals with these characteristics PDS completion only approaches 80% when the compliance time constraint is ignored, and this would be closer to approximately 50% when imposing the compliance time constraint. Therefore, while those who are more likely to have post-deployment mental health problems are more likely to undergo PDS, moderately large fractions with these characteristics do not. Additionally, the lower completion compliance that was observed among those with a past mental health problem and officers suggests that some service members may be hesitant to screen, perhaps because of some perceived barrier.43 44 Our findings would suggest that there is a need to reinforce the guidelines for PDS among both commanding officers and service members, perhaps as part of their mental health training, in order to optimise compliance with screening completion. This would help to ensure that all members receive the same direction post-deployment, possibly reducing some individuals’ hesitancy or avoidance. It may also help to reduce the number of individuals who screen too early or too late and possibly, erroneously, perceive their screening to be compliant.

In defining compliance with completion of PDS, we specified that non-compliance can occur when a screening is completed outside of the 90 to 180 day post-deployment period as well as when not completed at all. The rationale for a compliant screening period relates to implied disadvantages to either screening too early (eg, returnee’s reduced likelihood of reporting symptoms immediately post-deployment)20 22 24 or screening too late, which undermines the goal of early problem detection. We observed some indication that a fraction of late screeners were seeking mental health services after the 180 day post-deployment period, possibly supporting the notion of a delay-to-care disadvantage to screening late. However, a more detailed assessment of the timing of PDS completion and its influence on outcomes such as delay to care and both screening sensitivity and specificity are needed to further delineate the optimal period of completion. Additionally, while PDS is designed to identify potential mental health problems that may require a follow-up mental health assessment and care, further work is also needed to assess what happens after this. For example, it would be beneficial to determine whether certain screening-associated or individual characteristics influence outcomes such as a follow-up recommendation, compliance with a follow-up recommendation and ultimately, MHSU and a mental disorder diagnosis.