Law enforcement officers are routinely exposed to hazardous, disturbing events that can impose severe stress and long-term psychological trauma. Upward of 42% of public safety personnel (PSP) report one or more mental health symptoms [1]. Accumulated stress and posttraumatic stress injuries (PTSIs) result in chronic physical and mental health disorders, including anxiety, depression, substance abuse, and cardiovascular disease [2,3]. PTSI is also related to reduced occupational performance, absenteeism, and risky behavior, with implications for both police and public safety [4,5]. Recent empirical research and government reports have highlighted a mental health and suicide crisis among various PSP sectors in Canada [6-8]. The accumulated evidence forms an urgent call for evidence-based programs that build resilience and wellness capacity in order to prevent PTSI symptoms before they manifest as severe, chronic, diagnosable disorders in PSP.

Existing PTSI prevention and resilience promotion interventions tailored to PSP not only show limited effectiveness but also do not consider sex and gender as central determinants of health and variables of importance [5,9,10]. Further, most traditional interventions are focused on addressing cognitive, emotional, and behavioral components without addressing the underlying neurophysiological mechanisms that erode resilience [2,11-14]. Typically, when a person encounters stress (either real or imagined), they will experience characteristic changes in the autonomic nervous system (ANS), including increased sympathetic nervous system (SNS) activation that elevates heart rate (HR) and respiration and a withdrawal of parasympathetic nervous system (PNS) activation. When stress has abated, PNS activation will lower heart and breathing rates and blood pressure, helping the body to return to reparative and restorative functioning [15].

Advances in physiology and neuroscience demonstrate that resilience is maintained by healthy ANS functioning. Disruptions in ANS functioning can be observed by objectively and noninvasively measuring 2 biological indices [13,14,16]: heart rate variability (HRV) and respiratory sinus arrhythmia (RSA). HRV is the variation in the time between successive heartbeats and depends primarily on the extrinsic regulation of HR by the balance between the SNS and PNS branches of the ANS [17,18]. If the time between heartbeats is rather constant, HRV will be low, whereas if the time varies, HRV will be higher. When a person is at rest, a high HRV signals good health because it represents a physical indicator of autonomic flexibility in responding to challenges [18,19]. In contrast, low resting HRV signals an imbalance in autonomic function, such as in clinical cases of anxiety and psychopathology, where there is an overactivation of the sympathetic branch and a weakening of parasympathetic influence over the heart [20]. RSA refers to the synchronization of HRV with breathing such that HR speeds up during inhalation (shortening the time between heartbeats) and slows during exhalation (lengthening the time between heartbeats) [21,22]. Increasing RSA builds physical wellness capacity by strengthening an individual’s long-term reserves, which is theorized to occur via multiple pathways (ie, baroreflex control, blood pressure regulation, and efficiency of pulmonary gas exchange) [21,23].

There is evidence to suggest sex can influence ANS functioning, as indicated by differences in HRV and RSA at rest [24]. These findings have direct implications for evaluating and understanding biological resilience and the effectiveness of psychophysiological interventions. Policing is known to be both intensely stressful [25] and a hypermasculinized profession [26,27], both of which have significant consequences for women’s professional experiences of stress and health. What remains to be seen is whether sociocultural gender and related occupational stressors (eg, gender discrimination and harassment) might contribute to differences in biological functioning and the effectiveness of physiologically focused wellness interventions.

Research reveals a connection between respiration, arousal, and emotional control, such that how a person breathes sends signals to regions in the brainstem and forebrain that modulate arousal [28]. For example, periods of slow breathing are associated with SNS suppression and PNS stimulation, which may underlie the anxiety- and stress-reducing effects of RSA [28]. For nearly 3 decades, Lehrer and Gevirtz [21] have pioneered and refined a cardiorespiratory intervention called heart rate variability biofeedback (HRVBF). During HRVBF, a person is shown visual feedback of their beat-by-beat HR data while engaging in slow breathing with the goal of maximizing RSA. For clinically effective treatment, it is not sufficient to simply instruct a person to breathe slowly. Rather, HRVBF technology is required to identify the precise breathing rate for maximizing individual RSA and to condition this rate through training [29,30].

In collaboration with expert police practitioners as well as medical and academic experts in psychophysiological risk and resilience, the lead author developed the International Performance Resilience and Efficiency Program (iPREP), an in-person resilience program that improves occupational health and performance by addressing the unique acute and chronic stressors faced by police [31-33]. This program has been refined across 6 grant-funded multimethod studies with over 300 police participants measuring objective psychophysiological and behavioral data [31-40]. The components of the currently proposed intervention are based upon the same core scientific principles and validated intervention methods as iPREP, including mental health awareness and support [2] and psychophysiological techniques (ie, HRVBF) to promote resilience by conditioning adaptive HRV, RSA, and HR functioning [30,41,42]. Research has shown that the autonomic modulation techniques instructed during iPREP result in significant, long-term reductions in occupational errors and improvements in officers’ ability to recover quickly from acutely stressful events when completed during scenario-based, in-person training [31-33].

The aim of the proposed study is to develop (phase 1) and test (phase 2) if a web-based delivery of autonomic modulation training (AMT) can reduce symptoms of PTSI and build psychological wellness capacity and physiological resilience in police officers with similar effectiveness as in-person training. An additional novel scientific contribution of this protocol includes an examination of sex and gender in baseline biological and psychological presentations of PTSI among police and in response to a resilience-building intervention.

This protocol has been approved by the University of Toronto’s Social Sciences, Humanities, and Education Research Ethics Board (Protocol #39478).

Capitalizing on the numerous long-standing professional relationships between study coinvestigators and police populations of interest, phase 1—development of the AMT intervention—will use an integrated knowledge translation (iKT) approach to generate user-informed content [42]. To test the effectiveness of the AMT intervention (phase 2), a cluster randomized control trial with parallel assignment will be used on several cohort samples of active-duty Canadian police officers. This study has been registered with ClinicalTrials.gov (NCT05521360).

We aim to test the AMT intervention on a sample of approximately 250 frontline police officers in rolling cohorts of 10-15 participants (half AMT, half waitlist control) from geographically representative police services across Canada. A power analysis was conducted using G*Power software [43] and found that a minimum total sample size of 128 participants was required (effect size f=0.25, α=.05, power=0.80, number of groups=2, number of repeated measures=2). However, we plan to oversample officers by sex and mitigate possible attrition based on the principal investigator’s previous experience with conducting longitudinal research on police officers [31-33]. A small laboratory-based prepilot sample (n=10) will test equipment reliability, accuracy, and integration of HR monitors with HRVBF and learning management system (LMS) platforms. An initial pilot cohort 1 (n=30) will then test the AMT intervention in the population of interest (ie, law enforcement officers) using a combined laboratory-remote paradigm to validate the implementation of study procedures. These participants will also be invited to participate in a poststudy focus group as per the study protocol (see iKT Focus Groups in Procedures) to optimize the intervention (ie, content and delivery) and overall study procedures to provide stakeholder- and user-informed feedback. When all user feedback has been incorporated, all subsequent participants will be tested fully remotely and in accordance with the randomized control trial procedures and group allocation.

Participants will be recruited in several ways: (1) partnering police services will provide a link to the web-based study hosted on the confidential University of Toronto (ie, not police-related) servers to members via email, announcements, and internal media postings with encouragement to participate; (2) the research team will visit police services and other related fora (eg, industry meetings and conferences) to hold recruitment and information days where eligible officers may sign up for the study; (3) our knowledge user (MLV) and expert police practitioner collaborator (PS) will facilitate recruitment of law enforcement officers through targeted media attention, networking, and professional outreach activities; (4) supporting stakeholder groups will facilitate communication and recruitment with users and law enforcement officers. Based on delays due to the COVID-19 pandemic, the order in which partner organizations are recruited and tested is subject to change based on availability and current public health guidelines.

Interested participants will be directed to a web-based consent form on a secure site that will detail the study aims, eligibility criteria, intervention protocol, allocation and procedures for experimental and control groups, risks and benefits, and other information related to their rights as a research participant (ie, confidentiality and contact information for PIs and the institutional ethics board). Participants will provide their informed consent by clicking on a link that will take them to an eligibility screening survey that will collect the following identifiable information: full name, mailing address (for equipment shipping and incentive payment), personal email address (for correspondence with the research team that is not monitored or accessible by the participants’ police agency), and demographics including sex at birth, gender, age, ethnicity, police service (to verify active duty status against publicly available directories), role, years of service as a police officer, and confirmation that they are not currently on extended medical leave. Participants deemed eligible for the study will be provided with deidentified usernames and passwords to access the preintervention assessment (see phase 2 procedures).

Inclusion criteria included adults (aged 18 years or older) currently employed as active duty (ie, not retired or on an extended medical or disability leave) frontline law enforcement officers fluent in English. Exclusion criteria include non-Canadian law enforcement officers, police administrators, civilian employees (ie, nonsworn members), and officers that do not have regular access to a computer with internet access to complete the web-based AMT intervention. Scores on self-reported mental health screening tools completed during the preintervention sessions will be used to match the assignment of participants to experimental or control groups but will not be used to exclude participants based on established cutoff scores for high or severe PTSI symptoms.

In accordance with our hypotheses, PTSI (H1) will be measured by the PHQ-8, GAD-7, DASS-42, PCL-5, AUDIT, PSS, CBI, and self-reported stress visual analog scale. Self-reported resilience and well-being (H2) will be measured by the Brief COPE, RSA-33, BRS, WBSI, PTQ, UB-PSWQ, and UB-RRS. Functional wellness capacity and occupational stress will be measured by the SDS, GHPS, COVID-19 HRSI, WHODAS 2.0, PSQ, and LEAPS. Sex will be measured by self-reported sex at birth (male, female, or other); gender is measured by self-identification as male, female, nonbinary, or other (with the option to self-report) or “prefer not to disclose.” Sex and gender-related experiences will be measured by the WGDS and GEQ (H3 and H7). Changes in resting HRV (H4) will be measured by time and frequency domain parameters based on data collected from the wearable HR monitors (eg, high- and low-frequency HRV, low- or high- frequency ratio, and the Root Mean Square of Successive Differences, [RMSSD]) as recommended in the standards of practice for HRV analysis [17]. RSA (H5) will be measured by obtaining respiratory frequency. Greater power at the respiratory frequency indicates a higher level of RSA and will be analyzed using the gold standard HRV analysis program, Kubios [92]. Recovery from acute stress (H6) will be measured following completion of the web-based PASAT and emotional Stroop Tasks. Acute stress recovery will be assessed via a time-domain measure of HRV (RMSSD) [17].

Any personally identifiable information provided in the eligibility questionnaire will be downloaded and immediately deleted from web-based platforms, stored offline in encrypted files on password-protected computers at the University of Toronto accessible only by members of the primary research team approved by the lead principal investigator (JPA), and will not be shared with collaborators or participating police organizations. Data collected on the LMS and secure survey platforms (ie, pre- and postintervention assessments and responses to intervention exercises) will be matched to a deidentified user ID. Physiological data gathered by the wearable HR monitor and HRVBF app will be stored locally and will not be transferred to publicly accessible third-party cloud-based servers. Rather, participants will mail the equipment back to the research team, which will match the deidentified data offline. The key code linking participants’ personal information and deidentified user IDs will only be stored on secure servers at the University of Toronto and will be destroyed once the participant completes or withdraws from the study, has received their compensation, and all deidentified data files have been matched.

Physiological data will be transferred to a 2-time Canada Research Chair in Health Informatics and Big Data expert coinvestigator (CM) via a secure virtual private network (VPN) tunnel and housed on a private cloud-based Big Data analytics platform located within the Compute Ontario Advanced Research Computing equipment at the Centre for Advanced Computing (Queen’s University). CM will design the data management, cleaning, and analysis protocols necessary to construct the big data pipeline to process the large amount of continuous data, including matching the time stamps on the physiological data to significant event markers in the AMT intervention (eg, questionnaires, PASAT, and emotional Stroop tasks) and on the HRVBF app (eg, paced breathing exercise). Preprocessed data will then be sent back to the primary research team (JPA, PMD, and SCS) via a secure VPN tunnel and housed on the secure internal University of Toronto server (RES-NAS) for the statistical analyses described below using bespoke syntax and code on various software platforms (eg, Kubios, MatLab, and SPSS). Data from iKT focus groups will be stored in RR’s secure data web-based Al Fresco space and hard copies in her triple-locking filing cabinet in her office at the Memorial University of Newfoundland.

To test the effectiveness of the AMT intervention on reducing PTSI symptoms and improving resilience, we will conduct prepost analyses of the self-reported PTSI (H1) and resilience measures (H2) using separate repeated measures analyses of covariance (ANCOVAs) accounting for demographic covariates for normally distributed data, and ranked repeated measures ANCOVAs for nonnormally distributed data [93]. Variables including sex, gender (H3), age, years of service, and geographic service region (eg, urban or rural; East, Central, or West) will be included as covariates.

To test the effectiveness of the AMT intervention on improving biological indicators of resilience, prepost analyses of resting HRV parameters (H4), RSA (H5), and recovery from acute stress (H6) will be separately assessed using repeated measures ANCOVAs with sex, gender (H7), age, years of service, and geographic service region as covariates. To determine which factors significantly predict greater changes in biological outcome measures, hierarchical linear regression analyses will be undertaken with self-reported measures and demographic variables entered as predictors in the models. We will also compare weekly changes in HRV and RSA as measured during AMT module exercises with repeated measures ANCOVAs. A main effects analysis of the “time” variable will explore the potential incremental benefit of AMT training and identify at what point during the 6-week intervention significant biological changes occur. Computational approaches to analyzing and visualizing biological indices of resilience will be led by CM, including integration of cardiac (eg, HRV) and respiratory (eg, RSA) rhythms to elucidate patterns of desynchronization before, during, and following AMT.

To explore sex and gender-based contributions to self-report (H3) and objective indices (H7) of PTSI and resilience, the above analyses will include sex and gender as covariates or predictor variables with sex-stratified samples. These findings will clarify any sex-related differences in baseline and postintervention rates of PTSI and resilience among law enforcement officers and whether there are any sex-related differences in the magnitude of treatment effects (H3). Moderated regression analyses will test whether the AMT intervention affects the relationship between sex and PTSI symptoms and resilience. Multiple comparisons will be adjusted with Bonferroni corrections, where applicable.

Qualitative analyses of the transcribed data obtained from the iKT focus groups will be led by the sex and gender coinvestigator (RR) with the assistance of graduate students and postdoctoral fellows, with the aim of identifying sex- and gender-based considerations for developing effective occupational health interventions tailored to police officers and other PSP groups. Qualitative data analysis will use an inductive, comparative approach without initially arriving at any definitive, substantive, or theoretical conclusions about what the data reflect interpretatively with a lens on sociology and criminology [94]. Themes will emerge through the dynamic interaction of participants. Coding will be performed using NVivo qualitative analysis software (Lumivero), allowing for comparison across sessions [89] and the emergence of prominent themes through the tracking of coding “nodes” both across and within groups. Regular research team meetings will ensure that thematic development emerges in a consistent and reliable manner and help to ensure the hermeneutically attuned validity of the data [90].