Impact of new technologies on stress, attrition and well-being in emergency call centers: the NextGeneration 9–1-1 study protocol

Like the first responders they support, TC work is characterized by high demand and routine exposure to workplace stressors that are associated with a range of psychological and physical health consequences [7]. TCs are primarily involved in the initial response involved in an emergency and rarely play a part in final resolution or outcome of the call which may be a source of stress itself. TCs not only feel responsible for the safety of the public but also perceive themselves as responsible for the safety of the dispatched personnel [8]. And despite lack of a direct physical exposure to trauma, TCs also experience duty-related distress that enhances risk for adverse psychological outcomes. Peritraumatic stress—defined as the responses during or immediately after trauma exposure—have been shown to predict depression and post-traumatic stress disorder (PTSD) among firefighters, police officers, emergency service personnel [9, 10]. Only recently has this same peritraumatic association with depression and PTSD been identified in the TC workforce [11, 12]. In a study of secondary trauma exposure, Troxell (2008) reported high levels of peritraumatic distress and a positive relationship between peritraumatic distress and secondary traumatic stress in 9–1-1 TCs [5]. Similarly, high levels of peritraumatic distress and a moderate, positive relationship between peritraumatic distress and PTSD symptom severity have been reported in TCs, suggesting that direct, physical exposure to trauma may not be necessary to increase risk for PTSD in this group of workers [12, 13]. However, few studies have focused specifically on the association between TC workplace stress and psychological health.

Research on workplace stress in call centers has reported associations with attrition, burnout, absenteeism and early retirement. While limited to call centers in WA State, a recent survey found that 42% of TCs assess their job as “stressful to very stressful” and 60% reported their job as demanding (47%) or extremely demanding (14%), with no differences by gender or years of service, suggesting that stress levels do not abate over time [14].

The cost of this demanding workplace cannot be ignored. A 2012 survey found that 63% of emergency center managers associated job stress with TC turnover [15]. Although there is great variety in size, location and type, 83% of centers in a call center retention study reported an increase in the number of dispatched calls as compared to the previous three years while struggling with a national average annual turnover rate of 19% [16]. “Burnout” (a combination of exhaustion, cynicism, job detachment, a sense of ineffectiveness and lack of accomplishment or perceived lack of recognition/appreciation [17]) was identified as a primary cause of these lowered retention rates [16]. Research has also positively associated TC burnout with both sickness- and work-related absenteeism [18], an increased likelihood of lower job performance [19], depressive symptoms and lower job/life satisfaction [20].

Workplace factors, such as perceived manager support, sense of control over one’s schedule and social relationships with co-workers, including workplace climate characteristics such as bullying/harassment, can all influence stress levels and outcomes [18, 21, 22]. Individual characteristics, such as overcommitment (an exhaustive pattern of coping associated with the inability to withdraw from work demands) has been positively associated with high stress scores in TCs [14]. However, a variety of factors, both environmental and individual, may buffer the impact of negative outcomes associated with stress. Lifestyle factors, such as physical activity, and emotion-regulating skills can reduce the impact of environmental stressors as can work-life balance [21, 23]. In addition, individual characteristics and differences, such as coping strategies, self-efficacy, resiliency and compassion satisfaction can play a part in reducing or ameliorating high levels of work-related stress [24‐29].

To date, studies of technology adoption by emergency call centers have focused on the ability of new technologies to improve the speed of emergency response, patient health outcomes and mortality, and hospital costs [30]. “Technostress” is defined as stress created by adoption of information and communication technologies which can impact worker productivity, job satisfaction, commitment to the organization, and retention [31, 32]. Another issue specific to text-to-911 calls is TCs reliance on verbal communication to obtain and convey critical, clear, time-sensitive information, as well as take the caller’s language, voice, tone and auditory background and surroundings into account in order to successfully delivering lifesaving information [33]. Research has found that when communication challenges affect efficient information gathering, TC stress can be exacerbated [34, 35]. Other NG911 implementation impacts of concern are possible health impacts [33]. While 9–1-1 TCs will receive training to adapt to the new technologies deployed through the NG911 initiative, the focus is solely on technical skills, not human resource skills such as how to handle psychological stress due to hyper-tasking, exposure to multi-media images (to which TCs have not historically been exposed and which may enhance peritraumatic distress and thus, heighten risk for PTSD), or cumulative stress effects that could increase risk for stress and mood-related disorders. Exposure to photo uploads of graphic physical injuries or streaming videos of fatalities on the Interstate may increase occupational stress and distress. In addition, it is unknown how performance might be impacted, for example, when a TC needs to issue character-limited text message instructions for clearing a blocked airway or administering CPR.

The “NextGeneration 911: From Multi-tasking to Hyper-tasking” Project is designed to systematically describe the impact of the NG911 technologies implementation on TC stress and job performance, focusing on the first NG911 technology to be implemented: text-to-911 calling. The project’s Specific Aim is to:

Measure levels of stress, job satisfaction and job performance among 9–1-1 TCs before and after the NG911 technologies implementation.

Study objectives include:

We plan to measure levels of stress among 9–1-1 TCs before and after implementation of text-to-911; explore variables of potential influence on stress, such as social support; and evaluate the relationship between those variables on stress-related outcomes such as sick days and intent to leave employment. We hypothesize that implementation of text-to-911, one of the NG911 technologies, will have a significant effect on TC stress levels across multiple indicators.

Two conceptual frameworks inform the project approach: the Demands-Control model (DCM), currently considered one of the most influential models linking job characteristics (e.g., low levels of control and work overload) to job stress and stress-mediated health outcomes [36, 37] and the Job-Demands-Resources (JDR) Model which stipulates that both demands and resources (such as supervisor support) can have multiplicative or cumulative effects on job stress and stress-related outcomes such as attrition and turnover intention, burnout/exhaustion and absenteeism (illness and work-related) [38, 39].

The DCM focuses on the psychosocial job characteristics of job demands and job control (sometimes referred to as decision latitude) within the context of the interactions between a worker and the environment, and the importance of identifying extent of worker decision authority (control over work situation) and skill discretion (possibility of using skills and competencies). High demands and low control predict “high-strain” which leads to increased levels of depression, fatigue and cardiovascular disease and mortality. Workers with high control are buffered from the negative effect of demands on outcomes which leads to lower levels of illness [36]. Social support also can act as a buffer in high demand situations [40‐42]. In the context of NG911 technologies implementation, the DCM suggests that a technology change such as text-to-911 will create a “High Strain” work environment with TCs at higher risk for elevated stress and related health problems which would influence the likelihood of higher turnover intention and absenteeism and reduce job satisfaction and job performance. However, TCs already work in a high strain work setting and adding more stressors could exacerbate an already taxed and burnt out workforce.

In addition, the DCM cannot explain why the same levels of demand and control in two workers of similar background can lead to differential behavioral or health outcomes [43] or that there are different types of demand, some of which may be positive, such as the challenge of learning a new skill [44, 45]. The JDR Model expands on the DCM by including resources: aspects of the workplace that help with achievement of work goals, reduce demands or stimulate growth and development. In this model resources of all types can buffer the impact of demands on stress outcomes [38] and also incorporates personal characteristics that contribute to resiliency to stress, such as self-efficacy, organizational-based self-esteem, and optimism, among others [39].

The NG911 Project is an observational, prospective cohort study. Between January 2015–June 2017 we will recruit TCs employed at emergency call centers nationwide to participate in a study to measure the impact of NG911 technologies implementation on worker stress, job satisfaction and job performance. Recruitment will continue until sample size requirements are met (see “Sample Size and Power Calculations” below). To address project aims, we include three data collection time points: baseline (before text-to-911 implementation); early implementation (within the first month of implementation); and 6 months after implementation. The first NG911 technology being widely implemented is text-to-911 and will be the focus of our study. Given text-to-911 will be deployed at different times at different call centers, data collection will be staggered and tailored to the implementation timeline at each participating center. The study has received Institutional Review Board approval (#49412, University of Washington [UW] Human Subjects Division, Seattle, WA) for all research activities.

As stated, NG911 technologies implementation is a phased process in emergency call centers across the country. Most call centers are beginning this process by enabling the text-to-911 functionality. Emergency call centers that anticipate deploying any NG911 technology during the study timeframe will be eligible to participate. Call centers already using any NG911 technology will not be eligible. All call center types will be eligible: primary public safety access points (PSAPs) which focus on police calls; secondary PSAPs (medical and fire); and centers that function as both primary and secondary PSAPs. Some emergency call centers employ uniformed staff (i.e., EMTs), however the majority of centers employ civilian TCs. We aim to include representation of TCs across small and large emergency call centers. Recruitment will begin in WA State and will expand as needed throughout the region and nationwide to ensure representation and diversity in the sample.

9–1-1 TCs working at any emergency call center will be eligible to participate. Given hiring requirements, no individuals under age 18 will be included. We will limit study participation to TCs which have completed training and are operating as fully functioning, independent TCs. TCs still in training will not be eligible.

Recruitment will be conducted in coordination with managers at the participating call centers. TCs will be recruited by memos, flyers and other outreach methods at call centers. Information about the study and invitations to participate will also be posted through national emergency call center forums. All recruitment materials will include information about the study, its purpose and methods, and study investigator contact information so potential subjects can ask questions about the study prior to enrolling.

Our primary outcome is symptoms of stress as measured by the Calgary Symptoms of Stress Index (cSOSI) [46]. Our aim is to measure changes in stress that occur with the introduction of NG911 technologies in call centers. Of primary interest are the comparisons between pre-implementation stress levels and post-implementation stress levels, and between the immediate post-implementation stress levels and long-term post-implementation stress levels.

Calculations of sample size were performed using standard sample size calculation software for a paired t-test [47]. Based on estimates of within-participant standard deviation from a longitudinal study using the cSOSI [48] sample size calculations assumed an effect size of 0.25 and a type I error rate of 5%, which will require 160 participants with data. Assuming an 18% loss to follow-up will require a total of 200 participants. With this sample size we estimate that we will have 80% power to detect a change in cSOSI score of 11 points. Previous studies have used an effect size of 0.4 as clinically meaningful [49‐55].

All instruments will be built using the University of Washington’s Catalyst survey tool (www.​uw.​edu/​lst/​web_​tools/​webq) which has been approved by the UW Human Subjects Division for creating anonymous or confidential surveys. Participation in the study will be voluntary and confidential.

The following instruments will be used to collect data:

All analyses will be conducted in R [56].

The timing of the data collection at baseline, during and post-intervention will be determined by the staggered implementation of text-to-911 at the emergency call centers participating in the study. At each data collection point, study participants will be asked to complete a 30-min online survey around worker stress and demographics as described in “Study Variables and Measurement Instruments” above.

The study’s findings will be submitted for publication in peer-reviewed journals and disseminated at national conferences and through emergency call center quarterly journals and social media. We will share the actual data only on the basis of agreement that the data will be used solely for research, that no individuals will be identified in any manner, that data will be secured by electronic safeguards, and that once data analysis is complete, the data will be returned or destroyed. In short, we plan to share the data in accordance with NIH policy.

A primary issue for this project is scheduling. Our recruitment schedule is predicated on timely deployment of NG911 technologies. Emergency call centers across the country have been anticipating this implementation for several years and budget and equipment selection decisions have routinely delayed the timeline. We believe we can overcome this limitation as all data collection instruments are available online so wide expansion of recruitment efforts is easily facilitated. In addition, deployment when it occurs has been initially limited to text-to-911 technologies so our study findings may be limited to stress impacts of this one technology rather than the entire suite of technologies included in the NG911 initiative.

It is possible that emergency call centers that elect to participate could introduce a selection bias in the subject sample. However, given the widely expressed concerns about overall NG911 technologies implementation, we believe any generalizability issues will be overcome by our recruitment strategy, which includes sampling from a variety of call centers that differ in organizational structure, size, service area and population coverage. An additional selection bias is that those who self-select into the study may represent levels of resilience or other personal psychological characteristics that differ from the general 9–1-1 TC workforce. However, we believe our recruitment strategy across different call centers will mitigate this generalizability issue. A final limitation is that newly hired TCs, as compared to seasoned workers, may be more comfortable with new technologies than the current workforce.