Post-Code PTSD Symptoms in Internal Medicine Residents Who Participate in Cardiopulmonary Resuscitation Events: A Mixed Methods Study Open Access

Internal medicine (IM) residents participate in and lead cardiopulmonary resuscitation events or “codes,” which are unpredictable, stressful events that involve the death or near death of patients.^1–3  Experiencing traumatic situations can cause posttraumatic stress disorder (PTSD),⁴  and prior studies have linked physicians' experiences in some settings to elevated PTSD rates.^5–11  Physicians with PTSD symptoms may experience anxiety, depression, and burnout.⁶  Exposure to medical trauma has also been associated with decreased job performance and quality of life.^6,12  This is the first study, to our knowledge, to explore the occurrence of PTSD symptoms in IM residents after code participation (post-code PTSD symptoms).

This study used a mixed methods design. Phase 1 was a qualitative analysis of interview text from IM residents from 9 institutions. In phase 2, we administered a questionnaire to IM residents at a single site. The University of Wisconsin Health Sciences Institutional Review Board approved all facets of this study.

We interviewed 25 IM residents from 9 programs about their experiences leading codes.¹³  Participant recruitment, data collection, and initial analytic plan have previously been reported.¹³  “Codes as trauma” emerged as a content area, but was not further explored in the original study. For the present study, 2 researchers (C.K., M.B.) reanalyzed the text to identify remarks that described codes as trauma, referenced symptoms in 1 of the 3 symptom clusters of PTSD,¹⁴  or described factors that mitigated or aggravated post-code stress. Consensus was reached in classifying segments of text through repeated discussions. Dual analysis of 5 transcripts indicated ≥ 97% interrater agreement (Cohen κ coefficient = 0.647).¹⁵  Qualitative software (NVivo 10, QSR International Inc) was used for the analysis.

In October 2013, we invited all IM residents at the University of Wisconsin Hospitals and Clinics to participate in an anonymous, online survey. Participation was voluntary. Of 82 eligible residents, 51 (68% of female residents and 58% of male residents) completed the 5-item questionnaire, and 42 (19 women, 23 men) reported participation in a code and were included in analysis (table 1). The questionnaire included the 4-item Primary Care PTSD Screen (PC-PTSD),¹⁴  with the only modification being a focus on code participation. The questionnaire assessed whether residents had participated in a code during their residency that was so frightening, horrible, or upsetting that in the past month they (1) had nightmares about the code or thought about the code when they did not want to; (2) tried hard not to think about the code or went out of their way to avoid situations that reminded them of the code; (3) were constantly on guard, watchful, or easily startled; or (4) felt numb or detached from others, activities, or their surroundings.¹⁴  As recommended for the PC-PTSD,¹⁴  endorsing 2 or more symptoms indicated a positive screen for PTSD. Residents responded to a fifth question, which asked if they felt that their past participation in a code had “a negative impact on some aspect(s) of their performance as a resident.” The response used a Likert scale (1, not at all, to 5, extremely).

We assessed the association between screening positive for PTSD and reporting that code participation negatively impacted work performance through univariate analysis of variance (ANOVA), with PTSD screen as an independent variable, and the impact of code participation scale as a dependent variable. Regression analysis examined the association between the number of PTSD symptoms and the reported impact of code participation. Alpha values were set at α = 0.05 level for tests of significance. All statistical analysis used SPSS Statistics for Windows version 21.0 (IBM Corp).

A total of 7 of 25 (28%) residents (all postgraduate year [PGY]-3s, 5 women) characterized codes as traumatic events or described PTSD symptoms. In describing how they dealt with PTSD symptoms, residents described trying to suppress thoughts and images of a code experience. Some also described ways in which supervisors were helpful in reducing or contributing to the trauma of a code. Illustrative text from resident interviews is shown in table 2.

The 4-item screening questionnaire showed moderate evidence of reliability with a Cronbach α value of 0.65 (0.70 recommended for scale reliability).¹⁶  The majority of the sample (36 of 42, 86%) answered “no” to all 4 scale items. Six of 42 responders (14%) answered “yes” to 1 or more items, and 3 (7%) screened positive for PTSD symptoms by endorsing 2 or more scale items. All were women.

The ANOVA results suggested that residents who screened positive for PTSD reported that code participation had a significantly greater negative impact on their work performance (mean =  2.00, SD = 0) than residents who screened negative for PTSD (mean = 0.26; SD = 0.50; F_1,40 = 6.53; P = .02; 95% CI −1.33 to −0.16). Regression analysis showed the number of PTSD symptoms endorsed was significantly associated with the level of negative impact reported (t_3.62 = 40; P < .001; R² = 0.25; 95% CI 0.28–0.98). Residents who endorsed more PTSD symptoms were more likely to indicate that code participation had a negative impact on their work performance.

Our study suggests that some IM residents experience PTSD symptoms that may be associated with their participation on codes and that these symptoms may be associated with self-reported negative impact on work performance. Other studies have found that residents involved in medical traumas are susceptible to PTSD symptoms.^7–9  Our study is the first, to our knowledge, to have residents describe their experience, attribute symptoms of traumatic stress to code involvement, and report how they dealt with the traumatic stress of code participation. Like earlier studies, this highlighted the benefit of talking with others to mitigate stress,^17,18  and the role of attending physicians to mitigate the impact of traumatic code experiences.^19–21  Our data suggest some residents use maladaptive coping strategies, such as thought suppression, when experiencing PTSD symptoms. This points to the need for interventions to reduce the traumatic impact of codes for some IM residents.

Despite the low overall prevalence of PTSD and the small sample in our study, we found PTSD symptoms were more often endorsed by women. While female residents may be more willing than male residents to endorse PTSD symptoms on a questionnaire, our results align with reports showing that women are more likely to meet the criteria for PTSD and to report greater increased PTSD than men for a single type of traumatic event.^22,23  Previous work suggests the need to enact code leadership in a stereotypically masculine manner may induce additional code-related stress for female residents.¹³  Simply acknowledging the existence of sex-based assumptions and their potential impact could alleviate this added stress.²⁴ 

Our findings support the need for IM residency programs to further investigate PTSD symptoms. Training on how to prevent and cope with post-code PTSD symptoms could include

• •

  Acknowledging that participating in codes is traumatic and that experiencing powerful emotions is normal,

• •

  Encouraging faculty to provide supportive statements during and/or after a code to reduce (and never reinforce) residents' feelings of self-blame,

• •

  Affirming that attempts to suppress traumatic thoughts can exacerbate and prolong post-code PTSD symptoms,

• •

  Encouraging access to social support that provides opportunities for discussing code events with peers and supervisors, and

• •

  Other strategies might include providing forums for residents to write about or discuss their experiences.

Our study has several limitations. Qualitative data is valuable in exploring human experience but is not generalizable. The questionnaire relies on self-report, has no established validity evidence in this population, and used a single question to capture the residents' perception of impact on work performance. Our findings also are from a single site, and our response rate was low, introducing the potential for sampling bias.

This mixed methods study found that some IM residents may be at risk for developing symptoms of PTSD that could negatively affect their work performance, and these residents may benefit from prevention and supportive measures during training.