2014 Volume 56 Issue 6 Pages 430-443
Objectives: Coping with difficult care-related situations is a common challenge for health-care professionals. How these professionals deal with the regrets they may experience following one of the many decisions and interventions they must make every day can have an impact on their own health and quality of life, and also on their patient care practices. To identify professionals most at need for extra support, development and validation of a tool measuring coping style are needed. Methods: We performed a survey of physicians and nurses of a French-speaking University hospital; 469 health-care professionals responded to the survey, and 175 responded to the same survey one-month later. Regret was assessed with the regret coping scale developed for this study, self-report questions on the frequency of regretted situations and the intensity of regret. Construct validity was assessed using measures of health-care professionals' quality of life (including job and life satisfaction, and self-reported health) as well as sleep problems and depression. Results: Based on factor analysis and item response analysis, the initial 31-item scale was shortened to 15 items, which measured three types of strategies: problem-focused strategies (i.e., trying to find solutions, talking to colleagues) and two types of emotion-focused strategies, A (i.e., self-blame, rumination) and B (e.g., acceptance, emotional distance). All subscales showed high internal consistency (α >0.85). Overall, as expected, problem-focused and emotion-focused B strategies correlated with higher quality of life, fewer sleep problems and less depression, and emotion-focused A strategies showed the opposite pattern. Conclusions: The regret coping scale (RCS-HCP) is a valid and reliable measure of coping abilities of hospital-based health-care professionals.
(J Occup Health 2014; 56: 430–443)
Both affective1) and cognitive2) processes impact clinical decisions3), such as ordering an additional test or choosing a specific treatment. In particular, the experience of regret is ubiquitous among health-care professionals and plays an important role in clinical decision making4–6). Care-related regret7) is an emotion experienced when physicians and nurses make decisions or perform actions that retrospectively appear ineffective, futile, or inappropriate8).
Although care-related regrets are a common and most likely inevitable component of decision making, health-care professionals may differ in how they cope with regret, that is, in their coping style. A frequent classification of coping strategies9) contrasts problem- and emotion-focused strategies10). Problem-focused strategies aim to change the environment to rectify the situation or to avoid its recurrence. Emotion-focused strategies aim to integrate one's experience of regret and readapt thinking and behavior accordingly. Thinking about what happened on a regular basis or accepting one's limits11, 12) are good examples of these strategies. Seeking social support is considered either problem or emotion focused depending on its goal. For health-care professionals, talking to colleagues or superiors would probably be associated with problem-focused strategies, as it could lead to change in practice. While the best strategy is situation specific and no single strategy can be described as generally better than others, some strategies, such as rumination and self-blame, are more consistently associated with negative outcomes13).
How health-care professionals cope with the regrets they experience (i.e., regret regulation) may impact both patient care and the professionals' own health14). For instance, regulation strategies have an influence on the physical and mental health of decision makers15, 16), such as sleep disorders17, 18) and attention deficits7), which may in turn increase the risk of errors19). Among physicians and nurses, poorly regulated regrets have also been shown to be related to self-reported changes in practice, irrespective of the presence or not of a human error or patient complications7, 16).
The aim of this study was to develop a scale to measure habitual coping style to deal with care-related regret among health-care professionals. Existing regret coping scales with good psychometric properties are not appropriate for various reasons. First, most instruments have been created for patients facing a single serious situation (e.g., new cancer patient)20). In contrast, a scale for health-care professionals should be general enough to assess coping style in general (i.e., the whole set of strategies used). Second, some instruments only cover some strategies, such as emotion-focused strategies in their work11, 12). Third, health-care professionals will certainly use some specific problem-focused strategies (e.g., presenting excuses to a patient). Fourth, regret is related to decision-making, whereas many coping scales focus on stress in general and do not capture decision-related strategies.
The study was conducted at the [hidden to preserve blind review], a 2000-bed public teaching hospital that includes acute, primary care, psychiatric and geriatric facilities. Questionnaires and up to three reminders were sent to nurses and medical doctors at their home addresses in 2011. To explore the properties of a 31-item scale, a study sample of 310 health-care professionals was required (10 × number of items)21). Since care-related regrets may be a sensitive subject, we contacted a sample of 825 nurses and 825 physicians. Of these 1,650 questionnaires, the first 1,100 sent were identified by a subject-specific number and yielded a low response rate. We thus sent 550 additional questionnaires, which were completely anonymous, to determine if more thorough anonymity would increase the response rate. Comparative analysis of the two groups of respondents revealed a slight difference in response rate, a difference in regret intensity and a difference in 2 out of 31 regret coping questions but nearly no difference in quality of life, job and life satisfaction and self-reported health and self-esteem22). Thus, analyses were made on the pooled samples. Respondents were excluded if they had (a) not been working with patients for the last 5 years or (b) had recently retired. To assess test-retest reliability, a sample of volunteer respondents were asked to complete the questionnaire again one month later. The study was approved by the Research Ethics Committee of the University Hospitals of Geneva.
Scale developmentIn a previous qualitative study7), we explored the main strategies of regret coping8, 20) following care-related decisions or interventions among hospital-based physicians and nurses. Respondents described problem-focused strategies and four main emotion-focused strategies: rumination, suppression, acceptance, and self-attacking. Seeking social support was also frequently reported, and other strategies were reported infrequently.
Based on the published literature9–12) and on this qualitative exploration, we generated 73 items in French to measure problem-focused strategies, seeking social support, and emotion-focused strategies of regret among health-care professionals. A panel of six experts (two psychologists, three physicians, and one sociologist) examined these items, and 33 items were kept. The scale was pretested among 14 nurses and 5 physicians working at [hidden to preserve blind review]. We dropped 2 items because they were poorly understood, which resulted in a 31-item scale for assessment of coping style that included all strategies (full scale and instruction in French available from first author). After asking respondents to think of their regrets, the instructions clarified that respondents should indicate what they do to deal with their regrets (“Please think now of your regrets, major or minor, frequent or exceptional, regarding events or situations that happened with patients. In general, indicate how frequently these propositions characterize your experience. In general, when I regret events or situations with patients…). For each item, participants were asked to rate the frequency with which they used one of the strategies on a 4-point Likert-scale ranging from 1 (never or almost never) to 4 (always or almost always).
Questionnaire and variables for construct validityIn addition to the 31-item regret coping scale, single questions probing self-perception of the intensity of regret and the number of recent regret-inducing events (within the last month) were added: “In the last 30 days, at work, how many situations relating to patient care did you regret?” which the respondent answered in writing, and “What was the mean intensity you would give to these situations in the last 30 days?”, which the respondent answered with a visual analog scale ranging from 0 (none) to 10 (very high), respectively. The intensity of the most important regret experienced in the last five years was also measured on the 10-item regret intensity scale (RIS-10)16). Items were rated on a scale of 1 (strongly disagree) to 5 (strongly agree). This scale has shown very good internal consistency (α=0.87) and good test-retest reliability16), with an intraclass correlation (ICC2) of 0.7023).
To assess construct validity, we measured variables related to respondents' health status, job and life satisfaction and self-esteem. Satisfaction with life was assessed by the 5-item Satisfaction With Life scale (SWL) shown to have a high reliability (α>0.80 in three studies)24). The General Job Satisfaction Scale25) contains 5 items with the same rating scale as the SWL, with a relatively low reliability (α=0.61). The Insomnia Severity Index (ISI) consists of 7 items rated from 0 (none) to 4 (very much), and higher scores indicate more severe insomnia. The ISI evaluates sleep quantity, sleep quality, and consequences of insufficient sleep over a one-month interval, with a good internal consistency (α=0.80)26, 27). Self-reported health was measured by the first question of the SF-36 (“In general, do you think your health is…?”, rated from 1 (excellent) to 5 (poor)). Self-esteem was measured by 1 item (“I have a high self-esteem”), which was rated on a scale of 1 (completely disagree) to 5 (strongly agree). Both the self-reported health and the self-esteem single item measures have good construct validity, as self-reported health predicts mortality28) and self-esteem predicts depression and other physical and mental health constructs29). Depression was measured using the CESD-10, a 10-item scale with good internal consistency (α=0.85 in the general population) and test-retest reliability (r=0.71)30). Information on respondents' professional and sociodemographic statuses was also collected.
Based on previous literature15, 17–19) and on the results of a previous qualitative study7), our construct validity hypotheses were as follows. First, we expected all strategies to be associated with regret intensity (the number and intensity of regret episodes in the last month, and the intensity of the most regretted situation in the last five years). However, we hypothesized that some emotion-focused strategies (i.e., rumination, self-attacking) would be more often associated with negative outcomes, such as depression13), and that these strategies would be more strongly associated with very high regret intensity, as more appropriate strategies may be overwhelmed by very intense regret experiences. Secondly, we hypothesized that problem-focused and emotion-focused strategies (other than rumination and self-attacking) would be associated with higher job and life satisfaction, higher self-reported health and lower sleep problems and depression. Rumination and self-attacking strategies would show the opposite pattern of relationships and also be associated with lower self-esteem, since self-attacking leads to considering oneself as generally worthless.
AnalysisWe describe the socio-professional characteristics of the sample and the mean values of the scores measured by profession. Values for nurses and physicians were compared using chi-square tests for socio-professional characteristics and t-tests for scores.
Because nurses and physicians reported that the situations eliciting regrets were different between professions7), the structure of the scale and construct validity analyses were conducted separately for nurses and physicians. Item characteristics were examined to determine the percentage of missing answers. In the first step, principal component analysis (PCA) was used to determine the number of underlying dimensions, based on the scree plot, and Kaiser's criterion (eigenvalue>1). Description of the components was then facilitated by a factor analysis (varimax rotation) to interpret the loadings. Varimax rotation was used because coping strategies are supposed to be independent. A sensitivity analysis using oblique rotation confirmed that correlations between factors were low and loadings were similar.
In the second step, we used item response theory (IRT) using a graded response model31) on each subscale, as defined in the first step, to shorten the scale. IRT was used for each subscale to assess item information (i.e., item precision, based on difficulty and discrimination, for all values of the total scale), and item difficulty (a measure of how often an item is endorsed). Items were discarded if their information functions were low, or if their item difficulties or information were redundant with another item32). In the third step, a PCA including the remaining items allowed us to estimate the proportion of variance explained by the items. Finally, subscale scores were computed by taking the mean of the corresponding items, if more than half of them were answered.
Validity and reliabilityContent validity of the shortened scale was assessed by a panel of experts to examine if items still covered all types of strategies. Construct validity of each subscale, by profession, was examined using t-tests for categorical variables or correlations for continuous variables.
Global reliability of each subscale was assessed by Cronbach's alpha. Similarity of scores between the baseline survey and the one-month retest was assessed using the weighted kappa statistic for items and intraclass coefficient of correlation (ICC2) for subscale scores. A Bland-Altman plot of the differences between these scores (Y-axis) against the mean between the baseline and follow-up scores provided an estimate of the scores' variability over time (including measurement error). The limits of agreement are defined as 1.96 times the standard deviation of the difference.
As the structures of answers may substantially differ between physicians and nurses, PCA, factor analysis and IRT were also run separately for each profession as a sensitivity analysis. All analyses were done using R v.3.0.1 (R Core Team, Vienna, Austria) and with the ltm33) library for IRT analyses.
A total of 469 persons returned the survey (31.7% of the 1,484 persons eligible): 220 physicians, 240 nurses and 9 persons who did not indicate their profession. Among the 311 participants of the questionnaires identified by a number, 175 (56.3%) completed the retest one month after the initial survey.
The mean age of the participants was 39.2 (SD 9.2) years (Table 1), and most of the participants had several years of professional experience. While nurses and physicians differed in socio-professional characteristics, they were similar in regret intensity, number of regrets and coping strategies, except for emotion-focused strategies (excluding rumination and self-attacking), which were slightly more often used by physicians.
| Characteristics | Categories | Nurse, N (%) N=240 | Physician, N (%) N=220 | p |
|---|---|---|---|---|
| Socio-professional | ||||
| Sex | Male | 45 (18.8) | 96 (43.6) | <0.01 |
| Female | 195 (81.3) | 124 (56.4) | ||
| Age | <30 | 25 (10.4) | 28 (12.8) | <0.01 |
| 30–39 | 84 (35.0) | 126 (57.8) | ||
| 40–49 | 76 (31.7) | 45 (20.6) | ||
| >50 | 55 (22.9) | 19 (8.7) | ||
| Professional status | Nurse / resident | 217 (91.6) | 106 (48.4) | <0.01 |
| Head nurse / Board certified | 20 (8.4) | 113 (51.6) | ||
| Percentage of clinical activity | 0–50% | 12 (5.0) | 10 (4.6) | <0.01 |
| 51–80% | 100 (41.7) | 27 (12.4) | ||
| 81–100% | 128 (53.3) | 180 (82.9) | ||
| Years of experience | <3 | 15 (6.3) | 20 (9.3) | <0.01 |
| 3–5 | 8 (3.3) | 54 (25.1) | ||
| 6–10 | 50 (20.9) | 73 (33.9) | ||
| 11–20 | 67 (28.0) | 44 (20.5) | ||
| >20 | 99 (41.4) | 24 (11.2) | ||
| Scores, mean (SD) | ||||
| Regret intensity | 1.74 (0.65) | 1.70 (0.73) | 0.57 | |
| Nb regrets (last 30 d) | 1.36 (2.76) | 1.03 (1.58) | 0.11 | |
| Mean intensity (last 30 d) | 2.31 (2.78) | 1.94 (2.36) | 0.13 | |
| Coping strategies | Problem-focused | 2.92 (0.60) | 2.94 (0.59) | 0.72 |
| Emotion-focused A | 1.83 (0.57) | 1.86 (0.70) | 0.62 | |
| Emotion-focused B | 2.64 (0.51) | 2.76 (0.59) | 0.02 | |
| Job satisfaction | 4.45 (1.24) | 4.57 (1.28) | 0.32 | |
| Life satisfaction | 5.39 (1.06) | 5.38 (1.08) | 0.95 | |
| Self-esteem | 3.36 (0.84) | 3.54 (0.89) | 0.03 | |
| Self-reported health | 2.39 (0.86) | 1.89 (0.84) | <0.01 | |
| Sleep | 8.60 (6.03) | 5.66 (5.33) | <0.01 | |
| Depression | 1.88 (0.48) | 1.84 (0.39) | 0.37 |
Because analyses stratified by profession showed similar results for all PCA, factor analysis and IRT models (data not shown), only global analyses are reported here. The results of the PCA suggested a three-component solution based on the scree plot and eigenvalues greater than 1. The subsequent factor analysis showed that 4 out of 31 items had low loadings (<0.3) on all factors. The remaining 27 items loaded significantly on the three factors (Table 2): problem-focused (PF) regret coping, emotion-focused A (EFA) strategies (including rumination and self-attacking) and emotion-focused B (EFB) strategies (including reappraisal, distanciation and acceptance). Based on the IRT analyses for each scale separately, 16 of the 31 items were dropped because their information functions (Fig. 1, gray lines indicate dropped items) showed that they were not very informative. They were either redundant with other items (e.g., items PF6 and PF26 have similar information functions) or made use of only one or two response modalities. Item 14 loaded relatively highly on both the problem-focused and EFA factors and was initially included in the IRT analyses of both scales. It was finally dropped from both scales. A PCA on the selected 15 items found a clear separation between the three components, which explained 54.2% of the variance (loadings of the factor analysis in Appendix 1). The 15-item scale had good content validity, as it covered all types of strategies.
| Factor 1 | Factor 2 | Factor 3 | |
|---|---|---|---|
| Problem-focused regret coping | |||
| 1. I talk about it with colleagues, to be listened to or reassured. | 0.63 | ||
| 2. I discuss the problem again with the patient (or his family) | 0.57 | ||
| 3. I try to find concrete solutions to the situation | 0.48 | 0.11 | |
| 6. I talk with a supervisor to prevent these events from recurring | 0.71 | ||
| 28. I expose the situation to colleagues to improve our practices | 0.67 | ||
| Emotion-focused regret coping A | |||
| 11. I think I am no good | −0.15 | 0.62 | |
| 13. I turn these situations over in my mind all the time | 0.81 | −0.12 | |
| 16. I think about it so much that it becomes invasive | 0.73 | −0.14 | |
| 18. I have a tendency to blame myself | 0.75 | ||
| 27. I think about these situations all the time | 0.67 | −0.19 | |
| Emotion-focused regret coping B | |||
| 8. I try to accept the situation | 0.64 | ||
| 19. I tell myself that “to err is human” | 0.60 | ||
| 21. I try to create some emotional distance | 0.10 | −0.12 | 0.52 |
| 29. I try to see the positive side of things | 0.24 | −0.21 | 0.43 |
| 30. I try to put the situation in perspective | 0.66 | ||
| Additional 16 items | |||
| Problem-focused regret coping | |||
| 9. I present my excuses to the patient (or his family) | 0.45 | 0.17 | 0.13 |
| 14. I try to analyze the event | 0.35 | 0.12 | 0.26 |
| 17. I change things so the situation does not happen again | 0.27 | 0.18 | 0.22 |
| 22. I do not talk about it to anyone, friends or family | −0.35 | ||
| 26. I expose the problem to my supervisor if he doesn't know about it | 0.66 | ||
| 31. Since then, I avoid handling similar situations | 0.20 | −0.12 | |
| Emotion-focused regret coping A | |||
| 7. I tell myself that I have betrayed the patient's trust | 0.35 | ||
| 20. I feel responsible for what happened | 0.53 | 0.12 | |
| Emotion-focused regret coping B | |||
| 4. I prefer not to think about it | −0.37 | 0.33 | |
| 5. I try to draw a lesson from it | 0.30 | 0.23 | |
| 10. I rationalize in order to avoid being overwhelmed by emotions | −0.15 | 0.18 | 0.46 |
| 12. I talk to family or friends to get support | 0.28 | 0.29 | |
| 15. I tell myself that I (we) have done the best I (we) can | −0.13 | 0.41 | |
| 23. I tell myself that perfection doesn't exist | 0.52 | ||
| 24. I try to evacuate any thoughts about it | −0.21 | 0.47 | |
| 25. I tell myself I can do better | 0.19 | 0.35 |
Loadings with absolute values =0.1 were omitted to facilitate reading.
Based on the numbers indicated in front of the items in this table, the 15-item scale would present the items in the following order: 1, 2, 3, 6, 8, 11, 13, 16, 18, 19, 21, 27, 28, 29, 30.
Item information curves for the 31 items of the regret coping scale, presented by subscale. Numbers correspond to the item order in Table 2.
There were between 2.0% and 3.4% missing data per item (see Appendix 2 for item characteristics including missing values, percentages for the highest or lowest category and means and standard deviations). The reliabilities of the three dimensions were good, with all Cronbach's alpha coefficients close to 0.90 (PF=0.89; EFA=0.89; EFB=0.89). The correlations between the three dimensions were low (0.29 between PF and EFB strategies, −0.05 between PF and EFA strategies and −0.13 between EFB and EFA strategies).
Construct validity (Table 3)| Nurses | Physicians | |||
|---|---|---|---|---|
| coef | p | coef | p | |
| Determinants of problem-focused regret coping | ||||
| Sex (reference: male) | 0.27 | 0.008 | 0.20 | 0.01 |
| Number of regrets in the last month | 0.03 | 0.64 | −0.05 | 0.48 |
| Intensity of regrets in the last month | −0.02 | 0.77 | −0.15 | 0.02 |
| Intensity of most regretted situation (last 5 years) | −0.00 | 0.95 | −0.13 | 0.07 |
| Correlates of problem-focused regret coping | ||||
| Job satisfaction | 0.25 | <0.001 | 0.27 | <0.001 |
| Life satisfaction | 0.23 | <0.001 | 0.27 | <0.001 |
| Self-esteem | 0.20 | <0.01 | 0.17 | 0.01 |
| Self-reported health | −0.04 | 0.58 | −0.04 | 0.54 |
| Sleep | −0.12 | 0.07 | −0.14 | 0.04 |
| Depression | −0.08 | 0.24 | −0.07 | 0.29 |
| Determinants of emotion-focused regret coping A | ||||
| Sex | 0.16 | 0.10 | 0.29 | 0.003 |
| Number of regrets in the last month | 0.16 | 0.02 | 0.09 | 0.19 |
| Intensity of regrets in the last month | 0.12 | 0.06 | 0.30 | <0.001 |
| Intensity of most regretted situation (last 5 years) | 0.41 | <0.0001 | 0.53 | <0.0001 |
| Correlates of emotion-focused regret coping A | ||||
| Job satisfaction | −0.24 | <0.001 | −0.21 | <0.01 |
| Life satisfaction | −0.07 | 0.27 | −0.24 | <0.001 |
| Self-esteem | −0.27 | <0.001 | −0.44 | <0.001 |
| Self-reported health | 0.18 | 0.01 | 0.13 | 0.055 |
| Sleep | 0.26 | <0.001 | 0.43 | <0.001 |
| Depression | 0.31 | <0.001 | 0.30 | <0.001 |
| Determinants of emotion-focused regret coping B | ||||
| Sex | −0.00 | 0.98 | 0.01 | 0.90 |
| Number of regrets in the last month | 0.01 | 0.90 | 0.03 | 0.65 |
| Intensity of regrets in the last month | −0.13 | 0.051 | −0.06 | 0.38 |
| Intensity of most regretted situation (last 5 years) | −0.23 | 0.0004 | −0.17 | 0.01 |
| Correlates of emotion-focused regret coping B | ||||
| Job satisfaction | 0.18 | 0.01 | 0.15 | 0.03 |
| Life satisfaction | 0.17 | 0.01 | 0.19 | 0.01 |
| Self-esteem | 0.09 | 0.15 | 0.17 | 0.01 |
| Self-reported health | −0.04 | 0.50 | −0.17 | 0.01 |
| Sleep | −0.01 | 0.82 | −0.20 | 0.004 |
| Depression | −0.08 | 0.24 | −0.01 | 0.93 |
Coping strategies showed no association with years of professional experience (data not shown, all p-values>0.05). Within each profession, being a woman was associated with a more frequent use of problem-focused strategies, including seeking social support. Among physicians, being a woman was also related to a more frequent use of rumination and self-attacking (EFA) strategies. As expected, the number of regrets in the last month and the intensities of both recent regrets and the most important regrets were associated with a more frequent use of rumination and self-attacking strategies.
The use of problem-focused strategies, both among nurses and physicians, was positively correlated with life and job satisfaction, as well as with self-esteem, but was not associated with self-reported health and depression. It was also associated with fewer sleep problems among physicians, but not among nurses.
We also found that EFA strategies were related to lower job and life satisfaction and lower self-esteem, although the association between EFA (i.e., rumination and self-attacking) strategies and life satisfaction was not significant among physicians. Maladaptive strategies were also associated with lower health outcomes (including more sleep problems, higher depression score, and, for nurses, lower self-reported health). In contrast, the other emotion-focused (EFB) strategies were associated with higher life and job satisfaction but were not associated with depression. Among physicians, they were also associated with higher self-esteem, higher self-reported health and fewer sleep problems.
Test-retest reliabilityTest-retest agreement was satisfactory for all items of the RCS-HCP (weighted kappa: range 0.39–0.58, mean=0.49). For the 5-item subscales, test-retest agreement was 0.78 (95%CI: 0.71–0.84) for problem-focused strategies, 0.82 (95%CI: 0.76–0.87) for EFA strategies, and 0.79 (95%CI: 0.72–0.85) for EFB strategies. According to Bland-Altman plots (Fig. 2), the mean differences between the baseline test and retest were low and nonsignificant for all subscales, though the variances of the differences were relatively large. The differences between the baseline test and retest were correlated, so an increase in EFB strategies was associated with an increase in problem-focused strategies (r=0.20, p=0.008) and a decrease in EFA strategies (r=−0.23, p=0.002). Furthermore, of the four participants who had a large decrease in problem-focused strategies between the baseline survey and one-month follow-up, two had a corresponding large increase in rumination and self-attacking strategies and a decrease in other emotion-focused (EFB) strategies.
Bland-Altman plot of the regret coping scale (RCS-HCP), by subscale for the baseline and 1-month follow-up surveys.
PF, problem-focused regret coping; EFA, emotion-focused regret coping A; EFB, emotion-focused regret coping B.
In this study of health-care professionals (nurses and physicians), we developed and validated a scale of regret that can be used to assess the strategies used by professionals to cope with regrets following adverse or difficult situations. Using factor analysis, we found three types of strategies that can be adequately measured by 5 items per subscale: problem-focused strategies and two types of emotion-focused strategies, i.e., emotion-focused A strategies including rumination and self-attacking and emotion-focused B strategies including reappraisal, distanciation and acceptance. This finding is congruent with theories of emotion regulation10, 34). It can be argued that, while the best strategy depends on the situation, some strategies, such as rumination and self-blame, are more consistently associated with negative outcomes. These strategies have sometimes been called maladaptive13).
The internal consistency of each dimension was high (α>0.85), and so was the test-retest reliability (ICC2>0.75). Only a few respondents changed strategies between the baseline test and the follow-up survey, mostly substituting problem-focused strategies for rumination and self-attacking.
Most construct validity hypotheses were verified. With respect to the determinants of regret, women used problem-focused strategies more often than men, and this trend was driven by more frequent talking to colleagues and superiors, which is congruent with findings of other studies on emotion regulation35). Furthermore, female physicians also used rumination and self-attacking strategies more often than male physicians, a finding in line with a survey showing that female physicians report a higher tendency toward resignation36).
At least one of the measures of regret intensity was related to all strategies. The number and intensity of recent regrets were positively associated with the use of rumination and self-attacking, while no association was found with problem-focused or emotion-focused B strategies (e.g., acceptance). A likely explanation is that there may be a threshold for the frequency and intensity of regrets that one can efficiently cope with. Beyond this threshold, additional regrets might be dealt with in a maladaptive manner (i.e., self-blame/self-attacking or rumination). This interpretation of our findings is in line with previous research on the choice of emotion regulation37). In addition, emotion-focused strategies were more strongly related to the intensity of the most intense regret in the last five years than to the current regret intensity. While this is partly due to the fact that the regret intensity scale has less measurement error than the single questions, it may also indicate that the most intense regret has a stronger impact on regret coping strategies than less intense day-to-day events. Indeed, this finding echoes previous research showing that emotional responses to medical errors can linger for several years. 38)
As expected, problem-focused and emotion-focused strategies other than rumination and self-blame were both related to higher life and job satisfaction. Problem-focused strategies were also strongly associated with self-esteem, though this could be the result of reverse causation whereby individuals with higher self-esteem are more comfortable using problem-focused strategies. Among physicians, emotion-focused B strategies were also associated with higher self-esteem and better self-reported health, and use of these strategies also seemed to protect against sleep problems. This finding is in line with recent research on emotion regulation and polysomnography of sleep39). However, the expected negative associations of depression with problem-focused and emotion-focused B coping strategies were very small and nonsignificant. This may be due to the fact that depression is a more severe outcome than lower self-esteem or sleep problems. Furthermore, depression is influenced by many factors other than care-related regret coping strategies. However, the lack of association could also be the result of selection bias, since individuals with a high level of depression are much less likely to answer surveys.
Finally, the strongest associations with the outcomes were with rumination and self-attacking strategies. These strategies were associated with lower job satisfaction, lower self-esteem and more frequent sleep and depression problems, thus supporting the name of maladaptive strategies that has been used to describe them. These strategies were also associated with lower life satisfaction among physicians and lower self-reported health among nurses. Particularly among physicians, the associations of maladaptive strategies with self-esteem and sleep were relatively high.
An unexpected and interesting result was that profession and years of experience were not related to any of the strategies. This suggests that regret coping strategies are already quite established early on and do not evolve during health-care professionals' careers and exposure to care-related regret experiences. Rather than targeting specific professions or levels of training, training programs for regret coping strategies may need to target health-care professionals who are more prone to use rumination or self-attacking. Our questionnaire may help identify this higher risk group in future interventional studies, both among physicians and nurses and regardless of training level.
The main limitation of this study is the low response rate. The percentage of returned surveys (31.7%) was low and could have resulted in selection bias if the individuals who did not return the survey differed from those who did. This limitation was discussed more fully in a previous article22). In this study, the main argument against this limitation is that the structure of the scale and the construct validity associations are in line with what was expected based on previous findings in samples that did not suffer from a low response rate. An interesting aspect of this low response rate is that approximately 20% of the sample returned an empty envelope or indicated that they did not want to participate. This percentage of unwillingness to participate suggests that regret may be a sensitive topic. The response rate based on all returned surveys (even empty) would be 52.1%, a more usual response rate for mailed surveys among health-care professionals40).
Another important limitation of this study is that some strategies have to be assessed as they happen. In particular, reappraisal, which is described as one of the most effective emotion regulation strategies41), is not well evaluated in a retrospective self-report study. Indeed, this could lead health-care professionals to report that they have no regret, whereas they actually have some that have been reappraised, as has been shown in a study on managers8). A distinction between absence of regret and a complete reappraisal of initial regret into a “learning experience” becomes consequently impossible. Future research should measure feelings of regrets at very short intervals in order to capture regrets before the initial affective experience is transformed by reappraisal.
In addition to the regret intensity scale, the regret coping scale for health-care professionals will be invaluable in enabling very concrete and applied studies on the emotional difficulties of hospital personnel. An important step for this is to prospectively examine in cohort studies, whether regret intensity and regret coping are causally associated with the outcomes assessed in this study, such as sleep, and job satisfaction. An alternative and possibly more useful approach to examine causality would be to develop a training program to promote problem-focused and adaptive coping rather than maladaptive coping and to help health-care professionals recognize the latter and seek help. The effectiveness of such a program both on coping strategies and outcomes could then be tested in a randomized controlled trial. For both cohort studies and randomized controlled trials, a scale with a relatively low measurement error and good test-retest reliability is necessary to detect trends of regret coping style over time.
In conclusion, the regret coping scale (RCS-HCP) possesses sound psychometric properties for measurement of coping strategies of hospital-based physicians and nurses. Because it does not measure specific events but instead measures the overall regrets experience, it can be systematically used to assess regret coping strategies without the risk of changing scores when the respondent refers to a different event. Given that it is also short, it could be easily used to assess, on a regular basis, the impact of regret coping programs developed to help professionals deal with difficult experiences.
| In general, when I regret events related to patient care… | Missing (%) | % at lowest value | % at highest value | Mean (SD) |
|---|---|---|---|---|
| RCS-HCP | ||||
| Action-oriented regret regulation | ||||
| 1. I talk about it with colleagues, to be listened to or reassured. | 2.3 | 3.7 | 35.2 | 3.1 (0.84) |
| 2. I discuss the problem again with the patient (or his family) | 2.1 | 12.0 | 16.6 | 2.5 (0.91) |
| 3. I try to find concrete solutions to the situation | 3.0 | 0.7 | 46.8 | 3.4 (0.68) |
| 6. I talk with a supervisor to prevent these events from recurring | 3.0 | 12.8 | 23.1 | 2.7 (0.97) |
| 28. I expose the situation to colleagues to improve our practices | 1.9 | 3.0 | 31.3 | 3.0 (0.81) |
| Maladaptive regret regulation | ||||
| 11. I think I am no good | 3.0 | 48.8 | 3.3 | 1.7 (0.79) |
| 13. I turn these situations over in my mind all the time | 2.8 | 36.0 | 5.7 | 1.9 (0.86) |
| 16. I think about it so much that it becomes invasive | 2.3 | 65.7 | 3.1 | 1.5 (0.73) |
| 18. I have a tendency to blame myself | 2.8 | 27.9 | 7.9 | 2.1 (0.88) |
| 27. I think about these situations all the time | 3.2 | 22.7 | 5.1 | 2.1 (0.81) |
| Adaptive regret regulation | ||||
| 8. I try to accept the situation | 2.1 | 8.3 | 20.0 | 2.7 (0.87) |
| 19. I tell myself “to err is human” | 3.0 | 15.2 | 8.6 | 2.3 (0.83) |
| 21. I try to create some emotional distance | 2.6 | 2.2 | 25.6 | 3.0 (0.74) |
| 29. I try to see the positive side of things | 3.0 | 5.7 | 18.0 | 2.8 (0.81) |
| 30. I try to put the situation in perspective | 2.8 | 6.6 | 13.2 | 2.6 (0.8) |
| Additional 16 items | ||||
| Action-oriented regret regulation | ||||
| 9. I present my excuses to the patient (or his family) | 3.4 | 10.4 | 18.8 | 2.5 (0.91) |
| 14. I try to analyze the event | 2.3 | 1.1 | 44.5 | 3.3 (0.69) |
| 17. I change things so the situation does not happen again | 3.2 | 2.4 | 49.0 | 3.3 (0.76) |
| 22. I do not talk about it to anyone, friends or family | 3.4 | 74.0 | 2.7 | 1.4 (0.7) |
| 26. I expose the problem to my supervisor if he doesn't know about it | 3.0 | 11.4 | 28.6 | 2.7 (1.0) |
| 31. Since then, I avoid handling similar situations | 2.8 | 80.7 | 1.3 | 1.2 (0.56) |
| Maladaptive regret regulation | ||||
| 7. I tell myself I've betrayed the patient's trust | 2.6 | 51.4 | 3.3 | 1.6 (0.76) |
| 20. I feel responsible for what happened | 2.5 | 14.4 | 12.5 | 2.3 (0.88) |
| Adaptive regret regulation | ||||
| 4. I prefer not to think about it | 2.6 | 57.6 | 1.1 | 1.5 (0.65) |
| 5. I try to draw a lesson from it | 2.3 | 0.4 | 60.5 | 3.5 (0.62) |
| 10. I rationalize in order to avoid being overwhelmed by emotions | 2.8 | 10.3 | 16.9 | 2.6 (0.88) |
| 12. I talk to family or friends to get support | 2.3 | 22.9 | 19.9 | 2.4 (1.05) |
| 15. I tell myself I (we) have done the best I (we) can | 3.0 | 2.6 | 19.0 | 2.9 (0.73) |
| 23. I tell myself that perfection doesn't exist | 3.6 | 21.0 | 13.3 | 2.4 (0.96) |
| 24. I try to evacuate any thoughts about it | 3.6 | 22.8 | 6.6 | 2.2 (0.86) |
| 25. I tell myself I can get better | 2.8 | 0.9 | 39.0 | 3.2 (0.7) |
| In general, when I regret events or situations with patients… | Factor 1 | Factor 2 | Factor 3 |
|---|---|---|---|
| Problem-focused regret coping | |||
| 1. I talk about it with colleagues, to be listened to or reassured. | 0.69 | ||
| 2. I discuss the problem again with the patient (or his family) | 0.52 | ||
| 3. I try to find concrete solutions to the situation | 0.41 | ||
| 6. I talk with a supervisor to prevent these events from recurring | 0.56 | ||
| 28. I expose the situation to colleagues to improve our practices | 0.78 | ||
| Maladaptive regret coping | |||
| 11. I think I am no good | 0.56 | ||
| 13. I turn these situations over in my mind all the time | 0.83 | ||
| 16. I think about it so much that it becomes invasive | 0.77 | ||
| 18. I have a tendency to blame myself | 0.66 | ||
| 27. I think about these situations all the time | 0.72 | ||
| Adaptive regret coping | |||
| 8. I try to accept the situation | 0.51 | ||
| 19. I tell myself that “to err is human” | 0.47 | ||
| 21. I try to create some emotional distance | 0.52 | ||
| 29. I try to see the positive side of things | 0.55 | ||
| 30. I try to put the situation in perspective | 0.80 |
