Moral distress and its contribution to the development of burnout syndrome among critical care providers

Setting

This survey was conducted in the tertiary, private, teaching, 470-bed Hospital Sírio-Libanês, in São Paulo, Brazil. The institutional review board (IRB), called “Comitê de Ética em Pesquisa da Sociedade Beneficiente de Senhoras do Hospital Sírio-Libanês,” reviewed and approved this study (HSL protocol number HSL 2015-65, 07/14/2015).

ICU providers (physicians, nurses, nurse technicians and respiratory therapists) and also the SDU providers working for more than 6 months in the unit were invited to participate and all signed a written informed consent. The intensive care unit (ICU) comprises 22 private rooms for general medical–surgical care and eight private rooms for neurological care. The professional-to-bed ratios in the ICU is: nurse 1:4; nurse technicians 1:2; physician 1:6 (day shift) and 1:10 (night shift). The ICU has a 24-h visitation policy (day or night free entry, with possibility to change the visitor at any time, and option to sleep with the patient in an individual box).

The step-down unit (SDU) provides an intermediate level of care between the intensive care units (ICUs) and the general medical–surgical wards. In our hospital, it consists in a 31-bed unit with monitoring and presence of a physician belonging to the same team of intensivists (ratio of physician: patient 1:8 mornings and 1:15 during in the rest of the day; nurse 1:6 and nurse technicians 1:3).

Interviews

For data collection, a self-administered questionnaire for each critical care provider was used including basic demographic data and two more instruments: (1) the Maslach Burnout Inventory (MBI) to assess burnout and (2) Moral Distress Scale-Revised (MDS-R) to investigate moral distress. All workers who agreed to participate in this study returned the questionnaire in a sealed envelope.

The MBI is a 22-item questionnaire that has shown to be reproducible and valid [15]. The MBI evaluates three domains of burnout: The emotional exhaustion subscale (nine items) assesses feelings of being emotionally overextended and exhausted by work; the depersonalization subscale (five items) measures an unfeeling and impersonal response toward recipients of one’s service, care or treatment; the personal accomplishment subscale (eight items) assesses feelings of competence and successful achievement in one’s work with people.

Cutoff scores were defined for each dimension, and we adopted the following internationally established definition of burnout, according to the MBI manual: high levels of emotional exhaustion (score ≥ 27 points) and depersonalization (score ≥ 10 points) combined with low personal accomplishment (score ≤ 33 points) [16].

We also used the Standard Hamric Moral Distress Scale-Revised (MDS-R) [17]. We previously obtained the author’s authorization for use of the MDS-R. Prior to use, the scale was translated into Portuguese using the forward–backward method and then matched. All questions were translated into Portuguese and validated before being applied [18]. The instrument was translated by two bilingual experts, and then, a back translation was made by other two translators who had not seen the original scale, in order to verify the equivalence of terms between the two versions. After translation and adaptation, the authors ensured that the new version included the evaluation properties required for application.

The Moral Distress Scale-Revised (MDS-R) is a questionnaire that measures moral distress in specific situations: Respondents are asked to indicate both the frequency (“F”) and the level of disturbance (intensity = “I”) when the situation arises.

The items of MDS-R were categorized according to Likert scale from zero (never) to 4 (often) to assess the frequency of moral distress and from zero (none) to 4 (largely) to assess the intensity of moral distress, which can range from 0 to 16, where the items that are less distressing have low “f” × “i” scores versus more distressing items, which have higher “f” × “i” scores. First, the product of frequency and intensity is obtained: [“F” × “I”] means F (frequency) multiplied by I (intensity) and the final value = [(frequencies score) × (intensity score)] is obtained, allowing to identify individual items or situations that are distressing. Then, the Likert scale data can be computed into a composite score or actual moral distress using a two-part procedure. The final composite is obtained by summing each item’s [“F” × “I”] score, resulting in a range of 0–336, where less actual distress has low composite score and more actual moral distress yields higher composite score [17].

Statistical analysis

Descriptive analysis was performed using mean and standard deviation (or median and interquartile range for nonparametric variables) for quantitative variables and frequencies for categorical variables. We compared staff characteristics according to workplace using the Chi-square test, Student’s t test and Wilcoxon rank-sum test, when appropriated. We calculate prevalence and its respective 95% confidence interval (95% CI) for severe burnout among critical care providers. Prevalence ratios (PR), their respective 95% CI and Chi-square test were used to compare severe burnout prevalence according to subjects’ characteristics.

Moral distress scores among professional categories were compared using one-way ANOVA; to compare these scores between workplaces, we performed t test. We used Pearson’s correlation coefficient to examine the correlation between MBI domains of burnout and MDS-R overall score. Comparisons between moral distress individual questions answers according to the presence or absence of severe burnout were made using Wilcoxon rank-sum test. We performed a logistic model step by step and analyses by Hosmer–Lemeshow test. Variables with p value <0.20 in univariate analysis were tested in the final model. A p value < 0.05 was considered significant. Statistical analysis was performed using the Stata^® 13.1 software (Stata Corp LP, College Station, TX, USA).

Participants

A total of 283 of 389 (72.7%) critical care providers agreed to participate, and the survey was conducted in August and September 2015. Out of this total, 134 were from ICU and 116 from a step-down unit. Regarding physicians, the same team of intensivists attended both ICU and step-down unit and they were treated as an individual group. Thirty-three of them (67.3%) participated, representing 11.7% of the total respondents. Nurses 35/134 (26.1%) were from ICU and 28/116 (24.1) from SDU, nurse technicians 68/134 (50.7%) from ICU and 60/116 (51.75%) from SDU and respiratory therapists 30/134 (22.4%) from ICU and 27/116 (23.3%) from SDU.

Physicians characteristics

Regarding physicians, 66.6% were male; mean age was 38.71 ± 6.65 years; all of them have a personal income of more than $2.350; none spent more than 1 h on the way to work; 66.6% were working night shift and 63.6% had more than one job. They had 8.33 ± 7.06 years of work in the unit. We observed that the majority (84.8%) had more than 10 h per week of leisure; 75.7% had sexual activities more than once a week; 63.6% did regular physical activity; 48.4% had a regular hobby; and 12.1% were doing psychotherapy at the time of interview. Regarding religion, 51.5% were Catholics followed by atheists (21.2%) and Spiritualist religion (15.1%).

We observed that the categories of respiratory therapists (64%) and physicians (64%) have more than one job, significantly higher than the nurses (14%) and nurse technicians (24%), (p < 0.0001).

We did not find a difference in burnout prevalence in the healthcare providers who worked night shifts, compared with others 24.3% (n = 33) versus 20.4% (n = 28), p = 0.448. Regarding the number of working hours, due to local rules, within each category, professionals have similar workloads.

Prevalence of severe burnout among professional categories

Three respondents did not complete the MBI. Overall, the prevalence of severe burnout was 22.5% (95% CI 17.7–27.8%). Severe burnout was identified in 18.2% of physicians. Considering all other critical care providers (ICU + SDU), we identified that 23.1% (95% CI 18.0–28.8%) presented severe burnout. There was no statistical difference between the units: From ICU respondents, 29/131 (22.1%) had severe burnout and 27/114 (23.7%) from SDU (p = 0.774).

Additional file 1: Figure S1 shows the prevalence of severe burnout in all respondents, according to the professional category. The differences found in the burnout rate in the SDU were nurse = 39.3%, nurse technicians = 15.2% and respiratory therapists = 25.9% with p = 0.046 and in the ICU were nurse = 28.6%, nurse technicians = 20.9% and respiratory therapists = 16.6% with p = 0.490. We did not find any differences in prevalence of severe burnout between workplace when comparing nurses (p = 0.370), nurse technicians (p = 0.391) or respiratory therapists (p = 0.392). Although the frequency of burnout in ICU + SDU (Additional file 1: Figure S1) was higher in nurses (33.8%), no significant difference was found between the different categories (p = 0.093).

The score for emotional exhaustion ranged from 9 to 45 (mean = 26.4; SD = 8,1); for depersonalization, the score ranged from 5 to 22 (mean = 9.8; SD = 3.9) and for personal accomplishment from 21 to 40 (mean = 33.1; SD = 4.5). Whereas depersonalization was similar among all professionals, we observed high rates of emotional exhaustion in MBI subscale much more often among nurses (60%) compared with respiratory therapists (50.8%), nurse technicians (43.6%) and physicians (27.3%), p = 0.015. We also more frequently identified in the nurses category lower level of personal accomplishment (61.5%) compared with others professionals (p = 0.018) (Additional file 2: Figure S2).

Moral distress

Exploratory factor analyses were performed by Kaiser–Meyer–Olkin measure of sampling adequacy (0.852) and Bartlett’s test of sphericity (p < 0.0001). Also, the Cronbach’s alpha was calculated and presented a good performance value 0.897.

A total of 68 questionnaires were incomplete and excluded from analysis. The category with more incomplete questionnaire was nurse technicians (33.6%) followed by nurses (21.5%), respiratory therapists (12.3%) and physicians (12.1%), p = 0.004.

The correlation between burnout and moral distress

Several questions from the MDS-R questionnaire were significantly associated with severe burnout, as listed in Table 3. MDS-R score was moderately correlated with emotional exhaustion (r = 0.43; p < 0.001) (Fig. 1). There were an inverse weak correlation between MDS-R score and personal accomplishment (r = −0.37; p < 0.001) and also a very weak correlation between MDS-R score and depersonalization (r = 0.25; p < 0.001).

Moral distress questions	No burnout Median [IRa]	Severe burnout Median [IR]	p*
Provide less than optimal care due to pressures from administrators or insurers to reduce costs	0 [0–2]	1 [0–6]	0.009
Initiate extensive lifesaving actions when I think they only prolong death	6 [2–9]	7 [4–16]	0.005
Follow the family’s request not to discuss death with a dying patient who asks about dying	4 [1–9]	8 [4–12]	0.005
Continue to participate in care for a hopelessly ill person who is being sustained on a ventilator, when no one will make a decision to withdraw support	8 [4–12]	12 [7–16]	0.006
Avoid taking action when I learn that a physician or nurse colleague has made a medical error and does not report it	1 [0–4]	2 [1–4]	0.017
Assist a physician who, in my opinion, is providing incompetent care	4 [1–8]	4 [2–9]	0.013
Be required to care for patients I do not feel qualified to care for	0 [0–2]	1 [0–4]	0.041
Provide care that does not relieve the patient’s suffering because the physician fears that increasing the dose of pain medication will cause death	2 [0–8]	6 [2–12]	<0.001
Follow the family’s wishes for the patient’s care when I do not agree with them, but do so because of fears of a lawsuit	2 [0–6]	7 [1.5–12]	<0.001
Work with nurses or other healthcare providers who are not as competent as the patient care requires	4 [2–9]	6 [3–12]	0.035
Witness diminished patient care quality due to poor team communication	4 [2–9]	9 [3–16]	0.004

* Wilcoxon rank-sum test

^aInterquartile range

According to ROC curve analysis, we observed that the MDS-R score was equal or higher than 100 points had a sensitivity of 68% and specificity of 57%, with a relative prevalence of 2.0, 95% CI 1.16–3.44, p = 0.010. Therefore, there is a fairly good association between the two instruments.

The variable moral distress is an independent predictor for severe burnout with adjusted odds ratio 2.8 (1.5–5.2) with good performance for validation model adjusted Hosmer–Lemeshow test (p = 0.998). No other variables showed association with severe burnout at univariate analysis (Additional file 3: Table S1). See the MDS-R Scale in Additional file 4.