Evaluation of burnout among hospital and health-system pharmacists in North Carolina

In his inaugural address as president of the American Society of Health-System Pharmacists in 2017, Dr. Paul Bush emphasized the importance of promoting a resilient work environment and a healthy work-life balance.¹ He commented on how workforce stress and burnout serve as internal organizational threats for hospitals and health systems. Research suggests that employees who are burned out experience issues with job turnover, absenteeism, poor job satisfaction, increased malpractice lawsuits, earlier retirements, decreased productivity, and low morale.^2-4 Additionally, this can affect patients and cause safety issues and increased medical error^1,2,5

Burnout was first described in the mid-1970s by Herbert Freudenberger, an American psychologist. He observed that many young and motivated volunteers who worked with recovering drug addicts would experience a gradual loss of energy and commitment. Freudenberger would label what he was observing as “burnout,” and he would emphasize that this was a culmination of effects from the professional responsibilities and work environment of that volunteer role.²

Around the same time as Freudenberger, Christina Maslach, a social psychology researcher, would use the term burnout to describe the effects she was seeing among healthcare providers. Healthcare providers are in roles that often require interacting with patients who experience feelings such as anger, embarrassment, fear, or despair. These interactions would lead the healthcare providers to “a syndrome of emotional exhaustion and cynicism that occurs frequently among individuals who do people-work of some kind.” ⁶ Maslach introduced a survey called the Maslach Burnout Inventory (MBI) that serves as the gold standard for assessing burnout among individuals and organizations.²

To date, the large number of published articles assessing burnout in healthcare providers has been centered on physicians and nurses. The Agency for Healthcare Research and Quality estimates that burnout may affect anywhere from 10% to 70% of nurses and 30% to 50% of physicians and midlevel practitioners.⁷ The prevalence of burnout for US physicians has also been increasing, going up by 10% between 2011 and 2014.⁸

There is minimal published research regarding the incidence of burnout among pharmacists, especially in hospital or health-system settings. Much of what is published has been researched in other countries and describes burnout among community pharmacists. A Turkish study of community pharmacists found that a large percentage of respondents had low feelings of personal accomplishment, 1 of the 3 subscales of the MBI.⁹ A similar study of community pharmacists and pharmacy technicians in France found that 56.2% of respondents were at risk for burnout. Furthermore, 10.5% of the pharmacists were considered to have severe burnout.¹⁰ Another study in Japan focused specifically on hospital pharmacists and found that 49.2% of respondents were at high risk for burnout.¹¹

The minimal research that has been done among pharmacists who work in hospital or health-system settings in the United States found the burnout rate to be 61.2%, one of the highest levels among any medical specialty.³ This study was specifically targeted at clinical pharmacists who were members of the American College of Clinical Pharmacy. Certain factors such as age, too many nonclinical duties, and underappreciated contributions were revealed to be predictors of burnout. A separate study that targeted pharmacists who primarily work in university hospital or health-system settings found that 53.2% of pharmacists reported a high degree of burnout.¹² Burnout was seen from institutions of varying sizes and across service areas. Lower years of experience and absence of resources were also associated with increased burnout. Both studies used the MBI.

The results presented in those two studies revealed alarming rates of burnout, and additional research is needed to confirm these results in a diverse population of pharmacists. The purpose of this study was to assess the current state of burnout among hospital and health-system pharmacists by surveying all licensed individuals in the state of North Carolina. North Carolina has a wide range of hospitals and health-system settings with pharmacists serving in a variety of practice roles. These settings range from small, rural, critical access hospitals with less than 100 beds to community teaching hospitals with 500 beds to academic medical centers with more than 1,000 beds. By surveying all licensed hospital and health-system pharmacists in North Carolina, this study helps capture a diverse population of pharmacist employees.

Methods

The cross-sectional study was approved by the institutional review board at Cone Health in Greensboro, North Carolina. The MBI-Human Services Survey for Medical Professionals (MBI-HSS [MP]) was used to assess burnout among pharmacists who work in hospital or health-system settings in North Carolina. The MBI surveys are owned by Mind Garden (Menlo Park, CA), and a license to reproduce was obtained to distribute the specific survey. The MBI is a 22-item questionnaire, with responses measured on a 7-point Likert scale (0-6) that assesses 3 subscales that contribute to burnout: emotional exhaustion (EE), depersonalization (DP), and personal accomplishment (PA). EE describes the feeling of being emotionally overextended and exhausted. DP describes a lack of feeling and an impersonal response toward patients. PA describes the feeling of competence and achievement. Higher scores on EE and DP and lower scores on PA represent higher levels of burnout. Of the 22 questions, 9 questions relate to EE (range of scores, 0-54), 5 questions relate to DP (range of scores, 0-30), and 8 questions relate to PA (range of scores, 0-48). Reliability and validity of the MBI have been confirmed through multiple studies.¹³ Of note, after the survey was released, it was discovered that one of the questions for the PA subscale was mistakenly left off. As such, 3 points were added to the calculated sample mean for PA to represent the median of the Likert scale.

In addition to the MBI, other questions addressing various modifiable and nonmodifiable demographic factors were asked: age, gender, race and ethnicity, marital status, number of dependents, practice setting, employment status, education level, years practiced, years practiced in current role, awareness of burnout resources, use of burnout resources, postpharmacy school training, additional advanced degrees, board of pharmacy specialties certification, practice role, hours worked per week, months with learners, and presence of a formalized career ladder.

To distribute the survey, an email listserv of all hospital and health-system pharmacists who were licensed with the North Carolina Board of Pharmacy (NCBOP) was obtained. The anonymous survey was distributed via an internet link through the listserv in June 2018. A follow-up email encouraging participation was sent 2 weeks later. The survey was open for a total of 4 weeks. Data was collected through the SurveyMonkey platform (SurveyMonkey, Greensboro, NC).

Only results with completed MBI responses were analyzed, because it would be difficult to assess burnout based on the MBI if questions were unanswered. Answers for modifiable and nonmodifiable demographic questions outside the MBI were not required.

The primary objective was to determine the percentage of hospital and health-system pharmacists in North Carolina who are at risk for burnout. Similar to previous studies, risk for burnout was defined as having EE scores no less than 27 or DP scores no less than 10.^3,14,15 Low PA scores were analyzed but not considered in the definition, because certain studies show lower Cronbach alpha with this subscale’s reliability in predicting burnout.¹⁶

The data from the MBI were reported using Stata 15.1 (StatCorp LLC, College Station, TX). Descriptive statistics showing mean and median were reported to describe the overall levels of burnout among pharmacists. Univariate logistic regression was run for each demographic factor (any question that was not part of the MBI) as a potential predictor variable with burnout as the dependent variable. Potential predictors with a P value less than 0.2 in the univariate analysis were included in a preliminary multiple logistic regression model, and a sequential backward selection process was then applied. Demographic questions that had a “not available” response greater than 10% were excluded from this multivariate analysis because it would greatly reduce the total samples in the model. Only predictors with a joint P value less than 0.05 across variable levels were considered independent statistically predictors of burnout and were retained in the final multiple logistic regression model.

Results

The MBI was delivered to 2,524 hospital and health-system pharmacists licensed with the NCBOP. A total of 380 responses were received (15.1% response rate). Of the 380 individuals who responded, 357 completed the entire survey (93.9% completion rate). Modifiable and nonmodifiable demographic characteristics are presented in Table 1. The majority of respondents were younger than 45 years (60.8%), female (70.0%), white (90.5%), and married (73.1%); had no dependents (52.7%); and worked in a community hospital setting (51.8%) and worked full-time (92.7%). Additionally, most respondents had no post–pharmacy school training (52.7%), no other advanced degrees (84.0%), and no certification through board of pharmacy specialties (58.3%). The most common response for years of practice was 6-15 years (34.5%), but the most common response for years of practice in current role was no more than 5 years (40.6%). The main practice role for the respondents was clinical (60.5%), followed by distribution (16.8%) and administration (12.3%). The majority of individuals were not aware of any burnout resources at their institution (56.6%), and a large majority had not used any burnout resources in the last 12 months (95.8%).

A breakdown of each MBI subscale with the corresponding number of pharmacists who responded as having a low, moderate, or high score is presented in Table 2. EE represented the largest concern for burnout among the 3 subscales because 49.6% of respondents scored high.

Based on the previously mentioned definitions for burnout, 198 pharmacists (55.5%) were considered at risk for burnout. This data was further analyzed to determine if any of the modifiable or nonmodifiable risk factors significantly contributed to burnout. After initially including all factors with a P less than 0.2 from the univariate model and then using backward selection to determine the final multivariate model, 3 factors were statistically associated with increased burnout among pharmacists: female gender (P = 0.014), working in a primarily distribution role (P = 0.012), and longer hours worked per week. Two factors were significantly associated with decreased risk of burnout: awareness of burnout resources (P = 0.012) and the number of months with learners (Table 3). For hours worked per week, it was specifically noted that working 50-59 hours per week led to significantly higher risk for burnout when compared with working 40-49 hours per week (P = 0.001). Working more than 60 hours trended in the same direction, but the result did not reach statistical significance (P = 0.052). On the other hand, working less than 40 hours trended in the opposite direction. For months with learners, having 4-6 months with learners showed a significantly lower risk of burnout when compared to having 0-3 months with learners (P = 0.006). No significant differences were seen when comparing 7-9 months or 10-12 months of learners with 0-3 months of learners.

Discussion

This study used the validated MBI-HSS (MP) to assess the current state of burnout among hospital and health-system pharmacists across the state of North Carolina. The results of this project show that more than half of pharmacists in North Carolina are at risk of experiencing burnout. The 55.5% of respondents was comparable to numbers seen among pharmacists in previously published articles.^3,12 The response rate of 15.1% is also comparable to the previously aforementioned study.³ This reproducible elevated risk of burnout among hospital and health-system pharmacists is concerning.

A large part of the burnout was driven by high levels of EE, which was seen in 49.6% of pharmacists. Fewer respondents (35.0%) were experiencing high levels of DP, the other qualifying factor for burnout. Compared with previous studies, the percentage of pharmacists in North Carolina who have high levels of EE and DP and have low levels of PA are similar.^3,12

As for modifiable and nonmodifiable factors that may be associated with burnout among pharmacists, 5 factors of significance were seen in the multivariate regression analysis: female gender, working in a primarily distribution role, longer hours worked per week, lack of awareness of burnout resources, and the number of months with learners.

Previous studies have mostly failed to find gender differences between burnout rates, but one study found that female medical residents may experience more EE than their male counterparts.^9,17,18,19 It is unclear if the similar finding of female gender being associated with higher burnout in this study is a chance finding or something worth exploring further.

In regard to pharmacy practice, some previous research has shown that employees who work in primarily distributive roles have higher levels of job dissatisfaction and burnout.^20,21 This theme was also seen in the results of this survey because pharmacists in distribution roles had statistically higher levels of burnout when compared with pharmacists who worked primarily in clinical roles. One theory for this is that job variety may decrease burnout, and variety is more commonly associated with more clinical roles.^22,23

Workload is another issue that has been described in previous literature as possibly contributing to burnout.^2,5,9,24 This concept makes sense intuitively and was supported again by the results of this project because pharmacists who work 50-59 hours were significantly more likely to have burnout compared with pharmacists who worked 40-49 hours. Of note, working no less than 60 hours per week was not significant when compared with 40-49 hours. This is probably due to the lack of statistical power associated with the low number of respondents who worked this many hours.

Regarding burnout resources, one interesting point is that although 28.0% of pharmacists were aware of burnout resources at their institution, only 3.6% of pharmacists actually used any resources in the last 12 months. Still, even just the awareness of burnout resources showed a significant difference in burnout compared with those who were not aware. It is possible that this could be due to the culture of an organization. Organizations that have burnout resources and promote them to employees could be more likely to promote a general environment of well-being and resiliency. Thus, employees at these organizations may have reduced levels of burnout regardless whether they directly use any of the resources.

For the number of months with learners, it seemed that having 4-6 months of learners each year offered a protective effect against burnout when compared with 0-3 months of learners. No difference was seen for those with 7-9 or 10-12 months of learners. The lack of any significant difference in the respondents who have 7-9 or 10-12 months of learners could be due to the low number of respondents for that group. For the purpose of this survey, learners were described as either pharmacy students or residents. This was an interesting result because learners often require extra time and work from preceptors. On the other hand, learners who are prepared and engaged may remove some of the workload from the preceptor. Learners may also be able to provide some extra energy and excitement because each rotation is a new opportunity for them. Future studies would need to confirm this result.

There are a few limitations to consider when evaluating the results of this study. First, because the survey is being sent through the NCBOP database, it assumes the database is accurate and up to date. Some pharmacists may have changed jobs or moved to work in a different state. However, despite this limitation, the NCBOP database is still the most complete list available for this study population. A second limitation is in the phrasing of the MBI-HSS (MP). Some of the questions are geared toward direct patient care interactions, and not all pharmacists may interact with patients regularly. Third, as previously mentioned, one of the questions for PA was left off. As such, that subscale was based off of 7 questions instead of 8. The question that was not included was, “I feel exhilarated after working closely with my patients.” This error does not affect the definition of burnout because only EE or DP were used to classify whether burnout risk was present. However, it does limit the ability to assess the PA scores to the full extent. Although 3 points (the average of the 0-6 Likert scale) were added to each respondents’ score to partially correct for this error, this does not fully represent exactly what each individual may have answered. Finally, the response rate of this survey was on the lower end at 15.1%. Response bias may have played a role given that individuals who feel burned out may be more likely to respond to a survey addressing burnout.

Although the primary purpose of this project was to assess levels of burnout among pharmacists in North Carolina, it is important to recognize that the ultimate goal of future research will be to prevent and reduce burnout. One study suggested that addressing burnout needs to start early, and there is concern that pharmacy schools tend to overtrain technical components while undertraining social, cultural, and political realities.²⁵ By providing a more holistic and realistic training while in pharmacy school, one potential cause of burnout can be reduced: the gap between expectations and realities of the job.²⁶ Furthermore, hospitals and health systems will need to continue to support individuals experiencing burnout by providing open avenues of discussion and necessary resources. Other organizational and individual solutions that have been offered for preventing and reducing burnout include avoiding task redundancy, limiting professional work hours, sharing experiences between employees, decreasing workload, and learning cognitive-behavioral and relaxation techniques.^2,24,27

Conclusion

The results of this statewide survey revealed that more than half of hospital and health-system pharmacists are at risk of burnout. Certain factors such as female gender, working in a primarily distribution role, and longer hours worked per week may be associated with higher levels of burnout, whereas awareness of burnout resources and the number of months with learners may be associated with lower levels of burnout.

Acknowledgments

We appreciate the Eckel Fund at the University of North Carolina (UNC) Eshelman School of Pharmacy for its financial support of this research project.

Disclosures

Funding was provided by the Eckel Fund at the UNC Eshelman School of Pharmacy to purchase the license to use the Maslach Burnout Inventory. The authors have declared no potential conflicts of interest.

References