Loneliness in healthcare providers: Results from a mid-pandemic survey

Abstract

Background

Provider loneliness in healthcare can have devastating consequences. The causes of loneliness are not well known and workplace changes during the COVID-19 pandemic likely cause fewer opportunities for socialization.

Objective

We sought to explore the relationships between healthcare worker loneliness and isolation among various specialties and work environments. We evaluated factors that may contribute to increased loneliness/isolation, such as job satisfaction, time with peers, and work-related changes during COVID-19.

Design

An email containing an 18-item survey was sent out to Hartford Hospital’s medical staff.

Setting

Hartford Hospital

Participants

Hartford Hospital’s medical staff including physicians, psychologists, physician assistants, and APRNs

Measurements

Our survey included demographic information, questions regarding frequency of meaningful interactions with colleagues, isolation behaviors, as well as the UCLA 3-Item Loneliness Scale, Patient Health Questionnaire-2, Single-Item Burnout Scale, and Single-Item Measure of Job Satisfaction

Results

Of 1,015 respondents, almost half (48%) reported feeling lonely. Staff in procedure areas had significantly higher loneliness scores, while hospital-based floor/unit staff reported the lowest scores. Isolation was attributed to exhaustion from clinical work (36.1%), being too busy (39.6%), and being cautious due to COVID-19 (39.2%). Those who reported burnout, depression, and having few meaningful interactions all had increased loneliness scores. Those with higher job satisfaction reported lower loneliness.

Limitations

While our study had a 51% response rate, this rate is equivalent or higher than the response rate the organization receives to obligatory engagement surveys. It is likely that the ongoing pandemic affected these results.

Conclusion

We found significant levels of loneliness among our medical staff that appeared strongly related to work setting and environment. Social isolation behaviors were associated with higher loneliness scores. Future research should examine the efficacy of tailored strategies/interventions to reduce loneliness.

Primary Funding Source

None

Introduction

Physicians experience one of the highest rates of loneliness among all professionals, ¹ despite regularly being in the presence of others. The terms “loneliness” and “social isolation” are conceptually linked and often studied together. Social isolation refers to a lack of social connections, which can be objectively quantified (e.g., living alone). Loneliness, however, is a state of subjective emotional distress wherein individuals lack meaningful relationships regardless of their social network.^2-5

Research has extensively linked social isolation and loneliness to a variety of chronic health conditions, worse mental health outcomes, and even dementia.^3-6 It has been reported that loneliness can result in a number of different psychiatric and physical conditions, including depression, dementia, suicide, hypertension, heart diesease, diabetes, burnout and reduced longevity.^1,3,6,7 A widely cited 2010 meta-analysis suggests that the health effects of loneliness and social isolation are worse than smoking 15 cigarettes per day. ⁸

Loneliness is particularly concerning in healthcare, where pressure is high and the impact of poor provider health can have devastating consequences. Cigna’s 2020 report suggests that loneliness has professional implications including lower productivity, lower work quality, more missed work and higher turnover. ⁹ It is likely that a myriad of forces, including workplace conditions, culture and personal factors contribute to loneliness and social isolation in healthcare settings, especially during and after the COVID-19 pandemic. However, what is not often discussed is the issue of workplace loneliness and isolation, particularly among physicians and other healthcare workers.

It is increasingly evident that both loneliness and isolation play significant roles in physician burnout and work stress.^4,10-12 Concerningly, a recent study found that about one-quarter of physicians regularly feel isolated in their professional lives, ¹ while up to 43% experience loneliness. ³ Unfortunately, the rates of loneliness, stress, and burnout have been exacerbated during the COVID-19 pandemic. Concurrently, a recent survey found that about two-thirds of physicians reported increased burnout during the past year. ¹³

The causes for feeling loneliness are not well understood, and unfortunately, workplace changes during the COVID-19 pandemic (e.g., working from home, increased use of telehealth) are likely leading to fewer opportunities for meaningful socialization with colleagues and family members. This study sought to explore the relationships between physician loneliness and isolation among various specialties and work environments, as well as the perceived levels of social interaction associated with those work environments. We also aimed to evaluate possible factors that may contribute to increased loneliness/isolation in the workplace, such as job (dis)satisfaction, not enough time with peers, and work-related changes during COVID-19 (e.g., increased use of telehealth, less face-to-face interaction).

Methods

After gaining Hartford HealthCare Institutional Review Board approval (HHC-2021-0043), an email containing a link to an 18-item survey was sent to all members of Hartford Hospital’s medical staff (who work within various work environments such as inpatient, outpatient, and office). The medical staff includes physicians, psychologists, physician assistants, and APRNs. Data were collected from February 26, 2021 through/including April 30, 2021. Of 1990 medical staff members who were sent the email, 1015 of them completed the survey (a 51% response rate). This response rate to our voluntary survey is equivalent or higher than the response rate the organization receives to obligatory engagement surveys.

The survey’s content included items concerning demographic information, loneliness, factors contributing to isolation (e.g., increased work from home, increased use of telehealth, not enough time with peers, too busy to socialize, organizational culture does not allow it, exhaustion), depression, burnout, job satisfaction, and frequency of meaningful interactions with colleagues.

Loneliness was assessed using the UCLA 3-Item Loneliness Scale. ¹⁴ Respondents recorded their feelings of loneliness on a 3-point Likert scale corresponding to scores of 1 to 3 (hardly ever; some of the time; often; resp.) for each item. Those with a total score of 6 or above were considered to be lonely. The Patient Health Questionnaire-2 (PHQ-2) ¹⁵ was used to screen for depression. Scores ranged from zero to 6, with a probable depression cut-point score of 3. Burnout was measured via the Single-Item Burnout Scale, ¹⁶ and responses were measured on a 7-point Likert scale ranging from zero “Never” to 6 “Every Day.” Burnout was defined as a score equal to or greater than 3, indicating at least 1 symptom of burnout. A measure of overall job satisfaction was obtained using the Single-Item Measure of Job Satisfaction. ¹⁷ This was rated on a seven-point Likert scale (1 = extremely dissatisfied, 7 = extremely satisfied).

The respondents were separated into 4 groups based on the clinical work environment/locations and the perceived extent of social interaction in each of the work areas: hospital-based floor/unit, outpatient office/clinic, procedural area, and other (eg, non-procedure areas).

Statistics

Descriptive statistics were used to summarize the prevalence of loneliness, isolation, burnout, and depression among all respondents. Spearman rank correlation coefficient was used to examine the relationship between loneliness and each of burnout, depression and job satisfaction. Differences in continuous variables were evaluated between 2 groups with a Student’s t test (if normally distributed) or a Mann-Whitney U test (if non-parametric). For 3 or more groups, an analysis of variance (ANOVA) with a post hoc Scheffé test or Kruskal-Wallis (K-W) H test was used. Categorical variables were compared with a Pearson chi square test. All evaluations used an alpha level of .05 such that results yielding P < .05 were deemed statistically significant.

Role of the funding source

This study was supported by Hartford HealthCare via in-kind funding support.

Results

Of the respondents, 56.9% were female, 85.6% identified as white, and about half (49.8%) were physicians (MD, DO, MBBS). On average, respondents were 16 years post-training, with an average age of 47 years. Complete demographic data are shown in Appendix Table 1.

About half of respondents (48%) reported feeling lonely, and most respondents reported “often” or “sometimes” feeling a lack of companionship (60%), left out (57%), or isolated from others (64%) (Appendix Figure 1). Females (P = .018) and individuals who identify as African-American (P = .031) reported significantly greater degrees of loneliness. The prevalence of loneliness did not change significantly with age (P = .269). Those who work in procedure areas had significantly higher loneliness scores, while hospital-based floor/unit staff reported the lowest scores for loneliness (ANOVA P < .001, Appendix Table 2).

Regarding social isolation, 28.4% reported that they have meaningful work-related interactions with colleagues “rarely” or “once weekly”, and 43.1% have non-work-related conversations (e.g., talking about their weekend) at work once weekly or less. The majority reported feeling professionally isolated due to exhaustion from clinical work (36.1%), being too busy (39.6%), and/or being cautious due to COVID-19 (39.2%; Appendix Figure 2). Behavioral changes due to COVID-19 also contributed to increased social isolation; 31% of respondents reported increased work from home, 38.8% reported increased telehealth use in place of in-person meetings, and 47.9% endorsed self-isolation to protect family (Appendix Table 3).

The overall prevalence of depression in our sample was 18% (160/889) and almost half (47.8%; 426/892) reported feeling burned out. Despite the significant levels of distress, burnout, and loneliness, 68.3% of respondents reported being satisfied with their job. Those who reported feeling burnout and those with symptoms of depression both had increased loneliness scores (both P < .001). Those who reported fewer meaningful professional/clinical interactions as well as nonclinical/work related interactions had higher loneliness scores (both P < .001). The medical staff with higher job satisfaction scores had lower loneliness scores (P < .001). Lastly, certain COVID-19 behaviors, such as increased work from home (P = .002) and increased self-isolation (P < .001), were associated with increased loneliness scores.

Using a Spearman rank correlation coefficient (rho, R), the correlation between loneliness and each of burnout, depression and job satisfaction was statistically significant (P < .001). Burnout and depression showed a direct correlation (R = .637 and .688, resp.) with loneliness; job satisfaction showed an inverse correlation (R = -.645).

The four clinical work areas were divided based on work environment and the perceived extent of social interaction in each of the work areas. These groups were compared based on meaningful interactions at work, meaningful non-work interactions, burnout, depression and job satisfaction. Hospital-based staff had significantly higher scores for work and non-work meaningful interactions and job satisfaction, compared to procedural area staff, where quality personal interactions were reported less frequently. In addition, the hospital-based staff had lower scores for depression compared to procedural area staff, and lower burnout scores compared to procedural area staff. Professional degree and medical staff membership category did not affect these key outcomes – location and type of work appeared to be most influential.

The staff of the four clinical work areas were also compared by their responses to isolation behaviors. Staff from procedural work areas reported that the set-up of their office/clinical space precluded socialization, and the procedural area staff more frequently reported not having enough time with their peers.

Discussion

Loneliness is an important issue in society, ⁶ and in as early as 2017, former US Surgeon General Vivek Murthy called loneliness a public health epidemic. ¹⁸ In their 2020 Loneliness and the Workplace Report, health insurer Cigna surveyed over 10,000 Americans and found that 61% of respondents were lonely, and that loneliness had increased by 7 percentage points from 2018 to 2019. ⁹ Anchor et al. reported that those with professional degrees (law and medical degrees) were the loneliest, ¹ and many groups have reported loneliness in the medical field.^4,19-21

Primary findings

We found the prevalence of loneliness among our responding medical staff to be 48%. This level of loneliness is very similar to the prevalence reported by Ofei-Dodoo et al. in their study of the members of the Medical Society of Sedgwick County in Kansas, ⁴ and is lower than the prevalence of 61% from the Cigna 2020 national survey. ⁹ Our study surveyed all active members of the medical staff (physicians [MD, DO, MBBS] and non-physicians, including PhD or PsyD, APRN, PA, and MS/MPH), whereas the report from MSSC included only physicians. This indicates that while the geographic area and actual roles differ, the degree of loneliness among healthcare professionals is consistent. In contrast to the Ofei-Dodoo et al. study, we found a significant increase in loneliness among women and no significant difference by age. In addition, we found significantly greater degrees of loneliness among those who identified as African-Americans.

We found a number of relationships between loneliness and oganizational, cultural and personal factors. We believe there is utility in understanding the drivers of loneliness in a framework that can inform an organizational response to improve the provider experience.

Organizational

Organizational factors related to loneliness included work location or area, office setup, and frequency of remote work. During the COVID-19 pandemic, work environments shifted dramatically due to social distancing, remote work, and virtual meetings. We expected the physical environment to play a significant role in the prevalence of loneliness, and we found significantly greater degrees of loneliness among those who primarily worked in procedural areas. The hospital-based workers reported the lowest loneliness scores, likely due to the physical proximity of colleagues working toward a shared goal. Those working on inpatient units or floors, by the nature of the work, constantly interact with others, and friendships often develop as a result of these frequent interactions. This enhances the number and quality of interactions. The opportunities for similar interactions may not exist to the same extent based on the set up and work flow of procedural areas and therefore similar friendships may not develop.

Society intentionally reduced social interactions as a means by which to combat COVID-19. Those who reported not enough time with their peers, increasing the amount of work they did from home and self-isolating to protect their families all had significantly higher loneliness scores. As described above, loneliness is determined by the perception of quality and to a lesser degree, the quantity of social relationships. It makes sense that if there were a decrease in the opportunity for meaningful social interactions, then there would be an increase in loneliness.

Culture

Cigna’s Loneliness and the Workplace report identified a significant relationship between workplace culture and loneliness, with employees reporting less loneliness when they can leave work at work and have employers who promote work-life balance. Our results showed that loneliness was associated with cultural factors including decreased time with peers and the perception of an organizational culture that doesn’t promote socialization. If leadership does not consciously promote connectedness, particularly under isolating conditions such as those experienced during COVID-19, the healthcare workforce will not feel supported to do so themselves. Understandably and rightly so, healthcare providers and organizations prioritized patient care and general operations during the pandemic, and the felt experience of those who work in healthcare settings may be that socialization and connection with others is both physically risky and frivolous.

Personal

Personal factors related to loneliness included caution due to COVID-19, exhaustion (leading to increased self-isolation), feeling misunderstood by others, and symptoms of depression and burnout. We found a prevalence of depression of 18% and a 47.8% prevalence of burnout among the study participants, and the loneliness scores were significantly higher for both those who were depressed and those who reported feeling burned out. Those who reported feeling exhausted from work also reported higher loneliness scores. This correlation between depression and/or burnout and loneliness is consistent with other reports in the literature.^1,3,13 Relatedly, Seppala and King (2017) suggested that loneliness is not a result of social isolation, but instead is secondary to the emotional fatigue that results from workplace burnout. ⁷

We found an inverse correlation between job satisfaction and loneliness, such that those who reported higher job satisfaction had lower loneliness scores. This was consistent with Cigna’s Loneliness and the Workplace report, which showed that individuals less satisfied with their workload, relationships and work-life balance were significantly more lonely. ⁹ Perhaps those who have more meaningful social interactions at work have a resultant greater job satisfaction, with more meaningful interactions reducing the level of loneliness.

Limitations

Survey studies can be hindered by a low response rate. Over 50% of those asked to participate in this voluntary and anonymous study did – which is a rate equivalent to this group’s participation in engagement surveys. Given pandemic stress and fatigue, this response rate is impressive though not quite to our target. Given the similarity between responses to this study and other studies with this population and in this setting, there is no reason to suspect that our data are skewed or negatively affected by response rate.

We acknowledge that some participants may work in several clinical work areas and that we analyzed data based on primary work area. While this may have impacted our results to a degree, it is assumed that primary work area likely has a large effect on work-related experience and well-being.

It is important to note that this study was conducted approximately one year into the COVID-19 pandemic. The total number of COVID-19 or COVID-related patients in our 867-bed hospital during this time ranged from approximately 30 to 45 patients a day, and the infectivity rate in our state was from 2.5% to 5.2%. Clinical operations had increased from the low volume experience during the peak of COVID-19, but the clinical volumes were not yet back to pre-COVID-19 levels. Most of the medical staff had access to the COVID-19 vaccine by this time, and the strictest COVID-19 precautions were still in place, including no in-person gatherings and distancing. In addition, no one had an opportunity to travel, for business or pleasure. Considering that the medical staff (and all of society) had been subjected to these conditions for over a year, it is likely that this survey was conducted at the nadir of social and emotional well-being. It is highly probable that these circumstances affected the results of this study, and it would be important to conduct this survey again after COVID-19 related restrictions decrease. This would also help to further elucidate the effect of COVID-19 on the well-being of the medical staff. This said, the discrepancies between individual groups are in and of themselves interesting and important for further study and intervention.

Conclusion and next steps

We found significant levels of loneliness among our medical staff that appeared strongly related to work setting and environment. In addition, we found that social isolation behaviors were associated with higher loneliness scores. To our knowledge, this was among the first studies in the U.S. examining healthcare provider loneliness at this point in the COVID-19 pandemic. While the precise impact of COVID-19 remains unclear, the pandemic has provided us with a novel opportunity to examine the impact of increased social isolation on the emotional experience of loneliness.

Further studies are needed to determine the post-pandemic prevalence of loneliness and to better understand the underlying factors for the differences of the loneliness levels among the different work areas. Understanding these underlying factors will enable the institution to design programs or structure the various work environments in order to reduce loneliness and enhance well-being. If the causes of loneliness and social isolation can be understood in terms of organizational, cultural and personal factors, then the solutions should align with those drivers. Given the known impact of loneliness on personal health and work-related well-being, we have a responsibility and an opportunity to improve the experience of our healthcare workforce through these organizational, cultural and personal solutions. Immediate solutions could include increasing opportunities for meaningful social interactions at work (organizational), and prioritizing social connection as a pathway to post-COVID healing (cultural). We suggest future research examining the efficacy of these and other strategies (Appendix Table 4) within these three core domains.