Impact of COVID-19 on the mental health of Singaporean GPs: a cross-sectional study

Background COVID-19 has stressed healthcare systems and workers worldwide. GPs, as first points of contact between suspected cases and the healthcare system, have assumed frontline roles in this crisis. While the prevalence of mental health problems and illnesses arising in healthcare workers (HCWs) from tertiary care settings during the COVID-19 pandemic is well-examined,1 the impact on GPs remains understudied. Aim To describe the prevalence and predictors of anxiety, burnout, depression, and post-traumatic stress disorder (PTSD) among GPs during the COVID-19 pandemic. Design & setting Survey of GPs operating in Singapore primary care clinics. Method GPs completed a survey that comprised of four validated psychometric instruments. Open-ended questions asked about responders’ challenges and their envisaged support. Data were analysed with multiple logistic regression with demographic data as covariates; concepts of grounded theory were used to analyse the qualitative responses. Results A total of 257 GPs participated. Fifty-five (21.4%) met the scales’ criteria for anxiety, 211 (82.1%) for burnout, 68 (26.6%) for depression, and 23 (8.9%) for PTSD. Multivariate regression analysis showed working in a public primary care setting was associated with anxiety and depression. Qualitative analyses uncovered possible stressors: changes to clinical and operational practices; increased workloads; and financial difficulties. Conclusion Mental health issues were found to be present in Singaporean GPs during the pandemic. Prevalence of anxiety, burnout, and depression were found to be higher than those reported pre-COVID-19. The findings also provide determinants of the issues that serve as possible foci for targeted interventions.


Introduction
The COVID-19 pandemic has stretched healthcare systems worldwide. Within a span of less than a year, SARS-CoV-2 -the virus responsible for the disease -infected nearly 70 million and claimed 1.5 million lives. 1,2 Globally, clinical guidelines developed in response to this public health emergency have highlighted the importance of the GP's role in screening and detecting for suspected cases. [3][4][5][6] This assignment is appropriate because physicians in general practice are typically first points of contact for people with suspected COVID-19, who often exhibit influenza-like symptoms. 7 Consequently, GPs have experienced significant changes at their workplaces including the cessation of routine services, acclimatising to new clinical practices, rationing of personal protective equipment (PPE) supplies, and being redeployed to unfamiliar workplaces. HCWs who are exposed to similar occupational disruptions, coupled with the increased risk of contracting a highly infectious disease, have been documented as having increased rates of mental health illnesses such as anxiety, depression, and PTSD, and mental health problems such as burnout. 1,[8][9][10][11] A physician workforce stricken with high rates of mental health issues has significant downstream implications. In a model produced from a review of existing research, Wallace and colleagues posited that poor physician mental health leads to poor health system outcomes, 12 with suboptimum quality of patient care, 13 increased medication errors, 14 and lower productivity and efficiency. 15 This in turn creates more workplace stressors and further deterioration of the workforce's mental wellbeing. 12 Detecting for these issues in physicians amid this crisis is, therefore, an immediate priority.
Current HCW mental health research is focused on those from tertiary care settings, 8,[16][17][18][19] while the mental health impact of COVID-19 on GPs remains unclear. A meta-analysis of studies conducted in hospital-based HCWs from countries affected by COVID-19 found that the majority of responders suffered mild symptoms for psychological distress associated with delivering care to diseased patients. 1 Given the wide-ranging psychological effects of COVID-19, and that GPs have been put at the forefront of the fight in this pandemic, it is crucial to monitor rates and factors for mental health distress in these physicians to tailor timely interventions. The aim of this study, therefore, is to examine the prevalence of COVID-19-related mental health problems (burnout) and illnesses (anxiety, depression, PTSD) within a GP setting, as well as to compare the differences in sociodemographic and work environment characteristics between those with and without such issues.

Method Setting
Singapore is a high-density, multiethnic city state of 5.7 million residents. Its first COVID-19 case was detected on 23 January 2020. A nationwide lockdown, colloquially termed 'circuit breaker', was imposed on 3 April 2020 for approximately 2 months before being gradually relaxed. 20,21 At the time of writing (January 2021), the Singapore healthcare system has registered 58 542 cases and 29 deaths arising from the disease. 22 Primary health care in Singapore is delivered in either publicly or privately funded settings. 23 Public sector GPs practice at polyclinics that serve as one-stop outpatient centres with medical and allied health (diagnostic imaging and laboratory investigations, health education, and pharmacy) services. Private sector GPs typically operate in either individually owned or small-group practices. During the COVID-19 outbreak, private sector GPs who participate in the Public Health Preparedness Clinic (PHPC) scheme are provided with government support via the provision of PPE, subventions for administering national subsidy schemes related to COVID-19, grants for operating expenses incurred, and professional training to build outbreak management capacity. 24 Local COVID-19 guidelines urge citizens who develop symptoms to seek medical attention from participating GPs (rather than at hospital emergency departments) in both public and private sectors, where fully subsidised SARS-CoV-2 polymerase chain reaction (PCR) tests are available. 6

Study design
The cross-sectional study took place in Singapore between April and September 2020. An earlier notification regarding the study was conveyed via email to GPs working in National University Polyclinics (NUP), a public sector network that operates six polyclinics, as well as various private primary care networks comprising of individual practice or small-group GPs. GPs had to be involved in clinical roles in either public or private primary care settings to be eligible to participate. The voluntary, anonymous, paper-based, English questionnaires were mailed to GPs who indicated interest to participate. GPs outside of these networks were also recruited through snowball sampling. A gift card was disbursed to all responders after survey completion as a token of appreciation.
Four commonly used, validated instruments were employed in the survey; relevant literature was referred to when deciding cut-offs for these scales. All instruments were validated for use, 25,26 or previously used, [27][28][29][30] in a Singaporean setting. Each survey also included: a section on responders' demographic and workplace environment information; and two additional, open-ended questions asking of responders' major concerns and envisaged support structures.

Instruments
Generalised anxiety Disorder-7 (GaD-7) The GAD-7 is a 7-item scale that aims to measure anxiety. Items consist of anxiety-related symptoms recorded on a 4-point Likert scale. Responders were classed as being at risk of anxiety if they met the total cut-off score of 5. 31

Oldenburg burnout inventory (OLBI)
The OLBI is a 16-item scale that aims to measure burnout. 32 Items consisting of positively and negatively worded questions relating to disengagement and exhaustion are recorded on a 4-point Likert scale. Responders were classed as being at risk of burnout if they met the cut-offs of 2.25 and 2.1 for the exhaustion and disengagement subscales, respectively. 33 Patient health Questionnaire-9 (PhQ-9) The PHQ-9 is a 9-item scale that aims to measure depression. Items consist of depression-related symptoms recorded on a 4-point Likert scale. Responders were classed as being at risk of depression if they met the total cut-off score of 5. 34

Impact of event scale-Revised (Ies-R)
The IES-R is a 22-item that aims to measure PTSD. Items measured the extent of distress in the domains of intrusion, avoidance, and hyperarousal experienced in the past week. Responders were classed as being at risk of PTSD if they met the total cut-off of 24. 35

Sample size
The study aimed to recruit 250 responders, which is consistent with the sample sizes in existing literature 11,29,36,37 assessing mental health problems and illnesses among Singaporean GPs during pandemics, which have achieved between 200 and 300 responders.

Statistics and qualitative analyses
Statistical analyses were performed using SPSS (version 25.0). The primary outcome measures were anxiety, burnout, depression, and PTSD. Explanatory variables, such as age, sex, ethnic group, marital status, type of employment, and nature of workplace environment, were included in the analyses. Multivariate logistic regression was used to assess the sociodemographic and workplace environment correlates of each dependent variable. All statistical significance was set at P<0.05.
In the qualitative analyses, detailed reading of the responses was done to identify significant and recurring themes. Understanding and validating the expressed meanings of the words and phrases used in identified themes was done through research team meetings to achieve consensus. Then, verbatim quotes were assigned and arranged into themes and subthemes using the concepts of grounded theory for data analyses. 38

Results
Quantitative data

Qualitative data
Responses from open-ended questions revealed possible contributory factors to the stresses faced by GPs and support structures they found useful or hoped 39 to receive. Three main themes of stressors emerged from the responses: (i) changed clinical and operational guidelines; (ii) increased or changed nature of workload; and (iii) financial stressors. Four types of support structures were evidenced: (i) psychosocial; (ii) clinical guidelines-related; (iii) financial; and (iv) workforce and training.

stressor 3: Financial stressors
During the nationwide lockdown and the subsequent tiered re-opening phase, clinical procedures that were deemed 'non-essential' by local authorities, such as elective dermatological treatment or health screening services, were prohibited. 39 Private sector GPs and GPs who worked in locum arrangements reported this temporary cessation of service offerings as a financial stressor:

'My overriding stressor from COVID-19 is due to drop-in patient load since mid-January 2020. The actual infection & possibility of dealing with infected patients has had minimal impact.'
(Private sector GP)

Discussion Summary
The study's findings suggest that prevalences of two mental health illnesses (anxiety and depression) and one mental health problem (burnout) at the height of the COVID-19 pandemic were higher than those reported in similar demographics during earlier periods or during previous disease outbreaks. Working in public practice was identified as a significant factor for anxiety and depression. The qualitative findings evidenced that stressors for GPs were driven by changes in service delivery modes. Suggestions from GPs to better manage future outbreaks were also gathered.

Strengths and limitations
To the authors' best knowledge, this is the first study on the impact of the COVID-19 pandemic on the mental health of GPs. Additionally, apart from determining correlated factors of the mental health issues, causative factors were also probed through qualitative inquiry. Finally, conducting this study in the months of April-September 2020, in which Singapore experienced its highest COVID-19 incidence, transmission, and mortality rates, allowed sentiments to be captured in the healthcare system during its most stressed state.
Several limitations exist. There is a paucity of research using the same (and same series of) instruments to assess for such mental health issues during pre-pandemic periods in GPs. This was owing to the numerous rating tools available, and limited the authors' ability to make statistically meaningful, empirical comparisons. It is also acknowledged that the sample size of 257 may not be representative of the sentiments of Singapore's 8000 GPs. 41 Sampling bias may have occurred as GPs were recruited within selected primary care networks in Singapore; furthermore, only GPs in these networks who had spare time in their demanding schedules responded. Lastly, the cross-sectional design of this study allowed only for correlations, and conclusions regarding causality cannot be made.

Comparison with existing literature
Anxiety and depression were noted in 21.4% and 26.6% of responders, respectively. A global metaanalysis of studies investigating for these morbidities in HCWs amid the pandemic echoes the findings in the present study: pooled prevalence rates of 23.2% for anxiety and 22.8% for depression were found across 12 publications. 1 Notably, the present study's rates of anxiety and depression are observed to be 1.5 times and 3 times higher, respectively, when compared with rates found in a similar Singaporean study, 42 which was conducted in the early phases of the COVID-19 pandemic (February-March 2020). This suggests that a worsened COVID-19 climate leads to increased anxiety and depression prevalence in healthcare providers, although more longitudinal data will be needed to confirm this association.
The OLBI instrument measures burnout in two dimensions: disengagement and exhaustion. Prepandemic studies using this instrument on similar demographics found lower severities of burnout. Mean disengagement scores assessed in HCWs worldwide during non-crisis periods ranged between 2.17 and 2.38 (versus 2.72 in the current study), while mean exhaustion scores were between 2.42 and 2.48 (versus 2.57 in the current study). [43][44][45] The criteria for burnout was met by 82.1% of the responders; the following provisos are suggested when interpreting this finding. First, the prevalence should be referenced with studies using the same OLBI burnout thresholds (≥2.25 for exhaustion;≥2.10 for disengagement), because applying other thresholds found in the literature 46 to the data may produce different findings. Second, physician burnout rates are well-documented to be high even in pre-pandemic states; for example, during medical school (44%), 47 residency years (76%), 13 and in fulltime work (54%). 48 Finally, the findings are similar to those of a recent, multinational study measuring HCW burnout using OLBI during COVID-19, which found an overall 67% rate of burnout and that physicians were 2.1 times more at risk than other healthcare professionals. 9 Interestingly, rates of PTSD (8.9%) were found to be lower as compared with existing literature. Earlier studies using the IES-R to measure PTSD in HCWs during disease outbreaks estimate prevalence rates to range between 9.4% (Singaporean primary care workers during SARS) and 36.5% (Wuhanbased HCWs during the early phases of COVID-19). 8,29,[49][50][51][52][53][54][55][56] Two possible, interrelated explanations exist. First, Singaporean GPs have reported to face less social stigmatisation from friends and family owing to infection risks during COVID-19 than in previous outbreaks. 37 This may explain the lower rates of PTSD, given that feelings of uncertainty and rejection by loved ones over contagion fears have been associated with the disorder. 37 Second, it is postulated that GPs had sufficient PPE provided by employing institutions (for public practice) or the PHPC scheme (for private practice). This limited GPs' viral exposure during clinical duties, which is a known risk factor found in HCW PTSD-focused literature. 30,49,53,55,[57][58][59][60][61] This is evidenced by the fact that 88.6% of Singaporean GPs surveyed felt they have adequate PPE. 37 Irrespective of the exact cause, it is suggested that a tight infection control strategy can be a starting point to lower HCW PTSD rates in airborne disease pandemics.
Demographic factors were found to be significantly associated with anxiety and depression only. Unexpectedly, the findings revealed GPs working in public practice (polyclinics) were at higher risk for both morbidities, which is a comparison not previously tested for in extant research. A likely stressor unique to polyclinic GPs, which was identified from the qualitative findings, was treating a higher-thanexpected volume of suspected cases. It is argued this increased proximity to the disease may have had an effect on GPs, given early evidence suggesting HCW anxiety and depression to be associated with direct diagnosis, treatment, and care of patients with COVID-19. 56,62 While there is a high proportion of the sample that met the cut-off for burnout, the multivariate regression analyses did not find any significantly associated demographic factors. A possible explanation was the relatively small number of cases who did not meet the cut-off. This reduced the accuracy of estimation for regression coefficients and tests of significance in the logistic regression model. This effect has been documented by Peduzzi and colleagues in their work of power of predictors in low events per predictor settings. 63 Implications for practice Findings discussed above highlight vulnerable segments in the GP workforce. Together with stressors and suggestions pooled in the qualitative data, two future strategies to reduce stressors behind these mental health issues are recommended for consideration. First, offering polyclinic redeployment opportunities to private sector GPs is suggested. This mobilisation serves to meet increases in patient volumes faced at these large-scale, community clinics, which should allow private GPs, whose service offerings have been trimmed by pandemic safety measures, to contribute gainfully to the national healthcare system. It should be noted that a reserve, supplementary pool of healthcare workers (Singapore Healthcare Corps) began calling for HCW volunteers in April 2020, 64 but its effectiveness and sign-up rates from private GPs have not been published. Second, the findings also highlight the need for continued, transparent communication on pandemic-related policy changes, and assistance in managing the changed clinical guidelines from local authorities to GPs. Setting up communication channels, such as dedicated hotlines for GPs' (clinical and non-clinical) concerns, can also be considered. Similar approaches have contributed to the effective national response in countries such as New Zealand 65, 66 and Taiwan. 67

Funding
This study received funding from Singapore's Institute of Mental Health through the Chairman of Medical Board fund.

Ethical approval
Institutional ethics approval was obtained for data collection in Singapore by the National Healthcare Group Domain Specific Research Board (NHG DSRB Reference number 2020-00289).

Provenance
Freely submitted; externally peer reviewed.