Psychological distress among healthcare workers accessing occupational health services during the COVID-19 pandemic in Zimbabwe

Background Healthcare workers (HCWs) have experienced anxiety and psychological distress during the COVID-19 pandemic. We established and report findings from an occupational health programme for HCWs in Zimbabwe that offered screening for SARS-CoV-2 with integrated screening for comorbidities including common mental disorder (CMD) and referral for counselling. Methods Quantitative outcomes were fearfulness about COVID-19, the Shona Symptom Questionnaire (SSQ-14) score (cutpoint 8/14) and the number and proportion of HCWs offered referral for counselling, accepting referral and counselled. We used chi square tests to identify factors associated with fearfulness, and logistic regression was used to model the association of fearfulness with wave, adjusting for variables identified using a DAG. Qualitative data included 18 in-depth interviews, two workshops conducted with HCWs and written feedback from counsellors, analysed concurrently with data collection using thematic analysis. Results Between 27 July 2020–31 July 2021, spanning three SARS-CoV-2 waves, the occupational health programme was accessed by 3577 HCWs from 22 facilities. The median age was 37 (IQR 30–43) years, 81.9% were women, 41.7% said they felt fearful about COVID-19 and 12.1% had an SSQ-14 score ≥ 8. A total of 501 HCWs were offered referral for counselling, 78.4% accepted and 68.9% had ≥1 counselling session. Adjusting for setting and role, wave 2 was associated with increased fearfulness over wave 1 (OR = 1.26, 95% CI 1.00–1.60). Qualitative data showed high levels of anxiety, psychosomatic symptoms and burnout related to the pandemic. Mental wellbeing was affected by financial insecurity, unmet physical health needs and inability to provide quality care within a fragile health system. Conclusions HCWs in Zimbabwe experience a high burden of mental health symptoms, intensified by the COVID-19 pandemic. Sustainable mental health interventions must be multisectoral addressing mental, physical and financial wellbeing.


Introduction
Globally the COVID-19 pandemic has resulted in increased prevalence of psychological distress [1,2]. Actions to mitigate the spread of SARS-CoV-2 such as lockdowns, social distancing and school closures, and their knock-on effects on livelihoods have exacerbated psychological distress and anxiety.
In the context of the pandemic, healthcare workers (HCWs) are particularly vulnerable to psychological distress. [3][4][5] Risk factors include their perceived increased risk of SARS-CoV-2 infection, insufficient supplies of personal protective equipment (PPE), limited treatment options for patients with COVID-19, stigma and discrimination because of their profession, personal fear of infecting their loved ones, isolation from family members and being quarantined. [6][7][8][9] Yet HCWs are crucial in ensuring an effective response to COVID-19 including diagnosis and treatment of patients infected with SARS-CoV-2, implementation of appropriate infection prevention and control (IPC) measures, vaccination and continued service provision for other health conditions.
Most studies assessing the mental health of HCWs were rapid crosssectional surveys [3,4] providing a snapshot during a certain phase of the pandemic such as the "first wave" which was characterised by unprepared health care systems and high levels of uncertainty. Mental health needs of HCWs are likely to differ as the number of SARS-CoV-2 infections wax and wane, health systems adapt, and more information becomes available. Cross-sectional surveys among HCWs in sub-Saharan Africa, most of them distributed through online platforms and conducted during the first six months of the pandemic, revealed high prevalence of anxiety, psychological distress, insomnia and symptoms of depression. [10][11][12][13][14][15][16][17][18] To date, there is little evidence on the impact of interventions to benefit the resilience and mental health of HCWs during or after epidemics. [19] Many countries experience continuing SARS-CoV-2 transmission and substantial associated morbidity and mortality. Hence mental health programmes are being implemented under public health emergency circumstances with no resources for evaluation.
In July 2020 we set up an occupational health programme in Zimbabwe for HCWs offering screening for symptoms of common mental disorder (CMD) integrated with screening for SARS-CoV-2 and other infections including HIV and tuberculosis (TB) as well as common non-communicable diseases. We conducted a mixed-methods study to investigate changes in psychological distress and anxiety among HCWs accessing the programme over 12 months across three SARS-CoV-2 waves and evaluated a psychosocial support model that combined screening for CMD with referral for remote counselling.

Study setting and population
The study was conducted between 27 July 2020 and 31 July 2021 in public hospitals in Harare and Chitungwiza, primary health clinics in Harare and mission hospitals near Harare. Sites were selected on the basis of need and logistics, beginning with the only functioning COVID-19 unit, at Parirenyatwa Group of Hospitals in Harare. Services were offered at respective health facilities over several weeks until demand decreased. From 1 October 2020 to 31 January 2021 two teams were active allowing parallel service provision at two locations. Details of study procedures have been described elsewhere. [20] All employees of the health facilities where the service was offered could access the occupational health service free of charge. Those who accessed the service and consented to participate in the research were included in the analysis. Those who did not consent to participate in the research could also access the service.

Interventions and procedures
The occupational health programme was offered during weekdays on an appointment basis. It was advertised to HCWs through fliers and posters, via departmental heads, on work social media platforms and through word-of-mouth. Services were offered in an outside space wherever possible using tents to ensure good ventilation, with social distancing observed. HCWs accessing the service were provided with an information sheet and verbal consent was obtained. HCWs could opt-out of any screening test. Screening included measurement of height, weight, temperature, oxygen saturation, blood pressure, point-of-care HbA1c (SD Biosensor, Singapore) and HIV testing, either providerdelivered rapid blood test (Alere Determine HIV 1/2, USA) or an oral mucosal transudate self-test (OraSure Technologies, USA), selfadministered on-or off-site. HCWs were screened for symptoms of TB and COVID-19; those screening positive were offered sputum and/or nasopharyngeal swab testing for TB and SARS-CoV-2 respectively. Test results were returned to HCWs within 48 h. Negative SARS-CoV-2 and TB results were communicated either by phone or SMS/WhatsApp. HCWs with positive SARS-CoV-2 results were contacted by telephone and given advice on IPC measures for themselves and household contacts. Severity of symptoms was assessed, and referrals made for hospital admission if warranted.
HCWs were screened for CMD using the Shona Symptom Questionnaire (SSQ-14) that was developed and validated in Zimbabwe, with a score ≥ 8 suggestive of CMD. [21] HCWs with a SSQ score ≥ 8, having 'red flags' (suicidal ideation and/or visual or auditory hallucinations) or testing positive for SARS-CoV-2 were offered referral for telephone counselling, provided by the Harare-based Counselling Service Unit (CSU) free of charge. CSU is a registered non-governmental health facility established in 2003. All CSU counselling staff hold recognized qualifications in their respective professions with accredited universities/institutions of higher learning. HCWs who accepted counselling were asked for their telephone numbers. CSU counsellors contacted HCWs on the same day for red-flag referrals and within 2-3 days for other referrals. The number and frequency of counselling sessions were tailored to the needs of the HCW. Clients who needed further care were referred to government hospitals.

Quantitative data collection
While HCWs awaited screening, a trained research assistant administered a questionnaire, which included questions on past medical history, contact with patients infected with SARS-CoV-2 with and without appropriate PPE, perceived severity of the COVID-19 pandemic in Zimbabwe (known as concern score, on a 0-10 scale) and fearfulness about COVID-19 on a 4-point Likert scale. Data were collected on tablets with forms designed using SurveyCTO software, uploaded daily and saved to a Microsoft SQL Server hosted at the Biomedical Research and Training Institute (BRTI).

Qualitative data collection
To better understand the mental health stressors experienced by HCWs and how best to model our services we used in-depth interviews and participatory workshops. We began with two participatory workshops within the first month of service provision, held at one of the main hospitals, with department representatives invited to participate. The main purpose of the workshops was to learn what worked well and what could be improved, to inform changes to the service design. We used participatory workshops as a means of data collection because nationwide industrial action was ongoing and the workshop was the only opportunity to engage with low-level HCWs. We anticipated to complete 15-25 in-depth interviews with HCWs who accessed the service. Data analysis occurred concurrently with data collection, and interviews were stopped once data saturation had been achieved. Interview participants were purposively selected to include varying SSQ-14 scores (<8 and ≥8), SARS-CoV-2 test results, presence of diabetes and/or hypertension.
Semi-structured topic guides were developed. The workshops topic guide sought to elicit the contextual setting in which the intervention was embedded and gain an understanding of the occupational health services provided at the health facilities. The interview guide asked questions on the perceived impact and experiences of the pandemic, on the psychological wellbeing of HCWs including their feelings, fears and anxieties, as well as the stressors and protective factors for mental health. Interviews were conducted face-to-face or by telephone depending on lockdown restrictions at the time. CSU counsellors fed back regularly on operational issues, and wrote 2 reports summarising recurrent themes that emerged during the counselling sessions.
All interviews were conducted in the local language (Shona), audio recorded, transcribed and translated into English. We used thematic analysis to identify, analyse and interpret patterns of meaning within the data. The transcripts were reviewed several times by the research team, with important statements being extracted, coded and discussed. Codes were summarized into themes and a continuous comparison of codes and categories was carried out with the research team.

Choice of outcome measure
The primary outcome for the quantitative analysis was fearfulness about COVID-19. HCWs were asked 'How do you feel about COVID-19?',   Table 1) is widely used in Zimbabwe to screen for CMDs at population level. It consists of 14 yes/no items and takes <10 min to administer. A cutpoint of ≥8 has been validated against a diagnosis of anxiety and/or depression using the Structured Clinical Interview of the DSM-IV applied by psychologists. [21]

Statistical analysis
The primary exposure was time period, defined based on the national COVID-19 data indicating the first, second and third waves, with corresponding lull periods between waves (Fig. 1). The start of a wave was defined by a 7-day average of ≥50 cases excluding localised outbreaks, while the end of a wave was defined by a 7-day average of <50 cases.
Median and IQR of SSQ-14 score (range 0-14) and concern score (range 0-10) were calculated per week and lowess curves were fitted.
Definitions for body mass index (BMI) followed World Health Organization guidelines. [22] An individual with a BMI of 25 to <30 kg/m 2 was considered overweight, and of ≥30 kg/m 2 obese. Occupations were coded as patient-facing and non-patient-facing. Patient-facing occupations comprised: nurse, nurse aide, midwife, doctor, community health worker, radiographer, dentist, physiotherapist, social worker, counsellor, or student of any of the above. Domestics and cleaners were also coded as patient-facing due to their increased exposure risk when cleaning rooms and attending to linen and waste. Non-patient-facing occupations included security guards, pharmacists, administrators, and kitchen staff. Clinic setting was coded as tertiary hospital, district or mission hospital and primary health clinic. Analysis was performed using Stata v15 and graphs were created using R.

Ethics
Ethical approval was granted by the Institutional Review Board of the BRTI, the Medical Research Council of Zimbabwe (MRCZ/A/2627) and the London School of Hygiene & Tropical Medicine ethics committee (22514). HCWs were given an information sheet about the occupational health services. The study was granted a waiver allowing verbal rather than written consent to be obtained because the primary intention of the study was to provide an occupational health service. Participants provided written informed consent for the workshop and face-to-face in-depth interviews. If the interview was conducted by telephone the information sheet was sent to the participant via email prior to the call. At the time of interview the study procedures were explained, any questions were answered and recorded verbal consent was obtained.

Results
The following dates defined the SARS-CoV-2 waves: 16 July 2020-13 September 2020 (wave 1), 11 November 2020-25 February 2021 (wave 2), and 6 June 2021 -ongoing at the time of writing (wave 3). The service was accessed 3673 times between 27 July 2020 and 31 July 2021, by 3577 HCWs (90 HCWs came twice and three came thrice) from 22 facilities. The median age was 37 (IQR 30-43) years and 81.9% were women (N = 3007) ( Table 1). Most HCWs (N = 2800, 76.2%) worked in patient-facing roles with the majority employed as nurses (828/2800, 29.6%), midwives (408/2800, 14.6%) and cleaners (425/2800, 15.2%). Almost two-thirds worked in tertiary hospitals (N = 2279; 62.1%), A total of 1474/3539 (41.7%) said they felt fearful about COVID-19. HCWs were more likely to be fearful if they worked in patient-facing roles, at district or mission hospitals or primary health clinics, had contact with a patient or colleague infected with SARS-CoV-2 without PPE, had received no IPC training, or had known chronic health conditions (Table 1). In univariate analysis odds of being fearful was higher during and following the second wave compared to the first wave. After adjusting for setting and patient-facing role, the odds ratio of fearfulness during the second wave compared to the first was 1.26 (95%CI 1.00-1.60) ( Table 2).
HCWs' perception of the seriousness of the pandemic in Zimbabwe mirrored SARS-CoV-2 notifications over time, while median SSQ-14 score increased throughout the 12 months (Fig. 3). Fig. 4 displays the proportion of HCWs answering yes to 12 of the 14 SSQ-14 items (excluding the rarest 2 items) across the five time periods (three waves and two lull periods). The proportion of HCWs reporting failure to concentrate, being frightened and losing their temper over trivial matters, having difficulty deciding what to do and 'thinking too much' (a local idiom for depression) steadily increased over time. When asked if there were times during the past week when they felt life was so tough that they cried or wanted to cry, one in three HCWs responded with yes   Fig. 3. Level of concern and mental health symptoms over time. Rolling 7-day average of SARS-CoV-2 cases (grey), SSQ-14 score median (red line) and 95% confidence intervals, score of the level of concern about the COVID-19 situation in Zimbabwe median (blue line) and 95% confidence intervals. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) symptoms increased from 8.7% (95%CI 6.2-11.8), to 10.0% (8.4-11.8), 13.0% (11.0-15.2), 15.5% (12.7-18.6) and 15.6% (11.7-20.3) across waves and lull periods. After adjusting for setting and patient-facing role the adjusted odds ratio for CMD symptoms in wave 2 versus wave 1 was 1.43 (95%CI 0.97-2.12) ( Table 2). Only 3 (0.8%) out of 3673 HCWs did not complete the SSQ-14, opting out of mental health screening. Among those completing the SSQ-14 13.4% (N = 490) had an SSQ-14 ≥ 8 and/or red flags and 382 (78.8%) accepted referral for counselling. An additional 11 HCWs tested SARS-CoV-2 positive and agreed to counselling. CSU counsellors successfully contacted 87.7% (345/393) HCWs who were referred for counselling (Fig. 5).
Emerging themes were categorised into fear and anxiety, stressors, and protective factors improving mental health and well-being ( Table 3). The CSU counsellors noted that "clients' concerns originated from a combination of factors, ranging from general unmet social, economic, psychological, and medical needs; poor working conditions The following questions related to the course of the past week and each heading of a panel: Concentration: "Did you find yourself sometimes failing to concentrate?" Cried: "Were there times when you felt life was so tough you cried or wanted to cry?" Frightened: "Were you frightened by trivial things?" Hallucinations: "Did you sometimes see or hear things others could not see or hear?" Nightmares: "Did you have nightmares or bad dreams?" No choice: "Did you feel you had problems deciding what to do?" No sleep:"Did you sometimes fail to sleep or did you lose sleep?" Stomach pain: "Was your stomach aching?" Suicidal ideations: "Did you sometimes feel like committing suicide?" Temper: "Did you lose your temper or get annoyed over trivial matters?" Thinking too much:"Did you sometimes think deeply or think about many things?" Tired: "Did you feel run down (tired)?" Unhappy: "Were you generally unhappy with the things you were doing each day?" Work lagging: "Was your work lagging behind?" and the current COVID-19 concerns". Most interviewees expressed feelings of fear and anxiety related to becoming infected or potentially infecting their loved ones with SARS-CoV-2. As the number of SARS-CoV-2 cases and related deaths increased, death became a source of distress, especially among parents of dependent children. Fear and anxiety were enhanced when PPE was lacking, particularly among those with pre-existing health conditions who felt at increased risk of severe COVID-19.
Stress was common and some participants reported feelings of burnout. Work-related pressures included long working hours and inadequate resources hindering HCWs to provide quality patient care and to practise recommended IPC. Disordered sleep and overthinking were frequently reported. HCWs felt vulnerable and alone because of societal stigma and discrimination from their families and communities. They were "labelled" as "the person who works in the hospital". Some HCWs who looked after patients with COVID-19 or had been exposed to patients or staff with SARS-CoV-2 had to live separately from their families or temporarily relocate other household members. Lockdowns restricted socialising which was a source of distress for some participants as they were unable to attend church services, or funerals of their loved ones, and meet socially with family and friends. This aggravated stress levels. Relationship problems and family tensions were additional stressors. CSU counsellors reported that whilst "some of the clients scheduled for review were feeling much better, there were high levels of ongoing distress due to poor health, safety issues as well as economic challenges". The CSU highlighted the "need for additional support beyond crisis care" for HCWs.
HCWs mentioned several factors and coping mechanisms that promoted their mental well-being. During lockdown some HCWs tried to improve their physical health by "spending more time exercising" whilst others took the opportunity to strengthen relationships through "increased family time". HCWs also identified positive aspects in the health system as a result of SARS-CoV-2 such as increased IPC training and enforcement of IPC measures.
HCWs who received counselling expressed gratitude and thought it was helpful. The knowledge that somebody "cared about you" made them feel "recognized" and valued. Being able to talk to somebody, if needed, provided reassurance. A 59 year-old midwife said "This call to ask about one's wellbeing is of great importance to us, even if I get overwhelmed, I know I can call that person now and tell them I am not well". HCWs gave mainly positive feedback about the occupational health service: a "free comprehensive service" offered in "a friendly atmosphere". The service encouraged HCWs to be more health conscious, to "know their health status" and to prioritise their own wellbeing by improving their "selfcare and lifestyle".
CSU had initially planned to establish WhatsApp peer support groups for HCWs, but HCWs preferred one-to-one counselling sessions from external providers. Thus, CSU provided client-centred counselling tailored to individual needs. In the early months CSU counsellors often had to make several attempts to contact HCWs, and failed to reach some because of wrong numbers or phones not working. From January 2021 HCWs were asked for multiple phone numbers including a number of a trusted friend, which improved the contact success rate. Another challenge CSU encountered was the unavailability of HCWs during working hours. Some HCWs had difficulty finding a time and place when they could speak openly about their problems.

Discussion
Four in ten HCWs in Zimbabwe felt fearful about COVID-19 with a 1.26-fold increase during the second wave. The prevalence of insomnia, nightmares, somatisation (stomach-ache), tearfulness and fatigue were high throughout the study spanning three national SARS-CoV-2 waves. The prevalence of concentration difficulties, feeling frightened, shorttemperedness and indecisiveness increased steadily over the 12 months of the study as did median SSQ-14 score.
Other studies conducted among HCWs in Africa during the COVID-19 pandemic reported prevalence of anxiety including mild, moderate  Themes, codes and supporting quotations from reports, workshops and in-depth interviews.

Theme Codes Quotation
Fear and anxiety Fear of being exposed "My first encounter with a COVID-19 patient made me cry, I was asking God why. I was very shaken, I thought I was going to contract the virus and I was very scared." (nurse, female, 58) "My work mates tested positive for COVID-19…this made me feel unstable, as we did not know when the virus would end and we thought we will end up getting infected as well, so it affected our mental health, and we were very unstable and shocked because we would even share things with them." (nurse, female, 49) "I was afraid of the place I would be working; I would hear stories of people who had died there." (nurse, male, 63) Heightened risk perception due to lack of PPE "What made me overthink was we as frontline workers were asked to come to work without adequate protective equipment and this was at a time when COVID-19 was initially hitting the most, in the January-February period (2021), that is the time I was over thinking, I thought God what can we do, how will we live through this?" (midwife, female, age unknown) "At start it was not that adequate but as we progressed on and also the numbers of those affected with COVID-19 were increasing, they started to procure more and more PPE so that the staff was not exposed… that made me feel better" (nurse, male, 48     . "I was happy and hoped I stayed home more to spend time with my child." (nurse, female, 31). Family support "I was scared that he would contract COVID-19 but he would say…if I leave no one else will help you. I was happy that this man was loving and he stood by me." (nurse, female, 29). "It is these same people (family) who reassured me when they heard that mum has also tested positive, they sent lots of advice and what not… I got a lot of support and I never got stressed then" (midwife, female, 59) "Social support is helping us as well. As an individual I get support from my work colleagues, when you are at work you talk and help each other and even our bosses help us sometimes. This gives us hope that at least when we are working, we will not be too afraid." (nurse, female, 29). Access to psychological support "This call to ask about one's wellbeing is of great importance to us, even if I get overwhelmed, I know I can call that person now and tell them I am not well" (midwife, COVID-19 patient, female, 59) "CSU called me, they asked how I was feeling, what I was thinking about…they still call now to ask how I am feeling… I was called by 2 or 3 different people…The calls helped because I realized there are some people who are caring about you when you think you "I think all the services were helpful because it's rare to find people who actually explain the purpose, even when you are told you are positive it doesn't mean you are dying. They explained my results in relation to the manner in which one was living. I was told my weight and height was not proportional and I had to lose at least 6 kgs from 78. Yes, I am trying to watch my diet, I am increasing intake of vegetables." (mid-wife, female, 59) "All in all, everything was good, I was helped with my mental health which from my responses, it showed that I was not stable. This could be because of my work, we transport Covid-19 patients to Parirenyatwa and you start to think that you will get infected and this increases stress or BP […] So, it all went well, we hope you will keep coming back." (ambulance driver, male, 48) "I was happy with everything because I was actually one of the patients who got diagnosed as diabetic on those tests regardless of being covid-19 negative." (nurse, female, 29) Referrals to further care "The services are very good because when I met them, I was a bit depressed and I was helped, I managed to get counselling, because the time I was helped, I was suicidal. Mental health was relevant to (continued on next page) "When I visited the tents, the staff was polite, and they did their testing privately and confidentially. What actually disappointed me was that I think the staff were short staffed. Such a good service needs to expand and to be well staffed to meet our demand" (nurse, female, 29) "The service is good. We feel at ease to come here than the staff clinic because we know our privacy will be respected. On top of that having such free and comprehensive services should be the benefit given to us as healthcare workers. I would recommend you make two visits a year at each facility." (casualty nurse, male, 33) Feedback from CSU

Service provision & general assessments
"We [CSU] initially envisioned support groups for referred workers, however early assessments demonstrated that there was low interest from referred staff in this, with a preference for one-on-one support from an external provider" "Some also preferred using WhatsApp chat as a mode of counselling, which was offered as per request. Psychological support is ongoing on both whatsapp texts at a time requested by the client and telecounselling." (CSU, report 01) "Similar to 2020, clients' concerns originated from a combination of factors ranging from general unmet social, economic, psychological, and medical needs, poor health and safety working conditions to the current COVID-19 concerns." "It was noted that 30% of clients scheduled for review were feeling much better as revealed by SSQ8 score on re-assessment and their narratives. However, 70% have remained anxious citing poor health and safety issues as well as economic challenges. The high levels of ongoing distress in follow up cases demonstrates the high need for this group and indicates that additional support beyond crisis care may be required for this group of health workers." "The clients' mental state were disturbed due to operational environment, particularly those with underlying health conditions. They all bemoaned lack of equipment that protect them against health and safety risks during the course of their work (PPE)." and severe ranging from 25.5% to 90.5%. [10][11][12][13]17,[23][24][25][26] The prevalence of depressive symptoms ranged between 32.1%-94.0%. [10][11][12][13]17,[23][24][25][26] Most of these studies used the Generalized Anxiety Disorder 7 (GAD-7) [11,12,15,16,23,24,[26][27][28] and Patient Health Questionnaire 9 (PHQ-9) [12,15,23,24,26,28] to determine anxiety and depression. Other data collection tools included the 10-item Cohen Perceived Stress Scale [10,26,27,29], 14-item Shirom-Melamed Burnout measure [10], 7-item Insomnia Severity Index [12,26,28,30], 22-item Impact of Event Scale-Revised [12,28], 16-item Stanford Professional Fulfilment Index [12,28], Hospital Anxiety Depression Score [13], Coronavirus Anxiety Scale [17] and Primary Care PTSD Screen. [15] The differences in prevalence of anxiety and depression across studies even when using the same tools may be explained by different cutpoints, study populations (primarily doctors versus primarily nurses), settings (COVID-19 specialist services versus general medicine, obstetrics [31] and psychiatry [14]), health system capacity, data collection methods (online versus in-person interviewer or self-administered) and timing of the survey in relation to the global and national pandemic situation. Of note, while most studies used standardised data collection tools they were not necessarily validated for the context or the population. Also online surveys are vulnerable to selection bias that can affect prevalence estimates. [32] Compared to the depressive symptoms and anxiety reported by other studies the prevalence of CMD in this study was relatively low. This is not surprising as the SSQ-14 cutpoint of ≥8 is used for CMD case-finding and to identify individuals who should be referred for psychological assessment and intervention. [21] A recently published study reported a CMD prevalence of 11% among lay counsellors in Zimbabwe before the COVID-19 pandemic [33], similar to the prevalence among HCWs in this study.
In contrast to other rapid cross-sectional surveys aiming to describe the prevalence of mental health symptoms among HCWs, we used the SSQ-14 as a screening tool to identify those in need of psychological support. HCWs were free to choose any screening tools or tests provided by the occupational health service. Of note uptake of mental health screening was almost 100% indicating high acceptability. In contrast uptake of HIV (58.5%) and HbA1c (94.6%) testing was considerably lower (data not shown). HCWs felt that mental health screening was an important aspect of the service. More than two thirds of HCWs who were offered referral agreed to counselling. In-depth interviews and reports from CSU showed that counselling was helpful, but also highlighted the need for ongoing support beyond the pandemic. Pre-existing stressors including financial constraints and relationship problems, were aggravated by the pandemic because of increased costs of commodities and limited social interaction and support. Salaries of health care workers in Zimbabwe have not kept pace with inflation (255% in 2019 and 557% in 2020 against the US dollar), hence in 2020 a nurse earned the equivalent of USD 50 per month, while a junior doctor earned USD 70. These untenable conditions have led to repeated industrial action, a huge brain drain of health care workers, and further pressure on those who remain.
Similar to other settings, HCWs in Zimbabwe feared for the lives and  wellbeing of themselves and their loved ones. [9,15] Those with dependent children felt particularly vulnerable and anxious, as shown in other studies. [9] This study has several strengths. Services were offered over 12 months allowing capture of trends across three waves. The study included HCWs from all tiers of the healthcare system and a wide range of professions. The SSQ-14 has been designed, validated and used extensively in Zimbabwe. [21,[33][34][35][36] The mixed-method approach enabled a more in-depth understanding of aggravating and mediating factors of mental health and well-being.
The study also has limitations. HCWs taking up the occupational health service may have been more health-conscious than those who did not, introducing selection bias. However, the service was not focused exclusively on mental health and the decision of which "screening package" to take up was entirely voluntary. Service provision was guided by need and hence tertiary hospitals were prioritised during the peak of SARS-CoV-2 waves. We tried to adjust for that in the analysis, but residual confounding cannot be excluded.
In conclusion, HCWs in Zimbabwe experience a high burden of mental health symptoms, intensified by the COVID-19 pandemic. Financial insecurity, unmet physical health needs and inability to provide quality care within extremely limited resources impede on the mental well-being of HCWs. Sustainable mental health interventions must be multisectoral addressing mental, physical and financial wellbeing. agreement which is also part of the EDCTP2 programme supported by the European Union (grant number MR/R010161/1). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Contributors
KK, VS, RAF, CEN, HM, SR conceptualised the study. RMSC, FPN, ETM collected data. KK and VS led the statistical data analysis with input from IDO. RMSC led the qualitative data analysis. MM, TH and NX provided counselling services and contributed to qualitative data collection. AJVM, CP, MC and PC facilitated service delivery at health care facilities. KK, RMSC, VS and RAF wrote the first draft of the manuscript. All authors provided input to the draft manuscript and read and approved the final manuscript.

Data sharing statement
Individual, anonymised participant data and a data dictionary will be available through the London School of Hygiene & Tropical Medicine repository (Data Compass) 12 months after publication of results.

Declaration of Competing Interest
No conflict of interest.

Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi. org/10.1016/j.comppsych.2022.152321.