Portrait of Montréal healthcare workers infected with SARS-CoV-2 during the first wave of the pandemic: a cross-sectional study

Abstract

Objectives

During the first wave of COVID-19 in Québec, healthcare workers (HCWs) represented 25% of the cases in Montréal. A study was conducted to describe SARS-CoV-2-infected HCWs in Montréal, and certain workplace and household characteristics. Secondary objectives included estimating the associations between having had access to personal protective equipment (PPE) and training, and following self-isolation recommendations, and certain sociodemographic and workplace characteristics.

Methods

A cross-sectional study was conducted, based on a stratified random sample, among Montréal HCWs who tested positive for SARS-CoV-2 between March and July 2020. A total of 370 participants answered a telephone-administered questionnaire. Descriptive statistics were conducted, followed by log binomial regressions to estimate the associations.

Results

Study participants were mostly female (74%), born outside of Canada (65%), and identified as Black, Indigenous, and People of Colour (BIPOC; 63%). In terms of healthcare positions, most were orderlies (40%) or registered nurses (20%). Half (52%) of the participants reported having had insufficient access to PPE and 30% reported having received no training related to SARS-CoV-2 infection prevention, with large proportions being BIPOC women. Working evening or night shifts decreased chances of having had sufficient access to PPE (OR 0.50; 0.30–0.83).

Conclusion

This study describes the profile of the HCWs who were infected during the first wave of the pandemic in Montréal. Recommendations include collecting comprehensive sociodemographic data on SARS-CoV-2 infections and ensuring equitable access to infection prevention and control training and PPE during health crises, particularly those at highest risk of exposure.

Introduction

Montréal was one of the Canadian cities most affected by COVID-19 during the first wave of pandemic (between March and July 2020), with healthcare workers (HCWs) comprising 25% of the cases (Direction régionale de santé publique de Montréal, 2021). HCWs were particularly vulnerable to SARS-CoV-2 infection given the risk of occupational exposures (Pérez-García et al., 2020). Much has been learned about risk factors for HCW infection of SARS-CoV-2 including occupational role and sociodemographic characteristics (Garcia Godoy et al., 2020), especially during the early phases of the pandemic (Chou et al., 2021; Hill et al., 2022; Mediavilla et al., 2021).

One important barrier to protecting HCWs at the beginning of the pandemic was insufficient access to personal protective equipment (PPE) and inadequate use of PPE (De Serres et al., 2020; Nguyen et al., 2020). In Québec, a cross-sectional survey estimated that 40% of HCWs had to provide care to SARS-CoV-2-infected patients without proper PPE due to shortages (De Serres et al., 2020). In Ontario, a survey estimated that 38% of HCW respondents had less than half of their PPE needs met during the first wave (Oudyk et al., 2020). Another study showed that one’s capacity to properly remove PPE was associated with having recently received an IPAC (infection prevention and control) training (Piché-Renaud et al., 2020).

Certain HWC roles have been associated with increased risk for SARS-CoV-2 infection (Gómez-Ochoa et al., 2020; Sims et al., 2020), including orderlies, registered nurses (RNs) and licensed practical nurses (LPNs). In Montréal, long-term care facilities (LTCFs) were hotspots for COVID-19 cases and deaths; 75% of COVID-19-related deaths were LTCF residents as of March 2021 (St. Onge, 2021). In Québec, a study of SARS-CoV-2-infected HCWs estimated that half (48%) of these HCWs had worked in a LTCF, and 51% of participants believed that they had been infected at work (De Serres et al., 2020).

A greater risk of SARS-CoV-2 infection among Black, Indigenous, and People of Colour (BIPOC) HCWs has been shown in many countries (Greenaway et al., 2020; Nguyen et al., 2020; Turcotte & Savage, 2020). In Québec, the data collected during COVID-19 case investigations did not include race, ethnicity, or migration status. Therefore, it is difficult to estimate the proportion of infected BIPOC or immigrant HCWs, although there is some evidence from Québec and Canada that supports this association with a higher risk of SARS-CoV-2 infection (Adrien et al., 2020; Gouvernement du Canada, 2021a, 2021b). In Canada, it has been estimated that 34% of HCWs identify as BIPOC, compared to 21% in other sectors (Statistics Canada, 2021).

At the request of the Montréal Public Health Department, a study of HCWs who were infected with SARS-CoV-2 during the first wave was conducted. The aim was to describe the sociodemographic profile of the HCWs infected by SARS-CoV-2 in Montréal as well as certain workplace and household characteristics related to SARS-CoV-2 infections such as access to PPE, PPE training, and following isolation recommendations. Secondary objectives were to estimate the associations between PPE access and training, and following self-isolation with different sociodemographic and workplace characteristics.

Methods

Study population

This cross-sectional study was conducted among HCWs who resided in Montréal, tested positive for SARS-CoV-2 between March 2 and July 22, 2020, and had worked during the 14-day period preceding their positive SARS-CoV-2 test result. Potential participants were randomly selected from the Québec provincial database of all SARS-CoV-2 cases reported to public health, including an identifier for HCWs. In this database, date of reporting and basic sociodemographic characteristics, such as age and sex, were included. Potential participants were selected using proportionate stratified random sample based on the reported date of SARS-CoV-2 infection to public health (March to April 2020 vs. May to July 2020) and their neighbourhood of residence. The distribution of age, sex, neighbourhood, and date of RT-PCR-positive result was compared to that of all public health–reported infected HCWs during the first wave in Montréal (n = 6174) to ensure representativity of the sample.

Sample size

The sample size was calculated based on having sufficient power to estimate the prevalence of insufficient PPE access at work in our study population. Previous surveys estimated that only 41% of HCWs in Manitoba (n = 1877) self-reported they had received sufficient PPE (Canadian Union of Public Employees (CUPE)—Manitoba Regional Office, 2020), and 38% of HCWs in Ontario (n = 4245) self-reported that less than half of their PPE needs were met (Oudyk et al., 2020). Assuming a prevalence of 40%, with an accuracy of 5% and a confidence level of 95%, a total of 370 participants was needed (Naing et al., 2006). Anticipating a participation rate of 50%, the final required sample size was estimated at 740. Selection of an additional 300 potential participants (total sample size = 1040) was necessary after all initial participants had been contacted, given the higher than anticipated non-response rate.

Questionnaire

The questionnaire included questions on sociodemographic characteristics, job characteristics, COVID-19 symptoms, preventive measures at work (e.g., training and PPE), and following self-isolation instructions while infectious (most of the time or always vs. sometimes or never). The questionnaire was tested on the investigation team and was translated from French into seven languages: English, Spanish, Creole, Arabic, Punjabi, Hindi, and Tagalog.

Data collection

Data collection took place by telephone between August 3 and August 29, 2020. During the initial call, the objectives were briefly described to participants. If they were interested in participating, verbal consent was obtained and documented. Participants were able to indicate their preferred language for the interview and, if possible, the interview was completed during the initial call; otherwise, a follow-up call was scheduled. A minimum of three calls were made before considering a participant as not reached.

Data analysis

Data were collected using Microsoft Form and exported as a csv file for analysis. Categorization of BIPOC (or not) was based on the visible minority categories as described by Statistics Canada in addition to the category of “Indigenous” (Statistics Canada, 2015). Participants were also categorized by nativity (born in or outside Canada) and by their migration status (Canadian citizen, permanent resident, or other status). Descriptive analyses were performed to describe the sample, as well as chi-square tests at a 0.05 significance threshold when relevant. Differences between participants born in Canada and outside of Canada were of particular interest, as the state of epidemiological knowledge indicated disparities in working conditions at this level. In addition, the difference in proportions was explored for variables of interest in the context of the study, namely gender, education, and low income (less than $40,000). These variables were relevant to the Montréal Public Health Department teams involved in the project who were interested in social inequalities in health. Three separate multivariable log binomial regressions were developed to estimate associations between three different outcomes: having completed IPAC SARS-CoV-2 infection prevention training, perception of having sufficient access to PPE, and following self-isolation recommendations with certain workplace and sociodemographic characteristics. Covariates were selected based on expert advice and scientific literature. For all 3 outcomes, the dependent variable was regressed on years of experience, age, identifying as BIPOC, work shift, month of diagnosis, household income, job title, and country of birth. For the outcome “perception of having sufficient access to PPE”, the model included having received IPAC training as a covariate. For the outcome “following self-isolation instructions”, the model included having received IPAC training, having COVID-19 symptoms, and duration of self-isolation as covariates. To select the 3 final regression models, various models integrating interaction terms were explored and AIC was used to compare the starting models and models with interaction for all three outcomes. Multicollinearity of covariates was assessed using variance inflation factor. Residual deviance was used to assess the model fit. Regression analyses were conducted using R software.

Ethics

Ethics approval was obtained from the Science and Health Research Ethics Committee (CERSES) of the University of Montréal (CERSES-20–105-D). All personal information collected in the consent form and the questionnaire remained confidential and all collected data were depersonalized.

Results

Population

Between August 3 and August 29, 2020, 918 HCWs were contacted and 370 (40%) agreed to participate (Fig. 1). For the HCWs who answered the phone call from the research team (n = 490), there was a 76% participation rate.

Fig. 1.

Details of the sample that was contacted, agreed to participate, was excluded or refused to participate, and the reasons for exclusion or refusal

Most participants completed their interview in French (82%), 12% in English, and 5% in Spanish, and some participants completed their interview in Creole (n = 2), in Arabic (n = 1), in Tagalog (n = 1), in Punjabi (n = 1), and in Hindi (n = 1). One interview was conducted in Mandarin with the support of an interpreter. No participant was excluded because of their preferred language. The distribution of participants by neighbourhood of residence is presented in Appendix Table 7, as well as the distribution of the stratified random sample (prior to recruitment) and the total reported cases of HCWs in Montréal. A significant difference in geographic distribution was observed between participants and HCWs SARS-CoV-2 cases in Montréal reported to public health in the Québec provincial database during the first wave (χ² = 55.03, df = 32, p = 0.03). However, this difference was non-significant (χ² = 41.77, df = 30, p = 0.12) when Anjou and Montréal-Nord are excluded, suggesting an under-representation of these neighbourhoods in our study population, but an otherwise representative distribution for neighbourhoods in Montréal. Some key sociodemographic and work-related characteristics (age, gender, access to provincial healthcare (RAMQ), and workplace) were also compared between participants, those who refused and those who were not reached (Appendix Table 8). HCWs who were not reached, refused, or agreed to participate were not significantly different except for their workplace characteristics, with a lower proportion of unreached HCWs working in hospitals or LTCFs compared to participants (χ² = 34.29, df = 16, p = 0.005).

Table 7.

Distribution of infected Montréal HCWs (n = 6174), participants (n = 370), refusal (n = 120), and unreached HCWs (n = 416) by neighbourhood. The proportions of HCWs who refused to participate (n = 120) and who were contacted but not reached (n = 416) are related to the total HCWs reached (n = 490) and total participants and HCWs not reached (n = 786), respectively

Neighbourhood	All reported HCW cases in Montréal (1st wave) (n = 6174)		Participants (n = 370)		Refusal (n = 120)		Unreached (n = 416)
	n	%	n	%	n	%	n	%
Ahuntsic-Cartierville	523	8.5	36	9.7	10	2.0	34	4.3
Anjou	244	4	4	1.1	4	0.8	20	2.5
Baie-d’Urfé	1	0	0	0	0	0	1	0.1
Beaconsfield	13	0.2	0	0	0	0	2	0.3
Côte-des-Neiges-Notre-Dame-de-Grâce	424	6.9	28	7.6	6	1.2	28	3.6
Côte-Saint-Luc	51	0.8	3	0.8	1	0.2	4	0.5
Dollard-des-Ormeaux	93	1.5	2	0.5	4	0.8	7	0.9
Dorval	26	0.4	1	0.3	0	0	0	0
Hampstead	11	0.2	2	0.5	1	0.2	1	0.1
Kirkland	12	0.2	1	0.3	2	0.4	0	0
L’Île-Bizard-Sainte-Geneviève	36	0.6	4	1.1	0	0	1	0.1
Lachine	127	2.1	7	1.9	4	0.8	6	0.8
Lasalle	268	4.3	19	5.1	4	0.8	16	2.0
Le Plateau-Mont-Royal	152	2.5	12	3.2	1	0.2	9	1.1
Le Sud-Ouest	186	3	9	2.4	3	0.6	15	1.9
Mercier-Hochelaga-Maisonneuve	497	8	34	9.2	7	1.4	37	4.7
Mont-Royal	36	0.6	3	0.8	1	0.2	1	0.1
Montréal-Est	13	0.2	2	0.5	0	0	1	0.1
Montréal-Nord	695	11.3	33	8.9	18	3.7	50	6.4
Montréal-Ouest	7	0.1	0	0	1	0.2	2	0.3
Outremont	28	0.5	2	0.5	0	0	2	0.3
Pierrefonds-Roxboro	167	2.7	10	2.7	5	1.0	12	1.5
Pointe-Claire	46	0.7	2	0.5	0	0	3	0.4
Rivière-des-Prairies-Pointe-aux-Trembles	632	10.2	35	9.5	15	3.1	42	5.3
Rosemont-La-Petite-Patrie	405	6.6	24	6.5	9	1.8	26	3.3
Saint-Laurent	246	4	17	4.6	6	1.2	15	1.9
Saint-Léonard	362	5.9	24	6.5	5	1.0	26	3.3
Sainte-Anne-de-Bellevue	4	0.1	2	0.5	1	0.2	1	0.1
Senneville	1	0.02	1	0.3	0	0	0	0
Verdun	151	2.4	5	1.4	5	1.0	12	1.5
Ville-Marie	140	2.3	11	3	2	0.4	6	0.8
Villeray-Saint-Michel-Parc-Extension	556	9	36	9.7	5	1.0	34	4.3
Westmount	21	0.3	1	0.3	0	0	2	0.3

Table 8.

Distribution of infected Montréal HCW participants (n = 370), refusal (n = 120), and unreached HCWs (n = 416) by age group, gender, access to RAMQ, and workplace

Characteristics	Participants (n = 370)		Refusal (n = 120)		Unreached (n = 416)		χ2	df	p-value
	n	%	n	%	n	%
Age							6.10	6	0.413
≤ 20	4	1.1	4	3.3	4	1			-
21–40	161	43.5	52	43.3	171	41.1			-
41–60	185	50	59	49.2	225	54.1			-
≥ 61	19	5.1	5	4.2	16	3.8			-
Gender							0.39	2	0.821
Woman	274	74.1	91	75.8	315	75.7			-
Man	96	25.9	29	24.2	100	24			-
Access to RAMQ							0.26	2	0.879
Yes	357	96.5	116	96.7	404	97.1			-
No	13	3.5	4	3.3	12	2.9			-
Workplace							34.29	16	0.005*
LTCF	205	55.4	53	44.2	185	44.5			-
Hospital	123	33.2	44	36.7	120	28.8			-
RPA	43	11.6	10	8.3	33	7.9			-
RI-RTF	7	1.9	0	0	5	1.2			-
Readaption centre	5	1.4	0	0	2	0.5			-
Laboratory	1	0.3	1	0.8	5	1.2			-
CLSC	0	0	2	1.7	10	2.4			-
First respondent	0	0	1	0.8	5	1.2			-
Other	55	14.9	30	25	82	19.7			-

Sociodemographic and personal characteristics

Sociodemographic characteristics of the participants are presented in Table 1. The majority were born outside of Canada (65%), and of these, 73% were Canadian citizens. The median number of years since arriving in Canada was 10 years (IQR = 13). The distribution of sociodemographic factors was significantly different between participants born in Canada and those born outside of Canada, except for gender (Table 1).

Table 1.

Sociodemographic characteristics of the participants (n = 370)

Characteristics	All		Born in Canada (n = 129; 34.9%)		Born outside of Canada (n = 241; 65.1%)		χ2	df	p-value
	Number	%	Number	%	Number	%
Age, years							8.68	3	0.02
≤ 20	4	1.1	2	1.6	2	0.8
21–40	161	43.5	69	53.5	92	38.2
41–60	185	50.0	53	41.1	132	54.8
≥ 61	19	5.1	5	3.9	14	5.8
Gender							0	1	1
Woman	274	74.1	96	74.4	178	73.9
BIPOC participants							181.85	4	< 0.001
Not a BIPOC participant	135	36.5	107	82.9	28	11.6
Black	131	35.4	13	10.1	118	49.0
Arab	45	12.2	2	1.6	43	17.8
Latino/a/x	27	7.3	4	3.1	23	9.5
Other*	31	8.4	4	3.1	27	11.2
Immigration status							-	-	-
Canadian citizen	-	-	-	-	176	73
Permanent resident	-	-	-	-	43	17.8
Other**	-	-	-	-	20	8.4
Education level							23.58	4	< 0.001
No diploma or Elementary school	1	0.3	0	0	1	0.4
High school	67	18.1	15	11.6	52	21.6
CÉGEP	104	28.1	29	22.5	75	31.1
Undergraduate degree	126	34.1	44	34.1	82	34
Graduate degree	69	18.6	40	31	29	12
Preferred not to answer	3	0.8	1	0.8	2	0.8
Annual household income							37.06	5	> 0.001
< $20,000	13	3.5	4	3.1	9	3.7
$20,000 to < $40,000	67	18.1	13	10.1	54	22.4
$40,000 to < $60,000	75	20.3	20	15.5	55	22.8
$60,000 to < $80,000	56	15.1	18	14	38	15.8
$80,000 to < $100,000	31	8.4	14	10.9	17	7.1
≥ $100,000	58	15.7	39	30.2	19	7.9
Preferred not to answer	70	18.9	21	16.3	49	20.3

^*South Asian; Southeast Asian; West Asian; Filipino; Indigenous

^**Asylum seeker; student visa; work visa; visitor visa; awaiting status (other than refugee)

The participants were largely women (74%) between 21 and 60 years old (Table 2). The proportion of men (26%) in our study population was not significantly different from the proportion of male HCWs in the cases reported in Montréal (25%; χ² = 0.14, df = 1, p = 0.71) but was significantly different when compared to the proportion of male HCWs in the province of Québec (18%; χ² = 15.11, df = 1, p < 0.001). Most participants had post-secondary education, with the proportion of HCWs with a graduate degree significantly higher among those born in Canada (31% vs. 12%; χ² = 18.71, df = 1, p < 0.001). The distribution of average annual household income varied greatly between participants, with 26% of HCWs born outside of Canada having a household income of less than $40,000, compared to 13% of Canadian-born participants (χ² = 7.58, df = 1, p = 0.006).

Table 2.

Type of positions and related characteristics of the participants (n = 370)

Job characteristics	Number	%
Positions
Orderly	148	40.0
Registered nurse	74	20.0
Licensed practical nurse	37	10.0
Maintenance	17	4.6
Support aide	15	4.1
Physician	14	3.8
Manager	12	3.2
Administrative agent	10	2.7
Other1	43	1.6
Work shift
Day	208	56.2
Evening	87	23.5
Night	37	10.0
Rotation shift schedule2	31	8.4
Unpredictable (call list)3	7	1.9
Job status
Permanent position	256	69.2
On call	57	15.1
Temporary/replacement	18	4.9
COVID-19 specific program4	14	3.8
Mobile team	13	3.5
Private healthcare worker	13	3.5
Other5	11	3.0

¹Other position included physiotherapist, professor/researcher, food services worker, security agent, family helper, occupational therapist, transport agent, social worker, medical electrophysiology technician, criminologist, respiratory therapist, kinesiologist, dietician, speech therapist, recreation technician

²Rotation: Work shifts vary according to a previously determined schedule

³Unpredictable: Work shifts vary but are not pre-determined

⁴Je contribue, Red Cross, Army, etc.

⁵Temporary relieved from another department/facility, daily worker, and self-employed

Job characteristics

More than half of the interviewed HCWs (55%) had worked in a LTCF between March and August 2020, 34% in a hospital setting, 12% in a retirement home (RH), and 9% in a clinic. During that time, most participants worked in one facility (75%), with 25% having worked in more than one (16% worked in 2 facilities, 6% in 3, and 4% in 4 or more). Most participants occupied positions that required contact with patients, such as orderlies, RNs, LPNs, support aides, and doctors (Table 2). Among orderlies, the group most represented among the participants (40%; 149/370), 81% were born outside of Canada (121/149), 82% were women (122/149), and 80% identified as BIPOC (119/149). More so, orderlies who identified as BIPOC, women, and born outside of Canada represented 26% of all the HCW participants (95/370).

Preventive measures at work

Infection prevention and control training specific to COVID-19

A large portion of participants (30%) reported that they had not received any SARS-CoV-2 IPAC training, including 37% of the orderlies, 30% of LPNs, and 23% of RNs (Fig. 2). Among HCWs who received at least one training session (67%), most reported that their IPAC practices improved following their training (87%) and that they felt equipped to avoid contaminating themselves or others in their workplace (81%).

Fig. 2.

Percentage of participants who received none, one, or more than one training by position (n = 323). Only positions with at least 10 participants were included

Among participants, there was an association between job title and IPAC training for orderlies; the odds of having received training specific to SARS-CoV-2 infection prevention were 0.48 (95% CI 0.31–1.06; p = 0.08) for orderlies compared to RNs and LPNs (Table 3). A similar association was observed for HCWs born outside of Canada, whose odds of having received IPAC training were 0.51 (95% CI 0.24–1.08; p = 0.08) compared to HCWs born in Canada. In contrast, the odds of having received IPAC training for participants who did not identify as BIPOC were 0.58 (95% CI 0.19–0.9; p = 0.03) compared to those who identified as BIPOC. This result is difficult to interpret and may be associated with a lower number of non-BIPOC participants. No interaction terms improved exhaustivity in this model.

Table 3.

Odds ratios of HCW participants having received IPAC training specific to COVID-19 for certain personal and job characteristics

Variables*	OR	95% CI	p-value
Years of experience (per year)	0.92	0.75–1.13	0.43
BIPOC status
Identifying as BIPOC (n = 234)	Reference
Non-BIPOC (n = 135)	0.58	0.19–0.9	0.03
Country of birth
Canada (n = 129)	Reference
Outside of Canada (n = 241)	0.51	0.24–1.08	0.08
Month of infection
April or March (n = 190)	Reference
May (n = 155)	1.52	0.93–2.52		0.1
June or July (n = 25)	1.66	0.62–5.0		0.33
Job title
Registered nurse or licensed practical nurse (n = 111)	Reference
Orderly (n = 148)	0.58	0.31–1.06	0.08
Other job with direct contact with patients (n = 62)	1.06	0.48–2.45	0.88
Other job without direct contact with patients (n = 49)	0.52	0.23–1.17	0.11
Work shift
Day shift (n = 208)	Reference
Evening and night shift (n = 124)	0.66	0.39–1.11	0.12
Rotation shift schedule (n = 38)	0.78	0.35–1.84	0.56

The variables shown in bold are those that had a significant OR for a p-value of less than 0.05, as well as the non-significant ORs under a p-value threshold of 0.1

^*The model was also adjusted for age and income. The variables shown in bold are those that had a significant OR for a p-value of less than 0.05, as well as the non-significant ORs under a p-value threshold of 0.1

Access to personal protective equipment

About half (52%; 193/370) of participants reported they did not have access to sufficient PPE to protect themselves against SARS-CoV-2 infection. The proportion of BIPOC participants was higher (χ² = 5.07, df = 1, p = 0.02) in the group who did not have access to sufficient PPE (69%; 133/193) than in the group who did have access to sufficient PPE (58%; 102/177). Similarly, the proportion of participants who were born outside of Canada and who did not have access to sufficient PPE (71%; 137/193) was higher (χ² = 5.55, df = 1, p = 0.02) than in the group who did have access to sufficient PPE (59%; 105/177). The proportion of women was similar in both groups (χ² = 0.34, df = 1, p = 0.56), with 76% (134/177) who had access to sufficient PPE and 73% (141/193) in the group who did not. However, among women who had sufficient access (n = 134), only 24% (32/134) identified as Black, compared to 45% (64/141) identifying as Black in the group of women who did not have sufficient access to PPE (χ² = 13.99, df = 1, p < 0.01). More so, when looking at intersecting identities, women who identified as BIPOC and were born outside of Canada represented half (50%; 96/193) of the participants who did not have access to sufficient PPE, which is significantly higher (χ² = 9.50, df = 1, p < 0.01) compared to 34% (60/177) in the group who did have access to sufficient PPE. In addition, among those who did not have access to sufficient PPE, more participants (χ² = 13.55, df = 1, p < 0.01) worked in LTCFs (64%; 124/193) than among the group who had access (45%; 80/177). The most frequently identified non-available PPE was N95 respirators, isolation gowns, and face shields (Table 4).

Table 4.

Unavailable personal protective equipment (PPE) (n = 194)

Unavailable personal protective equipment	LTCF (n = 124)		Other facilities (n = 70)		Total (n = 194)
	n	%	n	%	n	%
N95 respirators	96	77.4	51	72.9	147	75.8
Gowns/isolation gowns	88	71.0	48	68.6	136	70.1
Face shields	91	73.4	41	58.6	132	68.0
Procedural/surgical masks	72	58.1	39	55.7	111	57.2
Protective glasses	78	62.9	28	40.0	106	54.6
Gloves	44	35.5	21	30.0	65	33.5
Other PPE*	19	15.3	11	15.7	30	15.5

^*Hand sanitizer, soap, disinfectant, cleaning products, hair cap, work shoes, trash cans

In multivariable regression analysis, there was a significant association between access to PPE and shifts worked; the odds of having sufficient access to PPE for evening and night shift workers were 0.50 (95% CI 0.30–0.83; p < 0.01) compared to day shift workers (Table 5). There was also an association with the month of infection, with those infected in May, June, or July having greater odds of having access to enough PPE compared to those infected in March or April. AIC did not invite to integrate any interaction terms in the regression equation.

Table 5.

Odds ratios of HCW participants having sufficient access to personal protective equipment for certain personal and job characteristics

Variables*	OR	95% CI	p-value
Years of experience (per year)	1.20	0.98–1.47	0.07
BIPOC status
Identifying as BIPOC (n = 234)	Reference
Non-BIPOC (n = 135)	1.17	0.56–2.44	0.67
Country of birth
Canada (n = 129)	Reference
Outside of Canada (n = 241)	1.21	0.58–2.57	0.62
Month of infection
April or March (n = 190)	Reference
May (n = 155)	1.91	1.19–3.08		0.007
June or July (n = 25)	4.82	1.83–13.93		0.002
Job title
Registered nurse or licensed practical nurse (n = 111)	Reference
Orderly (n = 148)	0.82	0.46–1.46	0.5
Other job with direct contact with patients (n = 62)	1.21	0.59–2.48	0.61
Other job without direct contact with patients (n = 49)	1.61	0.74–3.53	0.23
Work shift
Day shift (n = 208)	Reference
Evening and night shift (n = 124)	0.50	0.30–0.83	0.008
Rotation shift schedule (n = 38)	0.71	0.32–1.57	0.4
IPAC COVID-19 training
Did not receive training (n = 117)	Reference
Received training (n = 246)	1.51	0.92–2.52		0.1

The variables shown in bold are those that had a significant OR for a p-value of less than 0.05, as well as the non-significant ORs under a p-value threshold of 0.1

^*The model was also adjusted for age and income. The variables shown in bold are those that had a significant OR for a p-value of less than 0.05, as well as the non-significant ORs under a p-value threshold of 0.1

Following isolation recommendations while infectious

Among participants who experienced symptoms compatible with COVID-19 before being diagnosed (n = 325), 23% continued to work despite their symptoms. Those who continued to work did so for various reasons: they felt well enough (60%), their employer requested them to work (20%), they did not know they should not work (9%), and they thought they were protecting themselves well enough to not infect others at work (9%). Following their positive PCR-RT test result, almost all (99%) who were symptomatic did not work and followed the instructions to stay home during their self-isolation period. However, 14 symptomatic HCWs reported that their employer requested that they work despite their SARS-CoV-2 positive test result.

In multivariable regression analysis, following all isolation recommendations was negatively associated and statistically significant with having a position other than RN, LPN, or orderly that required direct contact with patients (OR 0.37, 95% CI 0.15–0.92) and also with having a household income of $100,000 and over (OR 0.23, 95% CI 0.09–0.71). Following isolation recommendations was positively associated and statistically significant with being diagnosed in May, June, or July compared to March or April (Table 6). There was also a significant association with IPAC training, with the odds of following self-isolation recommendations being twice as high for those who had received training (OR 1.9, 95% CI 1.0–3.61). The addition of an interaction term was also not needed in this model.

Table 6.

Odds ratios of HCW participants following self-isolation recommendations for certain personal and job characteristics

Variables*	OR	95% CI	p-value
Years of experience (per year)	0.89	0.68–1.16	0.4
BIPOC status
Identifying as BIPOC (n = 234)	Reference
Non-BIPOC (n = 135)	1.11	0.45–2.79	0.82
Country of birth
Canada (n = 129)	Reference
Outside of Canada (n = 241)	1.21	0.49–2.99	0.67
Month of infection
April or March (n = 190)	Reference
May (n = 155)	1.94	1.04–3.68		0.04
June or July (n = 25)	4.55	1.34–17.35		0.02
Job title
Registered nurse or licensed practical nurse (n = 111)	Reference
Orderly (n = 148)	0.96	0.44–2.09	0.93
Other job with direct contact with patients (n = 62)	0.37	0.15–0.92	0.03
Other job without direct contact with patients (n = 49)	0.73	0.27–2.03	0.54
Work shift
Day shift (n = 208)	Reference
Evening and night shift (n = 124)	1.0	0.53–1.9	1.0
Rotation shift schedule (n = 38)	1.47	0.52–4.59	0.48
IPAC COVID-19 training
Did not receive training (n = 117)	Reference
Received training (n = 246)	1.9	1.0–3.61		0.05
COVID-19 symptoms
Not having COVID-19 symptoms (n = 44)	Reference
Having COVID-19 symptoms (n = 326)	0.86	0.34–2.07		0.74
Duration of self-isolation
2–4 weeks (n = 166)	Reference
1–2 weeks (n = 99)	0.77	0.39–1.52		0.45
> 4 weeks (n = 105)	2.15	1.05–4.57		0.04

The variables shown in bold are those that had a significant OR for a p-value of less than 0.05, as well as the non-significant ORs under a p-value threshold of 0.1

^*The model was also adjusted for age and income. The variables shown in bold are those that had a significant OR for a p-value of less than 0.05, as well as the non-significant ORs under a p-value threshold of 0.1

Discussion

Our study provides a more complete portrait of HCWs infected with SARS-CoV-2 during the first wave in Montréal. Most participants self-identified as BIPOC and hold an orderly or RN position. We found differences in terms of IPAC training, access to PPE, and following isolation recommendations. Our findings suggest a negative association between having received IPAC training and working as an orderly or being born in Canada, and a negative association between having sufficient access to PPE and working evening or night shifts. We found a positive association between following isolation recommendations and being diagnosed later in the pandemic or having received IPAC training.

Working conditions at the beginning of the pandemic

Access to training has been identified as a critical aspect of proper PPE use and safe working conditions (Gómez-Ochoa et al., 2020). We found that not having received IPAC training was associated with working as an orderly and HCW’s place of birth. Compared to nurses, orderlies had 0.42 lower odds of having had a training, thereby reinforcing the precariousness often associated with this occupation.

Insufficient access to PPE was associated with month of infection and working evening or night shifts. Workers who were infected in June or July had almost 5 times the odds of having had sufficient access to PPE compared to workers infected earlier in March or April. This suggests a lack of preparation of healthcare facilities early in the pandemic in terms of PPE, but it also demonstrates improvement over time as the pandemic unfolded. Insufficient access to PPE did not differ across job titles. Similarly, in Spain, equivalent rates of insufficient access to PPE were observed across different healthcare settings (Mediavilla et al., 2021). In our study, we did have a large proportion of racialized women who did not have sufficient access to PPE or IPAC training. This could suggest gender and race disparities, which has been shown by other studies on HCWs during the pandemic in terms of unequal distribution and access to resources (Paremoer et al., 2021).

Time differences were also detectable in adhesion to self-isolation instructions. HCWs infected in June or July had 4.55 the odds of following self-isolation instructions compared to workers infected in April or March. This could reflect an improved understanding of SARS-CoV-2 transmission. Interestingly, HCWs who had IPAC training had almost twice the odds of following self-isolation instructions compared to those who did not receive the training, which supports the importance of IPAC training in the adoption of safe practices.

Sociodemographic portrait of Montréal HCWs infected by SARS-CoV-2

Our study did not identify SARS-CoV-2 infection risk factors for HCWs, as it did not include non-infected HCWs. However, our results raise interesting points of discussion and suggest that at the intersection of several identities, including nativity, job title, and possibly identifying as BIPOC, there were likely COVID-19-related disparities among HCWs in Montréal. Other studies have found increased risk of BIPOC HCWs being infected by SARS-CoV-2 (Nguyen et al., 2020; Shields et al., 2020). A study of HCWs in Québec also found an association between COVID-19 seropositivity and being of Black or Hispanic ethnicity (Brousseau et al., 2021). Among our BIPOC participants, most worked as orderlies (71%), RNs (18%), or LPNs (10%), which are positions that have been associated with a higher risk of SARS-CoV-2 positivity in HCWs compared to physicians (Brousseau et al., 2021). The nature of the jobs held by most BIPOC participants involved close contact with patients, and this may partly explain these results. However, complex social factors, such as living and housing conditions and population density, have also been identified as risk factors for SARS-CoV-2 infection of BIPOC HCWs in Montréal (Brousseau et al., 2021).

Limitations

One important study limitation was our low response rate, which may have resulted in biased estimates although we found that our study population did not significantly differ from the non-responders by age, gender, and access to provincial healthcare (RAMQ). The representativeness of our study population was assessed by comparing certain sociodemographic characteristics to the provincial database of all SARS-CoV-2 HCW cases in Montréal. This comparison was limited by the lack of data on race and ethnicity in the provincial database. Another limitation was that our analysis was based on self-reported data, and particularly the perception of inadequate access to PPE, which likely resulted in information bias including recall bias, which we would anticipate would be non-differential and likely biasing the estimates towards the null. Also, the regression analyses were limited by the small sample size and therefore influencing the precision of the estimates, as indicated by the wide confidence intervals.

Conclusion

Overall, our findings suggest that there were challenges in having access to IPAC training and PPE equipment during the first wave of the pandemic, particularly for certain groups of HCWs in Montréal. Collecting sociodemographic data, including race, ethnicity, nativity, and immigration status, in case investigations at the onset of future pandemics would allow for the early identification of vulnerable groups and to adapt preventive and protective measures accordingly. In addition, governments and employers should be prepared to provide IPAC training and ensure equitable access to PPE across all groups of HCWs in future health crises, while ensuring PPE and training for those who are at highest risk of exposure.

Contributions to knowledge

What does this study add to existing knowledge?

• In Montréal, a large proportion of healthcare workers (HCWs) infected with SARS-CoV-2 during the first wave in 2020 reported not having had access to any SARS-CoV-2 infection prevention and control (IPAC) training, nor to enough personal protective equipment (PPE).

• Among HCWs infected with SARS-CoV-2 during the first wave in Montréal who participated in the study, differences in access to SARS-CoV-2 IPAC training and PPE were observed based on nativity and job type.

What are the key implications for public health interventions, practice, or policy?

• It is relevant and necessary to include race, ethnicity, and immigration status in the collection of sociodemographic data, at the outset of epidemiologic outbreaks and case investigations to identify groups most at-risk, to inform prevention and control measures appropriately.

• Ensuring equitable access to IPAC training and PPE for all groups of HCWs at the onset of a health crisis should be a public health priority.

Appendix

Tables 7 and 8

Author contributions

Fortin, Munoz-Bertrand, and Zinszer contributed to the study conception and design. Material preparation was performed by Fortin, Narasiah, Munoz-Bertrand, and Zinszer. Data collection and analysis were performed by Fortin, Saucier, Yuan, Ante, and Zinszer. The first draft of the manuscript was written by Fortin and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

This study was funded by the Centre de recherche en santé publique (CReSP) with funds available for projects conducted in collaboration with Montréal Public Health Department (Direction régionale de santé publique de Montréal).

Availability of data and material

Not applicable.

Code availability

Not applicable.

Declarations

Ethics approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the Science and Health Research Ethics Committee (CERSES) of the University of Montréal (CERSES-20–105-D) and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Consent to participate and for publication

Informed consent was obtained from all individual participants included in the study.

Conflict of interest

The authors declare no competing interests.