Occupation and SARS-CoV-2 seroprevalence studies: a systematic review

Objective To describe and synthesise studies of SARS-CoV-2 seroprevalence by occupation prior to the widespread vaccine roll-out. Methods We identified studies of occupational seroprevalence from a living systematic review (PROSPERO CRD42020183634). Electronic databases, grey literature and news media were searched for studies published during January–December 2020. Seroprevalence estimates and a free-text description of the occupation were extracted and classified according to the Standard Occupational Classification (SOC) 2010 system using a machine-learning algorithm. Due to heterogeneity, results were synthesised narratively. Results We identified 196 studies including 591 940 participants from 38 countries. Most studies (n=162; 83%) were conducted locally versus regionally or nationally. Sample sizes were generally small (median=220 participants per occupation) and 135 studies (69%) were at a high risk of bias. One or more estimates were available for 21/23 major SOC occupation groups, but over half of the estimates identified (n=359/600) were for healthcare-related occupations. ‘Personal Care and Service Occupations’ (median 22% (IQR 9–28%); n=14) had the highest median seroprevalence. Conclusions Many seroprevalence studies covering a broad range of occupations were published in the first year of the pandemic. Results suggest considerable differences in seroprevalence between occupations, although few large, high-quality studies were done. Well-designed studies are required to improve our understanding of the occupational risk of SARS-CoV-2 and should be considered as an element of pandemic preparedness for future respiratory pathogens.


3, line 39
Eligibility criteria 6 Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rationale.

4, lines 39-45
Information sources 7 Describe all information sources (e.g., databases with dates of coverage, contact with study authors to identify additional studies) in the search and date last searched.

4, lines 39-40
Search 8 Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated.

Suppl. File 2
Study selection 9 State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis).

S3. Detailed eligibility criteria
This study included eligible studies from the SeroTracker database. Eligibility criteria for the database and also for this review specifically are outlined below:

Eligibility criteria for inclusion in SeroTracker database
Eligibility criteria for inclusion in this review Study performed serologic testing to determine the prevalence of SARS-CoV-2 antibodies in a human population over a specified time period.
Reported sample size, sampling date, location and prevalence.
Study published between January 01 and December 31, 2020.
Article in English or French or could be fully extracted using machine translation.
Article written in English or French or machine-translatable using Google Translate.
Article did not report identical information to previously included studies (peer-reviewed studies were prioritised over news stories and pre-prints where available).
Reported seroprevalence data that could be fit into the 23 major SOC 2010 occupation categories or combined categories for healthcare workers, first-responders or unemployed persons.
Studies that only reported seroprevalence for mixed occupation groups or workplaces rather than specific occupations (e.g., "hospital staff") were excluded.
Studies conducted only in people previously diagnosed with COVID-19 (molecular or antigen testing, or clinical or self-assessment).
Seroprevalence estimates did not include people <18 years (i.e., possibly affected by COVID-19 exposure at school, which could impact occupational seroprevalence estimates).
Cohort or cross-sectional design (case reports, case-control studies, trials, and reviews were excluded, as were dashboards not associated with a defined serology study).

S4. Tool for assessing study risk of bias
Item 1: Was the sample frame appropriate to address the target population?

Yes
Sample frame described and it approximated the target population No Sample frame did not approximate the target population (e.g., blood donors do not represent general population, doctors do not represent all health care providers) Exclude Sample frame not described

*Notes
The term "target population" should not be taken to infer every individual from everywhere or with similar disease or exposure characteristics. Instead, give consideration to specific population characteristics in the study, including age range, gender, morbidities, medications, and other potentially influential factors. For example, a sample frame may not be appropriate to address the target population if a certain group has been used (such as those working for one organisation, or one profession) and the results then inferred to the target population (i.e. working adults). A sample frame may be appropriate when it includes almost all the members of the target population (i.e. a census, or a complete list of participants or complete registry data).

Item 2: Were study participants recruited in an appropriate way?
Yes Probability sampling method (simple or stratified random) or entire sample (e.g., an entire town) was used Where n = sample size; Z = Z statistic for level of confidence (95%); P = expected prevalence (2.5% WHO global estimate); d = precision (1.25%) In cases where the sample size calculation was provided and the required sample for 80% power was below our threshold (n<599), this item was marked as yes.
Item 4: Were the study subjects and setting described in detail?

Yes
Average age and distribution of gender/sex provided No Neither age or gender/sex is provided, or only one of age and gender/sex is provided Yes Does all of the following: corrects for population characteristics or the sample is somewhat representative of the population (probability sampling), corrects for test characteristics), and provides the information necessary to determine the numerator, denominator, prevalence estimate, and confidence interval.

No
Does not correct for population characteristics and the sample is not likely representative of the population (non-probability sampling), does not correct for test or provide the information necessary to correct for test characteristics, or does not provide the information necessary to determine the numerator, denominator, prevalence estimate, and confidence interval.
Item 9: Was the response rate adequate, and if not, was the low response rate managed appropriately?

Yes
Response rate > 60% or the demographics of the sample were a reasonable match to those of the target population [5] No Response rate < 60% and the demographics of the sample were not a reasonable match to those of the target population Unclear Response rate not provided and it was unclear if the demographics of the sample differed from the target population Item 10: Overall risk of bias

Low
The estimates are very likely correct for the target population. To obtain a low risk of bias classification, all criteria must be met or departures from the criteria must be minimal and unlikely to impact on the validity and reliability of the prevalence estimate. These include sample sizes that are just below the threshold when all other criteria are met, reporting only some of characteristics of the sample, test characteristics below the threshold but corrections for the test performance, and response rates that are just below the threshold in the context of probability based sampling of an appropriate sampling frame with population weighted seroprevalence estimates.

Moderate
The estimates are likely correct for the target population. To obtain a moderate risk of bias classification, most criteria must be met and departures from the criteria are likely to have only a small impact on the validity and reliability of the prevalence estimates. High The estimates are not likely correct for the target population. To obtain a high risk of bias, many criteria must not be met or departures from criteria are likely to have a major impact on the validity and reliability of the prevalence estimates.

Unclear
There was insufficient information to assess the risk of bias.
BMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s) Occupation Computerized Coding System (NIOCCS). However, NIOCCS reports the probability of 8 correct classification to the six-digit level. After manually verifying a subset of records from the first 9 round of classification, we decided to manual perform a second round of classification for any 10 observations for which the probability of correct classification was <0.8. This cut-off was chosen based 11 on the observation that that most codes with a probability of correct classification to of >0.8 to the 12 six-digit level were correctly coded at the two-and three-digit level, which we used in our main 13 analyses and are more likely to be coded correctly than the more granular, 6-digit codes and 14 consideration of the number of records that could feasibly be verified manually 15 BMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s)