Aim of the study
The aim of this study is to synthesise the evidence about the association of work risk factors and CVDs by performing a systematic review of observational studies of adults exposed to job strain, effort–reward imbalance, long working hours, job insecurity, shift work and occupational noise to assess the association of these work risk factors and the development of cerebrovascular diseases, ischaemic heart diseases or hypertensive diseases.
This systematic review will underpin and inform a broader objective to examine the effectiveness of interventions to minimise the effects of risk factors for CVDs in workplaces, with the final aim of informing occupational health policies in the future.
Study design
Due to the complexity and ethical considerations, human observational studies are the gold standard of the evidence base in the field of environmental health (20), which includes studies of work risk factors. This is a problem if systematic review methodologies as Cochrane and the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system are to be applied because they have been developed based on evaluation of randomized controlled clinical trials, considering human observational studies to be “low” quality evidence (20). That is the reason why our systematic review will be carried out following the Navigation Guide framework, which is a systematic and rigorous approach to research synthesis, based on best practices in the evaluation of information in evidence-based medicine and environmental health (20) (see Figure 1). It assigns a “moderate” quality rating to human observational studies and allows combination of diverse evidence streams (20).
This protocol adheres to the recommendation of the Preferred Reporting Items for Systematic Reviews and Meta-analysis Protocols guidelines, published in 2015 (PRISMA-P 2015) (21) (see Additional file 1: PRISMA-P 2015 Checklist) and it has been registered on the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42020179972). All the updates will be registered on PROSPERO and reported in the systematic review itself.
Eligibility criteria for study selection
Following the first step of the Navigation Guide, we will specify the study question using the Population, Exposure, Comparators and Outcomes (PECO) framework (20). A summary of the PECO statement of this study in Table 1.
Table 1: PECO statement
Population
|
Adults in working age
|
Exposure
|
Work risk factors
|
Comparator
|
Unexposed group or control group specified
|
Outcomes
|
Cardiovascular diseases
|
Details on the components of the PECO statement will be addressed in the following sections.
Types of studies
We will include quantitative observational research studies, specifically cohort studies (both prospective and retrospective) and case-control studies.
We will exclude randomized controlled trials, quasi-experimental trials, cross-over controlled trials, controlled trials without randomization, single-case studies, review articles, short communications, letters with insufficient information to analyse the results, guidelines, dissertations, qualitative studies, scientific conference abstracts and studies on animals.
Population
Inclusion: Studies of adults of working age at the baseline (18-65 years), working in the formal economy.
Exclusion: Studies of children (aged < 18 years), unpaid domestic workers will be excluded. Individuals with previous CVDs.
No restrictions will be imposed on the setting of recruitment.
Exposures
The working conditions and risk factors that will be included are:
- Job Strain.
- Effort-Reward Imbalance.
- Long Working Hours.
- Job Insecurity.
- Shift Work.
- Occupational Noise.
For eligibility criteria, we will include studies which apply descriptions of the work conditions and risk factors according to our definitions (see Additional file 2: Definition of each Working Condition and Risk Factor for Inclusion).
Comparator
Unexposed group or control group specified, depending on the risk factor assessed.
Outcomes
Development of a cardiovascular disease according to our definitions in Table 2, which follows the International Classification of Diseases 11th Revision, 2019 (ICD11) and its equivalence in the 10th Revision (ICD10), within the following groups of diseases:
- Cerebrovascular diseases.
- Ischaemic heart disease.
- Hypertensive diseases.
Table 2: Inclusion criteria for diseases (International Classification of Diseases 11/10th Revisions, 2019) (22).
Group of diseases (Outcomes)
|
Diseases included
|
ICD11 Codes
|
ICD10 Codes
|
Cerebrovascular diseases
|
Intracranial haemorrhage
|
8B00-8B03, 8B0Z
|
I60-I69
|
Cerebral ischaemia
|
8B10, 8B11, 8B1Y, 8B1Z
|
G45, I63, I67.8, I67.8
|
Stroke not known if ischaemic or haemorrhagic
|
8B20
|
I64
|
Cerebrovascular disease with no acute cerebral symptom
|
8B21
|
I60-I69
|
Certain specified cerebrovascular diseases
|
8B22
|
I67
|
Hypoxic-ischaemic encephalopathy
|
8B24
|
G93.1
|
Cerebrovascular diseases, unspecified
|
8B2Z
|
I60-I69
|
Ischaemic heart disease
|
Acute ischaemic heart disease
|
BA40-BA43, BA4Z
|
I20-I25
|
Chronic ischaemic heart disease
|
BA50, BA51, BA5Z
|
I25.2, I25.5, I25
|
Ischaemic heart diseases, unspecified
|
BA6Z
|
I20-I25
|
Hypertensive diseases
|
Essential hypertension
|
BA00
|
I10
|
Hypertensive heart disease
|
BA01
|
I11
|
For studies that applied older revisions of the International Classification of Diseases, the equivalences of the diseases will be used according to the WHO references.
We have excluded cardiovascular diseases with strong congenital evidence and diseases secondary to others, such as cerebrovascular abnormalities, late effects of cerebrovascular disease, vascular syndromes of brain in cerebrovascular diseases, certain current complications following acute myocardial infarction, hypertensive renal disease and secondary hypertension. Diseases of intracranial, extracranial and coronary arteries have been excluded because we are focusing on their effects which are included in our list of cerebrovascular diseases and ischaemic heart diseases.
For eligibility of measurement of cardiovascular disease, we will consider the following sources used by the World Health Organisation (WHO) and the International Labour Office (ILO) in their latest protocols to systematically review work-related burden of disease and injury (17) (23):
- Diagnosis by a medical doctor with images.
- Hospital discharge registry.
- Other relevant administrative data (for example, absence record due to illness or disability with a diagnosis certified by a medical doctor).
- Treatment registration data for an eligible cardiovascular disease (including both drugs and medical procedures from register data).
- Medically certified cause of death.
Self- report measures of cardiovascular disease will not be considered to avoid bias in the measurement.
Specifically for hypertension, we will include only studies where the diagnostic was done after two or more measures on different days to avoid confusions with ambulatory hypertension.
Search strategy
Following the Navigation Guide recommendations (20) and to avoid publication bias, the first author will conduct comprehensive literature searches using electronic academic databases for potentially relevant records from published and unpublished studies.
For published studies, the following electronic academic databases will be used:
- Ovid Embase (1974 to 26 May 2020).
- Ovid Medline (1946 to 26 May 2020).
- PubMed (1946 to 26 May 2020).
- Scopus (1788 to 26 May 2020).
- Web of Science (1900 to 26 May 2020).
- Ovid APA PsycInfo (1806 to 26 May 2020).
(See Additional file 3: Draft of Search Strategy in Ovid Embase as an example)
For unpublished studies we will include grey literature from these electronic grey literature databases:
- OpenGrey (http://www.opengrey.eu/).
- Open Thesis (http://www.openthesis.org).
- Grey Literature Report (http://greylit.org/).
- ProQuest Dissertations & Theses Global™ (www.proquest.co.uk/products_pq/descriptions/pqdt.shtml).
Furthermore, we will include studies by hand searches from the following sources:
- Reference lists of included papers.
- Citing reference searching of included papers.
- Collections of the review authors.
The search will be conducted in English words but without a language filter. If an article is written in a language other than English or Spanish, the document will be translated into one of these languages.
The downloaded references will be stored in the reference managers Endnote and Mendeley.
Study selection
To support all the stages of the systematic review, we will use the web-based software platform Covidence (24), where the study records will be uploaded. Duplicates will be identified and deleted. One reviewer will screen the titles and abstracts of the studies retrieved during the searches for relevance.
Two reviewers will then assess independently the full texts of articles identified as being potentially eligible for inclusion against the predefined criteria. Any discrepancies will be resolved by consensus or by a third reviewer.
We will illustrate the process of study selection at different stages in a PRISMA-based flow diagram (25).
If the full text of the article is not available in the databases used, we will conduct a more extensive search on the Internet. If after that search full text is not found, authors will be contacted and asked for the document. References, where full text and contact information are not available after the extensive search, will be excluded.
Data extraction
Data will be extracted by two reviewers independently and any discrepancies will be resolved by consensus or by a third reviewer. The extraction will be done using a data extraction form (See Additional file 4: Data Extraction Form), which will be designed and piloted by all the reviewers on 10 references before the data extraction.
For the data extraction of exposure and its measures, we will include specific fields for job strain, effort-reward imbalance, long working hours, job insecurity, shift work and occupational noise, which will be applied according to the exposure assessed in the study in extraction.
Any change in the form will be updated in PROSPERO and registered in the systematic review itself.
Missing data will be requested from the authors by the email or phone provided in the principal study record.
Quality of individual studies
The risk of bias, defined as “characteristics of a study that can introduce systematic errors in the magnitude or direction of the results” (20), will be assessed for each selected article and will be done by two reviewers working in parallel by applying an adaptation of the Navigation Guide method for human studies tool (26)(23)(27), which judges the risk of bias by nine domains:
- Recruitment strategy
- Blinding
- Exposure assessment
- Outcome assessment
- Confounding (at least adjusted for age, sex and socioeconomic status)
- Incomplete outcome data
- Selective outcome reporting
- Conflict of interest
- Other sources of bias
For the risk of bias assessment we will use the instructions included in the Case Study 7 of Application of the Navigation Guide (26) (ongoing) for judging the risk of bias for human studies (Additional file 5: Instructions for Making Risk of Bias Determinations).
Every domain will be graded as “Low Risk”, “Probably Low Risk”, “Probably High Risk”, “High Risk” or “Not Applicable”. The worst rating in any bias domain for any outcome will define the overall risk of bias at study level (17). Disagreements will be solved by consensus or by a third review author.
Strategy for data synthesis
This review will provide a description of the occupational risk factors related to cardiovascular diseases and their level of association in the development of cardiovascular diseases in working-age adults. Due to the diversity in the populations and exposures in the studies, the synthesis of the data will be done by a synthesis without meta-analysis (SWiM) (28).
The studies will be grouped by exposure (job strain, effort-reward imbalance, long working hours, job insecurity, shift work and occupational noise) and sub-divided by outcome (cerebrovascular diseases, ischaemic heart disease, hypertensive diseases).
To synthesise direction of the effects for each outcome we will apply the “vote counting based on direction” method (29) to have a big picture of the evidence to make decisions. This will be possible because we will include studies where the direction of effect is reported and because we expect the inclusion of a large number of reports (29). Following the recommendations of this method, we will create a standardized binary metric, where each effect estimate is categorized as showing the harm or benefit based on the observed direction of effect alone to finally compare the number of studies (votes) showing harm with the number showing benefit using a sign test (29). For the sign test, if there is no effect, the study will be part of the null hypothesis of no difference, following the Cochrane recommendations (30). Studies with “high risk of bias” and “probably high risk of bias” won´t be included in the primary synthesis, however all the studies will be considered to draw the final conclusions about the strength of the evidence.
Quality of the overall body of evidence
For grading the quality of evidence of each outcome by working condition and risk factor we will use the Navigation Guide tool (26) (Additional file 6: Instructions for Grading the Quality and Strength of Evidence), which downgrades or upgrades the quality considering eight categories::
- Category 1. Quality of Study Limitations (Risk of Bias).
- Category 2. Indirectness of Evidence.
- Category 3. Inconsistency of Evidence.
- Category 4. Imprecision of Evidence.
- Category 5. Publication Bias.
- Category 6. Large Magnitude of Effect.
- Category 7. Dose‐response.
- Category 8. Confounding Minimizes Effect.
The quality of evidence will be assessed by two reviewers working in parallel using the Navigation Guide quality of evidence assessment tool, grading the evidence in “high”, “moderate” and “low” (26). Disagreements will be resolved by consensus or by a third review author.
Strength of the evidence
Finally, the overall strength of each body of evidence of each outcome by working condition and risk factor will be rated based on four considerations following the Navigation Guide criteria (26):
- Quality of body of evidence.
- Direction of effect.
- Confidence in effect.
- Other compelling attributes of the data that may influence certainty.
The assessment will be done by two reviewers applying the definitions in Additional File 6: Instructions for Grading the Quality and Strength of Evidence. The strength of the evidence will be rated in one of the following four categories:
- Sufficient evidence of toxicity/harmfulness
- Limited Evidence of Toxicity/harmfulness
- Inadequate Evidence of Toxicity/harmfulness
- Evidence of Lack of Toxicity/harmfulness
Disagreements will be resolved by consensus or by a third review author.