STRENGTHS AND LIMITATIONS OF THIS STUDY
-
This systematic review highlights the complexity of measuring health at work and provides an overview of existing instruments categorised according to the Jobs Demands Resources model assisting readers in selecting appropriate instruments for their specific context.
-
Data screening and extraction were piloted to minimise discrepancies and variations in interpretation between the different researchers.
-
The under-reporting of important information regarding measurement instruments resulted in manual cleaning of instrument names and constructs and introduced interpretation challenges, emphasising the need for uniformity and completeness in reporting.
-
Despite the fact we only included evaluated instruments, due to the large number of instruments, it was not feasible to asses validity and reliability outcomes in detail for each instrument.
Introduction
Health at work can be defined as ‘the creation of an environment that fosters contentment and allows employees to flourish and achieve their full potential, benefiting both themselves and their organisation’.1 Health at work for healthcare professionals (HCPs) is for achieving effective, safe and good patient care.2–4 Considering current rates of absenteeism/turnover among HCPs, it may be clear that health protection in the field has been insufficient.5 Lack of health at work is associated with consequences for individual HCPs, such as poor work–life balance,4 6 obesity,4 reduced quality of life,4 7 substance abuse and even suicide.4 8 At an organisational level it is related to high staff turnover,4 8 absenteeism4 9 and costs.4
It is understood that health contains psychological, physical and social health.10 Different models and concepts of health exist. One of these is the Self-Determination theory,11 suggesting that motivation is a mediator between work performance and well-being.11 For students, the Coping Reserve Tank was illustrated,12 showing a reservoir that can be replenished or drained by various aspects as stress, mentorship, demands and support.12 Potential outcomes as resilience versus burnout were described.12 Current study uses the Job Demands Resources model (JD-R model) since it facilitates communication about ‘work and health’.13 14 The model is comprehensively tested, frequently used in literature and fits within the context of hospitals. In essence, the JD-R model integrates two processes: the stress process, which is sparked by excessive ‘job demands’ and lack of resources and the motivational process, which is triggered by abundant ‘job resources’ and may lead to positive outcomes such as commitment, intention to stay and work performance.14 The different model components contribute to a more positive well-being (eg, Job satisfaction) or to a more negative well-being (eg, Burnout). The JD-R model components are ‘demands’ (eg, stress, workload, conflicts), ‘resources’ (eg, support, development opportunities, team atmosphere), ‘leadership’ (eg, inspiring, connecting) and ‘personal resources’ (eg, motivation, resilience), see figure 1.14 In summary, the JD-R model gives a clear structure to the various components and processes that influence health at work.
The majority of institutions remain focused on individual-level approaches when it comes to preserving and promoting personal health.15 16 There is often a mismatch between employees’ perception and organisations’ understanding of health at work.1 Moreover, working in healthcare is seen as a vocation encompassing compassion, relativeness and competence. This ideal is high stake and failing or impairment leads to stigmatising, shaming, blaming and humiliation.5 Uncertain effectiveness of interventions shifts responsibility onto individuals, posing potential harm.5 To bridge this gap, organisations should monitor health among HCPs to design and implement effective, tailored interventions to specific needs of HCPs.
However, it is unclear how to effectively measure and monitor health at work for HCPs, given its multidimensional nature and diverse elements.1 17 Health at work instruments vary for specific professions,18 specific settings19 or include only one or two aspects of health at work.14 20 Currently, there is a lack of comprehensive overviews of instruments specifically designed to assess and monitor HCPs’ health.
This study is part of a programme of the Netherlands Federation of University Medical Centers about finding ways to improve and monitor HCPs’ health in Dutch hospitals. This scoping review aims to provide a comprehensive overview of available health instruments specifically developed and validated for HCPs in hospital care. Addressing health at work for HCPs and give an overview of instruments for the purpose of timely health screenings is crucial for (1) ensuring the well-being of individuals, (2) prevent and act on negative health in workplaces,21 22 (3) gaining insight into the unique challenges and needs of these HCPs, (4) evaluating interventions’ effectiveness, (5) the sustainability of healthcare organisations and (6) the quality of patient care.21 22 Our scoping review of health instruments for HCPs serves as an essential step for hospitals on this transformative journey.23
Conceptual model of health at work, based on the energy compass of the Jobs Demands Resources model.
Methods
This scoping review was conducted using the six-stage framework of Arskey and O’Mally,24 as well as the Preferred Reporting Items for Systematic Reviews and Meta-Analyses for Scoping Reviews (PRISMA-Scr).25 Scoping reviews are helpful to explore a broader perspective, complex and heterogeneous literature.23 This review focuses on identifying suitable instruments, assessing their coverage of JD-R model components and measured constructs, and determining commonly used and comprehensive tools for monitoring HCPs’ health, leveraging its ability to uncover hidden gaps and inconsistencies. The protocol has been previously published.26
Eligibility criteria
Studies were considered eligible for inclusion when published in English from January 2011, and using or evaluating an instrument that assessed health of nurses or doctors working in hospitals. Health at work, or its aspects, was defined according to the JD-R model.14 Studies were excluded when they described instruments not evaluated or the sample consisted of students alone.
Information source
Studies were retrieved from MEDLINE, Embase and CINAHL. The first search was conducted in December 2021 and updated in January 2024.
Search
The search strategy was developed in collaboration with an information specialist. Several terms derived from the research aim were identified to develop search strings to find relevant literature. The search strategy is reported in online supplemental material 1. Keywords and MeSH terms related to the domain (HCPs working in hospitals), the determinants (instruments for monitoring) and the outcome (health at work) were used.
Supplemental material
Selection process
Independent screening and selection of studies were performed by AB, KB or KD using Rayyan (Rayyan Systems Inc, USA). After removing duplicate records, articles were screened on title and abstract regarding inclusion and exclusion criteria. To ensure consistent screening, the first 100 articles were pilot-screened on title and abstract by all three researchers until consensus was reached. The researchers proceeded to full-text screening with selected articles, from which the first 18 articles were again piloted until consensus. The domains of the energy compass (‘job demands’, ‘job resources’, ‘engaged leadership’, ‘personal resources’, ‘employees well-being’, ‘outcomes’) of the JD-R model14 were used to assess whether the reported instruments measured (aspects of) health at work. The JD-R model includes all aspects of health at work and is applicable to different employees in different settings.14 See figure 1 for the conceptual model used.
Data extraction
AB, KB, DI and KD performed data extraction and full-text screening simultaneously using a predeveloped data extraction form in Excel. They extracted eligible instruments and recorded instrument details separately if multiple instruments were mentioned. The form’s adequacy was tested on 18 studies, and refinements were made after discussing disagreements.
Data items
The data items charted were: (1) study characteristics (ie, year, authors, country and study aim); (2) sample characteristics (ie, type of HCP (nurses, doctors, other), setting and sample size); (3) details of measurement instrument (ie, instrument name as reported, main construct and subconstructs as reported by the article). If no constructs were reported, judgement was made by the researchers); and (4) reporting of psychometric properties (validity, reliability, responsiveness, quality references). Included instruments were categorised according to the six domains of the JD-R model (‘job demands’, ‘job resources’, ‘engaged leadership’, ‘personal resources’, ‘employees well-being’, ‘outcomes’14). Uncertainty in the selection and data extraction processes was resolved by consensus within the research team. Extracted instruments were checked for discrepancies. Instrument names and main constructs were cleaned on terminology (eg, equalising punctuation marks, abbreviations and capital letters). Additionally, the classification of the domains of the JD-R model was checked, discussed, and cleaned for similar instruments.
Critical appraisal of individual sources of evidence
As this scoping review aims to overview instruments measuring health at work, evaluating risk of bias was not of our interest.24 27
Synthesis of results
Results were synthesised by identifying unique measuring instruments. According to the JD-R model, instruments that measured more domains were valued as more comprehensive. Trends over time were analysed for main constructs, JD-R domains and the number of domains. Characteristics were descriptively analysed and visualised using Excel and Python, including a Sankey diagram representing instrument connections between JD-R domains, main constructs and population (nurses, doctors, or both).
Reflexivity
The research team comprises junior researchers supervised by experienced senior researchers. The interdisciplinary team consists of professionals with diverse backgrounds, experiences and perspectives, promoting self-reflection on biases, including medical doctors, a nurse, physiotherapist, pharmacoepidemiologist, former residency programme director and a review methodology specialist.
Patient and public involvement statement
Patients were not involved in our study.
Results
Selection of sources of evidence
The electronic database search yielded 7029 citations. After removing duplicates, 5751 records were screened based on their title/abstract. Among this dataset, several studies were excluded from the review. Main reasons for exclusion were absence of full-texts, language, date of publication, wrong population or did not fulfil outcome requirements. Data extraction was performed on 1204 articles, see figure 2.
Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram of the study search and selection for instruments measuring well- being among nurses and doctors.
Characteristics of sources of evidence
Last decade, the number of publications has significantly increased (almost quadrupled), suggesting interest in measuring health at work has also increased. Current study included studies published between 2011 and 2024. A total of 1204 studies included were reported using an instrument measuring health 2872 times. All instruments found were evaluated. Of the 2872 instruments, 75.4% (n=2166) were reported more than once, resulting in the identification of 986 unique instruments. Of the 986 instruments, in 569 instruments at least one study had reported that the instruments were valid. The reporting of the reliability was done for 678 instruments. An overview of all instruments is presented in online supplemental material 2. Half of the instruments were used to measure health at work in nurses only (n=515, 52.2%), 23.9% were used (n=236) in doctors only and the remaining 235 instruments (23.8%) were used for both professions. Studies from across six continents were included, and instruments were categorised per country according to the World Bank Classification 2022 (low-income, lower middle-income, upper middle-income and high-income countries). Most studies using instruments to measure HCPs’ health at work were conducted exclusively or partially in high-income countries (75.4%), see online supplemental material 3.
Synthesis of results
Main constructs and subconstructs
In total, we included 1204 studies and extracted 986 unique instruments. A wide variety of instruments have been found in terms of instrument characteristics (eg, aspects of health, main constructs and amount of question items (eg, 30 questionnaires reported as single item instruments)) and usage characteristics (eg, among which HCP, number of times applied and usage in various countries). Sankey diagrams were made to highlight connections between most frequently identified main constructs (and their underlying unique instruments) and JD-R domains among three groups of HCPs (nurses, doctors and both); see figure 3.
Sankey diagrams representing the number of unique instruments (752) and their connections between population, main constructs and Jobs Demands Resources domains.
The collected instruments represent a total of 251 different main constructs. Among these constructs, there were 9 that occurred 15 times or more, measured by a total of 379 instruments. The most commonly occurring main construct was Depression/anxiety/stress, which occurred in 90 instruments. Job satisfaction appeared in 77 instruments, Work environment in 69 instruments, Burnout in 37 instruments, Intention to leave in 29 instruments, Well-being in 24 instruments, Culture in 19 instruments, Quality of life in 19 instruments and Commitment in 15 instruments. Examples of other main constructs that occurred less frequently are Leadership, Self-efficacy and Belongingness. Additionally, subconstructs were often described in the included papers. For each instrument, we categorised which domains of the JD-R model were measured based on these subconstructs. Figure 3 shows a visual representation of relationships between main constructs and the six domains of the JD-R model. Sankey plots give insight into the heterogeneity and overlap among the main constructs, and their connections to the JD-R domains. The thickness of the flows in the plots corresponds to the frequency of occurrence of the various main constructs. However, it is important to note that the relative sizes of the nodes in different diagrams are not comparable. Consequently, it is suitable to compare thickness of flows within each plot, but not between different plots. For example, when looking at instruments measuring the construct Burnout, these instruments include questions that can be related mostly to the ‘well-being’ domain of the JD-R model (the light green line between Burnout and ‘well-being’ is the thickest), then to equally to ‘resources’ and ‘job demands’ domains of the JDR-model (red and orange lines between Burnout and ‘well-being’ are equally thick). Of the 986 unique instruments, we identified 294 instruments that at least measure ‘job demands’, 332 instruments that at least measure ‘job resources’, 78 instruments that at least measure ‘engaged leadership’, 256 instruments that at least measure ‘personal resources’, 450 instruments that at least measure ‘employees’ well-being’ and 124 instruments that at least measure ‘outcomes’. We observed significant overlap among instruments, as many of them measured multiple domains of the JD-R model. None of the extracted instruments covered all six JD-R domains comprehensively. The specific occurrences of constructs, instruments and main constructs are reported in more detail in the online supplemental material 2.
Most comprehensive instruments
The most comprehensive instruments were regarded as those which measure at least four out of six JD-R domains (based on what has been reported about the instrument in the included studies). We identified 32 comprehensive instruments, measuring 20 different main constructs. Five of them occurred more than 10 times and are shown in online supplemental material 4: (1) The ‘Moral Sensitivity Questionnaire’ measures Moral sensitivity, consists of 35 items and contains 4 domains of the JD-R model. It was extracted 40 times, and was applied to both nurses and doctors in Europe and Asia; (2) The ‘Short Form Health Survey’ measures General health, consists of 36 items and contains 5 domains of the JD-R model. This instrument was extracted 23 times, and was applied to both nurses and doctors in Europe, Asia, Africa, North/South America and Oceania; (3) The ‘Professional Fulfillment Index measures Professional fulfilment, consists of 16 items and contains 4 domains of the JD-R model. This instrument was extracted 18 times, and was applied to both nurses and doctors in Asia and North America; (4) The ‘Professional Practice Environment Scale’ measures Work environment, consists of 38 items and contains 4 domains of the JD-R model. This instrument was extracted 13 times, and was applied to both nurses and doctors in Europe, Asia, North America and Oceania; (5) The ‘Copenhagen Psychological Questionnaire’ measures Psychological health status, consists of 141 items and contains 4 domains of the JD-R model. This instrument was extracted 12 times, and was applied to both nurses and doctors in Asia and Europe.
Most common instruments
Online supplemental material 5 provides examples of instruments corresponding to common main constructs, with the assumption that frequently used instruments are generally more extensively evaluated, user-friendly, accessible and available in multiple languages. The five most commonly used instruments were the ‘Maslach Burnout Inventory’ (314), the ‘Professional Quality of Life Scale’ (60), the ‘Perceived Stress Scale’ (58), the ‘Patient Health Questionnaire (57) and the ‘Practice Environment Scale of the Work Nursing Index’ (56). All the instruments were used to measure health of both nurses and doctors in five continents (America, Africa, Asia, Europe, Oceania) except for the ‘Professional Quality of Life Scale’. This instrument was used in America, Asia, Europe and Oceania. The most common instruments are aggregated into four main constructs: Burnout, Depression/anxiety/stress, Quality of life and Work environment, and cover 56.9% (215) of all instruments (378) of the common main constructs.
Discussion
This scoping review provides a comprehensive overview of evaluated instruments to assess and monitor nurses’ and doctors’ health at work. The review identified 986 unique instruments, covering 251 different main constructs, which highlights the great variety of available instruments measuring health at work and its various constructs. In the review, a distinction is made between comprehensive instruments and common instruments. Comprehensive instruments can be used for conducting initial health screening, providing a broader assessment of overall health at work. Subsequently, evaluation of more specific health domains can be conducted using common instruments that measure the domains of interest. This approach allows for a comprehensive understanding of health at work by combining the use of broader and more targeted instruments.
A one-size-fits-all approach to measuring health at work is challenging due to its complexity, resulting in a wide variety of instruments, each addressing specific aspects. Interestingly, our analysis revealed that none of the identified instruments covered all six domains of the JD-R model, suggesting a need for combining instruments to fully capture all potential aspects. Using multiple multi-item questionnaires can burden staff, as it requires time and effort, alternatively combining single-item questionnaires may offer a more feasible option. This approach can also shorten surveys, potentially improving response rates and reducing attrition in longitudinal studies.28 Since we found 30 single-item questionnaires measuring the domains ‘job demands’, ‘personal resources’, ‘well-being’ and ‘outcomes’, by combining these single-item questionnaires, it can only be possible to screen health on some aspects. Nonetheless, given the broadness and diversity of health at work, a combination of instruments, including single-item questionnaires, offers a pragmatic approach to comprehensive assessment while minimising staff burden and maximising data collection efficiency. On top of that, single-item questionnaires have been shown to effectively assess many relevant constructs.28
When selecting a tool to measure health in HCPs, it’s important to note that some instruments are specific to doctors or nurses (likely to their distinct roles), while others can be used interchangeably. For example, Assertiveness and Belongingness are constructs measured exclusively among nurses, while Job fit and Career calling are specific to doctors. Considering the interdisciplinary nature of healthcare, a harmonised data collection system could facilitate in monitoring HCPs’ health across disciplines.29 Additionally, measurement properties and feasibility, alongside target groups and constructs, should be considered in instrument selection. Common instruments are presumed to be more accessible and user-friendly, that reflected in the visualised result presentation.
From all JD-R domains, well-being (eg, in terms of Burnout, Sleep and Job satisfaction) was the most represented domain with Burnout being the most measured main construct. The popularity of the Maslach burnout inventory, validated since 1981, may explain the prevalence of burnout as the most commonly measured construct.30
However, focusing solely on Burnout overlooks the broader concept of health at work. Most instruments in this study primarily emphasise negative outcomes (in contrast to the current trend in organisational psychology focusing on positive effects), while a more comprehensive approach should consider both positive and negative components of well-being. As Schaufeli reported, usually specific concepts of health are examined based on the organisations’ needs.14 Institutions planning to monitor HCP health should understand the comprehensive nature of workplace health and the lack of a single instrument. A holistic approach helps in understanding dynamics and guides future actions (eg, prioritising and implementing actions).14 Combining instruments or selecting relevant domains is crucial. Previous research also highlights the need for a broader view of health in healthcare contexts. 14 17
Lastly, the majority of included studies were conducted in high-income countries, which may limit generalisability. It is crucial to consider cultural nuances and stressors specific to health at work in diverse settings. Therewith, adapting instruments for broader use requires translation and validation for cultural sensitivity.
Strengths and limitations
This review has multiple strengths. It provides a comprehensive overview of various instruments rather than focusing on a single one. The figures and tables highlight the complexity and diversity of measuring health at work, guiding further research directions. The review followed a systematic approach and benefited from the expertise of review methodology specialists. The use of the JD-R model allowed for a clear understanding of how instruments align with the health construct. Categorising instruments according to JD-R domains assists readers in selecting appropriate instruments for their specific context. Consensus was reached through pilot screenings to ensure unbiased selection and data extraction. An audit trail documented and linked methodological choices, thoughts and uncertainties. Lastly, the review’s strength lies in its transparency, as it included a submitted protocol prior to data extraction. Limitations should be acknowledged. First, data screening and extraction were conducted by different researchers, which may introduce variations in interpretation. Pilots were conducted to minimise discrepancies. Second, the ‘well-being’ domain of the JD-R model was not further divided into positive and negative concepts due to not consistently measuring the same concepts of health at work. However, positive and negative outcomes are inter-related, and the absence of one can be seen as negative. These determinants can influence each other bidirectionally. Third, among different articles, instruments were often reported slightly different, or instruments were adjusted by the authors, either by shortening them or by choosing and studying subdomains of an instrument. Instrument names and main constructs were cleaned manually on terminology and were subject to our interpretation due to unclear reporting in the included studies. Also, literature mentions the phenomena of under-reporting of important information with regards to measurement instruments.31 32 Improving reporting increases transparency and decreases risk of bias.31 This allows accurate assessment and better reliable application.31 We therefore stress the importance of uniformity and completeness in reporting measurement instruments. Finally, for this review it was not feasible to assess validity and reliability outcomes for each instrument because of the vast numbers of studies and unique instruments. In this study, we extracted if studies reported or referred to other studies reporting measurement properties, which was a criterium for inclusion in this review. While validity and reliability were frequently reported, responsiveness was never mentioned. For screening and monitoring health at work, detecting (small) differences is essential. Therefore, insight in responsiveness is important since responsiveness refers to the ability to detect meaningful changes over time.31 By ensuring that instruments are responsive, we can enhance the accuracy and effectiveness of health assessments, ultimately leading to better support and interventions for HCPs. Our comprehensive overview of health instruments for HCPs serves as an essential step for hospitals on this transformative journey.
Practical implications
This review may guide healthcare organisations in selecting monitoring instruments tailored to their needs. To begin, organisations must first determine the domain of interest, that is, which aspect of health needs further enquiry. For zooming in on more specific domains more common and specific instruments can be used, like the examples presented in the results. Another option is to screen with a broader purpose, measuring more concepts of health at work, for which the examples of comprehensive instruments can be used. Last, organisations can combine these two types of instruments by screening broadly first to determine red flags. Thereafter, red flags can be explored by using specific instruments. Summarising, when hospitals assess and monitor nurses’ and doctors’ health with validated instruments, they can timely design and start tailored interventions to prevent negative well-being in workplaces, thereby contributing to sustainable employability and quality of care.
Conclusion
This scoping review provides an comprehensive overview of validated instruments, which can be used to assess and monitor nurses’ and doctors’ health at work. It reveals the broad variety in available instruments and their corresponding constructs. These instruments can be categorised into two groups: comprehensive instruments and most common instruments. For a more focused evaluation of specific domains, more common and specific instruments can be used. Another option is to screen with a broader purpose, aiming to measure multiple concepts of health at work. It is important for institutions planning to monitor health at work of their HCPs to be aware of the holistic nature of health at work and acknowledge the absence of a single comprehensive instrument that covers all the six domains of the JD-R model. Therefore, a combination of different instruments or a selection based on the most relevant domains should be considered. By taking into account the diverse dimensions of health at work and selecting appropriate instruments, institutions can gain a more comprehensive understanding of the well-being of their HCP.
Future research should prioritise investigating the generalisability of instruments for both nurses and doctors. Evaluating the feasibility and effectiveness of combining instruments to capture the concept of health at work according to the JD-R model comprehensively is crucial. Rather than developing new instruments, modifying, refining and shortening existing ones by incorporating subdomains is recommended to minimise confusion and inefficiency. Additionally, assessing psychometric properties, including responsiveness, of combined instruments is essential. These efforts will enhance the validity and applicability of health assessment tools in healthcare settings.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Ethics statements
Patient consent for publication
Ethics approval
Not applicable.
Acknowledgments
Special thanks to Neda Ansari, who helped with the data screening, and the information specialist Rene Spijker developing the search strategy.
This post was originally published on https://bmjopen.bmj.com