Do worksite health promotion programs (WHPP) influence presenteeism among employees? A systematic review 1

Abstract

BACKGROUND:

Worksite health promotion programs (WHPP) aiming to reduce presenteeism are rapidly increasing. However, program-efficacy seems impaired by lack of conceptual understanding and theoretical frameworks related to presenteeism.

OBJECTIVES:

To study (1) the concept of presenteeism, (2) the causes of the phenomenon and (3) which WHPP affect presenteeism.

METHODS:

We followed the PRISMA 2020 guidelines for mapping relevant literature in seven databases. The main search-terms were “worksite”, “WHPP”, “presenteeism”, “employee productivity” and “interventions”. Reviews and review of reviews were included and analysed according to research questions and ecological theory.

RESULTS:

Of 2196 studies identified, 14 reviews were included, comprising 1390 single studies. Conceptually, presenteeism was poorly defined, mostly related to productivity/performance limitations due to ill health. Factors causing presenteeism addressed individual health or lifestyle (78%), and, to a lesser extent work and organisational factors (management/job stress) (22%). Forty-five highly diverse WHPP were identified, about 90% focusing health outcomes. 22/45 WHPP affected presenteeism/productivity positively. Multi-component WHPP implemented at multi levels accounted for 55%. Single-component programs, implemented at individual level, were mostly targeting health-risk subgroups. No specific WHPP seems preferable. Methodological quality seems highly associated with worksite intervention-efficacy.

CONCLUSION:

Increased telework opportunities due to COVID-19 experiences has affected working conditions worldwide, for better and worse, requiring adaptions at employee, management, organisational and socio-cultural levels. Our results showed high program diversity, several adaptable to future needs (e.g., digital solutions), however hampered by narrow focus on individual health. Future research should focus on theory-based high quality multilevel WHPP comprising contextual, work, and organisational factors.

Keywords

Worksite health intervention occupational programs productivity loss cost sustainable physical exercise stress management

1 Introduction

The increasing incidence of unhealthy lifestyle e.g., sedentary behaviour, poor nutrition, and lack of stress management, has resulted in mental and physical health problems [1], also affecting the labour force, and often resulting in sickness presenteeism (e.g., attending work while ill) [2–5]. This emerges as a growing problem worldwide, with prevalence rates of presenteeism varying from 30% to above 90% [4, 5], showing even higher rates lately due to the extensive short- and long-term consequences of the COVID-19 pandemic on people’s family and private lives, health, work, and social security [6–8].

Most often sickness presence is about reduced work capacity of employees and productivity loss, due to psychological and physical health problems, not necessarily leading to health-related absence from work (i.e., absence from work due to illness or disability) [4]. However, a common agreement on definition, framework and tools used to measure its impact is not yet established [5, 9]. Our focus of interest is to further explore how the concept and causes of presenteeism are described in the scientific literature. The theoretical basis and conceptualisation of presenteeism influence efficacy of worksite interventions [4, 9], the latter, which is of particular interest in this study.

The term presenteeism is understood and used differently in scientific research [4, 9]. Firstly, it is closely related to individual lifestyle and health [10]. Secondly, presenteeism is related to environmental and psychosocial work factors (e.g., work demands, work control, social support) [11–13], including leadership [14] and finally, to productivity loss and financial cost (e.g., estimations of economic impact of presenteeism) [9, 15]. Economically, presenteeism seems far more challenging than health-related absence [5, 16] and more complicated to determine as it is not formally registered, being both invisible and subjective. Accordingly, it is often a hidden cost in businesses [16]. Due to its consequences for both employers and employees, the interest of the phenomenon has grown substantially during the last decades. Employees that frequently exhibit presenteeism are at greater risk of developing health issues in the future, which in return may lead to health-related absence [4, 10]. Absenteeism and presenteeism share a common ground, as they are both types of absence [17]. Meta-analytical data show strong correlations between health-related absence and presenteeism, also indicating that the relationship varies with contextual factors [9, 18]. This seems particularly true for employees in front occupations (e.g., hospital staff, teachers, public transport drivers) [6–8, 11].

The interest in the concept of presenteeism seems mainly to represent two distinct sources [3, 4]. One tradition, mostly American, assessed the productivity consequences of presenteeism as a function of various health conditions (e.g., musculoskeletal pain or mental health problems), while ignoring the causes of showing up ill. This resulted in a focus on measurements on productivity losses, individually oriented wellness and medical interventions aiming at reducing these losses [3]. In the other tradition, mainly European, research focused on presenteeism as a behavioural pattern of going to work with impaired health. Studies have investigated the causes, consequences, and the motives for this behaviour, frequently measured as the occurrence of presenteeism, and often reflecting job demands and other occupational characteristics causing stress and illness [3, 9]. Interventions related to this tradition may reflect e.g., job redesign and reduced job stress, as well as occupational health and rehabilitation programs [9]. Hence the concept of presenteeism is a complex phenomenon influenced by several factors.

Johns [3, p. 251] developed a theoretical framework containing elements of both traditions above, defining presenteeism as “attending work while ill”. He emphasized the importance of merging organisational theory with those of occupational health and epidemiology. This view has received special attention and is further developed in recent research [4, 9, 18]. For example, Lohaus and Habermann [4] presented a comprehensive framework for understanding presenteeism, integrating Johns’ [3] work. They emphasised environmental influences (e.g., culture, policy, legislation), organisational (e.g., absence policy), work – and person-related variables (e.g., workload and health) as the foundation for decision-making ability and the individuals’ possibility to choose either presenteeism or health-related absence. Most researchers in the field have focused on the negative consequences of presenteeism, however, presenteeism may also have positive outcomes [13, 19, 20]. Karanika-Murray and Biron [20] developed a framework representing four varieties of presenteeism, referred to as functional, therapeutic, dysfunctional, and overachieving presenteeism. Whether to choose presenteeism rather than sickness absence, is influenced by the availability to perform and the severeness of health impairment [20]. Functional and therapeutic presenteeism evolve positive outcomes. Employees managing work performance at a high level despite health impairments and simultaneously regaining one’s health, is described as functional presenteeism. Therapeutic presenteeism behaviour benefits health, but not necessarily enhancing performance. However, to facilitate these choices, a conscious and supportive management is required as well as resources to adjust related to work demands and psychosocial work conditions (e.g., reduced working hours) [19, 20]. Considering the occurrence of the different varieties of presenteeism is particularly important due to increased telework, lack of on-site management and changed health rules and regulations as a consequence of the COVID-19 pandemic [21].

Several researchers have investigated consequences of increased telework [21–23]. Steidelmüller et al. [22] reported associations between homebased telework and increased self-endangering behaviour, e.g., sickness presenteeism, during the pandemic. Gerich [24] further explored and elaborated these associations in a path-analysis on survey data from three economic sectors. Employees with pre-pandemic telework experience more often used telework for work intensification (e.g., working in weekends and days off) to meet changed work demands, (partly) resulting from new or changed leader behaviour (goal directed management/indirect work control). Indirect work control was associated with increased sickness presenteeism, indicating the importance of work organisation and leader behaviour on employee presenteeism [24]. Based on this research, it is reasonable to assume that both negative and positive effects of presenteeism will increase in most occupations in the future, depending on organisational, managerial, and individual adaptations to changes (e.g., more telework, increased indirect work control, changed social support, increased health impairments) [9, 20, 21, 25]. Hence, strategies to manage or reduce presenteeism behaviour are in demand [9]. Such interventions are still scarce [5], and characterized by conceptual differences, incomparable instruments to measure the phenomenon and lack of tailoring the programs to context, the program outcomes, or participants [26]. To meet the complexity and the many unforeseen challenges in the field, the current theoretical knowledge should be taken into consideration, focusing on complex, multilevel interventions concerning environmental, organisational, and individual determinants of health and health behaviour [27]. Lack of systematic implementation and process evaluation also seems to influence the outcome effects of WHPP [28]. Poor implementation affects methodological quality and impair transfer of evidence-based programs to real life worksite settings [29].

The purpose of this review is to increase knowledge of presenteeism and its significance. To thoroughly explore our main aims, we investigate (1) how the concept of presenteeism is described in the scientific literature, (2) what factors influence the concept and (3) which WHPP affect presenteeism. To address the complexity in the field of theory and interventions, we chose to analyse our results according to ecological theory, which include the numerous causes and levels of implementation [30, p. 41], specifically the individual (I), organisational (II) and socio-political (III) levels.

2 Methods

2.1 Literature search and study selection

To answer the research questions, a systematic review was conducted [31] and scientific literature systematically searched in seven databases (Business Source Elite, the Cochrane Library (Systematic Reviews and Trials), EMBASE (Ovid), MEDLINE (Ovid), PsycINFO (Ovid), Sport Discuss and Web of Science). In January 2019, a professional librarian along with the authors conducted the targeted search of literature in the databases from 2010 to the current date at the point. An extended track down of articles consecutively published throughout 2020 were executed. Inclusion criteria were worksite setting, employees aged 18 to 70 years, both sexes, employed and attending health promotion programs provided in work settings. The outcome variable of interest was presenteeism and/or productivity loss. We included systematic literature reviews as well as review of reviews from a North American-, European and Australian / New Zealand context written in English. Grey literature was omitted due to our interest primarily in peer-reviewed effect-studies and experimental designs. Exclusion criteria were studies published before 2010, not written in English, non-review designs and settings other than worksite. Asian, South American, and African studies were excluded due to different working environment legislation and labour laws.

To reflect the issue precisely and avoid bias, a customized Population-Intervention-Comparison-Outcome (PICO) -search was performed to identify and organise keywords. Terms were based on MeSH indexing and thesaurus’ as well as free text terms. The searches were mainly based on the following terms and procedures; (workplace* OR worksite* OR “white collar” OR employ* OR occupation*) AND (“physical activity” OR exercise OR “stress management” OR “health promotion” OR “organisational intervention*” OR “occupational health intervention*” OR “worksite health promotion program*” OR leader* OR “life skill”) AND (absenteeism OR “sick leave” OR presenteeism OR “employee productivity“ OR “productivity loss“ OR “cost“ OR “work while ill” OR “work while sick”). Furthermore, the results from the PICO-search were transferred to EndNote, a reference-mapping tool. Duplicates were removed prior to screening. The literature review was completed using The Preferred Reporting Items for Systematic Reviews and Meta-Analyses – PRISMA 2020 checklist. This checklist guides the researcher through all parts of the literature review that need to be considered and explained [32]. As the flow diagram (Fig. 1) shows, 2196 publications were identified in the main search. Both authors worked independently in separate Endnote files, and performed a blind title, abstract and key words screening, and evaluation of relevance before comparison of results. After the individual inclusion and exclusion processes, the results of both authors were compared and discussed. All titles identified were manually examined for relevance. Relevant titles lead us to consider the abstracts for applicability and eventually inclusion. Fifty-six articles, rated to be of importance, were retrieved in full-text and underwent scrutiny according to the inclusion criteria. Any doubt or disagreement about inclusion was resolved by discussion and consensus. Fourteen articles met the inclusion criteria. Periodic updating of the search between 2019 and 2020 did not identify additional publications (see PRISMA 2020 flow diagram, Fig. 1).

Fig. 1

PRISMA 2020 flow diagram of the literature search.

2.2 Data extraction and analysis of included material

Based on a narrative synthesis [31], key points of eligible studies were summarized in an overview of the included studies. Each study was examined for independent information about first author, publication year, nationality, aim and study outcome. In addition, we included information about method (study design, included number of databases searched, number of studies, designs, participants, setting and use of theoretical framework). To appraise risk of bias, methodological quality in each included review was assessed according to a questionnaire customized to this study. The form consisted of six questions (see Supplement 1). Each study was given a number: yes = 1, no = 0, and maximum total score was 6. The 14 studies were graded: high (score 6), moderate (score 4-5), or low methodological quality (score < 3). Studies lacking methodological information were termed inconclusive (IC). In addition, each single study identified (n = 44) reporting on relevant outcomes for our study, was appraised for reported internal methodological quality based on the review-authors own descriptions. A simple cluster analysis on reported positive/negative outcome effect versus reported graded methodological quality was completed for each study.

To answer our research questions, the results were analysed according to research questions and health promotion theory [30] emphasizing categorization into three implementation levels: (1) individual level, (2) organisational level and (3) socio-political level.

2.3 Risk of bias

Selection: The very few studies reporting on relevant outcomes (3%) and only considering studies written in English may both represent a selection bias. During the selection of the 14 reviews, misinterpretation can occur and the decision to retain or discard articles may fail.

Interpretation: Lack of a unified established framework and divergent definitions of presenteeism lead to different interpretations and varying quality of measurements. It requires valid narrative analysis. The absence of a valid and clear definition of central terms may lead to interpretation errors. Work-norms, social security and insurance systems worldwide vary extensively, and overall cultural and political guiding principles also influence how individuals interpret presenteeism. A literature review is a qualitative study that requires subjective interpretation. This is illustrated by Cancelliere et al. [35] and Jensen [36] reporting differently from the same included studies, see e.g., Mills’ intervention [37].

Measurement: An additional source of bias is the measurement of presenteeism. It is measured both objectively and subjectively, by costs and questionnaires [33]. There is no standard universal measuring method for presenteeism, and standardised objective measurement tools are lacking [9]. Authors measure and assess presenteeism differently. The construct validity is low and apparently affecting in particular measuring methods negatively. Therefore, valid measurements on presenteeism are still missing [4].

Methodological appraisal: The quality evaluation of included studies is based on customized selection criteria, e.g., number of RCTs, their methodological assessment and own reported quality grading of studies. In addition, quality appraisal of the single studies included in our analysis are based on review-authors own grading of these studies; hence the report of methodologically quality might be difficult to interpret and thereby associated with some uncertainty [34].

Duplicates: As literature reviews, and review of reviews, are included, some reviews and single studies are evaluated more than once. To avoid double processing, we excluded two reviews [36, 38] from the analysis of Schröer et al. [48], one review [35] from White et al. [49] and one [39] from Wagner et al. [50]. In addition, several single studies were excluded from the rest of the material (see Supplement 3). However, the review-authors presents their results in different detail, the possibility of double processing of information cannot be ignored.

3 Results

3.1 Description of the material

In the main search, 2196 eligible publications were identified, of which 56 studies were retrieved in full-text and underwent detailed assessment according to the inclusion criteria. Finally, 14 studies were included in this analysis (see Fig. 1). Eleven studies were systematic reviews containing 262 single studies [35, 36, 38, 40–47]. The remaining three studies were review of reviews [48–50] (see Table 1). After elimination of duplicates, the material comprised 53 reviews and 1390 single studies. However, we were only able to identify 44 single studies (3%) reporting on presenteeism or productivity as an outcome variable.

Table 1
Description of included studies

First author, publication year, nationality Aims and study outcomes Methods

1) Brown et al. [38] Australia Aim: To examine the impact of PA on employee well-being and presenteeism.
Outcomes:Workplace well-being (including presenteeism). Systematic review: Search in six databases.
Included number of studies: n = 20.
Included designs: 13 intervention trials (eight RCT, five comparison trials). Seven observational studies (three cohort, four cross-sectional).
Participants and settings: Mixed employees from private and public organisations.
Assessment of methodological quality: Not reported.

2) Cancelliere et al. [35]Canada Aim: (1) To investigate whether WHPP are effective at improving presenteeism among workers.(2) To identify components of successful WHPP and identify risk factors for presenteeism.
Outcomes:Presenteeism. Systematic review: Search in nine databases.
Included number of studies: n = 14.
Included designs: Five RCT, five cluster RCT, one interrupted time series study, one cross-over study, one pre-post study, one quasi-experimental design study.
Participants and settings: A variety of employees and settings, e.g., white collar, blue collar, service personnel in production, office, and distribution.
Assessment of methodological quality: The Quality Assessment Tool for Quantitative Studies, consisting of six criteria. Studies rated as “weak”, “moderate” or “strong”.

3) Cochrane et al. [40]Ireland Aim: To investigate whether early multicomponent interventions promote work participation (or reduced absenteeism) in people with regional musculoskeletal pain.
Outcomes:
– Duration of sick leave.
– Time to RTW.
– Pain, disability, psychological functioning.
– Quality of life and fatigue. Systematic review and meta-analysis: Search in seven databases.
Included number of studies: n = 20.
Included designs: 19 RCT’s, one cluster RCT.
Participants and settings: Employees or outpatients with LBP, BP or other musculoskeletal problems. Settings were hospital, health care centres, clinics, or other workplaces.
Assessment of methodological quality: According to Cochrane guidelines. Quality of evidence for outcome variables assessed by the GRADE System.

4) Davies [41]UK Aim: 1. Determine if HP programs reduce presenteeism and absenteeism.
2. Ascertain whether there is a causal relationship between health promotion and improvements in individual health, and
3. Identify which interventions have a positive, negative or nil effect on improving health.
Outcomes:
– Presenteeism.
– Absenteeism.
– Health outcomes. Systematic literature review: Search in seven databases.
Included number of studies: n = 10.
Included designs: Six RCT or CT designs. Four observational studies.
Participants and settings: Employees from a variety of workplaces are included.
Assessment of methodological quality: Customized for this study: Four pre-defined core criteria for CT’s. A STROBE (Strengthening of Reporting Observational Studies in Epidemiology) checklist used to assess robustness of the studies.

5) Howarth et al. [42]UK Aim: The impact of digital health interventions on health-related outcomes in the workplace.
Outcomes:Physical, psychological, biological, behavioural, or work measures. E.g., lifestyle parameters, mental and physical health, job stress, satisfaction and presenteeism. A systematic review: Search in five databases.
Included number of studies: n = 22.
Included designs: RCT only.
Participants and settings: A variety of employees from different workplaces e.g., public and private offices, academic or hospital settings and manufacturing plants.
Assessment of methodological quality: The Cochrane risk of bias tool used to allocate a risk of bias classification of each study.

6) Jensen [36]Denmark Aim: Whether or how worksite nutritional interventions can improve productivity and firm profitability.
Outcomes:
– Absenteeism.
– Presenteeism. A systematic review: Search in seven databases.
Included number of studies: n = 30. 13/30 studies reported on direct profitable variables. 17/30 on health promotion outcomes indirectly.
Included designs: RCT’s, quasi-experimental designs, observational cross-sectional designs.
Participants and settings:>120 940 employees included in 13 studies on profitability, 1469 employees reported on health outcomes (17 studies). Employees from a variety of settings.
Assessment of methodological quality: Customized to this study. A quality appraisal scheme based on study design and clarity in the definition of interventions and environmental and outcome variables.

7) Kröll et al. [43]Germany Aim: To investigate the effectiveness of primary preventive interventions (FWA; flexible work arrangements) or secondary preventive interventions (SMT; stress management training) at work.
Outcomes:
– Psychological health.
– Job satisfaction, job performance, absenteeism. A review and meta-analysis: Search in three databases.
Included number of studies: n = 43.
Included designs: Mostly quasi-experimental designs and field studies.
Participants and settings: 22 882 employees from multiple organisations, office, psychological health care or education.
Assessment of methodological quality: Quality of primary studies measured via DIAD (study Design and Implementation Assessment Device).

8) Neuhaus et al. [44]Australia Aim: To review systematically the impact of activity-permissive workstations on office workers sedentary time and other outcomes.
Outcomes:
– Overall sedentary time. Health-risk biomarkers.
– Work performance indicators (e.g., presenteeism). Cultural, organisational outcomes.
– Feasibility outcomes. Systematic review and meta-analysis: Search in nine databases.
Included number of studies: n = 38.
Included designs: Pre
– post design: 19 laboratory experimental designs, 19 field
– based trials.
Participants and settings: n = 984 employees: office workers, university staff, “adults” and medical practitioners.
Assessment of methodological quality: Based on a published standard scoring system. Eight criteria relating to the report of methods and results (0-1).

9) Oakman et al. [45]Australia Aim: To analyse the most effective interventions for workers with persistent musculoskeletal pain (PMP) to remain productively employed.
Outcomes:
– Productivity, sick leave.
– Pain, job loss, cost benefit. Systematic review: Search in four databases.
Included number of articles: n = 18 (14 studies).
Included designs: 12 RCT’s, one cohort, and one pre-post study.
Participants and settings: Employees with musculoskeletal pain, number and setting not reported.
Assessment of methodological quality: The Cochrane Handbook guidelines for risk of bias. The quality for each outcome assessed by the GRADE tool.

10) Osilla et al. [46]USA Aim: To analyse the impact of worksite wellness programs on health and financial outcomes, and the effect of incentives on participation.
Outcomes:
– Exercise, diet, smoking, alcohol.
– Physiological markers.
– Mental health, absenteeism. Systematic review: Search in five databases.
Included number of studies: n = 33.
Included designs: 17 RCTs, 10 observational designs, six comparisons with non-random assignment.
Participants and settings: Medium to large worksites (>100 < 50 000 employees). A variety of industries were represented, e.g., service and manufacturing.
Assessment of methodological quality: Customized to this study. Categorized according to published procedure, e.g., random designs, prospective studies, observational studies.

11) Rongen et al. [47]The Netherlands Aim: To evaluate (1) the effectiveness of WHPP aimed at a healthy lifestyle on self-perceived health, work absence due to sickness, productivity at work, and work ability. (2) the influence of population and study characteristics, intervention content, and methodologic quality.
Outcomes:
– Work ability, productivity, sickness absence.
– Self-perceived health. Systematic review and meta-analysis: Search in three databases.
Included number of studies: n = 18. Including 21 interventions.
Included designs: 18 RCTs.
Participants and settings:>5750 employees (range 40
– 860) from a variety of workplace settings.
Assessment of methodological quality: According to Cochrane guidelines: A predefined nine-item checklist based on the guidelines tool for assessing risk of bias.

12) Schröer et al. [48]Germany Aim: To summarise efficacy and cost-effectiveness of different workplace lifestyle interventions for promoting healthy lifestyle, preventing diseases, and reducing health care costs.
Outcomes:
– Weight related outcomes, PA, physical fitness, and dietary related outcomes.
– Economic outcomes. Systematic review of reviews: Search in six databases.
Included number of studies: n = 15 reviews, including 379 single studies.
Included designs: Systematic reviews, including a variety of interventional and observational studies. Of the 15 reviews, nine contained RCT’s only.
Participants and settings: The sample size varied from 10
– 48 835 participants from different work settings.
Assessment of methodological quality: According to the Cochrane Collaboration quality criteria for systematic reviews.

13) White et al. [49]Canada Aim: Identify PA and exercise interventions that most effectively reduce absenteeism, improve productivity, or increase financial outcomes.
Outcomes:
– Workplace absence.
– Worker productivity.
– Financial cost. Systematic review of reviews: Search in seven databases and grey literature. A Stakeholder-centred best evidence synthesis of systematic reviews.
Included number of studies: n = 18 reviews, including 511 articles from 490 primary studies.
Included design: Systematic reviews including a variety of controlled (R)CT, pre-post, and observational designs in the primary studies.
Participants and settings: Number of participants not reported. Participants grouped in “general workers” and “workers off-work” at baseline. Here, we report only “general workers”.
Assessment of methodological quality: Customized quality assessment form including 18 questions using a modified version of the Health-evidence.ca quality assessment tool, the EBM Glasgow Checklist for Systematic Reviews, and the AMSTAR methodological quality guidelines.

14) Wagner et al. [50]Canada Aim: To determine the level of evidence supporting mental health interventions as valuable to work outcomes.
Outcomes:
– Absenteeism.
– Worker productivity.
– Financial outcomes. Systematic review of reviews: Searches in seven databases and in grey literature. See [49]. This study is part of a larger synthesis appraising and summarizing workplace interventions to address previously identified modifiable risk factors of work absence across health conditions.
Included number of studies: n = 14 reviews, including 346 studies.
Included design: Systematic reviews including a variety of controlled (R)CT, pre-post and observational designs in the primary studies.
Participants and settings: Number of participants not reported. Varieties of worksites represented.
Assessment of methodological quality: As reported by White et al. [49].

First author, publication year, nationality	Aims and study outcomes	Methods
1) Brown et al. [38] Australia	Aim: To examine the impact of PA on employee well-being and presenteeism. Outcomes:Workplace well-being (including presenteeism).	Systematic review: Search in six databases. Included number of studies: n = 20. Included designs: 13 intervention trials (eight RCT, five comparison trials). Seven observational studies (three cohort, four cross-sectional). Participants and settings: Mixed employees from private and public organisations. Assessment of methodological quality: Not reported.
2) Cancelliere et al. [35]Canada	Aim: (1) To investigate whether WHPP are effective at improving presenteeism among workers.(2) To identify components of successful WHPP and identify risk factors for presenteeism. Outcomes:Presenteeism.	Systematic review: Search in nine databases. Included number of studies: n = 14. Included designs: Five RCT, five cluster RCT, one interrupted time series study, one cross-over study, one pre-post study, one quasi-experimental design study. Participants and settings: A variety of employees and settings, e.g., white collar, blue collar, service personnel in production, office, and distribution. Assessment of methodological quality: The Quality Assessment Tool for Quantitative Studies, consisting of six criteria. Studies rated as “weak”, “moderate” or “strong”.
3) Cochrane et al. [40]Ireland	Aim: To investigate whether early multicomponent interventions promote work participation (or reduced absenteeism) in people with regional musculoskeletal pain. Outcomes: – Duration of sick leave. – Time to RTW. – Pain, disability, psychological functioning. – Quality of life and fatigue.	Systematic review and meta-analysis: Search in seven databases. Included number of studies: n = 20. Included designs: 19 RCT’s, one cluster RCT. Participants and settings: Employees or outpatients with LBP, BP or other musculoskeletal problems. Settings were hospital, health care centres, clinics, or other workplaces. Assessment of methodological quality: According to Cochrane guidelines. Quality of evidence for outcome variables assessed by the GRADE System.
4) Davies [41]UK	Aim: 1. Determine if HP programs reduce presenteeism and absenteeism. 2. Ascertain whether there is a causal relationship between health promotion and improvements in individual health, and 3. Identify which interventions have a positive, negative or nil effect on improving health. Outcomes: – Presenteeism. – Absenteeism. – Health outcomes.	Systematic literature review: Search in seven databases. Included number of studies: n = 10. Included designs: Six RCT or CT designs. Four observational studies. Participants and settings: Employees from a variety of workplaces are included. Assessment of methodological quality: Customized for this study: Four pre-defined core criteria for CT’s. A STROBE (Strengthening of Reporting Observational Studies in Epidemiology) checklist used to assess robustness of the studies.
5) Howarth et al. [42]UK	Aim: The impact of digital health interventions on health-related outcomes in the workplace. Outcomes:Physical, psychological, biological, behavioural, or work measures. E.g., lifestyle parameters, mental and physical health, job stress, satisfaction and presenteeism.	A systematic review: Search in five databases. Included number of studies: n = 22. Included designs: RCT only. Participants and settings: A variety of employees from different workplaces e.g., public and private offices, academic or hospital settings and manufacturing plants. Assessment of methodological quality: The Cochrane risk of bias tool used to allocate a risk of bias classification of each study.
6) Jensen [36]Denmark	Aim: Whether or how worksite nutritional interventions can improve productivity and firm profitability. Outcomes: – Absenteeism. – Presenteeism.	A systematic review: Search in seven databases. Included number of studies: n = 30. 13/30 studies reported on direct profitable variables. 17/30 on health promotion outcomes indirectly. Included designs: RCT’s, quasi-experimental designs, observational cross-sectional designs. Participants and settings:>120 940 employees included in 13 studies on profitability, 1469 employees reported on health outcomes (17 studies). Employees from a variety of settings. Assessment of methodological quality: Customized to this study. A quality appraisal scheme based on study design and clarity in the definition of interventions and environmental and outcome variables.
7) Kröll et al. [43]Germany	Aim: To investigate the effectiveness of primary preventive interventions (FWA; flexible work arrangements) or secondary preventive interventions (SMT; stress management training) at work. Outcomes: – Psychological health. – Job satisfaction, job performance, absenteeism.	A review and meta-analysis: Search in three databases. Included number of studies: n = 43. Included designs: Mostly quasi-experimental designs and field studies. Participants and settings: 22 882 employees from multiple organisations, office, psychological health care or education. Assessment of methodological quality: Quality of primary studies measured via DIAD (study Design and Implementation Assessment Device).
8) Neuhaus et al. [44]Australia	Aim: To review systematically the impact of activity-permissive workstations on office workers sedentary time and other outcomes. Outcomes: – Overall sedentary time. Health-risk biomarkers. – Work performance indicators (e.g., presenteeism). Cultural, organisational outcomes. – Feasibility outcomes.	Systematic review and meta-analysis: Search in nine databases. Included number of studies: n = 38. Included designs: Pre – post design: 19 laboratory experimental designs, 19 field – based trials. Participants and settings: n = 984 employees: office workers, university staff, “adults” and medical practitioners. Assessment of methodological quality: Based on a published standard scoring system. Eight criteria relating to the report of methods and results (0-1).
9) Oakman et al. [45]Australia	Aim: To analyse the most effective interventions for workers with persistent musculoskeletal pain (PMP) to remain productively employed. Outcomes: – Productivity, sick leave. – Pain, job loss, cost benefit.	Systematic review: Search in four databases. Included number of articles: n = 18 (14 studies). Included designs: 12 RCT’s, one cohort, and one pre-post study. Participants and settings: Employees with musculoskeletal pain, number and setting not reported. Assessment of methodological quality: The Cochrane Handbook guidelines for risk of bias. The quality for each outcome assessed by the GRADE tool.
10) Osilla et al. [46]USA	Aim: To analyse the impact of worksite wellness programs on health and financial outcomes, and the effect of incentives on participation. Outcomes: – Exercise, diet, smoking, alcohol. – Physiological markers. – Mental health, absenteeism.	Systematic review: Search in five databases. Included number of studies: n = 33. Included designs: 17 RCTs, 10 observational designs, six comparisons with non-random assignment. Participants and settings: Medium to large worksites (>100 < 50 000 employees). A variety of industries were represented, e.g., service and manufacturing. Assessment of methodological quality: Customized to this study. Categorized according to published procedure, e.g., random designs, prospective studies, observational studies.
11) Rongen et al. [47]The Netherlands	Aim: To evaluate (1) the effectiveness of WHPP aimed at a healthy lifestyle on self-perceived health, work absence due to sickness, productivity at work, and work ability. (2) the influence of population and study characteristics, intervention content, and methodologic quality. Outcomes: – Work ability, productivity, sickness absence. – Self-perceived health.	Systematic review and meta-analysis: Search in three databases. Included number of studies: n = 18. Including 21 interventions. Included designs: 18 RCTs. Participants and settings:>5750 employees (range 40 – 860) from a variety of workplace settings. Assessment of methodological quality: According to Cochrane guidelines: A predefined nine-item checklist based on the guidelines tool for assessing risk of bias.
12) Schröer et al. [48]Germany	Aim: To summarise efficacy and cost-effectiveness of different workplace lifestyle interventions for promoting healthy lifestyle, preventing diseases, and reducing health care costs. Outcomes: – Weight related outcomes, PA, physical fitness, and dietary related outcomes. – Economic outcomes.	Systematic review of reviews: Search in six databases. Included number of studies: n = 15 reviews, including 379 single studies. Included designs: Systematic reviews, including a variety of interventional and observational studies. Of the 15 reviews, nine contained RCT’s only. Participants and settings: The sample size varied from 10 – 48 835 participants from different work settings. Assessment of methodological quality: According to the Cochrane Collaboration quality criteria for systematic reviews.
13) White et al. [49]Canada	Aim: Identify PA and exercise interventions that most effectively reduce absenteeism, improve productivity, or increase financial outcomes. Outcomes: – Workplace absence. – Worker productivity. – Financial cost.	Systematic review of reviews: Search in seven databases and grey literature. A Stakeholder-centred best evidence synthesis of systematic reviews. Included number of studies: n = 18 reviews, including 511 articles from 490 primary studies. Included design: Systematic reviews including a variety of controlled (R)CT, pre-post, and observational designs in the primary studies. Participants and settings: Number of participants not reported. Participants grouped in “general workers” and “workers off-work” at baseline. Here, we report only “general workers”. Assessment of methodological quality: Customized quality assessment form including 18 questions using a modified version of the Health-evidence.ca quality assessment tool, the EBM Glasgow Checklist for Systematic Reviews, and the AMSTAR methodological quality guidelines.
14) Wagner et al. [50]Canada	Aim: To determine the level of evidence supporting mental health interventions as valuable to work outcomes. Outcomes: – Absenteeism. – Worker productivity. – Financial outcomes.	Systematic review of reviews: Searches in seven databases and in grey literature. See [49]. This study is part of a larger synthesis appraising and summarizing workplace interventions to address previously identified modifiable risk factors of work absence across health conditions. Included number of studies: n = 14 reviews, including 346 studies. Included design: Systematic reviews including a variety of controlled (R)CT, pre-post and observational designs in the primary studies. Participants and settings: Number of participants not reported. Varieties of worksites represented. Assessment of methodological quality: As reported by White et al. [49].

BP: back pain, LBP: low back pain, PA: physical activity, (R)CT: (randomized) controlled trial, RTW: return to work, WHPP: worksite health promotion program(s).

About two-thirds of the single studies comprised a RCT or quasi-experimental design. Other frequently used designs were controlled trials (CT), cohort or observational/epidemiological designs and some cross-sectional studies. The included reviews originated from Australia [38, 44, 45], Canada [35, 49, 50], USA [46] and Europe [36, 40–43, 47, 48] (see Table 1). The participants were employees from a variety of worksite settings, both public and private, in health, educational and service settings. Two studies focused specifically on employee subgroups with musculoskeletal pain [40, 45] and Wagner et al. [50] studied employees with mental health problems (e.g., depression and anxiety).

3.2 Description of concepts and theoretical frameworks

In our material, all studies but two [41, 46] (86%) described the concept presenteeism, however in different ways (see Table 2). 7/12 studies described presenteeism related to health or lifestyle. Of these, 4/12 studies (33%) focused productivity loss due to reduced health or lifestyle [45, 47, 49, 50]. Brown et al. [38] and Cancelliere et al. [35] added a work performance aspect to their concept of presenteeism (e.g., performing below par) and Jensen [36] described presenteeism as attending work while ill, without focusing the consequences of the act. Cancelliere et al. [35] was the only study having presenteeism as the only outcome. 5/12 studies (42%) specified the productivity dimension as a work-related outcome (e.g., job performance), not solely related to health [42–44], while two studies (17%) reported economic consequences of productivity loss [40, 48] (see Table 2).

Table 2
Descriptions of the concept presenteeism

Level Categories Studies N (%) Description of the concept presenteeism

Individual level I Related to health and lifestyle 7 (58) • Being present at work, but limited in some aspects of job performance by a health problem [35]
• Attending work while sick [36]
• Being at work despite poor health and performing below par [38]
• Productivity loss due to muscle pain [45]
• Productivity loss due to unhealthy lifestyle/health behavior [47]
• Productivity losses due to different health conditions [49]
• Productivity losses due to mental health and psychological health conditions [50]

Organisational level II Related to work factors 3 (25) • Time spent at work with decreased levels of productivity [42]
• Decreased job performance e.g., job satisfaction and productivity due to work pressure [43]
• Categorized as work-related outcome/work performance indicator [44]

Related to cost 2 (17) • Cost of productivity loss due to musculoskeletal pain [40]• Economic consequences due to productivity losses [48]

Socio-political level III No categories identified

Level	Categories	Studies N (%)	Description of the concept presenteeism
Individual level I	Related to health and lifestyle	7 (58)	• Being present at work, but limited in some aspects of job performance by a health problem [35] • Attending work while sick [36] • Being at work despite poor health and performing below par [38] • Productivity loss due to muscle pain [45] • Productivity loss due to unhealthy lifestyle/health behavior [47] • Productivity losses due to different health conditions [49] • Productivity losses due to mental health and psychological health conditions [50]
Organisational level II	Related to work factors	3 (25)	• Time spent at work with decreased levels of productivity [42] • Decreased job performance e.g., job satisfaction and productivity due to work pressure [43] • Categorized as work-related outcome/work performance indicator [44]
	Related to cost	2 (17)	• Cost of productivity loss due to musculoskeletal pain [40]• Economic consequences due to productivity losses [48]
Socio-political level III	No categories identified

Only two studies (14%) described presenteeism/productivity loss in a theoretical context. Kröll, Doebler and Nuesch [43] used psychological stress theory (the Conservation of Resources (COR) theory) to predict and understand effects of primary (flexible work arrangements) interventions versus secondary preventive (stress management training) interventions, while Oakman et al. [45] used a macro-ergonomics framework, the Sociotechnical System Theory, to analyse their results on productivity in workers with persistent musculoskeletal pain (PMP) (see Supplement 3).

3.3 Factors affecting presenteeism

All but two studies [44, 46] reported different factors affecting presenteeism, in sum 75 factors. In this analysis, we categorized the factors according to three levels (see Table 3).

Table 3
Factors reported to affect presenteeism

Level Categories Factors affecting presenteeism / productivity/ cost

Individual level I Individual lifestyle factors16 (21%) • Behavioral factors: being overweight, poor diet, smoking, physical inactivity [35, 42, 47] and insomnia [42]• Life-style health issues affecting the economic position of organizations [48]• Decreased physical activity [49, 50]

Physical and mental health factors33 (44%) • Health conditions: a diversity of physical illness and diagnoses (e.g., CHD, diabetes, cancer, metabolic syndrome), mental health (e.g., depression, anxiety) [35]• Health risk factors leading to reduced cost and productivity [36]• Psychological, mental health and employee-wellbeing [38, 43]• Muscle skeletal pain, e.g., low back/shoulder/neck/forearm pain and knee pain [40]• Personal issues: chronic health problems and mental health e.g., depression [41, 42]• Persistent musculoskeletal pain with physical, emotional, and social impact [45]• Modifiable worker factors including e.g., decreased poor general physical and mental health [49, 50]

Immaterial,biological,psychological, factors10 (13%) • Money and power related to employee’s wellbeing [43]• Immaterial resources e.g., reserves like energy, time, knowledge [43]• Working age [45]• Negative health/disability perception or negative recovery expectations [49, 50]

Organisational level II Psychosocial-environmental, management, financial factors16 (21%) • Poor relations with co-workers [35]• Poor physical work environment [35]• Poor relations with leaders and management [35]• Emotional distress, high job stress [35, 43, 49, 50]• Lack of resources, financial stressors [41]• Psychosocial outcome measures, lack of family support [49, 50]

Socio-political level III Socio-political factors0 (0%) None reported

Level	Categories	Factors affecting presenteeism / productivity/ cost
Individual level I	Individual lifestyle factors16 (21%)	• Behavioral factors: being overweight, poor diet, smoking, physical inactivity [35, 42, 47] and insomnia [42]• Life-style health issues affecting the economic position of organizations [48]• Decreased physical activity [49, 50]
	Physical and mental health factors33 (44%)	• Health conditions: a diversity of physical illness and diagnoses (e.g., CHD, diabetes, cancer, metabolic syndrome), mental health (e.g., depression, anxiety) [35]• Health risk factors leading to reduced cost and productivity [36]• Psychological, mental health and employee-wellbeing [38, 43]• Muscle skeletal pain, e.g., low back/shoulder/neck/forearm pain and knee pain [40]• Personal issues: chronic health problems and mental health e.g., depression [41, 42]• Persistent musculoskeletal pain with physical, emotional, and social impact [45]• Modifiable worker factors including e.g., decreased poor general physical and mental health [49, 50]
	Immaterial,biological,psychological, factors10 (13%)	• Money and power related to employee’s wellbeing [43]• Immaterial resources e.g., reserves like energy, time, knowledge [43]• Working age [45]• Negative health/disability perception or negative recovery expectations [49, 50]
Organisational level II	Psychosocial-environmental, management, financial factors16 (21%)	• Poor relations with co-workers [35]• Poor physical work environment [35]• Poor relations with leaders and management [35]• Emotional distress, high job stress [35, 43, 49, 50]• Lack of resources, financial stressors [41]• Psychosocial outcome measures, lack of family support [49, 50]
Socio-political level III	Socio-political factors0 (0%)	None reported

Individual level (I): 12 studies (100%) reported factors influencing presenteeism / productivity at this level. Most of the factors (78%) addressed lifestyle (e.g., physical inactivity, unhealthy eating, smoking) [35, 42, 47–50], physical and mental health (e.g., musculoskeletal pain, and depression, anxiety, wellbeing) [35, 36, 38, 40–43, 45, 49, 50] and immaterial, psychological, and biological factors (e.g., money, working age, disability perception) [43, 45, 48–50].

Organisational level (II): 5/12 studies (42%) reported on this level. 21% of the reported factors affecting presenteeism addressed psychosocial and/or physical environmental factors (e.g., job stress factors, poor relations with co-workers) [35, 43, 49, 50] or work organisation and management factors (e.g., lack of resources, poor relation to leaders) [35, 41].

Socio-political level (III): no factors were found at this level (see Table 3).

3.4 WHPP affecting presenteeism or productivity loss

All studies except two [41, 46] described the intervention programs investigated to influence presenteeism/productivity, in total 45 programs (see Tables 4 and 5), of which 22 (49%) influenced our outcome variables positively. In particular, the study of Cancelliere et al. [34] contributed strongly reporting thirteen programs (31%) impacting presenteeism, of which nine were graded with high or moderate evidence of effect (see Table 4).

Table 4
Worksite health promotion programs affecting presenteeism positively

Level Programs Type of worksite health promotion programs Impact Quality

Individual level Single component programs10 (45%) • Worksite exercise program (health, work) [35]• Light change intervention (insomnia) [35]• Telephone therapy (depressed people) [35]• Changed rest break schedules (muscle pain, stress) [35]• Two treadmill desk interventions (work factors) [44]• HAD intervention (sedentary time, others) [44]• Ergonomic workplace assessment (LBP) [45]• Exercise based rehabilitation (LBP) [45]• Pedometer-based active walking on/off work (lifestyle) [47] Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+) HH/MH/MH/MMMLLF/M

Multi component programs None reported

Multilevel (mixed levels) Single component programs None reported

Multi component programs 12 (55%) • Educational job stress program for supervisors (subordinates mental health, performance) [35]• Multicomponent health program (health) [35]• Lifestyle interventions via e-mail (diet, pa) [35]• Multidisciplinary occupational health programs (health) [35]• Participatory processes interventions: teamwork with employees, managers, HES’ staff, researchers* (MSP) [35]• Health risk assessment. Multiple questions+biomarkers (health) [36] Awareness + weight management in groups+environmental initiatives. Health risk assessment (health) [36]• Web-based multicomponent PA, nutritional program+support (lifestyle) [38]• Digital CBT+health assessment, tailored feedback, online support against insomnia (three levels) (health) [42]• SMT, relaxation and mindfulness (work, health) [43]• Multicomponent intervention; nutritional and environmental components (lifestyle, cost) [48]• High intensity mental health intervention; CBT via telephoneor psycho-educational workbook. Access to clinical treatment outside worksite (depressed) [50] Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+) HH/MH/MH/MH/MH/MH/MH/MHHL/FH

Sum programs 22(100%) 19(86%)H/M3(14%)L/F

Level	Programs	Type of worksite health promotion programs	Impact	Quality
Individual level	Single component programs10 (45%)	• Worksite exercise program (health, work) [35]• Light change intervention (insomnia) [35]• Telephone therapy (depressed people) [35]• Changed rest break schedules (muscle pain, stress) [35]• Two treadmill desk interventions (work factors) [44]• HAD intervention (sedentary time, others) [44]• Ergonomic workplace assessment (LBP) [45]• Exercise based rehabilitation (LBP) [45]• Pedometer-based active walking on/off work (lifestyle) [47]	Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)	HH/MH/MH/MMMLLF/M
	Multi component programs	None reported
Multilevel (mixed levels)	Single component programs	None reported
	Multi component programs 12 (55%)	• Educational job stress program for supervisors (subordinates mental health, performance) [35]• Multicomponent health program (health) [35]• Lifestyle interventions via e-mail (diet, pa) [35]• Multidisciplinary occupational health programs (health) [35]• Participatory processes interventions: teamwork with employees, managers, HES’ staff, researchers* (MSP) [35]• Health risk assessment. Multiple questions+biomarkers (health) [36] Awareness + weight management in groups+environmental initiatives. Health risk assessment (health) [36]• Web-based multicomponent PA, nutritional program+support (lifestyle) [38]• Digital CBT+health assessment, tailored feedback, online support against insomnia (three levels) (health) [42]• SMT, relaxation and mindfulness (work, health) [43]• Multicomponent intervention; nutritional and environmental components (lifestyle, cost) [48]• High intensity mental health intervention; CBT via telephoneor psycho-educational workbook. Access to clinical treatment outside worksite (depressed) [50]	Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)Pos (+)	HH/MH/MH/MH/MH/MH/MH/MHHL/FH
		Sum programs	22(100%)	19(86%)H/M3(14%)L/F

CBT: cognitive behavioural therapy, HAD: fully height adjustable desk, HADM: height adjustable desk for monitor only, HES: health environment security, IC: inconclusive, LBP: low back pain, MSP: musculoskeletal pain, SMT: stress management training, PA: physical activity. H: high, M: moderate, F: fair, L: low methodological quality.

Table 5

Worksite health promotion programs not affecting presenteeism

Level	Programs	Type of worksite health promotion programs	Impact	Quality
Individual level	Single componentprograms 5 (22%)	• Equipment change: computer mouse with feedback (work) [35]• Two HADM interventions (sedentary time, others) [44]• Workplace rehabilitation (LBP) [45]• Web-based PA programs (lifestyle) [47]• Simple physical exercise intervention (lifestyle) [49]	No (–)No (–)No (–)No (–)No (–)	H/MMLFH
	Multi componentprograms 4 (17%)	• Specific resistance training and all-round PA [35]• HADM interventions+additional actions (sedentary time, others) [44]• Workplace rehabilitation (MSP) [45]• Web-based PA+weight loss program (health) [47]	No (–)No (–)No (–)No (–)	H/MHLF
Multilevel (mixed levels)	Single component programs	None reported
	Multi component programs 10 (54%)	• Multidimensional LBP program (health) [35]• Reduced work hours versus compulsory worksite exercise (biomarkers, health) [35]• Multilevel targeted occupational therapy (rheumatism) [45]• Multilevel intervention (rheumatism) [45]• Web-based PA, weight, health risk assessment, feedback and /or counselling (health) [47]• Multilevel stepped care (MSP) [40]• Rheumatologist treatment+CBT (MSP) [40]• Case manager-lead tailored work rehabilitation (MSP) [40]• Five FWA (flexible work arrangement) interventions (work) [43]• Complex psychosocial intervention (lifestyle) [49]	No (–)No (–)No (–)No (–)No (–)ICICICICIC	H/MH/MLLL/FLLLL/FH
		Sum programs	23 (100%)	8(35%)H/M15(65%)L/F

CBT: cognitive behavioural therapy, HAD: fully height adjustable desk, HADM: height adjustable desk for monitor only, HES: health environment security, IC: inconclusive, LBP: low back pain, MSP: musculoskeletal pain. SMT: stress management training, PA: physical activity. H: high, M: moderate, F: fair, L: low methodological quality.

The programs mostly (90%) focused individual health, lifestyle, or ergonomic conditions, and to a very small extent organisational factors (i.e., changed rest breaks [35]), containing a variety of components and active ingredients. 12/22 programs (55%) showing positive impact on presenteeism/productivity, were multicomponent programs implemented at two or more levels [35, 36, 38, 42, 43, 48, 50]. Mixed programs, involving individual and organisational levels seems to influence presenteeism favourably (e.g., participatory processes interventions implemented as teamwork including several organisational levels [35], digital cognitive behavioural therapy including health risk assessment, tailored feedback, and other actions on three different levels [42]).

The remaining 10/22 programs (45%) reporting positive effect were single component, implemented on an individual level only [35, 44, 45, 47] (see Table 4). Many of these programs were offered to risk populations e.g., exercise rehabilitation for musculoskeletal pain, light therapy for insomnia, telephone therapy for depressed and dietary behavioural interventions. However, 14/45 (31%) both single and multicomponent programs failed to influence presenteeism/productivity significantly. Nine studies (20%) reported inconclusive results [40, 43, 49] (see Table 5).

3.5 Methodological quality of included studies

According to our quality assessment, 8/14 (57%) studies were rated with high methodological quality [35, 40, 42, 45, 47–50], 3/14 (21%) with moderate quality [36, 43, 44], and 3/14 (21%) with low quality [38, 41, 46]. Most of our included studies, we evaluated to be moderate to high (see Supplement 2).

Further analysis revealed 30/45 (67%) single studies rated, by review-authors, as moderate to high in methodological quality. 14/45 (31%) studies were graded with low quality, often due to small samples, inadequate designs, or other implementation weaknesses (see Tables 4 and 5). Most of the studies 19/22 (86%) reporting positive effect on presenteeism were graded with moderate to high methodological quality (see Fig. 2), indicating a strong relationship between positive intervention effect and high methodological quality. Studies reporting no effect on presenteeism showed mixed methodological grading (see Fig. 2), however 8/9 studies reporting inconclusive results were graded with low methodological quality.

Fig. 2

Cluster analysis of studies reporting outcome effect related to methodological quality. N = 45. Q1: positive outcome effect/high methodological quality. Q2: positive outcome effect/low methodological quality. Q3: no or inconclusive outcome effect/low methodological quality. Q4: no or inconclusive effect/high methodological quality.

4 Discussion

4.1 Definitions of the concept presenteeism

In our material, few studies defined the concept of presenteeism precisely. In more than half of the studies presenteeism were related to productivity loss or performance limitations due to individual health or lifestyle impairments. These findings correspond with previous research in the field [3, 4, 9]. Few studies described presenteeism related to work factors or cost, and none addressed socio-political level e.g., policy or legislation. A major challenge facing this diversity of concept definitions, is low concept validity of presenteeism, leading to different measurements and lack of useful theoretical frameworks. As also shown in former research [3, 4, 13]. However, a significant increase in the research field of presenteeism focusing concept development and theoretically framework building, has emerged lately [3, 4, 13, 18–21, 24]. Fully agreement on the definition of presenteeism is still not received, but increasingly more researchers agree on a definition of presenteeism as a behaviour choosing attending work while ill [3, 9, 26]. The way the concept is defined and measured greatly influences the organisation and facilitation of work and working conditions for employees with impaired health [26]. Although the concept mostly has been viewed negatively [3], positive consequences of presenteeism have increasingly been demonstrated [18, 20]. This duality of presenteeism having both positive and negative consequences, are shown in recent research through different experiences with changed working conditions e.g., increased telework during the COVID-19 pandemic [19, 21, 24]. However, whether increased telework should be experienced positively or negatively depends on a number of factors linked to the severity of the illness, personality factors, type of work, working conditions and organisational factors (e.g., management, job control, job demands) [22, 26]. Karanika-Murray and Biron [20] elaborates this in their recently developed framework stating that presenteeism behaviour is a choice, adaptive in nature balancing health and work performance.

In the present study, there is a general lack of theory. Only two studies base their analysis on a theoretical framework [43, 45]. Kröll et al. [43] draw on conservation of resources theory (COR) in their meta-analysis of flexible work arrangements (FWA) and stress management training (SMT) reporting that both FWAs and SMT correlated positively with psychological health and job satisfaction. The COR theory provide a framework for understanding job stress and has proven lately to be an important contributor in the development of a theory of presenteeism [20, 26].

Recently, there has been a conceptual development of presenteeism as a dynamic, adaptive, optional behaviour, which is not reflected in our material describing presenteeism mainly by its negative consequences due to health or lifestyle, e.g., productivity loss. More theory-based research is needed to clarify the many questions still not answered related to the concept of presenteeism.

4.2 Factors identified to influence presenteeism

As reported the concept of presenteeism is multifactorial and inherently complex, which may explain the variety of factors affecting presenteeism revealed in this analysis. As expected, most factors addressed individual lifestyle and poor physical and psychological health conditions, also confirmed by others [18]. Meta-analytical findings emphasise in particular the importance of mental health (depression, stress, burnout) and general ill health as important in the decision of presenteeism behaviour [18].

In our analysis, few factors were linked to organisational and/or socio-political level, though many scholars [e.g., 3, 12, 13] highlight the importance of these factors as causes of presenteeism. Miraglia and Johns [18] reported several work factors to be of importance e.g., strict work absence policies, high workload, understaffing. However, some of our studies reported high job stress (e.g., lack of resources, poor relations with co-workers, management, and leaders) affecting the decision to choose either presenteeism or health-related absence. This corresponds with earlier findings, using the COR conceptual framework to analyse and explain presenteeism [20]. Chen et al. [25] showed that different job and personal resources (social support) could buffer the negative effects of job stressors on health-related productivity losses. Personal and contextual factors also become important, as shown in our analysis (e.g., money and power, expectations, knowledge, or physical work environment), being motivational components connected to presenteeism [3, 4]. Contextual factors may be related to work and organisational characteristics, like job satisfaction, expectations, belonging and rewards as well as job stress [3]. All these factors count in the decision of going to work or staying at home.

4.3 WHPP influencing presenteeism

The analysis of this review revealed a large number of different WHPP implemented to impact presenteeism or related outcomes. The programs varied widely regarding design, content, duration, outcomes, and implementation, also shown by other researchers [9, 51]. However, the literature investigating presenteeism interventions is scarce [5, 26], mostly targeting physical and mental health and stressful work environments, also found in our study. Half of the programs in our material showed positive effects on presenteeism. Cancelliere et al. [35] accounted for a substantial part of these effective programs, all rated with moderate to high methodologically quality. Other key features mentioned in their study were frequent use of health risk assessments, also reported by others in our sample [36, 42] and tailored programs specifically to the participants needs.

In addition, only multicomponent programs implemented at multi levels showed positive effects on presenteeism also supported in health promotion theory and research [3, 4] and scientifically based recommendations for health promotion programs [30, 52].

Most of our multicomponent studies showing positive effects were rated with moderate to high methodological quality. This is supported in implementation science claiming that high implementation quality increases the outcome effect of interventions [29].

In contrast, less than half of the implemented programs showing positive impact were single component, addressing presenteeism by often targeting individual health and/or lifestyle problems implemented on an individual level. Our study showed mixed results, while Oakman et al. [45] stated that “in general, individually focused interventions are likely to have no significant impact on improving productivity”, Cancelliere et al. [35] and Kröll et al. [43] reported beneficial effect on presenteeism of simple worksite programs addressing depression, job stress or other mental conditions. Both studies focused on simplicity by implementing single component programs, a strategy also supported by earlier research of Hutchinson and Wilson [53], finding that interventions focusing on one main area of change (e.g., health, diet, or PA) were associated with larger mean effect sizes. A reflection to make when focusing single component programs addressing small samples at risk, is the lack of focus on causality in the organisation, and thereby absence of strategies as prevention, maintenance, and sustainability of programs. Several studies did not contain identifiable program factors that were clearly pervasive. Reasons may include a lack of fundamental anchoring at the organisational and management level as well as methodological problems and barriers such as small groups or poor implementation of the programs [29]. Implementation failure might give negative results in an otherwise successful intervention (low participation rate, response rates, and intervention adherence) [35].

The implications of our analysis are highlighting presenteeism as subject to growing interest, practice, and financial control among leaders. In particular the relationship between employers and employees has received attention during the pandemic [14, 24, 26], e.g., Gerich [24] suggests supportive managerial strategies to handle employees’ self-endangering behaviour (sickness presenteeism) in the time of increased telework.

About half of the reported interventions (approximately 45%) in our study were implemented digitally (e.g., telephone therapy, web-based lifestyle programs, exercise programs) or could easily be adapted to a more remote working situation (height adjustable desks, exercise programs). This might be promising in the light of the pandemic, with increased numbers of employees working remote, or with flexible work arrangements.

In summary, our findings support a common assumption, that multicomponent interventions implemented at several levels, shows good evidence of positive effect. However, despite extensive research on WHPP to reduce worksite health problems, there is no consensus as to which WHPP have the best effect on presenteeism or what programs to recommend.

4.4 Strengths and weaknesses

This study has several strengths. Search in seven independent sources increase the precision of eligible study inclusion. Reporting all steps in the comprehensive review using PRISMA 2020 guidelines make the process reproduceable. Assessing and appraising methodological quality in included reviews as well as included relevant single studies also seem to be a strength. The in-depth analysis of the material revealing that only 3% of included studies actually define, describe, or measures the concept of presenteeism, seem on one hand, to represent a strength. On the other hand, this finding also represent a selection bias which is a weakness. Due to selection bias, there is not sufficient scientific basis to draw firm conclusions. The included studies cannot be directly compared due to validity considerations. Therefore, the interpretation must be done with caution and in general terms only, also affecting the consequences for both conclusions and ability to generalize our results.

Methodological quality in the included studies was found to be moderate to high, but some single studies within our material may not be of sufficient quality or they lack implementation strategies in accordance with recommended guidelines.

Presenteeism seems to have high “face-validity”, indicating that stakeholders, i.e., leaders, employees, and researchers, consider the concept to be a kind of health-related productivity loss. Our results reveal that few have a clear definition of the concept, and half of the studies describe the concept as “productivity loss due to health or lifestyle problems”. Furthermore, in our material, only two studies contained analysis of data within a theoretical framework, thus challenging internal as well as external validity.

Our sampling was mostly completed before the COVID-19 pandemic, and thus it does not include any reviews of WHPP impacting presenteeism from the pandemic. Recent research on telework, telecommuting and remote work, indicate increased sickness presenteeism and basic changes in working life. Our conclusions may be narrowed due to this.

4.5 Suggestions for further research

In the field of research, there are several challenges that invoke thorough elaboration: the first priority being to establish a common understanding and definition of the concept of presenteeism. Then choosing relevant theories appropriate to the context explored. Furthermore, the importance of developing a common unified language, measurements and understanding of what influences presenteeism i.e., competence to use the right interventions, is significant. In the light of the pandemic, changes in work conditions and contextual factors are observed, resulting in an expanding of the concept presenteeism, e.g., positively, and negatively effects of presenteeism. This should be further explored.

Since both single-component and multi-component program interventions are of importance, more research is needed on analysis of intervention programs to fully comprehend the concept. The incorporation of proper program designs remains a challenge.

In practice, there is insufficient knowledge of how to identify different needs at all levels (e.g., screenings) to enable tailored programs to specific subgroups.

Lack of knowledge and competence about implementation of sustainable interventions in complex problem areas and settings, still remains.

Single component programs implemented at individual subgroup level can be of great importance to the individual with health problems but must not be understood as “WHPP affecting public health”. Such measures do not bring the field further but should be interpreted as rehabilitation programs. They cannot be up-scaled and sustainable “in the real world”.

The field is complex and multileveled. The findings of this review are promising; reinforcement of interventions is to be multi-component and implemented in a multilevel fashion to be sustainable and upscaled from project to operation. To secure sustainability of programs, a stable culture must be created, i.e., implemented into business plans, strategies and into the companies’ operating budget.

The employer – employee relationship should be further investigated, acknowledging the development of changed working conditions, requiring new leadership competencies and inherently programs targeting supervisors and leaders.

5 Conclusion

Our study largely revealed a lack of theoretically informed use of the concept presenteeism, resulting in few clear definitions, and many vague descriptions of the concept. This may affect the validity and generalisability of results. A high program diversity were identified, several adaptable to future needs (e.g., digital solutions), however hampered by narrow focus on individual health. Future research should focus on theory-based high quality multilevel WHPP comprising contextual, work, and organisational factors. Increased telework opportunities due to COVID-19 experiences have affected working conditions worldwide. Hence, adaptions are required among employee and management, in addition to an organisational and socio-cultural level.

Footnotes

Ethical considerations

This study, as a literature review, is exempt from Institutional Review Board approval.

Informed consent

Not applicable.

Reporting guidelines

In this study, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) checklist was used.

Acknowledgments

We would like to thank Atle Hårklau, a professional librarian at the University of South-Eastern Norway for her guidance and contribution during the search process in the different databases of academic literature.

Conflict of interest

The authors declare that they have no conflict of interest.

Funding

The authors report no funding.

The supplementary materials are available online from: .

References

Inge

, Perera

, Orchard

, Golding

. Exercise as Medicine – Evidence for Prescribing Exercise for the National Health Priority Areas: An Umbrella Review. J Postgrad Med Edu Res. 2020;54(4):178–205.

Aronsson

, Gustafsson

. Sickness presenteeism: Prevalence, attendance-pressure factors, and an outline of a model for research. J Occup Environ Med. 2005;47(9):958–66.

Johns

. Presenteeism in the workplace: A review and research agenda. J Organ Behav. 2010;31(4):519–42.

Lohaus

, Habermann

. Presenteeism: A review and research directions. Hum Resour Manag Rev. 2019;29(1):43–58.

Kinman

. Sickness presenteeism at work: prevalence, costs and management. Br Med Bull. 2019;129(1):69–78.

Shoss

, Min

, Horan

, Schlotzhauer

, Nigam

, Swanson

. The Impact of Precarious Work on Going to Work Sick and Sending Children to School Sick During the COVID-19 Pandemic. 3 MDPI 2021; ECERPH-3. Available from: https://doi.org/10.3390/ECERPH-3-09091

Kinman

, Grant

. Presenteeism during the COVID-19 pandemic: risks and solutions. Occup Med. 2020;kqaa193.

Eisen

. Employee presenteeism and occupational acquisition of COVID-19. Med J Aust. 2020;213(3):140–40.

Ruhle

, Breitsohl

, Aboagye

, Baba

, Biron

, Correia Leal

et al “To work, or not to work, that is the question” – Recent trends and avenues for research on presenteeism. Eur J Work Org Psych. 2020;29(3):344–63.

10.

Skagen

, Collins

. The consequences of sickness presenteeism on health and wellbeing over time: a systematic review. Soc Sci Med. 2016;161:169–77.

11.

Gustafsson

, Marklund

, Leineweber

, Bergström

, Aboagye

, Helgesson

. Presenteeism, psychosocial working conditions and work ability among care workers – a cross-sectional Swedish population-based study. Int J Environ Health Res. 2020;17(7):2419.

12.

Rostad

, Milch

, Saksvik

PØ

. Psychosocial workplace factors associated with sickness presenteeism, sickness absenteeism, and long-term health in a Norwegian industrial company. Scand Psychol. 2015;2:e11.

13.

Thun

. The Dynamics of Sickness Presenteeism through the Lens of the Job Demands-Resources Theory. Doctoral thesis. Norway: Norwegian University of Science and Technology (NTNU) 2017;160.

14.

Dietz

, Zacher

, Scheel

, Otto

, Rigotti

. Leaders as role models: Effects of leader presenteeism on employee presenteeism and sick leave. Work & Stress. 2020;34(3):300–22.

15.

Strömberg

, Aboagye

, Hagberg

, Bergström

, Lohela-Karlsson

. Estimating the effect and economic impact of absenteeism, presenteeism, and work environment– related problems on reductions in productivity from a managerial perspective. Value Health. 2017;20(8):1058–64.

16.

Kigozi

, Jowett

, Lewis

, Barton

, Coast

. The estimation and inclusion of presenteeism costs in applied economic evaluation: a systematic review. Value Health. 2017;20(3):496–506.

17.

Arnold

, De Pinto

. How are work-related characteristics linked to sickness absence and presenteeism? Theory and Data, IAAEU Discussion Paper Series in Economics, No. 11/2015, University of Trier, Institute for Labour Law and Industrial Relations in the European Union (IAAEU). 2015. Available from: http://hdl.handle.net/10419/121352

18.

Miraglia

, Johns

. Going to work ill: A meta-analysis of the correlates of presenteeism and a dual-path model. J Occup Health Psychol. 2016;21(3):261–83.

19.

Lohaus

, Habermann

, El Kertoubi

, Röser

. Working while ill is not always bad – positive effects of presenteeism. Front Psychol. 2021;11:620918.

20.

Karanika-Murray

, Biron

. The health-performance framework of presenteeism: Towards understanding an adaptive behavior. Hum Relat. 2020;73(2):242–61.

21.

Ferreira

, Mach

, Martinez

, Miraglia

. Sickness Presenteeism in the Aftermath of COVID-19 Is Presenteeism Remote-Work Behavior the New (Ab)normal? Front Psychol. 2022;12:748053.

22.

Steidelmüller

, Meyer

, Müller

. Home-Based Telework and Presenteeism Across Europe. J Occup Environ Med. 2020;62(12):998–1005.

23.

Ruhle

, Schmoll

. COVID-19, Telecommuting, and (Virtual) Sickness Presenteeism: Working From Home While Ill During a Pandemic. Presenteeism in the Aftermath of COVID-19: New Trends and Contributions Regarding Sickness Presence at Work. Front Psychol. 2022;12:734106.

24.

Gerich

. Home-Based Telework and Presenteeism: New Lessons Learned from the Covid-19 Pandemic. J Occup Environ Med. 2022;64(3):243–9.

25.

Chen

, Lu

, Cooper

. The compensatory protective effects of social support at work in presenteeism during the coronavirus disease pandemic. Front Psychol. 2021;12:643437.

26.

McGregor

, Caputi

. Presenteeism Behaviour. Current Research, Theory and Future Directions. Palgrave Macmillan; 2022.

27.

Pieper

, Schröer

, Eilerts

. Evidence of Workplace Interventions – A Systematic Review of Systematic Reviews. Int J Environ Res Public Health. 2019;16(19):3553.

28.

Ammendolia

, Côté

, Cancelliere

et al. Healthy and productive workers: using intervention mapping to design a workplace health promotion and wellness program to improve presenteeism. BMC Public Health. 2016;16:1190.

29.

Grønningsæter

, Kiland

. Theory-based implementation of physical activity programs in worksite health promotion settings: a systematic review. Health Promot Int. 2022;1–13.

30.

Kemm

. Health promotion. Ideology, discipline and specialism. Oxford, United Kingdom: Oxford University Press; 2015;41–51.

31.

Snyder

. Literature review as a research methodology: An overview and guidelines. J Bus Res. 2019;104:333–9.

32.

Page

, McKenzie

, Bossuyt

, Boutron

, Hoffmann

, Mulrow

, et al. The PRISMA statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.

33.

Walker

, Tullar

, Diamond

, Kohl

, Amick

. Validity and Reliability of the 8-Item Work Limitations Questionnaire. J Occup Rehab. 2017;27(4):576–83.

34.

Savitz

, Wellenius

, Trikalinos

. The problem with mechanistic risk of bias assessments in evidence synthesis of observational studies and a practical alternative: assessing the impact of specific sources of potential bias. Am J Epidemiol. 2019;188(9):1581–5.

35.

Cancelliere

, Cassidy

, Ammendolia

, Côte

. Are workplace health promotion programs effective at improving presenteeism in workers? A systematic review and best evidence synthesis of the literature. BMC Public Health. 2011;11(395):2–11.

36.

Jensen

. Can worksite nutritional interventions improve productivity and firm profitability? A literature review. Perspect Public Health. 2011;131(4):184–92.

37.

Mills

, Kessler

, Cooper

, Sullivan

. Impact of a health promotion program on employee health risks and work productivity. Am J Health Promot. 2007;22(1):45–53.

38.

Brown

, Gilson

, Burton

, Brown

. Does Physical Activity Impact on Presenteeism and Other Indicators of Workplace Well-Being? Sports Med. 2011;41(3):249–62.

39.

Czabala

, Charzynska

, Mroziak

. Psychosocial interventions in workplace mental health promotion: an overview. Health Promot Int. 2011;26(Suppl 1):i70–84.

40.

Cochrane

, Higgins

, FitzGerald

, Gallagher

, Ashton

, Corcoran

, et al. Early interventions to promote work participation in people with regional musculoskeletal pain: a systematic review and meta-analysis. Clin Rehabil. 2017;31(11):1466–81.

41.

Davies

. Does a literature review prove that health promotion can improve employee health? Occupational Health Sutton. 2015;67(5):27–9.

42.

Howarth

, Quesada

, Silva

, Judycki

, Mills

. The impact of digital health interventions on health-related outcomes in the workplace: A systematic review. Digit Health. 2018;4:1–18.

43.

Kröll

, Doebler

, Nüesch

. Meta-analytic evidence of the effectiveness of stress management at work. Eur J Work Organ Psychol. 2017;26(5):677–93.

44.

Neuhaus

, Eakin

, Straker

, Owen

, Dunstan

, Reid

, Healy

. Reducing occupational sedentary time: a systematic review and meta-analysis of evidence on activity-permissive workstations. Obes Rev. 2014;15(10):822–38.

45.

Oakman

, Keegel

, Kinsman

, Briggs

. Persistent musculoskeletal pain and productive employment; a systematic review of interventions. Occup Environ Med. 2016;73(3):206–14.

46.

Osilla

, van Busum

, Schnyer

, Larkin

, Eibner

, Mattke

. Systematic Review of the Impact of Worksite Wellness Programs. Am J Manag Care. 2012;18(2):e68–e81.

47.

Rongen

, Robroek

SJW

, van Lenthe

, Burdorf

. Workplace Health Promotion: a meta-analysis of effectiveness. Am J Prev Med. 2013;44(4):406–15.

48.

Schröer

, Haupt

, Pieper

. Evidence-based lifestyle interventions in the workplace- an overview. Occup Med. 2014;64(1):8–12.

49.

White

, Dionne

, Wärje

, Koehoorn

, Wagner

, Schultz

, et al. Physical activity and exercise interventions in the workplace impacting work Outcomes: A Stakeholder-Centered Best Evidence Synthesis of Systematic reviews. Occup Environ Med. 2016;7(2):61–74.

50.

Wagner

, Koehn

, White

, Harder

, Schultz

, Williams-Whitt

, et al. Mental Health Interventions in the Workplace and Work Outcomes: A Best-Evidence Synthesis of Systematic Reviews. Int J Occup Environ Med. 2016;7(1):1–14.

51.

Lui

JNM

, Andres

, Johnston

. Presenteeism exposures and outcomes amongst hospital doctors and nurses: a systematic review. BMC Health Serv Res. 2018;18:985.

52.

WHO regional Office for Europe. Ottawa Charter for Health Promotion. Copenhagen: WHO regional Office for Europe. WHO (1948/2006). Constitution of the World Health Organization. NewYork; 1986. Available from: https://www.euro.who.int/__data/assets/pdf_file/0004/129532/Ottawa_Charter.pdf

53.

Hutchinson

, Wilson

. Improving nutrition and physical activity in the workplace: a meta-analysis of intervention studies. Health Promot Int. 2012;27(2):238–49. doi: 10.1093/heapro/dar035.