A healthy and productive workforce is critical for economic success and population health. Illness at the workplace can result in lost productivity, which arises from two sources: absenteeism and presenteeism. Absenteeism refers to an employee's time away from work due to illness or disability 1 . Presenteeism refers to the decrease in productivity in employees whose health problems have not necessarily led to absenteeism and the decrease in productivity for the disabled workers before and after their absence period 2 . It is defined as being present at work, but limited in some aspects of job performance by a health problem, and it is often a hidden cost for employers 3 . It includes time not spent on job tasks and decreased quality of work (e.g. product waste and product defects) 4 . Absenteeism and presenteeism are part of a continuum within which workers likely transition back and forth over time 5 .

Health promotion in the workplace is defined as preventing, minimizing and eliminating health hazards, and maintaining and promoting work ability 6 . Worker health and wellness is maintaining a balance of the physical, mental and social ingredients, as well as health habits associated with good physical condition, energy and vitality 6 .

Presenteeism emerged as a new business issue in the 1990's 7 . It is becoming a significant challenge to maintain a healthy and productive workforce for developed countries due to an increasing number of people affected by chronic health conditions and an aging workforce that is more likely to be affected by these conditions. Subsequent rising health care costs and an increasing awareness of presenteeism losses are escalating the demand for health promotion programs for working populations 8 .

Research on interventions to improve presenteeism is still relatively new compared with other workplace issues such as healthcare costs and absenteeism 9 10 11 . Most of the literature on presenteeism has investigated its measurement 3 .

Currently, there is no universal agreement on the most appropriate method for measuring or monetizing presenteeism 1 12 . It is typically measured as the costs associated with reduced work output, errors on the job, or failure to meet company production standards 13 . It is difficult to value presenteeism economically because past studies use different measures of presenteeism, include different populations of workers and use various methods to assign dollar values to their losses. Nevertheless, it appears that economic costs are considerable 1 14 15 .

Several self-report instruments have been developed to measure presenteeism across various types of jobs and organizations 1 16 17 18 . These are useful especially when it is difficult to obtain objective data regarding particular characteristics of a workplace or profession (e.g. the number of parts manufactured). Evidence of their psychometric properties has been reported in varying degrees 3 5 14 17 18 19 20 21 22 23 . Some common tools with good psychometric properties include the Work Limitations Questionnaire (WLQ) 20 24 , Work Productivity and Activity Impairment (WPAI) 25 26 , and the Stanford Presenteeism Scale (SPS) 20 27 . These tools assess presenteeism with the assessment of perceived impairment approach, where employees are asked how much their illnesses hinder them in performing common mental, physical, and interpersonal tasks and in meeting job demands.

WHP programs vary considerably in size and composition, and they have evolved significantly over the past 30 years 28 29 30 . However, whether or not programs can improve workplace productivity has yet to be determined. While containing health care-related costs and absenteeism have been important strategies for companies, greater gains may be realized by improving on-the-job productivity and investing in preventive and early intervention services 29 31 32 33 34 35 36 .

The primary objective of our study is to review and scientifically appraise the literature on WHP programs to see if they are effective in improving presenteeism among employees. The secondary objectives are to identify components of successful WHP programs and to identify risk factors for presenteeism.

The scientific literature published between 1990 and January, 2010 was systematically searched. Primary sources were the electronic databases of The Cochrane Library, Medline, Embase, CINAHL Plus, NLM Gateway, PsychINFO, Evidence in Health and Social Care, AMED and the Trip Database (Additional file 1). The search strategy was reviewed by a reference librarian. Additionally, the reference lists of all relevant studies were examined, and three of the most relevant journals were hand-searched between January, 2005 and February, 2010 (i.e. Journal of Occupational and Environmental Medicine, Ergonomics, and Journal of Industrial Medicine). A Google search for unpublished literature was conducted in January, 2010 using the search terms presenteeism, health promotion, and workplace. Government and other relevant websites were scanned. Experts and organizations involved with WHP and presenteeism were contacted (Additional file 2).

All citations identified through the search strategy were screened. These included English-language reports, published reports of original research (randomized controlled and controlled trials, cohort, pre-post, and ecological studies), systematic reviews, meta-analyses, conference proceedings, government reports, guidelines, and unpublished gray literature manuscripts. To be included, studies had to be original research that contained data on at least 20 human participants; focused on adults 18 years of age or older; and examined WHP programs including all types of measures aimed at promoting health and wellness, or reducing the risk of ill-health. These could be targeted at behavioural, physiological, organizational or environmental changes. The intervention had to be implemented at the workplace but activities were allowed to occur elsewhere. For example, risk factor screening and education could occur at the workplace, but workers were allowed to use external exercise facilities, or seek additional medical attention. The intervention also had to be described in sufficient detail. Exclusion criteria included studies examining only military personnel; return-to-work studies; narrative, editorial, or clinical reviews; opinion papers, editorials, and letters to the editor; studies where interventions were not implemented at the worksite; studies where productivity outcomes were not measured or specified; studies which grouped productivity outcomes together, such as presenteeism and absenteeism, and the results could not be evaluated specifically for presenteeism changes; and studies that measured productivity only in terms of lost earnings. All relevant systematic reviews and meta-analyses were screened to ensure that primary studies were not missed.

Relevant primary studies were assessed for methodological quality using a tool developed and tested by the Effective Public Health Practice Project (EPHPP) - EPHPP Quality Assessment Tool for Quantitative Studies (Additional file 3) 37 38 . It consists of six criteria: selection bias, allocation bias, control of confounders, blinding of outcome assessors, data collection methods, and withdrawals and dropouts. Reviewers were asked to rate each criterion as 'weak', 'moderate', or 'strong'. A final global rating of each study was subsequently determined. The tool has demonstrated reliability, content and construct validity, and the ability to adapt current methods for systematic literature reviews of effectiveness to questions related to public health. The test-retest reliability of the EPHPP is good (i.e. Cohen's Kappa of 0.74). Content validity was established by having six experts review the questionnaire, and it was independently tested on 10 primary studies by four experts in critical appraisal and community health. Construct validity was shown through comparisons with another highly rated instrument, the Guide to Community Preventive Services 39 . The EPHPP is relatively easy to use and an accompanying dictionary clarifies any questions related to the components. Two reviewers independently performed in-depth reviews for each study. A consensus method was used to solve disagreements about study selection and methodological quality. A third reviewer was consulted if disagreement persisted.

Studies were considered scientifically admissible if they were rated as moderate or strong, or scientifically inadmissible if they were rated as weak, had fatal biases or other important methodological flaws. Table 1 summarizes the criteria for this rating classification 38 . Information was extracted from each admissible study on: (1) country and workplace, (2) study design, (3) characteristics of participants, (4) inclusion and exclusion criteria, (5) intervention(s) and control(s), (6) outcome measurements and follow up periods, and (7) key findings and limitations (Additional files 4 and 5). The heterogeneity of the populations, interventions and outcome measures made it too difficult to compare the studies. A best evidence synthesis was performed and is based only on the results of the strong and moderate studies 40 . Interventions were deemed successful if they improved the outcome of interest and their program components were subsequently identified. Possible risk factors contributing to presenteeism were identified through the literature review.

Interventions were deemed successful if they improved the outcome of interest (i.e. presenteeism). Overall, ten interventions demonstrated preliminary evidence of promising effects on presenteeism in their respective employee populations and work settings (Additional files 4, 5 and 6).

Strong evidence for this was found in two studies (Additional files 4 and 6). The first involved worksite exercise 41 , and the second investigated the impact of a supervisor education program regarding mental health promotion 42 .

The remaining eight studies provided moderate evidence of positive intervention effects (Additional files 5 and 6). These interventions consisted of "A Lifestyle Intervention Via Email" (Alive!) 43 , extra rest break time for workers engaged in highly repetitive work 44 , a multi-disciplinary occupational health program 45 , a multi-component health promotion program 46 , participatory processes 47 48 , exposure to blue-enriched light (vs. white light) 49 , and a telephone intervention program for depressed workers 50 .

Interventions were deemed unsuccessful if they did not improve the outcome of interest (i.e. presenteeism). Four interventions were unsuccessful at improving presenteeism in their specific employee populations and work settings (Additional files 4 and 5). Two of these were rated as strong 51 52 and two as moderate 53 54 . These consisted of the implementation of a computer mouse with a feedback signal to prevent hovering behaviour 54 , a multi-dimensional program for low back pain prevention 51 , specific resistance training and all-around physical exercise 53 , and worksite exercise/reduced work hours 52 .

Of the 10 successful interventions, three were delivered by a health professional (Additional files 4, 5 and 6): physiotherapists, occupational health nurses 41 , psychologists 42 , or occupational health physicians 45 . Two were delivered using a participatory approach, which consists of teamwork among employees, managers, human resources personnel, and researchers 47 48 . One was delivered over the telephone by mental health clinicians 50 , one was delivered by email/internet 43 , one was an organizational change (i.e. rest break schedule) 44 , and one was an environmental change (i.e. lighting) 49 . A final successful intervention utilized a mixed delivery method consisting of email, paper-based packs, and worksite seminars 46 . Of the four unsuccessful interventions (Additional files 4 and 5), one was delivered by a health professional (i.e. physiotherapist) 51 , one was delivered by an experienced fitness instructor 53 , one was an equipment change (i.e. computer mouse) 54 , and the final one consisted of an organizational change (i.e. work hours allocated to mandatory physical exercise) 52 .

Some workplaces screened workers prior to intervention. One of the most common screening methods is the health risk assessment (HRA). It includes the assessment of personal health habits and risk factors, estimation of future risk of adverse health outcomes, and feedback in the form of education and counselling to alter risk factors 55 . It is valuable for identifying high-cost claimants; targeting individual program participation and measuring and tracking individual behaviours that effect health care costs, absenteeism and presenteeism 7 55 . Seven of the 10 successful interventions used screening methods (Additional file 6), consisting primarily of HRAs, or other questionnaires/assessments 41 43 45 46 47 48 50 . Two of these seven interventions utilized a participatory approach, whereby screening is inherent in the process (i.e. teams first identify problems and needs, then develop solutions) 47 48 . On the other hand, screening methods were used in only one of the four unsuccessful interventions 52 .

Seven of the 10 successful interventions tailored their programs to address participant needs (Additional file 6) 41 43 45 46 47 48 50 , whereas tailoring occurred in only half of the unsuccessful interventions 51 53 . Finally, only one of the 10 successful interventions used an incentive (i.e. lottery tickets) (Additional file 6) 46 . Incentives were not used in any of the unsuccessful interventions.

Four of the 14 admissible studies were rated as strong (Additional file 4). However, there still existed some bias with respect to thorough reporting of withdrawals and drop-outs 51 , data collection methods 42 , and participant selection 52 . All four studies also showed a risk of bias with respect to blinding of outcome assessors and/or participants, as well as intervention integrity 41 42 51 52 .

Three of these four studies were cluster RCTs and the other was a RCT. Two of the four interventions improved presenteeism (Additional files 4 and 6) and involved worksite exercises 41 and a supervisor mental health education program 42 . Both were delivered by health professionals. Workers were screened and the intervention was individually tailored in only one of them 41 . Employee incentives were not used in either intervention.

In two of the four strong studies, the interventions did not improve presenteeism (Additional file 4). These included a low back pain prevention program 51 , and a worksite exercise/reduced work hours program 52 . Only one of these was individually tailored and delivered by a health professional 51 . Neither intervention involved worker screening, or the use of incentives.

Of the 14 admissible studies, 10 were rated as moderate evidence and demonstrated similar biases although to a greater degree than the four higher quality studies (Additional file 5). These consist of four RCTs, two cluster RCTs, one crossover design, one interrupted time series, one pre-post study, and one quasi-experimental study. Eight of these 10 interventions improved presenteeism (Additional files 5 and 6). These include Alive (A Lifestyle Intervention Via Email) 43 , extra rest break time 44 , two participatory interventions 47 48 , lighting changes 49 , a multi-component health promotion program 46 , a telephone support program 50 , and an occupational health program 45 . Delivery methods varied and were executed by: health professionals 45 50 , participatory methods 47 48 , the internet 43 , mixed methods 46 , an organizational change 44 , and an environmental change 49 . Six of these eight successful interventions involved employee screening methods 43 45 46 47 48 50 . These six, in addition to the one by Blangsted and colleagues 53 , were individually tailored. Only one of the eight interventions used an employee incentive 46 .

Two of the 10 moderate studies were not successful at improving presenteeism (Additional file 5) and consisted of exercise 53 , and a computer mouse with a feedback signal to prevent hovering behaviour 54 . Neither of these were delivered by health professionals. One was delivered by a fitness intructor 53 , and the other involved an equipment change 54 . Only one intervention was individually tailored 53 . Workers were not screened and incentives were not used in either intervention.

A secondary objective was to identify components of WHP programs successful at improving presenteeism. We found preliminary evidence to support the use of one or more of the following: involving employees' supervisors/managers in WHP programs 42 45 47 48 , targeting organizational and/or environmental factors to influence behaviour 41 42 44 45 47 48 49 , screening workers prior to intervention using HRAs or other methods 43 45 46 47 48 50 , improving supervisor/manager knowledge regarding mental health in the workplace 42 , allowing physical exercise to occur during working hours 41 , and individually tailoring programs 41 43 45 46 47 48 50 . Grounding interventions in behaviour change models to help reinforce desirable lifestyle behaviours 43 , using participatory approaches with high employee involvement to develop interventions 47 48 , and increasing the frequency and duration of rest breaks for workers required to stand for prolonged periods 44 were also helpful. Interventions can be delivered by various modes including telephone, email/internet, seminars, paper-based literature 43 46 ; participatory teams 47 48 ; organizational changes 44 ; environmental changes 49 ; and by various health professionals 41 42 45 50 .

Another secondary objective was to understand risk factors affecting presenteeism. Several factors were reported in the literature such as being overweight, having a poor diet, smoking, a lack of physical exercise, high stress, poor relations with co-workers and management, and poor physical work environments 1 2 3 . Health conditions affecting worker presenteeism were also reported and included arthritis, allergies, migraine, chronic pain, diabetes, hypertension, gastro-intestinal conditions, gastro-esophageal reflux disease, musculoskeletal disorders, respiratory disorders, mental health problems such as depression and anxiety, cancer, cardiovascular disease, and metabolic syndrome 1 2 3 12 13 15 . In general, these risk factors and health conditions can decrease worker productivity by causing pain and fatigue, and by reducing physical and mental capacities.

This review has several strengths. The best available evidence was systematically gathered, appraised, and synthesized. Limiting the findings to studies of better methodological quality is a notable strength that builds on past results 9 10 . All steps of our comprehensive review were explicitly reported, thereby making the process more transparent and reproducible. Our findings are reported in compliance with PRISMA guidelines (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) 58 .

Our study has limitations. While the search was extensive, it is limited to the literature available as of January, 2010 and to English publications. Only 14 of 47 articles reviewed were scientifically admissible, which limits the number of higher quality studies to draw inferences from. The other studies were inadmissible because of bias. Included studies also had risk of bias, but to a lesser extent. Common problems encountered included selection bias, or the selection criteria to get into the study were not described 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 . Many of the inadmissible studies had poor response and participation rates (i.e. less than 60%) 59 60 61 67 69 72 77 82 . Other studies did not consider, or report on, common confounding factors such as gender, age, education, and health status 59 60 62 63 64 66 67 68 69 70 71 74 75 76 77 78 79 82 83 84 85 86 87 88 89 90 . Insufficient blinding of the assessors and/or the participants was observed in several studies 59 60 61 63 64 69 70 72 75 79 80 81 83 84 85 88 89 90 91 . Information bias is likely in a number of the studies as some of the data collection tools have not been shown to be reliable and valid 59 61 63 64 65 72 78 79 80 81 82 85 87 88 89 90 . Finally, various studies did not adequately report withdrawal and/or drop-out rates, or had unacceptably high rates (i.e. follow-up rate less than 60%) with little to no examination of differences between study participants and nonparticipants 59 60 61 62 64 65 68 69 70 71 73 74 75 82 83 88 89 .

Presenteeism is an evolving field and is difficult to measure. The uncertainties and inconsistencies surrounding its measurement may be one of the biggest limitations regarding the usefulness of this review. All of the current instruments have drawbacks and several significant measurement issues have been raised 1 3 13 14 18 92 ; Edington DW, personal communication, March 2010]. Only a few of the instruments have been validated against objective measures of productivity, such as the number of calls made in a call centre, or the number of parts assembled in a manufacturing plant. This is easiest to do when dealing with piecework, but is much more difficult when concrete, measureable output is not available, as with jobs in the knowledge sector, or when dealing with team work where the presenteeism of a single team member may affect the job performance of the group. Additionally, each instrument measures a different quality of presenteeism (e.g. quality or quantity of work), making results difficult to compare. Finally, indirect costs faced by employers having to respond to reduced productivity need to be taken into account, including hiring new or temporary staff, training staff, paying overtime to other employees, and dealing with lost revenue, to name a few.