* Outline the authors' ASSET model, which seeks to identify an evidence-based set of workplace factors affecting health and productivity outcomes.
* Discuss the new findings on workplace factors affecting productivity, including important sources of pressure to attend work while sick.
* Summarize the study implications for identifying and managing psychosocial workplace factors predictive of presenteeism.
Presenteeism has become a more visible construct in the occupational health and well-being fields,1 with considerable research effort devoted to its measurement2–4 and consequences.5–7 The core notion of presenteeism involves being at work but having impaired functioning due to mental or physical symptoms.1,8 Some authors have extended the concept to include factors such as working elevated hours and “putting in face time”9 or reduced productivity due to events that distract one from full productivity.10 This article takes a narrow view of presenteeism, in line with the majority of the research literature, and defines it as sickness presence or, more simply, attending work while ill.1,11 Using a sample of 6309 employees, we provide data on the prevalence of presenteeism across three industrial sectors, test a model of workplace causes of presenteeism (including health and productivity factors), and also examine perceived pressure on individuals to attend work while ill.
PREVALENCE AND COST OF PRESENTEEISM
Studies indicate that presenteeism is a common and costly workplace issue. A study of Swedish employees (n = 3096) found that 70% reported they had gone to work at least once while ill in the preceding year.11 Similarly, data collected from a random sample of Danish employees (n = 11,270) found that 73% attended work while ill in the preceding year, with 19% indicating at least four episodes of presenteeism.12 In addition to such studies, a body of research has focused on the work limitations (eg, impaired concentration) created by presenteeism4 and their cost to employers.13–15 For example, using a combination of data sources to assess the cost burden of health conditions (eg, arthritis, hypertension), Goetzel et al concluded that “anywhere from approximately one fifth to approximately three fifths of the total dollars attributed to common health conditions faced by [US] employers could be the result of on-the-job productivity losses.”16(p408) These researchers also note that their analysis “points to a potentially large category of expenses as-yet unaccounted for by many employers who are concerned about their health-related liabilities.”(p408)
We begin with analysis of the frequency of presenteeism across seven organizations (in the United Kingdom and Europe) in three industrial sectors: health, higher education, and government. The purpose of this analysis is to broaden appreciation of the prevalence of presenteeism, and in particular to provide data that may help facilitate future international and sector comparisons.
PSYCHOSOCIAL WORKPLACE FACTORS AND PRESENTEEISM
Research into work-related stress has already established many of the key factors that are associated with psychological well-being, health, and productivity at work.17,18 Evidence is beginning to accumulate that these same workplace factors, including job control,19 relationships with colleagues,20 and job insecurity,21 may also be associated with presenteeism. The ASSET (A Shortened Stress Evaluation Tool) model22–24 provides an evidence-based set of workplace factors that are related to health and productivity outcomes. Using the ASSET model, this study extends existing research by examining a broader range of potential causes or predictors of presenteeism. More specifically, we propose and test a model (see Fig. 1) in which workplace factors positively predict ill health,25,26 which, in turn, predicts higher presenteeism.27 In line with extant research,16 the model also contends that presenteeism negatively predicts productivity. Importantly, the statistical approach adopted (see later) permits examination of which workplace factors contribute most in determining ill health, presenteeism, and ultimately productivity, and also whether relationships between these factors vary by industrial sector.
Pressure to Attend Work While Ill
Psychosocial workplace factors and health may be important precursors of presenteeism, but other factors may also be relevant to the decision to work while ill. For example, in one study, the introduction of focused attendance management procedures was associated with an increase in presenteeism,28 suggesting that, when pressure on workplace attendance is increased, sickness presence may be substituted for sickness absence. Caverley et al investigated this “substitution” hypothesis and concluded that “The high rate of presenteeism relative to absenteeism observed in our sample is consistent with the substitution hypothesis.... If employees start coming to work when sick instead of staying away, then, for constant levels of health, there will be lower levels of absenteeism.”21(p317) Caverley et al also found that presenteeism increased during the organizational downsizing process that they examined. Other researchers have found similar results, suggesting that periods of downsizing or other changes that give rise to job insecurity may lead to increased presenteeism.29
In broad terms, the pressure to attend work when sick may come from at least three different sources: management, coworkers, and self. The authors could find no previous research that has focused explicitly on the extent to which these different sources of pressure may influence episodes of presenteeism. Through examination across industrial sectors of the perceived influence of these sources of pressure on those participants who reported attending work while ill, this study, therefore, further contributes to the understanding of presenteeism.
Study Context and Participants
All the data used in the following analyses were collected as part of well-being surveys conducted in the collaborating organizations. The data were collected within a 1-year period between November 2009 and November 2010 for 6309 individuals across seven different organizations in the United Kingdom and Europe. These organizations include four organizations from the health sector, two universities, and a European governmental department. Most study participants were invited to respond to the survey through e-mail and the vast majority of questionnaires were completed on-line (92%), with the remainder completed using a paper version of the questionnaire. The response rates for the seven organizations ranged from 27.6% to 64.6%. The average response rate, based on the total number of invites sent out, was 45.0%. Background information for the 6309 participants in this study is provided in Table 1.
A number of steps were taken to control for nonresponse bias and to maximize participation rates. For example, in each case the invitation e-mails were accompanied with a note from the organizations themselves to explain exactly why the surveys were being carried out; participants were actively encouraged to complete the questionnaire during work hours; at least two reminders were sent out during each data collection period to urge participants to respond; ongoing administrative and technical assistance was available to ensure that everyone who tried to complete the survey was supported to do so; and participants were continually assured that their responses would be confidential and that data would not be made available within the organization except in aggregate format. In addition to these steps during the data collection periods, the majority of surveys were followed up with randomly sampled focus groups. The response from focus group members to the findings presented supported the overall validity of the survey results.
Workplace factors and health-related outcomes were measured with the ASSET questionnaire.22,23 The questionnaire provides psychometrically reliable and valid multiple-item scales that measure job incumbents' experiences of the following specific workplace factors: Resources & Communication (sample item: “I am troubled that I do not have the proper equipment and resources to do my job”); Control & Autonomy (sample item: “I am troubled that I am not involved in decisions affecting my job”); Work Relationships (sample item: “I am troubled that my relationships with colleagues are poor”); Work Life Balance (sample item: “I am troubled that my work interferes with my home and personal life”); Workload (sample item: “I am troubled that I have unmanageable workloads”); Job Security & Change (sample item: “I am troubled that my job skills may become redundant in the near future”); Pay, Benefits & Job Conditions (sample item: “I am troubled that my pay and benefits are not as good as other people doing the same or similar work”). Response options for all of the above ranged from Strongly Disagree (1) to Strongly Agree (6). Johnson24 provides a full description of the ASSET model and the interpretation and relevance of the scales. ASSET scales to measure self-reported health include items related to physical health (sample item: “Over the last 3 months, have you experienced any of the following symptoms or changes in behaviour?—Indigestion or heartburn”) and items related to psychological health (sample item: “Over the last 3 months, have you experienced any of the following symptoms or changes in behaviour?—Difficulty in making decisions”).
In addition to the core workplace and health factors measured in ASSET, the study questionnaire also asked participants to report on their productivity with the following item: “Over the last three months, roughly how productive have you felt in your job?” Response options ranged from 0% to 9% in 11 bands to 100%.
Presenteeism was measured with the following item: “In the past 3 months, have you ever not felt well enough to perform your duties to your normal standard, but attended work regardless?” The 3-month time frame was chosen for its consistency with the time frame for other items in the ASSET scales. Response options were “Yes” or “No.” Respondents who reported presenteeism were also asked to report on the source of pressure that they believed contributed to their decision to attend work while they were ill: Self (“I put myself under pressure to attend work, regardless of my illness”), Manager (“I felt pressurised by my manager to attend work, regardless of my illness”), or Colleagues (“I felt pressurized by my colleagues to attend work, regardless of my illness”). The items described previously were again measured using response options ranging from Strongly Disagree (1) to Strongly Agree (6).
In addition, the questionnaire covered the demographic variables gender, age, disability, marital status, disability, employment basis, length of service, and the organizational variable type of organization.
To examine the relationships of presenteeism with workplace factors, health, and productivity (see Fig. 1), we applied structural equation modeling using AMOS 18.0.30 In the first stage of the analysis, we defined and tested a model with all workplace factors and the demographical control variables as exogenous variables, and with health, presenteeism, and productivity as endogenous variables. The aim of this exploratory analysis was to determine which of the seven workplace factors (resources & communication; control & autonomy; work relationships; work life balance; workload; job security & change; and pay, benefits & job conditions) uniquely contribute to the prediction of health.
Second, we tested the four-step sequence model that starts with the workplace factors (as identified in stage 1) and leads to productivity via health and presenteeism (see Fig. 2). To improve model fit, we modeled covariances between the exogenous variables where the respective exogenous variables could be theoretically expected to covary and where the corresponding modification index was above 10.0,31,32 indicating that including this covariance substantially improves model fit.
The acceptability of the overall fit of the model was assessed by three indices: the Comparative Fit Index (CFI) and the Tucker-Lewis Index (TLI) as two incremental fit indices and the root-mean-square error of approximation (RMSEA) as an absolute measure. The CFI is a noncentrality parameter-based index used to overcome the limitation of sample size effects.33 Comparative Fit Index values close to 0.9 indicate good model fit.34 The TLI compares a proposed model's fit to a nested baseline or null model. Again, a value close to 0.9 indicates good model fit.34 The RMSEA index constitutes a parsimony measure that depicts the discrepancy between the observed and estimated covariance matrices per degree of freedom. RMSEA values less than 0.08 are acceptable.34
The results indicate that nearly 60% of the sample reported presenteeism during the 3-month period prior to survey administration. In more detail (see Table 2), the results show that (1) presenteeism is lower for males than for females (χ2(1) = 38.92, P < 0.001); (2) there is a statistically significant association between marital status and presenteeism (χ2(5) = 20.45, P < 0.001), with those who are divorced, separated, or widowed being more likely to come to work while ill than single respondents and those who are married or in a civil partnership; (3) presenteeism relates to length of service (χ2(5) = 39.79, P < 0.001), increasing over time; (4) there is a significant association between disability and presenteeism (χ2(1) = 25.92, P < 0.001), that is, disabled people are more likely to come to work whilst ill; and (5) presenteeism relates to type of organization (χ2(2) = 77.68, P < 0.001), being highest for people working for the government and lowest for people working at the universities. Reported levels of presenteeism are unrelated to age or employment status (full- or part-time).
Relationship of Presenteeism With Workplace Factors and Outcome Variables
Figure 2 presents the structural model of presenteeism and the standardized pathway estimates. The goodness-of-fit measures indicate an acceptable fit for the model (CFI = 0.90; TLI = 0.88; and RMSEA = 0.05). The paths for the relationships between workplace factors (unique predictors, see the previous text), health, presenteeism, and productivity are significant at P < 0.01 after controlling for demographics. Greater levels of work life imbalance and workload predict poorer health (β = 0.18, P < 0.001, and β = 0.08, P < 0.01, respectively). Furthermore, less favorable job conditions predict poorer health (β = 0.38, P < 0.001). Poorer health, in turn, predicts presenteeism (β = 0.42, P < 0.001), which predicts lower productivity (β = −0.17, P < 0.001).
We cross-validated the model by testing it across the three industrial sectors: health, higher education, and government. The sample was split into three subsamples according to sector. No differences in model fit were observed, indicating that results are consistent across the different sectors.
Pressure to Attend Work While Ill
Having explored the key sources of ill health and presenteeism, we now examine the pressures that motivate individuals to attend work while ill. Table 3 shows the reported sources of pressure to attend work while unwell for the presenteeism group. As Table 3 shows, the overwhelming majority of people (67.3%) agreed or strongly agreed that pressure from themselves provided the main stimulus to attend work while ill. Pressure from managers also seemed relevant in nearly 20% of cases (19.1% agreed to some extent that they felt pressure from their manager to attend work regardless of their illness).
DISCUSSION AND CONCLUSIONS
The results indicate that almost 60% of participants reported a presenteeism incident in the 3 months before completing the survey. This finding indicates that the incidence of presenteeism episodes for health, higher education, and government employees in this study are broadly consistent with those observed in other studies11,12 and underline the gravity of the issue.
The analysis of workplace factors found that work life imbalance (eg, work interfering with home life), higher workload (eg, unmanageable work demands), and poorer work conditions (eg, unpleasant physical working conditions) are particularly important in predicting ill health, higher presenteeism, and lower productivity, in turn (see Fig. 2). Inherent in the model is the argument that ill health, presenteeism, and productivity are, to some degree, consequences of work factors over which managers have influence. Managers and supervisors have a key role in helping employees balance work demands and reduce uncertainty, conflict, and ambiguous situations, thereby improving well-being at work.35 There is substantial evidence showing that management and leadership impact employee psychological health36 and some research has already revealed links between leadership and presenteeism rates.37 The present study suggests that a supportive and encouraging leadership style is likely to reduce presenteeism and therefore enhance productivity.
This study also examined sources of pressure to attend work while ill. The majority of participants indicated themselves as the principal source of pressure. On the basis of the model (eg, work interfering with home life), it is possible that our sample of employees experienced high job demands, with responsibility for task completion, which engendered a sense of obligation to attend work while ill in order to maintain performance.38 A significant proportion of the sample also indicated that their manager was a source of pressure. This might be indicative of pressure that managers themselves experience39 or a lack of consideration and support for employees, or both.35 Manager/supervisor training aimed at helping employees to better appreciate the work situation40 might be of value in reducing the perception of pressure to attend work while ill.
Our analysis treats presenteeism as a voluntary behavior; that is, despite workplace factors and pressures, employees could decide not to attend work while ill. In much presenteeism research, there is also an underlying assumption that episodes of presenteeism are undesirable. Recent policy-driven reports and initiatives in the United Kingdom,41 however, have led to the development of a new version of the “sick note” for family doctors to use. The old version required patients to be certified as fit or unfit for work. The new version, referred to colloquially as the “fit note,” allows the physician to indicate what duties the patient may be able to undertake. One intended consequence of the new certificate is that people may return to work more quickly, but before they are 100% well again, in the expectation that they will carry out some of their duties—and therefore they will almost certainly not be 100% productive. In effect, this approach involves the active encouragement of presenteeism. An important consideration for future presenteeism definitions and measurement processes, therefore, is to develop a better understanding of why the respondent is at work, rather than absent. The short- and long-term consequences of feeling forced to work when unwell may be very different from working when unwell out of choice.
Limitations and Future Research
In this study, participants were asked to estimate their average level of productivity over the last 3 months. This study treats presenteeism as a dichotomous variable (yes/no) that is insensitive to, for example, frequency. Therefore, our results should be interpreted in the light of the study's coarseness. That is, this study gives only a very sharp picture of the prevalence, predictors, and outcomes of presenteeism. Nevertheless, it should be noted that while other studies11,42 used more sensitive measures of presenteeism, they dichotomized the scale for the purpose of analysis. The time frame used in this study is also an important issue. Obviously, the longer the time frame, the more individuals would have experienced presenteeism.43 Yet, the potential impact of recall bias might increase as the time frame gets longer.43 Given previous studies,11 we expected that sufficient individuals would report presenteeism with the 3-month time frame chosen for this study to permit analysis. Shortening the time frame to 2 to 4 weeks might have increased the accuracy of the assessment of presenteeism4; however, a time frame as short as this might have constrained examination of the consequences of presenteeism43 (ie, productivity loss). Accordingly, and given the fact that the participants were asked to assess whether they attended work while ill or not rather than remembering the exact amount of time they spent at work while ill, we expected that a 3-month recall period would be appropriate for the purpose of this study. In addition, the 3-month time frame was consistent with the time frame for other items in the ASSET scales.
The presenteeism group reported lower productivity (80.79%) compared with the no-presenteeism group (86.06%). The data collected as part of this research, however, did not allow for any consideration of the nature of the health conditions reported (eg, severity, chronicity). As such, we were unable to determine variation in presenteeism and productivity loss by medical condition. Furthermore, the study established the proposed relationships between the focal variables using data collected at a single moment in time. It would be worthwhile, therefore, to conduct studies over time (or longitudinal research designs) to obtain more robust conclusions about causality. For instance, on a daily basis over 3 months, employees could be asked to rate their health (indicate any condition and its severity) and supervisors asked to monitor performance. Such research designs would help overcome another shortcoming of this study: use of self-report data. For example, although self-reported productivity has been shown to be a valid measure of employee productivity in previous work,27,44 more objective data to validate the measures of productivity loss would have been desirable.
Research indicates that presenteeism is a major source of lost productivity. Our findings indicate that workplace factors represent important antecedents of ill health, presenteeism, and productivity. In this regard, we contend that a coaching, supportive, and encouraging leadership style may help employees better manage their work demands, thereby improving health and reducing the likelihood of undesirable presenteeism. Given the cost of presenteeism, it is vital that manger/supervisors are aware of their role in promoting employee health.
1. Johns G. Presenteeism in the workplace: a review and research agenda. J Organ Behav. 2009;31:519–542.
2. Sanderson K, Tilse E, Nicholson J, et al. Which presenteeism measures are more sensitive to depression and anxiety? J Affect Disord. 2007;101:65–74.
3. Turpin RS, Ozminkowski RJ, Sharda CE, et al. Reliability and validity of the Stanford Presenteeism Scale. J Occup Environ Med. 2004;46:1123–1133.
4. Lerner D, Amick BC, Rogers WH, Malspeis S, Bungay K, Cynn D. The Work Limitations Questionnaire. Med Care. 2001;39:72–85.
5. Burton WN, Chen CY, Conti DJ, et al. The association between health risk change and presenteeism change. J Occup Environ Med. 2006;48:252–263.
6. Bockerman P, Laukkanen E. Predictors of sickness absence and presenteeism: does the pattern differ by a respondent's health? J Occup Environ Med. 2010;52:332–335.
7. Collins JJ, Baase CM, Sharda CE, et al. The assessment of chronic health conditions on work performance, absence, and total economic impact for employers. J Occup Environ Med. 2005;47:547–557.
8. Hemp P. Presenteeism: at work—but out of it. Harv Bus Rev. 2004;82:49–58.
9. Worral L, Cooper C, Campbell F. The new reality for UK managers: perpetual change and employment instability. Work Employ Soc. 2000;14:647–668.
10. Whitehouse D. Workplace presenteeism: how behavioral professionals can make a difference. Behav Healthc Tomorrow. 2005;14:32.
11. Aronsson G, Gustafsson K. Sickness presenteeism: prevalence, attendance-pressure factors, and an outline of a model for research. J Occup Environ Med. 2005;47:958–966.
12. Hansen CD, Andersen JH. What personal circumstances, attitudes and work-related factors are associated with sickness presenteeism? Soc Sci Med. 2003;67:956–964.
13. Burton WN, Pransky G, Chen C-Y, Edington DW. The association of medical conditions and presenteeism. J Occup Environ Med. 2004;46:S38–S45.
14. Brooks A, Hagen SE, Sathyanarayanan S, Schultz AB, Edington DW. Presenteeism: critical issues. J Occup Environ Med. 2010;52:1055–1067.
15. Kessler RC, Ames M, Hymel PA, et al. Using the World Health Organization Health and Work Performance Questionnaire (HPQ) to evaluate the indirect workplace costs of illness. J Occup Environ Med. 2004;46(suppl):S23–S37.
16. Goetzel RZ, Long SR, Ozminkowski RJ, Hawkins K, Wang S, Lynch W. Health, absence, disability, and presenteeism cost estimates of certain physical and mental health conditions affecting U.S. employers. J Occup Environ Med. 2004;46:398–412.
17. Kuper H, Marmot M. Job strain, job demands, decision latitude, and risk of coronary heart disease within the Whitehall II study. J Epidemiol Community Health. 2003;57:147–153.
18. Donald I, Taylor P, Johnson S, Cooper C, Cartwright S, Robertson S. Work environments, stress and productivity: an examination using ASSET. Int J Stress Manag. 2005;12:409–423.
19. Johansson G, Lundberg I, Johansson G, Lundberg I. Adjustment latitude and attendance requirements as determinants of sickness absence or attendance. Empirical tests of the illness flexibility model. Soc Sci Med. 2004;58:1857–1868.
20. Grinyer A, Singleton V. Sickness absence as risk-taking behaviour: a study of organisational and cultural factors in the public sector. Health Risk Soc. 2000;2:7–21.
21. Caverley N, Cunningham J, MacGregor JN. Sickness presenteeism, sickness absenteeism, and health following restructuring in a public service organization. J Manag Stud. 2007;44:304–319.
22. Faragher EB, Cooper CL, Cartwright S. A Shortened Stress Evaluation Tool (ASSET). Stress Health. 2004;20:189–201.
23. Johnson S, Cooper C. The construct validity of the ASSET stress measure. Stress Health. 2003;19:181–185.
24. Johnson S. Organizational screening: the ASSET model. In: Cartwright S, Cooper CL, eds. The Oxford Handbook of Organizational Well-Being. Chichester, UK: John Wiley and Sons Ltd; 2009:133–155.
25. Michie S, Williams S. Reducing work related psychological ill health and sickness absence: a systematic literature review. Occup Environ Med. 2003;60:3–9.
26. Danna K, Griffin RW. Health and well-being in the workplace: a review and synthesis of the literature. J Manag. 1999;25:357–384.
27. Allen H. Using routinely collected data to augment the management of health and productivity loss. Occup Environ Med. 2008;50:615–632.
28. Munir F, Yarker J, Haslam C. Sickness absence management: encouraging attendance or “risk-taking” presenteeism in employees with chronic illness? Disabil Rehabil Int Multidiscip J. 2008;30:1461–1472.
29. Virtanen P. “An epidemic of good health” at the workplace. Br Nurs Index Arch Sociol Health Illness. 1994;16:394–401.
30. Arbuckle JL. Amos 18 User's Guide. Crawfordville, FL: Amos Development Corporation; 2009.
31. Denison DR, Hart SL, Kahn JA. From chimneys to cross-functional teams: developing and validating a diagnostic model. Acad Manag J. 1996;39:1005–1023.
32. Sabherwal R, Jeyaraj A, Chowa C. Information system success: individual and organizational determinants. Manag Sci. 2006;52:1849–1864.
33. Bentler P. M. Comparative fit indexes in structural models. Psychol Bull. 1990;107:238–246.
34. Schumacker RE, Lomax RG. A Beginner's Guide to Structural Equation Modeling. Mahwah, NJ: Lawrence Erlbaum Associates; 1996.
35. Gilbreath B. Creating healthy workplaces: the supervisor's role. In: Cooper CL, Robertson IT, eds. International Review of Industrial and Organizational Psychology. Vol 19. New York: John Wiley and Sons Ltd; 2004:93–118.
36. Robertson IT, Flint-Taylor J. Leadership, psychological well-being and organisational outcomes. In: Cartwright S, Coooper CL, eds. Oxford Handbook on Organisational Well-being. Oxford, UK: Oxford University Press; 2009:159–179.
37. Nyberg AH, Westerlund H, Magnusson Hanson LL, Theorell T. Managerial leadership is associated with self-reported sickness absence and sickness presenteeism among Swedish men and women. Scand J Public Health. 2008;36:803–811.
38. Demerouti E, Le Blanc PM, Bakker AB, Schaufeli WB, Hox J. Present but sick: a three-wave study on job demands, presenteeism and burnout. Career Dev Int. 2009;14:50–68.
39. Sparks K, Faragher B, Cooper CL. Well-being and occupational health in the 21st century workplace. J Occup Organ Psychol. 2007;74:489–509.
40. Morrison DL, Payne RL. Multilevel approaches to stress management. Aust Psychol. 2003;38:128–137.
41. Black C. Working for a Healthier Tomorrow. London: The Stationery Office; 2008.
42. Aronsson G, Gustafsson K, Dallner M. Sick but yet at work. An empirical study of sickness presenteeism. J Epidemiol Community Health. 2000;54:502–509.
43. Rantanen I, Tuominen R. Relative magnitude of presenteeism and absenteeism and work-related factors affecting them among health care professionals. Int Arch Occup Environ Health. 2011;84:225–230.
44. Allen HM, Bunn WB. Validating self-reported measures of productivity at work: a case for their credibility in a heavy manufacturing setting. J Occup Environ Med. 2003;45:926–940.
©2012The American College of Occupational and Environmental Medicine