Financial Impact Impact of a Health Promotion Program on Employee Health Risks and Work Productivity Peter R. Mills, MD Ronald C. Kessler, PhD John Cooper, MD Sean Sullivan, JD Abstract Purpose. Evaluate the impact of a multicomponent workplace health promotion program on employee health risks and work productivity. Design. Quasi-experimental 12-month before-after intervention-control study. Setting. A multinational corporation headquartered in the United Kingdom. Subjects. Of 618 employees offered the program, 266 (43%) completed questionnaires before and after the program. A total of 1242 of 2500 (49.7%) of a control population also completed questionnaires 12 months apart. Intervention. A multicomponent health promotion program incorporating a health risk appraisal questionnaire, access to a tailored health improvement web portal, wellness literature, and seminars and workshops focused upon identified wellness issues. Measures. Outcomes were (1) cumulative count of health risk factors and the World Health Organization health and work performance questionnaire measures of (2) workplace absenteeism and (3) work performance. Results. After adjusting for baseline differences, improvements in all three outcomes were significantly greater in the intervention group compared with the control group. Mean excess reductions of 0.45 health risk factors and 0.36 monthly absenteeism days and a mean increase of 0.79 on the work performance scale were observed in the intervention group compared with the control group. The intervention yielded a positive return on investment, even using conservative assumptions about effect size estimation. Conclusion. The results suggest that a well-implemented multicomponent workplace health promotion program can produce sizeable changes in health risks and productivity. (Am J Health Promot 2007 22[1]:45���53.) Key Words: Health Promotion Efficiency Absenteeism Health Status Indicators Prevention Research. Manuscript format: research Research purpose: intervention testing/program evaluation Study design: quasi-experimental Outcome measure: productivity, absenteeism, behavioral Setting: workplace Health focus: fitness/physical activity, nutrition, smoking control, stress management weight control Strategy: education, skill building/behavior change Target population: adults Target population circumstances: education/income level INTRODUCTION Epidemiologic research has consis- tently documented a positive associa- tion between a variety of health risk factors and direct costs of illness among working people in a variety of settings.1���4 Both the number and type of health risks have been shown to correlate directly with future medical and pharmacy claims costs.1,5,6 Indi- viduals with multiple concurrent health risk factors have been shown to incur approximately double the costs of those with few or no risks.1,7 In addition, specific lifestyle risk factors such as tobacco use (current and previous), obesity, stress, and lack of regular physical activity individually confer consistently greater risks of incurring higher costs.4,8���11 In addition to the relationship be- tween employee health risks and direct costs, there is a significant amount of evidence linking health risks and in- direct business costs in the form of absenteeism, workers��� compensation costs, and decreased work perfor- mance (presenteeism).2,12���16 A number of studies have shown that health promotion programs can have positive effects on both individual health risk status and the associated costs.17���21 Improvements in health risk status can yield significant reductions in medical claims costs, which are all the more relevant when one notes that increases in health risks over time lead to sharp increases in incurred costs.17 In addi- tion, participation in health promotion programs can reduce absenteeism compared with preprogram levels and in relation to nonprogram partici- pants.22���24 More recently researchers have fo- cused upon the area of presenteeism, Peter R. Mills, MD, is the Chief Medical Officer at Vielife Limited and is with the Department of Respiratory Medicine, The Whittington Hospital, London, United Kingdom. Ronald C. Kessler, PhD, is a Professor at the Department of Health Care Policy, Harvard Medical School, Boston, Massachusetts. John Cooper, MD, is the Global Head of Occupational Health at Unilever PLC, London, United Kingdom. Sean Sullivan, JD, is the President and Chief Executive Officer at the Institute for Health and Productivity Management, Phoenix, Arizona. Send reprint requests to Peter R. Mills, MD, Vielife Limited, 68 Lombard Street London EC3V 9LJ, United Kingdom p.mills@vielife.com. This manuscript was submitted December 18, 2006 revisions were requested March 26 and May 8, 2007 the manuscript was accepted for publication June 4, 2007. Copyright E 2007 by American Journal of Health Promotion, Inc. 0890-1171/07/$5.00 + 0 September/October 2007, Vol. 22, No. 1 45

and in particular, quantifying the decrements in work performance as- sociated with health risks.15,25���30 In- deed, the putative indirect workplace costs of illness have been the subject of growing interest in recent years as part of the larger interest in value-based purchasing.31 It has been suggested that the cost of presenteeism to a busi- ness may exceed the combined costs associated with medical claims and absenteeism.32 Another reason for the increase in interest in the area of work performance may be due to the shift away from manufacturing to service- based economies in the developed world, with attendant increases in competitive advantage based on hu- man capital investment.33 A number of self-report question- naires designed to assess work perfor- mance have been developed and vali- dated to help objectify research in this area.34���37 Studies using these instru- ments have documented strong asso- ciations between employee health and work performance.15,38���40 In addition, some uncontrolled studies have shown positive associations between improve- ment in health status and work per- formance.41,42 To our knowledge, there are no prospective controlled investi- gations that have looked specifically at the impact of workplace health pro- motion programs on presenteeism to date. The current report presents the results of such a study in which a quasi- experimental design was used to eval- uate the effects of a multicomponent health promotion program on change in health risk status and work perfor- mance in a large corporation over a 12- month intervention period, compared with changes in the same outcomes among propensity score-matched controls. METHODS Design This was a quasi-experimental 12- month before-after intervention-con- trol study. Participants in both the control and intervention groups com- pleted a health risk appraisal (HRA) questionnaire and the work perfor- mance section of the World Health Organization health and work perfor- mance questionnaire (WHO-HPQ) at baseline and in a follow-up survey 12 months after baseline. A multicompo- nent health promotion program was delivered to the intervention group during the intervening 12 months. No health promotion initiatives were pro- vided to the control population during the study period. The primary outcome variables were differences between the intervention and control groups in (1) cumulative count of health risk factors (derived from the HRA), as well as the WHO- HPQ derived measures of (2) work- place absenteeism and (3) work per- formance at the 12-month follow-up interview, adjusting for baseline differ- ences on these same outcomes. Because the control group was not completely comparable to the inter- vention group at baseline, a weighting adjustment was made to correct for these differences using the method of propensity score.43 The results re- ported here are based on those weighted data. Sample The intervention group consisted of the 618 full-time employees from three United Kingdom business units of Unilever PLC, a multinational manu- facturer of food, home care, and personal care products. Employees of the three business units were office based and engaged in service delivery rather than manufacturing. All inter- vention group members were asked to complete a secure online survey about their health and work performance both at baseline and 12 months after baseline. The response rate was 84.0% (n 5 519) at baseline and 43.0% (n 5 266) at follow-up. The control group consisted of a convenience sample of 1679 individuals, recruited from the community by a market research firm, who completed the baseline survey. A total of 2500 requests to participate were made by the firm, representing a 67.2% response rate. Control group members were selected to be employed full time in office-based service delivery jobs to match the intervention sample. The control group response rate at the 12-month follow-up was 49.7% (n 5 1242). Table 1 details the numbers of respondents within each group at baseline and 12 months. Measures The baseline and 12-month follow- up surveys both included an HRA questionnaire38 and the absenteeism and work performance questions from the WHO-HPQ.37 The full details of the validation research that each ques- tionnaire has undergone are published elsewhere however, briefly, the HRA is a 24-item instrument that was specifi- cally designed to capture health risk data within the corporate setting. The questions cover a number of domains, either as single items or as composites of multiple items, and capture data on risk status in 12 areas that have been shown to strongly predict future costs among employed people, either from medical care utilization, absenteeism, or presenteeism1,4,8,44 (see Table 2 for Table 1 Number of Participants in Each Group at Baseline and 12-Month Follow-up Intervention Group Control Group Total eligible population invited to participate 618 2500 Baseline questionnaire respondents (% of total population) 519 (84.0) 1679 (67.2) Study end questionnaire respondents 266 1242 Study end questionnaire respondents as % of total eligible population at baseline 43.0 49.7 Study end questionnaire respondents as % of baseline respondent population 51.3 74.0 46 American Journal of Health Promotion