To the Editor:
We have read the letter from Richard Mark Preece, MD, et al and appreciate their thoughtful review and comments on our research results.
We agree with the observation that the financial costs associated with a given amount of lost work performance vary across occupations, but we think it is important to point out that estimation of lost work performance (the focus of the health and work performance questionnaire, HPQ) can be separated from estimation of the financial implications of this lost performance for the employer. The HPQ is a population tool that measures time loss associated with ill health both for incidental absence and presenteeism. The monetization of that lost time is done separately and relies on models developed by Dr Sean Nicholson et al,1 which take into account different occupations, how the typical employer responds to lost time in those occupations, and the associated financial costs of those responses. The Nicholson monetization rules are designed to apply in the aggregate to workers of a given occupation and industry, but certainly should not be used by employers who have access to alternative rules that more closely approximate the real costs of lost work performance to their firms. Although it is unrealistic to think that any general-purpose monetizing rule will apply to all workers in all occupations in all firms, the Nicholson aggregate monetizing rules are nonetheless of considerable value in giving a general sense of the financial implications of lost work performance for research purposes.
Another measurement issue raised in the letter concerns the assessment of presenteeism using a metric that is anchored at the upper end by the work of a “perfect” employee. It is important to note that the measure of lost time because of presenteeism in the HPQ is not based on a person of “perfect” productivity, but on the actual scores of workers who do not have the medical conditions under study.
The letter also points out that the associations of specific conditions with self-reported absenteeism and presenteeism are not always intuitive. This is because of, at least in part, the fact that productivity loss is associated not only with the condition itself but also with severity and comorbity. For example, the strong association of obesity with presenteeism is likely because of the fact that obesity is often associated with multiple comorbid conditions. Future investigations are needed to examine the extent to which the gross effects of obesity vary as a function of occupation and industry, as well as the extent to which obesity is involved in mediating and modifying the impacts of related conditions. Furthermore, the treatment of certain conditions may be a source of health-related productivity loss, such as the use of diuretics to treat hypertension. Initial studies, like the one we presented, highlight broad associations and raise a great many secondary questions that go beyond the scope of a first report but provide valuable leads for future investigation.
Another issue in the letter concerns the fact that our study was unable to examine variation in the effects of medical conditions on work performance in all possible occupations. A much larger sample would be needed for that purpose, one that could estimate regression equations separately within individual occupations. Nevertheless, given that there are literally thousands of occupations listed in the O*NET system of job classification, it is unrealistic to think that any single study of finite size will ever be large enough to generate reliable estimates of this sort for each separate occupation. A more realistic goal is to investigate in a large general population sample of workers, the extent to which the effects of particular medical conditions on work performance vary as a function of one or more of the core job-demand dimensions underlying the O*NET classification (eg, analyzing data or information, making decisions, and solving problems). It might well be that useful specifications of this sort could be found, such as the possibly greater impact of conditions that adversely affect cognitive functioning on the performance of workers with cognitively demanding jobs. This kind of specification search is part of the long-term agenda for our ongoing studies of health problems and work performance with the HPQ surveys.
Another important area of generalization that remains to be explored in our HPQ studies is the effect of medical conditions on work performance in small firms. HPQ surveys, to date, have generally been performed in large firms because of the need for a large number of subjects to account for the relatively low prevalence of individual conditions. Nevertheless, there is no reason that similar surveys could not be performed in large samples made up of coalitions of small employers. Such studies would allow us to investigate the extent to which the results found in our large-employer studies generalize to small employers. We suspect that such a study will show that the effects of health problems on small employers will be greater than on large employers, because excess staffing and redundancy is typically not present in small firms and workers with specialized knowledge, skills, and job functions are difficult to replace temporality. Nevertheless, we agree with Dr Preece that this kind of study has not yet been performed and needs to be done in the future.
Our study also highlighted some additional future research needs regarding the relationship of health and work performance. These include a deeper understanding of the impact of illness severity, comorbidity, treatment variation, and medication adherence. Another intriguing question that we raised in the study article is “how much impairment can be reduced by improving health?” This is an issue related to interventions and their effects over time rather than the baseline measurement, so that we included in our study. Additional research is needed in this area as well.
We will be pursuing many of these issues in our future research and encourage others to do so as well. Together, we will continue to deepen our understanding of how employers can improve the health and productivity of their workforce.
Ronald Loeppke, MD, MPH
Michael Taitel, PhD
Vince Haufle, MPH
Ronald C. Kessler, PhD
Harvard Medical School
Thomas Parry, PhD
Kimberly Jinnett, PhD
Integrated Benefits Institute
San Francisco, Calif
1. Nicholson S, Pauly MV, Polsky D, Sharda C, Szrek H, Berger ML. Measuring the effects of work loss on productivity with team production. Health Econ
Readers are invited to submit letters for publication in this department. Submit letters online athttp://joem.edmgr.com. Choose “Submit New Manuscript.” A signed copyright assignment and financial disclosure form must be submitted with the letter. Form available atwww.joem.orgunder Author & Reviewer information.