Researchers create new index to measure employee productivity

Link demonstrated between risk factors, lower production rate

A new study in the Journal of Occupational and Environmental Medicine has shown a direct link between several health risk factors and the failure of employees to attain certain productivity standards. But perhaps even more fascinating than the study’s results were the methods used by the researchers to measure productivity.

"The costs attributed to employee health problems are usually measured by employers in terms of direct health costs, such as medical plans claims," they wrote. "Although it has been understood that employee health problems also produce indirect costs for employers, their measurement has been far less frequent."1

To ensure that indirect costs were given appropriate weight, the researchers created what they call a Worker Productivity Index (WPI), which incorporates time lost to employee absenteeism and disability, as well as on-the-job decreases in productivity that may be due to health risk factors. This latter measure, which the researchers call "presenteeism," creates a more complete picture of the effects of health risks on productivity.

". . . Absenteeism and disability costs should be recognized, at best, as a significant contributor to an incomplete estimate of the total loss of productivity resulting from health impairment," they wrote. "What are seldom measured are the decreases in productivity for the much larger group of employees whose health problems have not necessarily led to absenteeism and the decrease in productivity for the disabled group before and after the absence period."

Getting a handle on presenteeism

In order to measure presenteeism, there are a number of prerequisites. First, a company must have a "health data warehouse" so that it can integrate different measures of employee health costs with its productivity measures. What should this "warehouse" include?

"It should generally include demographic information, absenteeism and/or disability days lost, HRA [health risk appraisal] results, and health promotion participation," says Wayne N. Burton, MD, senior vice president and corporate medical director at Bank One in Chicago, and lead author of the paper. "More sophisticated systems include occupational health records, insurance claim information, and other databases."

"It should include as much data about the person as possible, especially related to the four key outcome measures — medical costs, absenteeism, disability, and productivity," adds Dee W. Edington, PhD, of the Health Management Research Center at the University of Michigan, Ann Arbor, and a co-author of the paper. "My feelings are that it should also include data on the family, especially if health care costs are an outcome measure."

This data were readily available to the researchers, as the 564 participants in the study were telephone customer-service agents of First Card, a subsidiary of Bank One. First Card is the fifth-largest credit card issuer in the United States. The population was drawn from an operations, marketing, and service center in Elgin, IL, that employs approximately 3,000 people.

Because management knew it took about three months for new employees to become comfortable in their position, only employees who have worked with First Card for at least 13 weeks were eligible to participate. In addition, employees who were pregnant were not eligible.

A greater challenge was to establish the productivity standards by which presenteeism would be measured. The authors began by focusing on the two major productivity measurements commonly used for that type of employee. One involves the correctness of information given. The second deals with time: the amount of time between calls, the time customers are kept waiting, and so forth.

After conducting interviews with managers in a number of different departments, two measures were chosen for the study. The first was called "Handle Time." This referred to the total time spent on each call, including hold time, and time for after-call work. The second was called "Aux Time," short for auxiliary time. This represented time the employee was not available to receive phone calls — i.e., "logged off" the system. In both cases, the lower the value reported, the higher the productivity.

Ultimately, the WPI was calculated by integrating two critical measures: lost productivity away from the job because of illness, and time lost because of failure to maintain the productivity standard. Time away from work included time lost due to scattered illness and short-term disability absences. The latter measure was an estimate of the time lost because of lowered productivity while the employee was on the job, or presenteeism.

Worker Productivity Index (WPI) for Health Risks
Total Time Lost (Mean) WPI
Total with HRA, nonpregnant
13+ weeks 89
Selected Health Risks (at risk):
Current Smokers 4.147 90
Physical Activities (<1/wk) 3.238 92
Seatbelt Usage (<90%) 3.443 91
Encountered Violent Events 5.722 86
Distress 5.396 87
Biological Risks
Diabetes 11.364 72
High Blood Pressure 5.068 87
Cholesterol 6.128 85
Body Mass Index at Risk 5.790 86
Number of Risks
0 to 1 Risk Factor 4.059 90
2 Risk Factors 4.635 88
3+ Risk Factors 5.565 86
Source: The role of health risk factors and disease on worker productivity. JOEM 1999; 41:863-877.

What the study showed

When the results were studied, three health risks — general distress, diabetes, and BMI (body mass index) were found to have a significant relationship to a failure to attain the productivity standards that had been set.

"Those employees who self-identified themselves as having diabetes showed significantly more hours lost to illness, STD [short-term disability] absences, and to the failure to meet the productivity standard. Those with higher general distress scores showed significantly more illness hours and lost hours due to the failure to meet the productivity standard. Employees at risk for BMI scores also showed significantly more illness hours and lost hours due to the failure to meet the productivity standard," wrote the authors. (See chart, p. 26.)

The study data further indicated that "as the number of health risks increases, an employee’s productivity decreases." However, that correlation did not meet what researchers call "statistical significance." They conjecture that this may have been due to the relatively small sample size, as only 87 employees actually fell into the high health risk category. They also believe they may have been too liberal in the setting of standards used to determine whether or not productivity goals have been met. "Defining the failure at a higher threshold may have more clearly marked the interaction of high health risk and lowered WPI by more clearly identifying those employees with chronic failure to maintain the productivity standard," they assert.

"Taken together, absenteeism, short-term disability, and presenteeism yield a far more accurate picture of lost productivity stemming from ill health," notes Burton. "Our research showed that risk factors from an HRA are associated with on-the-job lost productivity. The same group of risk factors which are associated with increased health are costs and increased disability days also are associated with reduced worker productivity, or presenteeism."

"I was surprised by the strength of the relationship [between absenteeism, disability, and presenteeism]," adds Edington. "Also, this nearly completes our investigation of outcome measures for lifestyle factors. We have now demonstrated the relationship between health behaviors and medical costs, absenteeism, disability, and now productivity. Taken together, this is the measure of productivity."

A universal application?

Assuming that their companies have suitable health data warehouses upon which to create a proper statistical base, would other health promotion professionals be able to create WPI measures? In other words, just how universal are the standards established in this study?

"Our study is applicable to jobs answering the telephone and working on a computer terminal," notes Burton. "The applicability to other types of jobs will need to be studied. This type of research requires a significant investment of time and resources. Efforts are under way to find easier ways of measuring presenteeism."

"This is an important, but open question that needs continued studying," adds Edington.

Of course, the next step, once such data have been gathered, is to translate them into action, i.e., new health promotion strategies. Can the WPI be used in such a fashion? "It is far too early to use this information to design wellness programs," says Burton. "A great deal of additional research is needed. However, the usefulness of health risk appraisal information in the development of work site programs is clear."

Edington does see some immediate possibilities. "First of all, make sure the productivity standards are agreed upon by management and the employees in that work environment," he advises. "Then, given individual risk profiles and job preferences, you could make selective job assignments."

More important, says Edington, is the study’s "take-home message" for wellness professionals: "Healthy behaviors and attitudes add value to the contribution of an employee every hour of the day."

[For more information, contact: Wayne N. Burton, Corporate Medical Director, Bank One, 1 Bank One Plaza, Chicago, IL 60670-0006. Telephone: (312) 732-6434.]


1. Burton WN, Conti DJ, Chen CY, et al. The role of health risk factors and disease on worker productivity. JOEM 1999; 41:863-877.