For some reason, there seems to be vastly more rigorous research published about the problems and costs of health behaviors and conditions than about how much “solving” them can yield in potential financial benefit to employers, or personal benefit to employees.  There are at least three reasons I can think of for the disparity:

* Problems are easier to measure
* When separate problems are totaled, they seem to add up to far more than the total potential available in their solution
* Once problems exist, their solution never saves as much as the problems cost

Measurement Ease

It is relatively simple to measure both the prevalence and negative effects of health problems, though it is far more difficult to measure the results of health-focused solutions in ways that are credible to employers.  A wide range of highly-qualified university professors and research centers have worked on identifying the size of the problem.  Academics are often exceptionally good at that, and it is far easier than coming up with or implementing a solution.

The size of the health insurance problem, for example, amounts to pretty much the entire employer-paid health insurance expense, or self-insured outlays of employers.  No question is normally raised as to whether that is all due to health.  The same applies to measuring workers compensation and disability insurance, since their expenses are directly and clearly related to injuries or illness among employees.  There is always some problem with measuring absence and impairment at work (“presenteeism”), but there are at least half a dozen well-validated survey approaches to estimating this problem.

By contrast, the literature on EHM solutions is replete with media stories and press releases, rather than rigorous research, with much of it provided by either happy customers who don’t want to look a gift horse in the mouth, or EHM providers who want to make their efforts look as successful as possible.   Much, if not most of such publications tend to overlook problems of regression to the mean and self-selection bias, reporting well-measured changes in employer finances, but not necessarily linking them scientifically to their own efforts.

Double Counting

When EHM productivity and performance problems are measured, the most usual practice seems to be to ask employees to describe how impaired they are, using some validated self-reporting survey, then look for the health or impairment conditions they also have.  Since the vast majority of employees have more than one impairment condition, this can result in an enormous degree of “double counting”, where the same impairment is counted every time a specific condition linked to it is totaled.

For example, I have one report that reflects the total impairment linked to fourteen separate conditions, including: risk behaviors such as smoking, inadequate sleep, poor diet and lack of exercise; risk conditions such as obesity, high blood pressure, high cholesterol; and chronic diseases such as diabetes and asthma.  When totaled, the sum of all the separate impairment-factor-specific amounts of impairment came to almost fifteen percent.  This was adjusted to account for an assumed “normal” level of impairment applicable to people without such factors, but amounted to a vary large overall loss of productivity and performance.

But when the prevalence of all the separate impairment factors was totaled, it came to almost 250%.  This clearly indicates that each impairment factor’s individual impairment effect was counted at least two-and-a-half times.  The total amount of impairment that could possibly be regained through EHM interventions was less than half the total reported, since each individual can only be improved once.  It might be that by treating the more than one impairment factor the average employee had, a greater recovery of lost productivity could be achieved, but there would be no way to recover more than 100% of it.

Solving Health Problems Doesn’t Restore Perfect Health

While there are no doubt many individuals, and even classes of health problems that can be solved so completely that they leave the individual affected as good as a perfectly healthy employee in terms of productivity and performance.  But since the average person tends to have many problems simultaneously, an EHM program that addresses only one of them is unlikely to restore anyone to perfect health.  And even if all health problems are eliminated, there is likely to be some residual damage that remains.

In the cited example, calculating a “normal” level of impairment is an effective way of dealing with this limitation.  But the most effective approach is simply to report how much productivity is improved by particular EHM solutions, merely keeping the reported impairment levels associated with the problem addressed as a likely limit.  Fortunately, most EHM experience indicates that interventions increase their results over time, so the initial degree of impairment may one day be recovered completely, though it may take longer to do so than the average employee remains with the same employer.

Reporting Results

If an employee is enrolled in a particular EHM intervention, and rigorous analysis (eliminating regression to the mean and self-selection bias, for example) shows that that employee reports a 10% improvement in productivity or performance, that’s a good start.  There is still the challenge of “proving”, or at least persuading employers, that this improvement is real, i.e. that a self-reported improvement equals a real one.  In one study, for example, the average migraine-impaired call center employee reported a 20% impairment, while objective data indicated that the actual loss of productivity was only 8%. [G. Pransky, et al. “Performance Decrements Resulting from Illness in the Workplace” JOEM 47:1 Jan 2005 34-40]

Added to this bias is the potential that when workers report productivity improvement, they may consciously or unconsciously over- or under-estimate their own recovery to a different degree than they did in reporting their impairment.  If so, then even the same instrument, validated through comparing self-reported to actual impairment, might not work for measuring recovery.  Employees whose impairment is estimated at 20% by a given tool, converted to “reality” by using a conversion factor of 40%, may even under-estimate their recovery when reporting that, so the 40% conversion would be way off.

But at least, once employees in a given EHM program report a reliably validated recovery or improvement, that is the true result they have achieved.  Unless they enroll in more than one EHM intervention, the result can be counted, and there will be no double counting.  And since there have been validated reports of recovery as great as 8% in as little as 180 days, the results, at least should be happy news for employers when they more than cover the cost of the intervention.