Using Prescription Data to Assess the Best (and Worst) Workplaces

Fortune's 2018 Best Places to Work list just came out, and I was surprised to see Salesforce take the #1 position. That's because when I did interviews with people for my forthcoming book, Dying for a Paycheck, several of the people I interviewed worked for Salesforce, and what they described was far from a healthy workplace. One person described facing pressure to return from maternity leave two weeks after delivering her child to make an important presentation. Another, several months pregnant at the time, found herself on a plane to Europe on Saturday night to be able to attend a Monday meeting--having been given little notice of the unexpected trip and not much choice to opt out. Of course, workplace health and being a "great place to work" may not be the same thing--but one would think that these two workplace measures would be highly correlated.

Maybe I just ran across some unrepresentative outliers in my research. That's possible. But I am increasingly convinced that much of the survey data used in constructing reviews is unreliable. We already know this is the case for consumer reviews of products and services. Because evidence shows that people rely on these reviews in their decision-making--the reviews are economically consequential--companies have incentives to get fake reviews posted. And they do. One peer-reviewed study reported that 16 percent of restaurant reviews on Yelp are fake, and a news story reported that one-quarter of the reviews submitted to Yelp are also fake, with similar problems on Facebook and Google. There is a burgeoning scholarly literature on ways of uncovering fake reviews, and all of the major websites that post reviews have ever-evolving algorithms attempting to discern review fraud. Nonetheless, scholarly research suggests that the validity of online user ratings is poor as assessed by several different methods.

Meanwhile, Glassdoor, which aggregates employee reviews for a list published in Forbes, and the Great Place to Work Institute which compiles the list for Fortune, are for-profit organizations that sell products and services to some of the same companies they are rating. The possibility for conflicts of interest abounds.

What to do to diagnose the real truth about workplaces? One of the Salesforce people noted they had gone on antidepressants a week after joining the company, and sort of half-facetiously suggested that maybe the proportion of people on such drugs might be a useful indicator of how toxic workplaces really are. It turns out that this suggestion actually has substantial research validity--prescription drug use can, if such data were publicly available, as they are in Denmark, provide excellent indicators of workplace stress and whether or not people are working in health-enhancing or health-destroying workplaces.

Michael S. Dahl, a professor in the department of management at Aarhus University in Denmark. has used prescription data to examine a number of interesting issues. For instance, he and colleagues found that first-time entrepreneurs, and their spouses, increased their use of sedatives/hypnotics in the first two years after becoming entrepreneurs, but their use of antidepressants declined. In another study, he reported that when wives outearned their husbands, the men were more likely to use erectile dysfunction medication while the breadwinner wives exhibited increased use of anxiety and insomnia medicines. And a third interesting research study found that stress-related medicine prescriptions increased for employees in firms undergoing significant organizational changes. His and other research suggests that prescription drug use can be a useful measure of social stress.

At an all-hands meeting at Salesforce last fall, employees saw data showing dramatic increases in medical claims and prescription drug use. I saw a screen shot of these data which, to my knowledge, have not been publicly shared even though they were presented to thousands of employees. If I had to rely on one indicator of what is going at the workplace, I would be more inclined to trust the medical and prescription claims changes than the GPTW rankings.

So next time you are interviewing for a job, don't just rely on the ratings. Ask, in a nice and polite way, the people interviewing you about what drugs they are taking. That may tell you much more about the conditions in the workplace you are contemplating joining.