We have all seen people exhibit flagrantly unhealthy behavior. Some of us–though we’d never admit it–derive a certain, smug satisfaction by observing them. At least I don’t do that!

Somewhere in the course of our daily lives though, most of us do exhibit behavior that suggests at least some disregard for our health. We don’t change our diet, though we know we should. We don’t floss, take medications as prescribed, or get the screening tests we’re supposed to.

Multiple intertwining causes underlie all unhealthy behavior, of course. I had always figured that one pervasive cause was the lack of a simple, observable connection between health-related behaviors and health outcomes. There is a long delay for example, between establishing unhealthy dietary preferences  and the sequellae of that behavior ( a heart attack, diabetes or whatever). The longer the delay between cause and effect, the more likely someone will be to exhibit unhealthy behavior.

On the other hand, if there’s a short interval between cause and effect—it only takes minutes for susceptible people to develop a severe allergic reaction after eating peanuts, for example—well, that’s where I’d expect high adherence to the required healthy behavior.

If I’m right, then we have a problem. For many chronic diseases (diabetes, heart disease, some cancers) the interval between cause and effect can be decades.

How might this reasoning apply to a person that has already been diagnosed with a chronic disease? Assuming providers have explained things to him or he has learned these things on his own, that’s a person that knows his behavior caused his predicament. And if he knows that, he should also know he has a fresh chance to rectify matters, at least to a degree. If he modifies his unhealthy behavior, then he can control the progression of, and indeed sometimes reverse the progression of his disease.

One way to test these assumptions is to study attrition rates from web-based and mobile behavioral intervention programs among people with chronic diseases. In an earlier post on the subject, I showed that many of these programs do work. That is, they foster more healthy behavior. But they are plagued by a massive attrition problem. In one study for example, nearly 70% of registered open access users to an online physical activity web site stopped using the site altogether within months after starting.

But what if participants weren’t healthy to begin with? What if they had diabetes, for example? If the above assumptions hold true, these people might be more engaged in such programs.

Alas, a study by Russell Glasgow and colleagues at Kaiser Permanente suggests this is not the case. Attrition from an Internet-based diabetes self-management program was a huge problem in this study.

Glasgow’s team looked at website engagement and attrition from a program known as My Path to Healthy Life (Spanish version: Mi Camino a la Vida Sana). All participants had diabetes. The team reported that 75% of all participants entered at least some self-monitoring data at least once per week. However, use of site extinguished dramatically with time. By 4 months after study onset, many users had stopped using it altogether, or visited it only sporadically. Overall, web site utilization dropped from 70% of participants visiting at least once per week during the first 6 weeks to 47% during the next 8 weeks.

Interestingly, factors like age, ethnicity, baseline computer use, health literacy and education were not related to site utilization or the risk of attrition.

What can we make of this?

It’s hard to know whether the initially high utilization rates were caused, as I hypothesized above, by increased awareness of cause and effect. Many other factors could have driven the initially favorable utilization rates.

The My Path to Healthy Life program was designed for example, to promote user interaction and choice specifically to encourage site utilization. Users could write their own alternatives to suggested recommendations and strategies, for example. The site also featured copious visual and auditory displays, customized feedback reports and prompts to keep users involved. In addition, rotating quiz questions and motivational tips, efforts to address issues of literacy, user forums, graphical displays and email reminders to return to the site could all have had salutary effects.

Still, the bottom line is that attrition from the program was ghastly: less than half the participants went to the site at least once per week by the end of the study, and the downward trend in utilization had not plateaued by then, either.

This speaks against my hypothesis. It calls attention to a key distinction behavioral psychologists like to make in explaining human motivation: there’s a difference between awareness of the need to act, and the volition to act. It appears that neither knowledge of what needs to be done, a clear understanding that doing something healthy can improve prognosis from a chronic disease, nor the positive stimulus provided by an engaging, informative self-management program are enough to trigger the volition to change behavior.

Glenn Laffel is a cardiologist who blogs at Pizaazz.

Submit a guest post and be heard on social media’s leading physician voice.