5 Ideas About Feedback & Behavior Change, Supported by Evidence

For several months I've been a regular at the Quantified Self meetups, where people of great passion for self-tracking present the results of their experiments in self-monitoring. These vary in everything from people who record themselves sleeping (in search of better sleep) to people who monitor the amount of coffee they drink to the mililiter - and then compare that to their productivity. One thing that I've noticed is that there's a general presumption, among the QS crowd, that self-monitoring - i.e., feedback - is an experiment with evidence behind it. That is, that self-monitoring works, insofar as it helps people reach a goal, whether it's to drink less coffee or get better sleep or lose weight. But there's not often any actual evidence presented, beyond these experiments (where N=1, often enough). So tonight, drawing on research from my forthcoming book, I'm presenting a few points of fact that support the notion that feedback can lead to improvements in health. This post will serve as a primer for those not in the audience, and for those in the audience a repository of links to the research cited.

Thus, here are five claims or presumptions about the benefits of feedback, with the evidence.

1) Observing our actions leads to better actions.

For this one I turn to the National Weight Control Registry, a nifty study out of Brown University that is tracking more than 5000 people who have successfully lost weight and kept it off. The registry basically studies what these people are doing right. And in addition to the common-sense things like exercise often and change to a healthier diet (less fat and sugar, more vegetables and whole grains), the registry has found a strong association between the simple act of stepping on a scale and sustained weight loss. In particular, the study found that 75% weighed themselves at least once a week, and 44% weighed themselves at least once a day. Tracking our behavior leads to better behavior.

2) Engaging with Our Health Leads to Better Health

The idea here is that self-tracking not only improves the direct behavior, but that it has broader health implications, and can be quite positive for overall health. The evidence comes from a wonderful study called ALIVE! (short for A Lifestyle Intervention Via Email), conducted by the research wing of Kaiser Permanente.

This study sent out weekly reminder emails to study members, each message tailored to specific health goals. So the email would nudge them to meet their goal of eating, say, 3 vegatables a week, or getting exercise twice a week, and so forth. The subject would respond whether or not they met these goals, thus tracking their progress and tailoring the results for next week's email. The results: study subjects reported more than 50% greater improvement in concentration and productivity than a control group, and were 50% more likely to successfully change their diet. More broadly, they reported significantly better physical and mental quality of life. A link to the research is here.

3) When patients participate, their outcomes improve.

This idea goes a step further than simple self-tracking - it argues that by giving people an opportunity not only to track their health but to use that data to make decisions regarding their health, there is an upside, in terms of better outcomes. This one speaks directly to physicians and care providers who are reluctant to open the door to their own decision making and involve the patient in the discussion of treatments and care choices. Well, the Centre for Studies in Family Medicine at the University of Western Ontario took up this question - Does involving patients in their care, giving them a participatory role in their healthcare, improve the outcomes?

Their research showed a clear YES - involved patients had better recovery, better emotional health, and had a surprisingly 50% fewer diagnostic tests and referrals. In other words, they were happier with their care and the care was more successful. Not bad.

4) When patients participate, healthcare can be more effecient (read: Cheaper).

This one is going one step further, claiming that not only do self-tracking patients have better health, but that they actually have cheaper healthcare, to boot. For the answer to this idea, I come to a Boston University study that enrolled nearly 30,000 employees of the EMC Corporation in an online tracking program called DASH for Health.

The patients tracked their weight, blood pressure, and other information, and received tailored recommendations for dietary choices that they may not be aware of (such as: adding cream to your coffee instead of milk amounts to nearly 8 sticks of butter a month). The study found that among patients with an increased risk for cardiovascular disease, employees using the DASH program had on average $814 less spent on their healthcare annually. For all patients, the costs were basically neutral.

5) Data means more when it's our data.

This one is a response to the idea that healthcare woes in the US like obesity are an information problem - that there's a lack of resources or research or information for people. That's pretty clearly not true - we don't lack for information or health advice. But what we do need is relevant information, information that's tailored to our specific circumstances and conditions and that spells out the choices that our own situation holds.

This requires matching up our data with the right research data at the right time - when it's more specific and most useful. This is relevance. Annette O'Connor at the Ottawa Hospital Research Institute has done some extraordinary work in this area, looking at the growing body of decision aids available to patients to help them understand their options. Increasingly, these decision aids are turning up online in the form of nomograms - decision tools that allow us to input our personal information - data gleaned from our physicians or our own tracking - and then use that information to guide us to a decision that's right for us.

A couple years ago, O'Connor conducted a meta-analysis - a study of studies - examinging several hundred decision aids, with the goal of determining what effect they had, if any, on outcomes. It turns out that decision aids can actually be quite effective in improving our decisions and our outcomes. Patients using decision aids, O'Connor found, had significantly better understanding of benefits and harms, they perceived they were receiving better care, and they chose significantly less surgery (about 25% less). When they understood the pros and cons, it seemed, patients chose to avoid the knife more often, and were happier for it.

This research shows that self-tracking isn't just for a geeky fringe - it's actually a sensible strategy that anybody can take advantage of to gain some control of their health. Yes, often it requires a computer - but that's a good thing, insofar as it's easier to track and monitor stuff that way.