The content here is relevant because many health professionals (okay, me) are not trained well in the causes of obesity, and the causes are not actually known in all cases. So this is a good review of what’s known about the physiology, plus the experiment itself, which I’ll discuss in a bit.
Social Media as a platform for academic exchange – finally
This study published in JAMA a few weeks ago (September, 2016), produced unexpected (and curious) results.
Overweight and obese younger people randomized to receive wearable devices as part of a weight loss program gained back more weight than users who did not receive wearables, after an initial 6 month weight loss.
Both sets of subjects did not have significantly different rates/intensity of physical activity over the 2 year study, and their dietary intake was not statistically significant from each other (calories taken in slightly less for the wearables group at the end). Specifically, the group with the wearable did not exercise more than the group without the wearables.
I was of course curious and decided to look more closely at the data. I produced some charts below.
Here are the things I noticed
Subjects were randomized at the very beginning of the study, not at the 6 month mark, when the wearables were initiated. Did they know which group they were in at the beginning and did this shape their behavior?
I ask the question above, because the one thing I noticed in charting the data is that the group with the wearables (EWLI – stands for “Enhanced Weight Loss Intervention”), experienced a visible plunge in MVPA: “Nonsupervised moderate-to-vigorous physical activity” even before they got the wearables, that continued well past the time they had the wearables. Overall, though, across the 24 months, there was not found to be a significant difference in physical activity.
The subjects were placed on what are essentially high carbohydrate diets with caloric restriction, which remained restricted throughout the 24 months.
Outrunning a bad diet?
I was recently introduced to the work of Tim Noakes (@ProfTimNoakes) (about 5 years behind the rest of the world, but maybe 1-2 years ahead of part of the world) and decided to look more closely at other factors.
Focusing on the diet of the subjects, here’s what it said in the study details (supplemental materials)
All subjects will be prescribed an energy restricted dietary intervention that we have shown to effectively reduce body weight by 8-10% within the initial 6 months of treatment. This will include reducing energy intake to 1200 to 1800 kcal/d based on initial body weight (<200 pounds = 1200 kcal/d; 200 to 250 pounds = 1500 kcal/d; >250 pounds = 1800 kcal/d). Data from our research studies [14, 15] and the National Weight Control Registry  indicated that macronutrient composition in the most successful participants consists of 20-30% dietary fat intake, 50-55% carbohydrate intake, and 20-25% protein intake. Therefore, a similar dietary composition will be recommended in this study. However, we do recognize that low carbohydrate/high protein diets are currently popular, have demonstrated some initial efficacy, and some participants may gravitate towards this macronutrient composition, and this will be acceptable provided that total energy intake is within the prescribed range. To facilitate the adoption of the dietary recommendations, individuals will be provided with meal plans (see Appendix B), that will allow them to plan for modifications in their daily and weekly meal plans, and a calorie counter book.
So they were permitted to lower their carbohydrate intake as long as they maintained the same amount of calorie restriction. As a group they did not do this, though. They stuck to their high carbohydrate diets over the long run.
For all its merits, however, exercise is not an effective way to lose weight, research has shown. In a cruel twist, many people actually gain weight after they start exercising, whether from new muscle mass or a fired up appetite.
This study is about wearables, not exercise, because both groups of people exercised about the same over time.
However, because both groups were on a high-carbohydrate diet throughout the intervention, it’s possible that even if the wearables “worked” (they exercised more), that the results would be the same.
Community Commons (@CommunityCommon) to the rescue. Earlier this year they so very nicely agreed to add Washington, DC ward boundaries to their most awesome mapping system , and some key health data points based on DC-level data. They even created a special hub “Center for Total Health” that’s invite only so I can bring community health activists to map their city, collaboratively.
Here’s the obesity map for Washington, DC:
Here’s the smoking status map for Washington, DC by Ward:
These are new; previously there was no way to understand DC’s health using an interactive system because all of the data is clumped at the county level. As you can tell from the above, if you see an obesity rate of 21 % for “Washington, DC” you’ll miss important distinctions.
Now, here’s the race / ethnicity map for Washington, DC, available by census tract:
Do you see a picture of different health status on the left side of the map vs the right side of the map?
Wanna play? Click on either map
The new capabilities provided by Community Commons allow us to map any sub-county level data over ward boundaries so we can understand our city better. I even created a few interactive ones that you can play with here. Just click through, you’ll need to create an account on communitycommons.org to make any changes. Up to you.
While I was at it, I also created a map using DC-data-whiz Michael Schade’s (@mvs202) interactive Google Places map, plotting presence of what Google labels “gyms” in a 4-metro station radius of Shaw/Howard University Station:
You can see that gym businesses in the Google places database encircle Dupont Circle (14.4 % obesity rate), with a lot less presence east of Shaw/Howard University Metro (17.4 % obesity rate heading into 35 % obesity rate). These distinctions are important – people who spend their time in the western half of the city may believe they own a culture of health, but it’s only they that do.
Yesterday, I happened upon a ribbon cutting ceremony by our Mayor (@MayorVinceGray) (who has done incredible things for human rights) in that neighborhood. I asked a bystander if she was here during the riots that destroyed this part of the city in 1968. She said, she was, and it was a scary and sad time. I believe her.
Photos: S Street, 1968 and 2014 – next to Shaw/Howard Street Metro
Photos of/from …
She told me that her friends told her she should have taken photos of the before and after, because so much had changed. I responded that there were still plenty of photos to take – even though the buildings look new, there’s plenty of “before” to be found, the maps show it.
Thanks a ton, again, to Community Commons for being responsive/interested/supportive and no negative vibes meant to County Health Rankings – we are all friends and the two resources go together in the most useful way. Sometimes you have to go to the places that no one else goes to find and create innovation in health. That’s what social innovators do 🙂 .
If you’d like to do some DC mapping, drop me a line/comment/tweet and I’ll invite you to the hub on Community Commons.
My answer to the above is yes. I have told people at the Center for Total Health (@kptotalhealth) that I consider a patient who presents to their doctor without any physical activity a medical emergency. I get looks of surprise when I say this, sometimes a slight chuckle.
Bob passed me this paper, which shows, in a study of 14,345 men, assessed over 7 years, that weight was not associated with death. Level of physical activity is, across all weights.
I have believed that lack of physical activity is a medical emergency long before this study. There are several reasons for this. Before I say what those are for me, I’ll see if you agree or what you think of this concept – post in the comments, or on your favorite social network, please…
As compared with the control group, the group with a randomly assigned opportunity to use a voucher to move to a neighborhood with a lower poverty rate had lower prevalences of a BMI of 35 or more, a BMI of 40 or more, and a glycated hemoglobin level of 6.5% or more, representing relative reductions of 13.0%, 19.1%, and 21.6%, respectively. The magnitudes of the associations with health were larger still for participants who moved with a voucher that was restricted to use in a low-poverty area than they were for the intention-to-treat estimates for all participants who received the restricted voucher and are consistent with the effect sizes reported in previous observational studies.3 Because we generated estimates for several BMI cutoff points, our estimates for the associations between program participation and extreme obesity may be marginally significant.
Is moving to a different neighborhood as powerful or more powerful an intervention than health care, health education, apps, reminders, games and other individually-directed behavior change strategies people are talking about with obesity? It seems to have an impact.
What an interesting study – randomizing families (via incentive vouchers) to either move to a neighborhood with a low poverty rate, a neighborhood of their choice, or no intervention, between 1994 and 1998, with a 10-12 year follow-up. How often does that happen.
Not every family offered the voucher to move actually moved, and the analysis takes this into account (via intention-to-treat). It appears to my eyes that all families that did move (low poverty or move-where-you-want) went to neighborhoods that were less-poor and higher education levels, and there was an association with the impact on their health (weight, HbA1c).
Interestingly, at interview, none of the study groups reported having better access to non-emergency room health care, so that part of their neighborhood experience didn’t change, while their health experience did.
I recommend looking at Tables 2-3 – I can’t reproduce here to respect copyright, but the article is open access.
The quick answer is: don’t focus on the technology.
This is not one, not two, but three papers published in the last six months, the last one last week, encompassing an impressive body of work around behavior change and weight loss, from the same research group led by Bonnie Spring, PhD, at Northwestern University.
There’s a lot of stuff in here – everything from financial incentives to the way you coach people about behavior change to the use of mobile devices. If you don’t read them all (or don’t read any of them all the way through), it’s highly likely that you’ll come to the wrong conclusion, in my opinion, so I’m going to break it down here:
1. The way you counsel people about behavior change makes a difference
This is actually the most important part of the first two papers linked here, not the up to $730 incentive (more on that later). What the authors studied was competing theories about how to talk to people. They looked at four different ways, which I’m going to summarize, because it’s important – these are things that are cost-free and don’t require any capital expenditure to deploy 🙂
Theory 1: Familiarity hypothesis, I’ll call it “MOTS: More of the Same” – “Decrease your saturated fat intake, increase your physical activity”
Theory 2: Behavioral Economic hypothesis: Crowd out an unhealthy behavior with a heathy one – “Eat more fruits and vegetables, decrease your sedentary activity” <- This was the winner
Theory 3: Low Inhibitory Demand, I’ll call it “Don’t be a downer, tell people what they CAN do rather than what they CAN’T” – “Eat more fruits and vegetables, increase physical activity”
People coached with the winning theory had significantly higher changes in a calculated “Diet-Activity” score compared to others. If you break it down a little more, it looks like it was far more likely that they could eat more fruits and veggies, than that they could increase their physical activity, across all groups. The winning group, though, dramatically decreased sedentary leisure, almost by half, which persisted 20 weeks later.
2. Paying people is a background activity to the above
In the first two papers, people were paid $175 to go through 3 weeks of intervention and point incentives along the way up to 20 weeks of follow up, for a total of $730. It seems like this got them through to recording their information. I don’t think it’s the most important feature of the study, and as the authors point out, probably not a realistic approach moving forward.
3. The mobile technology itself doesn’t help in isolation
Notice, I keep saying “mobile technology” not “smartphone.” That’s because these studies used PalmPilots (!) to support entry of data and target feedback. All of the study participants in the behavior change theory study improved their diet-activity score. There was no control group there, the goal was just to compare the theoretical approaches. The first study relied on self report of food intake and physical activity, which the authors sought to keep accurate by deploying a “bogus pipeline” approach where they told people to submit their grocery receipts and accelerometer data that were actually not used – clever.
The second study, which is in the third paper linked to this post, was more focused on weight loss itself rather than behavior change, and in it, the authors planned at the start that the mobile technology would be part of intensive coaching for all study subjects – they didn’t even try to have the mobile device make this happen for people. And in fact, the mobile device by itself didn’t make it happen for people – the one group who were randomized to get the mobile device and didn’t go to class actually gained weight. They gained more weight than the control group who had no mobile device.
What did happen for people was weight loss when they (a) used the mobile device to track and get feedback AND (b) they went to classes, in person. That was the requirement – that both happen, and when it did, this group lost more weight initially, and kept it off – average of 6.38 pounds at 12 months. The people with the mobile devices that didn’t go to classes, gained about 5 pounds at 12 months, which is more than the people who went to classes by themselves, and more than people who didn’t do the classes or get mobile devices.
There’s another really important piece of information in all of this, which is that the people who were selected to get mobile devices randomized into the study were selected after a two week trial of recording their information. About 35% of the people that went into this gate didn’t make it, so in the end, this was a study of people who can use mobile devices to record their information.
As I said, there’s a lot here
I’m looking at this from a population/social determinants perspective, and I would ask the question, “Should mobile devices or apps be deployed to the entire population to make weight loss happen compared to other approaches?” The answer for me would be “no.”
I think what the authors are demonstrating is the answer to this question, “Should mobile devices or apps be deployed to people who are motivated to use apps to lose weight and participate in intensive behavioral interventions?” to which the answer is more of a “yes.”
Also, “Should we get smarter in communicating with people, with technology and not, about their choices? – answer is “yes.”
The reason I love technology and have been invested in it for so long (and will continue to be) is because of its role in facilitating communication and connecting people to people. The best.app.ever is the human brain, the most important innovation in health information technology is listening. Oh, and prevention is the new HIT.
I communicated with Dr. Spring before writing this post to help me understand what we can take away from this research, which is very important/timely/useful (and I of course invited her to Washington, DC to the Behavior Change Summit in 2013 🙂 – More info on that here: Behavior Change: What can we learn from other industries? – EXAMPLES | Ted Eytan, MD). I learned a ton here that I didn’t know before, which makes me happy that talented behavioral scientists are working in this area.