72cf727474b8bf815fdc505e58cadfea
0

Paleo diet research study: pitfalls to avoid?

by (2626)
Updated June 27, 2014 at 6:22 PM
Created June 25, 2014 at 4:28 AM

I may have the opportunity to collaborate on a randomized, controlled study of the paleo diet. Is there any advice you would want to give, to make sure we avoid common pitfalls or misunderstandings? Are there flaws you've seen in past diet studies that we should watch out for? Here are some givens:

  • We'd be taking American adults with some kind of chronic health condition and randomizing them to go on either a paleo diet or some kind of conventional wisdom diet. We'd follow them for at least 4 months to get bloodwork and health outcomes.
  • Participants would get support and advice on following the diet, but ultimately it's up to them (no standardized meals or metabolic ward studies).
  • We would get some basic data on whether people were actually eating the diet they were assigned to.
  • We'd be sticking to pretty simple, food-focused guidelines. I'd roughly sum them up as: 1) Eat real food. 2) No dairy, grains, or legumes. 3) Don't be scared of fat (appropriately selected).

72cf727474b8bf815fdc505e58cadfea
2626 · June 27, 2014 at 6:22 PM

Thanks for the suggestions. Yes, we're getting a number of daily food logs. Based on our resources, it's simply impossible to get anything more precise at this stage.

Thumbnail avatar
10126 · June 27, 2014 at 6:56 AM

@AxialGentleman, once a year I fill out a survey form for the NWCR, which is run by Renata Wing of Brown University. The objective is to determine how weight loss is maintained after dieting, and it takes about half an hour to fill out. The food portion breaks down into servings per day on average, for the year. But many other things are covered, from exercise to meal frequency to sense of well-being. It's a population study with several thousand participants surveyed over decades. Keep your study groups as large as is practical.

32f5749fa6cf7adbeb0b0b031ba82b46
41312 · June 27, 2014 at 1:06 AM

Talk about completely unrealistic expectations for a dietary study.

0ba891d22837788c4d5ccf3f33f60329
30 · June 26, 2014 at 10:46 PM

I can see how coconut oil might be cited as too "unique" to use because of the mcfa's, although beyond a slight thermic effect I doubt it differs much from other sat fats. I figured butter was out since it's dairy, but I feel like that would be worth using as well. Tallow is probably the best option for "true paleo" credibility, but its pretty hard to get and not the tastiest thing. So in the end, using fatty meats is probably the best option to bridge the gap between theory and practice.

72cf727474b8bf815fdc505e58cadfea
2626 · June 26, 2014 at 6:11 PM

We're still in the design phase, but hopefully we'll be able to share them later. We will of course pre-register our hypotheses and have them publicly available when the study begins, if not sooner.

Cb9a270955e2c277a02c4a4b5dad10b5
10994 · June 25, 2014 at 11:54 PM

Do you have the formal Background/Objective/Design/Methods for this study?

72cf727474b8bf815fdc505e58cadfea
2626 · June 25, 2014 at 8:50 PM

I've made some small edits to make this clearer.

72cf727474b8bf815fdc505e58cadfea
2626 · June 25, 2014 at 8:44 PM

Diet studies are always tricky because you have to balance theoretical perfection with what your participants will be able to assimilate and practically stick to. I'll be very interested to see what kind of challenges participants report and what they do end up eating.

72cf727474b8bf815fdc505e58cadfea
2626 · June 25, 2014 at 8:39 PM

There are definitely well-defined, a priori outcome measures. I've left out a lot those details because I'm not a lead member on this project and don't want to be perceived as speaking for the study team. That's why my question was framed so broadly even though, as you point out, it means you can't give a detailed critique of our protocol.

72cf727474b8bf815fdc505e58cadfea
2626 · June 25, 2014 at 8:24 PM

Those aren't the guidelines we're giving participants, they're just how I was summarizing it here, for people who are already familiar with these ideas. Participants will receive detailed nutrition counseling. My point is that we're trying to stick to a few simple principles that won't overwhelm them, even if it allows some imperfection.

Cf08ad26759fdd206a2c9f9385080a57
1005 · June 25, 2014 at 7:44 PM

Sweet. It'll be tricky taking people who don't eat for the nutrition and seeing what they initially do for a few months shopping the sides of the grocery store. (eg. you could eat nut butters & almond flour pancakes every morning, grazing on high-GI foods & chocolate and paleo recreations of junk foods with confined animal farmed meats cooked in organic vegetable oils and think you're eating 'real food' if you didn't know better. I'd prefer to see lbs of veg / herbs with wild seafood / grass-fed offal and bone broths planned into optimally nutrient-dense meals with intermittent fasting.)

72cf727474b8bf815fdc505e58cadfea
2626 · June 25, 2014 at 6:56 PM

We'll definitely get some food logs. There's no way there will be enough participants to split up the paleo group based on specific food choices and so forth, but at least we'll be able to tell whether people are actually following the diet.

32f5749fa6cf7adbeb0b0b031ba82b46
41312 · June 25, 2014 at 11:40 AM

Supplying the fat of choice, that's interesting… coconut for paleo, canola for conventional, olive oil for Mediterranean. You're right, paleo in trials always is the most restrictive in nature, which is big reason why it works. Would be interesting to see if a less restrictive paleo, with fatty meats is as effective.

32f5749fa6cf7adbeb0b0b031ba82b46
41312 · June 25, 2014 at 11:37 AM

Self-reported food logs? Not sure how that would be helpful. The value in the study would be results over many people, where the average meals are what matters.

Total Views
645

Recent Activity
72cf727474b8bf815fdc505e58cadfea

Last Activity
88D AGO

Followers
6

Get Free Paleo Recipes Instantly

6 Answers

Bfa1c9eacfc94a1b62f3a39b574480c6
0
3690 · June 26, 2014 at 10:55 PM

Flaws:

Observational studies (correlation does not mean causation, good for generating ideas, not proving causality).

Food frequency diaries (most individuals have poor recall). Even daily food logs can vary unless one is weighing every gram of food they put into their mouths.

"Eat real food." Scientifically doesn't mean much. Breakdown should be highly standardized %s of lipids, carbohydrates, and protein. Sourced from real food, not concoctions or "chow" in any sort of manner.

Ideal experiment: Double blind, double cross-over randomized trials with standardized meals in a metabolic ward. Even then, causality is difficult to establish, but such is the nature of science.

32f5749fa6cf7adbeb0b0b031ba82b46
41312 · June 27, 2014 at 1:06 AM

Talk about completely unrealistic expectations for a dietary study.

Medium avatar
0
583 · June 26, 2014 at 10:02 PM

Since most people serious about paleo take it beyond "don't be scared of fat" I'd suggest being a little more straight forward: eat lots of fat.

Be specific: no industrially processed seed oils, lots of olive, coconut, and animal fats, a little from nuts and avocados...

I don't know the constraints of the study but if it's just another self-reporting set up, then it fails where so many other studies have. At the very least, participants should be asked to try and log what they eat in a daily food journal. If you're just asking at the end of the month "so, did you eat like we said to?" then you'll get quite a lot of reporting bias.

I'm assuming also that your dietary counseling will get reasonably specific without overwhelming or being too restrictive. They should be eating plenty of fresh veggies, for instance, and at least a bit of fruit and tubers. They should be told to avoid mainstream coffee creamers, opting for coconut cream or heavy whipping cream and MAYBE a dash of honey.

You should print out a cheat sheet or simplified infographic to guid daily eating. I personally like the PHD one:

http://perfecthealthdiet.com/wp/wp-content/uploads...

72cf727474b8bf815fdc505e58cadfea
2626 · June 27, 2014 at 6:22 PM

Thanks for the suggestions. Yes, we're getting a number of daily food logs. Based on our resources, it's simply impossible to get anything more precise at this stage.

Cb9a270955e2c277a02c4a4b5dad10b5
0
10994 · June 25, 2014 at 7:59 PM

1) Eat real food - This statement is meaningless.

2) No dairy, grains, or legumes. Okay.. So a dairy free, grain free, legume free diet then is what you're studying.

3) Don't be scared of fat. This is incredibly vague. What does this even mean? I'm not scared of fat and I don't eat very much fat. If I'm not 'scared' of fat then does that mean I'd indulge in it?

Honestly without more specifics as to what is and isn't allowed, what specific macros you'll be testing for, types of fat, etc I don't think you're going to have anything more than a jumbled mess of (probably small sample sized) data at the end of this study.

If you aren't very specific in what parameters you'll be changing and what parameters you'll be measuring then your study is going to be a pile of garbage when you finish. Maybe these parameters are already defined, if so please elaborate on them in your question so I can modify my answer/recommendation.

Seriously I'm not trying to be harsh, but I've read a ton of medical journals and am a high ranking member here on PH, this is my honest opinion.

72cf727474b8bf815fdc505e58cadfea
2626 · June 25, 2014 at 8:39 PM

There are definitely well-defined, a priori outcome measures. I've left out a lot those details because I'm not a lead member on this project and don't want to be perceived as speaking for the study team. That's why my question was framed so broadly even though, as you point out, it means you can't give a detailed critique of our protocol.

72cf727474b8bf815fdc505e58cadfea
2626 · June 25, 2014 at 8:24 PM

Those aren't the guidelines we're giving participants, they're just how I was summarizing it here, for people who are already familiar with these ideas. Participants will receive detailed nutrition counseling. My point is that we're trying to stick to a few simple principles that won't overwhelm them, even if it allows some imperfection.

Medium avatar
0
10126 · June 25, 2014 at 1:52 PM

I'm glad you didn't overstructure the paleo part. It's voluntary compliance, and requiring grass fed liver and kale reduces compliance. But be as specific as you can be in what you want to accomplish. If you don't have a specific diet for your control group then I suggest letting them carry on with whatever they're doing. That way it becomes paleo vs whatever made them sick and not paleo vs med. Since the purpose is not weight loss I'd suggest letting both groups eat and exercise ad libitum. Above all have specific markers to measure against, such as systolic blood pressure, HDL, A1C, fasting blood glucose, etc. The more the study is shotgunned the less it will show.

And since most paleos eat butter you should allow it.

Cf08ad26759fdd206a2c9f9385080a57
0
1005 · June 25, 2014 at 6:19 AM

It would be nice to see a study on health metrics modulated by food choices actually study the food choices. Any chance you can publish the logs of what the participants eat, when they eat it, and how this affects their health metrics? (food weights + micro/macro nutrient averages w/ timestamps.) This would be much more useful than seeing the results that follow from the advice to 'eat natural food and avoid grass-fed ghee and raw dairy / fermented low-glycemic grains / soaked and properly prepared legumes' when told to chronically ill people on a suboptimal diet and followed for a few months.

Thumbnail avatar
10126 · June 27, 2014 at 6:56 AM

@AxialGentleman, once a year I fill out a survey form for the NWCR, which is run by Renata Wing of Brown University. The objective is to determine how weight loss is maintained after dieting, and it takes about half an hour to fill out. The food portion breaks down into servings per day on average, for the year. But many other things are covered, from exercise to meal frequency to sense of well-being. It's a population study with several thousand participants surveyed over decades. Keep your study groups as large as is practical.

72cf727474b8bf815fdc505e58cadfea
2626 · June 25, 2014 at 6:56 PM

We'll definitely get some food logs. There's no way there will be enough participants to split up the paleo group based on specific food choices and so forth, but at least we'll be able to tell whether people are actually following the diet.

32f5749fa6cf7adbeb0b0b031ba82b46
41312 · June 25, 2014 at 11:37 AM

Self-reported food logs? Not sure how that would be helpful. The value in the study would be results over many people, where the average meals are what matters.

0ba891d22837788c4d5ccf3f33f60329
0
30 · June 25, 2014 at 5:50 AM

I think a lot of published paleo diet trials have utilized more restrictive diets than seems to be the norm. They favor lean meats for example, which seems to be quite contra to what you'll hear most practitioners advocate. In many of the studies saturated fat intake was quite low, often lower than the control diet. Given the common belief that SF is unhealthy I think helping to control for a significant difference in this respect would be good.

Perhaps allowing, maybe even supplying a cooking fat like coconut oil could help with this, I don't know. I think something like that might also help with dietary adherence as well.

Anyway, best of luck to ya.

32f5749fa6cf7adbeb0b0b031ba82b46
41312 · June 25, 2014 at 11:40 AM

Supplying the fat of choice, that's interesting… coconut for paleo, canola for conventional, olive oil for Mediterranean. You're right, paleo in trials always is the most restrictive in nature, which is big reason why it works. Would be interesting to see if a less restrictive paleo, with fatty meats is as effective.

Answer Question

Login to Your PaleoHacks Account