How Well Does the Shangri-La Diet Work? (part 2)

The Post Your Tracking Data Here section of the SLD forums now contains lots of data. In addition to the weight-vs.time graphs on the home page and in the forums, I have now analyzed this data in other ways. The graphs below (based on data up to November 2) show how the rate of weight loss varies with (a) time on the diet and (b) weight.

For each person reporting weights, I computed a rate of weight loss for every interval in their data. For example, if someone reported her weight at three different times, then there are two intervals: from Time 1 to Time 2, and from Time 2 to Time 3. For each interval I computed a rate of weight change. The scatterplots below are based on 820 intervals. Each point is a different interval.

The first graph shows how the rate of weight change varied with how long you have been doing the diet.

This shows that average weight loss slowed down from about 2.2 pounds/week to about 1 pound/week during the first few weeks and didn’t change much after that.

Another obvious factor that might affect weight-loss rate is weight: Perhaps people who weigh more lose faster. Because rate of weight loss changes during the first few weeks, I looked at this question two ways: using only data for Week 1 on the diet (early loss); and using only data after 4 weeks on the diet (later loss).


The top graph (early loss) shows that during Week 1, your weight has a big effect on your rate of weight loss. People who weigh more lose faster. The bottom graph (later loss) shows that after 4 weeks, your weight has much less effect on how fast you lose.

My explanation: During the first week or so of SLD, most of the weight loss is not fat or water but the food in your digestive system. Because the diet has reduced your appetite, you are eating less each day. But the speed (inches/day) at which food travels through your digestive system does not change; so relatively full digestive system is slowly replaced by a relatively empty one. After this replacement — which takes about one week — is complete, further weight loss is all due to loss of fat. You comfortably lose fat at the rate at which your set point goes down. The long-term rate of weight loss is about 1 pound/week because the set point is going down about 1 pound/week.

Data analyses like these have never been published for any weight-loss method. Not that they’re sophisticated or clever or surprising — they’re not. Given (a) the amount of damage caused by obesity and (b) the amount of money spent on health research (2006 NIH budget: $28 billion), it’s quite a gap. Possibly related to the misguidedness I discussed last week.