Note the methodology:
This study is a retrospective analysis of the records of patients who attended a physician-supervised 10-day residential program (the McDougall Program) from 2002 to 2011. This ongoing program used Internet marketing to attract patients from a wide geographic area to spend 10 days at a hotel in Santa Rosa, California, where they received dietary counseling and were fed a low-fat (~10% of calories) diet based on minimally refined plant foods ad libitum to satiety. The educational staff included a medical doctor, a registered dietitian, a psychologist, exercise coaches, and cooking instructors. Patients were also given opportunities for light to moderate exercise. No stress reduction techniques or meditations were included in the program.
If you want to test if starch reduction itself increases health, you can't change any of the other factors. You would have to take randomly selected people and have them stop eating starch (or whatever) for 10 days (although the title of the study says 7 days). The methodology above is completely invalid because:
The participants were "locked" in a hotel, and on "vacation" (ie, away from home). This means they didn't do what they would normally do. Additionally, "patients were also given opportunities [...] for exercise", but it doesn't mention what other types of entertainment were available. I think this study just bored people into exercising. However, regardless of whether that's true, this study shows nothing about what would happen if people living their everyday lives would remove starch for 10 days.
The results aren't particularly impressive either: some weight loss, and an actual increase in triglycerides.
The authors openly admit the study is scientifically invalid and note:
Uncontrolled studies like this one can provide valuable evidence in situations where randomized, blinded clinical trials have not been or cannot be performed
This sounds more like wishful thinking than science, although the word "can" mitigates the statement slightly.
Realistically, however, this study is so insanely different from a true scientific experiment, it should be considered garbage and its results completely worthless, which the author almost admits:
this study has some significant limitations. One concern is that the moderate exercise allowed during the program could have contributed to the observed weight loss. Another concern is the timeframe, which was too short to show whether the improvements in biomarkers would be sustained or would ultimately translate into decreases in morbidity and mortality
I believe this is the most important statement in the study and should be printed in large bold-face type at the top of the study (and every similar study). It's shorthand for "this study is bullshit, but I need to publish papers to keep my job/image".
Interestingly, the author doesn't include the other reasons the study is grossly invalid.
Perhaps the biggest limitation of the vegan diet studies comes from the selection of the participants. Subjects who choose to participate in these studies are willing to try a low-fat vegan diet to improve their health and/or appearance. Thus, they may be more likely than the general population to adhere to the diet.
This is far from the biggest limitation of the study, and attempts to imply that people who adhere to the diet will succeed. The biggest limitation is that the study is completely unscientific, and it doesn't matter if people adhere to the diet or not!
Epidemiologic [sic] evidence, reinforced by clinical and laboratory studies, shows that the rich Western diet is the major underlying cause of death and disability (e.g, from cardiovascular disease and type 2 diabetes) in Western industrialized societies.
People say this a lot, but the study doesn't provide sources. The statement isn't actually part of the study, but is misleading because it attempts to add unsourced credence to the study.
Retrospective analysis of measurements [...]
Given any set of data, you can always find coincidental anomalies (statistics almost guarantees that). The fact is that the researcher may have pored through dozens of studies to find one that matches his conclusions. That's why you do the study after you decide what you're actually testing.
The primary goal of health care should be to decrease all-cause morbidity and mortality.
It's not actually relevant, but I disagree with this statement. The primary goal of medicine should be to let us live our lives the way we want, not dictate how we should live our lives.
One quibble is that more and more research is showing that certain fats are not only necessary but can helps stave off issues with the nerves and brain.
I also think that the selection of vegan items is important. You can be very vegan and eat a ton of refined carbohydrates. Glycemic load is important in food selection.
It is compelling as heck. McDougall is part of a growing cadre of physicians who espouse plant-based nutrition. That diet has been conclusively shown to be effective against many of the medical ills afflicting modern society: obesity, heart disease (the leading cause of death in the US for men and women), strokes, cancer, type-2 diabetes, rheumatoid arthritis, other auto-immune disorders, etc. The range of diseases is remarkable. Food can be medicine, and of course it is inexpensive, without side effects, and pleasurable.
McDougall emphasizes starches (complex carbohydrates) in a plant-based diet to make sure there are ample calories. He has many exciting case histories on his website.