A randomized controlled trial shows the right diet can improve depression.
For the past seven years, I’ve been writing articles about food and mood, exploring how different diets and types of foods might create or help mental health problems. Despite a great amount of interest in this topic spawning dozens of popular books from Grain Brain to Eat Complete, the data we had was always limited. The vast majority of studies about food and mental health are observational, meaning some form of asking people what they eat and then tracking mental health variables.
These data are limited by different types of bias, food studies being particularly prone to bias, as they are often done with “food frequency questionnaires” or FFQs asking about how many hamburgers or vegetables you eat. Even one of the most validated FFQs in the world, the one used for the Nurses’ Health study, has severe limitations. People who lied more about hamburger intake were healthier than those who didn’t, for example. With observational data, you usually find that healthy people who care about their health and listen to health messaging are healthier. Large observational studies are mostly interesting if they have findings that are opposite what is expected (for example, coffee drinkers, despite smoking more and drinking more alcohol, score higher on several measures of good health).
The real meat of science is in the randomized controlled trial. That means taking two groups of people, putting one through an experiment and one through a control, and seeing if there is a difference in outcomes between the groups. When it comes to mental illness, we did have some data for the use of randomized controlled trials of certain diets and some mood outcomes. All of these studies had depression as one of the measured endpoints, but none of these were in a group of depressed people trying different diets to feel better. The measures of depression were just collected along the way of a trial looking at something else (such as heart disease).
In these trials, changing diet to various options (such as Mediterranean or lower cholesterol) didn’t worsen symptoms of depressed mood, but only diet trials that didn’t restrict red meat or weren’t described as “low cholesterol” diets were effective in lowering measures of depression by the end of the study. Another randomized therapy trial for early depression in the elderly used a nutritional instruction arm as the control (thinking that food instruction was neutral and wouldn’t help mental health), finding it equal to a type of community-based psychotherapy in preventing worsening of depression.
This year, finally, we have the SMILES trial, the very first dietary trial to look specifically at a dietary treatment in a depressed population in a mental health setting. Participants met criteria for depression and many were already being treated with standard therapy, meds, or both. The designers of this trial took the preponderance of observational and controlled data we already have for general and mental health and decided to train people using dietary advice, nutritional counseling, and motivational interviewing directed at eating a “modified Mediterranean diet” that combined the Australian Dietary Guidelines and the Dietary Guidelines for Adults in Greece. They recommended eating whole grains, vegetables, fruit, legumes, unsweetened dairy, raw nuts, fish, chicken, eggs, red meat (up to three servings per week), and olive oil. Everyone in the study met criteria for a depressive disorder.
The experimental arm of subjects were instructed to reduce the intake of sweets, refined cereals, fried food, fast food, processed meat, sugary drinks, and any alcohol beyond 1-2 glasses of wine with meals. There were seven hour long nutritional counseling sessions and a sample “food hamper” with some food and recipes. The control group had the same number of sessions in “social support,” which is a type of supportive therapy that is meant to mimic the time and interpersonal engagement of the experimental group without utilizing psychotherapeutic techniques.
A few more interesting tidbits about the design and the study…they randomized a group of people found to be depressed and to eat more like the average Australian (meaning fast food, refined carbohydrates, etc) and excluded people who already ate very healthfully. The researchers also looked at the cost, as many folks have concerns that eating healthy is more expensive and out of reach of many people, and found that the average experimental participant spent $138 on food per week before the trial and $112 per week during the trial. A major limitation to this study is the size: just 33 in the experimental arm and 34 in the control group. The trial was designed for a larger number.
Despite the small size, the results were still statistically significant and better than anticipated. The dietary group had bigger reductions in depression scores at the end of 12 weeks. Remission of depression symptoms occurred in 32.3 percent of the diet group as opposed to 8 percent of the control group. That means that the “NNT” (or number needed to treat) for this study was 4.1, which is similar to and even better than many trials of chemical antidepressants.