More null results for you to love today: Researchers found that in a community sample of Japanese adults, diet and depressive symptoms were not related. There has been a lot of suggestion in the literature recently that depression is related to many factors that, traditionally, have not been thought of as being relevant to mental disorders. For example, in the past few decades, depression has been repeatedly linked to inflammation and other medical diseases such as cardiovascular disease, osteoporosis, and diabetes. Because diet obviously plays a major role in overall physical health, diet has recently been put under the microscope in the search for risk factors of depression.
Although results from one single study are not conclusive, it is good to see this paper published showing that, at least in one type of sample, diet and depression don’t appear to be related. And I think it’s fantastic that the journal, Annals of General Psychiatry (which also happens to be an open-access journal), published these null results.
The issues I have with the paper are, firstly, that their measure of diet was a self-report brief questionnaire, and although that is a good start, I don’t think we can draw any conclusions without doing another study with a much more comprehensive and objective measure of diet. Secondly, and I think more importantly, depressive symptoms were also measured as a self-report questionnaire, and if participants had enough symptoms (above what is called a “clinical cut off”, which is the score at which other research says people are likely to be clinically depressed), they were considered to be depressed. The problem with this is that clinical depression cannot really be diagnosed by adding up a number of items on a questionnaire. Otherwise we could all diagnose ourselves. Instead, the diagnosis must be made by a qualified healthcare professional if treatment is required, or in the case of data for research, by a trained interviewer. For example, depressive symptoms can be an indicator of many other things, including other mental disorders such as PTSD or Bipolar Disorder, or even medical disorders such as hyperthyroidism. Or the wording of some questionnaire items can be confusing and people may endorse them when they are not experiencing them (or vice versa).
If the point of the study was to compare types of diet between people who were depressed vs. those who weren’t depressed, an interview should have been used. If the point was to see if levels of depressive symptoms were correlated with diet, then a continuum of depressive symptoms should have been correlated with another continuum, perhaps a score of how “healthy” the diet was considered to be. I understand that in a large study with lots of participants, the former option of conducting interviews is difficult (there were 791 adults in this study). But I also think that there are enough methodological concerns in this study that we cannot yet make any definitive conclusion that diet isn’t related to depression. To give the authors credit, they also discuss these limitations and suggest that we can’t “rule out” this relationship yet.