Micronutrients for Anxiety and Depression

In one of the better-designed trials covered in this episode of Sigma Nutrition Radio, the gap between the group taking a high-dose multivitamin and the group taking nothing extra came down to very little: a capsule, since the weekly text reminding people to fill in an online form was identical across both groups. Everything else was identical, the baseline interview, the end-of-trial interview, and oversight from the same clinical psychologist for both groups (13:23). That sameness is not an accident. It is the entire point of the trial, and it is the reason host Danny Lennon and nutrition researcher Alan Flanagan hold it up as a rare example of a nutrition study built to answer the question it actually asks.

The episode starts from a listener question. A Sigma Nutrition Premium subscriber named Rachel asked about the work of Professor Julia Rucklidge, whose lab studies high-dose micronutrient supplements, meaning broad-spectrum combinations of vitamins and minerals, for conditions like anxiety, depression, and irritability (00:15). Rucklidge had appeared on the podcast roughly a decade earlier, and her group has since published a string of new trials (03:45). Lennon and Flanagan pick three: the NOMAD-RCT, in medication-free adults with anxiety and depression; NutriMum, in pregnant women with depressive symptoms; and BEAM, in adolescents with irritability and emotional dysregulation (11:30). Before getting to results, they spend the first half of the conversation on something less glamorous and more important: how anyone would know if any of these trials worked.

The measurement problem

Flanagan's starting complaint is that the field of nutrition and mental health is crowded with headlines and thin on evidence, a pattern he and Lennon had already traced through the widely cited SMILES trial, an earlier diet-and-depression study referenced here as a cautionary example (02:58). Part of the problem is definitional. Depression or anxiety sound like fixed targets, but researchers measure them with a long list of different instruments, each with its own strengths and blind spots.

"This is an area with a dizzying array of different instruments that have been developed to nominally capture what we are interested in." (Alan Flanagan, 05:41)

Flanagan walks through why this matters. Scales differ in sensitivity, meaning how well they catch true cases, and specificity, meaning how well they rule out false ones. They also differ in how closely they track formal diagnostic criteria and in whether a clinician scores them or the patient scores themselves (06:54). A trial that looks positive on one scale can look flat on another, not because the treatment changed but because the ruler changed.

The second problem is structural, and it goes back to why randomized trials exist in the first place: to isolate an independent effect of one intervention, stripped of everything else that might be happening around it (07:24). Pharmacological trials are comparatively simple. Behavioral and nutritional trials are not. Flanagan points out that in many of these studies, the group getting the active intervention also gets more attention, more monitoring, sometimes informal counseling, while the control group is left with vague usual care.

"Imbalance between an intervention group and a control group... introduces enormous bias towards the intervention group." (Alan Flanagan, 08:04)

That imbalance means a positive result can reflect extra human contact rather than anything in a supplement or a diet plan. It is a bias built into the design, not something that shows up later as a footnote in the discussion section.

The third issue is who gets enrolled in the first place. Flanagan notes that drug trials for depression often require scores above a fairly high cutoff on instruments like the Hamilton Rating Scale of Depression before someone qualifies, which means the sample is clinically severe (08:54). Behavioral and nutrition trials frequently set a lower bar, enrolling people with mild or non-clinical symptoms, or a mix of severities in the same study. That choice has consequences: interventions tend to show larger, more convincing effects in people who are more severely affected, and smaller, murkier effects in milder or mixed samples (09:29). So two trials can test the same supplement and reach different conclusions for a reason that has nothing to do with the supplement itself: who walked in the door.

What the NOMAD trial got right

This is the checklist Lennon and Flanagan apply to the NOMAD-RCT, published in 2023, which tested a broad-spectrum micronutrient formula against what the researchers call functionally impairing anxiety and depressive symptoms in adults (12:00, 12:23). The sample was deliberately community-based rather than clinically severe: medication-free adults recruited outside a psychiatric setting (12:35, 12:52). On its own, that would predict the smaller-effect pattern Flanagan already flagged. But the design compensates for the other major source of bias. Both the supplement group and the control group were monitored by the same clinical psychologist, and contact was kept to a bare minimum on purpose: a baseline assessment, an end-of-trial interview, and weekly text reminders to complete online questionnaires (13:23, 13:45). Neither group received counseling or formal therapy. The contact that did happen was administrative, not therapeutic, and it was identical across both arms.

Flanagan calls this a genuine strength, because it means the trial was built to ask one specific question: does the micronutrient formula itself do anything, separate from the extra support, counseling, or dietary coaching that complicates so many other nutrition-and-mental-health studies (14:05)? That is a narrower and more honest question than the one headlines usually imply, and it is the reason the trial's design, not yet its results, earns the attention it gets in this episode.

The outcomes themselves were measured with two tools: the PHQ-9 and the GAD-7 (14:35). The PHQ-9, or Patient Health Questionnaire Nine, is a nine-item, self-reported scale used both to screen for depression and to track its severity over time (14:50). Flanagan singles it out as one of the stronger instruments available because it has high sensitivity and specificity for detecting major depressive disorder, meaning it tends to correctly flag people who have the condition and correctly clear people who do not (15:06).

This preview portion of the episode stops short of walking through the NOMAD trial's actual results, saving that for the full premium episode. What it delivers instead is a working method for reading any trial in this space: ask what instrument measured the outcome and who administered it, ask whether the control group was matched for contact and attention, and ask how severe the enrolled sample was to begin with. Apply those three questions to the SMILES trial, and its halo starts to look thinner. Apply them to NOMAD, and the design holds up noticeably better, even before anyone looks at a single number in the results table.

That is a useful discipline for anyone who reads a diet-and-mood headline and wants to know whether to believe it. The claim that a food or a supplement changes how people feel is easy to make and hard to test well. The way to tell the difference is not to argue about nutrients in the abstract, but to ask the duller, more specific questions this episode keeps returning to: what was measured, who was compared to whom, and how much contact did each group actually get. A capsule and a weekly text message, it turns out, can be the most rigorous thing in the room.