• What the reanalysis uncovered — In this study, the team reviewed the raw trial data — not just published outcomes — and reaudited everything using the original study protocol. They applied the RIAT model (Restoring Invisible and Abandoned Trials) to see what the results would look like if all the data were handled transparently and analyzed according to the original protocol.

What they found turned the original TADS conclusion on its head — the team found "little support to justify the routine use of fluoxetine as treatment for adolescent depression."⁵ On top of that, over two-thirds of all serious adverse events (22 of 32) happened in kids who took fluoxetine.

• Fluoxetine's benefits were inconsistent — The reanalysis found that fluoxetine's advantage over placebo depended on how it was measured. Specifically, on the main outcome measure — the Children's Depression Rating Scale-Revised (CDRS-R) — the difference between fluoxetine and placebo was statistically insignificant.

Fluoxetine did produce more treatment responders than placebo on the second main measure, the Clinical Global Impression-Improvement scale (64% vs. 39%), but it beat placebo on only two of 11 secondary measures — both rated by clinicians who weren't blinded.

What's worse, the original study authors had emphasized suicidal thoughts and behaviors as rare. But the reanalysis team identified 11 additional suicide-related events that weren't reported in the official TADS reporting.

They also documented 10 suicide attempts in 10 participants — nine of whom were taking fluoxetine — compared with the seven attempts reported in the original paper. There were no completed suicides. That's a major correction to the record and shifts the risk-benefit picture.

• Adverse events were abundant — The reanalysis noted that 369 adverse events occurred for 171 participants, and 66% of those occurred in participants taking fluoxetine. These weren't limited to suicide-related episodes.

Some events appeared in the trial's records but not in the published papers, including two hospitalizations that should have been classified as serious adverse events, and three suicide attempts that were logged in the study's attrition-prevention records but not in its adverse event database. The original reports also left out events that occurred in fewer than 2% of participants and adverse events in participants whose treatment was ended early.

• Even the combination of CBT with fluoxetine was misleading — While it did show the strongest symptom drop, the reanalysis points out a problem — those therapy sessions weren't blinded. Therapists and participants both knew who was receiving treatment.

That opens the door for expectancy bias, meaning they expected themselves to improve, so they reported feeling better, regardless of whether the drug helped. That calls into question how much of that symptom improvement came from the medication itself.

• Statistical shortcuts flattered the results — The reanalysis also pointed to flaws in the original statistical analysis. In the original trial, multiple comparisons were made between groups without correcting for how that inflates the chance of false positives.

Essentially, the more statistical tests you run, the higher the odds you'll find something significant just by chance. The RIAT team corrected for this. After the correction, fluoxetine alone still didn't outperform placebo on the main depression scale, while its edge on the responder measure held.

• Possible reason why side effects occur — The reanalysis didn't examine how fluoxetine causes side effects, but the theory behind selective serotonin reuptake inhibitors (SSRIs) raises questions of its own.

SSRIs are thought to work by increasing serotonin levels in the brain, which is linked to mood. But that theory doesn't explain why some teens on fluoxetine experience agitation or insomnia, both of which were noted in the early TADS reports.

• Complete disclosure is required for the public's safety — The study called for full access to clinical trial data moving forward, and concluded that fluoxetine's "inconsistent benefits do not outweigh the harms identified in this study" and that the increase in harms "warrants further circumspection when prescribing it to adolescents." As summarized by the team's closing words:⁶

"Our reanalysis confirms the original reported findings that superiority over placebo was not demonstrated for fluoxetine. Contrary to original TADS Team's reporting, we have uncovered a higher, clinically significant level of harm, including 11 additional suicide-related adverse events."

• The side effects of fluoxetine are downplayed — While previous reports have mostly emphasized fluoxetine's effects on depressive symptoms, the BMJ Open team turned the spotlight to what was missing. For example, well-known SSRI side effects such as weight and appetite changes and sexual problems were never mentioned in TADS-related publications, and others, such as insomnia, dropped out of the reports after the first 12 weeks.

Only suicidal behavior was consistently tracked through the 36-week period, leaving out a wide array of common and distressing adverse experiences that could have shaped treatment decisions for families and clinicians.

• A misleading classification was flagged — Nine cases of suicidal behavior were reported in the placebo group even though those teens were taking fluoxetine at the time and the placebo period had already ended.

As the reviewers explain, the risk of suicidal behavior won't appear higher for fluoxetine than for placebo if teens taking fluoxetine are counted in the placebo group, which made the drug's safety profile look cleaner than it was.

• Another issue was the reporting threshold itself — The researchers noted that the TADS team only counted events that caused clinically significant interference with functioning, required medical attention or led to a new medication. Milder events weren't counted as adverse events:⁸

"In the TADS, adverse events were defined as an unfavourable medical change that occurred after beginning or during the study that might or might not be related to or caused by the study drug or CBT treatment.

This was further specified as any medical event that caused clinically significant interference with functioning (e.g., headache that caused school absence or otherwise caused clinically significant activity restriction), any event that required medical attention, and any medical event associated with impairment in functioning and induced the patient to take a concomitant medication.

Conditions that did not lead to clinically significant interference with functioning or did not require medical attention were not defined as adverse events."

• The threshold kept milder problems off the record — The reviewers couldn't find published versions of the questionnaires used to collect side effect data, so it's unclear which symptoms were specifically asked about. They also noted that the threshold likely reduced the number of side effects reported:⁹

"The TADS protocol included a threshold limit on what would be considered an adverse event, specifying that the event must cause clinically significant interference with functioning, require medical attention or cause a need to take medication. As an example, emerging mania was not recorded unless symptoms exceeded this threshold.

It must be assumed that this reduced the number of reported adverse effects, which may not have been severe enough to reduce daily functioning or cause a need for additional treatment."

• There's also the issue of long-term follow-up — Even though the TADS trial lasted for 36 weeks, most publications only focused on the 12-week acute treatment phase. After that, side effect reporting all but disappeared:¹⁰

"Due to its long duration (36 weeks) and follow-up (1 year), the TADS could have provided valuable information on the long-term occurrence of adverse effects both in frequency and severity. The adverse effects profile of FLX [fluoxetine] in the TADS has only been reported in detail for stage 1, where approximately 200 patients received FLX for 12 weeks."

1. Strengthen your gut health — Improving your digestive health is a foundational step toward better mental well-being due to the gut-brain axis. To nurture this relationship further, incorporating homemade probiotic-rich foods and steering clear of ultraprocessed foods high in linoleic acid (LA) may support communication between your gut and brain.

Excess LA may harm your gut health. I recommend keeping your intake between 2 and 5 grams a day from all sources — LA is still an essential fat, so don't aim for zero. To help you monitor your intake, download the upcoming PAXapp. It has a feature called the Seed Oil Sleuth, which calculates the LA in your food to a tenth of a gram.

In addition, limiting processed foods and reducing exposure to antibiotics and agricultural chemicals can further protect your gut microbiome. A nutrient-rich diet fuels cellular energy and supports metabolic functions — both of which may help support a stable mood.

2. Embrace a lifestyle that supports mental resilience — Support your mental health through regular exercise and quality sleep. Physical activity enhances mitochondrial health and boosts natural energy levels. Even simple habits like walking about an hour each day and keeping a consistent sleep routine can help lift your mood.

3. Try non-pharmaceutical methods — Examples include counseling, talk therapy, and cognitive-behavioral therapy (CBT). These methods empower you to develop coping strategies, confront root causes of distress, and build mental resilience without depending on medications.

Another promising option is the Emotional Freedom Techniques (EFT), a form of psychological acupressure that some small studies suggest may help ease anxiety and depressive symptoms.¹¹

4. Prioritize healthy vitamin D levels — Ensuring your body has adequate vitamin D through safe sun exposure may support your mood. As noted in a previous article, research has linked low vitamin D to a higher risk of depression. Thus, increasing your levels naturally by spending time outdoors may help support emotional well-being.

However, it's not advisable to get out in the sun right away, especially if you've been eating a diet high in LA. When sunlight hits your skin, the LA embedded there may oxidize and promote inflammation.

To protect your health, it's advisable to avoid midday sunlight, especially between 11 a.m. and 3 p.m. for about four to six months after you cut out seed oils, then gradually build up exposure around solar noon. To hasten the process, I recommend getting C15:0 (pentadecanoic acid) from full-fat grass fed dairy.

Most people get only about 100 to 200 milligrams of C15:0 per day. I personally take 2 grams daily. At that level, research suggests your keratinocytes — the skin cells in your outermost layer — may start incorporating C15:0 in place of LA. For an in-depth guide on this process, read "The Fast-Track Path to Clearing Vegetable Oils from Your Skin."

5. Foster honest dialogue with the youth — When supporting a young person's mental health, creating a safe, non-judgmental space for open communication is important. Encouraging them to express their emotions and concerns helps them feel heard and understood. This approach also aids in detecting any early signs of distress, which can lead to timely, effective support or intervention.

6. Explore these other natural strategies — For those seeking natural relief from stress and anxiety, gamma-aminobutyric acid (GABA) supplements may be one option to explore. GABA is your brain's main calming neurotransmitter, and a systematic review of placebo-controlled human trials found limited evidence that oral GABA helps with stress.¹² Talk with your health care provider before trying it, especially if you take medication for anxiety, sleep, or mood.

Methylene blue supports mitochondrial energy production, and in one small 3-week trial in adults with severe depression, it outperformed placebo.¹³ It should only be used in pharmaceutical-grade form prescribed by a health care professional, and never alongside antidepressants that act on serotonin, since case reports have linked the combination to serotonin syndrome.¹⁴