Four studies, four claims

The Metropolitan Wellness Assessment was conducted between March and May 2021 with support from a state department of health prevention grant. Researchers mailed a twenty-eight item questionnaire to a randomly selected sample of 6,000 households in one metropolitan county; 4,200 adults returned a completed form, a response rate of 70%. Respondents were 54% female, with a median age of 44, and the sample was broadly similar to the county on age and sex, though renters and adults under thirty were somewhat underrepresented.

Everyone who responded answered the survey on a single occasion. Among the twenty-eight items, one asked how many days in the past week the person had exercised for at least thirty minutes, and another asked them to rate how often they had felt down, depressed or hopeless in the past two weeks. Both answers came from the same form, filled out at the same sitting. Of those who reported exercising four or more days a week, 11% reported frequent depressive symptoms. Of those who reported exercising once a week or less, 27% did. The difference was statistically significant. The authors note in their limitations section that the survey relied on self-report and that the findings may not generalize to rural counties.

The Northern Communities Health Study began enrollment in 2012 and was funded by a national research institute. Investigators enrolled 9,800 adults aged 25 to 64 across eleven counties, none of whom had ever been diagnosed with depression at the time they joined. Participants completed a baseline visit that included a physical exam, a household income questionnaire and a residential history.

At enrollment, each person's home address was recorded and the amount of park and green space within half a mile was calculated from municipal land-use records. The investigators then contacted every participant once a year for the next eight years; 8,430 participants, or 86%, completed at least six of the eight annual contacts. Over that period, 1,140 participants received a first diagnosis of depression. Participants who had lived in the greenest neighborhoods at enrollment developed depression at a lower rate than those who had lived in the least green neighborhoods. The gap held after the investigators accounted for income and age. The authors note that participants who moved during follow-up were retained under their original classification.

This analysis was carried out by a research team at a four-hospital system and reviewed by the system's institutional review board. Using admission records from 2019 through 2022, the team identified 460 adults who had been admitted for a first episode of severe depression. They then identified 460 other adults of similar age and sex who had been treated at the same hospitals for unrelated conditions — orthopedic injuries, gallbladder surgery, routine obstetric care — and had no history of depression in their record.

Every person in both groups was interviewed by a trained research assistant about the previous three years of their life, using a structured questionnaire covering housing, relationships, finances and employment, including whether they had lost a job. Interviews averaged forty minutes and were conducted by phone. Among those admitted for depression, 38% reported a job loss in that window. Among the comparison group, 19% did. The authors report that the two groups were well matched on age, sex and insurance type, and acknowledge in their discussion that interview data depend on accurate recall.

Six hundred adults on the waiting list of a community mental health clinic agreed to take part in a trial registered with a national clinical trials registry before enrollment began. Eligible participants were adults aged 18 and over who owned a mobile phone and had been waiting at least four weeks for an intake appointment. Recruitment ran from January to April 2023.

A computer assigned each person, by the equivalent of a coin flip, to one of two groups. One group received a daily supportive text message for twelve weeks; the other received the clinic's usual waiting-list care and no messages. Neither the participants nor the staff chose which group anyone went into. At baseline the two groups were similar in age, sex, and symptom score. At twelve weeks, both groups completed the same symptom questionnaire; 548 of the 600 participants, or 91%, completed it. The text-message group scored lower on average, and the difference was larger than would be expected by chance. The authors note that participants knew which group they were in, since receiving a daily message is difficult to conceal.