Which health interventions have a truly large impact, and which ones make hardly any difference?
Exercise and stable housing for homeless people have the strongest evidence base; antidepressants demonstrably work for severe depression, but the agents differ in their side-effect profile. Some popular approaches, such as population-wide fall prevention and apps, do not yet have convincing evidence.
Exercise is probably the most broadly effective intervention that exists for older adults. People who are physically active have a lower risk of premature death, cardiovascular disease, multiple forms of cancer, bone fractures, falls, dementia and depression, all at the same time. That association has been found repeatedly and across many different outcomes. Whether exercise itself causes those benefits, or whether healthier people simply exercise more, cannot be proven conclusively, but the biological mechanism is well understood and the breadth of the effect is impressive.
Stable housing for homeless people (the 'Housing First' principle, in which someone receives a home immediately without sobriety or treatment being a condition) also has a large and reliable effect. After one and a half to two years, participants were more than twice as likely to be stably housed as people in standard shelter care. The programme also led to fewer emergency department visits and fewer hospital admissions: the incidence ratio for ED visits was 0.63 and for admissions 0.76, representing clearly measurable reductions. A frequently heard concern, that people drink more or use more drugs when abstinence is no longer required, was not borne out by the studies.
Antidepressants work better than a placebo for severe depression; that holds for all 21 agents studied. The differences between agents are modest: amitriptyline, escitalopram and mirtazapine score best on efficacy, while reboxetine, fluoxetine and trazodone are somewhat less effective. For those who weigh side effects heavily, agomelatine, escitalopram and sertraline are better tolerated than amitriptyline or clomipramine.
Behavioural interventions in which a coach helps someone find their own motivation produce an average of around 1,300 extra steps per day, but that effect disappears after one year. As soon as such an approach is compared with equally intensive alternative support, no demonstrable difference remains. The attention itself therefore plays a large role.
Population-wide fall prevention for older adults, such as free gym memberships combined with home modifications and education, has an unclear effect. Some studies find fewer falls, others do not. There are currently too few good studies for a clear conclusion. Digital lifestyle tools (apps, online coaching) help with self-monitoring and goal setting, but the researchers did not report the size of that benefit in a consistent way.
This answer draws on several large review studies that brought together dozens of earlier studies. The evidence for exercise in older adults and for Housing First is the strongest: it comes from multiple independent studies with consistent outcomes. For antidepressants, an extensive comparison of 21 agents was used, but the researchers rated the quality of evidence as ranging from moderate to very low. The evidence for motivational interviewing is low to very low, and the researchers themselves noted that effects disappear after one year. For population-wide fall prevention, the strength of evidence was rated as very low by the researchers themselves, owing to inconsistent results and methodological problems. The evidence for digital interventions is limited because the review study used is more an inventory of methods than a measurement of effects.