Helped into Harm

Abstract

Background: Interventions can have harmful effects among subgroups they intend to help. The Moving To Opportunity experiment, in which families were randomized to receive a Section 8 housing voucher, was one example. Voucher receipt generally resulted in better long-term mental health and lower substance use and risk behavior outcomes among adolescent girls, but resulted in worse outcomes among adolescent boys. Reasons for this discrepancy and the unintended harmful health effects for boys are unclear. We used mediation analysis to estimate processes through which voucher receipt was hypothesized to affect adolescent mental health and substance use. Methods: We used longitudinal data (10-15 years) on boys enrolled in Moving To Opportunity. We estimated interventional (also known as stochastic) indirect effects of voucher receipt on mental health and substance use outcomes through mediators capturing aspects of the school environment, neighborhood poverty, and instability of the social environment. We also estimated interventional direct effects not operating through these mediators. We used a robust, efficient, nonparametric substitution estimator in the targeted minimum lossbased framework. Results: Housing voucher receipt increased long-term risk of any diagnostic statistical manual disorder, any mood disorder, any externalizing disorder, and cigarette smoking among boys. The majority (between 69% and 90%) of the total negative long-term effects could be explained by indirect effects through the mediators considered.Conclusions: This evidence suggests that, even though the intervention had the desired effects on neighborhood poverty and the school environment, these "positives" ultimately negatively impacted the long-term mental health and behaviors of boys.

Social Deprivation and Population Density Trajectories Before and After Psychotic Disorder Diagnosis

“…This finding provides vital clues for intervention research and suggests that ambitious trials are now warranted to investigate whether moving people out of more deprived environments can ameliorate SMI risk. To our knowledge, no trial has tested such interventions regarding SMI, although the Moving to Opportunity trial has shown evidence that moving to higher-quality neighborhoods resulted in lower psychological distress in adolescence, although this may also have introduced unintended harms for some groups, including increased mental health risks for boys …”

Mediating to Opportunity: The Challenges of Translating Mediation Estimands into Policy Recommendations

“…Intuitively, natural effects assess the causal effects of treatment mediated (indirect) or not mediated (direct) through some mediator(s) by fixing the mediator to the specific value it would have taken under treatment or no treatment. Despite natural effects' appealing qualities, Rudolph et al 7 eschew them for good reason. Natural effects require strong identification assumptions 9,11,12 and are difficult to interpret: 13 in this situation, the natural direct effect, for example, would be a contrast between the probability of a given behavioral or substance use outcome had every boy's family not received a voucher, and the probability of the outcome had all families received vouchers although each boy was exposed to the school, neighborhood, and social environments he would have lived in had his family not received a voucher.…”

“…In addition, controlled direct effects have limited utility if the goal is to quantify the mediated portion of the total effect, as there are generally no corresponding indirect effects, 15,16 although the contrast between the direct and total effects can be conceptualized as the portion of the harmful effect eliminated by a policy intervention. 17 Although natural effects can be used to ask and answer theoretical questions about mechanisms, and controlled direct effects about a possibly radical intervention on the mediator, the stochastic effects Rudolph et al 7 estimate fall somewhere in between. Instead of assigning each boy to the environment he would have experienced in the absence of a housing voucher or all boys to the same environment, these estimands instead apply distributional interventions on the mediator.…”

“…Although stochastic estimands correspond with possibly implementable policies, Rudolph et al's 7 conclusion from their analysis is a mechanistic one: most of the harmful effects of voucher assignment on boys' behavioral and substance use outcomes operated through the environmental characteristics they investigated. Adhering to this journal's recommendations on policy implications, 21 the authors refrain from commenting on how their findings could be applied in future housing interventions.…”

“…In this issue, Rudolph et al turned to mediation to reanalyze boys' MTO data in an attempt to understand what drove the harmful effects boys experienced, examining behavioral and substance abuse outcomes and a set of mediators describing boys' school, neighborhood, and social environments. 7 In the context of that work, this commentary highlights key methodologic and conceptual decision points for epidemiologists turning to mediation in an effort to autopsy harmful policy experiments and improve interventions. First, we propose distinct goals of mediation analyses in such situations.…”