Beyond the Individual: Measuring Household Spillover Effects in Depression Treatment

Mental health research has long focused on the individual as the unit of analysis. A participant enrols, is randomized, receives an intervention, and their outcomes are measured. The people who live with them — partners, children, parents, siblings — are background variables at best, absent entirely at worst.

This individual focus is methodologically convenient but empirically incomplete. Depression is not a condition that exists in isolation from its social context. It affects the quality of relationships, the capacity for caregiving, the functioning of households. When one person’s depression improves, the people around them are likely to experience changes in their own wellbeing — changes that conventional outcome measurement misses entirely.

The research on household spillover effects in depression treatment is attempting to correct this blind spot. Drawing on evidence from the WHO Step-by-Step program and broader economic wellbeing research, this post reviews what we know about spillover effects, how they’re measured, what they add to cost-effectiveness analysis, and what they suggest about how we should think about the value of mental health intervention.

The Theoretical Basis for Spillover Effects

The hypothesis that depression treatment generates household spillover benefits is grounded in well-established relational and economic theory.

From a relational psychology perspective, depression impairs the capacity for emotional availability, responsiveness, and connection. Depressed parents show reduced sensitivity to infant cues, less consistent discipline with older children, and higher rates of conflict in adult relationships. Partners of depressed individuals experience elevated stress, reduced relationship satisfaction, and — in longitudinal studies — elevated rates of depression themselves, a phenomenon sometimes called “emotional contagion” or “depression crossover.” When one person’s depression improves, the quality of these relational interactions improves, with downstream effects on partner and child wellbeing.

From an economic perspective, depression reduces productive capacity — both in formal employment and in household production (childcare, domestic labor, community participation). Household income and functioning suffer. When treatment restores productive capacity, economic benefits flow to the household as a whole, not just to the treated individual.

These theoretical pathways suggest that the benefits of depression treatment should extend beyond the individual in predictable and measurable ways. The question is whether the empirical evidence bears this out, and to what magnitude.

Measuring Spillover: Methodological Approaches

Measuring household spillover effects from mental health interventions requires a study design that goes beyond standard clinical trial methodology. Most RCTs collect outcome data only from enrolled participants; capturing spillover requires also collecting data from household members who did not themselves receive the intervention.

Several approaches have been used in the literature:

Partner and family member surveys. Some trials have included instruments measuring partner wellbeing, relationship quality, or child functioning, administered to household members of enrolled participants at follow-up. This approach directly measures spillover at the individual level but requires consent and cooperation from people who are not study participants themselves.

Economic production and labor force participation measures. Studies using administrative data or household surveys can measure changes in employment, income, and labor force participation among household members following treatment of the index participant. This captures the economic dimension of spillover but misses emotional and relational effects.

Structural modelling from population data. Using population-representative data on the relationship between depression and household wellbeing, researchers can estimate the expected spillover benefits of depression treatment without requiring direct measurement of household members in clinical trials. This approach draws on the established literature linking individual depression to household outcomes to project spillover at scale.

The approach used in the Ambitious Impact analysis of WHO Step-by-Step draws on the structural modelling method, using evidence from the economics of wellbeing literature to estimate spillover effects from the individual outcome gains measured in the five RCTs. This allows for systematic inclusion of spillover in cost-effectiveness analysis without requiring household-level data collection in clinical trials.

The Step-by-Step Spillover Estimate

The Ambitious Impact cost-effectiveness analysis estimated that household spillover effects account for approximately 16.24 percent of the total impact of the WHO Step-by-Step program.

To understand what this means: if the direct effect of the program on enrolled participants — measured in WELLBYs (wellbeing-adjusted life years) or DALYs — represents 100 percent of the individually measured impact, the household spillover adds approximately 16 additional units of benefit for every 100 units of direct individual benefit. The total impact of the program is therefore approximately 116 percent of what would be estimated from a standard clinical trial outcome analysis.

This 16 percent uplift has a direct effect on cost-effectiveness ratios. If the program costs $97 per DALY averted at the individual level, accounting for spillover reduces the effective cost per DALY to approximately $84. Similarly, the $25/WELLBY figure at the individual level becomes approximately $22/WELLBY when spillover is included.

These are not trivial adjustments. In cost-effectiveness analysis, a 16 percent improvement in the benefit side of the equation can shift a program from “highly cost-effective” to “exceptionally cost-effective” relative to relevant thresholds and comparators. For programs near the margin of funding decisions, spillover effects could be decisive.

The Evidence Behind the Estimate

The 16.24 percent figure derives from several lines of evidence in the wellbeing economics literature.

Studies of the relationship between individual depression and partner wellbeing consistently show that having a depressed partner reduces life satisfaction scores by meaningful amounts — estimates vary by study but typically fall in the range of 0.2-0.5 points on a 10-point scale. Applying population data on household composition and depression prevalence, the aggregate wellbeing cost to partners of depressed individuals can be estimated, and from this the expected benefit to partners from depression treatment can be projected.

Studies of parental depression and child outcomes provide a parallel line of evidence. Maternal depression, in particular, has well-documented effects on child developmental outcomes — cognitive, behavioral, and emotional. Estimates of the wellbeing impact on children of parental depression treatment vary but are generally positive and statistically significant in longitudinal studies.

The combination of partner spillover and child spillover effects produces the overall household spillover estimate. The 16.24 percent figure is derived from a weighted combination of these pathways, calibrated to the household composition and depression rates in the target population contexts.

Limitations and Uncertainties

The spillover estimate is subject to several sources of uncertainty that should be acknowledged in any rigorous analysis.

The structural modelling approach assumes that the relationship between depression and household wellbeing observed in population data generalizes to the specific populations and contexts served by Step-by-Step. This assumption is plausible but not directly validated. The socioeconomic, cultural, and household structure characteristics of Syrian refugees in Lebanon, rural communities in Pakistan, or urban populations in South Africa may produce different spillover magnitudes than those estimated from population-level studies in other settings.

The estimate does not account for potential negative spillover effects. There is limited evidence that some forms of depression treatment — particularly those that change individual behavior and relationship dynamics — can produce relationship strain or adjustment difficulties for partners. If some participants’ improvement comes at the cost of relationship disruption, aggregate spillover benefits would be overstated. This seems unlikely to be a major factor for a psychosocial program like Step-by-Step, but it represents a theoretical caveat.

The estimate relies on WELLBY-based calculations for spillover, but the primary cost-effectiveness analysis for Step-by-Step also reports DALY-based figures. Translating spillover effects measured in subjective wellbeing terms into DALY equivalents requires assumptions about the relationship between subjective wellbeing and disability-adjusted life years that are not straightforward. The WELLBY-based spillover calculation is more methodologically tractable but less directly comparable to the DALY literature that dominates global health cost-effectiveness analysis.

Direct measurement is lacking. The estimate is derived from structural modelling rather than direct measurement of household members in the Step-by-Step trials. Prospective collection of household member wellbeing data in future Step-by-Step trials would substantially strengthen the spillover evidence base and enable direct, rather than modelled, estimation of spillover effects.

Implications for Cost-Effectiveness Analysis

The spillover evidence raises a methodological question that extends beyond Step-by-Step: should standard cost-effectiveness analysis for mental health interventions routinely include household spillover effects?

The case for inclusion is strong. If the goal of cost-effectiveness analysis is to capture the full social value of an intervention — to provide an accurate estimate of total benefits relative to total costs — then excluding well-evidenced spillover effects produces systematically biased (downward) estimates of program value. Programs that affect household-level wellbeing, as mental health interventions clearly do, will be systematically undervalued by analyses that focus exclusively on enrolled participants.

The practical argument for inclusion is reinforced by the competitive context in which mental health programs are evaluated. Infectious disease interventions — which dominate global health funding priorities — do not typically generate household wellbeing spillover effects in the same way, because their benefits are primarily biological and are captured in the treated individual’s outcomes (though there are exception cases like HIV treatment reducing transmission). Systematically excluding spillover from mental health cost-effectiveness while implicitly capturing all benefits for infectious disease comparators creates a methodological asymmetry that disadvantages mental health in funding decisions.

The case against routine inclusion rests primarily on the uncertainty of spillover estimates and the risk of double-counting. If household members of program participants are themselves candidates for program enrollment (e.g., a partner who is also depressed could themselves enroll), their wellbeing improvement might be captured both as direct program benefit (if they enroll) and as spillover benefit (from their partner’s enrollment). Careful analysis is needed to avoid this double-counting, particularly at scale where program saturation in households or communities could create complex overlapping effects.

What Spillover Effects Tell Us About Mental Health Investment

Beyond the cost-effectiveness calculation, the spillover evidence has a broader implication for how we frame the value of mental health investment.

The standard framing — depression costs X DALYs per year, therefore investment in depression treatment is justified by Y DALYs averted — treats mental health as a private health problem with private health benefits. The spillover evidence shows that this framing is incomplete. Depression is a social condition with social consequences, and treating it has social benefits that extend beyond the individual.

This reframing matters for advocacy and funding. The case for mental health investment is often made in terms of individual rights and individual suffering — compelling arguments, but ones that can be heard as “special pleading” for a cause area that may seem to compete with other health priorities. The spillover evidence supports a different framing: mental health investment is family investment, community investment, and human capital investment. Its benefits are diffuse, social, and extend across generations through the effects of parental mental health on child development.

A funder or policymaker who thinks about mental health investment as “treating one person’s depression” is using a narrow frame that the evidence doesn’t support. The more accurate frame is: “improving the functioning of households, families, and communities by treating the individuals within them whose depression degrades collective wellbeing.” At 16 percent spillover, and potentially more, the difference between these frames is not marginal.

Research Priorities

Several research priorities emerge from this review.

Future RCTs of Step-by-Step and comparable programs should prospectively collect household member outcome data. The spillover effect should be measured directly, not estimated from structural models, in at least some trials. This would substantially improve the precision of spillover estimates and validate (or revise) the structural modelling approach.

Research should explore variation in spillover magnitude across household structures and cultural contexts. The degree to which spillover effects generalize from high-income, Western household structures to the diverse household arrangements found in target countries for Step-by-Step is an important open question.

Studies on child development outcomes in families where a parent participates in Step-by-Step would extend the spillover evidence in a particularly important direction. Parental mental health is one of the strongest determinants of child outcomes; evidence that Step-by-Step improves children’s wellbeing would significantly strengthen the case for its inclusion in integrated family health programs.

Finally, methods for incorporating spillover effects into DALY-based cost-effectiveness analyses — the dominant framework in global health — need development. The current WELLBY-based approach is valuable but creates a disconnect with the infectious disease comparators against which mental health programs are evaluated.

Conclusion

The household spillover evidence adds approximately 16 percent to the measured impact of the WHO Step-by-Step program — a meaningful adjustment that improves an already compelling cost-effectiveness case. More fundamentally, it reframes what we should understand mental health treatment to be: not a private benefit to an individual patient, but a social investment whose returns are distributed across families and households.

Methodologically, the field is in early stages of incorporating spillover into standard evaluation frameworks. Practically, the evidence is sufficient to justify including spillover in cost-effectiveness estimates for Step-by-Step and comparable programs, with appropriate acknowledgment of uncertainty. Strategically, the spillover evidence strengthens the case for mental health as a global health priority and challenges the narrow individual-health framing that has historically undervalued it.

Depression treatment is family treatment. The sooner the field adopts that frame, the more accurately it will measure and communicate the value of the programs it evaluates.

For more on the research behind the WHO Step-by-Step program and the work of Kaya Guides, visit besidehealth.org.