It is two in the morning and you are sitting with your phone in your hands, going over the conversation again. Your daughter is twenty-five. She has been in and out of treatment twice in three years, and what you have noticed โ and what no one has named clearly for you โ is that the drinking and the disappearing and the months of paralysis move together. The anxiety and the substance use travel as a pair, each one feeding the other, neither one quiet long enough to treat. You have looked up programs. You have printed articles. You have rehearsed conversations that felt brave in the dark and impossible by morning. What you have not been told โ and what the evidence is beginning to make clear โ is that your role in whether she ever walks into a treatment room is not what you have been led to believe.
A growing body of peer-reviewed research now points toward something both specific and, for many families, counterintuitive: for young adults living with both a substance use disorder and a co-occurring mental health condition, the path to treatment is not primarily a matter of individual readiness. It is built โ or blocked โ by the quality of the relationships surrounding them. Families who understand the mechanics of this are not simply watching from the sidelines. They are, whether they know it yet or not, participants in the architecture of help-seeking itself.
---
**The clinical reality most families were never told about**
The picture most families carry of addiction โ a discrete problem with a discrete treatment solution โ rarely matches what clinicians actually see when young adults arrive in crisis. Kiefer D. Cowie and colleagues published a naturalistic study in *Early Intervention in Psychiatry* in 2026 that documented what practitioners have long suspected but rarely quantified. Their sample consisted of 612 young adults who initiated services at a psychiatric specialty clinic between June 2016 and January 2020. These were not patients presenting at an addiction program; they were young people arriving for general psychiatric care. And yet self-report measures revealed a high rate of substance use and co-occurring nonsubstance mental health concerns across the sample as a whole (Cowie et al. 2026). Anxiety, depression, trauma, emotional dysregulation โ these did not wait outside the door while the substance use was treated. They came in together.
The clinical term for this overlap is concurrent disorders, and Cowie's sample suggests it is not a rare complication but a defining feature of how young adults in mental health crisis actually present. For families, this reframes what can feel like inexplicable cycles โ the sobriety that collapses into anxiety, the stable weeks that erode into using โ as the signature of a dual clinical picture rather than a failure of willpower or character.
What follows is an argument in three parts. First, the stakes: new spatial analysis of emergency services data reveals that the communities where overdose concentrates are also communities where younger adults are reaching the threshold for cardiac emergency at elevated rates โ a public health geography signaling that the environments most burdened by addiction are not single-problem environments. Second, the mechanism: recent qualitative research has clarified how help-seeking decisions actually work among people with substance use and mental health problems, and the findings challenge the assumption that readiness is purely internal. Third, the therapeutic landscape: Cowie's study offers both a caution and a point of hope about what happens when young adults do reach treatment โ and what keeps them from getting there.
---
**A public health geography of overlapping crises**
Jeffrey Pesarsick and colleagues published a spatial point process analysis in *Preventing Chronic Disease* in 2026, examining emergency medical services activations in Washington County, Pennsylvania, across seven years of data from 2017 to 2023. Their central question was whether cardiac EMS activation patterns varied by age group, and whether local overdose activity modified that variation.
The answer was yes on both counts. Younger adults between the ages of 18 and 45 showed significantly different patterns of cardiac EMS activation compared to those over 45. More importantly, local suspected overdose activity modified that age-dependent pattern in a statistically significant way (Pesarsick et al. 2026): in neighborhoods where overdose EMS calls were more prevalent, the geographic density of cardiac emergencies among younger adults was elevated at rates that substantially improved the model's fit when the interaction was included (deviance = 228.12, p < .001).
It is essential to state what this kind of spatial, population-level analysis cannot show. The cardiac EMS calls tallied in high-overdose neighborhoods may belong to people who use substances, to people who do not but live nearby and are subject to the broader stressors that cluster in the same communities, or to some combination of both. The study design aggregates data to the neighborhood level and cannot link any individual cardiac event to any individual's substance use history. No individual-level claim โ about what is happening inside any particular body โ follows from a neighborhood-level correlation.
What the study does establish is a public health geography: the neighborhoods where overdose concentrates are also the neighborhoods where younger adults are calling for cardiac emergencies at elevated rates. Pesarsick and colleagues describe this convergence as "key to developing age-appropriate public health intervention materials to lessen early onset cardiovascular disease illness and death" โ framing it as a structural community pattern worth targeting, not a claim about individual risk profiles (2026). For families, the pattern is a reminder that the environments most affected by addiction are not single-problem environments. Multiple crises appear in the same places, among the same age groups. The geography is not neutral.
---
**Help-seeking is not a solo act**
Given this landscape, it might seem that the path forward is straightforward โ show a young person the evidence, point them toward a program, and stand back. But Catriona Connell and colleagues, writing in *PLOS ONE* in 2026, map the terrain of help-seeking in ways that should change how families think about their own role.
Connell's team applied framework analysis to qualitative data from a mixed-methods social network study, examining help-seeking behavior among people in contact with the criminal justice system who were living in the community โ a population with high rates of both substance use and mental health problems. Their framework identified three essential elements: the desire to seek help, the ability to seek help, and the context in which seeking help occurs. The finding that cuts deepest is this: all three elements are shaped by relational influences. Help-seeking, Connell and colleagues conclude, "is not an individual behaviour, but strongly affected by relational influences that operate between individuals, across social networks, and via cultural norms" (Connell et al. 2026).
This is a qualitative study conducted among a specific population โ people in contact with the criminal justice system โ and its conclusions are best understood as theory-building: a framework for understanding behavior, not a universal clinical prescription. Whether these dynamics generalize to young adults in outpatient psychiatric settings, or to families without significant criminal justice involvement, is an open question the research cannot answer from within its own data. What the study offers is a carefully constructed account of how relational and contextual factors shape whether a person who privately wants help is able to act on that want.
In the populations Connell and colleagues studied, where substance use, mental health struggle, and institutional contact cluster in the same social networks, the ambient norms around seeking help are correspondingly narrow. A person who privately wants help and can technically access it may still not reach for it, because no one visible in their world has modeled that reaching is something people do. The network speaks louder than the internal resolve.
Connell and colleagues conclude that optimizing help-seeking requires "multi-level policy/practice interventions that go beyond individual factors" โ meaning that approaches focused solely on motivating the person in front of you are working against the actual structure of how these decisions are made (Connell et al. 2026).
---
**When treatment is reached, the data changes**
Returning to Cowie's data from the specialty clinic surfaces a finding with a difficult shape. Of the 612 young adults in the sample, 30.6% were referred to a dialectical behaviour therapy (DBT) skills training group โ a structured therapeutic modality designed to address emotional dysregulation and co-occurring distress. Of the total sample, only 15.4% actually attended. The majority of young adults who might have benefited from the intervention did not receive it (Cowie et al. 2026).
For those who completed the group, however, the pre-to-post change data was meaningful: effect sizes of d = 0.35 to d = 0.69 on measures of psychiatric symptoms and functioning. By Cohen's conventional benchmarks โ where d = 0.2 is small, d = 0.5 is medium, and d = 0.8 is large โ the lower end of this range (d = 0.35) falls in the small-to-medium band, and the upper end (d = 0.69) in the medium-to-large band. The study design was naturalistic and pre-to-post rather than controlled, which limits how firmly causal conclusions can be drawn. Within those limits, meaningful change was observed for the young adults who completed the intervention.
What makes these findings most instructive is what Cowie and colleagues found โ and did not find โ about who engaged. Young adults who participated in group programming differed from those who did not on several clinical and demographic variables. But in regression models, no individual variable significantly predicted group engagement (Cowie et al. 2026). A null result in regression can arise from many sources โ unmeasured variables, collinearity, sample limitations, measurement error โ and the study did not examine relational or contextual factors, so nothing in this finding confirms that such factors explain engagement. What can be said is narrower: the individual clinical profile, as measured, did not predict who showed up. Whether the relational environment surrounding these young adults played a role is a hypothesis the Cowie data is consistent with, but cannot confirm. It is Connell's framework that supplies the mechanism; Cowie supplies only the gap the mechanism would fill.
---
**The objection every family member already knows**
There is an objection every family member who has been through this more than once will recognize before it is stated: I cannot make her go. Forcing an adult into treatment against their will is, at best, ineffective and, at worst, destructive to the trust that makes any future help-seeking possible. At some point it has to be her decision.
This is true. Nothing in the research above argues for coercion or for manufacturing willingness where none exists. What the research does is clarify what kind of decision this is, and what conditions shape it.
If Connell and colleagues are right that help-seeking is relational โ that it rises or falls on the social context visible to the person making the choice โ then the quality of the relationship between a family member and a young adult is not incidental to the decision. A young adult who believes her family sees her as a problem to be solved, a source of embarrassment, an obligation to be managed, is making a decision about whether to seek help in a fundamentally different environment than one who believes her family sees her as a person in pain who has not yet found the right door.
The question is not whether you can make her go. The question is what relational conditions you are sustaining while she decides whether she wants to. That is not a passive position. It is a form of presence that โ if Connell's model applies beyond the population it studied โ may be among the most structurally significant things a family can offer. Not because it guarantees an outcome โ the evidence does not support that claim, and this article will not make it โ but because it changes the environment in which the decision is being made.
---
**The thread that holds**
Two in the morning. The phone. The daughter who moves through anxiety and substance use as through a paired current you cannot interrupt from the outside.
The research does not hand you a moment when this becomes simple, and it would be dishonest to suggest otherwise. What it offers is something smaller and more durable: a more accurate account of what you are actually doing when you stay present, when you maintain warmth, when you refuse to let the relationship be defined entirely by the crisis. You are not simply waiting. By Connell and colleagues' account, you are influencing the conditions under which a frightened young person will eventually decide whether seeking help is something people like her do.
When she reaches that treatment room โ and Cowie's data suggests that for the young adults who get there and stay, meaningful change is measurable โ your relationship will have been part of what made the decision possible. Not sufficient. Not guaranteed. But part of the architecture.
Facing addiction with hope and understanding is not mere optimism. The research suggests it is something structural: the kind of relational environment in which a person can eventually find their way toward help. For a family sitting awake at two in the morning, that may be the most important thing there is to know.