What thirty years has taught me about who actually recovers

Willpower, intelligence and the depth of remorse predict very little at intake. In thousands of admissions over thirty years, three plainer observations have shifted my prognosis more reliably: whether the person volunteers inconvenient facts unprompted, who drove them to the admission, and whether the consequences have landed on the person or been absorbed by the family. This article sets out how a first interview is read and what revises the assessment during treatment.

Reading the first interview

Four questions carry most of the prognostic information: when was the last use, how much, how many days in the past month, who else knows the full facts. Direct answers to all four, without mitigation, demonstrate the one behaviour treatment depends on. An answer built from work pressure and childhood demonstrates the opposite. A GP or EAP practitioner assessing possible dependence should apply the same rule: decline the narrative, require the specifics.

Two observations precede the questions. The first is who brought the person: the admissions that hold have almost always been driven by one steady relative who knows the full facts, while a person delivered by a delegation, or by a family that has spent years settling debts and explaining absences to employers, arrives with the consequences still outsourced. The second is a consequence audit: who paid the debts, who spoke to the employer, who cleaned up after the last incident. Where the answer to all three is a family member, the person has not yet carried the cost, and the earliest requests to leave the programme come from this group.

The polished patient is the higher-risk profile

Fluency at intake is routinely read as a good sign. It is the opposite. An intelligent, articulate patient acquires the treatment vocabulary within a week, speaks well in group, and produces accounts of triggers and childhood that sound like insight. The indicator is workload: the counsellor is working harder than the patient. This patient does not deny; he supplies context and half-truths until the family doubts what it saw. The plain admission, I drink my wages and I cannot stop, carries more prognostic weight than a twenty-minute account containing no admission at all. Intelligence becomes an asset only after honesty is established.

The base rates

A 2000 JAMA paper by McLellan and colleagues put relapse in substance dependence at roughly 40 to 60 percent, comparable to recurrence in type 2 diabetes, hypertension and asthma. The planning consequence is structure built for a condition managed over years, not an infection cured in 28 days. In a 2019 national study led by Kelly, people who had resolved a serious alcohol or drug problem had made an average of 5.35 recovery attempts, with a median of just two. Most who recover needed a few attempts; a first relapse sits inside the expected course.

Collapses concentrate months after a crisis, once the person has decided he is well. A worked example. A man sober for some months is promoted, concludes he is recovered, and stops attending his weekly meeting. Nothing happens for several weeks, which he reads as confirmation. The relapse arrives later, in unstructured evenings with nobody asking direct questions. The failure point was the decision to stop attending, not the eventual drink.

What revises the assessment mid-treatment

Prognosis is revised on behaviour, not speeches. Upward: the first unprompted disclosure, a sentence such as "I used on Saturday" or "I have been phoning that friend again", offered before anyone had proof; acceptance of structure without negotiation; return to routine after a bad weekend. Downward: treatment vocabulary arriving before any disclosure, and negotiation over rules. The strongest downward revision is disappearance after a slip. The people who do well are seldom impressive at intake; they arrive late, defensive and tired, then repeat ordinary behaviours: working, telling the truth, keeping appointments. Durable recovery without disclosure is rare; a person in early recovery needs two or three people who know everything and may ask direct questions.

The factors with evidence behind them

A 2020 Cochrane review of 27 studies and 10,565 participants found manualised twelve-step facilitation outperformed other active treatments on continuous abstinence, 42 percent versus 35 percent still abstinent at twelve months. The protocol consequence: residents are handed into AA and NA meetings before they leave, and the relative who drives someone to a Tuesday meeting is doing clinical work.

A 2023 Recovery Research Institute review found employment among the strongest predictors of remaining substance-free after treatment, which is why work reintegration is built into the halfway-house phase. The morning reason does not need to be noble; a child, a job, a debt to clear, any of these serves.

The last factor is agreed with the family before discharge: after a slip, the first phone call gets help arranged rather than a lecture, and the promise is kept. The purpose is to prevent concealment, which converts a single use into a hidden pattern. A person who reports a slip early is in a better clinical position than one who has begun concealing use.

Sources and further reading

Kelly JF, Humphreys K, Ferri M, Alcoholics Anonymous and 12-step facilitation, Cochrane Review CD012880, 2020.

McLellan AT et al., Drug dependence, a chronic medical illness, JAMA, 2000.

Kelly JF et al., National Recovery Study, Alcoholism: Clinical and Experimental Research, 2019.

Recovery Research Institute, Employment and recovery outcomes review, 2023.