Harm reduction: clinical reasoning beyond all or nothing

At the outpatient follow-up, the woman in her early thirties says she is still using opioids, but that she now always uses with someone present, that she got a naloxone kit and that she switched syringes at a program near her home. The resident writes "maintains use, no adherence to treatment" and suggests to the preceptor ending the follow-up until the patient "is ready". The preceptor asks: "Since the last visit, has her risk of death gone up or down?". The resident stops. The honest answer is that it went down, and the note describes the opposite.

NEXO is a daily medical diagnosis game: one clinical case a day, up to six progressive clues, and ICD-10 proximity feedback across five levels.

This text treats harm reduction as a problem of clinical reasoning. Addiction is the setting, but the question is general: how to decide when the ideal outcome is out of reach in the short term, how to avoid all-or-nothing thinking and how to talk in a way that keeps the person coming back. You will see definitions, principles for health services, a table of interventions read as risk reduction, the evidence and how to interpret it, plus phrases that open or close a conversation. Specific approaches and protocols follow the guidelines and the network of each place.

Key points

• Harm reduction is an approach that seeks to reduce the negative consequences of substance use without requiring abstinence as a condition for care.
• All-or-nothing thinking is a bias: it classifies any outcome other than abstinence as failure and ignores large changes in risk.
• Goals negotiated with the person, ordered by risk, turn a binary decision into a sequence of possible steps.
• The evidence on take-home naloxone, syringe exchange programs and treatment with opioid agonists measures outcomes such as deaths avoided and transmitted infections, and should be read with that outcome in mind.
• Keeping the person in care is, in itself, an intervention, because each visit is a new opportunity to change risk.
• The language of the clinical note defines how the next team will read progress.

What harm reduction is, in terms of reasoning

The most cited definition comes from Marlatt, who described harm reduction as a "come as you are" approach: accept the person at the point where they are and work to reduce the consequences of use, instead of conditioning care on a complete change. NIDA and the European Union Drugs Agency describe harm reduction as part of the continuum of care, alongside prevention, treatment and recovery.

For clinical reasoning, the definition can be rewritten like this: faced with a person whose ideal outcome is not available now, the goal becomes the best possible outcome now, with the door open for the next ones. This formulation applies to many areas of medicine. In diabetes, a gradual reduction in blood glucose is celebrated even when the final target is still far off. In heart failure, every admission avoided counts. Harm reduction applies the same logic to substance use.

Hawk and colleagues proposed principles of harm reduction for health services:

• Humanism: the person has value regardless of substance use.
• Pragmatism: care starts from what is possible, not from what would be ideal.
• Individualism: each person has their own needs, risks and priorities.
• Autonomy: the person decides about their own life, with adequate information.
• Incrementalism: any positive step counts, and setbacks are part of the journey.
• Accountability without punishment: choices have consequences, but care remains available.

Note that several of these principles are, in fact, principles of reasoning. Pragmatism and incrementalism describe how to choose goals under constraints. Individualism describes the need to adjust probability and risk to the person. Together they form a method for deciding in situations where the optimal answer is off the table.

All-or-nothing thinking as a bias

Dichotomous thinking, which classifies outcomes as total success or total failure, is a known cognitive pattern. In substance use care, it shows up in phrases like "there is no point treating him while he keeps using" or "she relapsed, she is back to square one". These phrases sound like descriptions, but they are choices of outcome: they define abstinence as the only measure that matters.

The problem is that this single outcome hides large changes in risk. A person who starts using with someone else present, with naloxone nearby and sterile equipment, has a very different risk of death and infection from someone who uses alone, without any of these resources. A chart that records only "continues use" treats the two situations as the same.

An analogy helps to see the bias. No one considers the treatment of a person with asthma a failure if they still have attacks, but have fewer attacks, fewer emergency visits and no hospital admissions in the past year. The asthma is still there, and control has improved. In substance use, the same improvement is usually described as "still using", and the word "still" erases the change. Replacing the single outcome with several outcomes is what makes control visible.

Outcome bias

Judging the quality of a decision by the immediate result, and not by the information available at the time, is another common error. If the person resumes use after a period of treatment, the conclusion "the treatment failed" may be wrong. The treatment may have reduced risk for months, and the resumption may be part of a chronic and fluctuating course, as in other chronic diseases.

Zero-sum thinking

There is also the idea that offering harm reduction competes with treatment, as if a person who receives naloxone had less motivation to stop. This idea works as a hypothesis, and hypotheses are tested. The literature reviewed by McDonald and Strang on take-home naloxone programs examined exactly this concern, and the authors concluded that the programs reduce overdose mortality among participants, with a low rate of adverse events.

Premature closure of prognosis

When the team decides that the person "does not want treatment", the prognosis is closed too early. The post on prognostic reasoning shows how to estimate trajectories, and the trajectory of substance use is usually long, with periods of change that are not always predictable at the start.

The correction for all these biases is the same: define intermediate outcomes that matter and track each of them. Deaths avoided, infections prevented, hospital admissions reduced, connection maintained. Each one is a real result, and all of them can improve before abstinence appears, or even if it never appears.

Goals, risk and prognosis: deciding with the person

If the ideal outcome is out of reach now, the question becomes: which next step reduces the most risk with the greatest chance of being carried out? This is a question of clinical decision-making, with the same format as decisions about tests and management: consider benefit, probability of success and the person's preference.

One way to structure the conversation is to rank the risks:

1. Risk of immediate death. Overdose, solitary use, combining depressant substances, loss of tolerance after a period without use, such as after leaving prison or after a hospital stay.
2. Risk of infection. Sharing equipment, injection practices, access to testing and treatment for HIV and viral hepatitis.
3. Risk of clinical complication. Abscesses, endocarditis, liver disease, lung disease.
4. Social risk. Loss of housing, work, relationships and access to the health service.

Ranking the risks turns a decision of "treat or not treat" into a list of targets. The person may not be willing to stop using, but may accept a naloxone kit, testing for infections, a hepatitis B vaccine or a conversation about not using alone. Each acceptance reduces a concrete risk.

The decision also involves uncertainty. You rarely know which intervention the person will follow, or when they will be ready for another. The post on diagnostic uncertainty describes strategies for deciding when information is incomplete, and they apply here: choose reversible actions, reassess often and state clearly what would make you change course.

Finally, time is part of the approach. Keeping the person connected to the service is an intervention with its own effect, because each visit allows risks to be reviewed, new options to be offered and changes to be noticed. A decision that ends follow-up until the person "is ready" trades a certain benefit, the connection, for an uncertain goal.

A fictional case in three clues

First clue: a forty-year-old man who injects opioids, recently discharged from a three-week stay for pneumonia. The team records "no use during the stay" and feels relief. Second clue: he says he plans to "use just a little" over the weekend to celebrate the discharge. Third clue: he lives alone and has no naloxone.

The all-or-nothing reading sees a success, the three weeks without use, threatened by a relapse. The risk reading sees something else: reduced tolerance after the period without use, intention to use, solitary use and no naloxone. It is a combination with a high risk of fatal overdose in the coming days. The approach that follows from this reading is concrete: talk about the loss of tolerance, offer naloxone, suggest that he tell someone he trusts and arrange a follow-up visit soon. None of these measures requires him to have decided to stop.

When the person's goal differs from the team's goal

Sometimes the person wants something the team considers secondary: keeping a job, regaining custody of a child, sleeping better, stopping having abscesses. These goals are valuable entry points. A person who wants to stop having abscesses is ready to talk about injection technique, sterile supplies and skin care. A person who wants to keep a job may be open to discussing use patterns and times of higher risk. Starting from the person's goal increases adherence and, over time, opens space for other goals.

The reasoning here is the same as in any shared decision. The team brings the risk estimate and the available options; the person brings values, priorities and knowledge of their own life. The decision that comes out of this conversation is usually less ambitious than the team would like and more realistic than any unilateral prescription. Recording the chosen goal, in the person's own words, helps the next team continue from the right point.

Table: intervention, risk it reduces and clinical question

The table organizes harm reduction interventions by the risk each one reduces and by the question the professional can ask to decide whether it applies. Availability and rules vary between countries and services.

Intervention	Main risk it reduces	Clinical question for the conversation
Take-home naloxone	Death from opioid overdose	Who would be nearby to use naloxone if you had an overdose?
Syringe and needle exchange programs	HIV and hepatitis C transmission	Where do you get sterile supplies and where do you dispose of used ones?
Opioid agonist treatment	Mortality, overdose and injecting use	What have you heard about this treatment, and what would you like to know?
Testing and treatment of infections	Progression of HIV, hepatitis and other infections	When was your last test, and can I offer one now?
Supervised consumption services, where they exist	Death from overdose and use in public places	Where and with whom do you usually use?
Drug content testing, where permitted	Unexpected exposure to more potent substances	Do you know what is in the substance you use?

The third column is the most important for the reasoning. Each question collects a piece of data that changes the risk estimate, and the answer guides which intervention to offer first. It also signals to the person that the purpose of the conversation is their safety, not a judgment about use.

Note also that the table mixes individual and service interventions. Some fit in a single visit, such as offering testing or talking about naloxone. Others depend on the local network, such as syringe programs and supervised services. Knowing what exists in your region is part of clinical competence, just as knowing where to refer a suspected cancer is.

What the evidence shows and how to read it

The evidence on harm reduction is extensive, and reading it well requires attention to the outcomes each study chose.

Mortality outcomes

The systematic review by Sordo and colleagues in the BMJ analyzed cohorts of people in treatment with opioid agonists and found lower mortality during treatment than outside it, with increased risk in the first weeks after leaving treatment. The clinical reading is twofold: treatment protects, and the transition out of it is a period of risk that deserves planning. The Cochrane review by Mattick and colleagues compared maintenance treatment options and showed benefit for retention in treatment and opioid use. In a US cohort, Wakeman and colleagues compared different treatment trajectories and found fewer overdoses and opioid-related hospitalizations among people treated with medication.

Infection outcomes

The Cochrane review by Platt and colleagues found an association between treatment with opioid agonists and lower risk of acquiring hepatitis C, with weaker but suggestive evidence for syringe programs, especially when combined. The meta-analysis by Aspinall and colleagues found an association between syringe exchange programs and reduced HIV transmission, with study quality limitations that the authors themselves discuss.

Service and community outcomes

The review by Kennedy and colleagues on supervised consumption services analyzed public health and public order outcomes, such as overdoses, use in public places and crime in the surrounding area. A qualitative study by Goldman and colleagues on drug checking shows how information about the substance can change use behaviors among young adults.

How to read these studies

Three precautions help avoid hasty conclusions, and the post on critical reading of articles develops each of them:

• Look at the outcome. A study that measures deaths avoided answers a different question from a study that measures abstinence. Comparing the two as if they measured the same thing leads to wrong conclusions.
• Look at the design. Many harm reduction interventions are assessed in observational studies, because randomizing access to protective measures raises ethical problems. This calls for attention to confounders, not automatic dismissal.
• Look at the context. Results depend on how the intervention was implemented, on access and on the population. A program that succeeds in one city may need adaptations in another.

The WHO reference documents on HIV, viral hepatitis and sexually transmitted infections in key populations and the WHO descriptive note on opioid overdose bring together international recommendations. The UNODC World Drug Report provides the epidemiological overview that contextualizes the scale of the problem.

Communication: questions that open and phrases that close

Harm reduction depends on conversation, and conversation depends on words. A poorly worded question can end the connection; a well worded question can bring the detail that changes the approach. The post on empathy and clinical reasoning in the consultation describes how the quality of the relationship improves the quality of the information.

Some phrases tend to close the conversation, even when said with good intentions: "you need to stop, otherwise I can't help", "again? you promised you wouldn't use anymore" and "come back when you're ready for treatment". All of them make abstinence a condition and turn the consultation into a demand. Phrases that tend to open:

• "What has changed since the last time we talked?"
• "What would you like to take care of first?"
• "What makes use more risky for you today?"
• "Can I explain some ways to reduce that risk?"

Listening attentively is also a technique. Reflecting back what the person said ("you want to keep working and you're afraid the use will get in the way of that"), summarizing their priorities and asking them to propose the next step themselves increases the chance that the step happens. These elements appear in person-centered interview approaches and combine well with the incremental logic of harm reduction. The professional stops being the one who demands and becomes the one who helps to think.

The structure of difficult conversations also helps. The SPIKES protocol, created to communicate bad news, brings steps that apply here: prepare the setting, find out what the person already knows, ask permission to share information, offer knowledge in small doses, respond to emotions and agree on a next step.

The clinical note is the last stage of communication. "Keeps using, refuses treatment" and "keeps using; started using accompanied; received naloxone kit; accepted testing; wants to talk about agonist treatment at the next appointment" describe the same person. The second note records progress and guides the next team. If the topic touched on something personal, seek the help you need from a trusted health service.

How NEXO trains this reasoning

NEXO presents fictional cases with a real clinical presentation, with up to six progressive clues. You propose a diagnosis, look for the ICD-10 code and receive proximity feedback on five levels. This format trains a central skill of harm reduction: recognizing that a partially correct hypothesis, in the right ICD-10 chapter, is already progress, and that the next clue can bring you closer to the target.

In cases involving substance use, the exercise is to treat use as a risk factor that changes the probability of certain diagnoses, not as a sufficient explanation. The Collection page shows the scope of the cases, and the limits of use are at /economy. Doctors who want to review cases find the path in NEXO Curator, researchers in NEXO Researcher, content creators in NEXO Creators and institutions in /closer. The design of the game is in what NEXO is.

Frequently asked questions

Does harm reduction mean giving up abstinence?

It means accepting that abstinence is one of the possible goals, not a condition for care. NIDA describes harm reduction as part of the continuum that includes treatment, and many people move from one approach to another over time.

Does offering naloxone encourage use?

The review by McDonald and Strang analyzed take-home naloxone programs and concluded that they reduce overdose mortality among participants. The concern about encouraging use should be treated as a hypothesis, and the available data do not support it.

How do you record progress when the person keeps using?

Describe the risks that changed: accompanied use, access to naloxone, sterile material, testing, connection with the service. These data show the trajectory and guide the next appointment better than "keeps using".

Is the evidence on harm reduction reliable?

There are systematic reviews for several interventions, such as those by Platt and colleagues and Aspinall and colleagues. Many studies are observational, which requires careful reading for confounders, as in other areas of public health.

What should be done when the person leaves a period without use?

Leaving treatment, hospitalization or prison is a period of increased risk, as the review by Sordo and colleagues shows. Planning this transition with the person and with the care network is part of the approach.

Does this text guide treatment?

The text is educational and deals with reasoning and communication. Approaches, medications and services follow the guidelines and the network of each place. If you need support, seek the help you need from a health service.

Final checklist

• I ordered the person's risks, starting with the risk of immediate death.
• I defined with them at least one possible next step now.
• I offered the interventions available at the service without making them conditional on abstinence.
• I read the evidence by the outcome each study measured.
• I used open questions and avoided phrases that end the conversation.
• I recorded in the note the risks that changed, not just the pattern of use.

Play the case of the day: open NEXO and notice how much each clue brings your hypothesis closer to the target, even before you get it right.

References

1. Marlatt GA. Harm reduction: come as you are. Addict Behav. 1996. PMID 8904943.
2. Hawk M, Coulter RWS, Egan JE, et al. Harm reduction principles for healthcare settings. Harm Reduct J. 2017. PMID 29065896.
3. McDonald R, Strang J. Are take-home naloxone programmes effective? Systematic review (Bradford Hill criteria). Addiction. 2016. PMID 27028542.
4. Sordo L, Barrio G, Bravo MJ, et al. Mortality risk during and after opioid substitution treatment: systematic review and meta-analysis of cohort studies. BMJ. 2017. PMID 28446428.
5. Mattick RP, Breen C, Kimber J, Davoli M. Buprenorphine maintenance versus placebo or methadone maintenance for opioid dependence. Cochrane Database Syst Rev. 2014. PMID 24500948.
6. Wakeman SE, Larochelle MR, Ameli O, et al. Comparative effectiveness of different treatment pathways for opioid use disorder. JAMA Netw Open. 2020. PMID 32022884.
7. Platt L, Minozzi S, Reed J, et al. Needle syringe programmes and opioid substitution therapy for preventing hepatitis C transmission in people who inject drugs. Cochrane Database Syst Rev. 2017. PMID 28922449.
8. Aspinall EJ, Nambiar D, Goldberg DJ, et al. Are needle and syringe programmes associated with a reduction in HIV transmission among people who inject drugs: a systematic review and meta-analysis. Int J Epidemiol. 2014. PMID 24374889.
9. Kennedy MC, Karamouzian M, Kerr T. Public health and public order outcomes associated with supervised drug consumption facilities: a systematic review. Curr HIV/AIDS Rep. 2017. PMID 28875422.
10. Goldman JE, Waye KM, Periera KA, et al. Perspectives on rapid fentanyl test strips as a harm reduction practice among young adults who use drugs: a qualitative study. Harm Reduct J. 2019. PMID 30621699.
11. Volkow ND. NIH Invests in a New Harm Reduction Research Network. National Institute on Drug Abuse, Nora's Blog, Dec. 16, 2022. nida.nih.gov.
12. European Union Drugs Agency. Harm reduction. euda.europa.eu.
13. World Health Organization. Consolidated guidelines on HIV, viral hepatitis and STI prevention, diagnosis, treatment and care for key populations. 2022. who.int.
14. World Health Organization. Opioid overdose: fact sheet. who.int.
15. United Nations Office on Drugs and Crime. World Drug Report 2024. unodc.org.