Shared Decision Making in 10 Minutes: How to Present Risk, NNT, and Treatment Options to Patients Without a Statistics Degree

A 58-year-old man sits in your consulting room. His QRISK3 is 14%. His LDL-C is 3.8 mmol/L. He has no symptoms. You want to start atorvastatin 20 mg. You explain that his 10-year cardiovascular risk is 14% and that statins reduce this risk by about 30%. He nods. You print the prescription. He takes it home, reads the patient information leaflet, and never fills it. Six months later, his wife tells you he stopped because "the side effects weren't worth it for a drug that only works 30% of the time." He understood the relative risk reduction as his personal probability of benefit. He thought there was a 70% chance the drug would not work for him. He was not wrong to be cautious — he was wrong to be uninformed. And the responsibility for that lies with the consultation, not the patient. Shared decision making is the clinical skill of translating evidence into the language of individual experience — so that patients can make genuinely informed choices about their own care. It is not about persuading patients to accept your recommendation. It is about giving them the information they need to make the decision that is right for them. This guide gives you the specific tools to do that in the time available in a real GP consultation.

Why Shared Decision Making Matters: The Evidence

Shared decision making (SDM) is not a soft skill or a communication nicety — it is a clinical intervention with measurable outcomes. A 2017 Cochrane review of 105 randomised controlled trials found that patients who received decision aids (structured tools to support SDM) were more knowledgeable about their options, had more accurate risk perceptions, felt less decisional conflict, and were more likely to choose treatments consistent with their values. Crucially, they were also less likely to choose elective surgery and more likely to choose conservative management — suggesting that SDM reduces overtreatment, not just undertreatment. A 2019 systematic review in the BMJ found that SDM improved medication adherence by 27% compared to standard consultations. The mechanism is straightforward: patients who understand why they are taking a medication, and who have actively chosen it, are more likely to take it.

The Three-Talk Model of Shared Decision Making

The most widely used SDM framework in UK primary care is the Three-Talk Model, developed by Glyn Elwyn and colleagues. It provides a practical structure for SDM consultations that can be adapted to any clinical scenario and any time constraint.

The Core Problem: Why Patients Misunderstand Risk

Before you can communicate risk effectively, you need to understand why patients misunderstand it. The research on health numeracy and risk perception identifies five consistent patterns of misunderstanding that affect even highly educated patients.

The Five Patterns of Risk Misunderstanding

Understanding these patterns allows you to anticipate and pre-empt misunderstanding. The most important correction is the first: always present absolute risk reduction, not just relative risk reduction. This single change — from "statins reduce your risk by 30%" to "statins prevent 1 heart attack per 50 patients treated over 5 years" — transforms the quality of the consultation.

Presenting Absolute Risk: The Tools That Work

Research on risk communication consistently identifies three formats that improve patient understanding of absolute risk: natural frequencies, icon arrays, and comparative risk statements. Each has specific strengths and is appropriate for different clinical contexts.

Natural Frequencies: The Most Powerful Format

Natural frequencies express risk as a count out of a defined population — "3 in 100 people" rather than "3%." A landmark study by Gerd Gigerenzen and colleagues demonstrated that natural frequencies improve correct risk interpretation by 4-fold compared to percentages, across all levels of numeracy. The reason is evolutionary: human brains evolved to process counts of events (3 mammoths seen today) not abstract probabilities (3% chance of seeing a mammoth). Natural frequencies are the format closest to how we naturally think about risk.

Icon Arrays: Seeing the Numbers

An icon array is a grid of 100 (or 1,000) small human figures, with a subset coloured to represent the proportion experiencing an event. They are the most effective visual format for communicating absolute risk to patients with low numeracy. A 2012 meta-analysis in Medical Decision Making found that icon arrays improved risk comprehension by 40% compared to bar charts, and by 60% compared to text alone. They are particularly effective for communicating the NNT — showing 100 figures with 2 highlighted (the 2 who benefit from treatment) makes the NNT of 50 viscerally real in a way that the number alone cannot.

• Free online tools: Cates Plot (www.nntonline.net) generates icon arrays from NNT data in seconds — bookmark this for use in consultations
• NHS-approved decision aids: NICE has approved several patient decision aids that include icon arrays — available at nice.org.uk/about/what-we-do/our-programmes/nice-guidance/nice-guidelines/shared-decision-making
• Printed handouts: The Winton Centre for Risk and Evidence Communication (Cambridge) produces free, validated risk communication materials for common clinical scenarios — available at wintoncentre.maths.cam.ac.uk
• Smartphone apps: The NHS-approved "Statin Choice" decision aid and "Anticoagulation Choice" tool include icon arrays and can be used on a tablet during the consultation
• Drawing it yourself: A simple hand-drawn grid of 100 squares, with a few coloured in, is more effective than a verbal explanation — keep a notepad in your consulting room for this purpose

Comparative Risk Statements: Anchoring the Abstract

Patients struggle to evaluate an isolated risk figure because they have no reference point. A 14% 10-year cardiovascular risk means nothing without context. Comparative risk statements anchor the abstract number to something the patient already understands — either a familiar risk (driving, flying) or a comparison between treated and untreated states. Research by Lipkus and colleagues found that comparative risk statements improve risk comprehension and treatment acceptance more than absolute risk figures alone.

Explaining the NNT: The Most Honest Number in Medicine

The Number Needed to Treat (NNT) is the most honest and clinically useful statistic for patient communication — but it requires careful framing. An NNT of 50 means that 50 patients need to take the medication for 5 years for 1 to benefit. The other 49 take the medication, experience its side effects and inconvenience, and receive no direct benefit. This is not a failure of the medication — it is the nature of population-level risk reduction applied to individuals. But patients need to understand this clearly, or they will feel deceived when they discover it.

The NNT Conversation: Scripts That Work

The following scripts are adapted from validated patient communication research and can be used directly in consultations. They are designed to be honest, non-coercive, and to elicit the patient's values rather than direct them towards a predetermined conclusion.

• Opening: "I want to be really honest with you about what this medication can and can't do — because I think you deserve the full picture, not just the headline."
• Presenting the NNT: "If we treat 50 people like you with this medication for 5 years, 1 of those 50 will avoid a heart attack that they would otherwise have had. The other 49 will take the medication without that direct benefit — though they may still benefit in other ways, like lower blood pressure."
• Acknowledging uncertainty: "We can't tell in advance which of the 50 you are. The medication doesn't know either. What we do know is that your risk profile puts you in the group where the benefit is real and meaningful."
• Framing the decision: "So the question is: does preventing 1 heart attack per 50 people treated feel worth taking a daily tablet, given what you know about the side effects? There's no right answer — it depends on what matters to you."
• Handling the "what would you do?" question: "That's a fair question. In your situation, I would probably take it — because the benefit is real and the side effects are manageable for most people. But I also know that you know your own life and values better than I do, and I'd support whatever decision you make."

NNT Reference Table for Common GP Decisions

Having these numbers ready before the consultation — not calculating them during it — is what makes SDM feasible in 10 minutes. The following table provides the key NNTs for the most common GP prescribing decisions, with patient-friendly framing.

Presenting Treatment Options: The Option Grid

The Option Grid is a one-page summary of treatment options and their key attributes, designed to be used during the consultation as a shared reference point. It was developed by Glyn Elwyn and colleagues at Dartmouth and has been validated in multiple clinical settings. The key principle is that all options — including the option of doing nothing — are presented with equal visual weight, without implicit recommendation. The patient's role is to identify which attributes matter most to them.

Building an Option Grid: The Template

An Option Grid has three to five columns (one per option) and four to six rows (one per attribute). The attributes are the questions patients most commonly ask about each option — not the questions clinicians think they should ask. Research on patient priorities consistently identifies the same five questions across most clinical decisions.

Pre-built Option Grids for common clinical decisions are available free at optiongrid.org — covering hypertension, atrial fibrillation, diabetes, osteoporosis, depression, and many other conditions. These can be printed and used directly in consultations, or displayed on a screen. The Winton Centre at Cambridge also produces validated "fact boxes" — one-page summaries of treatment benefits and harms in natural frequency format — for common clinical decisions.

Eliciting Patient Values: The Questions That Matter

Presenting the evidence is only half of shared decision making. The other half is understanding what the patient values — because the same evidence can lead to different decisions for different patients, and both decisions can be correct. A patient who values longevity above all else will make a different decision about a statin than a patient who values quality of life and is unwilling to take daily medication for a small absolute benefit. Neither is wrong. Both deserve to have their values respected.

The Values Clarification Conversation

• "What matters most to you about your health at the moment?" — open-ended; invites the patient to set the agenda
• "When you think about the future, what are you most hoping to avoid?" — identifies the patient's primary fear (stroke, disability, death, loss of independence)
• "How do you feel about taking regular medication?" — surfaces medication burden concerns before they become barriers
• "Is there anything about this treatment that worries you?" — invites concerns without leading; allows the patient to raise side effect fears
• "How important is it to you to avoid side effects, even if it means a slightly higher risk of [the condition]?" — explicitly surfaces the benefit-harm trade-off
• "Is there anything in your life at the moment that would make it difficult to [take daily medication / make lifestyle changes / attend appointments]?" — identifies practical barriers
• "Would you like to take some time to think about this, or talk to someone else before deciding?" — normalises deferral; reduces decisional pressure

Recognising and Responding to Decisional Conflict

Decisional conflict is the state of uncertainty about which course of action to take when the choice involves risk, loss, or challenge to personal values. It is normal, expected, and not a sign of patient irrationality. The Decisional Conflict Scale (DCS) is a validated 16-item questionnaire that measures decisional conflict — but in a busy GP consultation, the following three questions provide a rapid clinical assessment.

• "Do you feel you have enough information to make a decision?" — if no, provide more information or a decision aid
• "Do you feel clear about which option fits best with what matters to you?" — if no, explore values further
• "Do you feel supported in making this decision?" — if no, offer to involve family, provide written information, or arrange a follow-up appointment

If a patient is in significant decisional conflict, the right response is not to push for a decision in the current consultation. Deferral is a legitimate clinical outcome. "Let's give you some time to think about this — here's a summary of the options, and we can talk again at your next appointment" is not a consultation failure. It is good clinical practice. Decisions made under pressure are less likely to be adhered to and more likely to be regretted.

Special Scenarios: SDM in Complex Clinical Situations

Scenario 1: The Patient Who Refuses Evidence-Based Treatment

A 65-year-old woman with CHA2DS2-VASc score of 4 refuses anticoagulation for AF because her neighbour had a "terrible bleed" on warfarin. She has capacity. She understands the stroke risk. She has made an informed decision. What do you do? The answer is: you respect her decision, document it clearly, and ensure she has the information she needs to change her mind if she wishes. You do not prescribe anticoagulation without consent. You do not repeatedly pressure her at every appointment. You do ensure she knows that DOACs have a significantly lower bleed risk than warfarin, and that her stroke risk (approximately 4% per year) is substantially higher than her bleed risk on a DOAC (approximately 1.5% per year). You offer to revisit the decision at any time. And you document: "Patient informed of stroke risk (4% per year), bleed risk on DOAC (1.5% per year), and the net benefit of anticoagulation. Patient declines anticoagulation at this time. Decision respected. Will review at next appointment."

Scenario 2: The Patient Who Demands Treatment You Would Not Recommend

A 45-year-old man with QRISK3 of 4% has read about statins online and wants to start one. He has no symptoms, no family history, and his lipids are normal. The evidence does not support statin therapy at this risk level (NNT approximately 500 over 5 years). How do you respond? You present the evidence honestly: "At your current risk level, out of 500 people like you taking a statin for 5 years, 1 would avoid a heart attack. The other 499 would take the medication without that direct benefit. Given that, I don't think the benefit outweighs the small but real risk of side effects — but I want to understand what's driving your concern." This opens a conversation about the patient's underlying anxiety (family history? recent bereavement? health anxiety?) that is more clinically useful than simply prescribing or refusing.

Scenario 3: The Patient with Low Health Literacy

Approximately 43% of adults in England have health literacy below the level needed to understand standard health information. For these patients, standard risk communication tools (percentages, NNTs, confidence intervals) are not just unhelpful — they are actively harmful, creating the illusion of informed consent without the reality. The following adaptations make SDM accessible to patients with low health literacy.

• Use the "teach-back" method: "I want to make sure I've explained this clearly — can you tell me in your own words what you understand about the options?" This identifies misunderstanding without blame.
• Use visual aids exclusively: Icon arrays, simple diagrams, and physical demonstrations (e.g., showing 100 coins, with 3 coloured differently) are more effective than any verbal explanation for patients with low numeracy.
• Avoid medical jargon entirely: "Heart attack" not "myocardial infarction"; "blood clot in the brain" not "ischaemic stroke"; "muscle pain" not "myopathy".
• Use the "1 in X" format rather than percentages: "1 in 50 people" is more concrete than "2%".
• Involve a family member or carer if the patient consents — a trusted person who can help process information after the consultation significantly improves decision quality.
• Provide written information at the appropriate reading level — NHS patient information leaflets are written at a Grade 8 reading level; the Winton Centre materials are written at Grade 6.

Scenario 4: The Time-Pressured Consultation

The most common objection to SDM in primary care is time. "I have 10 minutes — I can't do a full shared decision making consultation." This objection is understandable but overstated. The evidence shows that SDM consultations are not significantly longer than standard consultations when GPs are trained in the approach. The key is preparation: knowing your numbers before the consultation, having decision aids available, and using a structured framework (Three-Talk) rather than improvising. The following time-efficient SDM protocol can be completed in 8 minutes.

Decision Aids: The Tools That Do the Work for You

Patient decision aids are evidence-based tools that present options, benefits, and harms in a structured, accessible format. They are not a replacement for the consultation — they are a preparation for it. Patients who review a decision aid before the consultation arrive better informed, ask better questions, and make decisions more consistent with their values. A 2017 Cochrane review of 105 trials found that decision aids reduced decisional conflict, improved knowledge, and reduced the proportion of patients who were passive in decision making.

Decision Aid Resources for UK GPs

Documenting Shared Decision Making: The Legal and Clinical Standard

Post-Montgomery, documentation of the SDM process is not optional — it is a medicolegal necessity. The documentation does not need to be lengthy, but it must demonstrate that the patient was informed of the material risks and alternatives, and that their decision was voluntary and based on adequate information. The following template provides a concise, legally adequate SDM record.

SDM Documentation Template

• Options discussed: [List all options presented, including no treatment]
• Benefits presented: [Absolute risk reduction and NNT in natural frequency format]
• Harms presented: [Key harms in natural frequency format, including NNH]
• Patient understanding confirmed: [Teach-back or equivalent]
• Patient values elicited: [Brief summary of what the patient said mattered to them]
• Decision aid used: [Name of decision aid, if used]
• Decision reached: [Patient's decision, or decision deferred to follow-up]
• Patient capacity: [Confirmed — patient has capacity to make this decision]
• Example: "Options discussed: statin vs lifestyle vs no treatment. Absolute benefit: 3 in 100 avoid MI over 5 years (NNT 33). Key harm: 1 in 200 develop myopathy. Patient understands and confirms. Patient values: wants to reduce risk for family reasons. Decision: start atorvastatin 20 mg. Capacity confirmed."

The Deeper Purpose: Why SDM Makes You a Better Doctor

Shared decision making is sometimes framed as a patient rights issue — something you do for the patient. But the evidence suggests it is also something you do for yourself. GPs who practise SDM report higher job satisfaction, lower rates of burnout, and fewer complaints. The reason is structural: SDM shifts the locus of responsibility for the decision from the GP to the patient-GP partnership. When a patient makes an informed decision and the outcome is not what either of you hoped for, the consultation record shows that the patient understood the risks and chose freely. This is not about avoiding blame — it is about practising medicine in a way that is honest, respectful, and sustainable.

There is also a deeper clinical reason. The patients who benefit most from treatment are not always the patients who accept it most readily. The patient who refuses a statin because of a neighbour's experience with warfarin has conflated two different drugs — but her underlying concern (fear of medication side effects) is real and valid. The SDM consultation that surfaces this concern, addresses it with evidence, and gives her the information to make a genuinely informed decision is more likely to result in treatment adherence than the consultation that simply prescribes and moves on. SDM is not slower medicine. It is more effective medicine.

Quick Reference: The SDM Consultation Checklist

• 1. Before the consultation: Know your numbers — absolute risk, NNT, NNH for the decision at hand. Have a decision aid available (printed or on screen).
• 2. Team Talk (60 seconds): Signal that a decision needs to be made. Invite the patient into the process. "There are a few options — I'd like to go through them with you."
• 3. Option Talk (3–5 minutes): Present all options including no treatment. Use natural frequencies, not percentages. Present both benefits and harms. Use a visual aid if available.
• 4. Values elicitation (1–2 minutes): Ask one open values question. Listen. Reflect back. "So it sounds like what matters most to you is..."
• 5. Decision Talk (2 minutes): Check understanding (teach-back). Elicit preference. Agree a decision or defer. "Does one option feel right, or would you like time to think?"
• 6. Documentation (1 minute): Record options discussed, benefits and harms presented, patient understanding confirmed, values elicited, decision reached, capacity confirmed.
• 7. Follow-up: If decision deferred, arrange follow-up. Send decision aid if not already provided. Document plan.