What incidence counts: new cases in a period
Incidence has a three-part definition: the number of new cases recorded in a population over a stated period, usually a year.
It counts new cases only — someone diagnosed three years ago and still unwell adds nothing to this year's figure. And it is per population and per period: "4000 cases" means little until it is "4000 a year per million".
It measures what gets recorded — the cases the system detects, not the illness present.
Two branches: rate change vs counting change
Sort every reason into one of two branches; using both reaches the top band.
Branch 1 — the rate changed. More people genuinely develop the problem: new life stressors, or greater availability of an addictive activity.
Branch 2 — the counting changed. The same illness is recorded more often: revised ICD criteria, falling stigma, better access, or a new counting method.
Two reasons from each branch, not a long list from one, is balance.
Use reasons that fit the named problem
The block names two problems, so a question names one — not generally.
Depression. The recorded figure is very sensitive to detection and disclosure: milder cases are missed, so wider criteria or less embarrassment raise it. Real change comes from population stressors.
Addiction. The figure responds to what counts as addiction and to availability: as a behaviour becomes recognised as a disorder, cases once described otherwise get counted.
Drawn from real examiner reports.
Give a conclusion, not just the result
Incidence questions bring a table or graph, the likeliest place to lose marks.
A result is what the figures say — new cases rose. A conclusion says what that means: more cases are recorded, though the rise may reflect diagnosis and reporting changes, not a real increase. A conclusion needs both halves — the figures, then their limit.
Examiner reports record this mis-read: an answer that was "a result, not a conclusion" scored zero (June 2023 P2 Q1f), and the same confusion recurred (June 2024 P1 Q5a, Q11a). No item is set directly on 3.1.2, but trend data is where the error appears.
Incidence vs prevalence: new vs all
Imprecision is punished. Incidence = the new cases in a population over a stated period, usually a year. Prevalence = the total present at a given time, new and old. Incidence can fall while prevalence rises — fewer new cases yearly, sufferers unwell longer. "How many have depression" defines neither: say new cases, population, period.
The topic's documented definition trap is imprecise listing: on the features of addiction the scheme wanted terms such as impaired control, "withdrawal symptoms" was not creditworthy, and most scored one of two marks (June 2023 P1 Q10).
Identify a reason, then justify it
Reasons are easy to name, hard to justify, so answers stall at identification. Identification: "less stigma now." Justified: "…so more people describe symptoms instead of hiding them, so once-undiagnosed cases get recorded — raising incidence even if the same number were unwell." Shape it: change → what people or services do differently → effect on the figure.
Identification without justification is named on strength, weakness and improvement items across Paper 1, most recently June 2024 P1 Q11b and Q15b, and the June 2023 and June 2024 P2 Summaries warn candidates must not only describe.
One branch is not balance
Balance means covering both branches, not writing at length. Reasons all from "the counting changed" — stigma, criteria, access — are still one-sided, and a list of stressors and availability is one-sided the other way.
Show that a rise could be real or a counting artefact: give one reason from each branch, then say the evidence cannot yet decide between them.
A difference needs a connective
Asked for the difference between incidence and prevalence, two definitions side by side score one mark — it is only implied. A connective fixes it: incidence counts only the new cases in a period, whereas prevalence counts every case present. "Whereas" names the axis they differ on and earns the second mark.
"Describe the difference" answered as two separate definitions with no connective is recorded at June 2019 P1 Q6 and June 2022 P1 Q7; connectives such as "whereas" earn the second mark.
Percentage change: divide by the original
Incidence trends invite a percentage-change sum, and the base is the trap. Percentage change = change ÷ original × 100, and the original is the earlier year: 240 to 300 is a change of 60, so 60 ÷ 240 × 100 = 25%, not 60 ÷ 300 = 20%. Two more slips: quoting the raw rise (60), and concluding it "became more common" when only the counting changed.
Myth: a recorded rise proves higher illness
A rise fits two worlds: a real rise (more people genuinely develop the problem) or a counting rise (the same illness detected more often after wider criteria, less stigma or easier access). One set of numbers cannot tell which — a limit on the figures' validity. Only other evidence settles it: the same symptom questions asked of a comparable sample.
Grounded in the general "result versus conclusion" failure (June 2023 P2 Q1f; June 2024 P1 Q5a, Q11a) rather than a topic-specific entry; the digest sets no item on 3.1.2.
Myth: a criteria revision made people ill
A revision of the ICD changes the rule for who counts, not health: when a wider criterion takes effect nobody becomes unwell, yet more now meet the definition. The rise shows as a step at the revision year, not a steady climb, and figures before and after are not comparable. Recognising a diagnosis is not inventing the problem: people had it before it had a name.
9-mark Assess: apply, balance, conclude
The 9-mark Assess is AO1 3 / AO2 3 / AO3 3 (bands 1-3, 4-6, 7-9). AO1: define incidence and the two branches. AO2: use the scenario's figures. AO3: weigh both branches, then conclude — more cases are recorded, but the figures cannot show the true rate rose.
Write AO2 before AO3
AO2 and AO3 are marked separately, but candidates who lead with AO3 leave AO2 too thin and forfeit both. Do AO2 first — quote a figure or a stated change and say what it does to the count — then evaluate. Order is free, but AO2-first is safer.
Application means figures, not the name
AO2 credit comes from the scenario's own detail. Naming the person, or copying the stem back, is not application. Quote a figure or a stated change — "self-referral opened in Year 3" — and say what it does to the count.
Continuous prose, not bullet points
Extended 9-mark answers must be in continuous prose. Bullet-pointed essays are not accepted, however good: the marks reward a connected argument, so disconnected points cap the level. Plan in notes, but write in full linked sentences.
Incidence — the number of new cases of a mental health problem recorded in a population over a stated period of time, usually one year.
Prevalence — the total number of cases present in a population at a given time, including people who already had the problem.
Recorded incidence — the cases the system actually detects and counts. This is what every published figure is, and it is not the same thing as the amount of illness in the population.
Full notes, flashcards, Q&A and the topic quiz for every premium subject.
Premium plans are US$8.99/month or US$49.99/year — first month free.
Studying with a parent's blessing? Show them this.
