ABDPH Dental Public Health common mistakes
ABDPH Dental Public Health · Understand the traps and how to avoid them.
Computing crude rates and comparing them across populations with very different age structures instead of age-standardising before making the comparison.
Crude prevalence and incidence figures are the easiest to pull from a survey table, and candidates forget that a population with more older adults will show higher caries or periodontal disease burden regardless of risk.
What to do instead
Before any between-population comparison, identify the standard population named in the stem (or the one implied), apply direct standardisation with the provided age-specific weights, and state that the comparison is age-adjusted. If weights are not given and the age structures differ, flag the comparison as invalid rather than computing a number.
Reading a screening or diagnostic test question and reporting sensitivity and specificity when the clinical decision in the stem requires positive and negative predictive values (or vice versa).
Sensitivity and specificity are the metrics most heavily drilled, so they are retrieved automatically even when the question asks what a patient or program should expect given a stated prevalence.
What to do instead
Anchor on the question being asked: 'how well does the test detect disease' means sensitivity/specificity; 'given a positive result, what is the chance of disease' means PPV. Rebuild the 2x2 table with the stated prevalence before selecting, and note that PPV falls as prevalence falls.
Treating a screening programme as justified because the test is accurate, without checking that the condition has an accepted presymptomatic phase, an effective early intervention, and adequate follow-up capacity.
Candidates conflate test validity with screening programme validity, and exam stems often supply a highly sensitive test to distract from missing WHO/JMCF screening criteria.
What to do instead
Run the stem against the classic screening criteria explicitly: important health problem, accepted treatment, detectable latent stage, suitable test, agreed policy on whom to treat, and cost-effective. If any criterion is unmet in the stem, the programme is not justified regardless of test accuracy.
Confusing primary, secondary and tertiary prevention when classifying a described intervention, labelling a fluoride varnish programme as secondary prevention or a sealant programme as tertiary.
The categories are learned as a list rather than as a test of disease status at the time of intervention, and stems describe activities that straddle boundaries.
What to do instead
Ask one question: has disease already occurred and, if so, is there damage? No disease and exposure reduced means primary; disease present but asymptomatic and detected early means secondary; disease with damage and rehabilitation or limitation of disability means tertiary. Classify on that single axis before reading the options.
Designing or critiquing a community programme without a logic model, so the stem's activities, outputs and outcomes are mixed and the evaluation measures the wrong level.
Candidates jump to outcome indicators because they sound rigorous, then cannot show whether the programme was implemented as intended, and process failures get misread as ineffectiveness.
What to do instead
Lay out inputs, activities, outputs, short-term outcomes and long-term outcomes before choosing indicators. Pair every outcome measure with a process measure (for example, sealants placed versus sealant retention at 12 months) so implementation failure and theory failure can be distinguished.
Selecting the most clinically effective intervention in a cost-effectiveness question without checking the incremental cost-effectiveness ratio against the stated willingness-to-pay threshold or comparing against the current standard.
Candidates equate 'most effective' with 'best value' and skip the arithmetic, and stems deliberately include a dominated option that is both less effective and more costly.
What to do instead
Compute ICER as the difference in cost divided by the difference in effect between each option and the comparator, discard dominated options outright, and then compare the remaining ICER to the threshold given. State the decision rule in one sentence before answering.
Answering an outbreak investigation question by moving straight to control measures while skipping descriptive epidemiology and hypothesis generation.
Clinical training rewards immediate intervention, and candidates feel that waiting to characterise person, place and time is a delay rather than a step.
What to do instead
Follow the sequence in the stem: verify the diagnosis and confirm the outbreak, establish a case definition, describe cases by person, place and time, generate and test hypotheses, then implement control. If the stem asks what to do 'next', the answer is the next step in the sequence, not the final one.
Misinterpreting a confidence interval or p-value, such as calling a non-significant result proof of no effect or treating overlapping confidence intervals as proof of no difference.
Statistical significance is read as a binary truth value, and candidates do not distinguish absence of evidence from evidence of absence.
What to do instead
Read the interval as the range of plausible effect sizes: if it includes the null but also includes clinically important effects, the study is inconclusive rather than negative. Report the estimate with its interval and avoid the word 'proves'.
Confusing the roles of Medicaid, CHIP, and state or local health department safety-net programmes when answering a financing or access question, or assuming Medicaid uniformly covers adult dental care nationwide.
Coverage rules are jurisdictional and change over time, and candidates generalise from one state's benefit or from the paediatric Early and Periodic Screening, Diagnostic and Treatment (EPSDT) requirement to all populations.
What to do instead
Separate the populations: EPSDT guarantees a dental benefit for children in Medicaid, while adult dental benefits are optional and state-specific. Where a stem depends on a specific jurisdiction's benefit, answer only from information given in the stem and treat unstated coverage as unknown.
Ignoring the legal and ethical framework in a public health decision, such as recommending mandatory screening or exclusion without addressing consent, confidentiality, or the authority under which the health department acts.
Candidates focus on the epidemiological merit of the measure and treat law and ethics as peripheral, then miss options that pair an effective measure with a lawful basis.
What to do instead
For any policy or population-level intervention, check four things in the stem: the legal authority, informed consent and confidentiality, the balance of benefits and harms including stigma, and the least restrictive alternative. Select the option that satisfies effectiveness and these constraints together.