Study guide
Professional Practice: Ethics, EBM/Biostatistics, Patient Safety & Public Health
Professional Practice is a cross-cutting PANCE domain (~part of the blueprint's professional-behavior/EBM tasks) testing medical ethics, informed consent, biostatistics/study design, patient safety, and public health rather than pure disease facts. Expect vignettes requiring you to pick the ethically/legally correct action or calculate/interpret a test-characteristic or study measure.
Medical Ethics & the 4 Core Principles
- ◆Autonomy: a competent adult may refuse any treatment even if life-saving; respect the informed refusal (e.g., Jehovah's Witness declining blood).
- ◆Beneficence (act in patient's benefit) vs. nonmaleficence ('first do no harm'); justice = fair distribution of resources and nondiscrimination.
- ◆Decision-making capacity is CLINICAL and decision-specific (understand, appreciate, reason, communicate a choice) — competence is a LEGAL determination by a court.
- ◆Emancipated minors (married, military, self-supporting, parenting, or court-declared) consent independently; most states allow minors to consent to STI, contraception, prenatal, and substance-use care without parental notification.
Informed Consent, Confidentiality & Surrogate Decisions
- ◆Valid informed consent requires disclosure of diagnosis, nature/purpose of intervention, risks/benefits, alternatives, and consequences of refusal — documented and voluntary.
- ◆Exceptions to consent: true emergency (implied consent), waiver by patient, therapeutic privilege (rarely), or lack of capacity with surrogate available.
- ◆Surrogate hierarchy when no advance directive (typical state order): healthcare proxy/durable power of attorney > spouse > adult children > parents > siblings; use substituted judgment (what the patient would want), then best-interest standard.
- ◆HIPAA/confidentiality exceptions (mandatory reporting): reportable communicable diseases, suspected child/elder/dependent abuse, gunshot/stab wounds, and Tarasoff duty to warn/protect identifiable third parties.
- ◆A parent CANNOT refuse clearly life-saving treatment for a child — override via emergency treatment or court order.
Test Characteristics (know the 2x2 table cold)
- ◆Sensitivity = TP/(TP+FN); a highly SeNsitive test with a Negative result rules OUT (SnNout) — best for screening.
- ◆Specificity = TN/(TN+FP); a highly SPecific test with a Positive result rules IN (SpPin) — best for confirmation.
- ◆PPV and NPV depend on PREVALENCE: higher prevalence raises PPV and lowers NPV; sensitivity/specificity are intrinsic and prevalence-independent.
- ◆Positive likelihood ratio = sens/(1−spec); negative LR = (1−sens)/spec; LR+ >10 markedly raises post-test probability, LR− <0.1 markedly lowers it.
- ◆Lowering a diagnostic threshold increases sensitivity (fewer FN) but decreases specificity (more FP).
Study Design, Risk Measures & Bias
- ◆RCT = highest-quality individual study for therapy; cohort (prospective) gives incidence and relative risk; case-control (retrospective, rare diseases) gives odds ratios only.
- ◆RR/OR >1 = increased risk, <1 = protective, and CI crossing 1.0 (or including it) = not statistically significant; for mean differences, CI crossing 0 = not significant.
- ◆Absolute risk reduction (ARR) = control event rate − treatment event rate; NNT = 1/ARR (round up); NNH = 1/attributable risk from harm.
- ◆Number needed to treat interpretation: lower NNT = more effective therapy.
- ◆Key biases: selection, recall (case-control), lead-time/length-time (screening), confounding (control via randomization, matching, multivariable adjustment).
Hypothesis Testing & Screening Principles
- ◆Type I error (α, false positive) = rejecting a true null; Type II error (β, false negative) = failing to reject a false null; power = 1−β, increased mainly by larger sample size.
- ◆p <0.05 means that IF the null hypothesis were true, the probability of observing a result at least as extreme as this one is <5%; statistical significance ≠ clinical significance.
- ◆USPSTF Grade A/B = recommend (offer/provide); Grade C = selectively offer; Grade D = recommend against; Grade I = insufficient evidence.
- ◆High-yield USPSTF screens: colorectal cancer start age 45 (Grade B for 45–49); breast cancer biennial mammography start age 40 through 74 (Grade B, per the 2024 update lowering the start age from 50); lung low-dose CT for 50–80 yr with ≥20 pack-years (current or quit <15 yr); AAA one-time US in men 65–75 who ever smoked.
- ◆Effective screening requires a detectable preclinical phase and a test that improves outcomes — beware overdiagnosis, lead-time, and length-time bias.
Patient Safety & Quality Improvement
- ◆Error types: near miss (caught before harm), adverse event (harm from care), sentinel event (death/serious harm requiring immediate investigation); disclose errors honestly to patients.
- ◆Root cause analysis (RCA) is RETROSPECTIVE (after an event); Failure Mode and Effects Analysis (FMEA) is PROSPECTIVE (before implementation).
- ◆System-based ('Swiss cheese' model) approach targets latent process failures, not individual blame — culture of safety encourages nonpunitive reporting.
- ◆QI cycle = Plan-Do-Study-Act (PDSA); use SMART aims and run/control charts to track measures.
- ◆High-reliability tools: read-backs, time-outs (Universal Protocol) to prevent wrong-site surgery, medication reconciliation, and standardized handoffs (e.g., SBAR).