Risk factor Study Guide

📖 Core Concepts Determinant – Any variable linked to higher disease risk; in public‑health language it’s a broad, often societal factor (e.g., poverty). Risk factor – An individual‑level exposure that raises disease likelihood (e.g., low vitamin C intake). Risk marker – A measurable variable associated with disease, but modifying it does not necessarily change disease risk. Correlation ≠ Causation – An observed association does not prove the exposure causes the outcome; causality needs extra evidence (biologic plausibility, temporality, etc.). Relative Risk (RR) – Ratio of disease probability in exposed vs. unexposed groups. Odds Ratio (OR) – Ratio of odds of disease between groups; useful for case‑control studies. Hazard Ratio (HR) – Ratio of event rates over time between groups (survival analysis). Confounder – A third variable that is linked to both exposure and outcome, distorting the true association. --- 📌 Must Remember Determinant vs. Risk factor: Determinant = broad, often non‑modifiable; Risk factor = individual, potentially modifiable. Risk marker ≠ Risk factor – Changing a marker may not alter disease risk. RR > 1 → exposure increases risk; RR < 1 → protective. OR ≈ RR only when outcome is rare (<10 %). HR incorporates time; a HR of 1.24 means a 24 % higher instantaneous event rate. Confounding control: stratify, multivariable adjustment, or randomization. Correlation → test with statistical methods (e.g., chi‑square, regression) plus biological evidence to infer causality. --- 🔄 Key Processes Assessing an Association Define exposure and outcome. Choose appropriate study design (cohort → RR/HR, case‑control → OR). Compute measure (RR, OR, HR). Test statistical significance (confidence interval, p‑value). Evaluate for confounding (stratify, adjust). Integrate biological plausibility → judge causality. Controlling Confounders Identify potential confounders (age, sex, socioeconomic status, etc.). Stratify data by confounder levels → compare within strata. Use multivariable regression to adjust simultaneously. Screening Using Risk Factors Verify risk factor is causal (solid evidence). Develop risk‑prediction model → set threshold → target high‑risk individuals for screening. --- 🔍 Key Comparisons Determinant vs. Risk factor Determinant: societal, often non‑modifiable (poverty). Risk factor: personal, potentially modifiable (smoking). Risk factor vs. Risk marker Risk factor: changing it changes disease risk. Risk marker: only signals risk; altering it may not affect outcome. Relative Risk vs. Odds Ratio RR: direct probability ratio, used in cohort studies. OR: odds ratio, used in case‑control; approximates RR only for rare diseases. Correlation vs. Causation Correlation: statistical association; may be spurious. Causation: exposure directly contributes to disease (requires additional evidence). --- ⚠️ Common Misunderstandings “RR = OR” – True only when disease incidence is low; otherwise OR overestimates effect size. “A significant p‑value proves causation” – Significance shows association, not causality. “All determinants are risk factors” – Determinants can be macro‑level (policy, environment) not directly actionable at the individual level. “A risk marker can be used for prevention” – Changing a marker alone may not reduce disease risk. --- 🧠 Mental Models / Intuition “Domino chain” – Think of a determinant as the board that sets the stage; a risk factor is the first domino you can push; a risk marker is just a painted domino that tells you a domino fell elsewhere. “Lens of time” – RR is a snapshot (probability now), HR is a movie (rate over time). “Confounder as hidden traffic” – When you measure travel time (exposure → outcome), traffic (confounder) can make the route look faster or slower; you must account for traffic to see the true effect. --- 🚩 Exceptions & Edge Cases Rare disease: OR ≈ RR → can report OR as effect size. High baseline risk: HR may diverge from RR if hazard rates change over follow‑up. Multiple interacting determinants: synergy or effect modification; stratified analysis needed. Reverse causation: when disease influences exposure (common in cross‑sectional studies). --- 📍 When to Use Which Relative Risk → prospective cohort or randomized trial with incidence data. Odds Ratio → case‑control or logistic regression, especially when outcome is rare. Hazard Ratio → time‑to‑event (survival) analyses, Cox proportional hazards model. Risk marker → use for risk prediction or screening, not for intervention unless proven causal. Stratification → simple confounders (few levels) and when sample size permits. Multivariable regression → many confounders or continuous variables. --- 👀 Patterns to Recognize Age‑gradient pattern: many diseases rise sharply with age → suspect age as confounder. Dose‑response trend: increasing exposure → increasing RR/OR suggests causality. Consistency across study designs → strengthens causal inference (e.g., cohort RR matches case‑control OR). Biologic plausibility wording in questions → hints that the association may be causal. --- 🗂️ Exam Traps Choosing OR instead of RR in a cohort question → will be marked wrong unless disease is rare. Labeling a risk marker as a risk factor – distractor; remember the marker may not be modifiable. Confusing “determinant” with “risk factor” – test often expects the broader public‑health definition for determinant. Interpreting HR > 1 as “30 % more cases” – HR reflects rate, not absolute number; the correct phrasing is “30 % higher instantaneous risk.” Assuming significance = causation – look for mention of biological plausibility, temporality, or Hill’s criteria. ---