Mortality rate Study Guide

📖 Core Concepts Mortality Rate – Number of deaths ($d$) in a defined population ($p$) per unit time; usually expressed per 1,000 persons per year. Crude Death Rate (CDR) – Mortality rate from all causes; calculated as total deaths ÷ mid‑interval population. Specific Mortality Measures – Rates limited to a subgroup (e.g., sex‑specific, neonatal, infant, maternal). Denominator is the relevant subgroup size. Incidence vs. Mortality – Incidence counts new cases of disease; mortality counts deaths. Life Table – Tabular summary of a cohort’s probability of dying at each age; basis for life expectancy calculations. Gompertz–Makeham Law – Mortality = age‑independent component (Makeham) + exponential increase with age (Gompertz). Compensation Law of Mortality – At older ages, mortality curves of different groups tend to converge. Micromort – One‑in‑a‑million chance of death; useful for comparing risks of activities. Risk‑Adjusted Mortality Rate – Mortality corrected for patient‑level risk factors, enabling fair provider comparison. Weekend Effect – Observed higher mortality for patients admitted on weekends, often due to staffing/resource differences. --- 📌 Must Remember Generic formula: $$\text{Mortality Rate}= \frac{d}{p}\times c$$ (e.g., $c=1{,}000$ for “per 1,000”). Units: deaths per 1,000 (or 100,000) individuals per year. Crude vs. Specific: CDR ignores age/sex structure; specific rates isolate groups and are more informative for targeted interventions. Developed vs. Developing: Developed → 90 % of deaths are age‑related, dominated by chronic NCDs. Developing → Larger share of deaths from infectious diseases, malnutrition, preventable conditions. Key preventable under‑5 causes: malaria, respiratory infections, diarrhoea, perinatal conditions, measles. Compensation law implication: Early‑life mortality differences shrink in old age. Micromort conversion: 1 micromort = $10^{-6}$ probability of death. --- 🔄 Key Processes Calculate a Mortality Rate Obtain deaths ($d$) for the period. Get appropriate population denominator ($p$) – total or subgroup. Choose conversion factor ($c$) (1,000 or 100,000). Apply formula $ \frac{d}{p}\times c$. Construct a Life Table (simplified) Start with radix (e.g., 100,000 births). For each age interval, record deaths ($dx$) and survivors ($lx$). Compute probability of death $qx = dx / lx$. Derive remaining life expectancy $ex$ from cumulative survivorship. Sisterhood Method for Maternal Mortality Interview women about survival of sisters of child‑bearing age. Count sister deaths due to pregnancy‑related causes. Adjust for reporting bias to estimate maternal mortality ratio. Sampling in Conflict Settings Cluster sampling: Randomly select whole clusters (e.g., villages) → survey all residents. Multistage sampling: First select clusters, then stratify within clusters for better representativeness. Recognize increased sampling error when displacement is high. --- 🔍 Key Comparisons Mortality vs. Morbidity – Death count vs. disease prevalence/incidence. Crude Death Rate vs. Sex‑Specific Rate – Overall population denominator vs. male/female denominator only. Developed Countries vs. Developing Countries – Chronic NCDs dominate vs. infectious/preventable diseases dominate. Gompertz Component vs. Makeham Component – Age‑exponential increase vs. constant age‑independent risk. Raw Mortality vs. Risk‑Adjusted Mortality – Unadjusted counts vs. mortality after accounting for patient risk factors. --- ⚠️ Common Misunderstandings “Incidence = Mortality” – They measure different events; incidence does not include deaths. Using CDR for age‑specific conclusions – Crude rates mask age structure; always check age‑specific rates for policy decisions. Assuming micromorts are “small” – Even a few micromorts can be meaningful when comparing high‑risk activities. Treating survey estimates as exact – Sisterhood, orphanhood, and widowhood surveys are prone to recall and selection bias. --- 🧠 Mental Models / Intuition “Death per head per year” – Visualize each person as a tiny bucket; mortality rate tells you how many buckets leak each year. Compensation law – Think of two runners starting at different speeds; they eventually run side‑by‑side on the long downhill (old age). Gompertz–Makeham – Picture a base “background” hazard (Makeham) plus a steeply rising hill (Gompertz) as age increases. --- 🚩 Exceptions & Edge Cases Conflict / disaster zones: Vital registration may be missing; rely on surveys with higher uncertainty. Mortality displacement: Spike in deaths after heat wave or epidemic, followed by a dip (short‑term “harvesting”). Weekend Effect: Not universal – some high‑resource hospitals show no difference; interpret cautiously. --- 📍 When to Use Which Crude Death Rate – Quick population‑wide overview; not suitable for comparing populations with different age structures. Sex‑Specific or Age‑Specific Rate – When evaluating gender‑ or age‑targeted interventions. Risk‑Adjusted Mortality – Comparing hospitals or providers with differing case‑mixes. Micromort – Communicating risk of a single activity (e.g., skydiving ≈ 10 micromorts). Gompertz–Makeham model – Modeling age‑specific mortality in demographic projections. --- 👀 Patterns to Recognize High under‑5 mortality + low‑income setting → expect preventable disease dominance. Sharp rise in mortality after age 65 in developed nations → chronic NCD pattern. Temporary mortality surge + subsequent dip → mortality displacement after a disaster. Higher weekend admission mortality → possible staffing/resource limitation clue. --- 🗂️ Exam Traps Distractor: “Incidence rate = deaths / population” – wrong; incidence counts new cases, not deaths. Wrong denominator: Using total population for a sex‑specific rate → under‑ or over‑estimates. Conversion factor misuse: Forgetting to multiply by 1,000 (or 100,000) → answers off by a factor of 1,000. Assuming surveys are unbiased – sisterhood, orphanhood, and widowhood methods each have systematic biases; exam questions may test awareness. Confusing compensation law with “mortality equals at old age” – it describes convergence of rates, not identical values. ---