Introduction to Epidemiology

Lingbin Kong Dept. Epidemiol, JNMC

What Is Epidemiology?
Epidemios in Greek~ among the people Epi~ above Demos~ population ~ology a study / science / discipline

Objectives of Medicine
To explore etiology and pathogenesis To study natural history of disease To apply artificial interventions To evaluate their effectiveness

Components of Medicine
• Basic sciences • Clinical sciences • Preventive medicine ( public health ) • Epidemiology

Position of Epidemiology in Biomedical Sciences
• • • • • • • • • • • • • ┌──── ─────────────────────────┐ │Molecular Cell Organ / Tissue Individuals Family Community └───────────────────────────── ┘ │Mol Biology│ └─────┘ │ Cell Biology │ └──────┘ │ Physiology + Pathology│ └──────────┘ │Clinical Medicine │ └────────┘ │ Epidemiology │ └──────────┘ │

Definition of Public Health
• Organized community efforts aimed at the prevention of disease and promotion of health, which links many disciplines and rests upon the scientific core of epidemiology • Epidemiology is the basic science of public health

Definition of Epidemiology
The study of the distribution and determinants of health-related status or events, and the application of this study to control of health problems in specific population

Definition of Epidemiology (cont’d)
• Epidemiology is the study of the health of human population with its functions as follows: • To define health problems (the diagnostic discipline of public health) and set the priorities • To discover the agent, host and environmental factors • To determine the relative importance of causes of illness • To identify those sections of the population at the greatest risk • To determine principles to guide progrmas for disease control • To evaluate the effectiveness of intervention and health program and services

The Premise Underlying Epidemiology
• Disease, illness, and ill health are not randomly distributed in a population. • Each people has certain characteristics that predispose us to, or protect us against, a variety of different diseases.

Introduction
• Concept • Brief History • Uses • Causes of Disease • Distribution of Disease • Indicators for Measuring Risk

Introduction (cont’d)
• Fetus of medicine • Rapid development • Related to social and economic development • Bridging medicine and public health • Interdisciplinary science (multidisciplinary nature)
– Basic sciences – Clinical sciences – Statistics – Computer technology – Sociology

Concept of Epidemiology
• To study health of human population • To study all diseases ( health-related events ) • To study distribution of disease and its determinants • To apply its principles for disease control and health promotion • To communicate epidemiologic findings to health professionals and the public

Specific Objectives of Epidemiology
• To identify the etiology, or the cause of a disease and the risk factors • To determine the extent of disease found in the community (burden of disease in public) and set priorities for intervention • To study the natural history and prognosis of disease • To evaluate new prevention and therapeutic measures • To provide the foundation for developing public policy and regulatory decisions relating to environmental problems

Who Needs To Study Epidemiology?
• public health workers • policy-makers at all levels of government • students in schools of public health and medicine • practitioners • people involving in designing, planning, monitoring and assessing larger-scale programs and services

Brief History(cont’d)
• Sanitary statistics era • Communicable disease era • Chronic non-communicable disease era • Molecular epidemiology era

Brief History (cont’d)
• Evolution of epidemiology : • From infectious to all diseases – From “source of infection, route of transmission and susceptibility” to “agent, host and environment” • From disease to health status • From epidemic to distribution • From specialty to subspecialties ( branches ) • From univariate to multivariate • From uni-disciplinary to interdisciplinary • From qualitative to quantitative • From manual calculation to computer

Branches of the Overall Science of Epidemiology
• infectious disease, non-communicable disease, clinical, cardiovascular disease, cancer, seroepidemiology, genetic, molecular, behavioral, occupational, environemntal, perinatal, obstetric, nutritional, metabolic, etc.

Selected Milestones in the Historical Development of Epidemiology
• 400 B.C.: Hippocrates suggested that the development of human disease might be related to lifestyle factors and the external environment (On Airsm Waters, and Places) • 1600s: Francis Bacon and others developed principles of inductive logic, forming a philosophical basis for epidemiology • 1662: John Graunt analyzed births and deaths in London and quantified disease in a population

Selected Milestones in the Historical Development of Epidemiology (cont’d)
• 1747: James Lind conducted a study of treatment for scurvy, one of the first experimental trials • 1839: William Farr set up a system for routine summaries of causes of death • 1849~1854: John Snow formed and tested a hypothesis on the origins of cholera in London, one of the first studies in analytical epidemiology • 1920: Joseph Goldberger published a descriptive field study showing the dietary origins of pellagra

Selected Milestones in the Historical Development of Epidemiology (cont’d)
• 1948: The Framingham Heart Study began, the first cohort study • 1950: Richard Doll and Austin Hill, and others, published the first case-control study of cigarette smoking and lung cancer • 1954: Field Salk polio vaccine conducted, the largest formal human experiment • 1959: N. Mantel and W. Haenszel developed a statistical procedure for analysis of case-control study with stratification

Selected Milestones in the Historical Development of Epidemiology (cont’d)
• 1960: Brian MacMahon published the first textbook of epidemiology with a systematic focus on study design • 1964: The US Surgeon General’s Advisory Committee on Smoking and Health established criteria for evaluation of causality • 1971~1972: North Karelia and Stanford Five Community studies launched, the first community-based cardiovascular disease prevention program • 1970s: New multivariate statistical methods developed, such as log-linear, logistic, proportional hazard regression models

Selected Milestones in the Historical Development of Epidemiology (cont’d)
• 1970s~to date: Invention and continuing evolution of microcomputer technologies allowing linkage and analysis of large databases • 1990s: Development and application of techniques in molecular biology to large populations (genetic and molecualr epidemiology)

Uses of Epidemiology
• To study health of population • To make community diagnosis • To study etiology and causes of epidemics • To discover natural history of disease ( individual and population ) • To search for prognostic factors • To prevent and control disease with interventions

Uses of Epidemiology (cont’d)
• To make health policy and decision
– To set priorities in health issues – To determine high-risk population – To determine main risk factors in population – To formulate hygienic standards – To determine strategy in disease control – To provide adequate health service based need and demands ( resource allocation )

• To evaluate effectiveness of interventions (to test new treatments)

Evaluation for Typhus Vaccine in Hospitalized Patients
• ━━━━━━━━━━━━━━━━━━━━━━━━━ • Immunization No. Cases No. Death Case-fatality • history with typhus (%) • ───────────────────────── • • Yes No 197 1 907 2 97 1.02 5.09

• Total 2 104 99 4.71 • ━━━━━━━━━━━━━━━━━━━━━━━━━

Evaluation for Typhoid Vaccine in Field Population
• Immunization • • • • Yes No Total No.Subject No.Cases Attack rate (1/106) with typhoid 10 40 50 93 410 240

108 000 98 001 206 001

Evaluation for Pirenzeping (Bisvanil)
• Pirenzeping (Bisvanil, 哌吡氮平 ) for duodenal ulcer • ━━━━━━━━━━━━━━━━━━━━━━━━━ Group No. No. % No. % No. w/o % No. % • case cure alleviantacid mouth • ated dryness • ─────────────────────────── • Pirenzeoung 47 • Placebo 42 27 19 57.4 45.2 31 16 66.0 38.1 35 30 74.5 71.4 23 46.9 13 31.0

• total 89 46 51.7 47 52.8 65 73.0 • ━━━━━━━━━━━━━━━━━━━━━━━━━━━ • * Cure rate of 70 ~ 79% in literature (diagnosed by endoscopy)

Evaluation for Whooping Cough Vaccine
• (British Medical Council, 1946) • ━━━━━━━━━━━━━━━━━━━━━━━━━━ • Group No. subj. No. dis. % No. non-dis. % Efficacy(%) • ────────────────────────── • Inoculated 3801 149 3.92 3503 93.08 78.56

• Not inocul. 3757 687 18.28 2383 81.72 • ━━━━━━━━━━━━━━━━━━━━━━━━━━ • * χ2=86.78, P<0.001

Examples of Etiological Studies
• Scurvy & orange and lemon • Scrotal skin cancer & soot cleaner • Cholera & water contamination • Mottled teeth & high fluorine in water • Minamata disease & methylmercuric • Ita-ita disease & cadmium • Phocomelia & thalidomide • Congenital defects & rubella infection • Retrolental fibroplasia & pure oxygen inhalation • Leukemia & radiation

Examples of Etiological Studies (cont’d)
• Lung cancer & smoking • Coronary heart disease & high blood cholesterol and sugar, hypertension, smoking, lack of physical activities • Vaginal adenocarcinoma & estrogen • Allergic encephalitis & rat brain tissue derived vaccine • Male infertility & crude cotton seed oil • Dermatitis & toxic hair of caterpillar • Breast enlargement & Fusarium -contaminated buckwheat • Botulism & fermented flour and soy paste

Examples of Etiological Studies (cont’d)
• Extrapyramidal abnormality & phenothiazines • Hepatitis A & blood clam • Periconceptional use of folic acid can reduce a woman’s risk of having a baby with a neural tube defect • Lung cancer & smoky coal burning in unvented indoor firepits • Excess of neoplastic, respiratory and vascular deaths (51%, 31% and 15%) & tobacco, which will kill about 100 million of the 0.3 billion males now aged 0~29, with half of these deaths in middle age and half in old age

Surveillance of Disease
• Occurrence of disease • ↓ • Recognition (by health-care providers) • ↓ • Notifying to health authority • ↓ • Analysis • ┌────────┴─────────┐ • ↓ ↓ • Strategies & measures Feedback to health care providers • and their evaluation & public health administrations

Evaluation and Surveillance
• • • • • • • • ┌─←─Implementation &─←──┐ │ Modification │ ┌───┐ ┌───┐ │ Strategy │ │ Disease │ │ & │ │ Surveil- │ │Measures│ │ lance │ └───┘ └───┘ └─→───Evaluation─→───┘

Surveillance of Disease
• Can help to identify the new breakout of an illness • Can provide clues to possible causes of the conditions • Can be used to suggest strategies to control or prevent the spread of disease • Can be used to measure the impact of disease prevention and control efforts • Can provide information on the burden of illness, necessary for determining health and medical care service

Factors Affecting the Count of New Cases
• Frequency with which the disease occurs • Definition of the disease • Size of the population out of which the cases develop • Completeness of the reporting of the cases

Epidemiologic Transition
• Urbanization, Industrialization, Income rising, Education expansion, Medical care & public health improvement • Demographic pattern transition • population aging • life expectancy at birth / at 65 • fertility decline • Disease & mortality pattern: dual burden • communicable disease decline / emerge / reemerge • chronic disease emerge • Risk factors • Community health problems

Epidemiology & Prevention
• A major goal of epidemiology is to identify subgroups in the population who are at high risk for disease, then to identify the specific factors that put them at high risk • Primary prevention~ultimate goal • Secondary prevention • Population-based approach (dietary advice, smokingquitting) • High-risk approach (screening for cholesterol)

Risk Approach
• •
• • 心血管事件减少率( 1/1000 ) 人群 心血管事件发生率 (%) 血压降低 20 / 10 mmHg 低危 中危 高危 极高危 < 15 15―20 20―30 > 30 <5 5―7 8―10 > 10

降压对不同危险性人群的效果
降压治疗后

血压降低 10 / 5 mmHg

• • • •

<8 8―11 12―17 > 17

Advantages & Disadvantages of Two Strategies
• Population Approach
• Advantages

Risk Approach
Appropriate to individuals Subject motivation Physician motivation


• • • • •

Radical Large potential for for whole population Behaviorally appropriate Benefit to risk ratio is favorable Small benefit to individuals Poor motivation of subjects Poor motivation of physicians Benefit to risk ratio may be low

• Disadvantages


• • • • •

Difficulties to identify high risk individuals Temporary effect Limited effect Behaviorally inappropriate

Comparison of Two Strategies
• ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ • 人群策略 高危策略 • ─────────────────────────────── • 针对整个人群的 针对高危个体的 • 适于行为或环境改变 不适于行为方面改变 • 效应是根本性的 效应是暂时性的 ( 权宜之计 ) • 群体效益好 效应范围小 • 对个体效益较小 个体受益 • 效益 (benefit) 与危险性 (risk) 之比高 • 个体积极性低 可以调动个体积极性 • 卫生人员 ( 医生 ) 积极性低 卫生人员 ( 医生 ) 积极性高 • 较舒服 不很舒服 • 创伤少 创伤较大 • 费用低 费用较高 , 但可承受 • 确定高危个体有困难 • ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Epidemiology & Clinical Practice
• Diagnosis is population-based, that is based on correlation of the auscultatory findings with findings of surgical pathology or autopsy in a group of patients • Prognosis is also based on population • Selection of therapy is also based on population

Epidemiology & Laboratory Study
• Biomarkers for exposure • Case definition • Genetic susceptibility markers • Early markers for disease • Genetic epidemiology • Molecular epidemiology • Seroepidemiology

Epidemiologic Approach
• Epidemiologic Reasoning • • • • • to determine whether an association exists between a factor and development of the disease in question to derive appropriate inferences regarding a possible causal relationship from the patterns of association to begin with descriptive data (disease distribution) (shoe-leather epidemiology~door to door direct inquiry) further to demonstrate a causal relationship from observational data to preventive action

Causes of Disease
• Etiological factors • Risk factors • Exposure : health-related – Environmental • External • Internal – Personal characteristics • Modifiable • Un-modifiable • Confounders • Effect modifiers

Genetic Cause of Disease
• • • • • • • • • Specified by a single gene Autosomal dominant Autosomal recessive X-linked (recessive or dominant) Y-linked Mosaicism Digenic diallelic Digenic trialleic Disorders caused by mutant allele(s) at a specific genetic locus

Genetic Cause of Disease
• • • • • • • • • • • • • Meiosis and mitosis errors Non-disjucction (e.g., trisomy 21) Reciprocal translocation X-autosomal trnaslocation Robertsonian translocation Inversions Abnormal chromosomes (deletion) Non-Medelian inheritance Imprinting Reciprocal duplication and eletion Mitochondrial disorders Comprelex disorders Multifactorial

Dynamic of Disease Transmission
• Interaction of host, agent and environment (epidemiologic triad), a vector often involved • Interaction of genetic and environmental factors • Disease can be transmitted in a direct or indirect fashion

Epidemiologic Triad
Host

• Environment

Agent

Modes of Disease Transmission
• Horizontal • • • • common vehicles (indirect) single, multiple, continuous exposure person-to-person (direct) vector

• Vertical (mother-to-infant): via placenta, birth canal, breast feeding

Horizontal Transmission
• droplet, nucleus of droplet, dust-borne • water-borne • food-borne • soil-borne • arthropod-borne (insect vector-borne) • contact, direct & indirect (sex, skin, bite, etc.) • blood-borne • iatrogenic • mixed

Modes of Disease Transmission (cont’d)
• • • • • • • • • • • Spectrum of diseases: clinical vs. subclinical apparent vs. inapparent carrier status endemic epidemic pandemic outbreak determinants of disease outbreak herd immunity incubation period

Process of Epidemic
Reservoir
Natural & Social Factors

Route of Transmission

Susceptible Population

Study Methods
• Observational
– – – – – Cross-sectional Ecological ( Correlational ) Case-control Cohort Proportionate mortality ratio

• Experimental
– Randomized clinical trials – Community-based / field intervention trials
• Individual • Population/Environment

– Quasi- / Semi-experiment

• Theoretical
– Mathematical model – Methodology

Logical Reasoning in Etiological Studies
• Observation of Phenomena • Formulating Etiological Hypothesis • Corroboration ( Verification ) of Hypothesis

Distribution of Disease
• Person~who • Place~where • Time~when • Dynamic

Host Factors That May Be Associated with Increased Risk of Human Disease
• Age • Race • Customs • Genetic profile • Family background • Immune status • Psychological Sex Religion Occupation Marital status Previous disease Stress Behavior

Environmental Factors That May Be Associated with Increased Risk of Human Disease
• Biological agents (bacteria, virus) • Chemical agents (air pollution, poison, alcohol, smoke) • Physical agents (temperature, humidity, altitude, trauma, radiation, noise, fire) • Nutritional (lack or excess) • Social (crowding, housing, neighborhood, water, food, health care provision)

Personal Characteristics
• • • • • • • • • • • • Age Gender Race / ethnic Marital status Maternal age Genetic susceptibility Biological markers Socioeconomic status Occupation Religion Behavior Personality / psychological

Geographic Characteristics
• Administration-based worldwide, the continent, nationwide, provincial, prefecture, county, township, village, municipality (city), district, sub-district, community, neighborhood • Geography-based plain, plateau, mountainous, forest, grassland, hills • Urban and rural • Endemic • Imported

Temporal Characteristics
• Century, decade, year, season, month, week, day, hour • Short-term fluctuation • Secular trend • Seasonal change • Periodicity

Involuntary Risks of Death in US (per person per year)
• • • • • • • • • • • • • Struck by automobile 1 in 20 000 Floods 1 in 455 000 Earthquake 1 in 588 000 Tornados 1 in 455 000 Lightning 1 in 10 million Falling aircraft 1 in 10 milliom Release from an atomic power station at site boundary 1 in 10 milliom at 1 km 1 in 10 milliom Bites of venomous creatures 1 in 5 milliom Leukemia 1 in 12 500 Influenza 1 in 5 000 Meteorite 1 in 100 billion

Voluntary Risks of Death in US (per person per year)
• • • • • • • • • • • • • • • • • Smoking Drinking Soccer, football Automobile racing Automobile driving Motorcycling Rock climbing Taking OC Power boating Canoeing Horse racing Aamteur boxing Professional boxing Skiing Pregnancy Abortion (< 12 wks) Abortion (>14 wks) 1 in 200 1 in 13 300 1 in 25 500 1 in 1 000 1 in 5 900 1 in 50 1 in 7 150 1 in 5 000 1 in 5 900 1 in 100 000 1 in 740 1 in 2 million 1 in 14 300 1 in 430 000 1 in 4 350 1 in 50 000 1 in 5 900

Migrant Study
• Rationale for comparison and inference (rate of disease) • ━━━━━━━━━━━━━━━━━━━━━━━ ━ • Explanation CO M CA • ──────────────────────── • Environmental M≠CO M≈CA • Genetic M≈CO M≠CA • ━━━━━━━━━━━━━━━━━━━━━━━ ━ • CO = country of origin • M = migrants • CA = native of the country of adoption

Example of Migrant Study
• Prevalence of hypertension ( % ) in Yi farmers, Yi migrant and Han urban residents in Liangshan, Sichuan, 1986 • ━━━━━━━━━━━━━━━━━━━━━━━ ━━ • Hypertension Borderline HBP • Group ───────── ───────── • men women men women • ───────────────────────── • Yi farmer 0.50 0.31 2.19 2.11 • Yi migrants 3.20 1.09 6.40 2.29 • Han residents 3.82 2.31 5.66 3.76

Example of Migrant Study
• Blood pressure ( mmHg ) in Yi farmers, Yi migrant and Han urban residents in Liangshan, Sichuan, 1986
• • • • • • • • • • • • ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ ━━━━ SBP DBP Group Gender ──────────── ──────────── Mean Increment / yr Mean Increment / ye ─────────────────────────────────── Yi farmer M 110.8 0.13 66.3 0.23 F 111.2 0.06 65.8 0.14 Yi migrant M 113.0 0.33 70.9 0.33 F 106.0 0.37 65.2 0.23 Han resident M 114.5 0.36 72.6 0.23 F 108.1 0.56 67.7 0.36 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ ━━━━

Birth-Cohort Analysis

• • • • • • • • • • • • • • • • • • •

Age- and sex-specific prevalence (1/ 106) of congenital deaf-mutism in the Census of 1911, 1921 and 1933 in New South Wales, Australia
━━━━━━━━━━━━━━━━━━━━━━ Year of Census Age (yrs) ──────────────── 1911 1921 1933 ────────────────────── 0~ 1.6 1.7 1.1 5~ 5.9 7.2 9.5 10 ~ 11.1 8.6 8.9 15 ~ 6.4 5.7 14.0 20 ~ 6.5 11.5 9.0 25 ~ 6.0 5.9 6.9 30 ~ 5.4 6.7 14.1 35 ~ 5.7 6.2 7.1 40 ~ 3.6 4.3 5.9 45 ~ 3.2 8.0 5.2 50 ~ 2.8 2.6 4.7 55 ~ 1.4 2.4 4.0 60 ~ 3.8 7.6 6.8 ━━━━━━━━━━━━━━━━━━━━━━

Birth-Cohort Analysis
• Years at high prevalence of congenital deaf-mutism in the Census of 1911, 1921 and 1933 in New South Wales, Australia

• ━━━━━━━━━━━━━━━━━━━━━━━━━━ • Year of Census • 1911 1921 1933 • ────────────────────────── • Age (yrs) at high prevalence 34 • Year of birth • ~ 1901 ~ 1901 ~ 1903 1899 ~ 1901 1899 • Common year of birth • Year of rubella epidemic 10 ~ 14 1897 20 ~ 24 1899 30 ~

1897

Indicators for Measuring Risk
• Absolute (rate, risk) • Proportion (percentage) • Relative (ratio) • Attributable

Absolute Indicators
• Incidence rate • Attack rate
– Secondary attack rate

• Mortality rate (death rate) • Cumulative mortality • Prevalence
– infection rate, positive rate – point vs period

• Case-fatality ratio • Survival rate
– Direct method – Life-table method

Some Considerations for Incidence
• Knowledge of the health status of study population • Time of onset • Specification of numerator: number of persons vs. number of conditions • Specification of denominator • Period of observation • Person-time denominator: unequal periods of observation • Constant k

Incidence Rate
• Number of new cases of a disease over a period of time

• ───────────────── × k

Population at risk


Attack Rate
Number of persons ill with the disease % Number of persons exposed

• ───────────── ×100

General Sources of Morbidity Statustucs
• Disease Reporting : communicbale disease, cancer regsitry • Data Accumulated as a by-product of insurance and prepaid medical care plans : group health and accident insurance, prepaid medical care plans, state disability insurance plans, life insurance companies, hosiptal insurance plans (Blue Cross), Railroad Retirement Board • Tax financed public assistance and medical care plans : public assistance, aid to the blind, aid to the disabled, State or federal medical care plans, armed forces, Veterans Administration

General Sources of Morbidity Statustucs
• Hospitals and clinics • Absenteeism records (industry and schools) • Pre-employment and periodica physical examinations in industry and schools • Case-finding programs • Selective service records • Morbidity surveys on population samples (National health Survey, National Cancer Surveys)

Mortality Rate
Number of deaths during the year • ──────────────── × k • Average (mid-year) population

Risk or Cumulative Incidence
• The proportion of unaffected individuals who, on average, will contract the disease of interest over a specified period of time. • Is estimated by observing a particular population for a defined period of time--the risk period. • The estimated risk ( R ) is a proportion, the numerator is the number of newly affected persons ( A ) , called cases by epidemiologists, and the denominator is the size ( N ) of unaffected population under observation. • • New cases A • R= =

Persons at risk

N

Case-fatality Ratio
• Number of deaths from a disease • ─────────────────×100% • Number of clinical cases of that disease

Prevalence Rate
Number of existing cases of a disease at a point of time • ─────────────────────── × k • Total population

• Point vs. Period prevalence • Life-time prevalence • Prevalence ≈ incidence ×duration of disease

Disease Rates
• Overall vs. Specific • Crude vs. Standardized ( Adjusted )

Adjustment ( Standardization ) of Rates
• Direct Method

• • • • • • •

Crude mortality rates in areas F and A ━━━━━━━━━━━━━━━━━━━━━━━━━━ Areas F A ────────────────────────── Population 12,335,000 524,000 No. deaths 131,044 2,064 Crude death rate ( 1/105 ) 1,062.4 393.9 ━━━━━━━━━━━━━━━━━━━━━━━━━━

Adjustment (Standardization) of Rates (cont’d) • Direct Adjustment for Rates
• • • • • • • • • • • • Age-specific mortality ( 1/105 ) F <5 5~ 20 ~ 45 ~ >65 total adjusted rate ( 1/105 ) 812.0 284 A 274 57 198 815 4425 4350 65 188 629 30.4 245.7 18.3 Reference population Expected No. deaths ( ×106 ) F 52,000 52.9 98.1 46.0 1,345,000 1,996,000 A 50,000 30,000 34,000 194,000 184,000 375,000 289,000 1,322,000 1,879,000

Age Groups

764.4

Adjustment ( Standardization ) of Rates (cont’d) • Indirect Adjustment of Rates

• • • • • • • • • • Age group A <5 5~ 61 20 ~ 388 45 ~ 674 >65 total Age-specific mortality in reference population F A 251.1 0.85 47.2 161.8 841.9 5104.8 882.0 2.20 0.02 0.06 2,134 2.28 0.13 4.41 2.60 0.24 0.08 151 1,076 7,135 21,889 1,021 2,295 Population ( ×106) ( 1/105 ) Expected deaths F

112,305 144,539

Relative Indicators
• Ratio – Relative risk ( RR ) – Relative odds ( OR ) – SMR ( SIR ) – MOR • Proportion or percentage – PMR – PCMR • Difference • Standardized ( Adjusted ) rates

Attributable Indicators
• Attributable risk proportion • Population attributable risk proportion • Attributable fraction ( etiologic fraction )

Terms Describing Epidemics
• Sporadic • Epidemic • Pandemic • Outbreak • Clustering

Large-scale Epidemiological Studies
• Community intervention trials of fluoride supplementation in water to prevent dental caries during 1940’s • Framingham Heart Study initiated in 1948 • Salk vaccine field trial in 1954 • Smoking and health studies during ’50s~’60s

Other Studies Attracting Public Attention
• • • • • • • • • • • • • • • • the efficacy of oral anti-diabetic medication the effect of diethylstilbestrol on offspring clustering and infectious transmission of Hodgkin’s disease reserpine and breast cancer Legionnaires’ disease low-level ionizing radiation and leukemia saccharin and bladder cancer swine flu vaccine and Guillain-Barre syndrome hormone drugs in pregnancy and birth defects tampons and toxic-shock syndrome hazardous waste disposal sites replacement estrogen therapy and endometrial cancer coffee drinking and pancreatic cancer passive smoking and lung cancer Agent Orange in veterans from Vietnam acquired immune deficiency syndrome

Cross-sectional Study
• Concept • Uses • Subject Selection • Type of Study – Census – Sampling study • Study Variables • Data Collection • Data Analysis and Interpretation • Limitations

Uses of Cross-sectional Studies
• Describe disease distribution and its determinants • Establish hygienic standards • Study sub-clinical, non-fatal changes or health effects and chronic disease • Study risk factors • Understand health level in population, e.g., nutrition, development, etc. • Understand current status in family planning and MCH • Evaluate effectiveness of intervention • Disease surveillance • Community diagnosis

Sampling Methods
• Simple random sampling • Stratified sampling – Equal proportion allocation – Unequal proportion allocation • Systematic sampling • Cluster sampling • Multi-stage sampling – Probability proportional to size sampling • Mixed sampling

Design for Cross-sectional Study
• Objectives • Subjects • Sample size • Questionnaire • Data collection • Quality assurance • Analysis

Data Analysis
• Prevalence
– Point – Period – Life-time

• Prevalence ratio
– Standardization

• Index of synergism • Etiologic fraction

Strengths and Limitations
• Simple , time- and cost-saving , easy • Etiological clues • No incidence data

Ecological Study
• Concept – Observational unit : population of group of individuals – Ecological fallacy ( inferential bias ) • Uses – Etiological clues – Disease surveillance for estimation of epidemic – Evaluation for intervention – Cumulative exposure assessment in population • Limitations – Association between exposure and disease not evaluated in individual – Confounding not be controlled – Average level unequal to individual level of exposure

Examples of Ecological Studies
• Pork consumption vs. Breast cancer • GDP vs. Overall all-cause mortality • Economic development vs. Colon cancer • Population smoking vs. Lung cancer • Alcohol consumption vs. Coronary heart dis. • Cholesterol vs. CHD • Protestant religion vs. Suicide • Oral contraceptive vs. CHD • Salt sale vs. Esophageal cancer mortality

Examples of Ecological Studies
• Asthma death vs. Anti-asthma drugs • Water hardness vs. Cardiovascular mortality • Pap smear vs. Cervical cancer mortality • Near-sighted vs. TV watching • Peptic ulcer vs. Lung cancer • Tuberculosis vs. AIDS • Saccharine consumption vs. Bladder cancer • Thalidomide vs. Seal-like deformity • Industrialization vs. Lung cancer

Steps in Investigation of An Acute Outbreak
• Define the epidemic • Examine the distribution of cases • by time, place, person • Look for combination (interaction) of relevant • variables • Develop hypothesis based on • • existing knowledge of the disease analogy to diseases of known etiology

Steps in Investigation of An Acute Outbreak
• Test hypothesis • • further analyze existed data collect additional data

• Recommend control measures • • control of present outbreak prevent future similar outbreak

Causal Inference
• Observed association? Yes • Could it be a result of chance? Probably not • Could it be due to selection or measurement bias? No • Could it be due to confounding? No • Could it be causal? Apply guidelines and make judgement

Guidelines for Causal Judgement
• Temporal relationship • Strength of the association • Dose-response relationship • Replication of the findings • Biological plausibility • Consideration of alternative explanations • Cessation of exposure (Reversibility) • Specificity of the association • Consistency with other knowledge • Study design

Case-control study

Cohort Study

Clinical Trials

Field Community Trials

Exposure Assessment

Bias

Effect Modification

Causal Inference

Survival Analysis

Life-Table Method

Kaplan-Meier Method

Log-Rank Test

Mantel-Haenszel Method

Multivariate Analysis
• Linear • Logistic • Analysis of Covariance • Cox ( Proportional Hazard Regression )

Questionnaire Design

How to Write A Scientific Paper
• 前置:题目、署名、摘要、关键词 • 主体:引言、材料与方法、结果、讨论 ( IMRaD )、小结、英语摘要、志谢 • 附录:公式、大批数据、重要照片、文献目 录、

Understanding John Snow on Cholera
• John Snow 曾在英国女皇 Victoria 分娩时为她 施行氯仿麻醉。但他对影响霍乱流行的因素 和传播途径的调查研究,使他名垂医史。 • 不完全相信传统的学说(瘴气学说)。 • 不盲目迷信权威的说法, William Farr (the
superintedndent of the Statistical Department of the Registrar General’s Office of England and Wales from 1839 to 1879) 认为霍乱流行与地势高低有关:

地势高,瘴气少;地势低,瘴气多。

Altitude and Cholera

Understanding John Snow on Cholera
• 从现象着手(观察 现象) • • ① 霍乱沿交通线传播,传播速度慢于人 的旅行速度。 ② 霍乱是通过接触病人传播的,潜伏期 平均为 24~48 小时。

Understanding John Snow on Cholera
• 根据现象提出假 设 • • • • ① 霍乱“病毒”必须有生命的,微小的、 肉眼看不见。 ② “ 病毒”可能是在肠子里生长繁殖,介 粪便传播。 ③ 生活在肮脏环境中的人得病者多。 ④ 水容易受污染。

Understanding John Snow on Cholera
• 通过调查验证假设 • • ① 调查伦敦宽街的流性情况及其与供水 的关系。 ② 调查伦敦各区的霍乱流行情况及其与 供水的关系。

Understanding John Snow on Cholera
• 病因研究的逻辑 思维方式 •
• ↓

观察现象 提出假设


验证假设

Understanding John Snow on Cholera
• 关于宽街的霍乱流行情况 • ① 将 St. James Parish1954 年秋季流行情况 与 1832 、 1848~1849 、 1853 年的情况进行 比较。 • ② 将 1854 年曼诺华广场( 9/ 万)、圣焉丁 区( 33 )、金广场( 217 )、伯伟克街 ( 212 )、圣全司广场( 16 )霍乱死亡率进 行比较。 • ③ 详细调查了该地区内可能引起霍乱流行 的因素:地势、土壤、街道、房屋、人口、 粪池、水坑、尘土、水井等。

Understanding John Snow on Cholera
• ④ 用标点地图画出病人的地区分布。 • ⑤ 调查了 St. James Parish 内病人的时间分布 , 8 月 27 日 ~9 月 2 日共登记 89 例病人,其 中 79 例是发生在 9 月 1 日 ~ 9 月 2 日, 4 例 发生于 8 月 31 日, 6 例发生于 8 月 27 日 ~30 日。

Spot Map

Epidemic Curve

Understanding John Snow on Cholera
• ⑥ 正面证据:回顾调查了 8 月 31 日 ~9 月 2 日发生的 83 例,有 73 例( 88% )都发生在离 宽街供水站不远处,另 10 例离供水站较远, 但其中 5 例经常到供水站挑水吃, 2 例儿童上 学时曾饮过供水站的水, 1 例据其父母说也可 能喝过供水站的水,只有 2 例情况不明。 • 正面证据:有几例死者虽然无证据表明其喝过 供水站的水,但可能喝过附近公共场所的饮料 ,而该公共场所的饮料恰恰是用该供水站的水 做的。

Understanding John Snow on Cholera
• ⑦ 反面证据:一家救济工厂有 535 名徒工, 只有 5 人死于霍乱,死亡率为 93.46/ 万。如果 按周围疫情( 8 月 19 日 ~9 月 30 日圣全司教 区共发生 616 例霍乱死亡,死亡率为 169.2/ 万),该厂至少有 9 人死于霍乱,调查显示该 厂有一口自备井。 • 另一家有 70 名职工的酒店,无一人死亡,因 该店有自备水井。

Understanding John Snow on Cholera
• ⑧ 正、反面证据:另宽街 38 号一家有 200 名徒工的工厂,有 18 人死于霍乱,他们均经 常喝供水站的水,死亡率为 900/ 万。而不喝 该供水站的水者,无一死亡。 • ⑨ 典型例子:
• 某先生:探亲 喝(含水)饮料 第二天发病 、第三天死亡。 • E 太太:住地并无霍乱、也多天未去宽街马车夫每 天从宽街提水回来 8 月 31 日喝水 9月2日 死于霍乱。

Understanding John Snow on Cholera
• ⑩ 干预:经过上述调查认为圣全司教区霍 乱与宽街水站有关, 9 月 7 日就说服当局, 在 9 月 8 日将取水的手把拆除。 9 月 10 日霍 乱死亡病例明显下降。 • ①① 流行曲线显示,宽街霍乱呈单峰型, 提示有共同传染来源。

Understanding John Snow on Cholera
• 分析伦敦各区霍乱死亡病例与供水的关系 ( 1849~1853 年间)
• 两家水厂( Southward & Vauxhaul 和 Lamberth )均设在 Thames River 下游,均无过滤设备,末梢水中可见杂物,提示 严重污染。 • 该两水厂供水区的霍乱死亡率为 1,276~1,622/10 万,其它地区 仅为 430/10 万,有显著性差异( u=160 , P<0.0001 )。 • 1848~1853 年间, Lamberth 公司将取水点移到 Thames River 上 游,增添了过滤设备, 1953 年 Lamberth 公司供水区霍乱死亡 率( 64.6/10 万)与 Southward 公司供水区的霍乱死亡率 ( 146/10 万)有显著差别( u=26.5 , P<0.001 )。 • 在 Southward & Vauxhaul 和 Lamberth 混合供水区,由 Southward & Vauxhaul 供水的地区霍乱死亡率为 315/ 万户, Lamberth 供水的地区为 37/ 万户,有显著性差异

Water Supply by Different Companies

Understanding John Snow on Cholera
• 从此以后,英国改进了供水设备、逐步增添 了抽水马桶,几十年后又开始了水的消毒, 霍乱再未在英国出现。

Understanding John Snow on Cholera
• 结论 • ① 在霍乱弧菌发现前 30 多年,通过朴素的流 行病学观点,根据分布的原理,同样可以较准 确地判断疾病的传播途径。 • ② 在病原体尚未清楚知道的情况下,通过切 断因果链中的任何一个环节,可以起到阻断疾 病流行的作用。 • ③ 自然状态下的疾病传播途径与实验室条件 下是不同的,例如,黑斑蚊与黄热,白蛉与黑 热病,虱子与斑疹伤寒。