Epidemiological Approaches

Classification of study design based on purpose, whether the investigator assigned exposure, and the direction of exposure-outcome (Creasy & Resnik's Maternal-Fetal Medicine)

1. Descriptive Epidemiology

• Person - who gets the disease (age, sex, occupation, socioeconomic status)
• Place - where the disease occurs (geographic distribution)
• Time - when the disease occurs (secular trends, seasonal patterns)

2. Analytical Epidemiology

A. Observational Studies (no investigator-assigned exposure)

i. Ecological Studies

• Compare disease frequencies and exposure data across different geographical regions or populations as a whole (not individuals)
• Relatively inexpensive; useful for generating hypotheses
• Subject to the ecological fallacy (group-level associations may not hold at the individual level)
• Cannot establish causality

ii. Cross-Sectional Studies (Prevalence Studies)

• A snapshot - measures both exposure and outcome simultaneously in a defined population at a single point in time
• Direction of exposure and outcome is simultaneous (Exposure ↔ Outcome)
• Can estimate disease prevalence
• Cannot determine temporal sequence (which came first - exposure or disease?)
• Useful for planning health services

iii. Case-Control Studies

• Subjects are selected based on disease status (cases = disease present; controls = disease absent)
• Looks backwards to find past exposures (Exposure ← Outcome; retrospective)
• Calculates Odds Ratio (OR) as the measure of association
• Advantages: fast, inexpensive, good for rare diseases, can study multiple exposures simultaneously
• Disadvantages: cannot calculate incidence rates, prone to recall bias and selection bias

iv. Cohort Studies (Follow-up / Longitudinal Studies)

• Subjects are selected based on exposure status (exposed vs. unexposed) and followed over time for the outcome
• Direction: Exposure → Outcome
• Can be prospective (follow forward in time from exposure) or retrospective (identify exposed group after outcome has occurred using records)
• Calculates Relative Risk (RR) / Incidence Rate
• Advantages: can directly calculate incidence, good for rare exposures, less prone to bias (especially prospective), can study multiple outcomes
• Disadvantages: expensive, time-consuming (for prospective), subject to loss to follow-up

3. Experimental (Intervention) Epidemiology

A. Randomized Controlled Trial (RCT) - Gold Standard

• Eligible participants are randomly allocated to receive different interventions
• Randomization minimizes confounding and bias
• Can be double-blind (neither participant nor investigator knows the allocation)
• Provides the strongest evidence for causality
• Example: the WHO clofibrate trial for primary prevention of coronary heart disease (involving >15,000 men across Edinburgh, Prague, and Budapest)

B. Non-Randomized (Quasi-Experimental) Trials

• Participants assigned to groups without randomization
• More prone to confounding than RCTs

Types of Experimental Studies in Epidemiology:

1. Preventive trials - test whether a preventive measure reduces disease incidence (e.g., vaccine trials)
2. Risk factor trials - investigator intervenes to modify a known risk factor (e.g., lowering blood cholesterol in CHD prevention)
   • Can be single-factor (e.g., smoking cessation only) or multi-factor (e.g., MRFIT in the USA - addressing cholesterol, smoking, and hypertension together)
3. Cessation experiments - evaluate removal of a suspected causal agent. If disease falls after removal, causal hypothesis is strengthened. Example: cigarette smoking cessation and lung cancer
4. Trial of aetiological agents - directly test whether a suspected cause produces disease. Classic example: the retrolental fibroplasia (RLF) trial confirming high oxygen therapy caused blindness in premature neonates

4. Other Key Approaches

Systematic Reviews and Meta-Analyses

• Sit at the top of the evidence hierarchy
• Systematically identify, appraise, and synthesize results from multiple primary studies
• Meta-analysis statistically pools data from several studies for a more precise estimate

Biomarker-Based Studies

• Use biological markers (blood levels, genetic markers) to link exposures to health outcomes
• Supplement traditional epidemiological approaches in environmental and molecular epidemiology

Hierarchy of Evidence

Systematic Reviews / Meta-Analyses  (highest)
         ↓
Randomized Controlled Trials
         ↓
Cohort Studies
         ↓
Case-Control Studies
         ↓
Cross-Sectional Studies
         ↓
Ecological Studies
         ↓
Case Series / Case Reports          (lowest)

Ultimate Aim

1. Describe the distribution and magnitude of health problems in populations
2. Identify aetiological (risk) factors in the pathogenesis of disease
3. Provide data for planning, implementing, and evaluating prevention and treatment services

Investigation of an Epidemic (Epidemic Investigation)

Objectives of Epidemic Investigation

1. Define the magnitude of the epidemic in terms of time, place, and person
2. Determine the conditions and factors responsible for the occurrence
3. Identify the cause, source(s) of infection, and modes of transmission to determine control measures
4. Make recommendations to prevent recurrence

Steps in Investigation of an Epidemic

Note: Some steps can be done concurrently. No single rigid "cook-book" approach applies to all situations.

Step 1 - Verification of Diagnosis

• The first and most important step
• Reports may sometimes be spurious, arising from misinterpretation of signs and symptoms by the lay public
• A clinical examination of a sample of cases usually suffices (not necessary to examine all)
• Laboratory investigations wherever applicable are most useful to confirm diagnosis
• Epidemiological investigations should NOT be delayed until laboratory results are available

Step 2 - Confirmation of the Existence of an Epidemic

• Compare disease frequency during the same period of previous years
• An epidemic exists when the observed frequency exceeds the expected frequency based on past experience
• An arbitrary limit of 2 standard errors above endemic occurrence is used to define the epidemic threshold (e.g., for influenza)
• Some epidemics (cholera, food poisoning, hepatitis A) are easily and obviously recognized as common-source outbreaks
• Modern epidemics (e.g., cancer, cardiovascular disease) require careful comparison with historical data

Step 3 - Defining the Population at Risk

• A detailed, current map of the area must be obtained or prepared
• Should show natural landmarks, roads, and locations of all dwelling units
• Area may be divided into segments using natural landmarks as boundaries, with dwelling units designated by number

• A complete census of the population by age and sex should be done via house-to-house visits
• Lay health workers may be employed
• This establishes the population at risk - essential for computing attack rates
• Without an appropriate denominator, attack rates cannot be calculated

Step 4 - Rapid Search for All Cases and Their Characteristics

• A medical survey is conducted in the defined area to identify all cases, including those who have not sought medical care and those possibly exposed to risk
• Lay health workers may be trained to administer epidemiological questionnaires

• Should be carefully designed based on findings of a rapid preliminary inquiry
• Collects: name, age, sex, occupation, social class, travel history, previous exposure, time of onset, signs and symptoms, personal contacts, foods eaten, exposure to common vehicles (water, food, milk), history of injections/blood products, attendance at large gatherings, etc.
• For food-borne outbreaks: detailed food histories are essential
• If the outbreak is large (e.g., influenza), a random sample may be examined instead of all cases

• Ask patients if they know of other cases at home, school, workplace, neighborhood
• Look for cases admitted to local hospitals
• This may reveal person-to-person spread
• Search for new cases continues daily until the area is declared free of epidemic - typically for a period equal to twice the incubation period since the last case

Step 5 - Data Analysis

• Prepare a chronological distribution of dates of onset and construct an epidemic curve
• The epidemic curve suggests:
  • A time relationship with exposure to a suspected source
  • Whether it is a common-source or propagated (person-to-person) epidemic
  • Whether it is a seasonal or cyclic pattern suggestive of a particular infection
• A sharp, explosive single-peaked curve with a short incubation period = common source
• A prolonged curve with multiple peaks = propagated/person-to-person spread

• Prepare a spot map (geographic distribution of cases)
• Map their relation to possible sources of infection: water supply, air pollution, foods eaten, occupation
• Clustering of cases indicates a common source
• Classic example: John Snow's spot map during the 1854 cholera outbreak in London's Golden Square district, implicating the Broad Street water pump

• Analyze data by age, sex, occupation, and other possible risk factors
• Determine attack rates and case fatality rates for those exposed vs. those not exposed
• In food-borne outbreaks: calculate food-specific attack rates for each food item eaten to identify the source

Step 6 - Formulation of Hypotheses

• (a) Possible source
• (b) Causative agent
• (c) Possible modes of spread
• (d) Environmental factors that enabled the epidemic

Step 7 - Testing of Hypotheses

• All reasonable hypotheses are considered and weighed
• Compare attack rates in various groups for those exposed vs. not exposed to each suspected factor
• Determine which hypothesis is consistent with all known facts
• This may involve analytical study designs (case-control, cohort) applied within the outbreak investigation

Step 8 - Evaluation of Ecological Factors

• Sanitary status of eating establishments
• Water and milk supply; breakdown in the water supply system
• Movements of human populations
• Atmospheric changes: temperature, humidity, air pollution
• Population dynamics of insects and animal reservoirs

Step 9 - Further Investigation of Population at Risk

• Medical examination, screening tests
• Examination of suspected food, faeces, or blood samples
• Biochemical studies, assessment of immunity status
• Serological studies may reveal clinically inapparent (subclinical) cases and illuminate pathogenesis
• Healthy individuals from the same population are studied in a case-control fashion to classify all members by:
  • (a) Exposure to specific potential vehicles
  • (b) Whether ill or not

Step 10 - Writing the Report

Control Measures

• Eliminating or reducing the source of infection (e.g., closing the contaminated water pump)
• Interrupting transmission (e.g., sanitation, vector control)
• Protecting susceptibles (e.g., vaccination, chemoprophylaxis)

Summary - The 10 Steps at a Glance

1.  Verify the diagnosis
2.  Confirm existence of an epidemic
3.  Define population at risk (map + census)
4.  Rapid case search + data collection
5.  Data analysis - time (epidemic curve), place (spot map), person (attack rates)
6.  Formulate hypotheses
7.  Test hypotheses (compare attack rates, case-control analysis)
8.  Evaluate ecological / environmental factors
9.  Further investigation of population at risk
10. Write the report + implement and evaluate control measures

Water Purification

1. Large-scale (municipal/community) purification
2. Small-scale (household/individual) purification

Sources of Water (and Treatment Needed)

Part 1: Large-Scale Purification of Water

I. Storage

• About 90% of suspended impurities settle by gravity within 24 hours (plain sedimentation)
• Water becomes clearer, allowing light penetration
• Reduces the load on filters

• Aerobic bacteria oxidize organic matter using dissolved oxygen
• Free ammonia content decreases; nitrates rise

• A tremendous drop in bacterial count occurs - total bacterial count of river water drops by up to 90% in the first 5-7 days
• Pathogenic organisms gradually die out

Optimum storage period for river water: 10-14 days. Longer storage risks development of algal growth (imparting bad smell and colour).

II. Filtration

A. Slow Sand Filter (Biological / Schmutzdecke Filter)

• A slimy, gelatinous biological layer that forms on the surface of the sand bed
• Consists of threadlike algae, plankton, diatoms, and bacteria
• Called "the heart of the slow sand filter"
• Functions: removes organic matter, holds back bacteria, oxidizes ammoniacal nitrogen to nitrates
• Takes several days to fully form ("ripening" of the filter) - the first filtrate is run to waste

B. Rapid Sand Filter (Mechanical Filter)

• Alum (aluminium sulphate) is commonly added to raw water
• Forms a gelatinous floc (Al(OH)₃) that traps colloidal and suspended particles
• The floc settles in a sedimentation tank before water enters the rapid filter

Comparison: Slow Sand vs. Rapid Sand Filter

III. Disinfection (Chlorination)

• Must destroy pathogens within the contact time available
• Must not leave toxic reaction products
• Must be available readily at reasonable cost
• Must leave a residual concentration to protect against re-contamination
• Must be easily detectable by simple tests

In water works practice, "disinfection" is synonymous with "chlorination".

Chlorination

• Kills pathogenic bacteria
• No effect on spores and certain viruses (e.g., polio, viral hepatitis) except in high doses
• Oxidizes iron, manganese, hydrogen sulphide
• Destroys some taste and odour-producing substances
• Controls algae and slime organisms
• Aids coagulation

H₂O + Cl₂ → HCl + HOCl   (hypochlorous acid)
HOCl → H⁺ + OCl⁻         (hypochlorite ions)

• Germicidal action is mainly due to hypochlorous acid (HOCl)
• HOCl is 70-80 times more effective than hypochlorite ions (OCl⁻)
• Chlorine acts best at pH around 7 (predominance of HOCl)
• At pH > 8.5, chlorine becomes unreliable (90% HOCl ionizes to OCl⁻)

1. Water must be clear and free from turbidity (turbidity impedes chlorination)
2. The "chlorine demand" of water must be estimated - the amount of chlorine needed to destroy bacteria and oxidize all organic matter and ammonia. The point at which this demand is met = "break-point" - beyond this, free residual chlorine appears
3. Adequate contact period must be allowed; free residual chlorine of 0.5 mg/litre must remain after 1 hour contact to confirm effective disinfection

• Gaseous chlorine (in large waterworks)
• Bleaching powder (CaOCl₂) - contains ~33% available chlorine when fresh; unstable on exposure to air/light
• Chlorine solution (5% solution from bleaching powder)
• High Test Hypochlorite (HTH/perchloron) - contains 60-70% available chlorine; more stable
• Chlorine tablets (e.g., halazone) - 0.5 g tablet sufficient for 20 litres of water

• Ozone - powerful oxidizing agent; no residual effect
• UV irradiation - effective but no residual action; used in point-of-use systems
• Iodine - for emergency use only; 2 drops of 2% iodine in 1 litre of clear water; contact time 20-30 minutes; not suitable for municipal supplies

Advanced Membrane Processes

Part 2: Small-Scale (Household) Purification

(a) Boiling

• Most reliable household method
• Water must be brought to a "rolling boil" for 10-20 minutes
• Kills all bacteria, spores, cysts, and ova - yields sterilized water
• Also removes temporary hardness (drives off CO₂, precipitates CaCO₃)
• Disadvantage: no residual protection against subsequent contamination
• Water should be boiled preferably in the same container used for storage

(b) Chemical Disinfection (Small Scale)

• Calculate the chlorine demand of the well
• Add bleaching powder in calculated dose
• Agitate by moving bucket vigorously
• Allow 1 hour contact period
• Test for free residual chlorine (≥0.5 mg/litre) using orthotolidine arsenite test
• During cholera epidemics, wells are disinfected daily

• A device using two nested pots filled with a mixture of bleaching powder + coarse sand
• Lowered into the well at least 1 m below water level
• Provides continuous, constant-dose chlorination for 2-3 weeks in household wells

(c) Filtration (Small Scale)

• Candle filters (Berkefeld, Chamberland type) - household ceramic filters
• Domestic sand filters

WHO Water Quality Standards (Guidelines for Drinking Water Quality, 2011)

1. Acceptability (physical) - turbidity, colour, taste, odour
2. Microbiological - bacteria, viruses
3. Chemical - heavy metals, nitrates, fluoride
4. Radiological - radioactive substances

Summary Flow Chart: Municipal Water Purification

Raw Water (River/Surface)
        ↓
    STORAGE (10-14 days)
    [Physical, chemical, biological self-purification; 90% bacteria killed]
        ↓
    COAGULATION + SEDIMENTATION
    [Alum added → floc forms → suspended particles settle]
        ↓
    FILTRATION
    [Slow sand filter (biological layer) OR Rapid sand filter (backwashing)]
    [98-99.99% bacteria removed]
        ↓
    DISINFECTION (CHLORINATION)
    [Free residual Cl: 0.5 mg/L after 1 hr contact]
        ↓
    DISTRIBUTION
    [Safe, potable drinking water]

Epidemiology of Non-Communicable Diseases (NCDs)

Definition and Classification

The Four Major NCDs (WHO Classification):

1. Cardiovascular diseases (coronary heart disease, stroke, hypertension)
2. Cancers
3. Chronic respiratory diseases (COPD, asthma)
4. Diabetes mellitus

Global Burden of NCDs

• 74% of all deaths worldwide (2022) were attributed to NCDs - Harrison's Principles of Internal Medicine (22e, 2025)
• More than three-quarters of NCD deaths occur in low- and middle-income countries (LMICs)
• LMICs account for 85% of all early NCD-related deaths (~17 million people) - exceeding the combined deaths from AIDS, tuberculosis, and malaria
• By 2030, NCDs are projected to cause 52 million deaths annually if no additional action is taken
• Developing countries carry a "double burden" - both communicable and non-communicable diseases simultaneously

The Epidemiologic Transition

The Four Shared Behavioural Risk Factors (WHO Framework)

• Raised blood pressure (hypertension)
• Overweight/obesity
• Raised blood glucose (hyperglycaemia)
• Raised blood lipids (dyslipidaemia)

The WHO estimates that 80% of cardiovascular disease and type 2 diabetes, and 40% of all cancers, can be prevented through healthier diets, increased physical activity, and avoidance of tobacco. - Harrison's, 22e

1. Cardiovascular Diseases (CVD)

Global Burden

• Leading cause of NCD death worldwide
• Coronary artery disease (CAD) and hypertension in the elderly are the dominant causes of heart failure in high-income countries
• In LMICs, rheumatic heart disease (RHD) is a major contributor - affecting ~40 million people globally, with ~2.8 million new cases/year and >300,000 deaths/year (almost all in developing countries)
• The Framingham Heart Study: prevalence of severe hypertension (>210/120 mmHg) fell from 1.8% in the 1950s to 0.1% by the 1960s following introduction of antihypertensive agents

Major Risk Factors for CVD:

• Age (risk increases with age)
• Male sex (premenopausal women relatively protected)
• Family history / genetic predisposition
• Ethnicity

• Hypertension (a leading risk factor)
• Dyslipidaemia (elevated LDL-cholesterol)
• Tobacco smoking
• Diabetes mellitus
• Obesity
• Physical inactivity
• Unhealthy diet
• Psychosocial stress

The Framingham Risk Score

Epidemiology in LMICs

• Hypertension goes largely undetected, untreated, and uncontrolled in sub-Saharan Africa
• Heart failure from non-ischaemic causes (RHD, peripartum cardiomyopathy, endomyocardial fibrosis, HIV cardiomyopathy) accounts for >90% of cardiac failure in Sub-Saharan Africa
• Eastern Europe and Central Asia: collapse of the Soviet Union was followed by a catastrophic surge in CVD deaths (ischaemic heart disease mortality nearly doubled in Russia between 1991 and 1994)

2. Cancers

Global Burden

• ~10 million cancer deaths worldwide in 2020
• Low- and middle-income countries accounted for ~70% of these deaths
• By 2040, annual mortality from cancer is expected to increase to >16 million deaths

Key Risk Factors by Type

Cancer-causing infections (HPV, HBV, H. pylori) are responsible for up to 50% of cancer cases in some Sub-Saharan African countries. - Harrison's, 22e

NCD-Cancer Paradox in LMICs

• Outcomes are far worse in developing countries
• Health systems cannot provide early detection - most patients at time of diagnosis already have incurable disease
• Treatment is available only to a very small, mostly wealthy minority

3. Diabetes Mellitus

Global Burden

• 537 million diabetic adults worldwide in 2022 (~1 in 11 adults) - International Diabetes Federation
• Expected to rise to 783 million by 2045
• 3 in 4 adults with diabetes live in developing countries
• Diabetics in LMICs are often younger, so micro- and macrovascular complications take a far greater toll
• GBD 2017 study: diabetes accounted for 1.6 million deaths in 2019; 84% in LMICs
• High fasting plasma glucose ranks third among risks for disability and global mortality

Risk Factors for Type 2 Diabetes

Complications of Diabetes (Sources of Mortality and Morbidity)

• Microvascular: Diabetic nephropathy (CKD), retinopathy (blindness), neuropathy
• Macrovascular: Coronary artery disease, stroke, peripheral arterial disease
• Diabetes + hypertension = dramatically accelerated renal failure

4. Chronic Respiratory Diseases

Chronic Obstructive Pulmonary Disease (COPD)

• Tobacco smoking is the most important cause in high-income countries
• Biomass fuel combustion (indoor cooking fires) is a major cause in LMICs
• Occupational dust and chemical exposure, ambient air pollution also contribute
• Characterized by progressive, irreversible airflow limitation

Asthma

• Multifactorial: allergen exposure, air pollution, obesity, reduced childhood infections (hygiene hypothesis)
• Increasing prevalence in urban populations worldwide

5. Mental and Neurological Disorders

• In 2019, 1 in 8 people (~970 million) worldwide lived with a mental health disorder

• 580 million people suffer from depression and anxiety disorders

• >700,000 deaths by suicide every year
• Depressive disorders are the second leading cause of years lost to disability worldwide
• Most LMICs devote <1% of health expenditure to mental health

Prevention of NCDs: WHO Framework

Three Levels of Prevention

The WHO "MPOWER" Package (Tobacco Control)

• Monitor tobacco use
• Protect from tobacco smoke
• Offer cessation support
• Warn about dangers of tobacco
• Enforce advertising bans
• Raise taxes on tobacco

The "Best Buys" (WHO Cost-Effective NCD Interventions)

• Tobacco taxation and smoke-free laws
• Alcohol taxation and restriction of advertising
• Salt reduction in food
• Replacement of trans-fats with polyunsaturated fats
• Aspirin therapy for heart attacks
• Multi-drug therapy for high CVD risk
• Vaccination against HPV and HBV (cancer prevention)

Epidemiological Study Approaches Used in NCD Research

Summary Table: The Four Major NCDs