Page 13 of 21

CM7.6-7 | CM7.6-7 | Screening Tests and Epidemic Investigation — SDL Guide

Learning Objectives

• Enumerate and evaluate the need for screening tests, including criteria for a good screening programme (Wilson-Jungner criteria) (CM7.6)
• Calculate and interpret sensitivity, specificity, positive predictive value, and negative predictive value from a 2×2 table (CM7.6)
• Explain how prevalence affects positive predictive value and the public health implications (CM7.6)
• Describe and demonstrate the steps in the investigation of a communicable disease epidemic (CM7.7)
• Interpret epidemic curves (point-source, propagated, and mixed) and describe principles of outbreak control measures (CM7.7)

INSTRUCTIONS

Screening detects disease before symptoms appear — enabling earlier treatment and potentially saving lives. Epidemic investigation stops disease spread — enabling rapid control of outbreaks that could otherwise become pandemics. Both skills require the same underlying toolkit: a clear case definition, systematic data collection, quantitative test evaluation, and evidence-based action. This module builds fluency in both, with worked calculations and a step-by-step investigation framework applicable to any communicable disease outbreak.

References

• Park's Textbook of Preventive and Social Medicine, 27th edition — Chapter 2: Epidemiology (Screening; Epidemic Investigation) (textbook)
• Wilson JM, Jungner G. Principles and Practice of Screening for Disease. WHO, 1968 (textbook)

---

Version 2.0 | NMC CBUC 2024

The Population Burden That Drives Screening: Finding Disease Before It Presents

Screening is the presumptive identification of unrecognised disease or defect among apparently healthy individuals by means of tests, examinations, or other procedures that can be applied rapidly. The key word is presumptive — a screening test is not a diagnostic test. A positive screening result requires confirmation by a definitive diagnostic test before treatment is initiated.

The rationale for screening is grounded in the natural history of disease: many diseases have a prolonged subclinical phase (the detectable pre-clinical phase or DPCP) during which they are present but not yet symptomatic. If the DPCP is long and the disease is serious and treatable, early detection during this phase can alter outcome. The classic example is cervical cancer, where the pre-cancerous phase (cervical intraepithelial neoplasia) may last 10–15 years before progression to invasive cancer — a long window for screening to detect and treat.

The burden justification for screening requires that the target disease meet several conditions: it must be a sufficiently serious health problem (high prevalence or severity); effective treatment must exist and be more effective when applied early (lead-time benefit must translate to real outcome benefit, not just earlier diagnosis); and the target population must be accessible and willing to participate. India's population-level screening priorities — cervical cancer (ASHA-delivered VIA), diabetes (opportunistic blood glucose in health and wellness centres), hypertension, and TB (active case finding) — reflect these criteria.

Types of screening: mass screening (entire population), selective (high-risk subgroups — e.g. HIV testing in high-risk groups), opportunistic (incidental testing during clinical encounter — e.g. measuring BP in every outpatient visit), multiphasic (multiple tests in one visit — e.g. national antenatal care protocol testing for anaemia, syphilis, blood group, blood pressure).

Evaluating Screening Tests: Sensitivity, Specificity, and Predictive Values

The performance of a screening test is evaluated using a 2×2 contingency table that cross-classifies test results against the true disease status (determined by a gold-standard diagnostic test):

``
True Disease+ True Disease−
Test Positive: TP FP → TP+FP (all positives)
Test Negative: FN TN → FN+TN (all negatives)
TP+FN FP+TN
``

Sensitivity (Sn) = TP / (TP + FN) × 100
The proportion of truly diseased persons correctly identified by the test as positive. A high-sensitivity test misses few true cases (few false negatives). When you must not miss disease (serious, treatable disease), maximise sensitivity. Mnemonic: SnNout — a highly Sensitive test, when Negative, rules OUT disease.

Specificity (Sp) = TN / (TN + FP) × 100
The proportion of truly non-diseased persons correctly identified by the test as negative. A high-specificity test produces few false positives. When a positive result will trigger a serious, expensive, or harmful intervention, maximise specificity. Mnemonic: SpPin — a highly Specific test, when Positive, rules IN disease.

Positive Predictive Value (PPV) = TP / (TP + FP) × 100
The probability that a person who tests positive actually has the disease. PPV is the clinically relevant question for a positive result: 'My patient tested positive — how likely is it that they truly have the disease?'

Negative Predictive Value (NPV) = TN / (TN + FN) × 100
The probability that a person who tests negative truly does not have the disease.

Critical relationship: Prevalence affects PPV and NPV — but NOT Sensitivity or Specificity.

Sensitivity and Specificity are intrinsic properties of the test, determined by the test's biological or chemical characteristics and the chosen cut-off point. They do not change with the population's disease prevalence. PPV and NPV, however, depend heavily on prevalence:

• When prevalence is HIGH → PPV increases (fewer false positives relative to true positives) → positive test results are more reassuring.
• When prevalence is LOW → PPV falls dramatically (many false positives relative to few true positives) → even a very sensitive, specific test will yield mostly false positives in a low-prevalence population.

Worked example: A tuberculin skin test has Sn = 95% and Sp = 90%. Applied to 1,000 persons in a high-TB community (prevalence = 10%, i.e. 100 truly infected):
- TP = 95, FN = 5 (5% of 100 true cases missed)
- TN = 810, FP = 90 (10% of 900 uninfected false-positive)
- PPV = 95 / (95+90) × 100 = 95/185 × 100 ≈ 51%

Applied to the same test in a low-TB community (prevalence = 1%, i.e. 10 truly infected per 1,000):
- TP = 9.5, FN = 0.5; TN = 891, FP = 99
- PPV = 9.5 / (9.5+99) × 100 ≈ 8.8%

The test is identical; the community prevalence drives the PPV from 51% to 8.8%. This is why mass screening with a low-specificity test in a low-prevalence population generates massive overdiagnosis and unnecessary investigation.

Provided image

Wilson-Jungner criteria (WHO, 1968): Ten criteria for evaluating whether a disease and test are suitable for a screening programme:
1. The condition should be an important health problem
2. There should be an accepted treatment for patients with recognised disease
3. Facilities for diagnosis and treatment should be available
4. There should be a recognisable latent or early symptomatic stage
5. There should be a suitable screening test
6. The test should be acceptable to the population
7. The natural history of the condition should be adequately understood
8. There should be an agreed policy on whom to treat as patients
9. The cost of case-finding should be economically balanced in relation to possible expenditure on medical care as a whole
10. Case-finding should be a continuing process and not a 'once and for all' project

Epidemic Investigation: From Alert to Control

An epidemic occurs when the incidence of a disease in a population exceeds what is normally expected for that place and time. An outbreak is epidemiologically equivalent to an epidemic but is often used for a more localised event. The investigation of an epidemic follows a systematic 10-step procedure, which can be conducted in a modified order depending on the circumstances:

Step 1: Verify the diagnosis — Confirm that cases have the disease suspected (clinical, laboratory, and/or epidemiological evidence). Laboratory confirmation prevents investigating a pseudo-epidemic (artefactual cluster due to changed reporting practices or diagnostic criteria).

Step 2: Confirm the existence of an epidemic — Compare current case counts to the expected baseline (seasonal average from prior years, or control population rates). If a seven-fold excess is confirmed, the epidemic is real.

Step 3: Define the case — A case definition is a set of standard criteria for classifying whether an individual should be counted as a case. It includes: clinical criteria (signs/symptoms); laboratory criteria (test results); epidemiological criteria (time, place, person — e.g. exposure to a specified event). Three levels: confirmed (clinical + laboratory), probable (clinical + epidemiological, no laboratory), suspect (consistent symptoms, no confirmation). Case definitions should be sensitive early in an outbreak (to find all cases) and narrowed as the investigation progresses.

Step 4: Find all cases (case ascertainment) — Actively search for cases beyond those already reported: alert local hospitals, review laboratory records, conduct community surveys. Calculate the attack rate for the outbreak (cases / exposed population × 100).

Step 5: Describe distribution by time, place, and person — Construct an epidemic curve (cases on Y-axis, date/time of onset on X-axis). The shape of the epidemic curve is diagnostically powerful:
- Point-source epidemic (common source, brief exposure): rapid rise, bell-shaped peak, rapid fall; the entire epidemic spans ≤1 incubation period; all cases were exposed at approximately the same time (e.g. food-poisoning at a wedding).
- Propagated (continuing source/person-to-person) epidemic: gradual rise, multiple irregular peaks, each peak approximately one incubation period apart (e.g. measles spreading through a school).
- Mixed epidemic: point-source outbreak followed by person-to-person spread (e.g. index case introduced pathogen, who then transmitted to contacts — typhoid with secondary cases).

Draw a spot map to show geographical clustering. Describe person characteristics (age, sex, occupation, immunisation status).

Types of Epidemic Curves

Step 6: Hypothesise source and mode of transmission — Based on the epidemic curve shape, the spot map, and person distribution, formulate hypotheses about the source and mode: Was it water? Food? Person-to-person? A specific event?

Step 7: Test the hypothesis — Design and execute an analytical study (usually a case-control or retrospective cohort study) comparing exposure histories of cases and controls. Calculate attack rates for different exposure categories; compute Relative Risk or Odds Ratio. A significantly elevated OR for consumption of a specific food item at a common meal, for example, confirms food as the vehicle.

Steps 8–10: Implement control measures, evaluate effectiveness, communicate findings — Control measures (described in detail in the next section) are implemented in parallel with investigation, not after it. Effectiveness is monitored by tracking new cases via the epidemic curve. A final outbreak investigation report is written and disseminated to health authorities.

Epidemic Control Measures and Communication

Control measures are implemented while the investigation is ongoing — they should not wait for complete analytical confirmation if the delay would allow further spread. Control measures are organised by the link in the transmission chain they target (reflecting CM7.2).

Source control: Remove or eliminate the contaminated vehicle (discard implicated food, close contaminated water supply, treat or isolate infectious cases). In the food-borne outbreak scenario, identifying and removing the implicated food from the canteen is the first control action — performed even before case-control results confirm which item was responsible.

Interrupting transmission: Disinfect water supply (hyperchlorination), enforce handwashing and food hygiene, implement infection control precautions in hospitals (standard and transmission-based precautions), trace and test contacts.

Protecting the susceptible host: Emergency mass vaccination during measles outbreaks; antibiotic prophylaxis for close contacts of meningococcal disease; passive immunisation (immunoglobulins) for post-exposure hepatitis A prophylaxis in close contacts.

Evaluating effectiveness of control: Plot a rolling epidemic curve during the response phase. If control is effective, new cases should decline and the curve should fall. If new cases cluster around a different source (a second wave), the investigation and control must be extended.

Risk communication: Public health authorities must communicate with multiple audiences simultaneously: affected communities (what to do, what to eat/drink, when to seek care), media (accurate, non-alarmist messaging), political authorities (resource needs, legal authorities for quarantine), and the clinical community (case definitions, treatment protocols, notification procedures). The WHO risk communication framework emphasises: Be first, be right, be credible, be empathetic.

Outbreak investigation report: The formal written report follows an IMRAD-like structure (Introduction/Background, Methods/Case definition, Findings/Epidemiological analysis, Discussion/Control measures, Recommendations). It is submitted to the district/state health authority and to IDSP for national surveillance purposes.