Measures of Disease Frequency

Measures of disease frequency quantify how often disease occurs in a population. They form the numerical backbone of epidemiology — every rate, ratio, and proportion you compute in outbreak investigation, surveillance, or screening research flows from these definitions. NEET PG loves scenario stems that hand you raw numbers and ask which measure fits and how to calculate it.

Core vocabulary: ratio, proportion, rate

Before frequency measures, fix the three basic mathematical forms, because examiners exploit the confusion.

Type	Definition	Numerator part of denominator?	Range	Example
Ratio	a / b (two unrelated quantities)	No	0 to ∞	Sex ratio, maternal mortality ratio, odds
Proportion	a / (a + b), a part of the whole	Yes	0 to 1 (0–100%)	Prevalence, attack rate
Rate	Events per unit population per unit time	Yes (with time)	Has time dimension	Incidence rate, mortality rate

High-yield: A true rate must contain a time element (per 1000 per year). Prevalence and attack rate are technically proportions, not rates, even though they are loosely called "rates" in practice. Maternal mortality "ratio" is a ratio because live births (denominator) are not part of the maternal deaths (numerator).

Incidence

Incidence measures new cases arising in a previously disease-free population over a defined period. It captures the risk / force of new disease and is the measure of choice for studying etiology and acute conditions.

Cumulative incidence (incidence risk / attack rate)

$$\text{Cumulative incidence} = \frac{\text{No. of NEW cases during a period}}{\text{Population at risk at the start}}$$

• A proportion (0–1); dimensionless, but a time period must be specified.
• Denominator = population at risk at the beginning (people already diseased or immune are excluded).
• Synonyms: incidence risk, incidence proportion, attack rate (when used in outbreaks over a short period).

Incidence rate (incidence density)

$$\text{Incidence rate} = \frac{\text{No. of NEW cases}}{\text{Total person-time at risk}}$$

• A true rate; denominator is person-time (person-years, person-months).
• Used when follow-up periods differ between subjects or when people enter/leave the cohort at different times.
• Units: cases per person-year (e.g., 5 per 1000 person-years).

High-yield: Person-time accounts for variable follow-up. If 100 people are followed but for differing durations, you sum each person's at-risk time. A subject contributes person-time only until they develop the disease, die, or are lost to follow-up.

Worked logic: 500 disease-free people followed for 2 years; 20 develop disease (assume cases occur at year-end). Cumulative incidence = 20/500 = 0.04 = 4% over 2 years. Person-years ≈ 500 × 2 = 1000 (minus time lost by cases) → incidence rate ≈ 20/1000 = 20 per 1000 person-years.

Prevalence

Prevalence measures all existing cases (old + new) at a point or over a period. It reflects the burden / load of disease and is the measure of choice for chronic conditions and for health-service planning.

Point prevalence

$$\text{Point prevalence} = \frac{\text{All cases (old + new) at a point in time}}{\text{Total population examined at that point}}$$

• "Snapshot" on a single day or single survey — the most commonly reported prevalence.

Period prevalence

$$\text{Period prevalence} = \frac{\text{All cases existing during a period}}{\text{Average (or mid-period) population}}$$

$$\text{Period prevalence} = \text{Point prevalence at start} + \text{Incidence during the period}$$

• Includes cases present at the start plus new cases arising during the interval.

Feature	Incidence	Prevalence
Counts	NEW cases only	OLD + NEW (existing) cases
Numerator timing	During a period	At a point / during a period
Nature	Risk / force of disease	Burden / load of disease
Best for	Acute disease, etiology, RCTs/cohorts	Chronic disease, health planning, resource allocation
Affected by duration?	No	Yes — longer duration → higher prevalence
Type	Risk (proportion) or rate	Proportion
Study design	Cohort	Cross-sectional survey

The pivotal relationship: P = I × D

When incidence and duration are stable (steady state):

$$\textbf{Prevalence} \approx \textbf{Incidence} \times \textbf{Duration}$$ (more precisely, P/(1−P) = I × D; for low-prevalence diseases P ≈ I × D)

Reasoning flow: New cases enter the prevalence pool by incidence → they stay in the pool for the duration of illness → they exit by recovery or death.

High-yield: A new effective treatment that prolongs life without curing (e.g., antiretrovirals for HIV, insulin for type 1 diabetes) increases prevalence while incidence is unchanged — because duration rises. Conversely, a treatment that cures rapidly or a disease that kills rapidly lowers prevalence (short duration). This is a favourite trap stem.

Factors that change prevalence

Increase prevalence	Decrease prevalence
Longer disease duration	Shorter duration
Prolongation of life without cure	Higher case fatality (rapid death)
Increase in incidence (new cases)	Rapid cure / recovery
In-migration of cases	Decrease in incidence
Better diagnosis / reporting	Out-migration of cases
In-migration of susceptibles? (no — that lowers it)	Improved cure rate

Attack rate and secondary attack rate

These are special cumulative incidences used in acute outbreaks (food poisoning, measles in a school, COVID in a household).

Attack rate (AR)

$$\text{Attack rate} = \frac{\text{No. who develop disease}}{\text{Total exposed (at risk)}} \times 100$$

• Expressed as a percentage; restricted to the epidemic period.
• In a food-borne outbreak, calculate a food-specific attack rate for each item; the food with the highest AR among eaters and the greatest difference (AR in eaters − AR in non-eaters) is the likely culprit.

Secondary attack rate (SAR)

$$\text{SAR} = \frac{\text{No. of NEW (secondary) cases among contacts}}{\text{Total susceptible contacts}} \times 100$$

• Measures infectivity / communicability of an agent and the effect of household/contact exposure.
• Crucial denominator rule: exclude the primary (index) case and exclude already-immune contacts from the denominator. Numerator counts only cases appearing within one incubation period of the primary case.

High-yield: SAR denominator = susceptible contacts MINUS the index case. A classic MCQ gives a family of 6 where the index case and 1 immune (vaccinated) member must be removed: denominator = 6 − 1 (index) − 1 (immune) = 4.

Worked SAR: Index case in a house of 5. Of the remaining 4 susceptible contacts, 2 fall ill within one incubation period. SAR = 2/4 × 100 = 50%.

Crude, specific, and standardised rates

Rate	Definition	Use
Crude rate	Whole population in denominator	Simple, single summary; but confounded by age/sex structure
Specific rate	Restricted to a subgroup (age-, sex-, cause-specific)	Detailed comparison within strata
Standardised (adjusted) rate	Adjusted to a standard population	Fair comparison between populations of differing structure

• Direct standardisation: apply the study population's specific rates to a standard population → needs age-specific rates of the study population.
• Indirect standardisation: apply standard population's rates to the study population → yields the Standardised Mortality Ratio (SMR). Used when age-specific rates of the study group are unavailable or numbers are small.

High-yield: Use indirect standardisation (SMR) when age-specific rates of the study population are not known or the population is small. SMR = (Observed deaths / Expected deaths) × 100.

Choosing the right measure — the exam algorithm

1. Are the cases NEW only? → Incidence. All existing cases? → Prevalence.
2. New cases, denominator = person-time? → Incidence rate (density). Denominator = population at start? → Cumulative incidence (risk).
3. Acute outbreak, short epidemic period? → Attack rate.
4. Spread among contacts of an index case? → Secondary attack rate.
5. Snapshot survey for burden of a chronic disease? → Point prevalence.
6. Comparing two populations of different age structure? → Standardised rate.

One-line flow: New cases → Incidence → if person-time → density; Existing cases → Prevalence → P = I × D.

Special / related measures (commonly bundled in MCQs)

• Crude birth rate = (Live births in a year / Mid-year population) × 1000.
• Crude death rate = (Deaths in a year / Mid-year population) × 1000.
• Case fatality rate (CFR) = (Deaths from a disease / Total cases of that disease) × 100 — measures virulence/severity, best for acute/short-course disease.
• Proportional mortality rate = (Deaths from a specific cause / Total deaths) × 100.
• Infant mortality rate = (Infant deaths under 1 yr / Live births) × 1000 — a sensitive index of overall health.

High-yield: Case fatality rate is really a proportion (deaths among cases). It is suitable for acute diseases (cholera, rabies) and is a measure of killing power / virulence, not of frequency of occurrence.

Common pitfalls and differentials

• Incidence vs prevalence confusion: A cross-sectional survey can only give prevalence (a one-time examination cannot distinguish old from new cases). Incidence requires follow-up over time (cohort design).
• Attack rate is an incidence, not a prevalence — it counts new cases over the epidemic period.
• Person-time vs persons: if follow-up durations differ, the denominator must be person-time, not number of persons.
• MMR is a ratio, not a rate (denominator = live births, which excludes the maternal deaths conceptually).
• Prevalence is unsuitable for studying causation because it is influenced by duration and survival, not just by the rate of new disease.

Recently asked / exam angle

• Scenario calculation of SAR with an index case and one immune family member to be excluded — repeatedly tested; watch the denominator.
• P = I × D interpretation: "A new drug prolongs survival in heart failure — what happens to prevalence and incidence?" Answer: prevalence rises, incidence unchanged.
• Food-specific attack rate tables in a food-poisoning outbreak — identify the implicated food by highest attack rate / largest difference in eaters vs non-eaters.
• Choosing the measure: chronic disease burden survey → point prevalence; etiologic cohort → incidence.
• Person-years calculation in a cohort with staggered entry — compute incidence density.
• Direct vs indirect standardisation — when age-specific rates are unavailable, choose indirect (SMR).
• Ratio vs proportion vs rate classification of MMR, sex ratio, prevalence, and incidence density.
• Crude vs standardised rate — why crude rates mislead when comparing populations with different age structures.

Rapid revision

1. Incidence = new cases; prevalence = all existing cases (old + new).
2. Cumulative incidence uses population at risk at start; incidence density uses person-time.
3. A true rate has a time dimension; prevalence and attack rate are proportions.
4. P ≈ I × D — prevalence depends on incidence and disease duration.
5. A drug that prolongs life without curing → prevalence increases, incidence unchanged.
6. Attack rate = special incidence for acute outbreaks, expressed as a percentage.
7. SAR denominator = susceptible contacts minus the index case (and minus immune contacts).
8. Highest food-specific attack rate points to the implicated food in an outbreak.
9. Cross-sectional survey → prevalence; cohort study → incidence.
10. Case fatality rate = killing power, suited to acute diseases; it is a proportion.
11. Indirect standardisation (SMR) when age-specific rates are unknown or numbers small.
12. Maternal mortality ratio and sex ratio are ratios, not rates.