Biostatistics Practice Questions

Q: An investigator wants to study the association between maternal intake of iron supplements (Yes or No) and the incidence of low birth weight ( = 2500 gms). He collects data from 100 pregnant women on the usage of iron supplements and the birth weight status of their newborns. What is the appropriate statistical test of hypothesis for this situation?

Chi-square test. ### Explanation The core of this question lies in identifying the **type of data** being analyzed. **1. Why Chi-square test is correct:** The investigator is looking for an association between two variables: * **Maternal Iron Intake:** Categorical/Qualitative (Yes or No). * **Birth Weight Status:** Categorical/Qualitative (< 2500g or ≥ 2500g). When both the independent and dependent variables are **qualitative (nominal/ordinal)**, the data is typically represented in a contingency table (e.g., a 2x2 table). The **Chi-square ($\chi^2$) test** is the standard non-parametric test used to compare proportions and determine if there is a statistically significant association between two such categorical variables. **2. Why other options are incorrect:** * **Paired t-test:** Used to compare means of two related groups (e.g., blood pressure before and after treatment in the same individual). It requires quantitative data. * **Unpaired (Independent) t-test:** Used to compare the means of two independent groups (e.g., comparing the actual mean birth weight in grams between supplement users and non-users). It requires quantitative data. * **Analysis of Variance (ANOVA):** Used to compare the means of three or more independent groups. It also requires quantitative data. **3. Clinical Pearls & High-Yield Facts for NEET-PG:** * **Quantitative Data (Means):** Use T-test (2 groups) or ANOVA (>2 groups). * **Qualitative Data (Proportions):** Use Chi-square test or Fisher’s Exact test (if any cell value in the 2x2 table is < 5). * **Correlation:** Use Pearson’s coefficient ($r$) for quantitative data to see the strength of a linear relationship. * **Golden Rule:** Always check if the question provides "Mean/Standard Deviation" (Numerical) or "Percentage/Incidence/Proportion" (Categorical) before choosing the test.

Q: A village with a population of 10,000 has a birth rate of 36 per 1000 population. In one year, there have been 5 maternal deaths. What is the maternal mortality rate (MMR) in this village?

None of the above. ### Explanation **1. Understanding the Correct Answer (Option D)** The Maternal Mortality Ratio (MMR) is defined as the number of maternal deaths per **100,000 live births**. To calculate it, we first need to determine the number of live births in the village: * **Population:** 10,000 * **Birth Rate:** 36 per 1,000 population * **Total Live Births:** $(36 / 1,000) \times 10,000 = 360$ live births. Now, apply the MMR formula: $$\text{MMR} = \frac{\text{Total Maternal Deaths}}{\text{Total Live Births}} \times 100,000$$ $$\text{MMR} = \frac{5}{360} \times 100,000 = 1,388.8 \text{ per 100,000 live births.}$$ Since 1,388.8 is not among the options, **Option D (None of the above)** is correct. **2. Why Other Options are Incorrect** * **Option A (0.5 per 1000):** This is a miscalculation likely derived from dividing deaths by total population ($5/10,000$), which is the Crude Death Rate, not MMR. * **Option B (5 per 1000):** This ignores the denominator of live births and uses an incorrect multiplier. * **Option C (7.2 per 1000):** This is a common distractor calculated by $(36/5)$, which has no statistical basis in epidemiology. **3. High-Yield Clinical Pearls for NEET-PG** * **Ratio vs. Rate:** MMR is technically a **Ratio**, not a rate, because the numerator (deaths) is not necessarily part of the denominator (live births; as one mother can have multiple births or a mother may die without a live birth). * **Denominator:** Always use **Live Births** for MMR. If the question provides "Total Pregnancies," it is used for the Maternal Mortality *Rate* (per 1,000 women of reproductive age), but for MMR, live births is the standard. * **Multiplier:** MMR is the only obstetric indicator that uses **100,000** as a multiplier; most others (IMR, NMR, CBR) use 1,000. * **Timeframe:** Maternal death is defined as death during pregnancy or within **42 days** of delivery.

Q: If the confidence limit is increased, what happens to the significance of the data?

Previously significant data becomes insignificant. ### Explanation **1. Why the Correct Answer is Right:** In biostatistics, the **Confidence Limit (CL)** and the **Level of Significance (α)** are inversely related. The formula is: **CL = 100 – α**. When you "increase the confidence limit" (e.g., moving from 95% to 99%), you are demanding a higher degree of certainty before rejecting the null hypothesis. This effectively decreases the allowable error (α) from 0.05 to 0.01. A wider confidence interval is more likely to include the "null value" (e.g., 0 for mean difference or 1 for Odds Ratio/Relative Risk). If a study was significant at the 95% level (p < 0.05), but the threshold is raised to 99% (p < 0.01), a result with a p-value of 0.03—which was previously significant—now fails to meet the stricter criteria and becomes **insignificant**. **2. Why Incorrect Options are Wrong:** * **Option A:** Increasing the confidence limit makes the "test" harder to pass. It cannot make insignificant data significant; it does the opposite by tightening the requirements for proof. * **Option C:** Significance is directly tied to the confidence interval width. Changing the CL changes the p-value threshold, thus directly affecting the interpretation of significance. * **Option D:** The relationship is mathematical and predictable. Increasing the CL always makes the interval wider, increasing the likelihood of encompassing the null hypothesis. **3. Clinical Pearls & High-Yield Facts:** * **95% Confidence Interval (CI):** Corresponds to a p-value of < 0.05. * **99% Confidence Interval (CI):** Corresponds to a p-value of < 0.01. * **Width of CI:** If the CI for Relative Risk (RR) or Odds Ratio (OR) includes **1**, the result is NOT significant. If the CI for the difference in means includes **0**, the result is NOT significant. * **Precision:** A narrower CI indicates greater precision and is usually achieved by increasing the sample size ($n$).

Q: If the correlation between weight and height is very strong, what will be the correlation coefficient?

1. **Explanation:** The correlation coefficient (denoted as **'r'**) measures the strength and direction of a linear relationship between two continuous variables (e.g., height and weight). **Why Option A is Correct:** The value of the correlation coefficient always ranges from **-1 to +1**. * A value of **+1** indicates a **perfect positive correlation**, meaning as one variable increases, the other increases in a perfectly predictable linear fashion. * In medical biostatistics, when a relationship is described as "very strong" or "perfect," the coefficient approaches or equals 1. Since height and weight generally increase together, they share a positive correlation. **Why Other Options are Incorrect:** * **Option B (Greater than 1):** This is mathematically impossible. The Pearson correlation coefficient cannot exceed +1 or be less than -1. Any value outside this range indicates a calculation error. * **Option C (0):** A correlation coefficient of 0 indicates **zero correlation** or no linear relationship between the variables. * **Option D:** Incorrect as Option A is the standard mathematical representation of a perfect strong correlation. **High-Yield Clinical Pearls for NEET-PG:** 1. **Direction:** Positive (+) means variables move in the same direction; Negative (-) means they move in opposite directions (e.g., exercise and body fat). 2. **Strength:** * 0.00–0.19: Very weak * 0.20–0.39: Weak * 0.40–0.59: Moderate * 0.60–0.79: Strong * **0.80–1.0: Very strong/Perfect** 3. **Coefficient of Determination (r²):** This represents the proportion of variance in one variable explained by the other. If r = 0.7, then r² = 0.49 (49% of the change is explained). 4. **Scatter Diagram:** A perfect correlation (r=1) forms a straight line on a scatter plot.

Q: What statistical method is used to calculate the death rate between two populations with different age groups?

Standardized death rate. ### Explanation **1. Why Standardized Death Rate is Correct:** The age structure of a population is a major determinant of its mortality; an older population will naturally have more deaths than a younger one, even if health conditions are identical. When comparing two populations with different age distributions, the **Standardized (Adjusted) Death Rate** is used to eliminate the "confounding" effect of age. It allows for a fair comparison by applying the observed death rates to a single **Standard population**, ensuring that any difference in mortality is due to actual health factors rather than demographic makeup. **2. Why Other Options are Incorrect:** * **Crude Death Rate (CDR):** This is the actual number of deaths per 1,000 population. It does not account for age distribution, making it misleading for comparisons between populations with different age structures (e.g., comparing Kerala with Bihar). * **Case Fatality Rate (CFR):** This measures the killing power of a specific disease (Deaths from disease / Total cases of that disease). it is a measure of **virulence**, not population mortality. * **Age-Specific Death Rate:** This calculates mortality within a specific age bracket (e.g., 5–14 years). While it accounts for age, it only looks at one segment at a time and cannot provide a single summary value to compare two entire populations. **3. High-Yield NEET-PG Pearls:** * **Direct Standardization:** Used when the age-specific death rates of the study population are known. * **Indirect Standardization:** Used when age-specific rates are unknown or the population is small. It yields the **Standardized Mortality Ratio (SMR)**. * **SMR Formula:** (Observed Deaths / Expected Deaths) × 100. * Standardization is the method of choice for comparing any vital statistics (birth rates, death rates) across different geographical areas.

Q: All are true regarding Student's t-test, except?

Standard error of the mean is not estimated.. ### Explanation The **Student’s t-test** is a parametric test used to determine if there is a significant difference between the means of two groups. **Why Option A is the Correct Answer (The "Except" statement):** In a t-test, the **Standard Error of the Mean (SEM)** is a fundamental component of the formula. The t-statistic is calculated as: $$t = \frac{\text{Difference between means}}{\text{Standard Error of the difference}}$$ Since the population standard deviation is usually unknown in t-tests, it is estimated using the sample standard deviation to calculate the SEM. Therefore, stating that the SEM is not estimated is **incorrect**, making it the right choice for an "except" question. **Analysis of Other Options:** * **Option B (Standard population is selected):** In a one-sample t-test, a sample mean is compared against a known "standard" or "target" population mean. * **Option C (Two samples are compared):** This is the most common application (unpaired t-test), comparing means between two independent groups (e.g., blood pressure in Group A vs. Group B). * **Option D (t-distribution table is required):** To determine the p-value and statistical significance, the calculated t-value must be compared against critical values in the Student’s t-distribution table, adjusted for **degrees of freedom (df)**. ### High-Yield Clinical Pearls for NEET-PG * **Sample Size:** t-tests are typically used for small samples (**n 30, the distribution approximates a Normal Distribution, and a **Z-test** can be used. * **Paired t-test:** Used for "before and after" studies on the same group (e.g., BP before and after a drug). * **Unpaired (Independent) t-test:** Used to compare two different groups. * **ANOVA (F-test):** Used when comparing means of **three or more** groups. * **Data Type:** t-tests are used for **quantitative (numerical)** data that follows a normal distribution.

Q: In which aspect is a case-control study superior to a cohort study?

Shorter study duration. ### Explanation **Correct Option: C. Shorter study duration** **Why it is correct:** A **Case-Control study** is retrospective in nature. It begins with the identification of "cases" (those with the disease) and "controls" (those without), then looks backward in time to assess exposure. Since the outcome has already occurred at the start of the study, data collection is rapid and inexpensive. In contrast, a **Cohort study** is typically prospective; it starts with healthy individuals and follows them over a long period to see who develops the disease, making it time-consuming and costly. **Why other options are incorrect:** * **A. Reduced likelihood of bias:** Case-control studies are actually *more* prone to bias, particularly **Recall Bias** (patients with the disease remember exposures differently) and **Selection Bias**. Cohort studies are generally more robust. * **B. Greater accuracy:** Cohort studies provide more accurate data regarding the temporal relationship between exposure and outcome (incidence), whereas case-control studies are better suited for generating hypotheses. * **D. Ability to calculate relative risk:** This is a classic NEET-PG trap. Case-control studies cannot calculate **Relative Risk (RR)** because they do not measure incidence. They can only estimate risk using the **Odds Ratio (OR)**. RR is calculated in Cohort studies. **High-Yield Clinical Pearls for NEET-PG:** * **Case-Control Study:** Best for **rare diseases** and diseases with long latency periods. * **Cohort Study:** Best for **rare exposures** and determining the incidence of a disease. * **Mnemonic for RR vs. OR:** **C**ohort = **R**elative **R**isk (both have 'R'); Case-C**o**ntrol = **O**dds Ratio (both have 'O'). * **Nested Case-Control Study:** A hybrid design that is more cost-effective than a full cohort study but reduces selection/recall bias compared to a standard case-control study.

Q: Susceptibility bias is eliminated by:

Randomization. ### Explanation **1. Why Randomization is Correct:** **Susceptibility bias** (also known as allocation bias) occurs when the groups being compared in a study have different baseline characteristics or prognostic factors, making one group more "susceptible" to the outcome than the other. **Randomization** is the "heart" of a Randomized Controlled Trial (RCT). It ensures that every participant has an equal chance of being assigned to any study group. This process distributes both **known and unknown (latent) confounders** equally between the intervention and control groups. By ensuring baseline comparability, randomization eliminates susceptibility bias at the start of the study. **2. Why Other Options are Incorrect:** * **Blinding (Single/Double):** Blinding is used to eliminate **Information/Observation bias**. It prevents participants or investigators from knowing the treatment assignment, thereby preventing subjective influence on reporting or measuring outcomes. It does *not* affect the initial allocation of patients. * **Matching:** Matching is primarily used in **Case-Control studies** to eliminate known confounders (like age or sex). However, matching cannot account for unknown confounders and can lead to "over-matching." It is less robust than randomization for eliminating susceptibility bias. **3. High-Yield Pearls for NEET-PG:** * **Randomization:** The best method to eliminate **confounding**. It ensures "comparability" of groups. * **Blinding:** The best method to eliminate **ascertainment/observer bias**. * **Confounding:** A situation where an external variable is associated with both the exposure and the outcome. * **Allocation Concealment:** A process used *during* randomization to prevent selection bias by ensuring the person enrolling participants does not know which group the next participant will fall into (e.g., opaque envelopes).

Question 1

An investigator wants to study the association between maternal intake of iron supplements (Yes or No) and the incidence of low birth weight (< 2500 gms or >= 2500 gms). He collects data from 100 pregnant women on the usage of iron supplements and the birth weight status of their newborns. What is the appropriate statistical test of hypothesis for this situation?

Accepted Answer

Chi-square test

Answer

Paired t-test

Answer

Unpaired or independent t-test

Answer

Analysis of variance

Question 2

A village with a population of 10,000 has a birth rate of 36 per 1000 population. In one year, there have been 5 maternal deaths. What is the maternal mortality rate (MMR) in this village?

Accepted Answer

None of the above

Answer

0.5 per 1000 live births

Answer

5 per 1000 live births

Answer

7.2 per 1000 live births

Question 3

An ELISA test has a sensitivity of 95% and a specificity of 95%. If the prevalence of HIV infection in a population is 5%, what is the positive predictive value of the test?

Accepted Answer

approx. 9.5%

Answer

95%

Answer

100%

Answer

75%

Question 4

If the confidence limit is increased, what happens to the significance of the data?

Accepted Answer

Previously significant data becomes insignificant

Answer

Previously insignificant data becomes significant

Answer

No effect on significance

Answer

Any change can happen

Question 5

If the correlation between weight and height is very strong, what will be the correlation coefficient?

Accepted Answer

1

Answer

Greater than 1

Answer

0

Answer

None of the above

Question 6

What statistical method is used to calculate the death rate between two populations with different age groups?

Accepted Answer

Standardized death rate

Answer

Crude death rate

Answer

Case fatality rate

Answer

Age-specific death rate

Question 7

All are true regarding Student's t-test, except?

Accepted Answer

Standard error of the mean is not estimated.

Answer

A standard population is selected.

Answer

Two samples are compared.

Answer

Student's t-distribution table is required for calculation.

Question 8

For international comparisons, what is the perinatal mortality rate calculated as the sum of late fetal deaths (28 weeks gestation or more) plus early neonatal deaths (first week) in a year per?

Accepted Answer

1,000 live births

Answer

100 live births

Answer

10,000 live births

Answer

1,00,000 live births

Question 9

In which aspect is a case-control study superior to a cohort study?

Accepted Answer

Shorter study duration

Answer

Reduced likelihood of bias

Answer

Greater accuracy

Answer

Ability to calculate relative risk

Question 10

Susceptibility bias is eliminated by:

Accepted Answer

Randomization

Answer

Single blinding

Answer

Double blinding

Answer

Matching

Biostatistics — MCQs

Biostatistics — MCQs

On this page

Practice by Chapter

Want unlimited practice?