Biostatistics Practice Questions

Q: What is true about the reliability of a test?

It gives the same results on repeated tests.. **Explanation:** **Reliability** (also known as precision, consistency, or reproducibility) refers to the ability of a test to yield the same results when repeated under similar conditions. In biostatistics, it measures how free a test is from random error. If a test is reliable, repeated measurements on the same stable subject will produce consistent values. **Analysis of Options:** * **Option C (Correct):** This is the literal definition of reliability. It ensures that the results are stable over time and across different observers (inter-rater reliability). * **Option A:** Incorrect. Consistency and reproducibility are the *core* components of reliability; they are the primary focus, not a "non-problem." * **Option B:** Incorrect. While an investigator's skill can reduce observer bias, reliability is a property of the test/instrument itself. High reliability should ideally minimize the impact of an individual investigator's subjective knowledge. * **Option D:** Incorrect. This describes "Validity." Validity (accuracy) is the extent to which a test measures what it is actually intended to measure. **High-Yield NEET-PG Pearls:** 1. **Reliability vs. Validity:** A test can be reliable but not valid (e.g., a weighing scale that consistently shows 5kg extra is reliable but inaccurate). However, for a test to be highly valid, it must be reliable. 2. **Evaluation:** Reliability is measured using the **Kappa Coefficient** (for qualitative data) or **Intraclass Correlation Coefficient** (for quantitative data). 3. **Factors affecting Reliability:** * **Observer Variation:** Intra-observer (same person) vs. Inter-observer (different people). * **Biological Variation:** Changes in the parameter being measured (e.g., BP fluctuations). * **Instrument Error:** Faulty equipment.

Q: A drug company is developing a new pregnancy-test kit for use on an outpatient basis. The company used the pregnancy test on 100 women who are known to be pregnant. Out of 100 women, 99 showed positive test. Upon using the same test on 100 non-pregnant women, 90 showed negative result. What is the sensitivity of the test?

99%. ### Explanation **1. Why the Correct Answer (B) is Right:** Sensitivity is defined as the ability of a test to correctly identify those **with the disease** (True Positives). It is calculated using the formula: $$\text{Sensitivity} = \frac{\text{True Positives (TP)}}{\text{Total Diseased (TP + FN)}} \times 100$$ In this scenario: * **Total pregnant women (Diseased):** 100 * **Positive results (True Positives):** 99 * **Sensitivity:** $(99 / 100) \times 100 = 99\%$ The test correctly identified 99% of the women who were actually pregnant. **2. Why the Incorrect Options are Wrong:** * **Option A (90%):** This represents the **Specificity** of the test. Specificity is the ability to correctly identify those **without the disease** (True Negatives). Here, 90 out of 100 non-pregnant women tested negative ($90/100 = 90\%$). * **Option C (Average):** Sensitivity and Specificity are independent properties of a diagnostic test. Averaging them has no statistical significance in evaluating test performance. * **Option D:** The data provided is sufficient. We have the "Gold Standard" status (known pregnant vs. non-pregnant) and the test results, which are the only requirements to build a 2x2 contingency table. **3. Clinical Pearls & High-Yield Facts for NEET-PG:** * **SNOUT:** **S**ensitivity rules **OUT** a disease when the result is negative (useful for screening). * **SPIN:** **S**pecificity rules **IN** a disease when the result is positive (useful for confirmation). * **Sensitivity** is also known as the **True Positive Rate**. * **False Negative Rate** = $100 - \text{Sensitivity}$. In this case, it is 1%. * **False Positive Rate** = $100 - \text{Specificity}$. In this case, it is 10%.

Q: Which of the following statements is true regarding epidemiological study designs?

Cohort studies involve a large number of subjects.. ### Explanation **Why Option B is Correct:** Cohort studies are **prospective** in nature, starting with a group of exposed and non-exposed individuals and following them over time to observe the development of an outcome (disease). Because the incidence of most diseases is relatively low, a **large sample size** is mandatory to ensure enough cases occur to achieve statistical significance. This makes them resource-intensive but ideal for calculating **Incidence** and **Relative Risk**. **Analysis of Incorrect Options:** * **Option A:** Case-control studies are actually the **study of choice for rare diseases**. Since they start with "cases" (people who already have the disease), researchers do not have to wait for rare events to occur naturally. * **Option C:** Case-control studies are **inexpensive** and easy to conduct because they use existing records or interviews and require smaller sample sizes compared to cohort studies. * **Option D:** Cohort studies are **time-consuming** because they require a long follow-up period (attrition/loss to follow-up is a major challenge). Case-control studies yield quicker results as the outcome has already occurred. **NEET-PG High-Yield Pearls:** * **Directionality:** Cohort is "Cause to Effect" (Prospective); Case-control is "Effect to Cause" (Retrospective). * **Measures of Association:** * Cohort Study $\rightarrow$ **Relative Risk (RR)** and Attributable Risk. * Case-control Study $\rightarrow$ **Odds Ratio (OR)**. * **Bias:** Case-control studies are highly prone to **Recall Bias**, while Cohort studies are prone to **Selection/Attrition Bias**. * **Nested Case-Control Study:** A hybrid design where a case-control study is conducted within an ongoing cohort study; it is more cost-effective than a full cohort.

Q: Selection bias occurs mainly in which type of study?

Case-control study. **Explanation:** **Selection bias** occurs when the study population does not accurately represent the target population due to the way participants are selected. **Why Case-Control Study is the correct answer:** In a case-control study, the researcher selects participants based on the **outcome** (presence or absence of disease) and then looks backward for exposure. This design is highly susceptible to selection bias because both the "cases" and "controls" are often chosen from hospital settings (Berkson’s bias) or through non-random methods. If the criteria for selecting cases differ systematically from those for selecting controls (other than the disease itself), the relationship between exposure and outcome becomes distorted. **Analysis of Incorrect Options:** * **Cohort Study:** While selection bias can occur (e.g., Healthy Worker Effect), these studies are more prone to **Follow-up bias (Attrition bias)** because they track participants over time. * **Randomized Controlled Trial (RCT):** The hallmark of an RCT is **Randomization**, which is specifically designed to eliminate selection bias by ensuring that every participant has an equal chance of being assigned to any group, thereby balancing both known and unknown confounders. * **All of the above:** While bias can exist in any study, it is "mainly" or most characteristically associated with Case-Control designs in epidemiological hierarchy. **High-Yield Clinical Pearls for NEET-PG:** * **Berkson’s Bias (Admission Rate Bias):** A specific type of selection bias occurring in hospital-based case-control studies. * **Neyman Bias (Prevalence-Incidence Bias):** Occurs when cases are selected from survivors (common in cross-sectional/case-control studies of fatal diseases). * **Recall Bias:** The most common type of **Information Bias** in case-control studies (not to be confused with selection bias). * **Gold Standard to eliminate Selection Bias:** Randomization.

Q: Which of the following statements about meta-analysis is FALSE?

It is easy to conduct.. ### Explanation **1. Why Option A is the Correct (False) Statement:** Meta-analysis is considered the highest level of evidence in the hierarchy of evidence-based medicine (EBM). Contrary to being "easy," it is a **complex, rigorous, and time-consuming** statistical process. It requires a systematic review of literature, stringent inclusion/exclusion criteria, assessment of study quality (risk of bias), and sophisticated statistical software to pool data. Therefore, the statement that it is "easy to conduct" is incorrect. **2. Analysis of Other Options:** * **Option B (Analysis of multiple analyses):** This is the literal definition of meta-analysis. It uses statistical methods to combine data from multiple independent studies (usually RCTs) to reach a single quantitative conclusion with higher statistical power. * **Option C (Funnel and Forest plots):** These are the hallmark graphical tools of meta-analysis. * **Forest Plot:** Displays the results of individual studies and the pooled estimate (represented by a diamond). * **Funnel Plot:** Used primarily to detect **publication bias**. * **Option D (Apples-and-oranges effect):** This refers to a common criticism/limitation of meta-analysis. It occurs when a researcher pools studies that are too clinically or methodologically diverse (heterogeneous) to be meaningfully combined, leading to invalid conclusions. **3. High-Yield Clinical Pearls for NEET-PG:** * **Hierarchy of Evidence:** Meta-analysis of RCTs > Systematic Review > RCT > Cohort > Case-Control > Case Series > Case Report. * **Heterogeneity:** Measured using the **I² statistic**. High I² (>50-75%) suggests studies are too different to pool. * **Publication Bias:** If the funnel plot is asymmetrical, it suggests publication bias (small negative studies are often not published). * **Fixed vs. Random Effects Model:** Used to pool data depending on the presence of heterogeneity.

Q: For a population of 10,000, what does a sex ratio of more than 1000 indicate?

Males are less than 5000. ### Explanation **1. Understanding the Concept** In Indian Demography and Biostatistics, the **Sex Ratio** is defined as the number of females per 1,000 males. * **Formula:** (Number of Females / Number of Males) × 1,000 If the sex ratio is **more than 1,000**, it mathematically implies that the numerator (Females) is greater than the denominator (Males). * In a population of 10,000, if the ratio is exactly 1,000, there are 5,000 males and 5,000 females. * If the ratio is **>1,000**, females must be **>5,000** and males must be **<5,000**. Therefore, Option D is the correct logical conclusion. **2. Analysis of Incorrect Options** * **Option A & B:** These options suggest a total population count far below the given 10,000. If males or females were less than 500, the other gender would have to be over 9,500, which is an extreme outlier and does not specifically define a ratio "more than 1,000." * **Option C:** If females were less than 5,000 in a population of 10,000, the males would be more than 5,000. This would result in a sex ratio of **less than 1,000** (a deficit of females), which contradicts the question. **3. High-Yield Clinical Pearls for NEET-PG** * **Child Sex Ratio:** Calculated for the 0–6 years age group. * **Global vs. Indian Definition:** In most Western countries, the sex ratio is expressed as males per 100 females. However, for NEET-PG, always follow the **Indian Census definition**: Females per 1,000 males. * **Highest Sex Ratio (Census 2011):** Kerala (1,084). * **Lowest Sex Ratio (Census 2011):** Haryana (879). * **Overall India Sex Ratio (Census 2011):** 943.

Q: Which statistical test is used for counting the direction of differences within a paired sample?

Sign test. ### Explanation **Correct Answer: D. Sign test** The **Sign test** is a non-parametric test used to analyze the **direction of differences** between paired observations (e.g., "before and after" measurements). It focuses solely on whether the difference is positive (+) or negative (-) rather than the actual magnitude of the change. It is used when the data is ordinal or when the distribution of differences is not normal, making it the simplest test for paired data. **Why other options are incorrect:** * **A. t-test (Paired):** This is a parametric test used for paired samples. Unlike the Sign test, it requires the data to be on an interval/ratio scale and assumes a **normal distribution**. It considers the magnitude of the difference, not just the direction. * **B. Z-test:** This is used for large samples (n > 30) to compare means or proportions. It is not specifically designed for "counting directions" in paired samples. * **C. F-test (ANOVA):** This test is used to compare the **variances** of two populations or to compare means among three or more groups. It does not analyze paired directional differences. **High-Yield Clinical Pearls for NEET-PG:** * **Sign Test vs. Wilcoxon Signed-Rank Test:** While both are non-parametric tests for paired data, the Sign test only looks at the **direction** (+/-), whereas the Wilcoxon Signed-Rank test considers both the **direction and the magnitude** (ranks). * **Non-Parametric Equivalents:** * Unpaired t-test → Mann-Whitney U test. * Paired t-test → Wilcoxon Signed-Rank test or Sign test. * One-way ANOVA → Kruskal-Wallis test. * **Rule of Thumb:** If the question mentions "direction only" or "nominal/ordinal paired data," think **Sign Test**.

Q: In the presence of unusually high outliers, which is the preferred measure of central tendency?

Median. ### Explanation **1. Why Median is the Correct Answer:** In biostatistics, the **Median** is the "positional average" that divides a distribution into two equal halves. Its primary advantage is that it is **robust to outliers** (extreme values). When a dataset contains unusually high or low values, the distribution becomes **skewed**. While the Mean is pulled toward the tail of the skew, the Median remains stable because it depends on the rank order of observations rather than their numerical magnitude. Therefore, for skewed data, the Median provides a more accurate representation of the "typical" value. **2. Analysis of Incorrect Options:** * **A. Mean:** This is the arithmetic average. It is highly sensitive to outliers because every value in the dataset is used in its calculation. A single extreme outlier can significantly inflate or deflate the Mean, making it a poor measure for skewed distributions. * **B. Mode:** This is the most frequently occurring value. While it is not affected by outliers, it is often unstable (a dataset can be bimodal or have no mode at all), making it less reliable than the Median for general central tendency. * **D. Range:** This is a measure of **dispersion** (spread), not central tendency. It is calculated as the difference between the maximum and minimum values and is, in fact, the measure *most* affected by outliers. **3. High-Yield Clinical Pearls for NEET-PG:** * **Normal Distribution:** Mean = Median = Mode. * **Positive Skew (Right-tailed):** Mean > Median > Mode (e.g., income levels, incubation periods). * **Negative Skew (Left-tailed):** Mode > Median > Mean (e.g., age at death in developed countries). * **Best measure for Nominal data:** Mode. * **Best measure for Ordinal data:** Median. * **Best measure for Interval/Ratio data (No outliers):** Mean.

Q: In a population, the total number of births in a year is 3050. There are 50 stillbirths, 100 neonatal deaths within the first 7 days, and 150 deaths between the 8th and 28th day of life. Calculate the Neonatal Mortality Rate.

83. ### Explanation **1. Understanding the Correct Answer (C: 83)** To calculate the **Neonatal Mortality Rate (NMR)**, we must understand its definition: the number of deaths of live-born infants during the first 28 completed days of life per 1,000 live births. * **Step 1: Identify Live Births.** The question gives "Total Births" (3050) and "Stillbirths" (50). Since Live Births = Total Births – Stillbirths, we have: $3050 - 50 = 3000$ live births. * **Step 2: Identify Neonatal Deaths.** This includes both Early Neonatal Deaths (0-7 days) and Late Neonatal Deaths (8-28 days). Total deaths = $100 + 150 = 250$. * **Step 3: Apply Formula.** $$\text{NMR} = \frac{\text{Total Neonatal Deaths}}{\text{Total Live Births}} \times 1000$$ $$\text{NMR} = \frac{250}{3000} \times 1000 = \frac{250}{3} \approx \mathbf{83.33}$$ **2. Why Other Options are Incorrect** * **Option A (250):** This is the absolute number of neonatal deaths, not the rate per 1,000 live births. * **Option B (100):** This only accounts for Early Neonatal Deaths (0-7 days). * **Option D (90):** This is a distractor resulting from calculation errors or using the wrong denominator (e.g., using total births instead of live births). **3. Clinical Pearls & High-Yield Facts for NEET-PG** * **Denominator Rule:** Most mortality rates in the first year of life (Infant, Neonatal, Post-neonatal) use **Live Births** as the denominator. The **Perinatal Mortality Rate** is a notable exception that uses **Total Births** (Live births + Stillbirths). * **Early vs. Late:** Early Neonatal Mortality (0-7 days) reflects maternal health and quality of obstetric care, while Late Neonatal Mortality (8-28 days) often reflects environmental factors and infections. * **Current Trend:** In India, Neonatal Mortality accounts for approximately **70% of the Infant Mortality Rate (IMR)**, making it the most critical target for reducing under-5 mortality.

Question 1

What is true about the reliability of a test?

Accepted Answer

It gives the same results on repeated tests.

Answer

Consistency and reproducibility of the test are not a problem.

Answer

Investigator's knowledge is important.

Answer

It is the extent of variation of measurement of contained behaviour.

Question 2

A drug company is developing a new pregnancy-test kit for use on an outpatient basis. The company used the pregnancy test on 100 women who are known to be pregnant. Out of 100 women, 99 showed positive test. Upon using the same test on 100 non-pregnant women, 90 showed negative result. What is the sensitivity of the test?

Accepted Answer

99%

Answer

90%

Answer

Average of 90% and 99%

Answer

Cannot be calculated from the given data

Question 3

Which of the following statements is true regarding epidemiological study designs?

Accepted Answer

Cohort studies involve a large number of subjects.

Answer

Case-control studies are not suitable for rare diseases.

Answer

Case-control studies are relatively expensive.

Answer

Cohort studies yield relatively quick results.

Question 4

Selection bias occurs mainly in which type of study?

Accepted Answer

Case-control study

Answer

Cohort study

Answer

Randomized controlled trial

Answer

All the above

Question 5

Which of the following statements about meta-analysis is FALSE?

Accepted Answer

It is easy to conduct.

Answer

It is an analysis of multiple analyses.

Answer

Funnel plots and forest plots are used in its interpretation.

Answer

An 'apple-oranges' effect can be observed in it.

Question 6

For a population of 10,000, what does a sex ratio of more than 1000 indicate?

Accepted Answer

Males are less than 5000

Answer

Males are less than 500

Answer

Females are less than 500

Answer

Females are less than 5000

Question 7

Which statistical test is used for counting the direction of differences within a paired sample?

Accepted Answer

Sign test

Answer

t test

Answer

Z test

Answer

F test

Question 8

In the presence of unusually high outliers, which is the preferred measure of central tendency?

Accepted Answer

Median

Answer

Mean

Answer

Mode

Answer

Range

Question 9

In a population, the total number of births in a year is 3050. There are 50 stillbirths, 100 neonatal deaths within the first 7 days, and 150 deaths between the 8th and 28th day of life. Calculate the Neonatal Mortality Rate.

Accepted Answer

83

Answer

250

Answer

100

Answer

90

Question 10

ELISA is performed on a population with low prevalence. What would be the result of performing double screening ELISA tests?

Accepted Answer

Increased specificity and positive predictive value

Answer

Increased sensitivity and positive predictive value

Answer

Increased sensitivity and negative predictive value

Answer

Increased specificity and negative predictive value

Biostatistics — MCQs

Biostatistics — MCQs

On this page

Practice by Chapter

Want unlimited practice?