Clinical Informatics Certification Data Analytics & Population Health 2 — Questions and Answers
Question 1: Which statistical method is most appropriate for identifying patient subgroups with similar clinical characteristics from EHR data?
- Logistic regression
- K-means clustering (Correct answer)
- Chi-square test
- Kaplan-Meier analysis
Correct answer: K-means clustering
K-means clustering is an unsupervised machine learning technique used to segment patients into homogeneous groups based on shared clinical features.
Question 2: A population health program wants to measure the rate of new diabetes diagnoses in a defined community over one year. Which epidemiological measure should be used?
- Prevalence
- Incidence rate (Correct answer)
- Odds ratio
- Attributable risk
Correct answer: Incidence rate
Incidence rate measures the number of new cases of a disease occurring in a population over a specific time period.
Question 3: In risk stratification models for care management, which metric best indicates a model's ability to distinguish high-risk from low-risk patients?
- Sensitivity alone
- Specificity alone
- Area Under the ROC Curve (AUC) (Correct answer)
- Positive predictive value
Correct answer: Area Under the ROC Curve (AUC)
AUC summarizes the model's discriminative ability across all classification thresholds, making it the standard metric for risk stratification performance.
Question 4: What is the primary purpose of a Master Patient Index (MPI) in population health data management?
- Store clinical notes from encounters
- Uniquely identify and link patient records across systems (Correct answer)
- Calculate risk scores for chronic disease
- Track immunization schedules
Correct answer: Uniquely identify and link patient records across systems
An MPI assigns unique identifiers to patients and links their records across disparate healthcare systems to ensure accurate patient matching.
Question 5: A health system notices its readmission prediction model performs well overall but poorly for Hispanic patients. This is an example of:
- Overfitting
- Algorithmic bias (Correct answer)
- Confounding
- Selection bias in training data only
Correct answer: Algorithmic bias
Algorithmic bias occurs when a model produces systematically skewed outcomes for specific demographic subgroups due to underrepresentation or proxy variables in training data.
Question 6: Which data visualization type is best suited for displaying the distribution of hemoglobin A1c values across a diabetic patient population?
- Pie chart
- Histogram (Correct answer)
- Scatter plot
- Gantt chart
Correct answer: Histogram
A histogram displays the frequency distribution of a continuous variable, making it ideal for showing the spread of HbA1c values in a population.
Question 7: Social determinants of health (SDOH) data is increasingly integrated into population health platforms. Which of the following is a recognized SDOH domain?
- Allergy history
- Surgical history
- Food insecurity (Correct answer)
- Medication adherence rates
Correct answer: Food insecurity
Food insecurity is a core SDOH domain under economic stability, as defined by Healthy People 2030 and incorporated into population health frameworks.
Clinical Informatics Certification Data Analytics & Population Health 2
Which statistical method is most appropriate for identifying patient subgroups with similar clinical characteristics from EHR data?