Clinical Informatics Certification Clinical Informatics Data Analytics 2 — Questions and Answers
Question 1: Which statistical measure is most appropriate for describing the central tendency of a skewed clinical dataset such as length of hospital stay?
- Mean
- Median (Correct answer)
- Mode
- Standard deviation
Correct answer: Median
The median is preferred for skewed distributions because it is not distorted by extreme outliers the way the mean is.
Question 2: In clinical NLP pipelines, what is the primary purpose of named entity recognition (NER)?
- Translate clinical notes into structured codes
- Identify and classify clinical concepts such as diagnoses and medications in free text (Correct answer)
- Summarize discharge summaries automatically
- Remove protected health information from documents
Correct answer: Identify and classify clinical concepts such as diagnoses and medications in free text
NER extracts and categorizes clinical entities like diagnoses, drugs, and procedures from unstructured text.
Question 3: A hospital wants to predict 30-day readmission risk. Which type of machine learning model outputs a probability score between 0 and 1?
- K-means clustering
- Logistic regression (Correct answer)
- Association rule mining
- Principal component analysis
Correct answer: Logistic regression
Logistic regression produces probability estimates between 0 and 1, making it well suited for binary classification tasks like readmission prediction.
Question 4: Which data quality dimension refers to how well data reflects the real-world construct it is intended to measure?
- Completeness
- Timeliness
- Validity (Correct answer)
- Concordance
Correct answer: Validity
Validity (also called accuracy or correctness) measures whether recorded values truly represent the underlying clinical reality.
Question 5: When building a predictive model with highly imbalanced classes (e.g., 1% positive cases), which evaluation metric is most informative?
- Overall accuracy
- Area under the ROC curve (AUROC) (Correct answer)
- Mean squared error
- R-squared
Correct answer: Area under the ROC curve (AUROC)
AUROC evaluates model discrimination across all classification thresholds and is robust to class imbalance, unlike overall accuracy.
Question 6: In a clinical data warehouse, a 'fact table' typically contains which type of information?
- Descriptive attributes about patients such as name and address
- Quantitative measurements and foreign keys to dimension tables (Correct answer)
- Master code lists for ICD and CPT codes
- Audit logs of database transactions
Correct answer: Quantitative measurements and foreign keys to dimension tables
In a star schema, the fact table stores numeric measures and keys that link to surrounding dimension tables.
Question 7: Which technique is used to handle missing clinical data by filling in plausible values based on observed data distributions?
- Oversampling
- Imputation (Correct answer)
- Normalization
- Tokenization
Correct answer: Imputation
Imputation replaces missing values with estimated values derived from the rest of the dataset, such as mean, median, or model-predicted values.
Clinical Informatics Certification Clinical Informatics Data Analytics 2
Which statistical measure is most appropriate for describing the central tendency of a skewed clinical dataset such as length of hospital stay?