Clinical Informatics Data Analytics Flashcards
7 cards from real Clinical Informatics Certification practice questions. Tap to flip, then mark Knew It or Still Learning — missed cards come back until you master them.
Read the first 7 Clinical Informatics Data Analytics flashcards as text
Which statistical measure is most appropriate for describing the central tendency of a skewed clinical dataset such as length of hospital stay?
Answer: Median
The median is preferred for skewed distributions because it is not distorted by extreme outliers the way the mean is.
In clinical NLP pipelines, what is the primary purpose of named entity recognition (NER)?
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.
A hospital wants to predict 30-day readmission risk. Which type of machine learning model outputs a probability score between 0 and 1?
Answer: Logistic regression
Logistic regression produces probability estimates between 0 and 1, making it well suited for binary classification tasks like readmission prediction.
Which data quality dimension refers to how well data reflects the real-world construct it is intended to measure?
Answer: Validity
Validity (also called accuracy or correctness) measures whether recorded values truly represent the underlying clinical reality.
When building a predictive model with highly imbalanced classes (e.g., 1% positive cases), which evaluation metric is most informative?
Answer: Area under the ROC curve (AUROC)
AUROC evaluates model discrimination across all classification thresholds and is robust to class imbalance, unlike overall accuracy.
In a clinical data warehouse, a 'fact table' typically contains which type of information?
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.
Which technique is used to handle missing clinical data by filling in plausible values based on observed data distributions?
Answer: Imputation
Imputation replaces missing values with estimated values derived from the rest of the dataset, such as mean, median, or model-predicted values.