Data Analytics and Informatics
Data Analytics and Informatics tests quantitative reasoning with healthcare data and knowledge of the tools that produce and exchange it. Items ask you to compute a rate or index, interpret a comparison, select a sampling method, or judge how much trust to place in an informatics output.
Measurement basics
- Data types: nominal, ordinal, interval, ratio. Means are appropriate for interval and ratio data; medians suit skewed distributions.
- Rates = numerator ÷ denominator × k, with matched time periods and populations. Inpatient mortality and readmission rates use discharges as the denominator, not admissions.
- Central tendency and spread: mean, median, mode, range, standard deviation, percentiles.
- Validity means measuring what is intended; reliability means producing consistent results on repeat measurement.
Analytics and benchmarking
Analytics maturity runs descriptive, diagnostic, predictive, prescriptive. Benchmarking compares your rate to an external standard, and dashboards track key performance indicators over time. Risk adjustment compares observed to expected events (O/E); case mix index reflects resource intensity. Control charts distinguish common-cause variation from special-cause signals; run charts show trend.
Audit sampling
Random sampling supports statistical inference; convenience samples do not. Statistical sample size depends on confidence level, margin of error, and expected error rate. Extrapolation of findings is defensible for statistical samples and is used by federal auditors; nonstatistical samples generally are not extrapolated.
Informatics landscape
EHRs, health information exchanges, clinical decision support, registries, telehealth, and computer-assisted coding using natural language processing. CAC suggests codes but the credentialed coder remains accountable for accuracy and for resolving conflicts.
Worked example
A hospital reports 36 deaths among 1,200 discharges: 36 ÷ 1,200 = 3.0% observed. The risk-adjusted expected rate is 2.5%, so O/E = 1.20 - 20% more deaths than expected. Before escalating, check that the denominator excludes observation patients and that the period matches; confirm the difference is statistically stable given small monthly counts; and verify whether increased secondary diagnosis capture raised the expected rate rather than outcomes changing.
Traps that cost marks
Reporting counts instead of rates; using admissions for a discharge-based measure; comparing means when outliers skew the data; drawing conclusions from a single month of small numbers; extrapolating a convenience sample; and accepting CAC output without coder review.