Data Analytics and Informatics
From raw data to decisions
Analytics questions reward candidates who can identify the type of analysis being described and the quality problem that would invalidate it. Learn the ladder: descriptive (what happened), diagnostic (why), predictive (what will happen), and prescriptive (what should we do). Most operational dashboards are descriptive; risk stratification and readmission models are predictive.
Know basic statistics vocabulary: mean, median, mode, range, standard deviation, rates, ratios, proportions, and percent change. Understand the difference between a measure and an indicator, and how benchmarking, dashboards, and scorecards differ. A dashboard monitors key indicators in real time; a scorecard tracks progress against strategic targets.
Informatics content includes the EHR, clinical decision support, computerized provider order entry, health information exchange models (directed, query-based, consumer-mediated), and interoperability standards. Know that clinical decision support is most effective when it is actionable, timely, and not excessive; alert fatigue is a recognized failure mode.
Data visualization and reporting traps are common: confusing correlation with causation, using the wrong denominator, or presenting a rate without its base. Population health analytics depends on attribution and risk adjustment, so questions may test whether you recognize that comparing crude rates across providers with different case mixes is misleading.
Worked example: A clinic reports 30 readmissions out of 600 discharges. The rate is 30 ÷ 600 = 5%. If a question offers "30/570 = 5.3%," the error is using post-discharge survivors as the denominator. Always ask what population is at risk.
Trap: Items that describe a predictive model but call it descriptive, or that treat a data quality problem (missing race/ethnicity fields) as a statistical problem rather than a documentation and workflow problem.