Data Management and Analytics
What to know
Data management covers the data lifecycle (capture, validation, storage, retrieval, analysis, reporting) and data quality dimensions: accuracy, completeness, consistency, timeliness, validity, uniqueness. Know the data warehouse vs data mart distinction and the role of data governance (ownership, stewardship, stewardship councils). Analytics tiers: descriptive (what happened), diagnostic (why), predictive (what will happen), prescriptive (what to do). Familiarity with quality measures (e.g., eCQM, HEDIS-like concepts), benchmarking, and statistical literacy-rates vs counts, numerators/denominators, sensitivity/specificity-is expected. Understand secondary use of data for research and population health, and de-identification (HIPAA Safe Harbor vs expert determination).
Traps
- Reporting a count when a rate is needed (or vice versa) and mis-stating the denominator.
- Confusing data governance (policy/accountability) with data management (operational execution).
- Treating predictive analytics as certain-models yield probabilities, not guarantees.
- Forgetting that de-identification must remove all 18 HIPAA identifiers under Safe Harbor.
Worked example / rule of thumb
A unit wants to reduce falls. Start with descriptive (fall counts and rates per 1,000 patient-days), move to diagnostic (which units/shifts), then predictive (risk model), then prescriptive (targeted interventions). Rule of thumb: always compute the rate = numerator ÷ denominator and state the denominator.
Key concepts
Data lifecycle, data quality dimensions, warehouse vs mart, governance vs management, analytics tiers, quality measures, de-identification.