Surveillance and Epidemiologic Investigation
Domain 2 - Surveillance and Epidemiologic Investigation
This domain tests the mechanics of counting infections correctly and of running an outbreak investigation: standardized case definitions, correct denominators, and a disciplined sequence of epidemiologic steps.
Core rules
- Use a written case definition (NHSN/CDC or locally validated) applied consistently; changing a definition mid-year breaks comparability.
- Incidence counts new cases in a period; prevalence counts all existing cases at a point in time. Never mix them.
- Use the right denominator: device-associated rates per 1,000 device-days; SSI by procedure with the NHSN risk index; MRSA bacteremia per 1,000 patient-days.
- Standardized infection ratio (SIR) = observed ÷ expected, risk-adjusted against a baseline; SIR <1 means fewer infections than predicted, not "safe."
- Outbreak steps: confirm the diagnosis and rule out laboratory error, verify the excess, define a case, build a line list, describe by time-place-person, generate hypotheses, test them (cohort or case-control), implement control, communicate, then evaluate.
Worked example
Two weeks of Serratia isolates in a 20-bed ICU: 6 cases among 75 admitted patients. Attack rate = 6 ÷ 75 = 8%. Baseline for the prior year was 1–2 per month, so the excess is real. The epidemic curve shows a sharp point-source pattern over 4 days, pointing to a common exposure - a shared item or fluid rather than person-to-person spread. Cases cluster in rooms supplied by one medication refrigerator, so the team samples the implicated product and stops its use while the investigation continues.
Traps
Using patient-days when the measure requires device-days; counting a positive culture drawn on admission as an HAI; reporting an SIR without its confidence interval; skipping "verify the diagnosis" and chasing a pseudo-outbreak caused by contaminated laboratory reagents or surveillance artifacts; waiting for perfect data before intervening.