NEW
Font size
WorksheetsClinicomm set 2
Total questions: 30
Worksheet time: 30mins
In a hospital outbreak investigation of carbapenem-resistant Klebsiella pneumoniae, which study design best determines the risk factors for colonization among ICU patients while controlling for confounders?
Cross-sectional
Case-control with matching
Prospective cohort
Ecological
During an RCT assessing a new rapid sepsis screening tool, interim analysis shows borderline benefit (p=0.049). Which approach avoids inflated Type I error?
Original p-value
Bonferroni
Sequential boundaries (O'Brien-Fleming)
Stop now
Non-inferiority trial of two HAP regimens: CI for difference in cure rates lies entirely within the non-inferiority margin. Interpretation?
Inferior
Non-inferior
Superior
Inconclusive
Early discharge protocol using troponin: NPV is 99% in low-risk chest pain. What most affects NPV?
Sensitivity
Specificity
Disease prevalence
Sample size
Hospital hand hygiene observations: Hawthorne effect will most likely…
Inflate observed compliance
Deflate
No effect
Increase random error
Oncology survival analysis violates proportional hazards. Best alternative?
Log-rank only
Cox with time-varying covariates/stratified Cox
Paired t-test
Ignore
ICU sepsis bundle trial: RR=0.85 (95% CI: 0.60-1.02).
Statistically significant
Clinically meaningful but not statistically significant
No relevance
Both
Retrospective decade-long liver transplant outcomes at a tertiary center: most likely bias?
Recall
Lead-time
Selection bias
Measurement
Time-series of monthly CAUTI rates before/after silver catheters. Best method for autocorrelation?
Chi-square
ARIMA
ANOVA
Logistic regression
Hip-rehab trial with large loss to follow-up in one arm. Most robust analysis?
Per-protocol
Intention-to-treat
On-treatment
Sensitivity only
Septic shock model: OR for prior statin use = 0.6 (p=0.049). Key consideration before inferring causality?
Lack of dose–response relationship
Small sample size/power issues
Residual confounding/model robustness
Selective reporting bias
Stroke recurrence model shows significant interaction (HTN × AF). Epidemiologic interpretation?
Confounding by indication
Effect modification
Measurement error in exposure
Multicollinearity in the model
ROC for new troponin assay AUC=0.55.
Excellent
Acceptable
Poor
Perfect
Perioperative beta-blocker trial RR=1.25 for hypotension. To compute PAR%, you need:
Se/Sp
Exposure prevalence and RR
Incidence density
p-value
ICU ventilation strategies: NNH=50 means…
1 extra harm per 50 treated
50% risk increase
Harm in 50%
No meaning
Most sensitive early metric for QI changes?
Mortality
Process indicators
Life expectancy
Crude incidence
Adjusted RR differs substantially from crude RR. Most likely reason?
Effect modification
Random error
Confounding
Chance
Diagnostic accuracy: LR+ = 10 implies…
Rule out
Minimal value
Strong rule-in
No change
Best for inter-rater agreement between radiologists?
Pearson r
Kappa statistic
Chi-square
Bland-Altman
Case-control: OR=4.5 for smoking and post-op pneumonia.
Protective
Odds ↑ by ~350%
RR=4.5
OR=RR always
ED AI triage during pandemic-primary ethical epidemiologic concern?
Selection bias
Algorithmic bias
Measurement bias
Recall bias
Stewardship outcomes pre/post policy: design to reduce secular trends?
Pre-post
Interrupted time series with control
Cross-sectional
Ecological
Clinical audit where data collectors know allocation-most likely error?
Information (ascertainment) bias
Selection
Confounding
Random error
Hospital reports zero CLABSI for a year. Best interpretation?
True elimination
Possible under-reporting/surveillance bias
Proof of success
Chance
Best example of lead-time bias?
Earlier detection without improved survival
Late diagnosis w/ improved survival
Screening reduces mortality
Drug RCT
New burns therapy improves PROs but not mortality. This reflects:
Surrogate success
Process success
QOL improvement
All of the above
Evaluating nurse staffing vs mortality while adjusting for severity-best analysis?
Stratified
Multivariable regression
Paired t-test
Chi-square
Most reliable source for nosocomial infection rates?
Admission register
Electronic infection surveillance system
Staff interviews
Patient self-reports
Teaching vs non-teaching hospital outcomes-main challenge?
Info bias
Confounding by case mix
Recall bias
Lead-time bias
Checklist reduces SSI 4% → 2% (p=0.04). This represents:
50% RRR
2% ARR
Both a & b
None
