Home | | Medicine Study Notes | Risks and Odds - Evidence Based Medicine

Risks and Odds - Evidence Based Medicine

Event Rate: proportion of patients in a group in whom an event is observed. Applied to Controls and Experimental groups -> CER and EER

Risks and Odds

·        Event Rate: proportion of patients in a group in whom an event is observed. Applied to Controls and Experimental groups ® CER and EER

·        Relative Risk = (A/(A+B))/(C/(C+D)) = EER/CER

·        Absolute Risk Reduction (ARR) = C/(C+D) – A/(A+B) = CER - EER

·        Relative Risk Reduction: percent reduction in events in the treat group event rate compared to the control group = (CER – EER)/CER * 100 = (C/(C+D) – A/(A+B))/(C/(C+D))

·        Risk Ratio = EER/CER Odds: ratio of events to non-events

·        Odds ratio: odds of an experimental patient suffering an adverse event relative to a control patient = (A/C)/(B/D)

·         Number needed to treat (NNT): number of patients needing treatment to achieve one favourable outcome = 1 /ARR – always rounded up to the nearest whole number and accompanied by the 95% CI

·        Number needed to harm (NNH): number of patients who need to be treated to achieve one adverse outcome = 1/Absolute Risk Increase (ARI = EER - CER)

·        RRR and OR do not say anything about absolute risk. An RR of 30% can mean a risk reduction from 60% to 20%, or from 3% to 1%. The ARR and NNT varies dramatically

·        Time frame: all measures (RR, RRR, ARR, OR) must be qualified by giving them a time frame (e.g. the length of the period of the study)

Study Material, Lecturing Notes, Assignment, Reference, Wiki description explanation, brief detail
Medicine Study Notes : Evidence Based Medicine : Risks and Odds - Evidence Based Medicine |