1. Are the interventions comparable?

What were the interventions?

1.a. intervention description, as II.D.1.a

Were there any co-interventions? Co-interventions may confound the attributable effect of the experimental intervention

1.b.i. co-intervention description, as II.E.2.a, and ii. co-intervention rates of use, as II.E.2.c

Were the interventions administered in such a way that any biases are comparable across the trials?

1.c.i. completion of assigned treatment, as II.D.3.a, and ii. actual compliance, as II.D.3.b-d

2. Are the study outcomes comparable?

What were the study outcomes?

2.a. outcome definitions, as II.F.1.a

Were the study outcomes assessed in such a way that any biases are comparable across the trials?

2.b.i. outcome assessment methods, as II.G.1.a, and ii. personnel, as II.G.1.b

3. Are the study populations comparable?

Were the entrance criteria sufficiently similar to select similar groups of subjects?

3.a. entrance criteria, as III.B.4.a-b

Were the baseline characteristics of the subjects relatively similar?

3.b. baseline characteristics, as III.B.4.c

4. Are the study designs and executions comparable?

Were the control groups similar? Differences in control groups affect the comparability of comparative summary statistics (e.g., OR)

4.a.i. description of control, as II.D.2.a, and ii. number of crossovers, as II.D.3.d

Were the settings of the studies sufficiently similar? Vastly different study settings or time of study can introduce unmatched confounders into the data synthesis

4.b.i. sites, as III.C.1.a, and ii. when study conducted, as III.C.3.a-b

Were the lengths of follow-up approximately the same? Different follow-up lengths can lead to essentially different outcomes

4.c. actual follow-up time, as II.H.2.d

Were the studies of similar quality? It is controversial how to adjust for trial quality (i.e., internal validity) in quantitative synthesis

4.d. all data requirements for judging internal validity, as II.A-K

Competency

B. Calculate summary statistic, for pairwise comparisons

A. Odds Ratio (OR)

1. Calculate OR

2. Calculate 95% confidence interval (ci) for OR

a. complete 2 X 2 contingency table

i. OR = a*d/b*c

ii. 95% ci formulas

iii. deduce 2*2 from necessary, sufficient data

B. Relative Risk Reduction (RRR)

1. Calculate RRR

2. Calculate 95% confidence interval (ci) for RRR

a. as IV.B.1-2.a

i. RRR = ARR / (c/c+d)

ii. 95% ci formula

C. Absolute Risk Reduction (ARR)

1. Calculate ARR 

2. Calculate 95% confidence interval (ci) for ARR

a. as IV.B.1-2.a

i. ARR = a/(a+b) - c/(c+d)

ii. 95% ci formula

D. Number Needed to Treat (NNT)

1. Calculate NNT

a. ARR

i. NNT= 1/ARR

C. Quantitative meta-analysis

A. Mantel—Haenszel, using odds ratio

1. Calculate OR for each trial

a. as IV.B.1-2.a

i. as VI.B.1-2.a.i

1. Calculate meta-analytic summary

a. ORs for all the trials

i. Mantel—Haenszel formula