CONSORT/RCT Triage Characteristics

Question 1

Participants

Answer 1

1. Description of participants, then check if conclusions are aimed at a similar population
   - reveals whether results apply to the target population or not
2. Baseline data table
   - reveals whether groups were similar at the start of the trial

Question 2

Outcome 1: Description

Answer 2

1. Look to see if a single primary outcome has been identified, then check to see if conclusions about efficacy are based on that outcome
   - tells you if the risk of a false positive was alpha (usually 5%), or it it was alpha times the number of outcomes measured

Question 3

Outcomes 2: Reporting

Answer 3

1. Look to see if all outcomes have been reported
   - reveals the possibility that undesired findings have been hidden from the reader
2. Check to see if results are presented as absolute numbers, not just %
   - % can leave exaggerated impression of efficacy
3. Check to see if confidence intervals have been provided, not just p-values
   - not possible to know if effect is medically important without a CI
4. Check to see if means in different groups have been compared to each other, or each mean compared to only baseline within same group
   - risk of false positive (50%) or higher, making results meaningless

Question 4

Sample Size Calculation

Answer 4

1. Look to see if they did a calculation, look for 4 variables as evidence that it was truly calculated
   - calculation to know risk of false positive result (fake; alpha), risk of false negative result (missed; beta), effect size, variance (omega)
2. Check to see if the groups used in the trial were at least as large as the calculation
   - tells you if trial is adequately powered or underpowered

Question 5

Randomization

Answer 5

1. Did they truly use random method to produce random sequence? If less than 100, did they use restricted randomization?
   - distribute confounding variables as evenly as possible between groups

Question 6

Allocation Concealment

Answer 6

1. Look to see if a third party was used to allocate participants into groups, ore used SNOSE
   - prevents the use of visual indicators of health to place healthier participants in one group

Question 7

Blinding

Answer 7

1. Look to see if the treatments were similar
   - figuring out who was recieving which intervention, may have biased the study
2. Mentioned who was blinded, or used vague “double-blind” terminology
   - participants and primary outcome assessors are important groups to be blind, to reduce hope & expectations that may distort the results
3. If no blinding, check if primary outcome was objective (facts)
   - if primary outcome was subjective (feelings) = hopes and expectations render worthless results; probably OK if outcome was objective

Question 8

Participants Flow and Losses

Answer 8

1. Look for flow chart showing losses from each group (with reason); if losses were included in the analysis (ITT)
   - important adverse effects have been ignored
   - non-random losses from each group = non-random groups
   - losses resulted in an underpowered trial (small sample size)

Question 9

Interpretation

Answer 9

1. Check to see if the conclusions match the data
   - helps to identify and reject exaggerated conclusions
2. Are limitations discussed and if serious limitations have been ignored in the conclusions
   - if limitation manes unjustified efficacy of conclusions = reject conclusions

Question 10

Harms

Answer 10

1. Look for table comparing adverse effects in each group, check is serious adverse effects (if any) have been taken into account in the conclusions
   - reduces risk of important adverse effects being missed/ignored/down-played

Question 11

Funding

Answer 11

1. Is funding for-profit (ex. Industry) or non-profit?
   - for-profit-funded studies are 4x more likely to exaggerate conclusions; take extra care in comparing conclusions with data
2. Do authors have financial gain from study? Are authors employees of the company? Have authors received fees from company in the past?
   - identify studies with a high risk of exaggeration due to conflicts of interest

Question 12

When an RCT is not needed?

Answer 12

1. Drugs that do not heal
   - for purposes other than healing
   - ex. General anesthetic to cause unconsciousness which the body is not trying to achieve —> no other reasonable cause
   - ex. Fluorescein dye to find damage —> illuminating under UV light was not the body’s own doing
2. When effects are immediate (seconds to minutes)
   - no need to question the cause of these effects as recovery did not occur naturally —> can be confident in responsibility of intervention
   - ex. Epinephrine increases/improve heart rate, contractility, and BP within seconds
3. When the body is incapable
   - body is not capable of achieving the outcomes naturally
   - ex. Patient with type I diabetes do not produce insulin = cannot utilize carbohydrates normally —> confident that insulin shots achieve outcomes of normal insulin levels