Critical Appraisal for Therapy Articles THERAPY STUDY: Are

the results of the trial valid? (Internal Validity) What question did the study ask? Patients Intervention Comparison Outcome(s) 1a. R- Was the assignment of atients to treatments randomised?
What is best? Centralised computer randomisation is ideal and often used in multi-centred trials. Smaller trials may use an independent person (e.g, the hospital pharmacy) to police the randomization. This paper: Yes Comment: What is best? If the randomisation process wor ed (that is, achie!ed compara"le groups) the groups should "e similar. The more similar the groups the "etter it is. There should "e some indication of whether differences "etween groups are statistically significant (ie. p !alues). This paper: Yes Comment: What is best? &part from the inter!ention the patients in the different groups should "e treated the same, eg., additional treatments or tests. This paper: Yes Comment: Where do I find the information? The Methods should tell you how patients were allocated to groups and whether or not randomisation was concealed.

1!. R- Were the grou s similar at the start of the trial?

Where do I find the information? The Results should ha!e a ta"le of #$aseline %haracteristics# comparing the randomized groups on a num"er of !aria"les that could affect the outcome (ie. age, ris factors etc). If not, there may "e a description of group similarity in the first paragraphs of the Results section.

"a. A # Aside from the allo$ated treatment% &ere grou s treated equally?
Where do I find the information? 'oo in the Methods section for the follow-up schedule, and permitted additional treatments, etc and in Results for actual use.

"!. A # Were all atients &ho entered the trial a$$ounted for? # and &ere they analysed in the grou s to &hi$h they &ere randomised?
What is best? 'osses to follow-up should "e minimal ( prefera"ly less than )*+. ,owe!er, if few patients ha!e the outcome of interest, then e!en small losses to follow-up can "ias the results. -atients should also "e analysed in the groups to which they were randomised ( .intention-to-treat analysis. This paper: Yes Comment: Where do I find the information? The Results section should say how many patients were /andomised (eg., $aseline %haracteristics ta"le) and how many patients were actually included in the analysis. 0ou will need to read the results section to clarify the num"er and reason for losses to follow-up.

'. ( - Were measures o!)e$tive or &ere the atients and $lini$ians ke t *!lind+ to &hi$h treatment &as !eing re$eived?
What is best? It is ideal if the study is .dou"le-"linded1 ( that is, "oth patients and in!estigators are unaware of treatment allocation. If the outcome is objective (eg., death) then "linding is less critical. If the outcome is subjective (eg., symptoms or function) then "linding of the outcome assessor is critical. This paper: Yes Comment: Where do I find the information? 2irst, loo in the Methods section to see if there is some mention of mas ing of treatments, eg., place"os with the same appearance or sham therapy. Second, the Methods section should descri"e how the outcome was assessed and whether the assessor3s were aware of the patients4 treatment.

University of Oxford, 2005

Critical Appraisal for Therapy Articles

What &ere the results?

1. ,o& large &as the treatment effe$t?
5ost often results are presented as dichotomous outcomes (yes or not outcomes that happen or don4t happen) and can include such outcomes as cancer recurrence, myocardial infarction and death. %onsider a study in which /6+ (*./6) of the control group died and /*+ (*./*) of the treatment group died after ) years of treatment. The results can "e e7pressed in many ways as shown "elow. What is the measure? Relative Risk (RR) 8 ris of the outcome in the treatment group 3 ris of the outcome in the control group. What does it mean? The relati!e ris tells us how many times more likely it is that an e!ent will occur in the treatment group relati!e to the control group. &n RR of 1 means that there is no difference "etween the two groups thus, the treatment had no effect. &n 99 : / means that the treatment decreases the ris of the outcome. &n 99 ; / means that the treatment increased the ris of the outcome. Since the 99 : /, the treatment decreases the ris of death. unclear Absolute Risk Reduction (ARR) 8 ris of the outcome in the control group - ris of the outcome in the treatment group. This is also nown as the absolute risk difference. In our e7ample, the &99 8 *./6 - *./* 8 *.*6 or 6+ The a"solute ris reduction tells us the a"solute difference in the rates of e!ents "etween the two groups and gi!es an indication of the "aseline ris and treatment effect. &n ARR of 0 means that there is no difference "etween the two groups thus, the treatment had no effect. The a"solute "enefit of treatment is a 6+ reduction in the death rate. unclear Relative Risk Reduction (RRR) 8 a"solute ris reduction 3 ris of the outcome in the control group. &n alternati!e way to calculate the 999 is to su"tract the 99 from / (eg. 999 8 / - 99) In our e7ample, the 999 8 *.*63*./6 8 *.>> or >>+ ?r 999 8 / - *.<= 8 *.>> or >>+ Number Needed to Treat (NNT) 8 in!erse of the &99 and is calculated as / 3 &99. The relati!e ris reduction is the complement of the 99 and is pro"a"ly the most commonly reported measure of treatment effects. It tells us the reduction in the rate of the outcome in the treatment group relati!e to that in the control group. The treatment reduced the ris of death "y >>+ relati!e to that occurring in the control group. unclear The num"er needed to treat represents the num"er of patients we need to treat with the e7perimental therapy in order to pre!ent / "ad outcome and incorporates the duration of treatment. %linical significance can "e determined to some e7tent "y loo ing at the @@Ts, "ut also "y weighing the @@Ts against any harms or ad!erse effects (@@,s) of therapy. Ae would need to treat )* people for ) years in order to pre!ent / death. unclear

In our e7ample, the 99 8 *./*3*./6 8 *.<=

In our e7ample, the @@T 8 /3 *.*6 8 )*

". ,o& re$ise &as the estimate of the treatment effe$t?

The true ris of the outcome in the population is not nown and the "est we can do is estimate the true ris "ased on the sample of patients in the trial. This estimate is called the point estimate. Ae can gauge how close this estimate is to the true !alue "y loo ing at the confidence inter!als (%I) for each estimate. If the confidence inter!al is fairly narrow then we can "e confident that our point estimate is a precise reflection of the population !alue. The confidence inter!al also pro!ides us with information a"out the statistical significance of the result. If the !alue corresponding to no effect falls outside the B6+ confidence inter!al then the result is statistically significant at the *.*6 le!el. If the confidence inter!al includes the !alue corresponding to no effect then the results are not statistically significant.

University of Oxford, 2005

Critical Appraisal for Therapy Articles Will the results hel me in $aring for my atient? (-.ternalValidity/A li$a!ility)

The Cuestions that you should as "efore you decide to apply the results of the study to your patient areD Is my patient so different to those in the study that the results cannot applyE no Is the treatment feasi"le in my settingE no Aill the potential "enefits of treatment outweigh the potential harms of treatment for my patientE no

University of Oxford, 2005