Journal Club: Fairfax Internal Medicine. Stephanie Shin PGY-2 Bianca Ummat PGY-2 July 27, 2012

Transcription

1 Journal Club: Fairfax Internal Medicine Stephanie Shin PGY-2 Bianca Ummat PGY-2 July 27, 2012 S

2 Objective Review EBM basic principles Learn how to apply the principles in critically appraising the article. S

3 Case report A 77 year old male with a history of chronic atrial fibrillation, diabetes mellitus type II, and ischemic cardiomyopathy (EF 30%). B

4 What anticoagulation do I give this patient? B

5 B

6 S Are the Results Valid?

7 Validity Group Questions S Group 1: Did intervention and control groups start with the same prognosis? Were patients randomized? Was randomization concealed? Were patients in the study groups similar with the respect to known prognostic factors? Group 2: Was Prognostic balance maintained as the study progressed? To what extent was the study blinded? Group 3: Were the groups prognostically balanced at the study's completion? Was follow-up complete? Were patients analyzed in the groups to which they were randomized? Was the trial stopped early?

8 Group 1 Did intervention and control groups start with the same prognosis? Were patients randomized? Was randomization concealed? Were patients in the study groups similar with the respect to known prognostic factors? S

9 RE-LY Trial Randomized to receive: 110mg or 150mg Dabigatran or Warfarin. Does not state if it was concealed randomization overall but, blinded administration of 110mg vs 150mg dose of Dabigatran. Warfarin given in unblinded way (dosing based on INR levels). Similar prognostic factors in control & treatment groups (Table 1). S

10 Randomization sometimes may not work. Check effectiveness of randomization by looking at patient characteristics of control vs treatment groups to see if prognostic factors are wellbalanced. Example: Aspirin use similar in 110mg vs 150mg vs Warfarin groups. Important if one of your endpoints is bleeding risk. S

11 Randomization? Allows control and treatment groups to be balanced in regard to prognostic factors thereby eliminate bias. Prognostic factors: determinants of outcome (i.e age/underlying illness/comorbidity) that affect target outcomes (i.e stroke/mi/death). Important because it will result in bias if unconcealed. Ex. recruiters may enroll sicker or less sick subjects to either control or treatment group. S

12 Group 2 Was Prognostic balance maintained as the study progressed? To what extent was the study blinded? S

13 RE-LY Trial Warfarin administered unblinded. INR used for dosing measure. Dabigatran blinded administration of either 110mg vs 150mg. Adjudicators: reviewed documents after blinding. Each primary and secondary outcome event was adjudicated by two independent investigators (unaware of treatment assignments). S

14 S Blinding? Definition: unaware of whether receiving experimental therapy or control therapy. Randomization early on in study Blinding strategy for maintaining prognostic balance. Becomes more important when more judgment involved in determining whether a patient has had a target outcome. (i.e. adjudicators determine if subject had certain outcomes stroke) Ultimate goal reduce bias.

15 Group 3 Were the groups prognostically balanced at the study's completion? Was follow-up complete? Were patients analyzed in the groups to which they were randomized? Was the trial stopped early? B

16 FOLLOW UP Follow up was complete: Follow-up visits occurred 14 days after randomization, at 1 and 3 months, every 3 months thereafter in the first year, and then every 4 months until the study ended. B

17 Were patients analyzed in the groups to which they were randomized? Yes. 29% of Dabigatran patients dropped out; 10% of Warfarin patients dropped out. Intention to treat : Everyone who begins the treatment is considered to be part of the trial, whether he or she finishes it or not. B

18 Early stop? The trial was not stopped early. B

19

20 S What are the results?

21 To answer that question 1. How large was the treatment effect? 2. How precise was the estimate of the treatment effect? **Focus on comparing Warfarin vs 150mg Dabigatran S

22 1. How large was the treatment effect? Simply means translating data/numbers to make it clinically applicable = Impact of treatment. Numerous ways: Relative Risk Relative Risk Reduction NNT (number needed to treat) S

23 Relative Risk The risk of events among patients receiving the new treatment relative to that risk among patients in control group. Relative Risk = [EER/CER] Control Event Rate (CER) Experimental Event Rate (EER) S

24 RE-LY Trial Event: Stroke or systemic embolism Control = Warfarin Experimental = Dabigatran 150mg Data: CER= Warfarin 1.69%/year EER= Dabigatran 1.11%/year Relative Risk = [EER/CER] = [1.11/1.69] = 0.66 S

25 What does this mean? If Relative Risk = 1, then no difference between Warfarin and Dabigatran 150mg in risk of strokesystemic embolism. If Relative Risk > 1, then Dabigatran 150mg has higher risk for stroke/systemic embolism compared to Warfarin. If Relative Risk < 1, then Dabigatran 150mg has lower risk for stroke/systemic embolism compared to Warfarin. Another way of looking at it: Relative Risk Reduction= 1 Relative Risk Example: = 0.33 Means Dabigatran 150mg reduces risk of stroke/systemic embolism by 33% relative to Warfarin group. S

26 33% decrease? S Which medication would you pick?

27 Dr. V s Example Sale 40% off on all mobile devices: Iphone $ 100 Flip-phone $ 10 Final price: IPhone = $100 (40%*100) = $60 Flip-phone = $10 (40%*10) = $6 Same SALE. But you save $40 vs $ 4 Question: Which one is a better deal? S

28 Going back to 33% relative risk reduction in RE-LY Trial Keep in mind that event rates were very low to begin with: Warfarin 1.69%/year Dabigatran 1.11%/year It can be misleading just looking at the relative risk reduction. S

29 Number Needed to Treat (NNT) Means number of patients one would need to treat to prevent an adverse event. NNT = 100/(ARR%) ARR= absolute risk reduction (risk difference) ARR = CER-EER S

30 NNT from RE-LY Event: Stroke/systemic embolism Control = Warfarin Experimental = Dabigatran 150mg Data: CER= Warfarin 1.69%/year EER= Dabigatran 1.11%/year ARR = CER-EER = = 0.58 %/year NNT = 100/ARR = 100/0.58% = 172 S

31 How precise was the estimate of the treatment effect? Precision is determined by P values and Confidence Intervals: P Value: whether trial results are likely to have occurred simply through chance. If p is small (<.05), findings are unlikely to have risen by chance. 95% CI: includes the true effect 95% of the time. The CI includes 2 standard deviations around the mean. B

32 Confidence Intervals The CI tells us 2 things: How precise the estimate is: the wider the confidence interval, the less precise because there is greater uncertainty in the data. Statistical significance: if the CI does not include 1, it is significant at p<.05. B

33 P values vs. Confidence Intervals Confidence intervals are preferable to p values, because they tell the range of possible effect sizes, whereas p values simply provide a cut off beyond which findings are considered statistically significant. B

34 Stroke and Systemic Emboli Table 2: There were fewer strokes of systemic emboli with Dabigatran (134) than for Warfarin (199). Therefore, the CI for Dabigatran preventing stroke or systemic embolism = , correlating to a p<0.005, making this effect highly statistically significant. B

35 B MI Table 2: However, there were more MIs in the Dabigatran group (89), compared to the Warfarin group (63). Therefore, the CI for Dabigatran preventing MI = , correlating to a p=.88, showing that this is effect is nonsignificant as the CI includes 1, which captures the value of no effect, ie a relative risk of 1.

36 How can I apply the results to patient care? Were the study patients similar to my patient? Were all patient-important outcomes considered? Are the likely treatment benefits worth the potential harm and costs? B

37 Were the study patients similar to my patient? Case patient: 77 years old, chronic atrial fibrillation, ischemic cardiomyopathy, diabetes. Relatively similar to study subjects. S

38 Were all patient-important outcomes considered? Primary outcome: Stroke and systemic embolism Secondary outcome: stroke, systemic embolism, death Safety outcome: hemmorhage Other outcomes: MI, PE, TIA, hospitalization B

39 Are the likely treatment benefits worth the potential harms and costs? Absolute Risk Reduction: the difference between the control group s event rate and the event group s event rate. (ARR = CER- EER) For stroke: Warfarin event rate (1.69%/year)- Dabigatran event rate (1.11%/year) = 0.68% 0.68% difference is very small value in light of low event rate. Something to consider when weighing benefits vs risks. B

40 $$$ Cost $$$ Cost-Effectiveness of Dabigatran for Stroke Prophylaxis in Atrial Fibrillation Circulation, March 4, 2011 Estimated Dabigatran to be $9/day. For patients at low stroke risk (CHADS <1), ASA most cost effective; medium stroke risk (CHADS 1-2), Warfarin most effective; high stroke risk (CHADS >2), Dabigatran most effective. Caveat: time in INR therapeutic range B

41 B $$$ Cost $$$ Total cost for Dabigatran s effect on decreasing stroke & systemic emboli = (difference in price of Dabigatran and Warfarin) x NNT= ($9 (based on trial estimate) -$0.13 (based on 4 dollar forumlary)) x (100/ARR%) = ($8.87) x (172)= $8.87 x 172 = $1525 more to prevent 1 stroke event with Darbigatran.??? Does this make sense? Im bad at math.

42 Conclusion RE-LY Trial: Randomized Large sample size Intention to treat Blinded study Well-designed Results: Dabigatran 150mg vs Warfarin in reducing stroke/systemic embolism statistically significant. Things to Consider: Concern for Lead-time Bias Low event rate NNT high S

43 Review Systemic approach to reviewing/appraising article/journals. Importance in study design and analysis in determining the validity of the study. Understanding treatment impact/effect is useful in applying to daily use. B

44 What would you give the patient: Dabigatran 150mg or Warfarin?