PHS 525: STATISTICS FOR

LABORATORY SCIENTISTS

Arthur Berg, PhD

profberg.com

SyllabusCourse syllabus is available at pro!erg.com

It will be updated with your feedback

We will attempt to cover all of the material in the textbook

‣ descriptive statistics, distributions, hypothesis testing, power analysis, analysis of variance, regression, correlation, et al.

The software Minitab, version 16, will be used in this course. You will need to bring a laptop to some of the classes (announced in advance).

StatisticsStatistics helps to sort out the good ideas from the bad.

Statistics allows us to quantify our level of uncertainty about what we know.

๏ Limitation: Statistics generally cannot explain mechanism. It cannot parse out correlation from causality.

★ Statistics can open the doors to new discoveries.

Statistics, like art and sport, takes practice to become skillful.

Riemann Hypothesis

Monte Carlo Method

Association vs Mechanism

In 1747 aboard HMS Salisbury, physician James Lind performed a controlled experiment to develop a cure for scurvy.■ A quart of cider every day■ Twenty five gutts (drops) of elixir vitriol (sulphuric acid) three times a day upon an empty stomach,■ One half-pint of seawater every day■ A mixture of garlic, mustard, and horseradish in a lump the size of a nutmeg■ Two spoonfuls of vinegar three times a day■ Two oranges and one lemon every day.

Mechanism of vitamin C made much later...

ONLINE FIRST

REVIEW ARTICLE

Dabigatran Association With Higher Riskof Acute Coronary EventsMeta-analysis of Noninferiority Randomized Controlled Trials

Ken Uchino, MD; Adrian V. Hernandez, MD, PhD

Background: The original RE-LY (Randomized Evalu-ation of Long-term Anticoagulant Therapy) trial sug-gested a small increased risk of myocardial infarction (MI)with the use of dabigatran etexilate vs warfarin in pa-tients with atrial fibrillation. We systematically evalu-ated the risk of MI or acute coronary syndrome (ACS)with the use of dabigatran.

Methods: We searched PubMed, Scopus, and the Webof Science for randomized controlled trials of dabiga-tran that reported on MI or ACS as secondary out-comes. The fixed-effects Mantel-Haenszel (M-H) test wasused to evaluate the effect of dabigatran on MI or ACS.We expressed the associations as odds ratios (ORs) andtheir 95% CIs.

Results: Seven trials were selected (N=30 514), includ-ing 2 studies of stroke prophylaxis in atrial fibrillation,1 in acute venous thromboembolism, 1 in ACS, and 3 ofshort-term prophylaxis of deep venous thrombosis. Con-trol arms included warfarin, enoxaparin, or placebo ad-

ministration. Dabigatran was significantly associated witha higher risk of MI or ACS than that seen with agents usedin the control group (dabigatran, 237 of 20 000 [1.19%]vs control, 83 of 10 514 [0.79%]; ORM-H, 1.33; 95% CI,1.03-1.71; P=.03). The risk of MI or ACS was similar whenusing revised RE-LY trial results (ORM-H, 1.27; 95% CI,1.00-1.61; P=.05) or after exclusion of short-term trials(ORM-H, 1.33; 95% CI, 1.03-1.72; P=.03). Risks were notheterogeneous for all analyses (I2=0%; P! .30) and wereconsistent using different methods and measures of as-sociation.

Conclusions: Dabigatran is associated with an in-creased risk of MI or ACS in a broad spectrum of pa-tients when tested against different controls. Cliniciansshould consider the potential of these serious harmfulcardiovascular effects with use of dabigatran.

Arch Intern Med.Published online January 9, 2012.doi:10.1001/archinternmed.2011.1666

D ABIGATRAN ETEXILATE WASapproved by the Euro-pean Medicines Agencyin 2008 for prophylaxis ofvenous thromboembo-

lism (VTE) in adults who have under-gone total hip or knee replacement and bythe US Food and Drug Administration in2010 for prevention of stroke and systemicembolism in persons with nonvalvularatrial fibrillation (AF).1 The largest study

of dabigatran, the RE-LY (RandomizedEvaluation of Long-term AnticoagulantTherapy) trial, followed more than 18 000persons with AF and compared the safetyand efficacy of dabigatran with that of war-farin in preventing stroke and systemic em-bolism for a median observation time of2 years. In its original publication,2 theRE-LY trial reported a significant de-

crease in stroke and systemic embolism of34% in the group receiving dabigatranetexilate, 150 mg, twice daily comparedwith adjusted-dose warfarin, but also re-ported a significant relative increase inmyocardial infarction (MI) of 38% (rela-tive risk, 1.38; 95% CI, 1.00-1.91; P=.048)in the group receiving dabigatran etexi-late, 150 mg, twice daily compared withwarfarin. A subsequent review3 of out-come and safety events revealed addi-tional events of stroke, bleeding, and MI,and the revised results no longer showeda significantly higher risk of MI with theuse of dabigatran (increased by 27%; rela-tive risk, 1.27; 95% CI, 0.94-1.71; P=.12).Dabigatran has also been studied for acuteVTE treatment, VTE prophylaxis after jointreplacement, and acute coronary syn-drome (ACS) in noninferiority random-ized controlled trials (RCTs). We system-atically evaluated the risk of MI or ACSwith the use of dabigatran for several clini-cal indications and against different con-trol arms.

See Invited Commentaryand Editor’s Noteat end of article

Author Affiliations:Cerebrovascular Center,Neurological Institute(Dr Uchino), Health Outcomesand Clinical Epidemiology,Department of QuantitativeHealth Sciences, LernerResearch Institute(Dr Hernandez), ClevelandClinic, Cleveland, Ohio; andQuantitative Research Division,BioEstadistica, S.C., Monterrey,Nuevo Leon, Mexico(Dr Hernandez).

ARCH INTERN MED PUBLISHED ONLINE JANUARY 9, 2012 WWW.ARCHINTERNMED.COME1

©2012 American Medical Association. All rights reserved. at Penn State Milton S Hershey Med Ctr, on January 10, 2012 www.archinternmed.comDownloaded from

ONLINE FIRST

REVIEW ARTICLE

Dabigatran Association With Higher Riskof Acute Coronary EventsMeta-analysis of Noninferiority Randomized Controlled Trials

Ken Uchino, MD; Adrian V. Hernandez, MD, PhD

Background: The original RE-LY (Randomized Evalu-ation of Long-term Anticoagulant Therapy) trial sug-gested a small increased risk of myocardial infarction (MI)with the use of dabigatran etexilate vs warfarin in pa-tients with atrial fibrillation. We systematically evalu-ated the risk of MI or acute coronary syndrome (ACS)with the use of dabigatran.

Methods: We searched PubMed, Scopus, and the Webof Science for randomized controlled trials of dabiga-tran that reported on MI or ACS as secondary out-comes. The fixed-effects Mantel-Haenszel (M-H) test wasused to evaluate the effect of dabigatran on MI or ACS.We expressed the associations as odds ratios (ORs) andtheir 95% CIs.

Results: Seven trials were selected (N=30 514), includ-ing 2 studies of stroke prophylaxis in atrial fibrillation,1 in acute venous thromboembolism, 1 in ACS, and 3 ofshort-term prophylaxis of deep venous thrombosis. Con-trol arms included warfarin, enoxaparin, or placebo ad-

ministration. Dabigatran was significantly associated witha higher risk of MI or ACS than that seen with agents usedin the control group (dabigatran, 237 of 20 000 [1.19%]vs control, 83 of 10 514 [0.79%]; ORM-H, 1.33; 95% CI,1.03-1.71; P=.03). The risk of MI or ACS was similar whenusing revised RE-LY trial results (ORM-H, 1.27; 95% CI,1.00-1.61; P=.05) or after exclusion of short-term trials(ORM-H, 1.33; 95% CI, 1.03-1.72; P=.03). Risks were notheterogeneous for all analyses (I2=0%; P! .30) and wereconsistent using different methods and measures of as-sociation.

Conclusions: Dabigatran is associated with an in-creased risk of MI or ACS in a broad spectrum of pa-tients when tested against different controls. Cliniciansshould consider the potential of these serious harmfulcardiovascular effects with use of dabigatran.

Arch Intern Med.Published online January 9, 2012.doi:10.1001/archinternmed.2011.1666

D ABIGATRAN ETEXILATE WASapproved by the Euro-pean Medicines Agencyin 2008 for prophylaxis ofvenous thromboembo-

lism (VTE) in adults who have under-gone total hip or knee replacement and bythe US Food and Drug Administration in2010 for prevention of stroke and systemicembolism in persons with nonvalvularatrial fibrillation (AF).1 The largest study

of dabigatran, the RE-LY (RandomizedEvaluation of Long-term AnticoagulantTherapy) trial, followed more than 18 000persons with AF and compared the safetyand efficacy of dabigatran with that of war-farin in preventing stroke and systemic em-bolism for a median observation time of2 years. In its original publication,2 theRE-LY trial reported a significant de-

crease in stroke and systemic embolism of34% in the group receiving dabigatranetexilate, 150 mg, twice daily comparedwith adjusted-dose warfarin, but also re-ported a significant relative increase inmyocardial infarction (MI) of 38% (rela-tive risk, 1.38; 95% CI, 1.00-1.91; P=.048)in the group receiving dabigatran etexi-late, 150 mg, twice daily compared withwarfarin. A subsequent review3 of out-come and safety events revealed addi-tional events of stroke, bleeding, and MI,and the revised results no longer showeda significantly higher risk of MI with theuse of dabigatran (increased by 27%; rela-tive risk, 1.27; 95% CI, 0.94-1.71; P=.12).Dabigatran has also been studied for acuteVTE treatment, VTE prophylaxis after jointreplacement, and acute coronary syn-drome (ACS) in noninferiority random-ized controlled trials (RCTs). We system-atically evaluated the risk of MI or ACSwith the use of dabigatran for several clini-cal indications and against different con-trol arms.

See Invited Commentaryand Editor’s Noteat end of article

Author Affiliations:Cerebrovascular Center,Neurological Institute(Dr Uchino), Health Outcomesand Clinical Epidemiology,Department of QuantitativeHealth Sciences, LernerResearch Institute(Dr Hernandez), ClevelandClinic, Cleveland, Ohio; andQuantitative Research Division,BioEstadistica, S.C., Monterrey,Nuevo Leon, Mexico(Dr Hernandez).

ARCH INTERN MED PUBLISHED ONLINE JANUARY 9, 2012 WWW.ARCHINTERNMED.COME1

©2012 American Medical Association. All rights reserved. at Penn State Milton S Hershey Med Ctr, on January 10, 2012 www.archinternmed.comDownloaded from

• 7 studies• 1.19% risk of MI/ACS with dabigatran•.79% risk of MI/ACS with control/warfarin•OR=1.33 (1.03--1.71), p=.03

The Practice of Statistics/Biostatistics

✤ Power calculations

✤ Sample size calculations

✤ Experimental design

❖ descriptive statistics

❖ graphics

❖ analysis

Statistics is...the analysis and interpretation of data

Population and Samples

Random Sampling...with care

Not an Independent Sample...

“Put me down for whoever comes out ahead in your poll.”