Therapy April 12 th , 2016 Manny Ribeiro, MD Larry - - PowerPoint PPT Presentation

Therapy

April 12th, 2016 Manny Ribeiro, MD Larry Young, MD

Objectives

 The evidence-based medicine cycle  Start with a case scenario  Ask the clinical question  Critical appraisal

• Allocation concealment
• Intention-to-treat
• Blinding
• Follow-up

 Results: making the math simple!

THE PATIENT ASSESS ASK ACQUIRE APPLY APPRAISE

Evidence- based Medicine Cycle

The 5 A’s

Case scenario

 A 67-year-old man presented with productive cough and fever for 3 days  Past medical history of coronary artery disease, with a myocardial infarction two years prior  Heavy smoker, 2 packs of cigarettes per day for 50 years (100 pack-year)

Bad pneumonia!

 Also had a new myocardial infarction

• Echocardiogram with a drop in

ejection fraction to 46% and new segmental wall motion abnormality

Case scenario

His daughter’s request

“Please start something for my father to quit smoking before he goes home, otherwise he will just go back to it right after discharge.”

This sounds like PICOTT... Can we PICOTT this? Yes, this is “PICOTT”able!

P I C O T T

Clinical question formation

• pulation

ntervention

• mparison

utcome ype of Question ype of (ideal) study design

P I C O T T

Clinical question

Smokers admitted to the hospital Inpatient strategies for smoking cessation No/other strategies Quit rate, pneumonia, myocardial infarction Therapy question Randomized controlled trial

Search strategy

Critical appraisal

= Randomized Controlled Trial

Goals of randomization

P

R Treatment Control Prognosis X Prognosis X

 To keep all known and unknown prognostic variables evenly distributed between the groups

Outcome A Outcome B

Group C Group D Group A Group B

Allocation concealment

Our population YOU!

List generation

Head: Allowed to ask questions (A) Tail: NOT Allowed to ask questions (NA)

Allocation concealment

 The person who is enrolling participants cannot know, predict, or manipulate the list  Trials with inappropriate allocation concealment are associated with larger estimates of treatment effect

Allocation concealment

P O

R

List generation Allocation concealment

Intention-to-treat

 AKA: Were patients analyzed in the groups to which they were randomized? Teaching method 1

Dead Alive Alive Dead Dead Alive Dead Alive R Treatment 1 Treatment 2

Dead Alive Alive Dead Dead Alive Dead Alive R Treatment 1 Treatment 2

Dead Alive Alive Dead Dead Alive Dead Alive Dead Alive Alive Dead Dead Alive Dead Alive R Treatment 1 Treatment 2

Intention-to-treat

Cerebro- vascular disease

R

Surgery + ASA ASA 200 100 100 Stroke Stroke 10 10 Stroke Stroke 10 10 Surgery Per Protocol 10/90 = 11% 20/100 = 20% ITT 20/100 = 20% 20/100 = 20%

RD = 9% RD = 0%

 Teaching method 2

Intention-to-treat… Why??

 Preserves balance between the groups  Reflects real life

Blinding

Patient Researcher

2 volunteers!

Shhhhhhh!!

Blinding

P O

R

List generation Allocation concealment

Blinding

Blinding

 Trials with inappropriate blinding are also associated with larger estimates

• f effect, but not as much as with

inappropriate allocation concealment

Allocation Concealment Blinding Who? Enroller Patients, caregivers, data collectors, adjudicators, analysts What? The list Group assignments When? Part of randomization After randomization

Allocation Concealment x blinding

Follow-up

 Patients who are lost often have different prognoses from those who are retained  Strategies to deal with lost to follow up: last observation carried forward, worst-case scenario  The best solution is to assure a good follow up

Therapy Math

All-cause hospital admissions

Step 2: Subtract: Step 3: Divide: Step 1: Intensive Usual 23% 41% 41% – 23% = 18% 23% / 41% = 0.56 Risk difference Risk ratio

Number Needed to Treat

 What is the risk difference of 18% telling you?

In order to save 18, you needed to treat 100 In order to save 1, how many do you need to treat? Formula NNT: 100 / RD

Number Needed to Treat

 Formula: NNT = 100 / RD NNT = 100 / 18 = 5.5 NNT = 6  You needed to treat 6 patients in

• rder to prevent one extra

hospitalization in 2 years

Relative Risk Reduction

 Teaching method 1

20 By how much (in %) did I reduce? Answer: 25% 15

All-cause hospital admissions

Step 2: Subtract: Step 3: Divide: Step 1: Intensive Usual 23% 41% 41% – 23% = 18% 23% / 41% = 0.56 Risk difference Risk ratio

Relative Risk Reduction

 Teaching method 2

Formula: RRR = 1 - RR Intensive Usual 23% 41% 23% / 41% = 0.56 Risk ratio

Relative Risk Reduction

 Teaching method 2

Formula: RRR = 1 - RR Intensive Usual 23% 41% 23% / 41% = 0.56 Risk ratio 41% 41% / 41% = 1

Relative Risk Reduction

 Teaching method 2

Formula: RRR = 1 - RR Intensive Usual 23% 41% 23% / 41% = 0.56 Risk ratio 1 RRR = 0.44 (44%)

Math time!

Mortality

Intensive Usual 2.8% 12% Risk difference: 9.2% Risk Ratio: 0.23 NNT: 11 RRR: 0.77 (77%)

Take-home points

 The evidence-based medicine cycle: everything starts and ends with a patient  Improper allocation concealment can

• verestimate the effect size

 For therapy papers, ITT is preferred to per-protocol analysis  Look for who was blinded in the study  Make the math simple! Remember to subtract and divide!

Teaching Take-home Points

• What strategies did we use to teach

these concepts?

Teaching Take-home Points

 Real clinical case – start with a patient.  Clinical question related to different specialties, and even non-medical learners  Group activities  Pre-mark article – saves time; reduces stress  Imperfect articles (you can teach with them)  Interactivity  Different teaching strategies (visual learners, math lovers)  Simplicity: one step at a time, with “no man left behind”  Triage: you can’t do it all

Back to our patient…

 One year after discharge:

Thank you!