Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
When
Wh can we consider RCT data
                    id         d
sufficient to address effectiveness
  and harm and when ...
Steps in evidence-based
                       healthcare

1.   Frame the question
2.   Find the best evidence
3.   Assess...
To understand how to recognize “best evidence” we need to
          understand the questions we are asking

Question      ...
One size does NOT fit all!
    Use y
        your q
             question classification to seek the
          appropriate...
One size does NOT fit all!
    Use y
        your q
             question classification to seek the
          appropriate...
You might ask
                    ask…..
• But what about observational data, can’t it
  be used to supplement what we kno...
Your questions
1. How should trial results be interpreted when the
   population studied is narrower than the population I...
I ask…
                     ask
• Is this population so different, is the
  intervention so different, is the setting so
 ...
Your questions
3.      Most scientists say to watch out for
     subgroup analysis, y in many cases
         g p        y ...
Mammographic screening in women 40-49




Nelson 2009 Annals Internal Medicine
Your questions (cont’d)
                        (cont d)

4.How should benefits and harms be
  balanced when different gro...
Evidence-Based Healthcare
  “The integration
  of best research
   fb t            h
  evidence with                      ...
Your questions
5.How can we translate to our constituents
  what to look for in a trial without a lot of
  jargon?
Your questions
6. Sometimes a trial may be well-designed in that
   it minimizes bias, but the comparison group is
   sele...
Your questions
8.How do we interpret results from head-to-
  head trials (ie, 2 active interventions
              ( ,
  c...
Your questions
• What is your assessment of the various
  adaptive designs proposed? How can
  advocates assess their rela...
Examples of adaptive designs
                                 g
Changes in:
• Study eligibility criteria
       y g       ...
Kay Dickersin
Kay Dickersin
Kay Dickersin
Upcoming SlideShare
Loading in …5
×
Upcoming SlideShare
Panel: CUE in partnership- Lisa Simpson
Next

Share

Kay Dickersin

Related Books

Free with a 30 day trial from Scribd

See all

Related Audiobooks

Free with a 30 day trial from Scribd

See all

Kay Dickersin

  1. 1. When Wh can we consider RCT data id d sufficient to address effectiveness and harm and when is more information needed? Kay Dickersin April 15, 2010 CUE Annual Meeting Washington, Washington DC
  2. 2. Steps in evidence-based healthcare 1. Frame the question 2. Find the best evidence 3. Assess the evidence 4. Apply the evidence to the individual Guyatt et al. Users' Guides to the Medical Literature: A Manual for Evidence- Based Clinical Practice. Chicago, IL: American Medical Association; 2002.
  3. 3. To understand how to recognize “best evidence” we need to understand the questions we are asking Question Classification/Type What should be done? What proportion of the population is newly Incidence diagnosed with this problem each year? How can I prevent further trouble? What proportion of the population is currently Prevalence living with Will there be any negative effects? this problem? What should be done it mean inthis problem? What will to treat the future? Therapy Will detecting this problem early, before Screening symptoms, make a difference in is wrong? What my health? How good is this test at detecting this Diagnostic Accuracy problem? How many more cases are there? What is the likely outcome of this problem? Prognosis How many cases are th ? H there? Will there be any negative effects (of an Harm intervention)? What causes this problem? What causes this problem? Etiology How can this problem be prevented? Prevention
  4. 4. One size does NOT fit all! Use y your q question classification to seek the appropriate type of evidence Question: Look for evidence from:: • Incidence, Surveys, followup studies prevalence l • Therapy Clinical trials • Screening Clinical trials • Diagnostic accuracy Clinical trials, cross sectional studies • Prognosis Clinical trials, followup studies • Harm Clinical trials, followup studies, case control studies t l t di • Etiology Follow up studies, case control studies • Prevention Clinical trials
  5. 5. One size does NOT fit all! Use y your q question classification to seek the appropriate type of evidence Question: Look for evidence from:: • Th Therapy Clinical trials Cli i l t i l • Screening Clinical trials • Diagnostic accuracy Clinical trials cross sectional trials, studies • Harm Clinical trials, followup studies, case control studies • Prevention Clinical trials
  6. 6. You might ask ask….. • But what about observational data, can’t it be used to supplement what we know from pp RCTs? • It’s complicated…..see workshop or I have slides if we have enough time.
  7. 7. Your questions 1. How should trial results be interpreted when the population studied is narrower than the population I am interested in (eg, they may be all white (eg (mammography), older, fatter, (Women’s Health Initiative) healthier)? ) ) 2. When is it ok to consider data from trials of healthy people? They may do better on the treatment and have fewer adverse events.
  8. 8. I ask… ask • Is this population so different, is the intervention so different, is the setting so , g different, that I would choose to ignore these findings in favor of no evidence?
  9. 9. Your questions 3. Most scientists say to watch out for subgroup analysis, y in many cases g p y , yet y there seems to be too much lumping and we really want to know about differences that might be present between subgroups – men and women blacks and whites old women, whites, and young. What should we think?
  10. 10. Mammographic screening in women 40-49 Nelson 2009 Annals Internal Medicine
  11. 11. Your questions (cont’d) (cont d) 4.How should benefits and harms be balanced when different groups assign g p g different weights to harms (eg, loss of fertility in women with and without kids)
  12. 12. Evidence-Based Healthcare “The integration of best research fb t h evidence with Best Research clinical expertise li i l ti Evidence and patient values.” EBHC Clinical Patient Expertise Values —Sackett et al, 2000 NY: Churchill Livingston
  13. 13. Your questions 5.How can we translate to our constituents what to look for in a trial without a lot of jargon?
  14. 14. Your questions 6. Sometimes a trial may be well-designed in that it minimizes bias, but the comparison group is selected to make the drug or intervention look better than it really is. 7. Sometimes a statistically significant outcome is reported, but other outcomes or time points that did not show a significant difference between interventions are not reported. How can we detect this and know when to be careful?
  15. 15. Your questions 8.How do we interpret results from head-to- head trials (ie, 2 active interventions ( , compared) versus results from an active intervention compared to a placebo or no intervention?
  16. 16. Your questions • What is your assessment of the various adaptive designs proposed? How can advocates assess their relative merits? ? • Adaptive design = changes in a trial’s design d i or analysis guided b l ki at th l i id d by looking t the trial data while the study is ongoing. Adaptive designs should ALWAYS preplan these “looks” and exactly what will be y done in various circumstances.
  17. 17. Examples of adaptive designs g Changes in: • Study eligibility criteria y g y • Randomization procedure • Treatment regimens of the different study groups (e.g., dose level, schedule, duration) • Total sample size of the study ( p y (including early g y termination) • Concomitant treatments used • Planned schedule of patient evaluations for data collection • Primary endpoint • Secondary endpoints • Analytic methods to evaluate the endpoints
  • whitepatches

    Jul. 13, 2010

Views

Total views

2,561

On Slideshare

0

From embeds

0

Number of embeds

1,138

Actions

Downloads

0

Shares

0

Comments

0

Likes

1

×