Systemic review in health care: chapter 6 interventions

1
Systemic review in health care chapter 6
interventions

• Wen-Hsuan Hou, MD
• Department of Physical Medicine and
  Rehabilitation, E-Da Hospital

2
The question

• Types of intervention used for systemic review
• Therapy for a specific disease
• A change in a risk factor
• Screening for earlier diagnosis
• Defining feature
• Specific activity undertake to improve or prevent
  adverse health outcomes

3
Study design

• Randomized controlled trials (RCT)
• Control for confounders
• Provide the best evidence for interventions
• Surveillance data
• Best on rare adverse effects
• Cohort studies
• Best on the prognostic factors needed to predict
  individual pretreatment risk

4
Finding relevant studies

• Finding existing systemic reviews
• Appendix A page 109
• Finding published primary studies
• CCTR (Cochrane Controlled Trials Registry)
• CENTRAL
• More recent studies handsearching key journals
  conference proceeding
• Finding unpublished studies
• Appendix in the Cochrane handbook
• Contact principle investigators

5
Standardizing appraisal and selecting studies

• Has selection bias (including allocation bias)
  been minimized?
• Have adequate adjustments been made for residual
  confounding?
• Have the final outcomes been adequately
  ascertained?
• Intention-to-treat analysis
• Degree of follow-up
• Has the measurement or misclassification bias
  been minimized?

6
Appraisal checklists
7
Should scales be generic or specific?

• Precise methods used for the outcome measure is
• Part of the quality of the study
• Vital for interpreting the results
• Keep the appraisal items to a few simple items
  which have proven to be important causes of bias

8
Presenting studies results by summary table
9
Presenting studies results by graphic presentation

• Forest plot

10
Synthesis of study results

• Two major aims of meta-analytic synthesis of
  controlled trial data
• Find a summery estimate of the overall effect
• Examine whether and how this average effect is
  modified by other factors
• Use single summery outcome measure
• A common metric (unit)
• A directly interpretable manner (e.g. reduction
  in the risk of death, days of the symptoms)

11
Outcome measures

• Discrete events
• Table 4.1 RD, RR, OR or average time to event
• LAbbe plot event rate in treatment v.s. control
  group
• RD varied most over different population RR and
  OR about equally stable
• Survival analysis for time-to-event varies
• Continuous outcomes
• The proportional improvement
• The standardized difference treated control/
  standard deviation in the control group

12
A common interpretable outcome measure

• We want to know both
• Average effect
• Variations from the average effect

13
Assessing heterogeneity

• Causal factors of the difference in the effects
• Disease feathers e.g. stage, severity
• Patient feathers e.g. age, ethnicity
• Intensity or timing of the intervention most
  important
• Study design feathers e.g. study duration,
  outcome measures
• Even the heterogeneity is nonsignificant,
  exploring for causes of variation is reasonable
  and useful

14
Economic evaluation

• Cost analysis
• Cost-effectiveness analysis
• Cost-utility analyses
• Cost-benefit analyses
• Cost-effectiveness ratio (C-E ratio)
• (Costs- Savings)/Effectiveness
• Direct systemic review of the C-E ratio is in
  appropriate

15
Further information

• Cochrane Handbook
• Full description of the process for Cochrane
  systemic reviews

Questions for part 2 interventions

• Antibiotics usage for acute otitis media in
  children aged under 2 years