Appraise
Why do we need to appraise the evidence behind ideas and suggestions?
Not all evidence is believable. Thus it will be necessary to appraise the evidence for validity, importance and relevance
How will we appraise evidence for validity?
For validity (“Can one believe the evidence?”): We will use established appraisal checklists. For examples, for therapeutic interventions, we will be looking for good quality (systematic reviews of) randomised trials, failing which controlled cohort studies, and in the absence of these, for observational studies. The aim will be to judiciously interpret the whole evidence. Appraisal checklists for various types of evidence can be found at, for example: http://www.cebm.net/critical_appraisal.asp
For importance and relevance: We will ask the following questions:
- Does the intervention address an important and common clinical problem?
- Is it effective? (magnitude of effect)
- Is it safe?
- In overdose
- Misuse
- Minimal interactions with other interventions - Is it affordable? (or affordable adaptations possible)
- Is it simple to administer? (oral/rectal/intramuscular)
- Is it simple to store? (e.g. no need for a fridge)
- Does it have limited alternative use? (to reduce the risk of abuse)
Other dimensions of appraisal will include:
- Consistency: The extent to which different studies found similar results
- Robustness: The extent to which minor alterations in the data do not change conclusions drawn from those data
How will we report the findings of evidence appraisal?
The appraisal of evidence will be reported on the Ammalife website as an “Evidence Appraisal Report” (EAR). Ideas and suggestions of interventions for which valid and important evidence is found will be considered for application. Such interventions will be grouped into “Implementation Packages” (IPs) – targeting the various conditions and the varied settings and audiences (e.g. General Public, Traditional Birth Attendants, Midwifes, Primary Care Physicians, Hospital Practitioners, Healthcare Managers and Planners).
