Article type
Year
Abstract
Background: The Cochrane Library is regarded as "the best single source of reliable evidence on the effects of health care". However, the majority of the reviews are conducted in developed countries and only the abstracts are freely available.
Objectives: To describe the relevance of recent reviews to developing countries and to assess the presentation of information necessary for application within a) the abstract; b) the text (introduction, discussion and implications for practice).
Methods: All completed reviews published in Issue 1 or 2, 2000 of the Cochrane Library were eligible for inclusion. The following data were extracted: i) estimated size of benefit and control event rate; ii) estimated size of harm and control event rate; iii) estimated size of benefit in specific sub-groups (or statement of "not assessed" if appropriate); iv) variation in estimated benefit with underlying risk in study populations. Results were coded as: present; partial information (quantitative); partial information (qualitative); absent; or not applicable.
Results: 130 reviews from 34 review groups were assessed. One review was excluded because it did not contain an abstract. Seven were regarded as directly relevant to developing countries by the authors. While most abstracts provided an estimate of the size of benefit, only around 20% provided direct or indirect information about the estimated control event rates. Around 50% of abstracts described harmful effects. Very few contained any information about the assessment of important sub-groups or variation in the estimates of benefit with underlying risk. Length of abstract did not appear to be a limiting factor.
Conclusion: Only a small number of reviews currently being published specifically address interventions most needed in developing countries. However, many others contain information that could be applied in these populations if presented appropriately. A standardised abstract with more explicit presentation of the overall benefits and harms (including estimated control event rates and variation with underlying risk) may increase the worldwide impact of reviews.
Objectives: To describe the relevance of recent reviews to developing countries and to assess the presentation of information necessary for application within a) the abstract; b) the text (introduction, discussion and implications for practice).
Methods: All completed reviews published in Issue 1 or 2, 2000 of the Cochrane Library were eligible for inclusion. The following data were extracted: i) estimated size of benefit and control event rate; ii) estimated size of harm and control event rate; iii) estimated size of benefit in specific sub-groups (or statement of "not assessed" if appropriate); iv) variation in estimated benefit with underlying risk in study populations. Results were coded as: present; partial information (quantitative); partial information (qualitative); absent; or not applicable.
Results: 130 reviews from 34 review groups were assessed. One review was excluded because it did not contain an abstract. Seven were regarded as directly relevant to developing countries by the authors. While most abstracts provided an estimate of the size of benefit, only around 20% provided direct or indirect information about the estimated control event rates. Around 50% of abstracts described harmful effects. Very few contained any information about the assessment of important sub-groups or variation in the estimates of benefit with underlying risk. Length of abstract did not appear to be a limiting factor.
Conclusion: Only a small number of reviews currently being published specifically address interventions most needed in developing countries. However, many others contain information that could be applied in these populations if presented appropriately. A standardised abstract with more explicit presentation of the overall benefits and harms (including estimated control event rates and variation with underlying risk) may increase the worldwide impact of reviews.