Data extraction and analysis

In step one, Data extraction and analysis basic details of the included papers (author, date, title, country) were indexed and organized into two categories-those representing the views and experiences of women and those representing the views and experiences of health care professionals.

For each category the papers were examined, and an index paper selected, chosen to best reflect the focus of the review.

The themes and findings identified by the authors of this paper were entered onto a spreadsheet.

This process was then continued for all the remaining papers, one at a time, iteratively building a set of themes that captured the quote material presented by the original authors, and their themes and statements, forming the ‘first order constructs’ of this stage of the review.

In step two the first order constructs were refined and merged to form second order constructs (Summary of Findings [SoF] statements), at one remove from the actual data in the included studies.

This process includes looking for what is similar between papers (‘reciprocal analysis’), and for what contradicted the emerging findings (‘refutational analysis’).

The disconfirming data identified in the on-going refutational analysis were used to refine the emerging constructs, so that the final analysis had high explanatory power for all the data.

The second order constructs were developed by KF and agreed by consensus between KF and SD.
In step three, the second order constructs (the SoFs) were assessed for confidence using the GRADE-CERQual tool.

This is a recently developed instrument, derived from the approach used in quantitative effectiveness reviews (GRADE).

CERQual’s assessment of confidence for individual review findings from qualitative evidence syntheses is based on four components: the methodological limitations of the qualitative studies contributing to a review finding, the relevance to the review question of the studies contributing to a review finding, the coherence of the review finding, and the adequacy of data supporting a review finding.

Based on these criteria, review findings were graded for confidence using a classification system ranging from ‘high’ to ‘moderate’ to ‘low’ to ‘very low’. As with study selection, the grades for each review finding were agreed by consensus and where there was disagreement a third reviewer (OTO) was asked to arbitrate.

In step four, the graded review findings were collapsed into over-arching interpretive themes as a means of synthesizing the data into an underlying ‘line of argument’ that describes the whole data set succinctly.