Background
A recent systematic review carried out by researchers at Leeds and Keele Universities was concerned not only with evaluating outcome - the effectiveness of written medicines information for patients - but also with process - the role and value of such information to patients and health professionals. Process evaluation can help to explain outcome findings1.

Methodologies of the studies reviewed for role and value.
It was anticipated that relevant studies for the role and value strand of the systematic review would be predominantly qualitative1. The difficulty of searching for qualitative studies has been highlighted by other systematic reviewers2. The literature search covered both strands of the review and was designed to be sensitive rather than specific, in order to find qualitative studies. This strategy produced over 50,000 references. Twenty eight papers (27 studies) were found to meet the inclusion criteria for the role and value strand of the review. Nine of the studies were qualitative, using mainly semi-structured interviews or focus groups. Seventeen were quantitative. Mostly these were surveys (12), but four were laboratory based studies which involved members of the public reviewing information about medicines. One was a mixed methods study. In eight of the quantitative studies and four of the qualitative some or all of the lay participants were not taking the medicine of the written information being evaluated nor had a related illness. These have been referred to by us as hypothetical studies.

Evaluation of the methodologies
Three different rationales for carrying out the research were identified. Twelve studies were carried out in relation to medicine information policy initiatives. Eight studies were set within a context of aiming to increase patient compliance and seven within a context of patients as active participants in their care. Questionnaires used in the surveys mostly had closed questions in which participants selected a response category predetermined by the researcher. Thus participants' responses were constrained to fit the a priori ideas of the researcher. In contrast semi-structured interviews and focus groups encourage respondent to dwell on what is important to them. In the quantitative studies the findings were more likely to reflect the starting position of the researcher. For example written information being secondary to spoken information from the doctor emerged in the qualitative studies but not in the quantitative. In hypothetical studies the context of illness and medicine taking was absent and respondents speculated on how they might use written information should they actually be prescribed a medicine. There was evidence to show that what patients anticipate doing and what they actually do are not the same.

Conclusion
Selecting an appropriate methodology in order to understand the use and usefulness of patients' written medicines information is crucial. Researchers' preconceived ideas influenced survey findings, and hypothetical studies could only answer certain questions.