The methods used to inform item generation in this study reflect best practice guidelines in the initial stages of questionnaire development [9], [10] and [11]. Gaining a rich and detailed understanding of the construct to be measured is best achieved from focused interviews with the relevant

population. Whilst this is particularly relevant for condition specific measures however, this generic measure needed to be applicable to people over a range of health conditions and roles (i.e. patients and carers). The opportunity to carry out BKM120 cost secondary data analysis using a large interview archive which spanned a range of conditions was therefore particularly useful for the development of this item pool. However, analysis of secondary data can be restrictive in comparison to primary research where the interviewer can focus their questions on the issues of most interest to their

own research agenda [15]. In some interviews the original reseracher had not probed into participants experiences of using health websites. Integrating secondary analysis of several, purposively selected collection of interviews with a conceptual literature review and using confirmatory sources of data was therefore vital in ensuring all check details potential themes were investigated thoroughly and assisted the triangulation of the findings. Secondary data analysis has also been critiqued for lacking relevant contextual knowledge when the researcher was not involved in the primary research. However, the availability of Fludarabine datasheet video and audio files of interviews largely overcomes this problem. Suitability of the data was also assessed through a number of steps before formal analysis commenced: (1) thematic summaries

and participant biographies prepared by the primary researchers were read, (2) primary researchers were consulted to gauge the appropriateness of the data for the research purpose, and (3) primary researchers coding books of relevant themes from their initial analyses were made available to the research team. Cognitive interviews also confirmed the relevance of the qualitative findings. Current studies evaluating ehealth interventions are limited by the lack of a suitable instrument to measure health-related effects associated with using a health website. A person may use guidance, filtering and accreditation tools [29] to help them assess health information on the internet. However, these instruments do not capture how a person may be affected through engaging with a website and users may be concerned of coming across factually correct, yet unwelcome information [30]. Furthermore, such accreditation tools fail to take into account that websites provide more than information, but can also be mechanisms of support. The potential effects of using health-related websites and support groups have been explored [31] using self-report measures which were not specifically developed to capture the range of effects associated with internet use.