The development of research tools used in the STORK Study (the Scottish Trial of Refer or Keep) to explore midwives' intrapartum decision making

Objectives: to develop appropriate tools to assess midwives' attitudes and behaviour in relation to decision making involving risk. Design: a questionnaire and series of vignettes were developed and testes to explore midwives' intrapartum decision making in relation to their attitudes towards risk. An innovative online computer package was developed specifically for use in the STORK Study which enabled the programme to be very tightly controlled with limited functions accessible to participants. Testing: a pilot study was conducted with over 50 midwives and nurses to ensure face and content validity of the vignettes and questionnaire. Initially designed to be a paper-based study, rigorous piloting highlighted the many difficulties in presenting it in that particular format. The solution to this problem was to develop the study as a secure online package. Findings: online data collection provided the researchers with a greater degree of control of the data collection process, not achievable using traditional paper survey methods. Another example of this control is the immediate entry of data from participants' responses to a background database which automatically stores and backs up data this means that no additional time is required for data entry. The cost of employing an information technology professional was easily offset by the financial savings made through the limited use of stationery and postage. Key conclusions: although the development and testing of the research tools for the STORK Study was labour and time intensive, ultimately a questionnaire and vignette package was produced that had been rigorously tested by over 50 midwives and nurses. The researchers are confident in the reliability of the questionnaire and vignettes, as well as the validity of the data collected. The use of an online survey is clearly indicated when the population has readily available internet access, and where controlling the process of data collection is required, as such control cannot be achieved in traditional survey and questionnaire implementation