The Development of a Holistic and Quantitative Tool for the Assessment and Improvement of Survey Quality.

There are a variety of guidelines and methods available to measure and assess survey quality. Most of these are based on qualitative descriptions. In practice, they are not easy to implement and it is very difficult to make comparisons between surveys. Hence there is a theoretical and pragmatic demand to develop a mainly quantitative based survey assessment tool. This research aimed to meet this need and make contributions to the evaluation and improvement of survey quality. Acknowledging the critical importance of measurement issues in survey research, this thesis starts with a comprehensive introduction to measurement theory and identifies the types of measurement errors associated with measurement procedures through three experiments. Then it moves on to describe concepts, guidelines and methods available for measuring and assessing survey quality. Combining these with measurement principles leads to the development of a quantitative based statistical holistic tool to measure and assess survey quality. The criteria, weights and subweights for the assessment tool are determined using Multi-Criteria Decision-Making (MCDM) and a survey questionnaire based on the Delphi method. Finally the model is applied to a database of surveys which was constructed to develop methods of classification, assessment and improvement of survey quality. The model developed in this thesis enables survey researchers and/or commissioners to make a holistic assessment of the value of the particular survey(s). This model is an Excel based audit which takes a holistic approach, following all stages of the survey from inception, to design, construction, execution, analysis and dissemination. At each stage a set of criteria are applied to assess quality. Scores attained against these assessments are weighted by the importance of the criteria and summed to give an overall assessment of the stage. The total score for a survey can be obtained by a combination of the scores for every stage weighted again by the importance of each stage. The advantage of this is to construct a means of survey assessment which can be used in a diagnostic manner to assess and improve survey quality.

Priorities for future intensive care research in the UK: results of a James Lind Alliance Priority Setting Partnership.

This James Lind Alliance (JLA) Priority Setting Partnership aimed to identify and prioritise unanswered questions about adult intensive care that are important to people who have been critically ill, their families, and the health professionals who care for them. Consensus techniques (modified Delphi and Nominal Group) were used to generate suggestions using online and postal surveys. Following verification and iterative editorial review, research topics were constructed from these suggestions. These topics were presented in a second online and postal survey for rating. A Nominal Group of 21 clinicians, patients and family representatives subsequently met to rank the most important research topics and produce a prioritised list. The project was coordinated by a representative Steering Group and independently overseen by the JLA. The initial survey and review of the literature generated over 1,300 suggestions. Preliminary editing and verification permitted us to encapsulate these suggestions within 151 research topics. Iterative review by members of the Steering Group produced 37 topic statements, subsequently rated by participants. Using the mode to determine importance, 19 topics were presented to the group from which a ‘top three’ intensive care research priorities were identified and a further nine topics were prioritised. By applying and adapting the JLA methodology to focus on an area of care rather than to a single disease, we have provided a means to ensure that patients, their families and professionals materially contribute to the prioritisation of intensive care research in the UK.