Most researchers do not approach survey data collection ab initio, but rather draw on "conventional wisdom", often encapsulated in existing textbooks or in what they have been taught at undergraduate or postgraduate level. We thought, therefore, that it would be appropriate to preface our review of empirical studies with a summary of current "expert opinion" and then to seek confirming or refuting evidence from primary research. Sources of expert opinion were key texts on survey methods (e.g. 1,5,13,40 ). These were identified through the personal and institutional libraries of the research team. DatabasesIn our search for primary research studies, previous literature reviews and theoretical articles, we initially prioritised four electronic databases for searching:This review should be reviewed as a guide to best practice, not a definition of best practice. It is a decision aid, not a substitute for critical appraisal of the options available.

As part of a large community-based study, we retrospectively examined the clinical course of epilepsy in an unselected population of people who had a recent history of seizures or were receiving antiepileptic drugs (AEDs). Clinical information was collected from medial records, and information about psychosocial functioning was obtained by means of postal questionnaires sent to identified subjects. The response rate to the postal questionnaire was 71%. There were some deficiencies in the recording of clinical data, which is not unusual since data were taken from records held by primary physicians rather than from hospital clinics. Nevertheless, findings regarding the clinical course of epilepsy corresponded to those of earlier studies. Fifty-seven percent of the sample had had at least a 2-year seizure-free period and 46% of subjects were currently in a remission of at least 2-year duration. There was a clear relationship between current seizure frequency and levels of anxiety and depression, perceived impact of epilepsy, perceived stigma, and marital and employment status. The relationship of seizure frequency and other clinical variables to psychosocial function was explored by multivariate analysis techniques. The amount of variation in scores on the various measures of function accounted for by the clinical variables was small. The most important predictor was current seizure activity, which was the first variable to enter the regression analyses for six of the eight measures of psychosocial function considered. Age at epilepsy onset also emerged as a significant predictor for depression, stigma, and marital status. In individuals with epilepsy in remission, there was little evidence that psychosocial functioning was associated with length of remission, a finding which may in part reflect the nature of this study population. The results indicate that there are several more important predictors of psychopathology and social dysfunction in epilepsy and suggest several implications for treatment interventions.

Objective To evaluate the use of a computerised support system for decision making for implementing evidence based clinical guidelines for the management of asthma and angina in adults in primary care. Design A before and after pragmatic cluster randomised controlled trial utilising a two by two incomplete block design. Setting 60 general practices in north east England.

Inappropriate analysis of cluster trials can lead to the presentation of inaccurate results and hence potentially misleading conclusions. We have demonstrated that adjustment for clustering can be applied to real-life data and we encourage more routine adoption of appropriate analytical techniques.

No abstract