Personal Social Services Research Unit
Cornwallis Building
George Allen Wing
University of Kent
Canterbury
Kent
CT2 7NF
01227 827672
ascot@kent.ac.uk
www.pssru.ac.uk/ascot
The data entry tools available on this site can be used to calculate the overall ASCOT well-being score for individuals on the basis of their responses to questions about their current social care-related quality life (SCRQoL), and their SCRQoL in the absence of services. Investigators can enter details from service users collected using ASCOT questionnaires (by interview or through self-completion). Current SCRQoL information is collected in either three or four levels across either eight or nine domains of SCRQoL (depending on the service option). The data entry tools then apply the relative importance or 'preference' weights estimated in the MOPSU and OSCA projects to each level within each domain to obtain the overall current SCRQoL score. This overall score is a number between zero and one which quantifies a service user's well-being in terms of the degree to which they experience a range of care needs. In developing these indicators, the research team experimented with a number of options for these measures which balance sensitivity and validity of the indicators against ease of use and general parsimony.
The indicators embodied in these data entry tools were developed in two projects which started in 2007 and investigated the outcomes of people using care home services and those using day care services. The three-level, nine-domain tool was developed for the day care MOPSU project. A four-level, eight-domain tool was subsequently developed in the OSCA project. A working version is now available, although some aspects of the measure are undergoing further testing and refinement. In particular, the preference weights in the data entry tools refer to a version with very minor wording differences to the current four-level version. We anticipate this making only minor differences, but new preference weights will be estimated in due course.
There is a great deal of common ground in these two measures. The intention was to assess whether adding a further level within each SCRQoL domain produced a justifiable increase in precision. The main difference between the three- and four-level versions is that the latter has more sensitivity in the top SCRQoL states: that is, for people with low needs. Nonetheless, the difference is small. We therefore advise users of these tools to select the version that best suits their specific needs.
The data entry tools are designed to calculate the scores for samples of people using care services. They are used in the expectation that an appropriate sample of services users are asked to provide well-being responses, where good surveying practice is adopted. In this way, investigators can use these tools to determine the sample average current SCRQoL score as well as information summarising the distribution of scores. The potential applications are many. One highly relevant example is where investigators compare the (current) SCRQoL of two samples of people: one where a new service is provided and the other where the current service is used. The difference in mean SCRQoL scores between these groups is an indication of the SCRQoL gain associated with the new service compared with the existing service.
The data entry tool for the three-level indicator also includes provision for investigators to enter service user condition, service use and background data collected from the day care self-completion questionnaire. These data can be used to predict the SCRQoL score of the person were they not to receive day care. A formula for this purpose was developed as part of the MOPSU project. By taking the difference between the current SCRQoL score and this predicted 'expected' SCRQoL score, predicted SCRQoL gain from day care service use can be calculated. This approach is useful because it precludes the need for investigators to compare day care service users with a group of people that did not use the service (as in the above example). Nonetheless, this prediction is an estimate. It is valid only for a sufficiently large and representative sample of people; individual person scores are generated, but the mean difference should be used, not individual person differences. The mean SCRQoL gain is calculated in the data entry tool along with the current SCRQoL scores.