Clinically useful outcome measures

First published: 10 May 2019 | Last updated: 20 May 2019

What might be a useful set of outcome measures rather depends upon the perspective taken.  Substance use, and particularly dependent use, may itself be a problem but is usually interwoven with other social, psychological or health harms.   People in treatment typically use and misuse a variety of substances and switch from one class of drug to another so that it is difficult to compare outcomes looking at substance use alone.  Some harms are clearly attributable to particular substances others much less so and some are associated with more harmful routes of administration than others.

A good outcomes package will have the following characteristics:  i) easy to collect – no more than 10 minutes of service user time; ii) selection of measures that are relatively independent of each other and adequately reflect aspects of life; iii)        measures that are of universal applicability; iv) rating scales should be completed by the service user or someone independent of their treatment; v) self completion scales must be written in plain English; vi) rating scales must have adequate psychometric properties and, preferably, published population norms.  RESULT is a package of separately developed scales all of which have published psychometrics:  a) dependence domain – Leeds Dependence Questionnaire (Raistrick et al., 1994); b) psychological domain – Clinical Outcomes in Routine Evaluation (Evans et al., 2002); c) social domain – Social Satisfaction Questionnaire (Raistrick et al., 2007); d) substance use domain – ICD 10 categorical codes (World Health Organisation, 1993); and e) costing domain – EQ5-D (EuroQol Group, 1990).  The scales selected have items that are useful in the clinical assessment and they are intended to be used as motivational tools in their own right.  Factor analysis has shown the LDQ, CORE, and SSQ to be essentially independent components of the outcomes package.  Using the RESULT measures in practice shows that most change across all outcome measures occurs within the first three months but does then continue through to 12 months.  Changes in dependence are the first to show while changes in social satisfaction occur much more slowly.

Clinically significant change is a concept that takes account of baseline scores and error due to the measuring instrument.  Clinically significant change can be said to have occurred where the change is a) statistically reliable and b) the end scores fall within a normal population range.  Using these criteria the RESULT measures have found around 60% of heroin users and 50% of alcohol users to reach clinically significant change at 12 months.


Download the presentation

Download the full presentation here

Dr Duncan Raistrick