Introduction

One of the major problems of opiate treatment evaluation research has been the lack of comparability of research findings. Studies vary both in the domains selected as outcome variables and in the criteria for "success" within these domains. For example, some major outcome studies concentrate exclusively on drug use and criminality as outcome variables (e.g. DeLeon, 1986; DeLeon et al, 1986), whereas others regard factors such as employment and psychiatric status as also constituting relevant outcome domains (Hubbard et al, 1983; McLellan et al, 1986; Simpson & Marsh, 1986).

Even when variable domains are comparable between studies, the manner in which these variables are measured renders comparability difficult. The area of drug use provides the most salient example of this. In many studies complete abstinence is used as the criterion for success, and the percentage of clients who achieve abstinence is what is reported (e.g. DeLeon, 1986). Other studies report frequency of use of substances (e.g. Hubbard et al, 1986), while still others report time to relapse (Fisher & Anglin, 1987; Simpson et al, 1982). The consequence of these discrepancies in methodology is the virtual impossibility of comparisons between studies.

One of the reasons for the non-comparability described above is the differing values of various researchers regarding the relevant goals of treatment. Another major contributor to the present situation has been the relative absence of standardized instruments in the area of opiate treatment outcome research. Researchers have, on the whole, employed questionnaires which are, unique to their studies, of unknown validity and reliability.

The aim of the present project was to provide a comprehensive, standardized set of measures for the evaluation of opiate treatment. In constructing the Opiate Treatment Index (OTI) as a comprehensive evaluative tool, the authors considered that a number of criteria should be met:

The primary consideration was that the index should be multi-dimensional in nature. Clearly, the aims of opiate treatments extend beyond the cessation or reduction of drug use. Opiate use is associated with a broad range of health, legal and social problems. Furthermore, there is evidence that the problems associated with opiate use are relatively independent (e.g. McLellan et al, 1981). An assessment instrument should reflect this heterogeneity in its structure.
The instrument should be based upon objective data rather than on the impressions of interviewers. This is to avoid the problems engendered by the differing criteria for "success" employed by different researchers. What such objective scales should provide are data on the recent behaviour of clients in a number of outcome domains. Obviously the interpretation of the data regarding success or failure will depend upon the ideology of the researcher. This should, however, be independent of data collection.
The variables employed should, if possible, be continuous rather than categorical, in order to maximize the sensitivity of the instrument to actual behaviour change.
To be of maximum utility, such an instrument should have both clinical and research applications. The scales should provide information which is of interest to clinical staff, as well as providing global research data.
To be of use in clinical settings, an assessment instrument should be relatively brief and easy to administer. The instrument must be able to be employed by both medical and non-medical personnel, given the diversity of staff engaged in both the clinical and research aspects of opiate treatment.
Such an instrument should be of proven reliability and validity.

The OTI was constructed to meet these criteria. A complete description of the OTI and data concerning its reliability and validity can be found in Darke et al (1992). The OTI has also subsequently been successfully tested for reliability and validity in the United Kingdom (Adelekan et al, 1996a) and New Zealand (Deering & Sellman, 1996).

As with all assessment instruments, there are problems in obtaining accurate data in a clinical setting that do not arise in an anonymous research setting. However, in a study addressing this issue (Adelekan et al, 1996b) had researchers and clinicians separately conduct OTI interviews among London methadone maintenance patients. There were no significant differences between reports to researchers and clinicians of heroin use, injecting frequency, crime or HIV risk-taking in the preceding month. It should be noted, however, that assurances were given to subjects that their reports would have no effect on their treatment, regardless of the interviewer. This issue is discussed further in the section on the psychometric properties of the OTI.

It should be emphasised that the OTI should not be considered as an alternative to a clinical assessment. Rather, it provides data on key variables over a range of outcome domains. There will always be other variables of interest to clinicians which are not covered in the OTI.