The test-retest reliability of the Opiate Treatment Index in nyaope users in Johannesburg

TL;DR

This study evaluated the reliability of the Opiate Treatment Index for nyaope users in Johannesburg, finding mixed results across different domains.

Contribution

The study provides the first assessment of the Opiate Treatment Index's reliability in a South African context.

Findings

The OTI showed excellent test-retest reliability for injecting and sexual behaviour domains.

Moderate reliability was found for criminality, general health, and social functioning domains.

The ICC for nyaope was 0.38, indicating low reliability for this specific drug domain.

Abstract

Epidemiological studies suggest that nyaope, a heroin-based drug, is widely used in South Africa. Yet few reliable research tools are available to assess treatment outcomes of users. The Opiate Treatment Index (OTI), a tool developed in Australia, could potentially facilitate research on context-specific South African treatment outcomes. However, we know little of its test-retest reliability.

This study aimed to assess the test-retest reliability of the OTI among a sample of nyaope users in Johannesburg.

This study was conducted across three substance use treatment facilities in Johannesburg.

The OTI was administered to 53 nyaope users at baseline and one week later. To determine the test-retest reliability of the OTI, the intra-class correlation coefficients (ICC) and the Brennan–Prediger coefficients of the two interviews were calculated.

The ICC of the Q-scores from the data sets along with the Brennan–Prediger coefficient for the substance use domain were calculated. The ICC for nyaope was 0.38. Brennan–Prediger coefficients were as follows: alcohol – 0.96, crack-cocaine – 0.89, cannabis – 0.92, methaqualone – 0.85 and crystal methamphetamine – 0.89.

A significant positive finding was the excellent test-retest reliability of the injecting and sexual behaviour domains and moderate reliability of the criminality, general health and social functioning domains.

The results of this study provide insight into the reliability of this tool and for its use in future studies in the South African context.

Figures