The pitfalls of prevalence estimation: the case of regular and dependent methamphetamine use in Australia

Abstract

Objectives: The prevalence of regular and dependent methamphetamine use in Australia in 2013/14 was recently estimated indirectly using multipliers at 268,000 regular (defined as >monthly use in the past 6 months) and 160,000 dependent (defined as >3 on the Severity of Dependence Scale) methamphetamine consumers, respectively. We replicated this approach using more contemporary multipliers and alternate data sources to explore identified limitations with this approach.

Design: Two multipliers were derived from surveys of methamphetamine consumption in Sydney (NSW) in 2003/04 (the original multiplier) and Melbourne (VIC) in 2010 (a more contemporary multiplier), and applied to the time series of treatment utilisation figures for NSW and VIC, along with ABS-estimated resident population figures.

Results: Using the more contemporary VIC multiplier, we estimated numbers of regular (67,250, 95% CI = 50,312-91,635) and dependent (43,232, 95% CI = 34,477-57,947) methamphetamine users for Australia in 2013/14. This is approximately one-third of the figures reported previously. Our national estimates using the 2003/04 NSW multiplier applied to NSW treatment data (175,452, 95% CI = 124,514-256,575 for regular, and 103,761, 95% CI = 73,576-154,699 dependent, methamphetamine users respectively) are approximately two-thirds of the figures reported previously.

Conclusions: We urge caution in using treatment multipliers to estimate the prevalence of illicit drug use, particularly during periods of dramatic changes in treatment system capacity, user behaviour, drug purity and policing practices.

Keywords:

• Prevalence estimation
• multiplier methods
• methamphetamine use

Acknowledgements

We gratefully acknowledge the Victorian Department of Health and Human Services for providing Victorian treatment data from the Alcohol and Drug Information System.

Disclosure statement

PD has received funding from Gilead Sciences Inc for work unrelated to this study. PD has received funding from Reckitt Benckiser for work unrelated to this study. BQ is the recipient of a Fulbright Fellowship. The Burnet Institute gratefully acknowledges the funding received through the Victorian Operational Infrastructure Scheme. These funders had no role in the work.

Notes

1 Degenhardt et al state that the survey was conducted in 2002/03 but published reports by the survey authors (see 8) suggest that it was conducted in late 2003/early 2004, so we have only used the multiplier prospectively from 2003/04 onwards and have not applied it retrospectively to 2002/03.

2 We cite the first question of the treatment service section of McKetin et al.'s (2005) survey as underpinning the multiplier as participants need to respond 'yes' to this question before being asked the main question for the multiplier “How many different times have you sought help for your drug use….”?. It is possible that those participants who received some substantial form of treatment as well as an 'assessment-only' may have included their 'assessment-only' in their responses. It is not possible to know the extent to which this is the case from McKetin et al.'s (2005) report.