A behavioral trial can improve a health outcome without demonstrating improvement on every drug-use measure. The historical ARTEMIS trial is an example: its HIV viral-load findings, self-reported stimulant-use findings and urine results answered different questions.
What comparison was randomized?
Carrico and colleagues randomized 110 sexual minority men living with HIV who had biologically confirmed recent methamphetamine use: 55 to five positive-affect intervention sessions and 55 to an attention-control condition. Both groups received three months of community-based contingency management, including incentives for stimulant abstinence.
The experimental contrast was therefore the added intervention versus the attention control on that shared background. It was not positive-affect training alone versus no care, nor a comparison assigning participants to take or stop methamphetamine.
What happened to HIV outcomes?
The intervention group had lower log10 HIV viral load at six, twelve and fifteen months. Risk of at least one unsuppressed measurement, defined as HIV RNA of at least 200 copies/mL, was lower: risk ratio 0.33, 95% confidence interval 0.15–0.69.
That endpoint concerns one or more qualifying viral-load measurements over follow-up. It is not a 67-percentage-point reduction in an individual’s probability, a measure of methamphetamine abstinence or proof of a particular behavioral mechanism.
Did urine testing tell the same story?
There were no significant between-group differences in the proportion with a reactive urine stimulant result at each follow-up assessment. For at least one reactive specimen over fifteen months, the risk ratio was 0.80, 95% interval 0.63–1.02, p = 0.056. This result did not meet the study’s significance criterion.
The endpoint concerned stimulant-reactive toxicology, rather than a quantified methamphetamine dose or a continuous record of abstinence. A favorable HIV outcome cannot be substituted for that laboratory drug-use outcome, and a non-significant urine result should not erase the viral-load finding.
Which other findings require separation?
The investigators reported decreases in self-reported stimulant-use frequency at selected follow-ups. However, the group-by-time test for methamphetamine craving was not significant, p = 0.100, and the corresponding CD4-cell analysis was also non-significant, p = 0.584.
These results show why a broad statement that the intervention improved “methamphetamine outcomes” is insufficient. A reader needs to know whether the claim concerns self-report, urine reactivity, craving, viral load or immune-cell count. Different endpoints can change differently.
How do selection and follow-up affect interpretation?
Of 136 eligible consenting participants, 110 completed a run-in and were randomized. That process selected people able to engage with several preliminary visits. Follow-up retention fell to 71% at fifteen months, and the viral-load table contained 74 measurements at that time, rather than results from all 110 randomized participants.
The study also took place in a well-resourced San Francisco setting and included only sexual minority men with HIV who used methamphetamine. The authors identified the need for broader effectiveness research and more objective assessment of antiretroviral adherence. Their proposed pathways were not established merely because several outcomes changed together.
What is the defensible conclusion?
This 2019 phase II trial supports a specific added behavioral intervention within an incentive-based program for a selected population. It does not establish universal abstinence, identify which component caused each benefit or provide current treatment instructions. Its scientific value is strongest when the positive primary result remains beside the null and uncertain secondary findings.
Source
Carrico AW and colleagues. Randomized controlled trial of a positive affect intervention to reduce HIV viral load among sexual minority men who use methamphetamine. Journal of the International AIDS Society. 2019;22(12):e25436. DOI: 10.1002/jia2.25436. Full text; bibliographic record.