Sometimes Less Is More: The Case of HIV Testing

June 27 was National HIV Testing Day. This got lost in the SCOTUS health reform fight, but it’s still worth noting. If you’re wondering, get yourself tested. It is important. It is quick. It’s easy to do. The oral rapid test is especially quick and painless. There are many good reasons to know your status.

This month’s American Journal of Public Health includes results from a nice randomized study led by my distinguished colleague and friend , Lisa Metsch. The analysis concerns a narrow but important set of challenges surrounding HIV-testing patients in substance abuse treatment. For obvious reasons, it’s important to regularly test these patients. Many practical obstacles hinder this effort. (I’ve helped to research these issues in various contexts, e.g. here).

One reality of such efforts is that the vast, vast majority of people tested will be HIV-negative. In one study, we visited testing programs at six safety-net emergency departments. Rates of unknown seropositivity were about 1% in each location. So the cost and administrative burdens are largely determined by the 99% of people tested who are free of HIV. The consent process, pre- and post-test counseling, the reporting of results, and whatever behavioral counseling are offered to this 99% will therefore exert a huge impact on an organization’s willingness and ability to conduct the kind of broad HIV testing that CDC and public health advocates would like to see.

The issue of sexual risk is particularly sensitive in substance use treatment. There are obvious public health concerns regarding HIV and other sexually-transmitted infections within this population. However, it’s unclear that post-test sexual-risk behavioral counseling does much good. The provision of such counseling raises significant logistical issues, and becomes, itself, a barrier to population screening. One can refer patients for specialized HIV screening and post-test counseling at health departments and other offsite facilities. Yet we know from experience that many people will never make it there. Many others, once tested, don’t return for their results.

For many treatment programs, the most appealing and economical approach to HIV testing is to offer on-site rapid oral testing, reporting the results quickly after these become known, and to provide only very limited post-test counseling to the 99% of patients with negative results. Whether one loses an important risk-reduction opportunity by curtailing post-test counseling for HIV-negative patients remains unknown.

Thus, Metsch and her colleagues implemented a randomized trial in which they randomized 1281 HIV-negative (or status unknown) adults who reported no past-year HIV testing to three different treatment arms: (1) referral to off-site HIV testing (n=429), (2) HIV risk-reduction counseling with on-site rapid HIV testing (n=433), and (3) verbal information about testing only with on-site rapid HIV testing (n=419).

So what happened?

There were two obvious results. On-site testing led many more people to be tested and to actually get their results. (18.4% off-site, vs. 79.7% on-site with risk-reduction counseling, and 84.8% on-site with information only). But mean self-reported rates of unprotected sex were essentially identical in the three groups. (There were some apparent differences in injection risks, with the post-test counseling group slightly more likely to report that they discontinued needle sharing. This latter finding suggests the potential value of individual-level counseling among injection drug users.)

For a broader population in substance abuse treatment, I read this study to say that broad-based rapid HIV testing is feasible, but that post-test counseling doesn’t seem to accomplish what we hope in reducing sexual risk. If it’s easy to provide such counseling– great. But it’s not always easy. And if the provision of such counseling hinders broader testing, I’d rather have less-intensive services reaching more people.

It’s nice to write about a nonpartisan issue in services research. And oh yeah. Please know your status, people. Give yourself some QALYs.

[Cross-posted at The Incidental Economist]

Harold Pollack

Harold Pollack is the Helen Ross Professor at the School of Social Service Administration at the University of Chicago.