In this week's BMJ, two analysis articles about testing for HIV argue for changes in policy that would expand the number of people routinely tested by promoting opt-out approaches.1 2 Both papers argue that this would increase the proportion of the population who know their serostatus and would decrease the number of late diagnoses of HIV. The papers agree about the benefits of swift diagnosis of HIV, including reduced mortality and morbidity, less onward transmission because treatment should reduce infectiousness, and reduced costs of acute treatment and lost productivity.

We argue that a more precise goal for any changes in policy should be to reduce the average time between HIV infection and diagnosis in people who become infected.3 This goal allows a range of measures of success beyond a CD4 count below 200×10⁶/l and acknowledges that the earlier HIV is diagnosed the better.

As the secretary general of the United Nations highlights, improved epidemiological outcomes are dependent on people being able to test in “a social and legal environment that is supportive and safe.”2 This needs to apply …