Lemons et al. Respond

Wiewel et al. suggest that misclassification may account for differences between Medical Monitoring Project (MMP) estimates of the proportion of US transgender HIV patients that are transgender men and HIV case surveillance estimates of this proportion among persons living with diagnosed HIV in New York City and California. The risk of misclassification attributable to data entry is reduced by MMP’s electronic data collection, which has programmed logic checks based on responses to questions about sex at birth and gender to prevent errors. Furthermore, in certain sections of the interview, transgender persons are asked interview questions that would not pertain to cisgender persons, providing additional protection against data entry errors. We attribute the differences Wiewel et al. note to the method of measurement used to calculate estimates and substantial differences in populations.

We are not surprised that our methods estimate a higher proportion of transgender men than found in HIV case surveillance. Technical notes in New York City’s surveillance report acknowledge that transgender status may not reflect the individual’s self-identification¹; MMP relies on self-report, whereas surveillance data are obtained from diagnosing providers, medical chart reviews, and case report forms, in addition to self-report. In fact, Wiewel et al. found that between 25% and 48% of transgender persons receiving services at community agencies were not identified in New York City’s HIV surveillance system as transgender.² Furthermore, transgender status was not routinely collected by New York City HIV surveillance until 2005, which could result in substantial misclassification of transgender status in surveillance data among persons reported to the system before 2005.¹

Populations also differ between National HIV Surveillance System and MMP with regard to care status, geography, and timing. If transgender women are less likely than others to be retained in HIV care,^3–5 this could result in a higher percentage of transgender men among transgender HIV patients compared with those with diagnosed HIV. National US surveillance data estimated that 15% of newly diagnosed transgender persons were transgender men,⁶ in contrast to the 2% to 4% reported by Wiewel et al. for New York City and California.^1,7 Finally, transgender status is often collected by HIV surveillance systems at HIV diagnosis; however, gender identity can change over time, so this can account for differences when comparing to current identity.

Adjudicating between estimates with substantial differences in populations and methods of measurement is difficult, but we agree with Wiewel et al. on the importance of collecting accurate data on transgender persons and support the need for ongoing data collection to help understand the needs of this important population.