Statistics from Altmetric.com
Editor,—Wen et al1 should be applauded for their attempt to address the key question of whether or not condoms protect people from genital warts. However, some of the major study variables need clarifying, as they did not match up with my knowledge of the Sydney Sexual Health Centre (SSHC) database.
The article discussed the issue of “acquisition of genital warts” and was presented as an incidence study. Cases were defined as: “All patients with a new diagnosis of macroscopic genital warts who attended SSHC [in] 1996.” However, many of these patients had been previously diagnosed with genital warts elsewhere while others had recurrent lesions. In Australia, most genital warts are managed by general practitioners.2 Consequently, the experience of specialist services is biased towards recurrent and difficult cases. “New diagnosis” in this situation means new to the clinic but not necessarily new to the patient. This means that the main outcome measure was a mixture of incident, prevalent, and recurrent cases, with the possibility that the warts may have affected the behaviour of many of the study subjects.
The SSHC database does document whether a person has previously been diagnosed with HPV infection. To me, the study would have had more validity if patients with a past history had been excluded.
The diagnostic grouping for warts at SSHC does not distinguish between genital and anal lesions. The readers of the journal need to know that many of these male “genital wart” cases would have been homosexually active men with anal warts. This is important as risk factors for penile and anal warts may differ, potentially confusing the results of the present study.
Originally developed as an HIV risk measure, the condom use variable at SSHC only refers to the previous 3 months or since the last registration/disease episode. Wen et al's article1 failed to mention that this variable was time limited. As 3 months is the median duration before the appearance of exophytic warts,3 up to half of the relevant sexual behaviour may have been overlooked.
Finally, the referent group in the table describing condom use deemed as “Not applicable, no sex” should have been more accurately described as “No vaginal or anal sex in the previous 3 months.” Many of these people would have practised oral sex or other sexual acts during those 3 months. Others may have ceased practising vaginal or anal intercourse up to 3 months earlier because of their persistent or recurrent warts.
Large relational quality assured clinical databases can be powerful tools for health service evaluation, surveillance, and the generation of research questions. It may be prudent for researchers to engage the people responsible for designing and maintaining those databases to minimise errors of interpretation.
Editor,—We are grateful to Dr Dayan for her helpful and constructive comments. The major criticism of our paper relates to the selection of cases, and the possible inclusion of prevalent and recurrent cases as well incident cases. However, our concern with this possible bias at the outset of the study led us to exclude all patients with a history of previous genital warts. This included those previously diagnosed at SSHC, and those who gave a history of having their warts managed elsewhere. Consequently, when we state a new diagnosis of genital warts, this is precisely what we mean.
With regard to the conduct of the study, this was performed with the assistance of the current data manager responsible for the SSHC data base, whose help and assistance were duly acknowledged.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.