36. Letter From Michael Peterson of the Office of the Assistant Secretary of Defense for Health Affairs to the Director of Research and Development, Department of the Navy (Melaragno)1
Thank you for the opportunity to review the proposed article, Epidemiology of HIV Infection Among Prostitutes in the Philippines,2 by Hayes, et al.
The authors used a case-control methodology to compare HIV infected prostitutes to prostitutes not infected with HIV. The authors need to clarify whether there were any differences between cases and controls with respect to age. If, for example, the HIV infected prostitutes were older than the non-infected prostitutes, this age difference could possibly account for increased exposure potential to HIV among the older prostitutes. The cases, if older, may have had more sexual partners than the controls prior to becoming prostitutes.
The overall prevalence rate needs to be better defined (p. 6). The denominator should be the number of prostitutes tested during a specified period and not the number of tests. Another helpful piece of information would be the inclusion of the approximate percent of prostitutes tested by region. Were various regions under or over sampled? If Regions II, IV, V, VIII, IX, X, XI, and XII were undersampled, this might account for the negative findings.3
The authors report that 93% of the infected prostitutes were working in areas adjacent to military bases (p. 7). This is a comparison of numerator data only and is misleading. The Region VI rate is the same as that adjacent to the military bases, but the authors offer no explanation for this.
A brief explanation of the matching ratio would be helpful (61 controls: 34 cases—why not 1:1 or 2:1?).
Some comments on the reliability of the subjective data are also necessary. For example, are there other data to support the reports by the prostitutes that there is no IV drug use among them? If no data [Page 108] are available, some comment needs to be made with reference to IV drug use in the Philippines. No reader will be naive enough to believe Philippine prostitutes do not or never did use IV drugs without some supporting data or frame of reference.
The authors comment that blood transfusion experience was similar among cases and controls. Is blood screened for the presence of antibody to HIV? Were there differences between cases and controls for indications for transfusion (e.g. were more illegal abortions done in cases than controls?)?
The conclusions in the discussion section are not totally supported by the data presented by the authors. Without better data on IV drug use, the assertion that transmission is by heterosexual intercourse is tenuous. Without better information on the age of the prostitutes and mobility of the prostitutes (i.e. are prostitutes brought to areas around military bases from other geographic regions?, could they have been infected elsewhere?), allusions to U.S. servicemen as a primary source of infection are not well founded.
The paper presents data that need to be published. The conclusions should be based on the data. As the paper now reads, that is not the case. I would suggest some changes as I have outlined to insure a fair representation of the data.4
Sincerely,
Lt. Colonel, USAF, BSC
Senior Policy Analyst
Preventive Medicine and Health Promotion
- Source: Washington National Records Center, OSD Files: FRC 330–91–0088, Box 4, Loose Document. No classification marking. Drafted by Peterson on April 11.↩
- A March 31 draft of the manuscript is ibid. The role that U.S. military bases in the Philippines played in AIDS transmission is discussed in Document 32.↩
- A map of the different regions is in Washington National Records Center, OSD Files: FRC 330–91–0088, Box 4, Loose Document.↩
- The article was printed in the Journal of Acquired Immune Deficiency Syndromes, 1990, vol. 3: pp. 913–920, and included the assertion that “HIV was introduced by the heterosexual route,” but the article omitted discussion about U.S. military bases.↩