Secondary Logo

Journal Logo

More of the Same is Not Validation

Potterat, John J. BA*; Brody, Stuart PhD

Sexually Transmitted Diseases: January 2000 - Volume 27 - Issue 1 - p 60–61
Letter to the Editor

*From the El Paso County Department of Health and Environment, Division of STD/HIV Programs, Colorado Springs, Colorado; and the University of Tubingen, Tubingen, Germany

Correspondence: John J. Potterat, BA, Director, STD/HIV Programs, El Paso County Department of Health and Environment, Colorado Springs, CO 80910-3123.

Received for publication May 26, 1999, accepted August 6, 1999.

TO IMPROVE DATA ACCURACY, Griffin and colleagues advocate that the monitoring of HIV cases reported to local health departments be dynamic versus static, using multiple sources of information.1 Although they consult up to five sources to “validate” HIV risk factors, these sources are essentially records-based and rely principally on the same thing: patient self-reports. More of the same is not validation; at best, disparate yet self-similar sources can be used to gauge reliability. Researchers often confuse consistency of reports, a form of reliability, with validity.2

Griffin et al correctly propose ongoing monitoring, yet it seems that trustworthy validation stems more from quality than quantity. Several proactive tools to improve HIV risk assessment that are available to local health departments are not considered by the authors. These tools include obtaining serologic and clinical markers of infections associated with HIV3 and its two most important risk behaviors: receptive anal intercourse and use of contaminated injecting devices. For example, at least in the developed world, hepatitis B infection is highly associated with both high-risk behaviors, and hepatitis C infection is associated with the use of contaminated injecting devices. No patient infected with HIV who claims “heterosexual” transmission as a sole risk factor can be classified as such without, at a minimum, the absence of evidence of hepatitis B or hepatitis C infection. Clinical evidence of rectal infection or trauma can serve a similar purpose.2 Neither line of evidence relies on patient self-reports; patients may lie, but clinical and serological evidence does not. Such evidence is not often available in medical records of recently diagnosed or recently reported cases of HIV and AIDS. Therefore, we advocate that before HIV surveillance workers resort to “heterosexual” classification as a default category, HIV transmission-associated clinical and serological evidence be actively sought as part of HIV case follow-up procedures. We have elsewhere argued that sex bias in risk categories should be eliminated by explicitly seeking histories of receptive anal intercourse2,4-something implicitly covered for men in the “men who have sex with men” category, but which defaults to “heterosexual” transmission for women, a rubric often misunderstood to refer to vaginal intercourse only.

The authors fail to explain what “DIS interview records” are; we assume that they refer to HIV counseling rather than contact interviews. Interviewing patients infected with HIV for information about their HIV-exposed sex partners enhances risk factor validity, because it permits health-department workers to locate and interview these partners for details about the index patient when he or she is named by one of their partners (cluster interviewing). Disease intervention specialists (DIS) know that cluster information is often more accurate than information provided by the index patient. Finally, comparing the geographic and sociodemographic attributes of patients infected with HIV who report questionable risk factor information (e.g., “heterosexual” and “no identified risk”) with patients possessing similar characteristics who report classic risk factors may increase the validity of the epidemiologic picture.5 Because such evidence is inferential, risk factor attribution can be classified as “suspected” or “probable” versus “confirmed.” Subsequent information may falsify the epidemiologically suspected risk for specific individuals, without damaging the population-level picture.

It has long been an axiom of contact interviewing that interviewing success is made or broken by the willingness of the interviewer to accept the patient's story. Superficially challenged patient reports produce questionable data that usually reflect the patient's desire to convey socially desirable impressions or fear to discuss pejorative practices or partner choices. Therefore, HIV-infected patients with dubious risk histories should be interviewed in person by skilled DIS who are willing to persistently probe for accuracy, shy of offending the patient. This ancient DIS skill-common during the “Golden Age” of syphilis interviewing-has been discouraged since the beginning of the AIDS epidemic6 for fear of alienating persons whose risk behaviors place them at the margins of society, yet whom are often at the center of HIV transmission. This unfounded fear has placed more emphasis on civility than on scientific accuracy among HIV public health workers. We think that society's right to know the true magnitude and direction of the epidemic outweighs the small risk of possibly offending patients. Given that the public health establishment's first responsibility is to protect the uninfected, it is our duty to use all available tools- providing that they be ethical-to obtain valid epidemiologic information; these tools are available at local health departments. In short, we agree with the authors' article title more than with the adequacy of the essentially self-similar, records-based process they present as “validation.” There is no substitute for the personal interview with the patient and with their partner(s), when possible, or for obtaining biomarkers of HIV-associated risk behaviors.

Back to Top | Article Outline


1. Griffin RG, Wilkinson TH, Hoff GL. HIV surveillance: a dynamic, not static, process to assure accurate local data. Sex Transm Dis 1999; 26: 291–295.
2. Brody S. Sex at Risk: Lifetime Number of Partners, Frequency of Intercourse, and the Low AIDS Risk of Vaginal Intercourse. New Brunswick: Transaction Publishers, 1997.
3. Potterat JJ, Muth JB, Markewich GS. Serological markers as indicators of sexual orientation in AIDS-virus infected men. JAMA 1986; 256:712.
4. Potterat JJ. HIV infection in rural Florida women. N Engl J Med 1993; 328:1351–1352.
5. Fulton R, Kennedy RE. Are bisexuals under-recognized among white men with AIDS in the United States (Abstact TU.D.107)? In: Program and abstracts of the VII International Conference on AIDS. June 1991; Florence, Italy.
6. Potterat JJ, Muth SQ, Muth JB. ‘Partner notification’ early in the AIDS era: misconstruing contact tracers as bedroom police. In Margolis, E, ed. AIDS Research/AIDS Policy: Competing Paradigms of Science and Public Policy. Greewich, CT: JAI Press, 1998; 6:1–15.
© Copyright 2000 American Sexually Transmitted Diseases Association