ACCURATE AND TIMELY surveillance data are critical for developing sound sexually transmitted diseases (STDs) prevention strategies and have been the rationale for the communicable disease reporting system. Besides the disease-specific morbidity data, there has also been substantial recent interest in developing systematic data on behavioral risks, such as the Centers for Disease Control and Prevention (CDC)-funded state-based Behavioral Risk Factor Surveys. Behavioral risk data have become increasingly important in evaluating prevention programs for which the clinical incident outcome is a relatively rare event such as human immunodeficiency virus infection prevention programs. Data on sexual behavior and drug use are, by necessity, almost always obtained through self-report, and the validity and reliability of these data are therefore difficult to estimate. Biologic indicators, or “bio-markers,” such as incident STDs, may be useful in validating self-reported sexual behaviors, such as condom use in populations where STD incidence is high. However, even in high-incident populations, where using incident STD as an outcome measure is epidemiologically feasible, measuring incident STD has, until recently, required uncomfortable physical examinations, was expensive, and requires access to clinical laboratory facilities.
In the accompanying article, Ku and colleagues1 have comprehensively reviewed the issues of disease and behavioral surveillance and, based on their operational field research, offer exciting insights into integration of these disciplines. Most studies of sexual behavior, including the majority of articles on the subject published in this Journal, are limited by their generalizability. For example, can behavioral risk and intervention findings from STD clinic-based studies be extrapolated to the general population? By focusing their review on the on the large population-based national surveys, Ku and colleagues control the impact of this bias. Although most of the studies described in their review are cross-sectional and do not have biologic validators, there appear to be consistent, broad-based changes, such as increase in condom use, especially in adolescents. These findings are encouraging and suggest that the behavioral norms in the high-risk groups are changing-albeit slowly.
From a clinical standpoint, Ku et al describe the possibilities of integrating well-designed behavior studies with population-based STD surveillance activities. The bias of traditionally collected STD surveillance data is well recognized. For example, using gonorrhea as a current and historical example, the data illustrate several problems, all of which have been recognized for years. Previous authors have estimated that the degree of underreporting for gonorrhea is between 30% to 90%.2,3 In 1995, the rate in blacks was 40 times higher than for whites (1,087/100,000 compared with 29/100,000 for the unadjusted crude rate).4 Nearly 60% of all disease is reported in public STD clinics, which service a discrete component of the population. This is lower than in previous years, a trend that may be caused by decreasing disease morbidity-or decreased clinical resources and hours at the public clinics. Private sector underreporting has been well documented and probably underlies the disparate male-female ratio seen in whites. It is well recognized that in the STD context, race may be a surrogate for socioeconomic status and/or health care access and that this rate differential may be overrepresented in the national surveillance data.
For chlamydia, there are additional problems in ascertaining disease trends. Testing was not available until 1985, and reporting was not required until recently. Many patients are treated syndromically. In addition, because testing activity is focused on screening women, the number of reported cases in women has been increasing exponentially and is six times the number of reported cases in men. These trends are widely recognized as increased ascertainment rather than increased burden of disease. Traditional chlamydia surveillance data are therefore similarly limited in their generalizability and in the ability to use these data to assess intervention outcomes on a broad population basis. For some diseases, population-based cross-sectional data are available. For example, using the National Health and Nutrition Examination Survey data, CDC investigators elegantly defined the epidemiology of herpes,5 but these results required phlebotomy and were reported more than 10 years after the survey was completed and were therefore not useful for intervention evaluation.
For STD epidemiologists and behavioral scientists, the development of new noninvasive ligase chain reaction and polymerase chain reaction DNA amplification tests for gonorrhea and chlamydia, which can be performed on urine, offers tremendous promise in addressing these survey bias issues. Using urine as the test specimen eliminates the need for urethral swabs in men and pelvic examinations in women. Specimens collected do not need any special handling or storage condition and therefore could be collected in field settings. The possibility of integrating large-scale, population-based behavioral studies with disease ascertainment has thus become a reality. Ascertainment of true population-based gonorrhea and chlamydia is now also possible.
In the pilot study for the National Survey of Adolescent Males, Ku et al performed important operational research in ascertaining whether the concept of field-based STD ascertainment can be translated into reality. The high compliance rates with urine collection (>80%) are an exciting finding. These findings are similar to those recently reported from the Rakai study in rural Uganda, in which 94% of 9,819 women provided urine for ligase chain reaction, and 95% provided self-administered vaginal swabs.6 These findings, especially the success of the operational pilots, suggest that application of the new diagnostic techniques to the field setting will be a useful and practical adjunct. This opens the possibility of developing new paradigms for STD surveillance and intervention evaluation activities.
1. Ku L, Sonenstein FL, Turner CF, Aral SO, Black CM. The promise of integrated representative surveys about STDs and behavior. Sex Transm Dis 1997;24:298–308.
2. Anderson JE, McCormick L, Fichtner R. Factors associated with self-reported STDs: Data from a national survey. Sex Trans Dis 1994;21:303–308.
3. Hinds MW, Gale JL. Male urethritis in King County, Washington, 1974–1975: II. Diagnosis and treatment. Am J Public Health 1978;68:26–30.
4. Centers for Disease Control. 1995 STD Statistics. Division of Sexually Transmitted Diseases/HIV Prevention, Atlanta, 1996.
5. Johnson R, Nahmias A, Magder L, Lee FK, Brooks CA, Snowden CB. A seroepidemiologic survey of the prevalence of herpes simplex type 2 infection in the United States. N Engl J Med 1989;321:7–12
6. Wabwire-Mangen F, McNaim D, Wawer M, et al. Feasibility of comprehensive, community-based sample collection to assess STD prevalence. IX International Conference on AIDS, Vancouver, Canada. Abstract TuΛ097.