Jacobs, Heather S. MPH*; Samuel, Michael C. DrPH†; Chow, Joan M. DrPH†; Bolan, Gail MD†
Gonorrhea (GC) and chlamydia (CT) are the most commonly reported communicable diseases in the United States.1 As with many other sexually transmitted diseases (STD), GC, and CT case rates are highly associated with young age.2 Adolescents and young adults, particularly females, have an increased risk of STD infection, due to biologic factors, including cervical ectopy,3 and behavioral factors, such as frequent changes in sex partners,4 as compared to factors in other periods in life. In order to assess differences in GC and CT rates by demographic factors, we maintain surveillance case report data systems with information on age, gender, and race/ethnicity.
At the national level, the Centers for Disease Control and Prevention reports gender- and age-specific rates for STD data in the following 5-year categories: 10 to 14, 15 to 19, 20 to 24, . . ., 60 to 64, and 65 and above.5 This level of aggregate, age-specific reporting presents 2 challenges. First, when aggregate age categories are used to report case data, finer differences in risk by smaller age categories may be masked. Second, lack of standard age categories used by some state or local agencies for the youngest and oldest ages can create difficulties in comparing GC and CT case rates across regions. For example, Hawaii's youngest category includes ages 0 to 14 years,6 while South Carolina reports case rates for 0 to 4 years, 5 to 9 years, and 10 to 14 years.7
Age-specific categories may also be different depending on whether the rates are used for tabular or graphical displays. In addition, lack of standard age categories used for other conditions makes it difficult to assess comorbidities within a given population. For example, the National Vital Statistics System uses the following age categories, in years, for pregnancy data: under 15, 15, 16, 17, 18, 19, 20 to 24, 25 to 29, 30 to 34, 35 to 39, and 40 to 44.8 Thus, comparisons of pregnancy and STD rates in adolescents are somewhat limited. The issue is further complicated by the lack of standardization of race/ethnicity data across reportable diseases (e.g., STD vs. HIV).
Although using aggregate age categories in the presentation of age-stratified GC and CT case rates is common, little has been published about using single-year age rates. Limited reports do address how STD prevalence and incidence vary significantly by age and how these differences may be influenced by the way age categories are constructed.9–14 Beyond STD surveillance, other agencies, such as the National Cancer Institute Surveillance, Epidemiology, and End Results Program, report case rates by single age, rather than by aggregate categories.15 From the epidemiologic perspective of assessing associations, deconstructing age categories may be important because the aggregation of a continuous age variable into a categorical variable can incompletely control for age-related confounding.14 The objective of this analysis is to more carefully examine age patterns among STD cases in California, with a focus on comparing single-age rates to 5-year categorical age rates by gender, assessing mean age at diagnosis by race/ethnicity, and examining trends in GC and CT rates over time.
In California, GC and CT cases are reported to local health jurisdictions by both medical providers and laboratories; reports are then forwarded for state and national surveillance.2
This analysis was based on California GC and CT case reports with a date of diagnosis from January 1, 1998 through December 31, 2007, and age-, gender-, and race/ethnicity-specific population estimates from the California Department of Finance 2000 to 2050 Projections.16 Age was calculated based on date of birth and date of diagnosis. Subgroups with small numbers of cases were excluded from analyses including cases with missing gender or age at diagnosis, transgender cases, cases with a reported race/ethnicity of Native American/Alaska Native (NA/AN), and cases with “other” or unknown race/ethnicity. For the calculation of rates, cases older than 100 years at diagnosis were excluded because population estimates for Californians are available only for ages 0 to 100 years. When we calculated mean age at diagnosis, cases aged 90 to 99 years were coded as “99” since we assumed that all cases with this age were actually missing. Cases with missing race/ethnicity were excluded from the calculation of race-specific rates but were not excluded from the calculation of single-age rates.
Separate analyses were conducted for GC and CT cases by gender. For each gender, age-specific rates were calculated by dividing the number of GC or CT cases within a single age at diagnosis, and within 5-year age groups by the total number of Californians of that age/age group from the California Department of Finance projections multiplied by 100,000 population. Detailed analyses of rates focused on 3 main age groups: under 10 years, 10 to 65 years, and over 65 years. Single-age rates were compared with 5-year categorical rates graphically. In a subanalysis, the single age denominators for adolescent rates (ages 15–18 years) were adjusted for percent sexually active by high school grade and gender, based on data from the 2007 Youth Risk Behavior Survey, to assess the impact of the smaller denominator on single age rates.17
Case rates were also stratified by region (northern California, San Francisco Bay Area, central California, Los Angeles, and southern California), but no differences were found and those results are not reported. Mean ages and respective 95% confidence intervals were calculated, and standard t tests were conducted to compare age by racial/ethnic group, with whites as the reference group. While this analysis focused primarily on 2007 cases and rates, trends in mean age over time from 1998 through 2006 were also assessed, with the use of standard linear regression. All data analyses were conducted with the use of SAS, version 9.1 (Cary, NC).
Characteristics of GC and CT Cases
In 2007, 31,177 cases of GC were reported in California: 14,480 female, 16,572 male, 0 transgender, and 125 with unknown gender. Three cases whose age exceeded 100 years, and 126 cases with unknown age were excluded from analysis. Race/ethnicity was unknown for 32.9% of GC cases. Of cases with race/ethnicity information, 40.1% were black, 29.9% were Hispanic, 24.8% were white, 3.6% were Asian/Pacific Islander (API), 1.1% were “other” race/ethnicity, and 0.5% were NA/AN. Because of small numbers, NA/AN were not included in further analyses.
In 2007, 142,243 cases of CT were reported in California: 101,401 female, 40,357 male, 2 transgender, and 483 with unknown gender. Excluded from analysis were 2 cases whose age exceeded 100 years, 623 cases with unknown age, and the 2 transgender cases. Race/ethnicity was unknown for 33.5% of CT cases. Of cases who had race/ethnicity information, 49.2% were Hispanic, 22.9% were black, 20.4% were white, 5.8% were API, 1.2% were other race/ethnicity, and 0.5% were NA/AN.
Figure 1A shows the distributions of male and female 2007 GC case rates by single age from 10 to 65 years; Figure 1B shows these distributions for CT. For both males and females, CT rates were substantially higher than GC rates at all ages. Among females, the shape of the distribution of rates across ages was very similar for CT and GC. Both CT and GC rates for females peaked at age 19 (3640 per 100,000 for CT and 497 per 100,000 for GC) and then quickly declined with increasing age. GC rates for males peaked at age 21 (328 per 100,000), while CT rates peaked at age 22 (994 per 100,000). Among males, GC rates remained relatively high in the older age groups.
Figures 2A, B compare the distributions of GC and CT 2007 rates for females aged 10 to 65 years, by both single-age and 5-year categories. For both GC and CT, the highest categorical rates were observed in the 20- to 24-year-old age group. However, the highest single-age rates were not in this category, but in the next younger category, 15 to 19 years old. There was high variability in the single-age rates between ages 15 and 19 (inclusive), including the 2 highest single-age rates for GC, at ages 18 and 19, and the highest and third-highest age rates for CT, also at ages 18 and 19, respectively. To a similar but lesser extent, this masking effect of aggregate rates is seen in the 10- to 14-year-old group, with 14-year olds displaying much higher rates than those for the other single ages in this category. Also, in the 25- to 29-year-old group, the CT rate for 25-year-old women is nearly 2 times higher than that for 29-year-old women.
In the subanalysis for adolescent rates calculated with the use of denominators that were adjusted for percent sexually active, rates were still observed to increase sharply from ages 15 to 18 years, with the adjusted single-age rates being higher than the unadjusted rates, as expected, but with the same general trend.
Older Adult Cases and Child Cases
A similar masking effect was also seen among adults aged 45 years and older. While these cases are collapsed into an aggregate age category, the single-age rates for individuals aged 45 to 55 years are significantly higher than those for individuals aged 56 years and older, and STD rates for persons in their 80s were extremely low, compared to those for adolescents and young adults.
Only 25 GC cases and 58 CT cases aged 0 to 9 years were reported in California in 2007. For both GC and CT, most of cases aged 0 to 9 years (12, or 48%, and 34, or 59%, respectively) were younger than 1 year of age at diagnosis and therefore most likely acquired their infection at birth. Between ages 1 year and 9 years, few cases were diagnosed during any single year of age. The numbers of male and female cases diagnosed during the first year of life were nearly equal, but slightly more female cases were diagnosed later in childhood.
Age Distribution of GC and CT, by Race/Ethnicity
When analyzed by gender and diagnosis, blacks had younger mean age at diagnosis than did any other race/ethnicity group. Compared to that for whites, this younger age at diagnosis for blacks was statistically significant for all comparisons (P < 0.0001; Table 1). In contrast, APIs tended toward older mean age at diagnosis, significantly so for female CT cases and male GC cases (P < 0.0001). White male GC cases had a mean age at diagnosis that was at least 3 years older than that for males of other races/ethnicities.
Secular Trends for GC and CT, by Age
Figure 3 shows that, from 1998 through 2007, there was a strong and mostly steady increase in the mean age at GC and CT diagnosis for both male and female reported cases; the mean age at diagnosis was older for both male and female GC cases, compared to CT cases. The increase in mean age over time was greatest (average, 1.95-month increase per year) for male CT cases (P < 0.001); the increase was also significant for female CT cases (0.55-month increase per year, P < 0.001) and for female GC cases (1.28-month increase per year, P < 0.001). These increases in mean age over time were seen in all race/ethnicity groups (data not shown).
In order to assess how 5-year age categories fit single-age CT rates over time, a separate analysis was conducted for the females with the highest age-specific rates (e.g., ages, 15–24 years). All single-age rates in 15- to 24-year-olds increased over time; the peak remained at age 19 years, with the second-highest rate at age 20 years, and the third-highest rate at age 18 years. In each year from 2001 through 2007, the 5-year age-specific rates of 15 to 19 years and 20 to 24 years masked the differences in single age at diagnosis, with observed increases in the masking effect over time (data not shown).
This analysis shows that the standard practice of using aggregate age categories for the analysis and dissemination of STD data can obscure important points, which are revealed when the data are examined more carefully by single-age groups. For instance, among GC and CT cases, rates for 18- and 19-year olds were considerably higher than those for 15-year-olds, thus masking the true range of risk of infection within this narrow age group. Additionally, the 2 highest single-age GC rates were found in the 15- to 19-year category, although the highest aggregate rate was in the 20- to 24-year age group. Analysis by single age may improve interventions for those at highest risk. Using single-age rates by gender and race/ethnicity for target age groups could focus interventions and policies related to STD control, and could help educate public health practitioners about the precise groups with greatest inequities, leading to the development of earlier and more targeted prevention and treatment efforts in settings serving older adolescents and young adults, e.g., high school General Educational Development programs and junior colleges.
This analysis also found significant differences by race/ethnicity through comparisons of mean ages at diagnosis and single-age-specific and race/ethnicity-specific case rates. In 2007, black males and females had a younger age at diagnosis than did cases of other race/ethnicity groups; this finding was consistent with data from 1998 through 2006. Our findings also emphasize the importance of looking at race/ethnicity-specific rates, rather than crude rates, because of the highly differential rates among race/ethnicity groups in all ages, particularly those ages with the highest rates, the teens and early 20s.
Last, this analysis found evidence from 1998 through 2007 of a significant increase in the mean age at diagnosis for both male and female CT cases and for female GC cases. This increase was also seen for all race/ethnicity groups. These results could be due to various factors including increased screening of a relatively older age group (e.g., age 20–24 years, as compared to age 15–19 years), cohort effect of aging of young high risk individuals, and possibly increased high-risk behavior among older individuals.
However, it is not always necessary or appropriate to disseminate data by single-age groups, and too much tabular or graphical presentation of single-age data would overly burden reports or other presentations of data. The appropriateness of analyzing STD data by single age groups should be determined based on the numbers of cases available for analysis, applicability of results to program interventions and planning, and availability of resources. With small numbers of cases, interpretation of data based on single age groups may be unreliable. For exploratory analysis, small jurisdictions could consider aggregating single age rates over multiple years to visualize important patterns in single age groups that might be masked by the typical age groupings. The use of aggregate groups is often reasonable in general, and particularly for the extremes of the age distribution, where rates are generally low. Some public health agencies have already begun to use age data aggregated into nonstandard categories as a result of examining single-age rates in finer detail. For example, the California Adolescent Sexual Health Work Group presents STD rates using the age categories 15 to 17 years, 18 to 19 years, and 20 to 24 years, recognizing that 15-year-olds are not developmentally at the same stage or the same level of risk as are 19-year-olds or 24-year-olds.18 Additionally, HIV/AIDS data are analyzed and presented by single age categories19,20; this type of analysis can inform STD control strategies for GC and CT as well.
There were limitations to this analysis. This analysis was based on data collected for surveillance purposes to monitor trends in GC and CT infection. However, these case rates may not reflect the true incidence of disease in the population because the analysis was based on reported cases, as well as those who sought health care and were tested. Second, calculation of rates for all analyses shown was based on total age-specific population denominators and was not restricted to those sexually active. However, in a subanalysis, when these denominators were adjusted for the proportion sexually active based on national behavioral data, similar trends were observed. It should be noted that these behavioral data were not California-specific and may not be representative; further investigation of this limitation including the use of race-specific behavioral data may be warranted. Third, while one-third of cases were excluded due to missing race/ethnicity data, based on a prior analysis by Mark and Gunn using provider interview to validate GC case information, the racial distribution of the cases with missing race/ethnicity data are similar to the distribution of cases with complete race/ethnicity data.21 Thus, conclusions based on cases with known race/ethnicity can be generalized to cases without this information. Lastly, while higher rates of repeat CT/GC infection among young females may explain some of the difference in age-specific rates, we did not adjust for repeat cases in the calculation of age-specific rates. It is more likely that most of the difference in age-specific rates is because of changes in the level of sexual risk behaviors over time.
This analysis of the detailed changes in STD risk over the lifespan calls for a greater awareness of the important differences that can be obscured by aggregation of data by age, particularly for adolescents and young adults, which were noted across race/ethnicity groups. Current STD screening guidelines recommend GC and CT testing in females up to age 24; these guidelines are supported by the high STD 5-year aggregate age-specific rates for adolescents and young adult females.22 Using single-age rates could assist in the development of more effective guidelines. This analysis indicated that the mean age at STD diagnosis is increasing, but that rates among adolescent and young adult Californians remain very high, emphasizing the great need for further prevention, testing, and successful treatment of young people. Youth of color are particularly at risk during this period of life, so the success in STD control relies on the ability to address disparities by race/ethnicity, as well as on targeting prevention toward age-specific groups with the highest rates and the least awareness and access to care. With this detailed age information, we will have the ability to better prevent the morbidity associated with early STD acquisition and to reduce overall case rates.
1. Centers for Disease Control and Prevention. Sexually Transmitted Disease Surveillance, 2006. Atlanta, GA: Department of Health and Human Services (US); 2007.
2. Sexually Transmitted Diseases in California, 2006. California Department of Public Health, STD Control Branch, November 2007.
3. Jacobson DL, Peralta L, Graham NM, et al. Histologic development of cervical ectopy: Relationship to reproductive hormones. Sex Transm Dis 2000; 27:252–258.
4. Kraut-Becher JR, Aral SO. Gap length: An important factor in sexually transmitted disease transmission. Sex Transm Dis 2003; 30:221–225.
5. Centers for Disease Control and Prevention. Chlamydia-Reported Cases and Rates per 100,000 Population by Age Group and Sex: United States, 2002–2006. Atlanta, GA: Centers for Disease Control and Prevention; 2007.
6. Hawaii State Department of Health. STD/AIDS Prevention Branch Epidemiology: Data and Statistics. Honolulu, HI: Hawaii State Department of Health; 2008.
7. South Carolina Department of Health and Environmental Control. 2007 STD/HIV Division Surveillance Report. Columbia, SC: South Carolina Department of Health and Environmental Control; 2008.
9. Bodley-Tickell AT, Olowokure B, Bhaduri S, et al. Trends in sexually transmitted infections (other than HIV) in older people: Analysis of data from an enhanced surveillance system. Sex Transm Infect 2008; 84:312–317.
10. Boscoe FP. Subdividing the age group of 85 years and older to improve US disease reporting. Am J Public Health 2008; 98:1167–1170.
11. Kaestle CE, Halpern CT, Miller WC, et al. Young age at first sexual intercourse and sexually transmitted infections in adolescents and young adults. Am J Epidemiol 2005; 161:774–780.
12. Burstein GR, Gaydos CA, Diener-West M, et al. Incident Chlamydia trachomatis infections among inner-city adolescent females. JAMA 1998; 280:521–526.
13. Robinson AJ, Watkeys JE, Ridgway GL. Sexually transmitted organisms in sexually abused children. Arch Dis Child 1998; 79:356–358.
14. Reijneveld SA. Age in epidemiological analysis. J Epidemiol Community Health 2003; 57:397.
16. State of California Department of Finance. Race/Ethnic Population With Age and Sex Detail, 2000–2050. Sacramento, CA: State of California Department of Finance; 2007.
17. Centers for Disease Control and Prevention. Youth risk behavior surveillance-United States, 2007. MMWR 2008; 57:table 61.
19. Foster G, Levine C, Williamson J. A Generation at Risk: The Global Risk of HIV/AIDS on Vulnerable Children. Cambridge, MA: Cambridge University Press; 2005:147–149.
20. Midwest AIDS Training and Education Center. HIV/AIDS Case-Finding in Family Planning Clinics. Health Care Education and Training Inc.; 2002.
21. Mark KE, Gunn RA. Gonorrhea surveillance: Estimating epidemiologic and clinical characteristics of reported cases using a sample survey methodology. Sex Transm Dis 2004; 31:215–220.
22. US Preventive Services Task Force. Screening for chlamydial infection: US Preventive Services Task Force recommendation statement. Ann Intern Med 2007; 147:128–133.