Increasing Chlamydia Positivity in Women Screened in Family
Planning Clinics: Do We Know Why?
DAVID FINE, PHD,* LINDA DICKER, PHD,† DEBRA MOSURE, PHD,† STUART BERMAN, MD,† AND REGION X
INFERTILITY PREVENTION PROJECT
increased 46% from 1997 through 2004 among young sexually active
women screened in Region X family planning clinics. The objective of
this analysis was to systematically examine the influences of risk
factors, changing laboratory test methods, and interclinic variability
on chlamydia positivity during this period.
We analyzed data from 520,512 chlamydia tests
from women aged 15 to 24 years screened in 125 family planning
clinics. Multivariate logistic regression modeling was used to adjust the
annual risk of chlamydia for the demographic, clinical, and sexual risk
behavior characteristics associated with infection and for the increas-
ing use of more sensitive laboratory test methods. A generalized linear
mixed model was used to adjust for interclinic variability.
Results: We found a significant 5% annual increase in the risk of
chlamydia even after adjusting for risk factors including laboratory
test characteristics (odds ratio 1.05, 95% confidence interval: 1.04,
1.06). Variability among the clinics where screening occurred did not
account for the increase.
Conclusions: Based on a review of all available data, we concluded
that there was a true increase in chlamydia positivity over the 8-year
Following a 9-year 60% decline, chlamydia positivity
CHLAMYDIA TRACHOMATIS IS THE MOST common bacterial
sexually transmitted infection (STI) in the United States, with an
estimated 2.8 million new cases occurring each year.1Chlamydial
infections are often asymptomatic, can persist for a prolonged
period, and are an important preventable cause of reproductive
sequelae in women, including pelvic inflammatory disease (PID),
ectopic pregnancy, and infertility.2Screening for chlamydia has
been shown to reduce the incidence of PID.3
In 1988, the first widespread screening and treatment program
for chlamydia began in US Public Health Service Region X
(Alaska, Idaho, Oregon, and Washington). The focus of the Region
X Infertility Prevention Project is to screen all young sexually
active women seen in the region’s Title X family planning clinics.4
During the first 9 years of the program, chlamydia positivity
among women aged 15 to 24 years declined over 60%, from 10.3%
in 1988 to 4.0% in 1996.5This decline corresponded with signif-
icant reductions in self-reported sexual risk behaviors.6Similar
declines in chlamydia positivity were also seen in other areas of
the United States where broad-based screening programs were
instituted during the same time period.7–9However, there was a
46% increase in chlamydia positivity, from 3.9% in 1997 to 5.7%
in 2004, among young women screened in Region X family planning
clinics.5There have been many questions and much speculation about
the reasons for the increases in positivity, including changes in labo-
ratory test technology and screening higher-risk women; however,
there have been no analyses systematically evaluating potential
causes for these increases. The 3 objectives of our analysis were to
examine: 1) demographic, clinical, and sexual behavioral risk
characteristics associated with chlamydial infection and their in-
fluences on chlamydia positivity in women aged 15 to 24 years
seen in Region X family planning clinics from 1997 through 2004;
2) the impact of changing laboratory test methods on the increases
in positivity; and 3) the effect of interclinic variability on chla-
mydia positivity using a generalized linear mixed model.
Materials and Methods
We analyzed data from 520,512 chlamydia tests from women aged
15 to 24 years screened in 125 family planning clinics participating in
the Region X Infertility Prevention Project from 1997 through 2004
(average 65,000 tests/yr). Women aged 24 years and younger were
routinely screened for chlamydial infection at least annually as rec-
ommended by the Centers for Disease Control and Prevention and the
US Preventive Service Task Force.10,11
All Region X family planning clinics used a common medical
record form. Information collected included age; race; ethnicity;
specimen collection date; clinical findings (ectopy, friable cervix,
PID, cervicitis); self-reported sexual risk behaviors (having had a
new sex partner in the past 60 days, having had multiple sex
partners in the past 60 days, having had a symptomatic sex partner
in the past 60 days, having had a sex partner who was diagnosed
with chlamydia, and condom use during last sex); having had
chlamydia in the past year; laboratory test type; and chlamydia test
result. We included clinics that routinely screened for chlamydia
during the entire 8-year period and had performed 50 or more
chlamydia tests in at least 7 of the 8 years (about 90% of all
chlamydia tests performed). We used data from the US Census to
Supported by Region X IPP, HHS/OPA and CDC/DSTDP.
Use of trade names and commercial sources is for identification only
and does not imply endorsement by the Centers for Disease Control and
Correspondence: David Fine, PhD, Center for Health Training, 1809
Seventh Avenue, Suite 400, Seattle, WA 98101-1313. E-mail: dfine@jba-
Received for publication March 16, 2007, and accepted June 12, 2007.
From the *Center for Health Training, Seattle, Washington;
and †Division of Sexually Transmitted Disease Prevention, Centers
for Disease Control and Prevention, Atlanta, Georgia
Sexually Transmitted Diseases, November 2007, Vol. 34, No. 11, p.000–000
Copyright © 2007, American Sexually Transmitted Diseases Association
All rights reserved.
estimate the population density of the town or city where each
family planning clinic was located and to examine the distribution
of tests by clinic location (urban vs. rural).
Four state public health laboratories (Alaska, Idaho, Oregon,
and Washington), a county health district laboratory (Spokane),
and the University of Washington Chlamydia Laboratory per-
formed chlamydia testing for the 125 family planning clinics.
From 1997 through 2004, these laboratories switched to newer,
more sensitive test methods, i.e., nucleic acid amplification tests
(NAATs). Over the time period, non-NAATs included (in order of
decreasing usage) enzyme immunoassays (MicroTrak II, Syva and
Behring Diagnostic Products, Cupertino, CA); nucleic acid hybrid-
ization tests (Pace 2, Gen-Probe, San Diego, CA); nucleic acid
hybridization assays (Hybrid Capture 2, Digene, Gaithersburg,
MD); and cell cultures. The majority of NAATs used in Region X
were ligase chain reaction tests (LCx, Abbott, Abbott Park, IL) and
target capture transcription-mediated amplification assays (Aptima
Combo 2, Gen-Probe, San Diego, CA).
Observed chlamydia positivity was calculated by dividing the
number of positive tests by the total number of tests that were
either positive or negative. We excluded tests with unsatisfactory
or indeterminate results. For our first objective, among the vari-
ables collected, potential predictors of chlamydial infection were
identified by odds ratios (OR) and 95% confidence intervals that
were significant at P value ?0.05. Stepwise logistic regression
modeling was used to assess the risk of chlamydial infection for
these predictors including demographic characteristics, clinical
findings, self-reported sexual risk behaviors, type of laboratory test
(NAAT vs. non-NAAT), and year of test. A P value ?0.01 was
used for determining statistical significance. To assess the trend in
risk of chlamydia by laboratory test type, we included an interac-
tion term, year-laboratory test type, in the multivariate model.
Our second objective focused on changes in laboratory test
methods and estimating the true disease prevalence. A switch to a
more sensitive laboratory test method can result in an increase in
observed chlamydia positivity even with no increase in true dis-
ease prevalence.12We used the sensitivity and specificity of each
laboratory test method to calculate an adjusted positivity [test-
specific adjusted positivity ? (test-specific observed positivity ?
test specificity ?1)/(test sensitivity ? test specificity ?1)].13We
calculated the overall adjusted positivity as a weighted sum of the
adjusted positivities for each test type. The sensitivity and speci-
ficity estimates used for the adjustment were culture (sensitivity
0.747, specificity 1.000); enzyme immunoassay tests with negative
gray zone confirmation (sensitivity 0.810, specificity 0.996); other
non-NAATs including Gen-Probe Pace 2 and Digene Hybrid Cap-
ture 2 assay (sensitivity 0.619, specificity 0.997); and, NAATs
(sensitivity 0.855, specificity 0.997).14,15We compared the differ-
ences in trends for observed and adjusted chlamydia positivity.
Our third objective was to account for the variability in chla-
mydia positivity across the 125 family planning clinics. We used
a generalized linear mixed model to examine the probability of a
positive chlamydia test controlling for the random selection of the
family planning clinics and the fixed effects of each of the iden-
tified risk factors. Specifically, the model assumed that total num-
bers of positive chlamydia tests at each level of a risk factor were
independent binomial random variables, and the levels of the risk
factor were linearly related to the logit of the probabilities of a
positive chlamydia test. The model assigned all observations from
the same clinic the same adjustment to the intercept term, and these
adjustments varied randomly from clinic to clinic. For the model-
ing we used the GLIMMIX procedure in Version 9.1 of SAS.16
Who Was Screened for Chlamydia?
Of the 520,512 chlamydia tests performed from 1997 through
2004, 52% occurred in the state of Washington (Table 1). Twenty-
two percent of the chlamydia tests were performed in women aged
15 to 17 years, 26% in women aged 18 to 19 years, and 52% in
women aged 20 to 24 years. The majority of the tests occurred
among non-Hispanic white women. Twenty-two percent of the
tests were done in clinics located in small towns with populations
?25,000; 44% were done in clinics located in cities with 25,000 to
under 100,000 residents, and the remaining 34% were done in
clinics located in cities with populations of 100,000 or greater.
Twenty-six percent of tests were in women who reported having
1 or more sexual behavior risks in the past 60 days. Ninety-three
percent of the tests were in women who had no clinical findings for
chlamydial infection on physical examination. Three percent of the
tests were among women who reported having had a previous
chlamydial infection in the past year.
Did the Demographic, Clinical, or Sexual Behavioral Risk
Characteristics of the Women Screened Change Over Time?
The percent of women aged ?18 years decreased from 24% to
18% from 1997 through 2004; the percent of women aged 18 to 19
years remained stable, 25%; and the percent of women aged 20 to
24 years increased from 51% to 56% (Fig. 1). The race/ethnicity
distribution remained stable over time. Although chlamydia
screening volume increased 38% from 1997 through 2004, the
distribution of tests performed across clinics located in different
population densities, i.e., urban versus rural, did not change over
time (data not shown). The proportion of women reporting sexual
risk behaviors or who had a clinical finding on physical examina-
tion remained stable or decreased over the time period (Fig. 2).
The proportion of women reporting having had a previous chla-
mydial infection in the past year increased from 2.6% in 1997 to
3.7% in 2004.
Did the Change in Laboratory Test Methods Account for the
Increase in Chlamydia Positivity?
There was an increase in the use of NAATs from 13% in 1997
to 60% in 2004 (Fig. 3). The observed chlamydia positivity in-
creased to 46%. After adjusting chlamydia positivity to account for
the use of more sensitive laboratory test methods and to better
estimate true chlamydia prevalence, there continued to be an
increase in positivity (65%) over the time period. This increase in
adjusted positivity persisted when a range of laboratory test sen-
sitivities and specificities for each laboratory test method were
used (data not shown).
Did Changes in Risk Factors Account for the Increase in
Chlamydia positivity was associated with young age, nonwhite
race, clinic city size ?100,000, having 1 or more self-reported
sexual behavioral risks, having had a sex partner with chlamydia,
not using a condom at last sex, having 1 or more clinical findings,
having had a positive chlamydia test in the past year, use of
NAATs, and year (Table 2, Crude OR; model 1 adjusted ORs).
Using multivariate logistic regression modeling to adjust for all the
Sexually Transmitted Diseases ● November 2007
FINE ET AL
above risk factors, there remained a significant 5% increase in the
risk of chlamydial infection each year (OR 1.05, 95% confidence
interval: 1.03, 1.05) (Table 2, model 1 adjusted ORs).
To assess the trend in risk of chlamydia by laboratory test method,
we included an interaction term, year-laboratory test method, in the
multivariate model. The year-laboratory test method interaction term
was significant (P ?0.01), indicating differences in the slopes of the
trend lines for NAATs versus non-NAATs (data not shown). How-
ever, there was an increase in the slopes of the trend lines for both
NAATs and non-NAATs. The risk of chlamydia associated with
NAATs increased 3%/yr, and the risk associated with non-NAATs
increased 5% per year; thus, there was an increase in the risk of
chlamydia over the time period regardless of the laboratory test
Did the Increases in Chlamydia Positivity Persist After
Adjusting for Clinic?
To account for the variability across the 125 family planning
clinics, we used a generalized linear mixed model to assess the
probability of a positive chlamydia test controlling for clinic as a
random effect and each risk factor as a fixed effect (Table 2, model 2).
The model would not converge when trying to include the 125 family
models did converge when separate analyses were performed that
Chlamydia Positivity in Region X Family Planning Clinics,
Characteristics of Women Aged 15–24 Yrs. and
Age group (yr)
Clinic city size, persons
New sex partner, past 60 d
More than one sex partner,
past 60 d
Symptomatic sex partner, past
One or more sexual behavioral
risks, past 60 d*
Sex partner with chlamydia
Condom use, last sex
One or more clinical findings†
Positive chlamydia test, past yr
*Includes having had a new sex partner, multiple sex partners, or a
symptomatic sex partner in the past 60 d.
†Includes cervicitis, friable cervix, ectopy, and pelvic inflammatory
Fig. 1. Trends in characteristics of women aged 15 to 24 years
screened for chlamydia, Region X family planning clinics, 1997–2004.
? race/ethnicity, white;
? 20–24 yrs.
? 15–17 yrs;
Fig. 2. Trends in self-reported sexual behavioral risks and clinical
findings among women aged 15 to 24 years screened for chla-
mydia, Region X family planning clinics, 1997–2004.
findings (includes cervicitis, friable cervix, ectopy, and PID);
? multiple sex partners, past 60 days;
partner, past 60 days;
? positive chlamydia test, past year;
? symptomatic sex partner, past 60 days.
? new sex
Vol. 34●No. 11
INCREASES IN CHLAMYDIA: DO WE KNOW WHY?
included the 125 clinics as a random effect and a single risk factor as
a fixed effect. The OR generated from these 10 generalized linear
mixed models were similar to the adjusted OR from the multivariate
logistic regression model that included all 10 of the risk factors (Table
2, model 1 and model 2). After adjusting for variability across the 125
clinics using the mixed model, there remained a significant 4%
increased risk of chlamydial infection each year (OR 1.04, 95%
confidence interval: 1.03, 1.04) (Table 2, model 2).
From 1997 through 2004 over half a million young sexually
active women, of whom over 90% were asymptomatic, were
screened for chlamydia at 125 family planning clinics participating
in the Region X Infertility Prevention Project. Over that time
period, chlamydia positivity increased 46%. We systematically
evaluated potential reasons for the increase, including changes in the
characteristics of the women screened over time, the use of more
sensitive laboratory testing methods, and the consistency of the in-
crease across the 125 family planning clinics where the screening
occurred. Based on the available data, we concluded that there was a
true increase in chlamydia positivity over the 8-year period.
The proportion of women with demographic, clinical, or sexual
behavioral risk characteristics associated with an increased risk of
chlamydial infection remained stable or decreased over time. We
found that the percent of women in the age group (15-to 19-year-
old) at highest risk of chlamydial infection decreased. Overall, the
proportion of women reporting any sexual risk behaviors also
decreased. There was little change in the proportion of women who
had an abnormal clinical finding on physical examination. The
decline in the proportion of women with these risk factors should
have resulted in a decrease, not an increase, in observed chlamydia
positivity during this 8-year period. The proportion of women who
reported a positive chlamydia test during the year before their
clinic visit increased 42%. This may have been due to increasing
efforts by some clinics to rescreen previously infected women.
However, the proportionally large increase reflects a small abso-
lute change—from 2.6% of testing in 1997 to 3.7% in 2004. This
increase could not account for the overall increase in chlamydia
positivity across the region. Thus, changes in the demographic,
clinical, and sexual behavioral risk characteristics of the women
screened could not account for the increase in positivity over time.
We adjusted chlamydia positivity to account for the use of more
sensitive laboratory tests and to better estimate true chlamydia
prevalence. Although statistical issues have been raised concerning
the estimation of sensitivity and specificity for NAATs, the adjusted
positivity increased over time even when a range of laboratory test
Fig. 3. Trends in observed and adjusted chlamydia positivity
among women aged 15 to 24 years and percent of tests that were
nucleic acid amplification tests (NAATs), Region X family planning
? percent NAATS;
? observed positivity.
? adjusted pos-
TABLE 2.Risk of Chlamydial Infection in Women Aged 15–24 Yrs. Seen in Region X Family Planning Clinics, 1997–2004
CrudeModel 1Model 2
OR95% CIAOR 95% CI AOR*95% CI
Age group (yr)
Clinic city size, persons
One or more sexual behavioral risks†
Sex partner with chlamydia
No condom use, last sex
One or more clinical findings‡
Positive chlamydia test, past year
Chlamydia test type
Per year since 1997
1.58–1.67 1.54–1.64 1.51–1.60
*Adjusted odds ratio based on generalized linear mixed model that included the risk factor as a fixed effect and clinic as a random effect.
†Includes having had a new sex partner; multiple sex partners or a symptomatic sex partner in the past 60 d.
‡Includes cervicitis, friable cervix, ectopy, and PID.
OR indicates odds ratio; CI, confidence interval; AOR, adjusted odds ratio.
Sexually Transmitted Diseases ● November 2007
FINE ET AL
sensitivities and specificities were used.17A multivariate logistic
regression model was used to adjust for all of the demographic,
clinical, sexual behavioral risk, and laboratory test characteristics
associated with an increased risk of chlamydia. Even after taking
into account of all those risk factors, there remained a significant
5% increase in the risk of chlamydia each year. We used a
generalized linear mixed model to assess if differences in factors
across the 125 individual clinics explained the increase in positiv-
ity. We concluded it was not likely that the variability among the
clinics where screening occurred accounted for the annual increase
in chlamydia positivity in the region.
Although we used all of the data available to systematically
evaluate possible reasons for the increase in chlamydia positivity,
we were unable to examine a broader set of individual and com-
munity-level factors that could have affected disease. Unmeasured
individual characteristics, e.g., nonsexual risk behaviors or socio-
economic status, could not be evaluated as potential risk factors for
chlamydia. In addition, we were unable to evaluate patients’ sexual
network characteristics, e.g., concurrency, and their influence on
the risk of chlamydia during this time period. Previous research
has suggested that partner selection over time, interrelationships
among groups of individuals, and the disease and treatment status
of network members affect the risk of chlamydia.18,19
Clinic screening coverage among young women, i.e., the pro-
portion of women seen at the clinics and who received a chlamydia
test during a calendar year, may impact chlamydia positivity.
Previous studies found that screening coverage was between 50%
and 60% in Region X.20,21Although trend data for screening
coverage were not available, it is unlikely that there were dramatic
changes in coverage that could have accounted for the increases in
chlamydia positivity during this time period. Furthermore, it is
unlikely that coverage reached 100% in the region. Low screening
coverage could result in a number of undetected infections, con-
tributing to a continuing reservoir of chlamydial infections.
A recent study of surveillance case report data in British Co-
lumbia, Canada, documented an increase in chlamydia case rates
from 1998 through 2003.22The authors hypothesized that early
treatment of chlamydial infection may increase a population’s
susceptibility to reinfection after the introduction of a chlamydia
control program. However, this analysis did not consider sexual
networks, screening coverage, or changes to more sensitive test
methods and their effects on the increase in chlamydia case rates.
A study that evaluated azithromycin treatment on trachoma infec-
tions in Vietnam also concluded that treatment might interrupt the
duration of infection necessary for developing immunity, thus
increasing the number of individuals susceptible to reinfection and
adversely affecting the prevalence of disease over time.23
Increases in chlamydia prevalence that could not be explained
by changes in laboratory test methods or changes in the demo-
graphic characteristics of the women screened have been docu-
mented in other areas of the United States and in Sweden.24,25
Likewise, our findings suggest that there was a true increase in
chlamydia positivity among young women screened in Region X
family planning clinics from 1997 through 2004. However, our
findings also suggest that future research on trends in chlamydia
prevalence must focus on data beyond the traditional risk factor
information currently collected. For example, simulation modeling
has been used in the Netherlands to study the spread of chlamydia
and to investigate the effects of several aspects of a screening
program, namely sex and age selection for screening and partner
referral, on the incidence and prevalence of the disease.26,27Adapt-
ing such a model to populations within the United States could be
useful in assessing how varying levels of screening coverage,
treatment rates, and partner referral might contribute to changes in
chlamydia prevalence over time. In addition, research is needed to
understand how more detailed and comprehensive individual risk
measures, social network characteristics, and broader community
factors influence the trends in chlamydia prevalence. In the mean-
time, our analysis confirmed that younger sexually active women
continue to be at a high risk for chlamydial infection. Continued
screening of these women is critical to detect and treat infections
and prevent future adverse reproductive sequelae.
1. Weinstock H, Berman S, Cates W Jr. Sexually transmitted diseases
among American youth: Incidence and prevalence estimates, 2000.
Perspect Sex Reprod Health 2004; 36:6–10.
2. Cates W Jr, Brunham RC. Sexually transmitted diseases and infertility.
In: Holmes KK, Sparling PF, Mardh P-A, et al., eds. Sexually Trans-
mitted Disease, 3rd ed. New York: McGraw-Hill, 1999; 1079–1087.
3. Scholes D, Stergachis A, Heidrich FE, et al. Prevention of pelvic
inflammatory disease by screening for cervical chlamydial infection.
N Engl J Med 1996; 34:1362–1366.
4. Britton TF, DeLisle S, Fine D. STDs and family planning clinics: A
regional program for chlamydia control that works. Am J Gynecol
Health 1992; 6:80–87.
5. Centers for Disease Control and Prevention. Sexually Transmitted
Disease Surveillance, 2004. Atlanta, GA: US Department of Health
and Human Services, September 2005.
6. Fine D, Mosure D, Berman S. Behavioral risk trends for women
screened for chlamydia at family planning clinics. Presented at the
International Congress of Sexually Transmitted Disease, Seville,
Spain, October 19–22, 1997; Abstract P606:163.
7. Mertz KJ, Ransom RL, St. Louise ME, et al. Prevalence of genital
chlamydial infection in young women entering a national job train-
ing program, 1990–1997. Am J Public Health 2001; 91:1287–1290.
8. Mertz KJ, Voigt RA, Hutchins K, et al. Findings from STD screening
of adolescents and adults entering corrections facilities: Implications
for STD control strategies. Sex Transm Dis 2002; 29:834–839.
9. Addiss DG, Vaughn ML, Ludka D, et al. Decreased prevalence of
Chlamydia trachomatis infection associated with a selective screen-
ing program in family planning clinics in Wisconsin. Sex Transm
Dis 1993; 20:28–35.
10. Centers forDiseaseControlandPrevention.Sexuallytransmitteddiseases
treatment guidelines. MMWR Morb Mortal Wkly Rep 2006; 55:40.
11. Guide to Clinical Preventive Services, 2005. AHRQ Publication No.
05-0570. Agency for Healthcare Research and Quality. Rockville,
MD, June 2005. Available at: www.ahrq.gov/clinic/pocketgd.htm.
12. Dicker LW, Mosure DJ, Levine WC, et al. Impact of switching
laboratory tests on reported trends in Chlamydia trachomatis infec-
tions. Am J Epidemiol 2000; 151:430–435.
13. Rogan WJ, Gladen B. Estimating prevalence from the results of a
screening test. Am J Epidemiol 1978; 107:71–76.
14. Newhall WJ, DeLisle S, Fine D, et al. Head-to-head evaluation of five
chlamydia tests relative to a quality-assured culture standard. J Clin
Micro 1999; 37:681–685.
15. Black CM, Marrazzo J, Johnson RE, et al. Head-to-head multicenter
comparison of DNA probe and nucleic acid amplification tests for
Chlamydia trachomatis infection in women performed with an im-
proved reference standard. J Clin Micro 2002; 40:3757–3763.
16. Production GLIMMIX Procedure, 2006. Cary, NC: SAS Institute.
Available at www.support.sas.com.
17. Hagdu A, Dendukuri N, Hilden J. Evaluation of nucleic acid amplifica-
tion tests in the absence of a perfect gold-standard test: A review of the
statistical and epidemiologic issues. Epidemiology 2005; 16:604–612.
18. Foxman B, Newman M, Percha B, et al. Measures of sexual partner-
ships: Lengths, gaps, overlaps, and sexually transmitted infection.
Sex Transm Dis 2006; 33:209–214.
19. Doherty IA, Padian NS, Marlow C, Aral SO. Determinants and con-
sequences of sexual networks as they affect the spread of sexually
transmitted infections. J Infect Dis 2005; 191(suppl 1):S42–S54.
20. Gunter D, Fine D, Dicker L, Mosure D. Methods for estimating
chlamydia screening coverage among women attending family plan-
Vol. 34●No. 11
INCREASES IN CHLAMYDIA: DO WE KNOW WHY?