Pelvic inflammatory disease (PID) has been associated with infertility and has been identified as a potential sequela of untreated sexually transmitted infections (STIs) such as chlamydia and gonorrhea.1–3 A study reporting data from 3 national surveys found decreases in PID diagnosed in hospital and ambulatory visits from 1985 to 2001.4 In addition, the study found that 91% of PID cases were diagnosed in ambulatory settings.4 More recently, PID cases seemed to decline from 244,000 to 100,000 patient visits to physician offices for the condition from 2000 to 2009, respectively.5 The cause of the apparent decline is unknown. Also, a substantial portion of PID cases may not be diagnosed because the symptoms may be mild or subclinical; therefore, established estimates for PID may be lower than true incidence.6 In addition to infertility, the potential complications of PID include ectopic pregnancy and chronic pelvic pain.5 The medical cost per case of PID, including these complications, has been estimated to be approximately $3200 (2010 dollars).7
Given the substantial burden related to PID in women, several studies have examined the correlates or risk factors for PID. Correlates that have been frequently associated with PID across multiple studies include age (mixed findings),6,8–10 younger age at first sexual intercourse,8–13 number of sex partners (multiple partners in a specific time frame or a higher number of partners),6,8–10 and a history of STIs.6,8,9,12,13 Other factors associated with PID that have been identified in some studies are nonuse of contraceptives including barrier methods,6,8,11,12 lower education level or socioeconomic status,10–12 and vaginal douching.8,12 Finally, several studies have found disparities in PID incidence and prevalence by racial or ethnic groups. Specifically, these studies have found that black or nonwhite women are more likely to have PID than white women.2,8–10,14,15
However, few studies have examined recent trends and correlates of PID in the United States (US).5 Given the recently observed decreases in PID, it is important to examine correlates of PID over time. Therefore, the objectives of this study were to 1) examine temporal trends in self-reported receipt of PID treatment and 2) examine correlates of receipt of PID treatment in 1995 and 2006–2010.
We used data from the National Survey of Family Growth (NSFG), a multistage national probability sample of men and women aged 15 to 44 years living in US households. From 1973 to 2002, NSFG was conducted periodically (i.e., interviews conducted over a period of several months every 3 to 7 years). From June 2006 to June 2010, NSFG switched to continuous administration, with interviews conducted throughout this period. For this study, we used female data from the 1995, 2002, and 2006–2010 NSFG. The design of the NSFG has been described in detail elsewhere16,17; however, sample sizes and response rates for female respondents were as follows: 1995 (n = 10,847; response rate, 79%),18 2002 (n = 7643; response rate, 80%),19 and 2006–2010 (n = 12,279; response rate, 78%).20 In all years, NSFG included oversamples of black and Hispanic women, and in 2002 and 2006–2010, it also included oversamples of adolescents. National Survey of Family Growth was approved by the ethics review board at the National Center for Health Statistics, and respondents provided informed consent (adolescents provided assent after parental consent).
We used data from the computer-assisted personal-interview where highly trained female interviewers administer survey questions. Our outcome of interest—ever received treatment for PID—was defined using the following survey question: “Have you ever been treated for an infection in your fallopian tubes, womb, or ovaries, also called a pelvic infection, pelvic inflammatory disease, or PID?” We examined several demographic characteristics and sexual behaviors previously identified as predictors or correlates of PID. Demographic factors included in this study were as follows: race/ethnicity (Hispanic, non-Hispanic white, and non-Hispanic black); age (15–19, 20–24, 25–34, 35–44 years); current marital status (married to a male, cohabiting with a male partner, formerly married, never married/not cohabiting); highest education level (less than high school/no general educational development [GED], high school/GED, some college, bachelor’s degree or higher); family income as percentage of federal poverty level (FPL), referred to as income-poverty ratio level (<150%, 150%–299%, and ≥300%), and region of country where the respondent resides (northeast, midwest, south, west). Sexual and related behaviors included were as follows: used a vaginal douche regularly (1995) or in the past 12 months (2002 and 2006–2010) (douching was recoded as yes, no), age at first vaginal intercourse (<15, 15–17, 18–19, ≥20 years), and number of vaginal intercourse partners in lifetime (recoded9 into the following categories: 1, 2–3, 4–9, ≥10 partners).
We used the SURVEY procedures in SAS (release 9.2; SAS Institute, Cary, NC) for all analyses to account for the multistaged sampling procedures used by NSFG. Data were weighted to represent the US female population aged 15 to 44 years in 1995, 2002, and 2006–2010. All analyses were limited to sexually experienced women (i.e., reported at least 1 vaginal intercourse partner in lifetime) and the 3 racial/ethnic groups for which we had large sample sizes or were oversampled—Hispanic, non-Hispanic white, and non-Hispanic black. We conducted several analyses to examine trends in receipt of PID treatment and to examine correlates of PID treatment to determine if they changed over time. First, we used χ2 tests to compare the differences in receipt of PID treatment overall and by race/ethnicity across time (1995, 2002, and 2006–2010). Next, we examined the relationship between demographic characteristics and sexual and related behaviors and receipt of PID treatment in 1995 and 2006–2010, separately. Because douching was not consistently measured over time, we did not merge the data from different survey years into one analysis that included survey year as a correlate. We used χ2 tests for bivariate analyses and conducted separate adjusted logistic regression models for the 2 survey periods (1995 and 2006–2010). Given that the variables included in our analyses were found to be significantly associated with PID in previous research, we used all variables in the adjusted analyses.
Trends in PID Treatment: 1995, 2002, 2006 to 2010
We examined differences in the receipt of PID treatment overall and by race/ethnicity (Hispanic, non-Hispanic white, non-Hispanic black) and during the 3 periods included in this study. Overall, receipt of PID treatment among sexually experienced women significantly declined from 8.6% in 1995 to 5.7% in 2002 (P < 0.0001) and then leveled off to 5.0% in 2006–2010 (P = 0.16). For 1995, we found significant differences in PID treatment among the 3 racial/ethnic groups, with 11.6% of non-Hispanic black women, 8.7% of non-Hispanic white women, and 8.0% of Hispanic women reporting receipt of PID treatment in their lifetime (P < 0.001; Fig. 1). The racial/ethnic differences seemed to decline in 2002; however, a significant difference remained, with 7.4% of non-Hispanic black women, 6.4% of non-Hispanic white women, and 5.1% of Hispanic women reporting receipt of PID treatment in their lifetime (P < 0.05). In 2006 to 2010, there was no statistically significant differences in receipt of PID treatment by race/ethnicity (P = 0.22).
Correlates of PID Treatment: 1995 and 2006 to 2010
In 1995, nearly all of the demographic characteristics and sexual behaviors that we examined had significant differences in receipt of PID treatment in bivariate analyses (Table 1). For race/ethnicity (P < 0.001), age (P < 0.0001), and current marital status (P < 0.0001), women with the highest reports of receiving PID treatment were non-Hispanic black (11.6%), 35 to 44 years old (10.5%), or formerly married (12.8%). Both socioeconomic status variables that we examined, education level (P < 0.0001) and poverty income level (P < 0.0001), were significantly associated with receipt of PID treatment, with highest reports among women with less than a high school education (10.1%) or an income less than 150% of FPL (10.7%). In addition, there were significant differences by US Census region of residence, with highest reports of receipt of PID treatment in the south (10.0%). Finally, all 3 sexual and related behaviors were significantly associated with receipt of PID treatment, with highest reports among women who douched (12.1%, P < 0.0001), had first vaginal intercourse before age 15 years (12.5%, P < 0.0001), or had 10 or more lifetime vaginal intercourse partners (15.1%, P < 0.0001).
Similar patterns were found in bivariate analyses for the 2006 to 2010 time frame. Age (P = 0.04), current marital status (P < 0.01), education level (P < 0.0001), and poverty income level (P < 0.0001) remained significant correlates of receipt of PID treatment, with the highest reports among women who were 35 to 45 years old (5.6%), were formerly married (8.0%), had less than a high school education (6.7%), or had an income less than 150% of the FPL (7.5%). Conversely, although significant in 1995, there were no significant differences in receipt of PID treatment by race/ethnicity (P = 0.22) or region of the United States (P = 0.94) in 2006–2010. Similar to 1995, vaginal douching (P < 0.0001), younger age at first vaginal intercourse (P < 0.0001), and a higher number of lifetime vaginal intercourse partners (P < 0.0001) were significantly related to receipt of PID treatment in 2006 to 2010, with highest reports among women who douched (7.7%), had first vaginal intercourse before age 15 years (10.3%), or had 10 or more lifetime vaginal sex partners (8.0%).
Adjusted Analyses: 1995 and 2006–2010
All variables were entered into separate logistic regression models by the 2 survey periods (Table 2). In 1995, 2 demographic variables remained significantly associated with receipt of PID treatment in adjusted analyses—age and education level. As compared with women aged 35 to 44 years, women aged 15 to 19 years (AOR, 0.43; 95% CI, 0.25–0.71), 20 to 24 years (AOR, 0.64; 95% CI, 0.47–0.86), 25 to 29 years (AOR, 0.61; 95% CI, 0.47–0.80), and 30 to 34 years (AOR, 0.76; 95% CI, 0.61–0.96) were less likely to have received PID treatment. For education level, women having less than a high school diploma (AOR, 1.49; 95% CI, 1.06–2.11) and those with a high school diploma or GED (AOR, 1.32; 95% CI, 1.02–1.70) had higher reports of receipt of PID treatment than did women with a bachelor’s degree or higher. All 3 sexual and related behaviors remained significant in adjusted analyses. Women who used a vaginal douche (AOR, 1.49; 95% CI, 1.24–1.80), had first vaginal sex before age 15 years (AOR, 1.59; 95% CI, 1.12–2.27), and had 4 to 9 partners (AOR, 1.67; 95% CI, 1.31–2.13) or 10 or more partners (AOR, 2.61; 95% CI, 1.91–3.55) were more likely to report receipt of PID treatment than women who did not douche, had first vaginal sex at 20 years or older, or had one lifetime vaginal sex partner, respectively.
Findings from the 2006–2010 logistic regression analysis had some similarities but also had some key differences to findings from the 1995 analysis. Similar to 1995, women with less education (having less than a high school diploma [AOR, 1.95; 95% CI, 1.14–3.35], a high school diploma or GED [AOR, 1.89; 95% CI, 1.23–2.90], or some college [AOR, 2.27; 95% CI, 1.46–3.55]) were more likely to report receipt of PID treatment as compared with women with a bachelor’s degree or higher. Also similar to 1995, women who douched (AOR, 1.33; 95% CI, 1.01–1.75) were more likely as those who did not douche to report receipt of PID treatment in 2006–2010. In addition, women who had first vaginal sex before age 15 years (AOR, 2.57; 95% CI, 1.36–4.85) were more likely to report receipt of PID treatment than women who had first vaginal sex at 20 years or older. Those who had 4 to 9 partners (AOR, 1.76; 95% CI, 1.03–3.00) or 10 or more partners (AOR, 2.09; 95% CI, 1.20–3.66) were more likely to report receipt of PID treatment than women who had one lifetime vaginal sex partner. Conversely, findings for age and poverty-level income differed from 1995 to 2006–2010. In 2006–2010, there was only 1 age group (instead of 4 in 1995) affected by a significant difference for age—women who were 15 to 19 years old (AOR, 0.38; 95% CI, 0. 21–0.68) were less likely than women 35 to 44 years old to report receipt of PID treatment. In addition, women with an income less than 150% of the FPL (AOR, 2.60; 95% CI, 1.79–3.76) and 150% to 299% of FPL (AOR, 1.84; 95% CI, 1.20–2.82) were more likely than women with an income 300% of the FPL or higher to report receipt of PID treatment.
Given that many women younger than 22 years are in the process of continuing their education, we conducted post hoc adjusted analyses on 22- to 44-year-old women to confirm that our results for education would be similar when limited to this somewhat older subpopulation. In both 1995 and 2006–2010, our findings for education were similar when examining 15- to 44-year-old and 22- to 44-year-old women.
Trends in PID Treatment
Findings from this study demonstrate a significant decline in self-reported receipt of PID treatment among reproductive-aged women in the United States. Extending earlier research on decreases in PID across time,4 we found a continued decline in receipt of PID treatment between 1995 and 2002; however, we found no difference between 2002 and 2006–2010. In addition, racial differences in PID diagnosis or treatment were frequently found in earlier research.2,8–10,14,15 However, although we found a difference by race/ethnicity using bivariate analyses in 1995 and 2002, race/ethnicity was not a significant correlate of receipt of PID treatment in our multivariable analysis of 1995 data.
It is possible that the continued decline in receipt of PID treatment may be related to changes in STI rates, screening for STIs such as gonorrhea and chlamydia (e.g., through identification of asymptomatic cases), or issues related to health care access. However, the number of reported cases of gonorrhea varied little during our study time frame.5 In addition, reported chlamydia cases seemed to increase during this time, but research suggests that this increase may be related to increases in screening coverage or improvements in laboratory test technology.21 A randomized controlled trial of chlamydia screening found fewer cases of PID in women in the screening arm,22 and a more recent analysis of monthly medical insurance claims data for women with private health insurance found that higher screening rates for chlamydia and gonorrhea were associated with a decreased rate of PID 4 months later.23 In addition, chlamydia screening of 16- to 24-year-old women enrolled in commercial or Medicaid plans has consistently increased since 2001.24 An analysis of 15- to 25-year-old women in 2006–2008 NSFG found that non-Hispanic black women were more likely than non-Hispanic white women to report receiving a chlamydia test.25 Thus, improvements in chlamydia screening, including treatment of the infection that is often asymptomatic, may be related to the observed declines in PID treatment over time. Although it is possible that changes in health care access played a role in the decline in PID, a previous analysis of NSFG data found no relationship between health insurance mandates and infertility service use.26 The percentage of uninsured in the United States did not seem to change from 1995 to 2007; however, there were slight shifts from private to public insurance coverage during this time.27 Therefore, additional research is needed to determine if health care access issues such as type of insurance were related to the observed decreases in PID treatment.
Correlates of PID Treatment
In both time frames, consistent with some previous research, receipt of PID treatment was related to douching28 and sexual behaviors8,9,12; however, our analysis identified different patterns for age and poverty status across the 2 time frames. In 2006–2010, only adolescents were less likely than women 35 to 44 years old to have ever received treatment for PID, whereas in 1995, all age groups were less likely to have received treatment for PID. It is possible that other factors such as a significant decrease in age disparities for reported cases of gonorrhea from 1981 to 200529 and the increasing importance of socioeconomic status and PID treatment may have played a role in the changing age-related patterns over time. As noted, socioeconomic status is an important predictor of PID treatment. In adjusted analyses, women whose education level was less than a bachelor’s degree were more likely to report receipt of PID treatment in both time frames; however, poverty level was only significant in 2006–2010, where women with lower household incomes were more likely to have received PID treatment.
There are some limitations to this study. First, our measure of receipt of PID treatment was subject to possible bias (e.g., recall and diagnosis biases). Diagnosis of PID can be difficult given that symptoms are similar to those of other reproductive health problems.6,30 Also, one study found inconsistencies in PID diagnoses across health care providers in Australia, and authors speculated that it may be related to cases of PID missed by some doctors.31 It is also possible that PID has become increasingly subclinical over time, leading to a decrease in diagnoses. Another limitation is that we did not have a consistent measure of health insurance and health care access with appropriate time frames across the datasets. There were limitations with our douching measure, given different time frame anchors and question wording used in 1995 and 2006–2010. In addition, there are limitations in adolescents’ knowledge about household income; therefore, our measure of poverty income level may be less valid for this subpopulation. Finally, we did not include a measure of STI history because questions about chlamydia and gonorrhea diagnoses used a shorter time frame (past 12 months) that did not match well with our lifetime PID treatment variable. However, the strength of our study is that we used one of the few existing national data sources that include information on PID treatment for reproductive-aged women.
In summary, we found a decline in receipt of PID treatment and associated racial disparities from 1995 to 2006–2010. However, in 2006–2010, at least 2.5 million reproductive-aged women have received treatment for PID with the burden of disease falling on women of lower socioeconomic status. Therefore, public health and other medical efforts to reduce PID in these women are needed. Finally, additional research is needed to determine the cause of the apparent decline in PID.
1. Westrom L, Joesoef R, Reynolds G, et al.. Pelvic inflammatory disease and fertility: A cohort study of 1,844 women with laparoscopically verified disease and 657 control women with normal laparoscopic results. Sex Transm Dis 1992; 19: 185–192.
2. Cates W, Rolfs RT, Aral S. Sexually transmitted diseases, pelvic inflammatory disease, and infertility: An epidemiologic update. Epidemiol Rev 1990; 12: 199–220.
4. Sutton MY, Sternberg M, Zaidi A, et al.. Trends in pelvic inflammatory disease hospital discharges and ambulatory visits, United States, 1998–2001. Sex Transm Dis 2005; 32: 778–784.
6. Beigi RH, Wiesenfeld HC. Pelvic inflammatory disease: New diagnostic criteria and treatment. Obstet Gynecol Clin N Am 2003; 30: 777–793.
7. Owusu-Edusei K, Chesson HW, Gift TL, et al.. The estimated direct medical cost of sexually transmitted infections in the United States, 2008. Sex Transm Dis 2013; 40:197–201.
8. Aral SO, Mosher WD, Cates W. Self-reported pelvic inflammatory disease in the United States, 1988. JAMA 1991; 266: 2570–2573.
9. Grodstein F, Rothman KJ. Epidemiology of pelvic inflammatory disease. Epidemiology 1994; 5: 234–242.
10. Simms I, Stephenson JM, Mallinson H, et al.. Risk factors associated with pelvic inflammatory disease. Sex Transm Infect 2006; 82: 452–457.
11. Lidegaard O, Helm P. pelvic inflammatory disease: The influence of contraceptive, sexual, and social life events. Contraceptin 1990; 41: 475–483.
12. Jossens MOR, Eskenazi B, Schachter J, et al.. Risk factors for pelvic inflammatory disease: A case-control study. Sex Transm Dis 1996; 23: 239–247.
13. Suss AL, Homel P, Hammerschlag M, et al.. Risk factors for pelvic inflammatory disease in inner-city adolescents. Sex Transm Dis 2000; 27: 289–291.
14. Aral SO, Mosher WD, Cates W. Self-reported pelvic inflammatory disease in the US: A common occurrence. Am J Public Health 1985; 75: 1216–1218.
15. Cox S, Dean T, Posner SF, et al.. Disparities in reproductive health-related visits to the emergency department, Maryland, 1999–2005. J Womens Health 2011; 20: 1833–1838.
16. Lepkowski J, Mosher W, Davis K, et al.. National Survey of Family Growth, Cycle 6: Sample design, weighting, imputation, and variance estimation. Vital and Health Statistics, Series 2, No. 142. Hyattsville, MD: National Center for Health Statistics, 2006.
17. Lepkowski JM, Mosher WD, Davis KE, et al.. The 2006–2010 National Survey of Family Growth: Sample design and analysis of a continuous survey. Vital and Health Statistics Series 2, No. 150. Hyattsville, MD: National Center for Health Statistics, 2010.
18. Potter FJ, Iannacchione VG, Mosher WD, et al.. Sample design, sampling weights, imputation, and variance estimation in the 1995 National Survey of Family Growth. Vital Health Stat 1998; 2 (124): 1–63.
19. Mosher WD, Martinez GM, Chandra A, et al.. Use of Contraception and Use of Family Planning Services in the United States, 1982–2002. Advance Data From Vital and Health Statistics; No. 350. Hyattsville, MD: National Center for Health Statistics, 2004.
20. Chandra A, Billioux VG, Copen CE, et al.. HIV Risk-Related Behaviors in the United States Household Population Aged 15–44: Data From the National Survey of Family Growth, 2002 and 2006–2010. National Health Statistics Reports; No. 46. Hyattsville, MD: National Center for Health Statistics, 2012.
21. Satterwhite CL, Tian LH, Braxton J, et al.. Chlamydia prevalence among women and men entering the National Job Training program: United States, 2003–2007. Sex Transm Dis 2010; 37: 63–67.
22. Scholes D, Stergachis A, Heidrich FE, et al.. Prevention of pelvic inflammatory disease by screening for cervical chlamydial infection. N Engl J Med 1996; 334: 1362–1366.
23. Owusu-Edusei K, Bohm MK, Chesson HW, et al.. Chlamydia screening and pelvic inflammatory disease: Insights from exploratory time-series analysis. Am J Prev Med 2010; 38: 652–657.
25. Tao G, Hoover KW, Leichliter JS, et al.. Self-reported chlamydia testing rates of sexually active women aged 15–25 years in the United States, 2006–2008. Sex Transm Dis 2012; 39: 605–607.
26. Bitler M, Schmidt L. Health disparities and infertility: Impacts of state-level insurance mandates. Fertil Steril 2006; 85: 858–865.
27. Cohen RA, Makuc DM, Bernstein AB, et al.. Health Insurance Coverage Trends, 1959–2007: Estimates From the National Health Interview Survey. National Health Statistics Reports; No. 17. Hyattsville, MD: National Center for Health Statistics, 2009. Available at: http://www.cdc.gov/nchs/data/nhsr/nhsr017.pdf
28. Martino JL, Youngpairoj S, Vermund SH. Vaginal douching: Personal practices and public policies. J Womens Health 2004; 13: 1048–1065.
29. Chesson HW, Zaidi AA, Aral SO. Decreasing age disparities in syphilis and gonorrhea incidence rates in the United States, 1981–2005. Sex Transm Dis 2008; 35: 393–397.
30. Haggerty CL, Ness RB. Diagnosis and treatment of pelvic inflammatory disease. Women Health 2008; 4: 383–397.
31. Doxanakis A, Hayes RD, Chen MY, et al.. Missing pelvic inflammatory disease? Substantial differences in the rate at which doctors diagnose PID. Sex Transm Infect 2008; 84: 518–513.