Risk Factors for Incident Polycystic Ovary Syndrome Diagnosis.

Objective: While highly prevalent, risk factors for incident polycystic ovary syndrome (PCOS) are poorly delineated. Using a population-based cohort, we sought to identify predictors of incident PCOS diagnosis. Materials and Methods: A matched case-control analysis was completed utilizing patients enrolled in Kaiser Permanente Washington from 2006 to 2019. Inclusion criteria included female sex, age 16-40 years, and ≥3 years of prior enrollment with ≥1 health care encounter. PCOS cases were identified using International Classification of Diseases codes. For each incident case (n = 2,491), 5 patients without PCOS (n = 12,455) were matched based on birth year and enrollment status. Potential risk factors preceding diagnosis included family history of PCOS, premature menarche, parity, race, weight gain, obesity, valproate use, metabolic syndrome, epilepsy, prediabetes, and types 1 and 2 diabetes. Potential risk factors for incident PCOS diagnosis were assessed with univariate and multivariable conditional logistic regressions. Results: Mean age of PCOS cases was 26.9 years (SD 6.8). PCOS cases, compared with non-PCOS, were more frequently nulliparous (70.9% versus 62.4%) and in the 3 years prior to index date were more likely to have obesity (53.8% versus 20.7%), metabolic syndrome (14.5% versus 4.3%), prediabetes (7.4% versus 1.6%), and type 2 diabetes (4.1% versus 1.7%) (p < 0.001 for all comparisons). In multivariable models, factors associated with higher risk for incident PCOS included the following: obesity (compared with nonobese) Class I-II (body-mass index [BMI], 30-40 kg/m2; odds ratio [OR], 3.8; 95% confidence interval [CI], 3.4-4.2), Class III (BMI > 40 kg/m2; OR, 7.5, 95% CI, 6.5-8.7), weight gain (compared with weight loss or maintenance) of 1-10% (OR, 1.7, 95% CI, 1.3-2.1), 10-20% (OR, 1.9; 95% CI, 1.5-2.4), and >20% (OR, 2.6; 95% CI, 1.9-3.6), prediabetes (OR, 2.7; 95% CI, 2.1-3.4), and metabolic syndrome (OR, 1.8: 95% CI, 1.5-2.1). Conclusion: Excess weight gain, obesity, and metabolic dysfunction may play a key role in the ensuing phenotypic expression of PCOS. Treatment and prevention strategies targeted at preventing weight gain in early reproductive years may help reduce the risk of this syndrome.

Incidence, prevalence, and trends in polycystic ovary syndrome diagnosis: a United States population-based study from 2006 to 2019.

BACKGROUND: Polycystic ovary syndrome is the most common endocrine disorder in women of reproductive age, yet US incidence estimates do not exist, and prevalence estimates vary widely. OBJECTIVE: A population-based US study estimated the incidence, prevalence, and trends of polycystic ovary syndrome by age, race and ethnicity, and diagnosing provider type. STUDY DESIGN: A retrospective cohort study of patients enrolled in Kaiser Permanente Washington from 2006 to 2019 was conducted. All members identified as female, aged 16 to 40 years with at least 3 years of enrollment and at least 1 healthcare encounter during that time, were eligible for inclusion. Individuals were excluded if they had a history of oophorectomy or hysterectomy. Polycystic ovary syndrome cases were identified using the International Classification of Diseases diagnosis codes (International Classification of Diseases, Ninth Revision, 256.4 or International Classification of Diseases, Tenth Revision, E28.2). Individuals with a polycystic ovary syndrome diagnosis before study entry were excluded from incidence rate estimations. The incidence rates were adjusted by age using direct standardization to the 2010 US census data. Temporal trends in incidence were assessed using weighted linear regression (overall) and Poisson regression (by age, race and ethnicity, and provider type). Prevalent cases were defined as patients with a polycystic ovary syndrome diagnosis at any time before the end of 2019. Medical record review of 700 incident cases diagnosed in 2011-2019 was performed to validate incident cases identified by International Classification of Diseases codes using the Rotterdam criteria. RESULTS: Among 177,527 eligible patients who contributed 586,470 person-years, 2491 incident polycystic ovary syndrome cases were identified. The mean age at diagnosis was 26.9 years, and the mean body mass index was 31.6 kg/m2. Overall incidence was 42.5 per 10,000 person-years; the rates were similar over time but increased in individuals aged 16 to 20 years from 31.0 to 51.9 per 10,000 person-years (P=.01) and decreased among those aged 26 to 30 years from 82.8 to 45.0 per 10,000 person-years (P=.02). A small decreasing temporal trend in incidence rates was only observed among non-Hispanic White individuals (P=.01). The incidence rates by diagnosing provider type varied little over time. Among the 58,241 patients who contributed person-time in 2019, 3036 (5.2%) had a polycystic ovary syndrome International Classification of Diseases diagnosis code; the prevalence was the highest among the Hawaiian and Pacific Islander group (7.6%) followed by Native American and Hispanic groups. Medical record review classified 60% as definite or probable incident, 14% as possible incident, and 17% as prevalent polycystic ovary syndrome. The overall positive predictive value of polycystic ovary syndrome International Classification of Diseases diagnosis code for identifying definite, probable, or possible incident polycystic ovary syndrome was 76% (95% confidence interval, 72%-79%). CONCLUSION: Among a cohort of nonselected females in the United States, we observed stable rates of incident polycystic ovary syndrome diagnoses over time. The incidence of polycystic ovary syndrome was 4- to 5-fold greater than reported for the United Kingdom. The prevalence of polycystic ovary syndrome (5.2%) was almost double before the published US estimates (2.9%) based on the International Classification of Diseases codes. Race and ethnicity and provider type did not seem to have a major impact on temporal rates. Incident diagnoses increased over time in younger and decreased in older age groups, perhaps related to shifting practice patterns with greater awareness among practitioners of the impact of polycystic ovary syndrome on long-term health outcomes and improved prevention efforts. Moreover, increasing obesity rates may be a factor driving the earlier ages at diagnosis.

Incidence, prevalence, and trends in endometriosis diagnosis: a United States population-based study from 2006 to 2015.

BACKGROUND: Accurate estimates of incidence and prevalence of endometriosis among nonselected cohorts are lacking in the United States, and earlier reports have produced varying results. OBJECTIVE: This study aimed to define endometriosis incidence and prevalence in a US population and evaluate factors influencing these estimates over time. STUDY DESIGN: A 10-year retrospective cohort study using Kaiser Permanente Washington electronic health records database was completed. The primary analysis included women enrollees aged 16 to 60 years, from January 2006 to December 2015, who had a uterus, were continuously enrolled for at least 2 years before cohort entry and had at least 1 healthcare utilization. Secondary analysis included all women enrollees aged 16 to 60 years during this time. Incident endometriosis was identified using the International Classification of Diseases, Ninth Revision and Tenth Revision, diagnosis codes. Annual incidence rates were age-adjusted by direct standardization to the 2015 study population. Secular trends in incidence overall and by 5-year age group, race and ethnicity, diagnosis modality, and practitioner type were assessed using Poisson regression analyses. Prevalent cases were defined as women enrolled in 2015 and had an endometriosis diagnosis before the end of 2015. The prevalence rates of chronic pelvic pain and dysmenorrhea defined by the International Classification of Diseases, Ninth Revision and Tenth Revision, diagnosis codes in 2006-2015 were estimated. RESULTS: Among 332,056 eligible women who contributed 1,176,329 person-years during the 10-year study period, 2863 incident endometriosis cases were identified for an average incidence of 24.3 cases per 10,000 person-years. In our primary analysis, incidence rates declined over the study interval from a high of 30.2 per 10,000 person-years in 2006 to 17.4 per 10,000 person-years in 2015 and were highest among women aged 36 to 45 years in most years. Incidence rates were similar across race and ethnicity groups. The distribution of the 2863 incident cases by the diagnosis modality was as follows: 45.5% surgical, 5.7% imaging, and 48.8% clinical. Endometriosis incidence rates per 10,000 person-years were similar in women who were surgically and clinically diagnosed and decreased significantly from 2006 to 2015 (surgically diagnosed endometriosis dropped from 13.4 to 7.4 and clinically diagnosed endometriosis dropped from 16.1 to 8.9; P value of <.001 for linear trend over time for each). Incident case distribution by diagnosing provider was as follows: 73.6% obstetrician and gynecologist, 15.7% primary care provider, and 10.7% "other." Incidence of endometriosis diagnosed by an obstetrician and gynecologist and primary care provider decreased over the study interval (P<.001 for linear trend over time for each). Method of diagnosis and provider type did not differ by race and ethnicity. Among 135,162 women who contributed person-time in 2015, 2521 women were diagnosed with endometriosis, a prevalence rate of 1.9%. In our secondary analysis, the frequency of chronic pelvic pain diagnosis increased over the study interval from 3.0% in 2006 to 5.6% in 2015. CONCLUSION: The incidence rates of endometriosis declined over the 10-year study interval and did so uniformly across age groups, races and ethnicities, and the main diagnosing modalities and providers. Declining rates may reflect a shift in practice patterns in the United States away from the diagnosis of endometriosis both clinically and surgically, rather than favoring more general diagnoses of chronic pelvic pain. The prevalence of endometriosis in 2015 in the United States is in keeping with data from recent studies outside the United States using health record data.

Association between triglycerides, known risk SNVs and conserved rare variation in SLC25A40 in a multi-ancestry cohort.

BACKGROUND: Elevated triglycerides (TG) are associated with, and may be causal for, cardiovascular disease (CVD), and co-morbidities such as type II diabetes and metabolic syndrome. Pathogenic variants in APOA5 and APOC3 as well as risk SNVs in other genes [APOE (rs429358, rs7412), APOA1/C3/A4/A5 gene cluster (rs964184), INSR (rs7248104), CETP (rs7205804), GCKR (rs1260326)] have been shown to affect TG levels. Knowledge of genetic causes for elevated TG may lead to early intervention and targeted treatment for CVD. We previously identified linkage and association of a rare, highly conserved missense variant in SLC25A40, rs762174003, with hypertriglyceridemia (HTG) in a single large family, and replicated this association with rare, highly conserved missense variants in a European American and African American sample. METHODS: Here, we analyzed a longitudinal mixed-ancestry cohort (European, African and Asian ancestry, N = 8966) from the Electronic Medical Record and Genomics (eMERGE) Network. We tested associations between median TG and the genes of interest, using linear regression, adjusting for sex, median age, median BMI, and the first two principal components of ancestry. RESULTS: We replicated the association between TG and APOC3, APOA5, and risk variation at APOE, APOA1/C3/A4/A5 gene cluster, and GCKR. We failed to replicate the association between rare, highly conserved variation at SLC25A40 and TG, as well as for risk variation at INSR and CETP. CONCLUSIONS: Analysis using data from electronic health records presents challenges that need to be overcome. Although large amounts of genotype data is becoming increasingly accessible, usable phenotype data can be challenging to obtain. We were able to replicate known, strong associations, but were unable to replicate moderate associations due to the limited sample size and missing drug information.

Adenomyosis incidence, prevalence and treatment: United States population-based study 2006-2015.

BACKGROUND: Adenomyosis symptoms are disabling. Population-based data on incidence and prevalence of adenomyosis are lacking that could guide future evidence-based treatments and clinical management. OBJECTIVE: To evaluate the incidence, 10-year secular trends, and prevalence of adenomyosis diagnoses and to describe symptoms and treatment patterns in a large U.S. cohort. STUDY DESIGN: We performed a retrospective population-based cohort study of women aged 16-60 years in 2006-2015, enrolled in Kaiser Permanente Washington, a mixed-model health insurance and care delivery system. Adenomyosis diagnoses identified by ICD codes from the International Classification of Diseases 9th and 10th editions and potential covariates were extracted from computerized databases. Women with prior hysterectomy, and for incidence estimates women with prior adenomyosis diagnoses, were excluded. Linear trends in incidence rates over the 10-year study period were evaluated using Poisson regression. Rates and trend tests were examined for all women adjusting for age using direct standardization to the 2015 study population, by age groups, and by race/ethnicity. Chart reviews were performed to validate diagnostic accuracy of ICD codes in identifying adenomyosis incidence. Symptoms and treatment patterns at diagnosis and in the following 5 years were assessed. RESULTS: A total of 333,693 women contributed 1,185,855 woman-years (2006-2015) for incidence calculations. Associated symptom-related codes (menorrhagia or abnormal uterine bleeding, dysmenorrhea or pelvic pain, dyspareunia, and infertility) were observed in 90.8%; 18.0% had co-occurrent endometriosis codes and 47.6% had co-occurrent uterine fibroid codes. The overall adenomyosis incidence was 1.03% or 28.9 per 10,000 woman-years, with a high of 30.6 in 2007 and a low of 24.4 in 2014. Overall age-adjusted estimated incidence rates declined during the 10-year study interval (linear trend P < .05). Incidence was highest for women aged 41-45 years (69.1 per 10,000 woman-years in 2008) and was higher for black (highest 44.6 per 10,000 woman-years in 2011) vs white women (highest 27.9 per 10,000 woman-years in 2010). Overall prevalence in 2015 was 0.8% and was highest among women aged 41-45 years (1.5%). Among the 624 potential adenomyosis cases identified by diagnostic codes in 2012-2015 and with sufficient information in the medical record to determine true case status, 490 were confirmed as incident cases, yielding a 78.5% (95% confidence interval, 75.1%, 81.7%) positive predictive value of adenomyosis ICD-9/ICD-10 codes for identifying an incident adenomyosis case. Health care burden was substantial: 82.0% of women had hysterectomies, nearly 70% had imaging studies suggestive of adenomyosis, and 37.6% used chronic pain medications. CONCLUSION: Adenomyosis burden to the individual and the health care system is high. Incidence rates are disproportionately high among black women. These findings are of concern, as currently available long-term medical therapies remain limited beyond hysterectomy. Our data and methodologies are novel and could serve as a foundation to guide clinicians and health care systems to develop clinical management plans and track outcomes for women with adenomyosis.

Identification and validation of uterine perforation, intrauterine device expulsion, and breastfeeding in four health care systems with electronic health records.

OBJECTIVE: To validate algorithms identifying uterine perforations and intrauterine device (IUD) expulsions and to ascertain availability of breastfeeding status at the time of IUD insertion. STUDY DESIGN AND SETTING: Four health care systems with electronic health records (EHRs) participated: Kaiser Permanente Northern California (KPNC), Kaiser Permanente Southern California (KPSC), Kaiser Permanente Washington (KPWA), and Regenstrief Institute (RI). The study included women ≤50 years of age with an IUD insertion. Site-specific algorithms using structured and unstructured data were developed and a sample validated by EHR review. Positive predictive values (PPVs) of the algorithms were calculated. Breastfeeding status was assessed in a random sample of 125 women at each research site with IUD placement within 52 weeks postpartum. RESULTS: The study population included 282,028 women with 325,582 IUD insertions. The PPVs for uterine perforation were KPNC 77%, KPSC 81%, KPWA 82%, and RI 47%; PPVs for IUD expulsion were KPNC 77%, KPSC 87%, KPWA 68%, and RI 37%. Across all research sites, breastfeeding status at the time of IUD insertion was determined for 94% of those sampled. CONCLUSIONS: Algorithms with a high PPV for uterine perforation and IUD expulsion were developed at 3 of the 4 research sites. Breastfeeding status at the time of IUD insertion could be determined at all research sites. Our findings suggest that a study to evaluate the associations of breastfeeding and postpartum IUD insertions with risk of uterine perforation and IUD expulsion can be successfully conducted retrospectively; however, automated application of algorithms must be supplemented with chart review for some outcomes at one research site due to low PPV.

Identification of Incident Uterine Fibroids Using Electronic Medical Record Data.

INTRODUCTION: Uterine fibroids are the most common benign tumors of the uterus and are associated with considerable morbidity. Diagnosis codes have been used to identify fibroid cases, but their accuracy, especially for incident cases, is uncertain. METHODS: We performed medical record review on a random sample of 617 women who received a fibroid diagnosis during 2012-2014 to assess diagnostic accuracy for incident fibroids. We developed 2 algorithms aimed at improving incident case-finding using classification and regression tree analysis that incorporated additional electronic health care data on demographics, symptoms, treatment, imaging, health care utilization, comorbidities and medication. Algorithm performance was assessed using medical record as gold standard. RESULTS: Medical record review confirmed 482 fibroid cases as incident, resulting a 78 percent positive predictive value (PPV) for incident cases based on diagnosis codes alone. Incorporating additional electronic data, the first algorithm classified 395 women with a pelvic ultrasound on diagnosis date but none before as incident cases. Of these, 344 were correctly classified, yielding an 87 percent PPV, 71 percent sensitivity, and 62 percent specificity. A second algorithm built on the first algorithm and further classified women based on a fibroid diagnosis code of 218.9 in 2 years after incident diagnosis and lower body mass index; yielded 93 percent PPV, 53 percent sensitivity, and 85 percent specificity. CONCLUSIONS: Compared to diagnosis codes alone, our algorithms using fibroid diagnosis codes and additional electronic data improved identification of incident cases with higher PPV, and high sensitivity or specificity to meet different aims of future studies seeking to identify incident fibroids from electronic data.

"It would be so much easier": health system-led genetic risk notification-feasibility and acceptability of cascade screening in an integrated system.

Assess the feasibility and acceptability of health system-led genetic risk notification in a US integrated health system. We conducted semi-structured phone interviews with individuals age 40-64 years who had undergone genetic sequencing, but had not yet received their results, assessing attitudes to direct outreach to relatives. During each interview, we collected contact information for adult relatives identified as members of the same system and attempted to identify each relative in administrative data. We conducted 20 interviews. Most participants expressed support for Kaiser Permanente Washington involvement in familial risk notification. Direct outreach to relatives received the most unqualified support; outreach to the relatives' physician or interaction with the relatives' electronic medical record received more tempered support. Support was motivated by the desire to have risk communicated accurately and quickly. The most common caveat was a desire to alert relatives before the health system contacted them. Of 57 named relatives who were members of the same health system, we retrieved a single match for 40 (70.2%) based on name or birthdate. Health system involvement in familial risk notification received support in a sample of patients in a US integrated health system, and identification of relatives is feasible.

A US population-based study of uterine fibroid diagnosis incidence, trends, and prevalence: 2005 through 2014.

BACKGROUND: Despite considerable public health burden, uterine fibroid population-based incidence estimates are few. Secular trends over time are even more limited. OBJECTIVE: We sought to evaluate the incidence, 10-year secular trends, and prevalence of uterine fibroid diagnoses and describe the proportion of symptomatic women. STUDY DESIGN: We performed a retrospective population-based cohort study of women, aged 18-65 years, enrolled 2005 through 2014 in Kaiser Permanente Washington. Uterine fibroid diagnoses identified by International Classification of Diseases, Ninth Revision codes and potential covariates were extracted from computerized databases. Women with prior hysterectomy and, for incidence estimates, women with prior fibroid diagnoses were excluded. Linear trends in incidence rates over the 10-year study period were evaluated using Poisson regression models. Rates and trend tests were examined for all women, by age groups, and by race/ethnicity. RESULTS: Associated International Classification of Diseases, Ninth Revision symptom-related codes were observed in 90% of incident cases. Incidence rates for fibroid diagnoses were highest for the age group 45-49 years, 240.3 per 10,000 woman-years in 2014, and for black women across all years. Overall age-adjusted estimated incidence rates declined during the 10-year study interval, from 139.4 per 10,000 woman-years in 2005 to 101.4 in 2014 (P value trend .0008). Overall prevalence in 2014 was 9.6%, and was highest among women aged 50-54 years (15.9%). Black women had higher prevalence (18.5%) than other racial/ethnic groups. CONCLUSION: We found a decreasing trend of new uterine fibroid diagnoses among predominantly symptomatic women ages 18-65 years in a recent 10-year interval. This finding was due, perhaps in part, to secular trends of decreasing hysterectomies. Nonetheless, uterine fibroids remain a common health burden, with a prevalence of nearly 10%. Rates are disproportionately high and occur at younger ages for black women, and possibly for other non-white racial/ethnic groups. These findings are of concern, as current available long-term medical therapies remain limited.

A GWAS Study on Liver Function Test Using eMERGE Network Participants.

INTRODUCTION: Liver enzyme levels and total serum bilirubin are under genetic control and in recent years genome-wide population-based association studies have identified different susceptibility loci for these traits. We conducted a genome-wide association study in European ancestry participants from the Electronic Medical Records and Genomics (eMERGE) Network dataset of patient medical records with available genotyping data in order to identify genetic contributors to variability in serum bilirubin levels and other liver function tests and to compare the effects between adult and pediatric populations. METHODS: The process of whole genome imputation of eMERGE samples with standard quality control measures have been described previously. After removing missing data and outliers based on principal components (PC) analyses, 3294 samples from European ancestry were used for the GWAS study. The association between each single nucleotide polymorphism (SNP) and total serum bilirubin and other liver function tests was tested using linear regression, adjusting for age, gender, site, platform and ancestry principal components (PC). RESULTS: Consistent with previous results, a strong association signal has been detected for UGT1A gene cluster (best SNP rs887829, beta = 0.15, p = 1.30x10-118) for total serum bilirubin level. Indeed, in this region more than 176 SNPs (or indels) had p<10-8 spanning 150Kb on the long arm of chromosome 2q37.1. In addition, we found a similar level of magnitude in a pediatric group (p = 8.26x10-47, beta = 0.17). Further imputation using sequencing data as a reference panel revealed association of other markers including known TA7 repeat indels (rs8175347) (p = 9.78x10-117) and rs111741722 (p = 5.41x10-119) which were in proxy (r2 = 0.99) with rs887829. Among rare variants, two Asian subjects homozygous for coding SNP rs4148323 (G71R) were identified. Additional known effects for total serum bilirubin were also confirmed including organic anion transporters SLCO1B1-SLCO1B3, TDRP and ZMYND8 at FDR<0.05 with no gene-gene interaction effects. Phenome-wide association studies (PheWAS) suggest a protective effect of TA7 repeat against cerebrovascular disease in an adult cohort (OR = 0.75, p = 0.0008). Among other liver function tests, we also confirmed the previous effect of the ABO blood group locus for variation in serum alkaline phosphatase (rs579459, p = 9.44x10-15). CONCLUSIONS: Taken together, our data present interesting findings with strong confirmation of previous effects by simply using the eMERGE electronic health record phenotyping. In addition, our findings indicate that similar to the adult population, the UGT1A1 is the main locus responsible for normal variation of serum bilirubin in pediatric populations.

A randomized trial of strategies to increase chlamydia screening in young women.

OBJECTIVE: Despite the recommendations of numerous clinical practice guidelines, testing of at-risk women for Chlamydia trachomatis infection remains low. We evaluated an intervention to increase guideline-recommended chlamydia screening. METHOD: In a two-by-two factorial design randomized trial conducted in 2001-2002, 23 primary care clinics at Group Health Cooperative in Washington State were randomized to either control (standard) or intervention (enhanced) guideline implementation arms. Clinic-level intervention strategies included use of clinic-based opinion leaders, individual measurement and feedback, and exam room reminders. A second patient-level intervention, a chart prompt to screen for chlamydia, was delivered in a random sample of 3509 women. The outcome measure was post-intervention chlamydia testing rates among sexually active women ages 14-25. RESULTS: The clinic-level intervention did not significantly affect overall chlamydia testing (odds ratio (OR) = 1.08, 95% confidence interval (CI) 0.92-1.26, P = 0.31). However, testing rates increased significantly among women making preventive care visits (OR, Pap test visit = 1.23, 95% CI, 1.01-1.51, P = 0.04; OR, physical exam visit = 1.22, 95% CI 1.06-1.42, P = 0.009, intervention vs. control clinics). The chart prompt intervention had no significant effect (OR = 1.08, 95% CI 0.94-1.23, P = 0.27). CONCLUSIONS: Interventions to improve guideline-recommended chlamydia testing increased testing among women making preventive care visits. Additional organizational change and/or patient activation strategies may improve plan-wide testing, particularly among asymptomatic women.

Is provider continuity associated with chlamydia screening for adolescent and young adult women?

BACKGROUND: Longitudinal patient-provider relationships are a cornerstone of primary care. For many prevention services, better continuity of provider has been associated with better adherence to recommended practice. Our objective was to examine the relationship between continuity of care and chlamydia screening in adolescent and young women, a preventive service where large performance gaps exist. METHODS: The study population included 4117 sexually active women aged 14-25 years continuously enrolled at a large U.S. HMO. Administrative data from 2000 to 2002 were used to document chlamydia testing, provider continuity, and selected covariates. We used logistic regression to examine the relationship between provider continuity and chlamydia testing after controlling for potential confounders. RESULTS: 57.2% of eligible young women received a chlamydia test over the 2-year period. After controlling for utilization and other confounders, we found women in the lowest continuity of care quartile had 41% greater odds of being tested than those in the highest quartile (OR 1.41, 95% CI 1.14-1.76). CONCLUSIONS: For adolescents and young women, the likelihood of testing for chlamydia was reduced when care was concentrated with a usual provider. Potential implications for health service delivery are discussed.

Risk factors for urinary tract infections in postmenopausal women.

BACKGROUND: Urinary tract infections (UTI) occur frequently in postmenopausal women and account for substantial morbidity and economic costs. Although extensively studied in younger women and older, debilitated women, the risk factors for UTI among healthy community-dwelling postmenopausal women have not been well described. METHODS: We conducted a population-based, case-control study of women aged between 55 and 75 years enrolled in a large, staff-model health maintenance organization. Cases were identified using computerized laboratory and outpatient records. Controls were randomly selected from the plan's enrollment files. We interviewed subjects regarding their habits, general health, and potential risk factors for UTI. RESULTS: We interviewed 899 study subjects and 911 controls. Sociodemographic characteristics were similar in subjects and controls. Most women were insured and white. Like younger women, postmenopausal women with current UTI were more likely to be sexually active (odds ratio [OR], 1.42; 95% confidence interval [CI], 1.07-1.87) and have a history of UTI (OR, 4.20; 95% CI, 3.25-5.42). Like older debilitated women, study subjects were more likely to have diabetes mellitus (OR, 2.78; 95% CI, 1.78-4.35) and to be incontinent (OR, 1.36; 95% CI, 1.03-1.78). Oral estrogen replacement did not reduce UTI risk. CONCLUSIONS: In this population, the risk factors of healthy community-dwelling postmenopausal women reflect the health status of women as they transition toward old age. Sexual activity, history of UTI, treated diabetes, and incontinence were all associated with a higher risk of UTI. The therapeutic role of oral estrogen remains uncertain. Prospective studies in different patient populations are needed to better understand the risk factors of UTI.