IBM Watson Health MarketScan commercial claims data from 2016 to 2017 were used for this epidemiological study. The MarketScan commercial claims database consists of approximately 26.1 million covered lives from about 100 large employers [12,13]. The claims data capture patient medical encounters for health services over time, both inpatient and outpatient, demographics, service date and location, diagnostic codes (International Classification of Disease, Tenth Revision) and procedural codes (Physician’s Current Procedural Terminology [CPT] and Healthcare Common Procedure Coding System), pharmaceutical drug use, and other billing-related information. The patient receiving each service is identified by a unique enrollee ID. The commercial claims data include geographic variables, such as region and Metropolitan Statistical Area (MSA) [14]. The data provided are deidentified; therefore, this study is not considered human subject research and does not require institutional review board review.

Using the MarketScan database, women aged 15–24 years who received an LARC insertion (either IUD or implant) during 2017 were identified using International Classification of Disease, Tenth Revision codes (Z30.017, Z30.014, and Z30.430), CPT codes (58300 and 11981), and Healthcare Common Procedure Coding System codes (J7300, J7301, J7296, J7297, J7298, and J7307). We considered the earliest date of diagnosis or procedure for an LARC insertion as the date of LARC insertion for each woman. The analytical sample for women with LARC insertion was limited to those continuously enrolled for ≥12 months before the date of LARC insertion. Using Healthcare Effectiveness Data and Information Set (HEDIS) on chlamydia screening, women were considered as sexually active in the 12 months before the day of LARC insertion if they had diagnosis and procedure codes associated with reproductive or sexual health service visits. The reproductive and sexual health services in the HEDIS measure include cervical cancer screening (Papanicolaou test or smear) or pelvic examination, contraceptives, pregnancy, STI, or infertility services [15]. We used CPT codes to identify chlamydia or gonorrhea tests performed (chlamydia: 87490–87492, 87110, 87810, 86631, 87320, 86632, 87270 and gonorrhea: 87590–87592, 87081, 87077, 87205, and 87850).

We identified two measures of chlamydia and gonorrhea screening: (1) overall annual screening performed anytime during the 12 months before IUD or implant insertion and/or on the day of an IUD or implant insertion; (2) among those not screened in the 12 months before insertion, proportion of them screened on the day of insertion. Some women with overall annual screening might be screened both on the day of insertion and in the 12 months before the day of insertion. We used the MSA variable to identify women who resided in precoded MSA (MSA>0; urban area) or non-MSA (MSA = 0; rural/nonurban area) localities. We categorized women by age groups, MSA, and geographic region and performed bivariate chi-square analysis to evaluate the association between each of our measures of chlamydia and gonorrhea screening and these demographic variables. We used SAS software version 9.4 (SAS Institute, Cary, NC, USA) for all statistical analyses and considered p < .05 to be statistically significant.

We identified 47,918 women aged 15–24 years who had received IUD and implant insertion, respectively, between January 1, 2017, to December 31, 2017, and had ≥12 months of insurance coverage before the date of LARC insertion (Figure 1).Of those women, 37,331 (77.9%) were identified as sexually active based on reproductive or sexual health claims in the past 12 months before the day of LARC insertion.

Overall annual chlamydia and gonorrhea screening rates within 12 months of LARC insertion including the day of insertion for sexually active women aged 15–24 years were 74.3% and 76.9%, respectively. Chlamydia testing rates were significantly higher among IUD initiators than implant initiators (77.8% vs.67.8%; p < .001), and a similar pattern was observed for gonorrhea testing (80.0% for IUD initiators vs. 71.2% for implant initiators) for gonorrhea (Table 1).

Our analysis showed that MSA and region were associated with chlamydia or gonorrhea testing among sexually active women with IUD insertion, and women’s age, MSA, and region were associated with chlamydia and gonorrhea testing among sexually active women with implant insertion (Table 1).

Similarly, chlamydia and gonorrhea screening rates were significantly higher among IUD users (13.2% and 13.4%, respectively) than implant users (6.1% and 6.4%, respectively) on the day of insertion and during the 12 months before the date of LARC insertion among sexually active women (Figure 2).

Among 11,455 sexually active women without chlamydia screening within the 12 months before the date of insertion, only 20.9% of IUD users and 9.2% of implant users were tested for chlamydia on the day of LARC insertion, and these estimates were significantly different (p < .001; Figure 2). Similarly, among 10,223 sexually active women without gonorrhea screening within the 12 months before the date of insertion, only 20.0% of IUD users and 8.9% of implant users were tested for gonorrhea on the day of LARC insertion, and these estimates were significantly different (p < .001; odds ratio: .64, confidence interval: .60–.68).

Approximately three in four sexually active women with LARC insertion received recommended chlamydia and gonorrhea screening in the prior 12 months and/or day of LARC insertion. Screening rates were higher among IUD initiators compared with implant initiators as well as among sexually active women resided in MSA area versus non-MSA area and in northeast region versus west region. The results of our analysis of commercially insured cohort of women aged 15–24 years showed that overall annual chlamydia screening among sexually active women with LARC insertion was higher than the rates reported for sexually active women aged 16–24 years enrolled in commercial plans that report chlamydia measures to HEDIS, where approximately half were screened [16]. Among those without screening within the 12 months before the date of insertion, screening on the day of insertion was significantly higher among IUD users (20.9% and 20.0% for chlamydia and gonorrhea, respectively) than among implant users (9.2% and 8.9% for chlamydia and gonorrhea, respectively). Our findings suggest that providers should provide STI screening for all IUD and implant users rather than IUD users only if they were not screened in the 12 months before the date of LARC insertion based on the recommendation. Our findings highlight that there is a missed opportunity for chlamydia and gonorrhea screening, an integration of family planning and STI services, during the LARC initiation visits, particularly for implant users.

The American College of Obstetrics and Gynecology notes that delivery of comprehensive sexual health care includes assessing sexual history, need for contraceptive service, and recommended screening, including for STIs [17]. Although evidence suggests increased risk of STI among LARC initiators [10,11], delivery of comprehensive sexual health care during family planning visits, such as taking sexual history and providing recommended annual screening for STI, would further improve sexual health among women [18].

There may be several reasons that may explain why providers did not screen young women for chlamydia and gonorrhea during the LARC initiation visits. These reasons may include clinician’s lack of knowledge on patient’s past-year screening status, additional privacy concern for STI/HIV services, clinicians’ prioritizing services in the limited-time encounter, and low reimbursement for STI testing [17,19–22]. For example, a study concluded that parent-insured young women were less likely to receive reproductive health services or chlamydia testing than self-insured women [23]. Further studies may elucidate the reasons why many women were not screened for chlamydia and gonorrhea during the LARC initiation visits if screening in the past 12 months was not performed.

The size of the MarketScan database, although it constitutes a sample of persons who have employer-based health insurance, presents a unique opportunity to longitudinally investigate screening patterns because of its size (i.e., 40 million patients). In addition, it allows for comprehensive assessment of the health outcomes using medical claims data as opposed to self-report to document detailed observation of the event by provider on what, when, and how the medical services rendered. Health care outcomes and services rendered reported on administrative claims data help measure health care quality and adherence to standards of care given to patients [24]. The database has its limitations. The MarketScan databases include commercially insured patients; and therefore, our results may not be representative of women who receive STI screening or LARC services, particularly those who are publicly insured. Many adolescent and young women obtain sexual health services, including contraception and STI screening, through school health centers or publicly funded clinics [25]. Some women may have been offered chlamydia and gonorrhea screening per guidelines but declined testing. Accuracy of the administrative data may have impact on the algorithms that are used to identify sexual activity and LARC insertion. Although the proportion of women aged 15–24 years who were identified as sexually active by HEDIS measure has been shown to be not much different from self-report [16,26], the HEDIS criteria to determine sexually active women may misclassify women’s sexual activity status if women seek contraceptives for menstrual bleeding issues or received Pap test for reasons other than sexual activity or if sexually active women did not have any reproductive health–related visit in the past 12 months [27]. Because of the changes in cervical screening guidelines [4], the cervical screening criteria of HEDIS measure may be an outdated approach to classify sexually active patients. The data also have limited laboratory testing result, limited information on provider specialty, and no information on patient’s race/ethnicity or education level [12,13]. Our analysis was further restricted to those patients with continuous enrollment for ≥12 months before IUD insertion to capture previous year screening service encounters; therefore, a small number of patients with lapses in coverage were excluded. Furthermore, the data lack patient’s sexual history information such as whether in a monogamous relationship. Although such sexual risk factors may influence STI screening and contraceptive counseling practices, CDC STI guidance recommends that all sexually active women aged 15–24 years receive chlamydia and gonorrhea screening annually.

Among sexually active women aged 15–24 years with an LARC insertion, approximately one in four had no documented screening for chlamydia or gonorrhea in the 12 months before or at the health care visit at which they received their LARC insertion. These results suggest missed opportunities for STI screening at the time of LARC insertion: among those with no indication of screening before insertion, less than one in six were screened on the day of insertion. When women present for contraceptive services, providers are encouraged to follow recommended STI screening guidelines and, when indicated, use the opportunity to appropriately screen women for chlamydia and gonorrhea. Future analyses may assess barriers to providing recommended screening, assess annual STI screening during routine preventive services for continuous LARC users, and identify opportunities for improvement of integration of contraceptive and STI services.