Syphilis testing adherence among women with livebirth deliveries: Indianapolis 2014-2016

Congenital syphilis (CS) results from infection in infants in utero with the bacterium, Treponema pallidum [1]. Untreated, pregnant women with syphilis are at high risk of spontaneous fetal abortion, fetal death, and intrauterine growth retardation [1]. A fetus infected with syphilis can develop severe physical and neurological sequelae, including cardiac malformation, skeletal deformities, blindness, deafness, Hutchinson triad disorder, and hepatosplenomegaly [1]. However, the severity of these adverse health effects depends on the duration of the syphilis infection in the pregnant women and if and when treatment is initiated. As a result, early diagnosis, and treatment of syphilis in all pregnant women is imperative.

Although timely diagnosis and adequate treatment of syphilis infected pregnant women can prevent CS, the number of reported CS cases in the United States has more than tripled since 2013, rising from 362 in 2013 to 1306 in 2018, the highest number of cases in the past twenty years [2]. Similarly, CS has increased in Indiana, rising from zero cases between 2008 and 2013 to 30 cases between 2014 and 2018 [3]. Specifically in Marion County, the state’s most populous county, CS cases increased from zero in 2008 through 2015 to nine cases in 2016 through 2017 [4].

The increase seen in CS among infants is correlated with the increase in syphilis rate seen among women of reproductive age [2]. The Centers for Disease Control and Prevention (CDC) recommends screening all pregnant women for syphilis during their first prenatal visit as this is critical in CS prevention [5]. Women who engage in risky sexual behaviors, use illicit drugs, have STIs (sexually transmitted infections) or multiple sex partners during pregnancy or reside in communities with increased prevalence of syphilis infection are considered high risk and recommended to be screened again between 28 to 32 weeks and also at delivery [5]. These recommendations are supported by the American Congress of Obstetricians and Gynecologists (ACOG). Most states, including Indiana, require screening during the first trimester and subsequent screening in the third trimester if the women are considered high risk [6].

Despite established guidelines, pregnant women may still not be tested for syphilis. The CDC found that in 2018, 28.2% of mothers of infants with CS did not receive timely prenatal care. Of those women that did, 8.9% were not screened appropriately [7]. Late serologic diagnosis was not uncommon as 11.2% of women with CS infants had a negative test early in pregnancy and a positive one less than 30 days before or within 90 days after delivery. Alarmingly, 30.7% of women diagnosed with syphilis during pregnancy lacked timely treatment [7]. Studies have also shown that healthcare providers’ adherence to testing guidelines varies significantly from one infectious disease to the other, and that maternal factors such as insurance type and geographical location are factors that contribute to low adherence [8, 9]. Several studies have also noted significant differences in providers’ adherence level based on data source [8, 10‐14]. The use of administrative data such as Medicaid data has shown providers’ adherence to prenatal syphilis testing to be 60% [13] while studies utilizing clinical records have found higher adherence rate over 80% [10, 11]. In a previous study we found that among confirmed stillbirth cases, just over 50% received any syphilis test [17]. In this study, using the same electronic medical record system, we examined providers’ adherence to CDC screening guidelines by examining the proportion of women with a livebirth delivery who received appropriate syphilis screening. Additionally, this study examined the factors associated with syphilis screening and non-screening among women with a livebirth delivery.

We conducted a cross-sectional cohort study on prenatal syphilis testing among women with a livebirth delivery in Marion County, Indiana, between January 1^st, 2014 and December 31^st, 2016. Pregnancies were identified using birth certificate data from the Marion County Public Health Department (MCPHD), which is responsible for collecting vital records on all births in the county. We linked birth certificate data with the mother’s demographic, clinical encounter, insurance, laboratory, census tract, and medical procedure data stored in the Indiana Network for Patient Care (INPC). The INPC, described previously [15], is a statewide network of electronic medical records that includes all major hospitals, laboratories, and many outpatient clinics in Indiana. The INPC is routinely used for health services and public health research [16], including prior studies on STIs [17‐20]. Probabilistic matching techniques, described previously [21, 22], were used to match the mother’s social security number, last name, first name, date of birth, and gender from the birth certificate to her medical records.

For the purpose of this study, all live births with gestation age missing or greater than forty-two weeks were excluded (201 mothers) (See Supplemental Digital Content). Maternal age was restricted to the reproductive years of 15-44 years, 23 mothers were excluded based on this criteria. In addition, women without clinical encounter data, 94 mothers, in the INPC were excluded from the study, because these women lack pertinent data needed for analysis. Race categories were condensed to reflect the main racial strata in Indianapolis. During this investigation, many African Americans were erroneously coded as Native Hawaiians in one health system, although Native Hawaiians account for <.05% of the Indianapolis population [23]. Therefore we combined African Americans and Native Hawaiians into a single Black race category. Additionally, given low proportions in the underlying population, we combined Multiracial and Asian categories into a single Other race category.

The CDC recommends testing pregnant women for syphilis at the diagnosis of pregnancy and additional syphilis testing early in the third trimester and again at delivery for women who are at increased risk [24]. For this study, providers’ adherence to syphilis testing is determined by the timing of the test performed following testing recommendation from the CDC. Because individual-level risk factors were not available from the INPC, we classified women as high-risk using the syphilis prevalence associated with the zip code of the mother’s home address. During the study period, four zip codes consistently had the greatest number of syphilis cases, with prevalence ranging from 56 to 95 cases per 100000 compared to the overall rate in the Indianapolis Metropolitan statistical area (MSA) which ranged from 4.6 to 8.3 cases per 100000 during the study period. Mothers residing in these zip codes were designated high risk. A woman is classified as having syphilis (by serology) if a non-treponemal test and a treponemal test were both positive. In addition, a few cases were also classified as having syphilis by serology, where an IgG was positive with a positive (reactive) confirmatory test (± 7 days) such as an fluorescent treponemal antibody absorption (FTA-ABS) or T pallidum particle agglutination (TP-PA), even without a non-treponemal test.

The unit of analysis for this study is pregnancy; therefore, deliveries with multiple births (twins, triplets, etc.), were counted as one pregnancy. Additionally, two pregnancies in a year by the same mother was counted as two unique pregnancies. The conception date of each pregnancy was calculated by using the estimated gestational day and birthdate of the baby provided on the birth certificate. The trimester of the infant’s delivery was calculated using the estimated gestational days of the infant. We deduced the trimester of each lab test by taking the numbers of days from conception to the test date. A laboratory test was classified to be within first trimester if it was done at 13 weeks or earlier, second trimester if it was performed between 14 to 27 weeks, and third trimester as lab tests performed between 28 weeks and 42 weeks. Tests performed at delivery were defined as lab tests carried out seventy-two hours before or after the birth. To avoid overlapping of timelines, third trimester lab tests were classified as lab tests performed between 28 weeks and 72 hours before the infant’s birthdate.

Descriptive statistics were calculated to examine the characteristics of mothers with a live birth delivery. Testing rates were calculated by trimester, 30 days before delivery, and at delivery. Results were stratified by year, high risk status, testing status, and mothers with syphilis based on serology. Statistics included mean and standard deviation for continuous variables and frequency and proportion for categorical variables. To evaluate the association between non-testing of women during pregnancy while adjusting for potential confounders, we constructed a multiple logistic regression model with race, age, delivery status (preterm or full-term), risk status, and insurance. The statistical software SAS, version 9.4, (SAS Institute Cary, NC) was used for all analyses, and a P-value < 0.05 was used for the level of significance. The study received approval by the Institutional Review Board at Indiana University (Study No. 1311659626 and 1611016230).

Prenatal syphilis testing is an essential preventative measure to halt the transmission of maternal syphilis infection to the fetus, which can lead to congenital syphilis in infants and stillbirths [1]. In this study, providers tested 82% of pregnancies that resulted in live births from 2014 to 2016. Providers’ adherence increased over time from 71.1% in 2014 to 86.6% in 2016.

Although women who were classified as high risk were screened more frequently, adherence to screening at 28 weeks and delivery was poor. In agreement with previous studies, we found that those with government insurance such as Medicaid and Medicare were tested more frequently than women with commercial insurance [25].

Providers’ adherence in this study was lower than previous studies [8‐10] including one from Indianapolis [13] (95-98.2%). These studies were published over 15 years ago. Although there were differences in study populations, it is also possible that providers decreased testing practice after CDC’s efforts to eliminate syphilis beginning in 1999 [26].

We found that 19% of women were not tested for syphilis at any time during pregnancy. These are missed opportunities to detect and treat syphilis infections. A Louisiana study examining opportunities to prevent CS found that one-third of CS cases could have been prevented if prenatal syphilis testing was done [27]. Another study found that over one-third of women who had a stillbirth were not screened for syphilis [17, 28]. Diagnoses must be followed by timely treatment. However, Kidd et al. (2018) found that most infants with congenital syphilis were born to mothers who tested positive and were diagnosed with syphilis but not treated [29]. Our study was unable to assess treatment because treatment data is incomplete in the INPC.

In addition, our study also indicates that 41 women had syphilis infection based on serology from 2014 to 2016. In Marion County, during this time, there were 365 cases of early syphilis, of which 24 were women, and 189 late latent syphilis cases, of which 66 were women [4]. In the state, a total of 23 congenital syphilis cases were reported during the study period [3, 30]. Specifically, in Marion County, a total of 5 CS cases were observed between 2014 and 2016, with these cases occurring in 2016. Although we found five pregnancies with serologic diagnosis of syphilis less than thirty days before delivery, we cannot confirm that these are the same CS cases reported because our data were de-identified.

Contrary to Schrag et al. (2003), we found that providers were more likely to test women at higher risk, as defined by zip code of residence, than the overall population [8]. The increase in testing observed within the high-risk population is consistent with the CDC prenatal testing guidance [24]. However, despite the recommendation from the CDC and ACOG, only 27% of high-risk women were tested at delivery, and 9% were tested only at delivery. Alarmingly, our study found that over 7.3% of the women that tested positive for syphilis during the study period were tested only at delivery. These data indicate that providers do not fully adhere to testing recommendations when attending to women that reside in areas with high rates of syphilis. This low compliance with the CDC and ACOG guidelines on high-risk prenatal syphilis testing suggests that more providers education is needed.

Providers were more likely to test women with government-issued insurance, had a preterm delivery, resided in high-risk zip codes, and self-identified as Black. Similarly, Sheikh et al. (2008), found socioeconomic factors such as maternal residence, lack of healthcare insurance, and incarceration to be associated with lack of prenatal syphilis testing [9]. We found that those with government insurance such as Medicaid and Medicare were tested more frequently than women with commercial insurance, which is on par with the 85% testing rate among pregnant women found in a study using MarketScan data [25]. Government-funded health insurance afford women the opportunity to seek and receive care during pregnancy. Women who delivered preterm were also tested more often, possibly because of more frequent follow-up appointments associated with preterm risk. We were, however, unable to assess the reasons for preterm birth.

There were several limitations in this study. First, syphilis testing may have taken place in a lab not covered by the INPC, in which case a woman would appear not to be tested [13]. Additionally, results are dependent on institutions entering data correctly. Another limitation of this study is that we could not ascertain the stages of the syphilis diagnosis. Thus, it is possible that cases were from a previous infection that occurred before pregnancy that might not represent active syphilis. Additionally, the definition of high risk for this study did not encompass the entire definition of high risk as outlined by the CDC which includes individuals with risky sexual behavior, illicit drug use, STIs, or multiple sex partners during pregnancy [5]. Thus, it is possible for some women that were not classified as high risk based on the definition used in this study (high morbidity zip codes) to, in fact, be high risk. Lastly, assigned Logical Observation Identifiers Names and Codes (LOINC) coding to some laboratory tests was unclear as to which syphilis test was performed. Although all LOINCs with positive results were examined manually, it is still possible that misclassification could have occurred.

Prenatal syphilis testing is crucial to reduce infant mortality and morbidity. In comparison to earlier studies, providers’ adherence to prenatal syphilis screening is relatively low. However, our study found an overall increase in testing over time which aligned with CDC’s call to action for providers to increase testing due to increase in congenital cases [31]. It is imperative to increase screening among high-risk populations, particularly at 28 weeks and at delivery. Future studies to examine treatment data and explore the proportion of pregnant women with syphilis that meet CDC’s high-risk definition is warranted.