Background: To prevent congenital syphilis, the Centers for Disease Control and Prevention and professional organizations recommend universal prenatal syphilis screening. State-level or larger-scale evaluations of adherence to these guidelines have relied on administrative data. We measured prenatal syphilis screening rates in Indiana women with prenatal Medicaid coverage and also used electronic medical records to examine the completeness of syphilis screening claims in Medicaid administrative data.
Methods: In statewide Indiana Medicaid claims data, diagnosis and procedure codes were used to identify women who delivered an infant between October 1, 1998, and September 30, 2002. Claims for prenatal (that is, during the 40 weeks before and including the delivery date) syphilis screens, including the “obstetric panel” of tests, and for prenatal visits were extracted. A subset of the study population received prenatal care in a large public hospital and its affiliated clinics served by an electronic medical records system. For these women, claims data were compared with laboratory reports.
Results: Among 74,188 women with one delivery in Medicaid claims data, 60% had at least 1 prenatal syphilis screening claim, and 15% had 2 or more. Women with continuous Medicaid enrollment during pregnancy or with at least one prenatal visit claim had higher rates. Among the 3960 women for whom Medicaid claims and laboratory data were available, 49.8% had at least one prenatal syphilis screen in Medicaid claims, but 99.3% had at least one laboratory report of a syphilis screen.
Conclusions: Measurements made using Medicaid administrative data appear to substantially underestimate true prenatal syphilis screening rates.
This study linked Indiana Medicaid administrative data with electronic medical records data and found that measurements made using Medicaid administrative data appear to substantially underestimate true prenatal syphilis screening rates.
From the *Department of Pediatrics, Indiana University School of Medicine; †Regenstrief Institute, Indianapolis, Indiana; ‡Division of STD Prevention, Centers for Disease Control and Prevention, Atlanta, Georgia; and Departments of §Epidemiology and ∥Pediatrics, Boston University Schools of Public Health and Medicine, Boston, Massachusetts
This study was funded by the Centers for Disease Control and Prevention and also was supported by the Association of American Medical Colleges (Cooperative Agreement No. MM-0710).
The findings and conclusions in this report are those of the authors and do not necessarily represent the Centers for Disease Control and Prevention or the Association of American Medical Colleges.
Barbara E. Mahon, MD, MPH, is currently at Novartis Vaccines, Cambridge, Massachusetts.
Correspondence: Marc Rosenman, MD, 410 W. 10th Street, Suite 2000, Indianapolis, IN 46202-3012. E-mail: email@example.com.
Received for publication August 17, 2007, and accepted October 12, 2007.
CONGENITAL SYPHILIS IS A DEVASTATING disease, which can be prevented by timely prenatal screening and treatment. Syphilis during pregnancy can lead to premature birth, spontaneous abortion, stillbirth, or perinatal death; it also increases the transmissibility of human immunodeficiency virus.1 In congenitally infected infants who survive the neonatal period, the consequences can include severe neurodevelopmental retardation, deafness, seizures, and skeletal deformities. Because symptoms can be minimal or absent in early stages of syphilis, screening is central to efforts to control the disease.
The Centers for Disease Control and Prevention, the US Preventive Services Task Force, and professional organizations have recommended universal prenatal screening.2–5 However, surveys of clinicians,6,7 hospitals,8 and jails,9 reviews of prenatal records from urban hospitals10 and from a health plan,11 and studies of pregnancies that led to congenital syphilis12–19 demonstrate that prenatal screening for syphilis and other sexually transmitted diseases often does not occur as recommended. Inadequate prenatal screening has been associated with outbreaks of congenital syphilis,12,20,21 whereas improved rates of prenatal screening may lower rates of congenital syphilis or its sequelae.22–24
Recent studies using large administrative databases have reported prenatal syphilis screening rates of only about 60%. In an analysis of Florida Medicaid data, 57% of women in 2000 (compared with 26% in 1995) had claims for syphilis screening.25 A multistate study of administrative data for women continuously enrolled in a commercial health plan during pregnancy found a prenatal or delivery syphilis screening claim in 63% of 13,250 pregnancies; the authors speculated that this lower-than-expected rate might reflect incomplete capture of syphilis screening in claims data.26 Analyses of clinical records have shown much higher screening rates. A cohort study of 550 pregnancies in Philadelphia27 found that 98.7% of women had at least one prenatal syphilis screen. A multistate review of approximately 5000 labor and delivery records found a syphilis screening rate of 98.4%.28 However, a review of hospital records for almost 2000 pregnant women in Miami-Dade County found that only 83% had prenatal syphilis screening.10
To examine prenatal syphilis screening practices in the Indiana Medicaid population, this study used both the statewide Medicaid claims database and an electronic medical records system for an Indianapolis public hospital and its clinics. Linking these 2 data sources for a subset of the state’s population enabled direct comparison of the capture of syphilis screening data in claims versus in electronic medical records.
Materials and Methods
The study was approved by the Institutional Review Boards of the Centers for Disease Control and Prevention and of Indiana University and Purdue University, Indianapolis. Women in the Indiana Medicaid database who delivered an infant between October 1, 1998, and September 30, 2002, were eligible for the study.
Analysis of Indiana Medicaid Data
The National Committee for Quality Assurance has developed a list of delivery-related diagnosis and procedure codes for the Health Plan Employer Data and Information Set (HEDIS) measure for timeliness of prenatal care.29 The list of prenatal care codes in the HEDIS measure was compared with Indiana Medicaid claims data to identify women who had delivered an infant. To simplify the interpretation of the data, the study was limited to women with one unambiguous delivery in the Indiana Medicaid database between October 1, 1998 and September 30, 2002; women were excluded if they had delivery codes more than 31 days apart during the study period. For women with more than one delivery code within 31 days, the latest of these dates was assumed to be the delivery date. Women under 15 or more than 50 years old at delivery were excluded.
The occurrence of syphilis screening was based on an Indiana Medicaid claim with any of the following procedure codes: 86592 (qualitative RPR), 86593 (quantitative RPR), 80055 (obstetric panel, which includes blood type, complete blood count, hepatitis B surface antigen test, Rubella antibody test, and qualitative rapid plasma reagin), or 86781 (confirmatory test for Treponema pallidum antibody). Routine prenatal visits were defined using the list of diagnosis and procedure codes in the HEDIS 2004 measure for timeliness of prenatal care. Because gestational age at birth is not recorded in Indiana Medicaid administrative data, all pregnancies were assumed to be full term, lasting 281 days—with the first trimester as the first 94 days, the second trimester as the next 93 days, and the third trimester as the next 94 days (the delivery date was the 94th day of third trimester). Claims for syphilis screening or prenatal visits were examined during these 281 days. The “labor-and-delivery” period was approximated as the 40th week of pregnancy—the week before delivery, plus the delivery date (8 days). The reason for analyzing the 40th week separately was that a high proportion of syphilis screens during the final week of pregnancy were thought to occur in inpatient settings, where Medicaid claims may be reimbursed at a standard daily rate rather than based on itemized services such as laboratory tests.25
Continuity of enrollment in Medicaid was measured by counting the number of calendar months that each woman was in Medicaid during the prenatal period. “Continuous” enrollment was defined as 9 to 10 calendar (prenatal) months in Medicaid; short enrollment gaps within a calendar month could not be discerned.
Analysis of Electronic Medical Records Data
A subset of women in the statewide Indiana Medicaid cohort receives health care in an Indianapolis public hospital—Wishard Memorial Hospital—and its clinics, jointly termed Wishard Health Services (WHS). WHS is served by the Regenstrief Medical Records System (RMRS), a state-of-the-art electronic medical records system that includes laboratory results, medications, orders, radiology reports, discharge summaries, registration information from inpatient and outpatient visits, and other data. The RMRS also contains laboratory and visit data for 3 nearby hospitals affiliated with Indiana University School of Medicine.30 To identify women with a delivery at Wishard Memorial Hospital during the study period, the WHS RMRS was searched for the HEDIS 2004 diagnosis and procedure codes for delivery that were used in the Indiana Medicaid data query, or for the presence of dates in the RMRS fields “delivery date” or “estimated date of confinement.” The RMRS delivery date for this study was defined as the value of the RMRS “delivery date” field.
Exact matching of Medicaid recipient identification numbers or social security numbers was used to link RMRS data to Medicaid data. After matching, delivery dates in the 2 systems were compared. Among women with one RMRS delivery during the study period, 95.4% had an RMRS delivery date within 30 days of the Medicaid claims delivery date; based on this observation, pregnancies were excluded from the comparative analysis if the RMRS delivery date was not within 30 days of the Medicaid claims delivery date. Some women with one delivery in Medicaid claims data had more than one delivery in RMRS during the study period. These women—one pregnancy for each—were included in the comparative analysis if they had a delivery date in RMRS within 30 days of the delivery date in Medicaid claims data. For women with a matching delivery in both the Indiana Medicaid data and the WHS RMRS data, the dates of all laboratory tests for syphilis were extracted from the RMRS from the 281 days up to and including the delivery date (whichever of the 2 delivery dates was later). Other variables extracted were gestational age at birth, birth weight, gravidity and parity, dates of WHS prenatal visits, and insurance status at those visits. Insurance status was analyzed to examine whether women in the statewide cohort might have received care at visits not covered by Medicaid, for which Medicaid claims would not have been generated.
Syphilis screening rates measured in Medicaid claims versus in RMRS were compared in several ways. First, the simple proportion of women with at least one prenatal (that is, during the 40 weeks before and including the delivery date) syphilis screen in each system was examined. Second, the prenatal period was divided into the first 39 weeks (273 days) and the 40th week—the “labor and delivery” period (8 days)—to compare the proportion of women with at least one syphilis screen in these time periods. These 2 comparisons were then repeated for a narrower subset of women—those who had 9 or 10 calendar months of prenatal Medicaid enrollment and whose Medicaid claims and RMRS delivery dates were matched within 2 days. In both Medicaid claims and RMRS data, additional analyses were conducted to stratify by number of months of prenatal Medicaid coverage and to analyze how prenatal syphilis screening rates varied with routine prenatal visits by trimester. The RMRS cohort was also analyzed for evidence of other forms of insurance during pregnancy.
Analysis of Indiana Medicaid Data
In total, 89,907 women had at least one Indiana Medicaid claim for delivery between October 1, 1998 and September 30, 2002. From these, 14,767 women who had 2 or more delivery codes more than 31 days apart during the study period and 952 women who were under 15 years or over 50 years old were excluded, leaving 74,188 women in the study. Their median age at delivery was 22.9 years; 71% were white, 15% black, and 12% Latino. Fifty-five percent were continuously enrolled in Indiana Medicaid during the entire prenatal period.
At least one prenatal (as defined above) syphilis screening claim was present in Indiana Medicaid data for 44,178 (60%) of the women: 32,890 (44%) had 1 prenatal syphilis screening claim; 10,711 (14%) had 2; and 577 (1%) had 3 or more. At least 1 prenatal visit claim was present for 58,811 (79%) of the women; 22,882 (31%) had a claim in the first trimester, 44,560 (60%) in the second trimester, and 57,064 (77%) in the third trimester. Women with a prenatal visit claim were more likely than women without a prenatal visit claim to have a prenatal syphilis screening claim (69% vs. 22%, respectively).
Analysis of Electronic Medical Records Data
Among the 74,188 women described above, 3796 of them had 1 delivery in the WHS RMRS during the study period, 448 had 2, and 19 had 3 (N = 4263 women in total). Comparison of the Medicaid-based and RMRS-based delivery dates for the 3796 women with 1 delivery in both systems showed that 83.4% matched exactly, 92.5% matched within 1 day, and 95.4% matched within 30 days. From the 4263 women with a delivery recorded in RMRS, for the comparison of prenatal syphilis screening in Medicaid claims versus RMRS laboratory data, 181 women were excluded because they did not have an RMRS delivery date within 30 days of the Medicaid claims-based delivery date. An additional 122 women were excluded because they had an RMRS estimated date of confinement or a HEDIS 2004 diagnosis or procedure code for delivery, but no date in the RMRS “delivery date” field.
Thus, there were 3960 women, each with 1 delivery, in the cohort for comparison of Medicaid claims and RMRS laboratory data; their median age at delivery was 23.0 years; 47% were black, 27% Latino, and 23% white. Mean gravidity and parity at delivery were 2.5 and 1.2, respectively. Mean gestational age and birth weight were 38.7 weeks and 3213 g, respectively. Based on Medicaid administrative data, 57% of the 3960 women had a prenatal visit claim. The 3960 women had a total of 26,911 prenatal months in Medicaid (mean 6.8 months). Forty-seven percent of the months involved pregnancy-only Medicaid coverage; 53% involved full Medicaid coverage in various aid categories.
Based on RMRS laboratory data, 3933 (99.3%) of the 3960 women had a prenatal syphilis screen—39% in the first trimester, 46% in the second trimester, and 95% in the third trimester. By comparison, only 1971 (49.8%) of the women had a prenatal syphilis screening claim in Indiana Medicaid data—10% in the first trimester, 14% in the second trimester, and 39% in the third trimester.
The number of prenatal months that a woman was in Medicaid, based on the monthly Medicaid eligibility file, was strongly associated with the prenatal syphilis screening rate measured in Medicaid claims data, whereas the prenatal syphilis screening rate measured in RMRS data was close to 100%, no matter how many prenatal months the woman was in Medicaid (Table 1). Continuous coverage during the first trimester was also associated with the prenatal syphilis screening rate measured in Medicaid claims (Table 2).
Among the 3960 women, 1221 were in Medicaid (based on Medicaid data) for 9 to 10 calendar months during pregnancy and had Medicaid-based and RMRS-based delivery dates within 2 days of each other. For these 1221, syphilis screening screens obtained during the labor and delivery period were much more likely to be recorded in the RMRS than in Medicaid claims data (Table 3).
In visit-level insurance data stored in RMRS, 74.5% of the 3960 women had at least one RMRS outpatient encounter covered by Medicaid during pregnancy, and an additional 15.9% had at least one RMRS outpatient encounter with “Medicaid pending” as the coverage status. It was also observed that many of the women had prenatal health care encounters that were covered by insurers other than Medicaid or perhaps were not covered by insurance at all. For the cohort of 1221 women, Table 4 shows, for each trimester, the distribution of primary insurance for each encounter, in 4 insurance categories.
In this study, Indiana Medicaid administrative data appear to substantially underrepresent true prenatal syphilis screening rates. Only 60% of pregnant women in the statewide Medicaid cohort and only 50% of the subset who delivered at WHS had a claim for prenatal syphilis screening. However, more than 99% of women who delivered at WHS had a laboratory report for prenatal syphilis screening, which in all likelihood indicates that they were, in fact, screened. Two purposes of the study were to use Medicaid claims to describe prenatal syphilis screening practices across the state and to construct a predictive model for timely screening. These goals were abandoned once the low rate of capture of prenatal syphilis screening in Medicaid claims was discovered. The focus then became another of the a priori goals—to use electronic medical records to help detail the shortcomings of administrative data as a measurement method.
The results of this study help reconcile the disparate findings of previous studies that used different approaches to measure prenatal syphilis screening rates. The 60% prenatal syphilis screening rate measured with Indiana Medicaid data are similar to rates measured in 2 recent studies of administrative databases—57% in Florida Medicaid25 and 63% in multistate women continuously enrolled in a commercial health plan.26 The 99% prenatal syphilis screening rate measured using electronic medical records for a large Indianapolis cohort is consistent with recent studies of clinical records in Philadelphia27 and in a multistate sample.28 Because women are frequently screened for syphilis at delivery, whether the prenatal period is defined as including labor and delivery has the potential to greatly affect screening rates. A study of medical records in Miami-Dade County found an 83% syphilis screening rate during prenatal care before labor and delivery, and a 93% syphilis screening rate during labor and delivery; overall, 98.8% of women had at least one record of a syphilis screen during pregnancy.10 The rates measured in the Miami-Dade County study might reflect suboptimal adherence to screening guidelines or, as the authors suggest, might reflect incomplete data.10 Some administrative data studies, such as in Florida Medicaid data,25 reflect screening claims from outpatient visits only, whereas we analyzed screens from both inpatient and outpatient visits during the entire 40-week period ending with the delivery date.
Continuity of prenatal health insurance coverage and timeliness of routine prenatal care are valuable in their own right and also are associated with adherence to the recommended prenatal schedules for sexually transmitted disease screening,24,26,31 and both were associated with higher prenatal syphilis screening rates in Indiana Medicaid claims data in this study (72% of women continuously enrolled in Medicaid had a prenatal syphilis screen, vs. 5% of women with 1 prenatal month in Medicaid). But in studies that rely on administrative data alone, periods of coverage in a plan are the only periods during which claims can be generated and outcome(s) ascertained. Perhaps because many women who receive prenatal care in WHS have limited or no access to other systems of prenatal care but have good access to WHS, the linked administrative and clinical data in this study showed that the true prenatal syphilis screening rate was not associated with continuity of prenatal Medicaid coverage. Almost all women had at least 1 prenatal syphilis screen no matter how many prenatal months they were in Medicaid.
By matching Medicaid claims and electronic medical records data at the individual level, the study was able to examine when and under what circumstances administrative data are most likely to underestimate actual screening rates. Although Medicaid claims underestimated the syphilis screening rate throughout pregnancy, they were especially unlikely to capture syphilis screening during labor and delivery—18% of women had a claim then, whereas 79% of the same women had a laboratory report for syphilis screening during labor and delivery. The especially low proportion of claims for screening performed during labor and delivery probably reflects the way that these inpatient stays are reimbursed—when reimbursement is not based on a highly itemized list of services provided, the presence or absence of a syphilis screening claim would not alter the payment that a hospital receives from Medicaid. The dearth of syphilis screening claims during the preceding 39 weeks might reflect several factors. Lack of significant financial incentive to bill might play a role here, too. Sometimes, routine prenatal services throughout pregnancy are billed at one all-inclusive price; in other situations clinicians or delivery systems are reimbursed under a Medicaid risk-based managed care (capitated) arrangement.
The analyses suggest an additional explanation for having prenatal syphilis screening recorded in clinical but not in Medicaid data: for many women with Medicaid coverage, a complex and varied set of payers is actually involved. Table 2 shows that, based on Medicaid administrative (monthly eligibility) files, 25% of women statewide (18,765 of 74,174) had no first trimester Medicaid coverage; this rate was as high as 44% (1758 of 3960) among the women in the RMRS subset. Still, even among the N = 1221 RMRS subset of women listed in Indiana Medicaid eligibility files in 9 to 10 calendar months before delivery, the RMRS-stored insurance fields revealed that at many encounters (19%, 12%, and 7% of encounters in the first, second, and third trimester, respectively), women registered for care with a primary insurance other than Medicaid. Some had county-financed health insurance; others had commercial or private insurance, and still others received care under the auspices of the prison system. For a larger percentage of encounters in the first and second trimesters, the RMRS insurance data were missing; missing values probably represent some uninsured women and others for whom RMRS simply did not record the insurance status. Medicaid might have covered some of these visits retroactively. In this public hospital and clinic system, Medicaid or county-financed coverage is often applied for at a visit, with the help of a financial counselor.
The matching of women with deliveries in Medicaid data and RMRS data proved to have high specificity. Undoubtedly the sensitivity of this matching was less impressive, for several reasons. For simplicity, this study included in the Medicaid claims data cohort only those women with a single unambiguous delivery during the study period. In addition, if Medicaid claims data in the study were incomplete not only for syphilis screening but also for delivery of an infant, the denominator of pregnant women would have been further underestimated. Clinical data in RMRS (e.g., estimated date of confinement based on ultrasound or last menstrual period) afford more opportunities than claims data to identify pregnant women. Still, even if there was incomplete inclusion or matching of the women with a delivery in both Medicaid and RMRS, the conclusions about measuring prenatal syphilis screening are based on individually matched data and would still hold. Syphilis screening rates for pregnant women whose data could not be matched would probably be similarly low in claims and high in RMRS data.
Although the 60% prenatal syphilis screening rate measured in Medicaid claims data are likely very much an underestimate, the 99% prenatal syphilis screening rate measured in WHS does not necessarily reflect medical practice across Indiana. Indianapolis was the site of a syphilis outbreak that began in 1998, peaked in 1999 with a reported annual incidence of 50 per 100,000 persons—compared with a national rate of 2.5 per 100,000 persons32—and ended by 2001. The outbreak elicited a multifaceted public health response,33 which included recommendations for intensified prenatal syphilis screening2 and a series of corresponding interventions at WHS and elsewhere. The US Preventive Services Task Force recommends “routine serologic testing for syphilis … for all pregnant women. All pregnant women should be tested at their first prenatal visit. For women at high risk of acquiring syphilis during pregnancy, repeat serologic testing is recommended in the third trimester and at delivery.”3 The American Academy of Pediatrics recommends that syphilis screening be conducted “early in pregnancy … and preferably again at delivery. In areas of high prevalence of syphilis and in patients considered at high risk of syphilis, a nontreponemal serum test at the beginning of the third trimester (28 weeks of gestation) and at delivery is indicated.”34
In response to the syphilis outbreak in Indianapolis in 1999, the Marion County Health Department encouraged 3 prenatal syphilis screens for women considered at high risk (including many women in the WHS population). WHS is also affiliated with Indiana University School of Medicine and its residency training programs. For these reasons, compared with other systems in Indiana, WHS might have above-average prenatal syphilis screening rates. Clinician adherence to screening guidelines might also be above average in health systems that deliver care to populations with high risk or high prevalence of syphilis.
Administrative data are widely used in health care research and have noteworthy advantages as well as disadvantages. Advantages include efficiency and unobtrusiveness in analysis, given large populations distributed across many providers, clinics, and delivery systems, and standardized diagnosis and procedure codes. Pharmacy claims data document prescriptions and refills dispensed at any pharmacy. Eligibility files reflect members’ coverage status even during months in which the member had no health care utilization. (In this study, RMRS insurance data were available only when patients registered at a health care encounter.) Although electronic medical records have “richer” data (such as the results of laboratory and radiology studies), the key data elements can be lacking. (In this study, current procedural terminology codes were available in Indiana Medicaid data but not in RMRS.) The principal disadvantage of administrative data are incompleteness; not all tests, procedures, and encounters get billed, coded, processed, or stored (or stored correctly).35,36 In addition, standard “rules” that use administrative data to define an illness or an episode of care (such as the HEDIS definition of a prenatal care visit), are themselves necessarily approximations of reality, approximations with a “sensitivity” and a “specificity.” A query for HEDIS-defined prenatal care visits misses some true prenatal care visits. Given the strengths and weaknesses of both medical records and administrative data,35,36 it can be desirable to augment one with the other. Risk-adjustment algorithms, historically claims-based, are increasingly being supplemented with clinical data.37 The development of electronic medical records systems is informed by an understanding of the limitations for research of (their component) administrative data.38 At the same time, new sources of administrative data are valuable additions to an electronic medical records system, for research but also for clinical purposes such as point-of-care delivery of medication lists. Indiana Medicaid data are being built into the local health information infrastructure (the next generation RMRS) in Indianapolis.39 Pay-for-performance initiatives also benefit from analyzing both types of data.
The results of this study help document the shortcomings of administrative data as a method of measuring prenatal syphilis screening rates. An electronic medical records system documents prenatal syphilis screening much more completely but also has its own shortcomings. Future analyses, using electronic medical records data or data from more than one source, may demonstrate that even when there is already good adherence to guidelines, improvements in prenatal syphilis screening practices still can be made that will further the goal of eliminating prenatal and congenital syphilis.
1. Mwapasa V, Rogerson SJ, Kwiek JJ, et al. Maternal syphilis infection is associated with increased risk of mother-to-child transmission of HIV in Malawi. AIDS 2006; 20:1869–1877.
2. CDC. Sexually transmitted diseases treatment guidelines 2002. MMWR Recomm Rep 2002; 51(RR-6):1–78.
3. U.S. Preventive Services Task Force. Recommendation statement: screening for syphilis infection. November 15, 2004. Available at: http://www.aafp.org/afp/20041115/us.html
. Accessed November 12, 2006.
4. American Academy of Pediatrics and American College of Obstetricians and Gynecologists. Guidelines for Perinatal Care, 5th ed. Washington DC: American Academy of Pediatrics and American College of Obstetricians and Gynecologists 2002.
6. Hogben M, St. Lawrence JS, Kasprzyk D, et al. Sexually transmitted disease screening by United States obstetricians and gynecologists. Obset Gynecol 2002; 100:801–807.
7. St. Lawrence JS, Montano DE, Kasprzyk D, et al. STD screening, testing, case reporting, and clinical and partner notification practices: a national survey of US physicians. Am J Public Health 2002; 92:1784–1788.
8. Martin DL, Bertrand JR, McKegney CP, et al. Congenital syphilis surveillance and newborn evaluation in a low-incidence state. Arch Pediatr Adolesc Med 2001; 155:140–144.
9. Parece MS, Herrera GA, Voigt RF, et al. STD testing policies and practices in US city and county jails. Sex Transm Dis 1999; 26:431–437.
10. Trepka MJ, Bloom SA, Zhang G, et al. Inadequate syphilis screening among women with prenatal care in a community with a high syphilis incidence. Sex Transm Dis 2006; 33:670–674.
11. Gazmararian JA, Cogswell ME, Koplan JP. The relationship between characteristics of obstetrician-gynecologists and rates of recommended prenatal care screening tests. HMO Pract 1996; 10:108–113.
12. Southwick KL, Guidry HM, Weldon MM, et al. An epidemic of congenital syphilis in Jefferson County, Texas, 1994–1995: Inadequate prenatal syphilis testing after an outbreak in adults. Am J Public Health 1999; 89:557–560.
13. Warner L, Rochat RW, Fichtner RR, et al. Missed opportunities for congenital syphilis prevention in an urban southeastern hospital. Sex Transm Dis 2001; 28:92–98.
14. Sheffield JS, Sanchez PJ, Morris G, et al. Congenital syphilis after maternal treatment for syphilis during pregnancy. Am J Obstet Gynecol 2002; 186:569–573.
15. CDC. Epidemic of congenital syphilis—Baltimore, 1996–1997. MMWR Morb Mortal Wkly Rep 1998; 904–907.
16. Risser WL, Hwang L-Y. Congenital syphilis in Harris County, Texas, USA, 1990–92: Incidence, causes, and risk factors. Int J STD AIDS 1997; 8:95–101.
17. Webber MP, Lambert G, Bateman DA, et al. Maternal risk factors for congenital syphilis: A case-control study. Am J Epidemiol 1993; 137:415–422.
18. McFarlin BL. Bottoms SF, Dock BS, et al. Epidemic syphilis: maternal factors associated with congenital infection. Am J Obset Gynecol 1994; 170:535–540.
19. Schulte JM, Burkham S, Hamaker D, et al. Syphilis among HIV-infected mothers and their infants in Texas from 1988 to 1994. Sex Transm Dis 2001; 28:315–320.
20. Outbreak of primary and secondary syphilis—Guilford County, North Carolina, 1996–1997. MMWR Morb Mortal Wkly Rep 1998; 47:1070–1073.
21. Hibbs JR, Ceglowski WS, Goldberg M, et al. Emergency department-based surveillance for syphilis during an outbreak in Philadelphia. Ann Emerg Med 1993; 22:1286–1290.
22. CDC. Congenital syphilis—United States, 2000. MMWR Morb Mortal Wkly Rep 2001; 50:573–577.
23. Beltrami J, Berman S. Congenital syphilis: A persisting sentinel public health event. Sex Transm Dis 2006; 33:675–676.
24. Gust DA, Levine WC, St. Louis ME, et al. Mortality associated with congenital syphilis in the United States, 1992–1998. Pediatrics 2002; 109:e79.
25. Wilson EK, Gavin NI, Adams K, et al. Patterns in prenatal syphilis screening among Florida Medicaid enrollees. Sex Transm Dis 2007; 34:378–383.
26. Tao G, Patterson E, Lee LM, et al. Estimating prenatal syphilis and HIV screening rates for commercially insured women. Am J Prev Med 2005; 28:175–181.
27. Jessop AB, Watson B, Mazar R, et al. Assessment of screening, treatment, and prevention of perinatal infections in the Philadelphia birth cohort. Am J Med Qual 2005; 20:253–261.
28. Schrag SJ, Arnold KE, Mohle-Boetani JC, et al. Prenatal screening for infectious diseases and opportunities for prevention. Obstet Gynecol 2003; 102:753–760.
29. National Committee for Quality Assurance. HEDIS 2004: Technical Specifications. Washington DC: National Committee for Quality Assurance, 2003.
30. McDonald CJ, Overhage JM, Tierney WM, et al. The Regenstrief medical records system: A quarter century experience. Int J Med Inform 1999; 54:225–253.
31. Hessol NA, Vittinghoff E, Fuentes-Afflick E. Reduced risk of inadequate prenatal care in the era after Medicaid expansions in California. Med Care 2004; 42:416–422.
32. Primary and secondary syphilis—United States, 1999, MMWR Morb Mortal Wkly Rep 2001; 50:113–117.
33. Rosenman MB, Kraft SK, Harezlak J, et al. Syphilis testing in association with gonorrhea/chlamydia testing during a syphilis outbreak. Am J Public Health 2004; 94:1124–1126.
34. American Academy of Pediatrics. Syphilis. In: Pickering LK, Baker CJ, Long SS, et al., eds. Red Book: 2006 Report of the Committee on Infectious Diseases, 27th ed. Elk Grove Village, IL: American Academy of Pediatrics, 2006.
35. Steinwachs DM, Stuart ME, Scholle S, et al. A comparison of ambulatory Medicaid claims to medical records: A reliability assessment. Am J Med Qual 1998; 13:63–69.
36. Semroc GN, Tierce JC, Fridman M. Assessing resource use in oncology patients: A comparison of analyses based on claims data and medical chart review. J Manag Care Pharm 2003; 9:6–9.
37. Pine M, Jordan HS, Elixhauser A, et al. Enhancement of claims data to improve risk adjustment of hospital mortality. JAMA 2007; 297:71–76.
38. Weiner MG, Lyman JA, Murphy S, et al. Electronic health records: High-quality electronic data for higher-quality clinical research. Inform Primary Care 2007; 15:121–127.
© Copyright 2008 American Sexually Transmitted Diseases Association
39. McDonald CJ, Overhage JM, Barnes M, et al. The Indiana Network for Patient Care: A working local health information infrastructure. Health Affairs 2005; 24:1214–1220.