BACKGROUND AND OBJECTIVES: Despite recommendations for universal HIV testing during routine prenatal care, maternal HIV status is not always available at the time of delivery, which may lead to missed opportunities for antiretroviral prophylaxis. We completed a quality improvement project focused on increasing the availability of maternal HIV status documentation at our perinatal facility. Our primary aim was to improve documentation rates from 50% to 100% between 2009 and 2013. Our secondary aim was to identify predictors of documentation.
METHODS: After an initial needs assessment, we performed a multidisciplinary quality improvement effort to address lack of HIV documentation in perinatal charts. The interventions included a switch to a verbal-only consent process, a rapid HIV testing protocol, and a simplified newborn admission document. To assess the impact of our intervention, we audited 100 charts per month and formally analyzed a second random sample of 200 charts in the postimplementation phase.
RESULTS: Rates of HIV status documentation improved between 2009 and 2013, from 55.5% to 96.5%. Multivariable models revealed that before our interventions, mothers receiving care at freestanding offices (versus community-based outreach clinics) and those privately insured (versus publicly) were less likely to have HIV status documented. In 2013, neither ambulatory site nor insurance type predicted documentation.
CONCLUSIONS: We demonstrated improvement in maternal HIV status documentation on admission to labor and delivery after implementation of a 3-pronged intervention. Next steps include investigating persistent barriers to achieving universal screening and documentation.
- AZT —
- CDC —
- Centers for Disease Control and Prevention
- CI —
- 95% confidence interval
- OR —
- odds ratio
Despite global advances and proven success of antiretroviral prophylaxis to prevent mother-to-child transmission of HIV, perinatal HIV acquisition remains a major public health issue worldwide. In the United States alone, there were 162 confirmed cases of perinatally acquired HIV in 2010.1 The Centers for Disease Control and Prevention (CDC) estimates that 5000 to 6000 infants are born to women with HIV in the United States every year.2
With effective prophylaxis, the perinatal transmission rate is <2%.3 Thus, the number of expected annual perinatally acquired HIV cases should range from 50 to 120 per year. Because infants at the highest risk of contracting HIV from their mothers are those born to women who do not know they are HIV positive, missed opportunity for prophylaxis is one of the modifiable factors probably contributing to a higher-than-expected transmission rate. To address this problem, several public health and clinical governing bodies have recommended routine, universal HIV screening during pregnancy.4–7 Despite these recommendations, not all women are tested in pregnancy, and therefore their infants remain at risk for perinatally acquired HIV.
In 2009, during the clinical care of newborns at our center, admitting physicians in the NICU and newborn nursery noticed that several infants’ charts lacked documentation of maternal HIV status. We brought the issue to the attention of the physician and nursing leadership of the labor and delivery floor. Together, we decided that a needs assessment and potential improvement effort were warranted. Whether HIV testing had actually occurred, pediatric providers were often unaware of maternal HIV status at the time of delivery. This lack of information was in contrast to nearly universal documentation of other prenatal laboratory results, whose early identification would prompt specific neonatal management, including hepatitis B surface antigen, syphilis (rapid plasma reagin), blood type and antibody, and group B Streptococcus status. To better understand this observed discrepancy, we performed a needs assessment to determine the frequency and predictors of documentation of maternal HIV status in perinatal patient charts. We then initiated a quality improvement project, with the goal of increasing maternal HIV documentation rates to 100% by 2013.
We performed a quality improvement initiative to improve maternal HIV documentation rates at Beth Israel Deaconess Medical Center in Boston, Massachusetts, a perinatal facility with ∼5000 deliveries annually. This study was approved by the institutional review board at Beth Israel Deaconess Medical Center. There were no ethical concerns beyond privacy, which was safeguarded by standard policies and procedures at our institution.
Data Collection for Needs Assessment
Our initial analysis included a chart review of 200 mother–infant pairs from 2009 to determine the frequency of HIV documentation and to identify risk factors for lack of documentation. We randomly selected 100 charts of infants admitted to the newborn nursery and 100 charts of infants admitted to the NICU for at least 24 hours. In cases of multiple gestations, only 1 infant was chosen. For each mother–infant pair, we reviewed all available inpatient and outpatient clinic records related to prenatal care and delivery, thereby using the same documentation accessible to a rounding pediatrician or neonatologist. We defined HIV status as “documented” if HIV status was “positive,” “negative,” or “declined testing.” We collected maternal demographic data including age, parity, race or ethnicity, insurance type (private versus public), use of in vitro fertilization, multiple versus singleton gestation, gestational age, maternal smoking, hepatitis B or C positivity, ambulatory care setting, and obstetric provider group. We dichotomized ambulatory settings into community-based outreach clinics and freestanding offices. Community-based outreach clinics and freestanding office settings differ in that the former care for both insured and uninsured patients and tend to care for women of lower socioeconomic status, whereas the freestanding offices care for women who are insured and of generally higher socioeconomic status. Three different groups of physicians (regardless of setting) deliver infants at our medical center, and we compared these groups with one another also.
After our initial needs assessment that demonstrated low HIV status documentation rates, we initiated a multidisciplinary quality improvement effort to resolve 3 factors we identified as key drivers of lack of documentation (Fig 1): First, we switched to a verbal-only informed consent (compared with the previous requirement for written informed consent) process for testing. This change was in accordance with 2012 state regulations and formal protocol modifications enacted by hospital laboratory leaders. Second, we instituted a rapid HIV testing protocol for mothers with undocumented HIV status at time of admission to labor and delivery (Supplemental Appendix 1), after discussion by obstetric and pathology leaders regarding utility of testing. Third, we simplified the newborn admission documentation of prenatal laboratory tests by including only laboratory results our multidisciplinary group deemed “compulsory” on the form that initiates the newborn’s chart (Supplemental Appendix 2). This step was accomplished via input from the labor and delivery nursing staff in addition to unit leaders educating nursing about the revised form. This education emphasized the expectation that HIV status should be completed for each woman on admission and designated as “positive,” “negative,” or “unknown.” If “unknown,” an expectation was set for offering rapid testing (“rapid HIV sent or declined”), thus empowering nurses to prompt physicians to obtain verbal consent for testing in cases of unknown HIV status.
In August 2012, after implementation of the 3 interventions, we began auditing 100 charts per month to closely monitor HIV documentation practices. The labor and delivery and postpartum medical director was notified of every instance of incomplete documentation, which she then communicated to the primary obstetric provider as an opportunity for improvement. Additionally, during monthly Newborn Nursery and NICU Leadership Committee (NNLC and NLC) Meetings, the chief of neonatology reported the percentage of charts missing HIV status documentation to nursing and physician leaders (15–20 attendees), including staff of the labor and delivery, postpartum, and NICU teams. Results were shared in unit staff meetings and posted in unit sites. In 2013, after allowing 6 months for adoption of our interventions, we again sampled 200 charts in the same fashion as we had done in the preintervention period. We also queried the electronic medical record for delivery admissions during which an HIV test was done during the inpatient stay by year since 2009. Figure 2 depicts the timeline of our study, including policy changes.
We examined unadjusted associations of covariates, such as race or ethnicity, age, insurance type, and ambulatory setting, with HIV status documentation in the preintervention and postintervention periods by using χ2 tests. We then performed multivariable logistic regression to analyze predictors of HIV status documentation. Based on our monthly audits, after the implementation of interventions, we created graphic representation of changes of HIV documentation rates over time (run charts). We also analyzed the rate of HIV testing declinations in the preintervention and postintervention periods to ensure that any improvement in documentation rates was not a result of increased documentation of mothers declining HIV testing.
Our needs assessment demonstrated that in 2009, only 55.5% (111/200) of mother–infant pair charts included documentation of maternal HIV status. There were no differences in age, race, or parity between women whose HIV status was documented and those whose status was not (Table 1). However, publicly (versus privately) insured women and women who received prenatal care in community-based outreach clinics (versus freestanding offices) were significantly less likely to lack HIV status documentation at time of admission (Table 1). In 2009, just 0.3% (14/4626) of women were tested for HIV during their admission, compared with 1.8% (86/4691) in 2013 (Supplemental Table 3). Multivariable logistic regression models revealed that women were less likely to lack HIV status documentation if they were publicly (versus privately) insured (odds ratio [OR] 0.28; 95% confidence interval [CI], 0.10–0.77) or if they were receiving prenatal care from physicians in our largest freestanding office site (compared with other freestanding offices) (OR 0.19; 95% CI, 0.06–0.65) (Table 2).
Demographics of sampled patients from 2009 and 2013 are presented in Table 1. There was no difference in baseline characteristics of patients between the preintervention and post intervention periods (data not shown), with the exception of obstetric provider group (P < .0001) and insurance type (P = .001). In the postintervention period, there were more patients (21%) from the multispecialty physician group, compared with just 11% in the preintervention period, because of shifts in referral patterns. Only 3 women were known to be HIV positive in the entire cohort (2 in 2009 and 1 in 2013).
After implementation of the 3 interventions (verbal-only consent, rapid HIV testing, simplified documentation), documentation rates improved substantially. In 2013, 96.5% (193/200) of records included documentation of maternal HIV status. Monthly audits of HIV status documentation, starting in 2012, reveal that this improvement has been sustained for >2 years (Fig 3). The most recent chart audit (October 2014) revealed 99% documentation. Documentation of declined testing was not significantly different in the 2 eras (17 [8.5%] in 2009 and 23 [11.5%] in 2013) (P = .32). Documentation of HIV test results (excluding declination documentation) increased from 47% to 85% in the preintervention and postintervention periods. Adjustment for confounding variables that changed between the preintervention and postintervention periods (obstetric group and race or ethnicity) did not attenuate the association with women in the postintervention period having much higher odds of documentation (adjusted OR 26.0; 95% CI, 11.1–60.8). In addition, when compared with the documentation rates of other routine prenatal laboratory results, there was a clear rise in HIV documentation rates, compared with high rates of documentation for other key laboratories throughout the study period (Fig 4).
In our perinatal center, we found that HIV antibody test results often were not available at the time of admission in 2009. Incomplete HIV documentation happens for 2 reasons. First, women might not be tested. Second, if testing occurs or a patient declines, the result or documentation of declination might not be transmitted to the inpatient record. After implementation of a multidisciplinary, 3-part quality improvement effort (verbal-only consent, rapid HIV testing, simplified documentation), we observed dramatic improvements in HIV status documentation from 55.5% to 96.5%. This improvement was achieved in the setting of local improvement efforts and secular trends, which included statewide initiatives.
Our findings of low documentation rates for maternal HIV status are consistent with other studies of prenatal laboratory results. For example, in 2006 Kumar8 demonstrated similar findings when charts of newborn infants were reviewed for documentation of maternal risk factors and prenatal screening test results. Physician documentation was found to be incomplete in most charts. Kumar also noted that nearly half of all infants were discharged without documentation of maternal hepatitis B, syphilis, and group B Streptococcus status. Specifically for HIV, 174 out of 400 maternal charts (43.7%) and 154 out of 400 (38%) infant charts lacked HIV documentation. Other similar studies, including a statewide study of Illinois birth hospitals9 and another from the Philadelphia Birth Cohort10 from the early 2000s, also demonstrated low rates of HIV documentation. Although these rates may have improved in the post-2006 CDC recommendation era, follow-up studies are lacking. Even if prenatal HIV testing rates improve, documentation may not. Webber et al11 interviewed 653 HIV-negative women postpartum who were found to be HIV negative when rapid HIV testing was sent, and 63% of women reported having previous HIV testing during the index pregnancy that was not available on admission to labor and delivery.
Massachusetts state law concerning HIV testing remains unusual in the United States. Massachusetts remains the only opt-in state, requiring health care providers to offer HIV testing to the patient and for the patient to actively give permission before testing can occur. This is in contrast to opt-out HIV testing, recommended by the CDC, whereby the health care provider can perform an HIV test after notifying the patient that the test is normally performed, but the patient can elect to decline or defer. As of 2011, all states except for Massachusetts had adopted either opt-out or mandatory HIV testing in pregnancy.12 In February 2012, however, the Massachusetts Department of Public Health released a statement reiterating a recommendation for routine HIV counseling and testing for all pregnant women, without reference to risk profile.13 In July 2012, the state legislature passed a law replacing the requirement of obtaining written informed consent before HIV testing with a verbal-only consent process.14 However, the new law stipulates a requirement for separate, distinct language in the written maternal and newborn care consent for the release and documentation of maternal HIV test results for the newborn’s medical record. Despite these limitations, the adoption of the Massachusetts-approved verbal-only consent process at our hospital probably contributed to higher rates of testing for HIV in pregnancy.
Although it is not possible to know how much of the HIV status documentation improvement resulted from each component of our initiative, the implementation of our streamlined newborn admission document was probably another factor in improving HIV status documentation, by increasing the likelihood of a test result making it into the inpatient record. In the past, the newborn admission document included several optional tests as part of the prenatal diagnostic result section, which were usually filled in only when a result was positive or patient was deemed to be at high risk (Supplemental Appendix 2). We thought that the inclusion of less important clinical data on this form contributed to an inappropriately low prioritization of HIV status. As part of our multidisciplinary team discussions, we reached a consensus about prenatal tests deemed “compulsory” and included only those tests on the form. The identified tests included blood type and antibody status, hepatitis B surface antigen, rubella immunity, syphilis (rapid plasma reagin), group B Streptococcus, and HIV. In addition, we incorporated a prompt for positive, negative, or unknown HIV status. In the case of “unknown” status, the nurse is now prompted to check a box labeled either “rapid testing sent” or “rapid testing declined.” For the tests that were not deemed “compulsory,” such as tuberculosis status, the staff was instructed to include relevant data on the “other maternal conditions” line.
Finally, we speculate that the initiation of a rapid HIV testing protocol on the labor and delivery floor of our hospital improved documentation in 2 ways. First, some women who had no documented testing on admission were tested. However, testing in labor and delivery was responsible for a small portion of increased documentation. Just 0.3% of women were tested in 2009 while admitted for delivery, compared with 1.8% in 2013. Second, in the case where outpatient obstetricians either had been offering testing without sharing documentation of results or had not been testing altogether, HIV testing on the labor floor probably motivated providers to provide results to spare their patients a stressful discussion and counseling on admission to the labor floor. Improvement in HIV status documentation after the initiation of rapid HIV testing is consistent with previous research conducted by Wong et al,15 which demonstrated improvement in statewide HIV documentation in Illinois after the initiation of a rapid HIV testing protocol on admission to labor and delivery.
Strengths of our quality improvement initiative include its pragmatic approach. We had access to the same documentation available to a pediatric provider in the newborn nursery or NICU. Our ongoing monitoring of documentation rates allows us to continuously monitor our progress as we strive for 100% documentation of perinatal HIV status every month. We also demonstrated the benefits of approaching an identified institutional problem area in a multidisciplinary, interdepartmental manner. However, our study does have several limitations. Between 2009 and 2013, secular trends such as improved transparency of electronic medical records linking outpatient charts to inpatient charts and the shift in providers may have influenced the availability of test results and contributed to increasing test rates. We speculate that reduction in the stigmatization of patients with HIV may also have led to increased testing and reporting of results, although we were unable to measure this potential confounding variable. Additionally, our small sample was selected from a low-risk population for HIV infection, and therefore it was not powered to decrease rates of perinatal transmission of HIV. In fact, in all of Massachusetts there are <3 mother-to-child transmissions of HIV per year.16 Another limitation of our study was potential bias that could result from sampling just 100 of the ∼400 deliveries per month for our HIV documentation audits. However, because of random sampling, they were probably representative of our delivery population. The generalizability of our HIV status documentation improvement initiative to other birth hospitals will vary across settings and states because of variation in the connectivity of outpatient and inpatient documentation systems and state laws.
If an increase in declination rates between 2 eras were responsible for the improvement in documentation rates, then no true potential benefits to infants would be experienced. Reassuringly, although there was a slight increase in declination, this difference was not statistically significant, nor was it responsible for the substantial rise in HIV status documentation. The remaining challenge and target for future improvement efforts for our team is to encourage women who decline HIV testing to reconsider testing in an effort to avoid missed opportunities for perinatal prophylaxis to prevent vertical transmission of HIV. We plan to continue our monthly HIV documentation audits for at least the next 2 fiscal years.
Given the complexity of transmitting outpatient test results from several office settings to the inpatient obstetric service and then to the newborn care setting, HIV test result documentation can serve as a quality indicator. We now use perinatal HIV status documentation as a quality metric that we report monthly in our multidisciplinary leadership quality reports.
Our quality improvement initiative led to a significant increase in the documentation of HIV status available to providers at the time of labor and delivery. A multidisciplinary, interdepartmental approach that capitalized on statewide initiatives increased our HIV status documentation from 55.5% to 96.5%. Such efforts will decrease the chance of missing an opportunity to prevent maternal-to-child HIV transmission in the future.
We thank the Beth Israel Deaconess Medical Center Post-Partum/Nursery and Labor and Delivery nursing staff for their input and active participation in the project. We thank Susan Crafts, MS, RN, and Janine Caruso, RN, BSM, MEd, for their dedicated and consistent monthly chart audits since the implementation of our study. We also thank Medgine D. Nau for collecting and organizing the chart audit data. We thank David Miedema for programming the query that randomly selected the medical records to be reviewed. We thank Gary Horowitz, MD, from the clinical laboratories for guiding the electronic medical record query to retrieve inpatient testing results. We thank Munish Gupta, MD, for his guidance and input in creating our key driver diagram.
- Accepted March 6, 2015.
- Address correspondence to Niloufar Paydar-Darian, MD, Boston Children’s Hospital, 300 Longwood Ave, Boston, MA 02115. E-mail:
Dr Paydar-Darian conceptualized and designed the study, performed data collection, and drafted the initial manuscript; Drs Pursley and Golen contributed to conceptualizing and designing the study and reviewed and revised the manuscript; Ms Haviland performed data collection, contributed to data analysis, and reviewed and revised the manuscript; Ms Mao carried out the analyses and reviewed and revised the manuscript; Dr Burris conceptualized and designed the study, including data collection instruments, supervised data analysis, and reviewed and revised the manuscript; and all authors approved the final manuscript as submitted.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Mason SM,
- Kaufman JS,
- Emch ME,
- Hogan VK,
- Savitz DA
- Cooper ER,
- Charurat M,
- Mofenson L,
- et al.,
- Women and Infants’ Transmission Study Group
- Ruiz MS,
- Gable AR,
- Kaplan EH,
- Stoto MA,
- Fineberg HV,
- Trussell J
- Human immunodeficiency virus screening. Joint statement of the American Academy of Pediatrics and the American College of Obstetricians and Gynecologists. Pediatrics. 1999;104(1 pt 1):128pmid:10390276
- Jessop AB,
- Watson B,
- Mazar R,
- Andrel J
- ↵The 188th General Court of the Commonwealth of Massachusetts. An Act Increasing Screening for HIV. Acts of 2012. Chapter 84. Available at: https://malegislature.gov/Laws/SessionLaws/Acts/2012/Chapter84. Accessed October 6, 2014
- ↵Massachusetts Department of Public Health. Massachusetts STD, HIV/AIDS and viral hepatitis surveillance report: 2012. 2013. Available at: www.mass.gov/eohhs/docs/dph/cdc/aids/std-surveillance-2012.pdf. Accessed October 6, 2014
- Copyright © 2015 by the American Academy of Pediatrics