Understanding The Impact Of Prenatal Care: Improving Metrics, Data, And Evaluation
Prenatal care has been largely left out of the growing national conversation about the rise in maternal morbidity and mortality and the stark racial disparities in maternal health outcomes. This notable absence may be a consequence of how little is understood about the content and quality of prenatal care services and their relationship to maternal and infant health. Increased understanding of what happens during labor, delivery, and the postpartum period is essential to improving outcomes: the Centers for Disease Control and Prevention (CDC) estimates that roughly two-thirds of maternal deaths occur during childbirth and the first year thereafter. The remaining third of deaths occur during pregnancy. Prenatal care—spanning most of a year and consuming substantial time and resources from patients and providers alike—may represent an important opportunity to prevent these deaths, as well as to identify and mitigate risks of subsequent mortality or morbidity.
There are three primary impediments to rigorous research in this area:
- Reliance on blunt quality metrics that do not reflect important dimensions of care;
- Limited access to data on what occurs during prenatal care; and
- Empirical challenges to evaluating the impact of prenatal care on maternal and infant outcomes.
Limitations Of Current Quality Measures
The widely used measures of prenatal care quality capture only two dimensions of care: when it began and how many times it happened. These two features are combined into a single index, which compares the number of visits received to the recommended number of visits given the patient’s gestational age at initiation of care and at delivery. The most commonly used indices are the Kessner Index and the Kotelchuck/Adequacy of Prenatal Care Utilization (APNCU) Index (see exhibit 1). While both the timing of initiation and number of visits may matter for maternal and infant health, these adequacy indices have important shortcomings.
Exhibit 1: Adequacy of Prenatal Care Utilization (APNCU) Index in the United States, 2016
Source: Osterman MJK, Martin JA. Timing and adequacy of prenatal care in the United States, 2016. National Vital Statistics Reports. 2018 May 30. Notes: Recommendations on number of visits are established by the American College of Obstetricians and Gynecologists (ACOG). Currently, the ACOG recommends visits every four weeks for the first 28 weeks of gestation, every two weeks until 36 weeks’ gestation, and then weekly visits from 36 weeks’ gestation until delivery.
Adequacy indices do not capture whether the patient received guideline-recommended care, such as laboratory screenings, psychosocial support, and education on childbirth and parenting, during pregnancy. Instead, these indices regard each prenatal visit as a homogenous service. As a result, there is very little evidence to indicate whether women are receiving the full range of prenatal care that they need and how quality of care may vary across demographic groups.
In addition to masking potential heterogeneity in the quality of care received, existing adequacy measures classify care as “adequate” when it may in fact reflect an unnecessary number of visits. The American College of Obstetricians and Gynecologists (ACOG) recommends a higher number of visits for low-risk pregnancies (required to achieve “adequacy”) than other developed countries (for example, the United Kingdom and Australia). The ACOG’s recommendation—which amounts to 12–14 in-person visits at an increasing cadence over the course of a pregnancy—has persisted since the mid-twentieth century, despite calls for a reduced visit schedule. Evidence from the United States and five other countries shows that fewer visits leave health outcomes unchanged and, in fact, may reduce unnecessary interventions in pregnancy and childbirth.
Moving Beyond Adequacy Indices
Clinicians and researchers have recognized the shortcomings of adequacy indices and are developing improved metrics of prenatal care quality. One novel approach is to measure receipt of “guideline-concordant” prenatal care, or whether a patient receives the routine tests and screenings recommended by the ACOG, the US Preventive Services Task Force, and the CDC (for example, screenings for sexually transmitted infections, group B streptococcus, and gestational diabetes). These process measures move beyond adequacy to capture whether prenatal care was consistent with clinical guidelines.
To measure guideline-concordant prenatal care, we are partnering with OptumLabs under the Health Data for Action program sponsored by the Robert Wood Johnson Foundation. We will leverage their rich claims and laboratory data to document variation in receipt of guideline-concordant care, how it varies across demographic lines, and its association with traditional measures of prenatal care adequacy. This work is part of a research agenda that will provide some of the first evidence on core components of prenatal care in the United States; findings will highlight areas in need of improvement.
In addition to measuring receipt of screenings and tests, new metrics should capture whether patients were satisfied with the psychosocial support and anticipatory guidance provided during their care. For example, the International Consortium for Health Outcomes Measurement released a suite of patient-reported pregnancy outcomes, such as patient satisfaction with birth experiences and with their role as active participants in health care decisions. These outcomes capture critically important experiences of care that are not available in traditional data sources. Research has documented stark racial disparities in patient experiences with maternal care, underscoring the importance of documenting and analyzing these patient-reported outcomes.
Measuring guideline-concordant prenatal care and patient-reported outcomes are important steps to move beyond adequacy indices and understand variation in receipt of high-quality prenatal care.
The Need For Richer Data
Researchers’ continued reliance on adequacy indices can be partially attributed to limited availability of data that capture more dimensions of prenatal care. The primary source of information on the quality of prenatal care has been birth certificates. Although birth certificate data are relatively easy for researchers to access, there are concerns about their accuracy in measuring prenatal care. Additionally, birth certificates capture only the number of visits and timing of initiation, thus limiting researchers’ understanding of the care provided prenatally.
Moving beyond measuring adequacy of care will require richer data that capture key components of prenatal care visits. Health insurance claims data from commercial payers and Medicaid could be fruitful to this end. However, it is common for payers to reimburse providers at a global or bundled rate for both childbirth and prenatal care. Because providers often do not bill individually for each prenatal service, it is difficult to analyze process measures of care based on claims data. Moreover, billing practices vary across payers and states, making it challenging to use claims data to understand variation across those lines. Analysis of electronic health record (EHR) data from prenatal care visits would be very valuable in the process of developing new measures of prenatal care quality; however, such data typically are not available to researchers, especially in a consistent way across a wide spread of time and geography.
Some national surveys of pregnant women, such as the CDC’s Pregnancy Risk Assessment Monitoring System (PRAMS), capture more detailed elements of care, such as whether a provider discussed sexually transmitted infections during a prenatal visit. Still, these data lack sufficient granularity to serve as meaningful signals of prenatal care quality. They focus predominantly on anticipatory guidance and psychosocial support, and they do not capture whether recommended tests, such as screening for sexually transmitted infections, gestational diabetes, or group B streptococcus, were done. In addition, research has shown that the women surveyed may not accurately recall and report medical procedures and diagnoses on PRAMS.
Leveraging New Data Sources
If structured EHR data were captured consistently across health care providers, especially across institutions serving diverse patient populations, we could much better understand variations in prenatal care. The ADVANCE Clinical Data Research Network, led by OCHIN, is making strides toward this effort by assembling structured EHR data from federally qualified health centers nationwide. This effort and others like it are new in the field and represent an important opportunity for researchers. EHR data on prenatal care would be a particularly novel input to quality measures if they included information commonly absent from claims, including social and psychosocial determinants of health, such as food or housing insecurity, threats of violence, and low health literacy; and interventions for these risk factors, such as referrals to community organizations and educational activities.
Although the utility of claims data generally is limited by global billing practices, sources that integrate claims data from multiple payers, such as state all-payer claims databases, may be valuable to advance understanding of variations in quality of care across demographic groups. To our knowledge, these databases have not been leveraged to measure variation in quality of prenatal care, such as receipt of guideline-concordant screenings.
Finally, states and the federal government could add detail about prenatal care services to birth certificates and the PRAMS survey. For example, either of these data sources could be expanded to include information on which services were delivered (for example, testing for gestational diabetes, an ultrasound to check for fetal abnormalities), when, and by whom.
Evaluating The Impact On Outcomes
Although the efficacy of typical guideline-recommended prenatal care often goes unquestioned, evidence from both the epidemiological and economic literature is mixed.
There are few randomized controlled trials (RCTs) of prenatal care in high-income countries, since it would not be considered ethical to withhold these standard services. Instead, researchers are using RCTs to evaluate alternative models of prenatal care, such as group prenatal care, nurse home visiting, and telemedicine. Evaluating the impact of these programs relative to standard prenatal care, on clinical as well as patient-centered outcomes, may shed light on which components of the traditional model are or are not necessary. If these studies enroll diverse groups of women, they may also provide evidence on which subgroups (by demographic or health characteristics) benefit most from particular components of care.
In addition to these RCTs, researchers must make judicious use of observational data, paying particular attention to selection bias because women with different patterns of prenatal care utilization may differ in important ways. For example, women with higher-risk pregnancies generally receive more intensive prenatal care and are at greater risk for adverse birth outcomes than women with low-risk pregnancies. Women with unwanted pregnancies, who are uninsured or on Medicaid, or who have experienced discrimination, are more likely to delay their initiation of prenatal care. For these reasons, it is not possible to infer the impact of prenatal care by comparing birth outcomes among women who did and did not receive adequate prenatal care. Controlling for observed patient characteristics, such as insurance status, age, race, and baseline health conditions is likely inadequate to address selection bias, as it omits any unobserved or unmeasured factors that influence both health outcomes and use of prenatal care.
Instead of controlling for observed covariates, researchers should consider quasi-experimental study designs when using observational data to understand the effects of prenatal care. These studies often center on a change in policy that impacts women’s ability to receive prenatal care. For example, a recent study found that the Affordable Care Act’s dependent coverage provision was associated with modest improvements in adequacy and early receipt of prenatal care and a modest reduction in preterm birth. Still, these studies are limited in their ability to draw a definitive causal link between prenatal care and birth outcomes, given the myriad factors that may change after the relevant policy shift and also may affect health outcomes (for example, maternal stress may be alleviated by gaining health insurance and also may directly affect perinatal outcomes).
The challenges to conducting both randomized and observational studies have limited our understanding of the causal relationship between prenatal care and perinatal outcomes. These challenges are compounded by the previously described issues in quality measurement; our reliance on adequacy indices may mask important information about what kinds of prenatal services are most associated with outcomes of interest. In addition, there is a lack of consensus on which perinatal outcomes are most likely to be sensitive to prenatal care services. Other researchers have proposed a composite outcome of “potentially avoidable maternity complications” that attempts to capture outcomes most sensitive to prenatal care (analogous to the Agency for Healthcare Research and Quality’s widely used ambulatory sensitive care condition definition). Although additional refinement and validation are required, a measure such as this should be used in future work to understand what components of prenatal care are most essential for positive outcomes. Future work should also examine the impact of prenatal care on under-explored outcomes such as postpartum depression, patient satisfaction, and patients’ sense of self-respect and dignity.
Conclusion
To determine how prenatal care can be leveraged to reduce maternal mortality and morbidity in the United States, we must first understand current variations in quality of prenatal care and their relationship to maternal and neonatal outcomes. Legislation recently introduced in Congress, including the MOMMIES Act and the Maternal CARE Act, would establish demonstration programs for prenatal care innovation as one policy lever to address racial disparities; improve psychosocial support; and help better prepare women for labor, delivery, and the postpartum period. Although these proposals represent important legislative attention to pressing issues, the lack of rigorous research on prenatal care leaves uncertainty as to whether the services proposed in the legislation, including pregnancy medical homes and telemedicine, will be effective in improving perinatal outcomes and reducing disparities.
Understanding the role of prenatal care in maternal and infant well-being will require developing more meaningful quality metrics, leveraging new data sources, and finding new and creative ways to conduct evaluations with careful attention to selection bias. Together, providers, researchers, heath care organizations, and funders of health services research can make progress in understanding and improving the quality of prenatal care.
