After many rounds of review, our paper on the ARRIVE trial using interrupted time series analysis is finally published!
Guidelines recommend screening for GDM using a series of tests that involve several blood samples and consumption of a ‘measured’ glucose drink. Depending on how people are screened, this may involve fasting. Not everyone chooses to complete the ‘recommended’ screening – anecdotally, I supected that this could be much more common in midwifery care.
For this study, I compared ‘non-standard GDM screening’ to ‘recommended screening’ among midwifery clients and physician in BC. I wanted to understand, first, whether the incidence of non-standard screening was more common in midwifery care, and second, what this might mean for infant and maternal health.
A version of this work is being presented as a poster at the UBC Pharmaceutical Sciences Graduate Research Forum on May 4, 2023. The abstract is below – and the full poster is also available here: Poster download
Perinatal Outcomes after Non-standard Screening for Gestational Diabetes Among Midwifery Clients in British Columbia, Canada
Presenter: Elizabeth Nethery
Elizabeth Nethery1,2, Laura Schummers1, Luba Butska2, Michelle Turner2, Jennifer A Hutcheon3,4, Patricia A Janssen3
Midwife-led care may be associated with lower completion of guideline-based prenatal tests and procedures. Routine gestational diabetes mellitus (GDM) screening, using guideline-based methods, is recommended from 24-28 weeks gestation, as diagnosis and treatment can reduce risks of adverse outcomes.1 Qualitative data suggests non-standard screening is more common among midwifery vs physician care.2,3 Rates of non-standard screening are not well understood. We evaluated risks of adverse maternal and neonatal outcomes comparing those unscreened to screened negative for GDM among midwifery clients.
We used a population-based linked cohort of all singleton births (>28 weeks gestational age) from 2005-2019. GDM screening status was obtained from billings data via a validated method.4 We reported frequency of non-standard screening (no screening or no guideline-based screening) compared to screening negative for GDM, crude and adjusted relative risks (RRs) for maternal and perinatal outcomes, adjusted for confounders, among midwifery clients. Potential confounders for adjustment (age, parity, pre-pregnancy body mass, antepartum risk composite, gestational age, health region, scheduled cesarean birth, planned home birth) were selected based on directed acyclic graph approach and informed by previous literature.
Among 91,317 births (2005-2019), 33% of midwifery clients had non-standard gestational diabetes screening. Non-standard screening for GDM was associated with decreased risks compared to those who screened negative for cesarean delivery (RR 0.87, 95%CI 0.84-0.89), shoulder dystocia (RR 0.83, 95%CI 0.77-0.90), large for gestational age (RR 0.88, 95%CI 0.85-0.91) or neonatal hypoglycemia (RR 0.78, 95%CI 0.65-0.92). There was an increased risk of small for gestational age (RR 1.12, 95%CI 1.05-1.19).
Within the context of a midwife-led model of care, non-standard GDM screening (no screening or alternative screening) was not associated with a higher risk of adverse perinatal outcomes normally associated with untreated gestational diabetes; however, an increase in small for gestational age warrants further study.
1. Feig, D. S. et al. Diabetes Canada Clinical Practice Guidelines Expert Committee. Can. J. Diabetes 42, S255–S282 (2018).
2. Murray-Davis, B. et al. A framework for understanding how midwives perceive and provide care management for pregnancies complicated by gestational diabetes or hypertensive disorders of pregnancy. Midwifery 115, 103498 (2022).
3. Stoll, K., Wang, J. J., Niles, P., Wells, L. & Vedam, S. I felt so much conflict instead of joy: an analysis of open-ended comments from people in British Columbia who declined care recommendations during pregnancy and childbirth. Reprod. Health 18, 79 (2021).
4. Nethery, E., Hutcheon, J. A., Law, M. R. & Janssen, P. A. Validation of Insurance Billing Codes for Monitoring Antenatal Screening. Epidemiology 34, (2023).
I have had a bunch of new manuscripts published in the past few months. These are all based on work from my PhD dissertation (completed in November 2022).
Most of the work is on gestational diabetes in pregnancy and I also did a second project on COVID-19 and pregnancy weight gain.
Gestational diabetes, screening and diagnosis
This is the bulk of the work that I did for my PhD. A series of 3 linked studies have so far been published, and I continue to examine other work using this cohort.
The overarching goal of this project is to explore how gestational diabetes screening in BC has changed over the past 15 years, how screening changes interact with diagnoses, how these changes might effect perinatal outcomes, and examine any interactions with health care provider type (ie. physician v midwifery care). This work uses data from British Columbia, Canada.
- Validation study – In order to study gestational diabetes screening, I needed to be able to obtain valid data on screening. At the advise of one of my committee members (shout-out to the brilliant Dr. Jennifer Hutcheon), I did a chart abstraction study and then compared the results of a random sample of medical records to the administrative data I planned to use for the larger population-based study. The manuscript for this work is available (published) in the journal Epidemiology – see here: Validation of Insurance Billing Codes for Monitoring Antenatal Screening
- Descriptive study – Gestational diabetes policies and practice guidelines have shifted in BC and Canada since 2004. This study describes incidence of screening across different population groups and presents this in context with relevant policy and guideline changes. This is published in the Journal of Obstetrics and Gynecology of Canada (JOGC) – see here: Trends in Gestational Diabetes Screening Practices in British Columbia from 2005–2019
- Screening changes and effects on diagnosis – This paper takes the next step of looking at how the screening changes in BC impacted gestational diabetes diagnoses. Specifically, incidence of gestational diabetes in BC in recent years has almost doubled since the early 2000’s and appears to be much higher than the Canadian average. Is this increase because of population demographic changes or other underlying risk factors changing? Or is this because of the screening practice changes that I found. This manuscript was published in the Canadian Medical Association Journal (CMAJ) – see here: The effect of changing screening practices and demographics on the incidence of gestational diabetes in British Columbia, 2005–2019
Covid-19 and pregnancy weight gain
This is an unrelated manuscript using Washington State data from the OB-COAP study population. We used an interrupted time series analysis to examine the time point of the COVID-19 pandemic and whether this impacted pregnancy weight gain and/or infant birth weight after the onset of the pandemic. This was published in the American Journal of Clinical Nutrition – see here: Weight gain in pregnancy and infant birthweight after the onset of the COVID-19 pandemic: an interrupted time series analysis – ScienceDirect
OBJECTIVE: To describe rates of maternal and perinatal birth outcomes for community births and to compare outcomes by planned place of birth (home vs statelicensed, freestanding birth center) in aWashington State birth cohort, where midwifery practice and integration mirrors international settings.
METHODS: We conducted a retrospective cohort study including all births attended by members of a statewide midwifery professional association that were within professional association guidelines and met eligibility criteria for planned birth center birth (term gestation, singleton, vertex fetus with no known fluid abnormalities at term, no prior
cesarean birth, no hypertensive disorders, no prepregnancy diabetes), from January 1, 2015 through June 30, 2020. Outcome rates were calculated for all planned community births in the cohort. Estimated relative risks were calculated
comparing delivery and perinatal outcomes for planned births at home to state-licensed birth centers, adjusted for parity and other confounders.
RESULTS: The study population included 10,609 births: 40.9% planned home and 59.1% planned birth center births. Intrapartum transfers to hospital were more frequent among nulliparous individuals (30.5%; 95% CI 29.2–31.9) than multiparous individuals (4.2%; 95% CI 3.6–4.6). The cesarean delivery rate was 11.4% (95% CI 10.2–12.3) in nulliparous individuals and 0.87% (95% CI 0.7–1.1) in multiparous individuals. The perinatal mortality rate after the onset of labor (intrapartum and neonatal deaths through 7 days) was 0.57 (95% CI 0.19–1.04) per 1,000 births. Rates for other adverse outcomes were also low. Compared with planned birth center births, planned home births had similar risks in crude and adjusted analyses.
CONCLUSION: Rates of adverse outcomes for this cohort in a U.S. state with well-established and integrated community midwifery were low overall. Birth outcomes were similar for births planned at home or at a state-licensed, freestanding birth center.
Quick summary of the paper?
Probably easiest to see this #tweetorial I put together.
Or my slide-deck.
Why blog about this?
Well, this has been a long project – many years in the making. I have been a volunteer with the Midwives Association of Washington State as part of the data and research committee. We have been working to identify ways that midwifery data from their midwives can be used for research and quality improvement. I have been presenting to the midwives, via a data webinar, for the past several years on midwifery outcomes.
Finally, in 2020 we had enough births to really work on trying to get a manuscript together for publication. There were a lot if issues – though – that we needed to work through in terms of checking the data. The data transfer from chart-abstraction (via MANAStats) to the import into OB COAP data (which is how we accessed the data) included mapping fields from the original data abstraction form to the OB COAP data fields – and that was set up years ago, before I started being involved with OB COAP and MAWS.
So, before I was able to start the analysis, we needed to do some internal validation of the data. That included cross-checking MANAStats data (for a non-random sample of local midwives) against the same data records in OB COAP. And, all perinatal mortality cases were also reviewed individually at the data level and then compared to cases in MAWS Quality management program and in the MANAStats death reviews (where available).
About the study
But most people plan to birth in the hospital? So why should we care about this small ‘fringe’ group of people who birth at home or in birth centers?
Yes – this is only a relatively small percent of all birthing people. But, there are a few reasons why this is still important. First, more and more people ARE planning to birth “in the community” – ie. at home or at a freestanding (not attached to a hospital) birth center. Across the US, this is increasing every year, although still only about 1-2% of all births. In states like Washington, in 2020, 4% of people actually birthed in the community. And this was a big jump up from 2019 – when it was only 3.5%. So, Covid has made a lot of people rethink where they want to birth.
Research on planned community birth is compelling – people have fewer cesarean deliveries, less interventions overall and high levels of satisfaction when compared to people who plan to birth in the hospital.
Some people choose a community birth because they want more control over their birth environment. Some want to have a low-intervention birth, avoiding pressure to have pain medication or IVs. Others want to have a waterbirth, which relatively few hospitals will allow. Others want to have midwifery care from a provider they know – who also did their prenatal care – and someone they have developed a relationship with. Some want to have family attend and hav more support people present during and after birth – which might not be allowed in the hospital.
So – what did we actually do?
We had data for over 11,000 planned home and birth center births that were intended to be with MAWS member midwives in Washington State from 2015-2020. In the supplemental for the paper, we report on outcomes for the whole study population – but in the main paper, we focus on a subgroup who met ‘eligibility critieria’ for Washington birth center birth and were within the MAWS Guidelines Indications document criteria. In essence, this included term, vertex births and people without previous cesarean delivery. We really didn’t want to make a strong comment on who is “eligible” for community birth or on “risk status” – that is up to an individual person via a shared decision-making approach. However, there are guidelines which exist – as GUIDELINES – to ensure that midwives have evidence-based recommendations to discuss and review with birthing people. These guidelines can be and are used to recommend either a transfer or care (to a hospital-based provider), or a consultation depending on what medical or obstetric conditions arise during prenatal, intrapartum, postpartum or neonatal care. However, a pregnant person may elect to decline the care recommendations made by the midwife – this is a key aspect of patient autonomy.
For example, in the case of midwifery care for VBAC (Vaginal birth after cesarean) – while this is a recommendation for a consultation with a physician in BC, and in WA State MAWS has a separate document for VBAC guidelines – there are many people who still elect to plan a VBAC (also referred to as TOLAC – Trial of Labor after Cesarean, or, my preference – simply, Labor after Cesarean (LAC)). People may elect to plan LAC at home because they are unable to find a supportive local hospital-based provider who will “allow” them to plan to labor at all. This is an especially important issue in smaller, rural areas where few hospitals support LAC.
People also want midwifery care. While attending VBAC patients is within scope for Certified Nurse Midwives (and LMs as well), many hospitals (even in WA State) have restricted CNMs from attending VBACs – forcing these patients to have OB care only. While this may be a well-meaning approach, in effect, some people who strongly value midwifery care may consider a planned community birth – because they can have a health care provider who they know, and who will support their desire to have a low-intervention birth. Greater availability of in-hospital midwifery options with CNMs might decrease the likelihood of pregnant people choosing to plan a home VBAC.
This study was mostly descriptive – this means we described the rates of outcomes for both birthing person and the neonate. But, we also did a statistical comparison of planned birth centers and home births.
We had two primary goals:
- Report outcomes data from this cohort and compare to other studies of community birth in similar settings
- Compare planned home to planned birth center births.
Low rates of perinatal and maternal (birthing person) complications.
Low Cesarean rates, low NICU admissions. Perinatal mortality (combined intrapartum and neonatal deaths 0-6d) was low and comparable to international benchmarks from other published community birth studies. Importantly, these other studies were in settings where midwifery and home birth is held up as a ‘model’ for better integration compared to the US. We make the case that WA state midwifery *is* better integrated – and hence outcomes are comparable to other settings.
There was no detectable difference in outcomes for planned home births compared to planned birth center births.
But what about those first time pregnancies – 30% of them transferred to hospital!!
One of the other key findings was that we reported 30% of first time birthing people transferred to hospital in labor. What is important is that this is right in line with what is reported from the UK (also 30%), from Canada (more like 40-50%), and slightly higher than other US studies (20-28%).
There are a few really important things about this.
First, even after a transfer, 63% went on have a vaginal delivery (~37% had a cesarean). So, given that many US hospitals have a baseline cesarean rate for first time birthing people that is in the 30%-35% range, this is still a pretty good chance of having a vaginal birth.
Second, transferring to hospital is not a “bad outcome” or a failure. This is an important part of informed choice, safe community birth and midwifery care risk management. People who plan to birth in the community may choose to go to hospital because they want medical pain management or simply because they decide they want to be in the hospital setting. A midwife may recommend a transfer as well – for slow or stalled progress (by far the most common reason), for therapeutic rest (especially in the case of a long prodromal labor). Or, much less commonly, a midwife may recommend a transfer for fetal reasons (heart rate), or for symptoms or vital signs in the birthing person (blood pressure, temperature, bleeding).