Menu
Menu
Menu
Menu
Background As part of World Health Organization (WHO) 2016 updated antenatal care (ANC) guidelines routine ultrasonography is recommended, including to detect congenital anomalies. The Ghana Health Service (GHS) developed an in-service midwifery ultrasound training course in 2017, which includes fetal anomaly detection. Training rollout has been very limited. We sought to determine proportions of anomalies among neonates presenting to Tamale Teaching Hospital (TTH) that should be prenatally detectable by course-trained midwives in order to determine training program potential utility. Methods We analyzed data from a registry of neonates admitted to TTH with congenital anomaly diagnoses in 2016. We classified ultrasonographic detectability of anomalies at ≤13 and 14–23 weeks gestation, based on GHS course content and literature review. Secondary analysis included 2011–2015 retrospective chart review data. Results Eighty-five neonates with congenital anomalies were admitted to TTH in 2016. Seventy-three (86%) mothers received ≥1 ANC visit; 47 (55%) had at least one prenatal ultrasound, but only three (6%) were interpreted as abnormal. Sixteen (19%) and 26 (31%) of the anomalies should be readily detectable by course-trained midwives at ≤13 and 14–23 weeks gestation, respectively. When the 161 anomalies from 2011–2015 were also analyzed, 52 (21%) and 105 (43%) should be readily detectable at ≤13 and 14–23 weeks gestation, respectively. “Optimal conditions” (state-of-the-art equipment by ultrasonography-trained physicians) should readily identify 53 (22%) and 115 (47%) of the anomalies at ≤13 and 14–23 weeks gestation, respectively. Conclusion Training Ghanaian midwives could substantially increase second trimester anomaly detection, potentially at proportions nearing highly resourced settings. Our data also highlight the need for refinement of the WHO antenatal ultrasonography recommendation for a scan before 24 weeks gestation for multiple purposes. Gestational dating accuracy requires first trimester scanning while fetal anomaly detection is more accurate during second trimester. Further specification will enhance guideline utility.
All neonates (<28 days) with congenital anomalies were enrolled into a registry at the time of admission to the TTH NICU in 2016. This unit and, by extension the hospital, serves seven of the 16 administrative regions of Ghana, encompassing a population of approximately 7 million people [11]. An estimated 200,000 annual births occur in this catchment area, approximately 22% of births nationally [11]. An abstraction template was used to record maternal and neonatal data, including prenatal ultrasound and maternal characteristics from ANC cards, and delivery history from inpatient notes. Based on the GHS midwifery training manual [12], a literature review [13–15], and maternal-fetal medicine specialist expert opinion (EEF) we classified anomalies as those that should be “readily detectable”, “potentially detectable” or “not detectable” by midwives trained in the course as well as under “optimal” circumstances (Table 1). The latter was defined as use of state-of-the-art transabdominal ultrasound by physicians trained in prenatal ultrasonography, but excluding advanced techniques (e.g., transvaginal ultrasound, nuchal translucency measurement). a“Optimal Conditions” were defined as using state-of-the-art ultrasound technology by an ultrasonography-trained physician. bMidwifery training was defined per Vance C., Jeanty P. Limited Obstetric Ultrasound: Course Manual. General Electric Healthcare; 2016 [8]. The 2-week GHS midwifery training is based on use of the General Electric (GE) V-scan ACCESS model to determine fetal number, estimate gestational age, assess for placental conditions, measure amniotic fluid levels, and identify internal and external fetal structures–including to detect anomalies [12]. While not intended to comprehensively identify all anomalies, the training provides explicit instruction on assessment for a number of specific anomalies (e.g., gastroschisis, hydrocephalus, spina bifida) as well as visualization and inspection of specific fetal structures (e.g. lower extremities, brain, genitourinary tract). While some anomalies were not specifically referenced by name in the training manual, examination techniques and anatomical coverage should lead to detection (e.g., scan of the genitourinary tract should reveal bladder exstrophy even though this condition was not specifically named in the manual). Anomalies in such scenarios were coded as “potentially detectable”. Other circumstances in which the “potentially detectable” code was applied included conditions that progress during gestation (e.g., microcephaly) or present with varying severity (e.g., osteogenesis imperfecta). Details of detectability coding are presented in Table 1. The ability to detect most anomalies varies by gestational age, hence we reported detectability under the following scenarios: 1) by 13 weeks gestation under “optimal” circumstances, 2) by 13 weeks gestation by course-trained midwives, 3) between 14–23 weeks gestation under “optimal” circumstances, and 4) between 14–23 weeks gestation by course-trained midwives. If a child was diagnosed with more than one condition, we included the most readily detectable anomaly in our classification count. As a secondary analysis, we included data collected by retrospective chart review and published in 2017 that enumerated neonates admitted to the TTH NICU with congenital anomalies over five years (2011–2015) [16]. We applied the same framework described above to these data. Descriptive statistics were calculated using Excel (Microsoft Corporation, Bellevue, WA). We also explored whether a priori determined demographic and clinical variables (number of ANC visits (classified as any vs none and ≥4 vs <4), history of prenatal ultrasound (including if at <24 weeks gestation vs ≥24 weeks and if by the first 2 trimesters vs not), and advanced maternal age) were associated with an anomaly amenable to detection by prenatal ultrasonography at <24 weeks gestation. These demographic and clinical data were captured in the 2016 registry, but not available for patients admitted from 2011–2015. We also tested whether there was a difference in potential detectability based on the midwifery training course compared to “optimal conditions”, using chi-square and Fisher exact (if cells contained <five observations) tests in Stata 16.0 (StataCorps, College Station, TX). P-values were two-tailed and alpha defined as 0.05. This study was approved by the TTH Ethical Review Committee (TTHERC/19/06/18/18) and exempted from University of Washington Human Subjects Division review. As data was abstracted from routine clinical records anonymously without any identifiers, consent was not required.
