Effect of telehealth-integrated antenatal care on pregnancy outcomes in Australia: an interrupted time-series analysis

Background: During the COVID-19 pandemic, rapid integration of telehealth into antenatal care occurred to support ongoing maternity care. A programme of this scale had not been previously implemented. We evaluated whether telehealth-integrated antenatal care in an Australian public health system could achieve pregnancy outcomes comparable to those of conventional care to assess its safety and efficacy. Methods: Routinely collected data for individuals who gave birth at Monash Health (Melbourne, VIC, Australia) during a conventional care period (Jan 1, 2018, to March 22, 2020) and telehealth-integrated period (April 20, 2020, to April 25, 2021) were analysed. We included all births that occurred at 20 weeks' gestation or later or with a birthweight of at least 400 g (if duration of gestation was unknown). We excluded multiple births, births for which private antenatal care was received, and births to individuals transferred from other hospitals or who had no antenatal care. Baseline demographics, telehealth uptake, and pregnancy complications (related to pre-eclampsia, fetal growth restriction [FGR], gestational diabetes, stillbirth, neonatal intensive care [NICU] admission, and preterm birth [<37 weeks' gestation]) were compared using comparative statistics and an interrupted time-series analysis. Results were stratified by care stream, with high-risk models consisting of obstetric specialist-led care, and all other streams categorised as low-risk models. The impact of the integrated period on outcomes was also assessed with stratification by parity. Findings: 17 873 births occurred in the conventional period and 8131 in the integrated period. Compared with the conventional period, women giving birth during the integrated period were slightly older (30·63 years vs 30·88 years) and had slightly higher BMI (25·52 kg/m² vs 26·14 kg/m²), and more Australian-born women gave birth during the integrated period (37·37% vs 39·79%). There were no significant differences in smoking status or parity between the two groups. 107 (0·08%) of 129 514 antenatal consultations in the conventional period and 34 444 (45·94%) of 74 982 in the integrated period were delivered by telehealth. No significant differences between the conventional and integrated periods were seen in median gestational age at pre-eclampsia diagnosis (low-risk models 37·4 weeks in the conventional period vs 37·1 weeks in the integrated period, difference –0·3 weeks [–0·7 to 0·1]; high-risk models 35·5 weeks vs 36·3 weeks, difference 0·3 weeks [–0·3 to 1·1]), incidence of FGR below the 3rd birthweight percentile (low-risk models 1·62% vs 1·74%, difference 0·12 percentage points [–0·26 to 0·50]; high-risk 4·04% vs 4·13%, difference 0·089 percentage points [–1·08 to 1·26]), and incidence of preterm birth (low-risk models 4·99% vs 5·01%, difference 0·02% [–0·62 to 0·66]; high-risk models 15·76% vs 14·43%, difference –1·33% [–3·42 to 0·77]). Parity did not affect these findings. Interrupted time-series analysis showed a significant reduction in induction of labour for singletons with suspected FGR among women in low-risk models during the integrated period (–0·04% change per week [95% CI –0·07 to –0·01], p=0·0040), and NICU admission declined after telehealth integration (low-risk models –0·02% change per week [–0·03 to –0·003], p=0·018; high-risk models –0·10% change per week, –0·19 to –0·001; p=0·047). No significant differences in stillbirth rates were observed. The proportion of women diagnosed with gestational diabetes was significantly higher in the integrated period compared with the conventional period for both low-risk care models (22·28% vs 25·13%, difference 2·85 percentage points [1·60 to 4·11]) and high-risk care models (28·70% vs 34·02%, difference 5·32 percentage points [2·57 to 8·07]). However overall, when compared with the conventional period, there was no significant difference in proportion of women with gestational diabetes requiring insulin therapy (low-risk models 8·08% vs 7·73%, difference –0·35 percentage points [−1·13 vs 0·44]; high-risk models 14·81% vs 15·71%, difference 0·89 percentage points [−1·23 to 3·02]), or proportion of women with gestational diabetes who gave birth to a baby with macrosomia in the integrated period (low-risk models 3·16% vs 2·33%, difference –0·83 percentage points [−1·77 to 0·12]; high-risk models 5·58% vs 4·81%, difference –0·77 percentage points [−3·06 to 1·52]). Interpretation: Telehealth-integrated antenatal care replaced around 46% of in-person consultations without compromising pregnancy outcomes. It might be associated with a reduction in labour induction for suspected FGR, particularly for women in low-risk models, without compromising FGR detection or perinatal morbidity. These findings support the ongoing use of telehealth in providing flexible antenatal care. Funding: None.