Distance matters: barriers to antenatal care and safe childbirth in a migrant population on the Thailand-Myanmar border from 2007 to 2015, a pregnancy cohort study

Abstract

Background: Antenatal care and skilled childbirth services are important interventions to improve maternal health and lower the risk of poor pregnancy outcomes and mortality. A growing body of literature has shown that geographic distance to clinics can be a disincentive towards seeking care during pregnancy. On the Thailand-Myanmar border antenatal clinics serving migrant populations have found high rates of loss to follow-up of 17.4%, but decades of civil conflict have made the underlying factors difficult to investigate. Here we perform a comprehensive study examining the geographic, demographic, and health-related factors contributing to loss to follow-up.

Methods: Using patient records we conducted a spatial and epidemiological analysis looking for predictors of loss to follow-up and pregnancy outcomes between 2007 and 2015. We used multivariable negative binomial regressions to assess for associations between distance travelled to the clinic and birth outcomes (loss to follow-up, pregnancy complications, and time of first presentation for antenatal care.) RESULTS: We found distance travelled to clinic strongly predicts loss to follow-up, miscarriage, malaria infections in pregnancy, and presentation for antenatal care after the first trimester. People lost to follow-up travelled 50% farther than people who had a normal singleton childbirth (a ratio of distances (DR) 1.5; 95% confidence interval (CI): 1.4 - 1.5). People with pregnancies complicated by miscarriage travelled 20% farther than those who did not have miscarriages (DR: 1.2; CI 1.1-1.3), and those with Plasmodium falciparum malaria in pregnancy travelled 60% farther than those without P. falciparum (DR: 1.6; CI: 1.6 - 1.8). People who delayed antenatal care until the third trimester travelled 50% farther compared to people who attended in the first trimester (DR: 1.5; CI: 1.4 - 1.5).

Conclusions: This analysis provides the first evidence of the complex impact of geography on access to antenatal services and pregnancy outcomes in the rural, remote, and politically complex Thailand-Myanmar border region. These findings can be used to help guide evidence-based interventions to increase uptake of maternal healthcare both in the Thailand-Myanmar region and in other rural, remote, and politically complicated environments.

Keywords: Access to care; Geography; Healthcare delivery; Malaria; Maternal health; Pregnancy; Prenatal care.

Fig. 1

Satellite images showing environmental changes over time for two of the antenatal clinics: Mawker Tai (top panel: A and B) and Wang Pha (bottom panel: C and D) clinics. The first columns (A for Mawker Tai and C for Wang Pha) show the larger geographic area whereas the second columns (B and D) show greater detail in the immediate clinic area (zoomed areas are indicated with the black boxes in A and C). The top rows show historical images, and the bottom rows show more recent images. Deforestation is apparent, especially in the Mawker Tai images and on both the Thailand and Myanmar side of the international border. Infrastructure and development have also increased at both clinics, with increased year-round roads, housing, and increased agricultural fields. Images come from Google Earth

Fig. 2

Map of the catchment areas for the four antenatal clinics (ANCs) from 2007 to 2015. Each of the four clinics is indicated by a different color scheme. The ellipses are standard deviational ellipses, with 2 and 3 standard deviations, explained in detail in the supplementary materials. The darker circle represents roughly 98% of people’s home villages for that specific clinic and the lighter circle representing 99.9% of people’s home villages. WPA (green) and MKT (blue) provided both antenatal care and skilled birth attendants, and MRC (yellow) and WAL (orange) provided antenatal care. WPA and MKT had the largest catchment areas whereas MLC and WAL served a more local population. Maps indicating changes in catchment area over time are presented in Supplementary Fig. 2. Maps and layers were created by DMP using QGIS software