Table 9 Observational studies studying the impact of ANC on stillbirth and perinatal mortality

Bhardwaj et al. 1995 [77]

India (Uttar Pradesh). Rural setting.

Longitudinal study. 4 rural villages, 1987–88. N = 212 women.

Within the context of a home-based ANC program, assessed how a composite measure of maternal care receptivity (MCR), a weighted score based on initiation of ANC, frequency of home-based visits accepted, number of doses of tetanus toxoid, and place of and type of attendant at delivery, impacted perinatal outcomes. Subjects' MCR was graded as poor (N = 36, 17%), moderate (N = 161, 75.9%), or high (N = 15, 7.1%).

SB rate: 30/1000, 25/1000, 0/1000 in poor, moderate, and high MCR groups, respectively.

PM rate: 90.9/1000, 86.9/1000, and 0/1000 in poor, moderate, and high MCR groups, respectively

Neonatal mortality rate (NMR): 93.8/1000, 63.7/1000, and 0/1000 in poor, moderate, and high MCR groups, respectively.

High MCR group significantly different from low/moderate MCR groups (Z = 5.46, P < 0.0001).

Dyal Chand et al. 1989 [73]

India (Aurangabad, Maharashtra). Rural setting.

Community-based surveillance and monitoring, 1979–80 and 1987–88. 50 rural villages. N = unspecified.

Evaluated the impact of maternal health services on perinatal and neonatal mortality, delivered by TBAs, community health volunteers, and female workers.

Fetal deaths: 27% reduction [NS]

[1979–80 = 15.6/1000; 1987–88 = 11.4/1000]

Fauveau et al. 1990 [75]

Bangladesh (Matlab).

Prospective cohort study. 1979–1982. N = 13818 cases, N = 16781 controls.

Assessed the impact of the Intensive Family Planning and Health Services Programme on pregnancy outcomes, compared to controls given routine ANC.

PM rate: 21% reduction among intervention group over 8 years of study (P < 0.001)

[82/1000 at start vs. 65/1000 8 years later]

Fawcus et al. 1992 [54]

Zimbabwe (Harare). Hospitals setting.

Case control study. N = 195 unbooked recently delivered mothers (cases), N = 196 booked mothers (controls).

Compared the impact on pregnancy outcomes of having had or not had ANC (booked vs. unbooked mothers).

PMR: 72% reduction in children of booked vs. unbooked mothers (P < 0.001)

[35.9/1000 vs 129.7/1000 in booked vs. unbooked mothers, respectively].

Booked mothers also had lower MMR.

Goldenberg et al. 2007 [6]

51 countries (developed and developing).

Retrospective analysis of data from WHO and other sources.

Assessed how the number of antenatal visits impacted intrapartum stillbirth rates.

SBR (intrapartum): For each 1% increase in the percentage of women with at least 4 antenatal visits, the intrapartum stillbirth rate decreased by 0.16 per 1,000 births (P < 0.0001).

Gunter et al. 2007 [53]

Germany.

Retrospective study. Data from the Perinatal Registry of Lower Saxony.

Compared odds of stillbirth for pregnancies without any ANC vs. pregnancies with ANC.

SBR: OR = 6.089 (95% CI: 4.7–7.8, P < 0.01) for pregnancies without vs. pregnancies with ANC.

Kumar et al. 1997 [56]

India (Ambala, Harayana). Rural Rajpur Rani.

Cross-sectional survey. 4 rural villages with varying health services. N = 600 married women age 15–45.

Assessed how health care availability impacted utilisation of maternity care and pregnancy outcome, comparing 2 villages without any health centre (HC) to 1 village with a sub-centre (SC) and another village with a primary health centre (PHC).

PMR: 76.0/1000 in villages without HC

87.4/1000 in SC village

38.9/1000 in PHC village

Difference between village with PHC and all other villages was statistically significant (P < 0.01).

Kwast et al. 1995 [79]

Guatemala and Bolivia (also Indonesia and Nigeria, but these projects did not involve ANC)

Before-after studies measured by cross-sectional survey to evaluate MotherCare demonstration projects.

In Guatemala, the Quetzaltenango maternal and neonatal health project involved training 400 TBAs (to manage a population of 150,000), improving TBA-to-hospital referral services and posting a neonatologist. In Bolivia, the Warmi project engaged women's groups in problem prioritisation and action to reduce neonatal health, including improved training for traditional birth attendants and education for mothers during pregnancy. ANC attendance increased from 45 to 77% over course of project.

PMR:

Guatemala: 47% reduction in PMR among referred women in intervention area after intervention implementation (P = 0.003)

[22.2% before vs. 11.8% after]

Bolivia: 64% overall reduction in PMR

[105/1000 before vs. 38/1000 after]

Maternal deaths declined from 11 to 7 in the Bolivian study population (sample too small to calculate MMR).

McCaw-Binns et al. 1994 [76]

Jamaica.

Retrospective cohort study. Pregnant women included in the Jamaican Perinatal Mortality Survey, including all deliveries Sept-Oct 1986 and all perinatal deaths (N = 9919).

Assessed the timing of ANC initiation and its association with pregnancy outcomes, particularly perinatal mortality. Those who initiated ANC during the 2^nd trimester served as the reference group.

PMR:

Began in 1st trimester: OR = 0.67 (95% CI: 0.54–0.83)

Began in 2nd trimester: OR = 1.00 [reference]

Began > 29 wks: OR = 1.04 (95% CI: 0.82–1.31) [NS]

Protective effect of early initiation of ANC (χ² = 14.5, P < 0.001)

McClure et al. 2007 [64]

188 countries (low, middle, and high-income).

Retrospective regression analysis using WHO data.

Assessed the association of number of ANC visits with stillbirth incidence.

SBR: Regression analysis results: an increase of 1% of women with ≥ 4 antenatal visits decreased SB by 0.22/1000 (P < 0.0001) [all countries].

0.18/1000 (P = 0.0002) [low- and middle-income countries] 0.04/1000 (P = 0.5789) [high-income countries]

McCord et al. 2001 [124]

India (Ahmedagar & Pune districts).

Prospective cohort study. Pregnant women (N = 2905) in 25 villages in Ahmedagar district; controls drawn from neighboring Pune district.

A comprehensive rural health project was set up in a rural community with predominantly home births and limited access to emergency obstetric care. 64% of perinatal deaths were infants delivered at home.

SBR: 4% reduction [no significance data], [18.9/1000 vs 19.6/1000 in intervention group vs. controls, respectively]

PMR: 20% reduction [no significance data], [36/1000 vs 45.2/1000 in intervention group vs. controls, respectively]

MMR: 28% reduction [no significance data]. [70/100,000 vs 97/100,000 in intervention group vs. controls, respectively]

Nilses et al. 2002 [55]

Zimbabwe (Gutu, Masvingo Province). Rural setting.

Cross-sectional survey in 12 villages. N = 1213 women aged 15–44 years (N = 889 women had completed 3601 pregnancies).

Assessed self-reported reproductive outcome and utilisation of care to identify associations with perinatal outcomes.

PMR: 23/1000 among women who used ANC services vs. 40/1000 national figures [NS]

ENMR: 8.4/1000.

Panaretto et al. 2007 [74]

Australia (Queensland). Community-based study.

Before-after design. N = 865 (N = 781 after, N = 84 before).

Evaluated the impact of the Mums and Babies program, a community-based quality improvement intervention providing collaborative ANC care, in a cohort of women attending Townsville Aboriginal and Islanders Health Service (MB group), compared with a historical control group (PreMB group).

PMR: 77% reduction (P = 0.014)

[14/1000 vs. 60/1000 in MB group vs PreMB group, respectively]

Salinas 1997 [82]

Mexico. Hospital records.

Retrospective analysis using hospital records maintained by the National Institute for Perinatology, Mexico City, comparing avoidable perinatal death cases (N = 181) to non-avoidable deaths that served as controls (N = 341).

Assessed the relationship of quality of care to perinatal mortality by comparing avoidable perinatal deaths with non-avoidable perinatal deaths.

PMR: 24.8/1000 overall, possible 35% reduction if all avoidable perinatal deaths were prevented.

16% of the deaths presented structural and 31.2% process deficiencies; both predominated among avoidable perinatal deaths (35.4% vs 5.3%, P < 0.000; and 79.3% vs 5.9%, P < 0.000, respectively). Structural deficiencies increased risk avoidable perinatal death (OR = 11; 95% CI: 4.1–26.9. P < 0.001), as did process deficiencies (OR = 88, 95% CI: 37.2–204.5, P < 0.001).

Shah et al. 1984 [52]

India.

Prospective community-based study. N = 3151 women with live births, N = 90 women with stillbirths.

Compared the impact on perinatal outcomes between women who had had ANC vs. women who had had no ANC.

SBR: 35.1/1000 vs 20.8/1000 among women without ANC vs women with ANC, respectively. (P < 0.05)

67% (60/90) of mothers with stillbirths had no ANC, compared with 54% (1707/3151) women who had live births.

Southwick et al. 2007 [51]

Russia. Multisite study.

Prospective cohort study. Studied women with syphilis (N = 1071).

Compared the impact on perinatal outcomes between women who had had ANC vs. women who had had no ANC.

SBR: OR = 9.5 (95% CI: 4.0–23.5) among women with inadequately treated current syphilis who had no ANC vs those who had ANC.

[25% of those with no ANC had a stillbirth, vs. 3% of those with ANC].