Is There a Causal Relationship Between Maternal Health

Is There a Causal Relationship Between Maternal Health Care Utilization and Subsequent Contraceptive Use? : Evidence from Kenya and Zambia Mai Do and David Hotchkiss Tulane University

Background § Service integration receiving increased attention from governments and donors as way of improving efficiency and access to services § Several reasons why use of maternal health services might influence post-partum contraceptive use § Few studies have examined family planning (FP) within the context of reproductive health service delivery § Mixed evidence on linkages between maternal health care (ANC, delivery, and PNC) and post-partum FP

Research questions § Is post-partum modern FP method use related to the use of ANC and PNC relating to the index childbirth? § If so, what can be said about the linkages between these services?

Data § Most recent DHS: 2008 -09 in Kenya and 2007 in Zambia § Selected Kenya and Zambia because: § DHS within the last three years § DHS included a birth and contraceptive calendar § Substantial contraceptive use among married and cohabiting women § Study sample: married and cohabiting women who had a live birth within five years of the survey § Kenya: 3, 667 women § Zambia: 3, 587 women

Methods (1) § Statistical method: Cox proportional hazard model § Dependent variable: duration (months) from childbirth to modern contraceptive adoption

Methods (2) § Independent variable: ANC and PNC service intensity score, constructed from 6 questions § Timing of first ANC visit, if any § Number of ANC visits (4 or more) § Received tetanus vaccination § Received ANC from trained provider § Whether specified procedures were carried out during ANC visit (measuring weight and height, blood pressure, taking urine and blood sample, breastfeeding counseling, being told about signs of complications) § Received PNC from trained provider

Methods (3) § Main control variables § Knowledge of contraceptive methods § Woman visited and talked about FP with a field worker last 12 months § Woman visited and talked about FP at a health facility last 12 months § Desire for more children § Prior use of modern contraceptive methods § Recall of FP messages in the mass media § Tests performed of exogeneity of ANC and PNC service intensity score

Country findings: Kenya 46% adopted modern FP post-partum Characteristic Distribution ANC/PNC service intensity Post-partum modern FP use % or mean (s. e. ) Coef. (s. e. ) Hazard ratio (s. e. ) ANC/PNC service intensity (range: -2. 75; . 90) 0 (1) 1. 11 (. 04)* Age at first birth 19. 2 (3. 4) . 02 (. 00)** - Desire for more children (ref=No) 49. 7 - . 88 (. 05)* Number of modern methods known (knowledge) 6. 7 (2. 6) - 1. 07 (. 01)*** Visited and talked about FP at health facility last 12 months (ref=No) 20. 9 - 1. 21 (. 07)** Heard FP messages on the radio last few months (ref=No) 71. 6 - 1. 15 (. 08)* * p<. 05; ** p<. 01; *** p<. 001 Controls for: socio-demographic characteristics, durations of breasfeeding and amenorrrhea

Kenya: Influences of ANC and PNC services on post-partum modern FP use Characteristic Distribution Post-partum modern FP use % or mean (s. e. ) Hazard ratio (s. e. ) ANC service intensity (range: -2. 72; . 88) . 1 (. 9) 1. 10 (. 04)** PNC service intensity (range: 0; 2. 00) . 6 (. 6) 1. 03 (. 05) * p<. 05; ** p<. 01; *** p<. 001 Controls for all women’s characteristics mentioned before.

Country findings: Zambia 45. 9% adopted modern FP post-partum Characteristic Distribution ANC/PNC service intensity Post-partum modern FP use % or mean (s. e. ) Coef. (s. e. ) Hazard ratio (s. e. ) ANC/PNC service intensity (range: -4. 07; 1. 27) 0 (1) 1. 08 (. 03)* Age at first birth 18. 6 (3. 0) . 04 (. 01)* - Desire for more children (ref=No) 66. 4 - . 95 (. 06) Number of modern methods known (knowledge) 6. 8 (2. 1) - 1. 04 (. 01)** Visited and talked about FP by a field worker last 12 months (ref=No) 7. 8 - 1. 20 (. 10)* Visited and talked about FP at health facility last 12 months (ref=No) 32. 4 - 1. 23 (. 07)*** * p<. 05; ** p<. 01; *** p<. 001 Controls for: socio-demographic characteristics, durations of breasfeeding and amenorrrhea

Zambia: Influences of ANC and PNC services on post-partum modern FP use Characteristic Distribution Post-partum modern FP use % or mean (s. e. ) Hazard ratio (s. e. ) ANC service intensity (range: -4. 06; 1. 26) . 0 (. 99) 1. 08 (. 03)* PNC service intensity (range: 0; 2. 00) . 6 (. 7) . 95 (. 04) * p<. 05; ** p<. 01; *** p<. 001 Controls for all women’s characteristics mentioned before.

Conclusions § Evidence of MCH service use as mediator for individual characteristics to influence post-partum modern FP use § Results suggest that maternal health care use and FP use are not influenced by common unobserved factors § Several observed individual factors influence maternal health care use, which then influence FP use § Evidence that ANC, not PNC, service intensity related to post-partum modern FP use § Need for service integration, esp. in public sector § PNC may be among the weakest aspects of RH program

Limitations § Not generalizable to all women of childbearing age § Only married, cohabiting women included in the sample § Possible endogeneity between FP use and exposure to FP messages in the media and visit by FP a field worker § No data on community-level and programmatic factors

MEASURE Evaluation PRH is a MEASURE project funded by the United States Agency for International Development (USAID) through Cooperative Agreement GHA-A-00 -08 -0000300 and is implemented by the Carolina Population Center at the University of North Carolina at Chapel Hill in partnership with Futures Group International, Management Sciences for Health, and Tulane University. Views expressed in this presentation do not necessarily reflect the views of USAID or the U. S. Government. MEASURE Evaluation PRH supports improvements in monitoring and evaluation in population, health and nutrition worldwide.