Household characteristics affecting where mothers deliver in rural Kenya

Abstract

Data from a household survey were used to analyse the distribution of newborn deliveries in a rural area of Kenya. It was found that 52% of deliveries occurred at home or with traditional birth attendants. Using regression techniques, the most significant predictors of choosing an informal delivery setting are the household's distance from the nearest maternity bed and whether a household member has insurance. The results suggest that travel time is an important barrier to access. Therefore, quality improvements at existing facilities may not result in greater use of modern sector delivery, particularly if improvements are partially offset by user fees.

PIP: In October 1989 in the rural South Nyanza district in Nyanza province of Kenya, interviews were conducted with households where any member delivered at least 1 infant in the last year. The purpose was to study the distribution of deliveries right before public health facilities implemented user fees for deliveries (20-50 Ksh). There were 86 health facilities actually or potentially accessible by sampled households. There were maternity beds at 29 of 33 health centers, 1 of 6 subhealth centers, and 3 of 44 dispensaries. 52.3% of deliveries took place at home or at the place of a traditional birth attendant (TBA). TBAs attended 14.1% of all deliveries. 47.5% of deliveries took place at health facilities, especially government health facilities (28.8%). Missionary health facilities accounted for 16.7%. The average cost of delivery was much higher in health facilities than for TBAs (Ksh 31.5-311 vs. Ksh 28.8). In fact, almost half of TBAs charged nothing. The variable having the highest influence on delivery location was distance to nearest maternity bed (coefficient estimate = -0.0974). The greater the distance to a health facility with maternity beds, the less likely women would deliver at a health facility (p 0.01). A 1 km increase in an average household's distance from the nearest maternity bed would reduce the probability of choosing the formal sector by 3.4%. Health facilities were also less likely to be used if the head of the household was a male or if the household had no health insurance (p 0.05). These findings suggest that travel time is a considerable barrier to access to delivery facilities. They do not indicate that quality improvements at existing facilities would overcome the barrier of distance and travel time, especially if user fees partially offset improvements.