Estimates of contraceptive prevalence based on service statistics and surveys in Gujarat State, India
Visaria, L.; Visaria, P.; Jain, A.
Studies in Family Planning 25(5): 293-303
1994
ISSN/ISBN: 0039-3665 PMID: 7871554 DOI: 10.2307/2138060Document Number: 293168
A comparison was made between contraceptive use statistics gleaned from service statistics and from survey statistics in four districts (Bharuch and Kheda as matched districts and Panchmahals and Rajkot) in Gujarat state, India. Service statistics showed increased use of reversible methods. The first survey was conducted in 1989 among two pairs of districts. A follow-up survey was conducted in 1991 among 25% of the villages in two districts among women who were service recorded users of sterilization and reversible methods during 1989-1990. The follow-up sample excluded users outside the health areas, those from the Muslim community, and duplicate names. The response rate ranged between 76% and 90%; nonresponses predominated among women aged 15-24 years, which would lead to a higher contraceptive prevalence rate (CPR). Adjustment was made for sampling error. The conclusion was that CPR may have been overstated by 1.0% to 1.5% in the follow-up survey. The findings showed that 18% and 23% of all acceptors in Bharuch and Panchmahals were duplicated names or misreports of users who were in fact nonusers. About 38-39% of pill acceptors in the two districts were found to be nonusers. 24-26% of recorded condom acceptors were found to be nonusers. Health workers misreports of acceptors in Bharuch for IUD users was 15% and over 27% in Panchmahals. The follow-up survey found about 21% unavailable for interviews (6-24% in Bharuch and 10-16% in Panchmahals). The follow-up sample was reduced from 452 acceptors to 292 acceptors in Bharuch and from 583 to 345 in Panchmahals. Contraceptive prevalence rates were thus 41% in Bharuch and 36% in Panchmahals. Sterilization records were the most accurate. Service statistics rates for the 4 districts in 1987 and 1989 were reported as higher than survey statistics. The findings suggest that reversible methods reported in service statistics were highly overstated. Target setting was thought likely to be responsible for the overstatement. Service statistics data may be useful only in follow-ups of users for home visits or in planning supply quantities, rather than as measures of service performance.