This study used month-by-month calendar data, similar to those collected in many Demographic and Health Surveys (DHSs). However, unlike DHSs that rely on a 5-year recall period, data were collected prospectively at 4-month intervals, which should lead to much more accurate dating of key events, such as resumption of menses and initiation of contraceptive use. In addition, in contrast to DHS, the study focused on women who had recently delivered and who resided in urban informal settlements.
The use of longitudinal data allowed us to control our analysis of contraceptive use and non-use for exposure to sex and risk of pregnancy on a month-to-month basis, which is typically not possible using cross-sectional data. The results show that women in the urban slum communities resume sexual relations quite early (50% by the third month), but relatively few initiate contraceptive use during the first six postpartum months. While this may not be an issue for women who are not sexually active or who are partially protected from pregnancy by postpartum amenorrhea, particular attention needs to be paid to the group of women who have experienced a return of menses, are sexually active, and are not using any form of contraception. This group tends to peak between the third and sixth month of postpartum. The proportion of women in this group remains stable up to the 12th postpartum month, but the proportion of those becoming pregnant increases steadily, with close to 12% being pregnant by the end of the first year after giving birth. It may be important to further understand the social context of such risks and whether being a resident in an informal settlement presents particular problems of access to reproductive health services for postpartum mothers.
The observed pattern where resumption of menses among postpartum women acts as a trigger or reminder to start using contraceptives shows that women in slums understand this postpartum sign of return of full fecundity and the associated increase in the probability of having another pregnancy thereafter. This relationship between the adoption of contraception and resumption of coitus has previously been observed in studies analyzing calendar data elsewhere.43
However, by associating the return of menses with the risk of another pregnancy, women may forget that ovulation precedes the appearance of menses with even higher likelihoods of ovulation occurring as the postpartum period gets longer.43,44
On the other hand, the cautious group of women who initiate contraception before the return of menses acquire both natural and contraceptive protection against unwanted pregnancies. However, this advantage only counts if the selected method is a permanent one or if consistent use of the method is achieved without discontinuation. In a study of Peru and Indonesian women, results showed that women who initiated the use of pills and IUDs within the first 6 months were more likely to be pregnant after 2 years of childbirth than women who initiated similar contraception after 6 months postpartum.43
Therefore, early adoption of a contraceptive method may not necessarily translate into adequate birth spacing if continuation rates are low. This is a likely occurrence in settings such as informal settlements where education levels for women are low and constant supply or access to a given contraceptive method is not guaranteed. Following additional data collection, the analysis of the association between contraceptive initiation and continuation or discontinuation will be explored.
Among this study population, the leading choices of contraception, namely injectables and pills, are consistent with the pattern observed in the recent 2003 Kenya Demographic and Health Survey.30
However, for slum communities where the HIV prevalence is about 11.5%,45
the reported use of condoms during postpartum remains very low. Among communities where extramarital relations are common around the time of pregnancy and postpartum, low condom use may have implications for HIV/AIDS transmission.46,47
Condoms generally provide dual protection against pregnancy and sexually transmitted infections.23,48
The results from the 2003 Kenya Demographic and Health Survey show that approximately 5.3% of women reported having used a condom during their recent sexual intercourse in the previous 12 months.30
This percentage was 1.9% for women who were married or cohabiting. Therefore, the observed rates in the current study compare well with Kenya national figures, since about 80% of the women in the current study were married or cohabiting. Better ways to promote condom use among married women need to be explored. For instance, in Ivory Coast, women who believed strongly in the cultural taboo that sexual relations during postpartum would harm the breastfeeding infant were more likely to accept the use of condoms with their husbands to minimize the risk of extramarital sex, as long as the semen remained in the condoms without direct contact with the breastfeeding woman.25,49
Alternatively, postpartum mothers may be reluctant to use condoms that have higher failure rates as well as the difficulties in maintaining the consistent use of condoms due to gender power imbalances in sexual relationships especially when compared to hormonal contraceptive methods such as pills and injectables.50
Despite this limitation, the results demonstrate the increased need for promoting the dual role of condoms as tools for HIV/STD and pregnancy prevention by HIV control and family planning programs in Africa.
In Kenya, just like any other sub-Saharan African country, contraceptive prevalence rates remain low but with considerable differences across educational and socioeconomic groups.35
In the current study, among the category of the exposed months, 63% were protected by a modern contraception. These findings were unexpected, although slums have previously shown lower fertility rates, at four births per woman in 2000, than the rest of Kenya as a whole at 4.9 births in 2003,35
and this may point to concerted efforts of many programs targeting provision of better reproductive health services to slum communities. Overall, maintaining access and availability of modern contraceptive methods especially hormonal contraceptives which are suitable for spacing remains a key challenge for meeting the need among postpartum women from less privileged societies such as those in informal settlements in Nairobi.
There are several limitations that we wish to highlight. One of the key features of carrying out longitudinal research in urban settlements is the high attrition rates due to migration. About 21% of the first cohort was lost to follow-up by the second wave. The corresponding figures for cohorts 2 and 3 were 25% and 27%, respectively. The political instability that affected Kenya during the early months of 2008 is largely responsible for the relatively high attrition for the first and second cohort. Other reasons accounting for loss to follow-up include deaths of mothers and internal changes of residence that often delays linking residents to their new locations. However, by limiting the study to the first 12 months, the effects of attrition were limited since most women were covered for this period at the time of the first or second interview.
In conclusion, the main lesson from these results is that a large proportion of women in slum communities run the risk of early postpartum conception because of delayed uptake of contraception. A distinctive feature of the period surrounding childbirth is the high motivation and intensity of contact between women and health care providers. Periods of antenatal or postpartum or routine child care visits for vaccinations need to be explored more since they present opportunities when women may be particularly receptive to messages concerning their reproductive health and that of the child. Therefore, supporting policies and programs for the integration of family planning services with postpartum contraception services present a valuable prospect to reach a large number of slum women with an unmet need during postpartum. Equally, including postpartum contraception in the training materials for traditional birth attendants is the key, since majority of the women especially in low-resource settings such as informal settlements do not deliver at designated health centers but report having delivered at home or home of a traditional birth attendant.