Exclusive Breastfeeding Rates and Factors Associated with Exclusive Breastfeeding Practices in Northern Tanzania: Measurement using Two Different Methodologies—24 Hours Recall and Recall Since Birth

1. Introduction

Exclusive breastfeeding (EBF) has been reported to be the best way of feeding infants from birth to 6 months of age.1 EBF has been defined by the World Health Organization (WHO) as giving only breast milk to the infant, without mixing it with water, other liquids, herbal preparations or food, with the exception of vitamins, mineral supplements or medicines.1,2 Exclusively breastfed infants have been shown to have lower rates of respiratory tract infections and diarrhea, better physical growth, neuro-developmental outcomes and in some countries higher intelligent quotient compared to infants who were not exclusively breastfed.3–5 Despite the benefits of EBF, this practice is still low compared to the WHO recommendations that at least 90 % of all infants less than 6 months should be exclusively breastfed6. It is currently estimated that only 40 % of infants less than six months are exclusively breastfed globally and 39 % in developing countries.7

The prevalence of exclusive breastfeeding has been improving over the years. However, there may be large variations from country to country or within a country. Studies from Ethiopia have shown a variation in EBF prevalence from 36 % in Northwest Ethiopia to 82 % in Oromia region.8–12 In Tanzania different studies have reported EBF prevalence ranging from 20 % in Kilimanjaro region, 25 % in Tanga region to 58 % in Kigoma.13–15 Using information from Demographic and Health Surveys (DHS), the prevalence of EBF in Tanzania has been improving over the years from 26 % in 1991-92, 41% in 2004-05, 50 % in 2010 and 59 % in 2015-2016.16

The method that has been used in many studies to present prevalence of EBF has been 24 hours recall. Many questions have been raised regarding this method and its ability to provide clear figures to present prevalence of EBF. A paper by Greiner discussed this issue in detail and showed that the method which has been used over the years has been exaggerating the actual EBF rates. This has been supported by other studies comparing data from 24 hours recall with recall since birth when calculating prevalence of EBF. In Eastern Uganda, for example, Engerbresten et al. showed the EBF prevalence at 1, 3 and 6 months to be 96 %, 81 % and 52 % when using 24 hours recall method compared to 45 %, 7 % and 0 % respectively when using recall since birth.17 However, Agampondi et al. reported that all retrospective evaluation methods overestimate the duration of EBF, whether using mothers recall (77.7 %) or calendar month (41 %) compared to prospective cohort design (23.9 %)18. There is thus contrasting information regarding the best method to measure EBF rates.17,18

Different studies have reported factors that influence the practice of EBF. The factors include that maternal age, education status, employment status, marital status, HIV status, place of delivery, and knowledge on breastfeeding, counselling on breastfeeding, antenatal care attendance and type of delivery.13,15,19–21 None of the previous studies have assessed if using different methods makes a difference when assessing factors associated with EBF within the same population. Given that there is a discussion on which method is better for measuring EBF rates there is a need for more knowledge concerning the choice of method when assessing predictors of EBF.

This study presents data from a cohort study where women were enrolled during third trimester of pregnancy in Moshi urban, Tanzania, and followed at delivery, 7 days and thereafter monthly up to nine months after delivery, and at each visit, information on breastfeeding using the two methods: 24 hours recall and recall-since-birth was collected.

2. Methodology

2.3. Breastfeeding cohort information

This paper is based on data from mother-infants pairs collected from delivery up to six months. Figure 1 depicts a flow chart for participants. Out of the 536 enrolled women, we had delivery information on 441 and 95 were lost to follow up. A total of 430 women formed a cohort for analyzing exclusive breastfeeding rates using the recall since birth and 24 hours recall and 364 women completed all the interviews up to 6 months.

Close

Figure 1

Flow chart of the study participants

Export to PPT

The reasons for loss of follow up were; women moved from one location to another without giving notice and giving wrong contact information. At enrollment, women were requested to give cell/mobile telephone numbers so as to contact them in case they miss the scheduled follow up. For women who missed the follow up, when we called the cell numbers, some were not in service and some belonged to their relatives/husbands. The relatives were given message to deliver to them which we are not sure if it was delivered. Of the 441 women with delivery information, 10 had still birth and one neonatal death which occurred 10 days after delivery (Figure 1).

3. Results

Comparison of women who came back even once (n=441) and those who did not come back at all after delivery (n=95) is presented in Table 1. The younger mothers (15-24 years) were significantly less likely to come back for follow up than the older women (p=0.01) and women who had one pregnancy were significantly less likely to come back than the others (p=0.004).

Table 2 shows the delivery information of the 431 women who came back after delivery at least once and had live birth. Most babies were born with weight above 2500gm (96.6 %). Most women delivered at a health facility (99 %). Ninety nine percent of women choose exclusive breastfeeding as their infant feeding option and 85 % and 95 % were able to initiate breastfeeding within the first hour after delivery and give colostrum respectively.

Table 3 presents the prevalence of EBF for each month using the two methods. There is a difference between the methods; 24 hours recall shows a higher prevalence rate than recall since birth throughout the six months. Predictors of EBF using recall since birth and 24 hours recall method are shown in Tables 4 and 5. A bivariate logistic regression using recall since birth method shows that women who have received counselling on infant feeding have increased odds [COR =2.1; 95 % CI (1.2, 3.5)] of practicing EBF compared to those who did not receive counseling (Table 4). In multivariate analysis, these results remained significant when adjusting for knowledge on EBF definition and appropriate breastfeeding knowledge [AOR=2.3; 95 % CI (1.2, 3.7)] (see Table 4).

In bivariate logistic regression using 24 hours recall, women who were unemployed had increased odds of practicing EBF compared to those who were employed [COR=1.5; 95 % CI (1.1, 2.5)]. Furthermore, women aged 35 - 49 years had decreased odds of practicing EBF [COR=0.28; 95% CI (0.1, 0.7)] than women of younger age and these associations remained significant after adjusting for all other factors in the model (see Table 5).

4. Discussion

In this study, it was found that exclusive breastfeeding rates decline before the children reach six months when using both the recall since birth method and 24 hours recall. However, prevalence and factors associated with EBF were different when using the two methods. Counseling had a positive effect on practicing EBF when using the recall since birth method, whereas employment and maternal age had a significant effect when using 24 hours recall.

The results showed that EBF rates were consistently lower when using the recall since birth method than when using 24 hours recall method. At 1, 3 and 6 months, the EBF rates were 96%, 74.4% and 24 % respectively, when using recall since birth and 100%, 88% and 39% respectively, when using 24 hours recall.

These differences are slightly less than what was found in a study from Eastern Uganda where EBF rates at 1, 3 and 6 months were 45 %, 7 % and 0 % when using the recall since birth method compared to 96 %, 81 % and 52 % when using 24 hours recall.25 Other studies conducted in Ethiopia and Mansoura reported significant differences in EBF rates using the recall since birth and 24 hours recall showing that the 24 hours methods overestimated the EBF rates.26,27 The results from these studies and our study indicate the importance of using both methods in estimating EBF rates. There may be many explanations to the differences in EBF rates when using the two methods. When using the 24 hours recall method the mother only has to recall for the past 24 hours and if the child was not given anything apart from breast milk in that time period, this mother will be classified as practicing EBF. This is regardless of whether the child is given water, other liquids, herbal preparations or food on other days. Thus this method is likely to overestimate the rates of EBF. The recall since birth methods as used in the preset study went more into detail on breastfeeding and introduction of foods and drinks and questions were repeated monthly. The moment a mother started to introduce any other food or drink apart from breast milk it was considered as the discontinuation of EBF. This way of measuring EBF by recalling all the breastfeeding practices since birth, seems to be more in line with WHO definition of EBF.

A standard method for assessing exclusive breastfeeding, used in the Demographic and Health Survey (DHS), is 24 hours recall at 6 months. DHS data present point prevalence from 0 - 6 months, where mothers were asked about breastfeeding practices in the past 24 hours preceding the survey.28 There are many considerations behind the choice of methods connected to the aim of the study and available resources. It should be kept in mind that the choice of the method will have relevance for the results and potential for appropriate action. In cases where the EBF rates are substantially overestimated, this might not trigger the effort to be taken to improve the situation. Thus, there is a need to rethink the best way to report matters that are important to health like EBF.

With regards to the area-specific issues around the study, there seems to be a slight improvement in EBF practices, when comparing the data from the present study with another study which was also done in Moshi, Tanzania. That study was based on cohort data collected in 2002 – 2006; EBF was measured using recall since birth at one, three and six months and the rates were 49%, 22% and 0.2% respectively.29 The improvement can be due to the counseling the mothers receive from health providers on the importance of EBF.

In the present study, those women who received counseling were more likely to practice EBF for 6 months than those who did not receive any counseling, as shown when using the recall since birth method. This has also been observed in studies from other countries, such as Nigeria, Egypt, and Ethiopia.11,30,31 This may be because every time women come for antenatal visits they receive information about infant feeding and the importance of EBF. It is important to continue giving antenatal counseling to all pregnant women that attend clinic since it has been shown to have benefits to the promotion on breastfeeding.

Women who were aged 35 to 49 years were found to be less likely to practice EBF compared to women of younger age when using 24 hours recall. This is different from other studies where mother’s age had a positive effect on EBF. A Canadian study indicated that women who gave birth at older age were more likely to practice EBF.32 This was also seen among Norwegian women in another study.19 In the present study, the negative effect might be because many women tend to use their own experience to a larger extent than what they have learned in the antenatal clinics. The older women have given birth many times; hence, it is possible that they are following whatever they were practicing before without taking into consideration the new messages that were taught in the clinic.

Women who are unemployed have increased odds or practicing EBF up to 6 months than women who are employed in the formal sector. This has also been seen in studies from three different parts of Ethiopia where unemployed women were more likely to practice EBF.10,11,33 This shows that policies about maternity leave need to be looked at so as to promote EBF for employed women. There is also a need to improve work place environment locations, such as infant nurseries, where women with infants can leave their children and come to breastfeed their infants without any problems.

5. Conclusion and Global Health Implications

This study has shown that there is a difference in the prevalence of EBF when using the two methods: the 24 hours recall gave higher EBF rates than recall since birth. Measuring the duration of EBF is complicated, since factors like timing, recall bias and method of analysis can affect the outcome. Based on the results from this study, and considering the way the data were collected, it can be argued that the higher EBF obtained with 24 hours recall represents an overestimation and thereby an overly positive picture of the situation. The 24 hours recall method has been used in many studies due to the fact that it is cheap to conduct in large population groups. It is important to keep in mind that the type of method used can have implications for both prevalence and predictors of EBF. However, both methods have their strengths and limitations and may reveal important points. Based on data from the recall since birth method, this study indicated those mothers who have received counseling on infant feeding are more likely to practice EBF. Therefore, the authorities should ensure that all primary health care facilities are equipped to be able to deliver such counseling. The data from 24 hours recall indicated that mothers that are employed had more chances of feeding their infants exclusively up to six months; hence the employers should consider ways to enable employed women to exclusively breastfeed their infants by considering the introduction and or strengthening of paid maternity leave.