A Path Analysis to Identify Factors Influencing the Provision of Water in Addition to Breast Milk by Mothers of Infants under Six Months of Age in Conakry and Kindia Regions, Republic of Guinea

Water provision to infants under six months of age (IU6M) can hamper exclusive breastfeeding (EBF). Understanding factors and their relationships influencing this practice is needed to tailor EBF promotion programs. Using a validated questionnaire, this study aims to identify pathways in which individual factors and the environment interact to affect the provision of water in addition to breast milk among 300 mothers of IU6M. Our finding shows that 75% of mothers intended to provide water in addition to breast milk to their IU6M and that about 60% reported doing it. Results of the final path show that the subjective norm/SN (β = 0.432, p < 0.001), the attitude (β = 0.349, p < 0.001), and to a lesser extent the perceived control/PC (β = 0.141, p = 0.005) predict the intention of mothers to provide water in addition to breast milk to their IU6M. The environment scores predict the attitude (β = 0.210, p = 0.001) and the SN (β = 0.284, p < 0.001). Having the mother practicing early breastfeeding initiation at birth positively predicted the PC score (β = 0.157, p = 0.017) and predicted an increasing score of SN (β = 0.221, p = 0.003). Even though predicting the final behavior is complex, this research provides directions to nutrition education programs to tailor their content to the context and be more efficient in reducing the proportion of women providing water to their IU6M, hence contributing to the improvement of EBF.


INTRODUCTION
Breastfeeding is a cost-effective, proven public health intervention with significant positive impacts on child morbidity and mortality while also providing health benefits to mothers. It is estimated that widespread breastfeeding would prevent worldwide 823,000 deaths per year, or 13.8% of deaths of infants under 24 months in 75 countries, in addition to averting 20,000 maternal deaths annually [1].
Despite progress in recent years, only 42% of children under six months are exclusively breastfed globally [2]. In the West and Central Africa region, the rate is lower at 32% [2]. One practice that undermines exclusive breastfeeding is the provision of water to babies aged below six months. In the Republic of Guinea, 35% of children under six months are provided water in addition to breast milk, a practice that may be detrimental to their nutritional and health status and development [3].
Individual and environmental factors determine the mother's decision to provide water to their baby under six months, hence not exclusively breastfeeding [4,5]. To address the provision of water to infants under the age of 6 months, it is essential to understand the underlying factors of this practice for effective behavior change interventions [6]. As such, behavioral theories are of valuable use to investigate which factors, either individual or environmental-related, determine the provision of water [7].
The theory of planned behavior (TPB) has been used to understand what predicts or explains breastfeeding-related behaviors [8]. According to the TPB, the attitude, the subjective norm, and the perceived control over a behavior are key determinants of behavior [9]. Research using the TPB has shown a positive association between the attitude and the intention to breastfeed [10][11][12][13][14][15][16][17][18][19][20]. Regarding the perception of the subjective norm, results are mixed as some studies have shown a positive relationship to breastfeeding [12,14,16,20] while others do not [11,15,21]. Some studies have revealed that perceived behavioral control has a significant positive effect on exclusively breastfeeding intention [14][15][16]18,20,22]. Environment-related factors such as the existence of community-based interventions (e.g., provision of group counseling or education) and work conditions (e.g., access to maternity leave and pay breaks) also appear to be conducive to exclusive breastfeeding [4].
To our knowledge, there is no comprehensive study that has investigated at the same time individual and environmental factors underlying the provision of water in addition to breast milk among infants under six months of age, neither the interrelationships between these factors. Therefore, using an extended version of the theory of planned behavior [23], this study aims to identify pathways by which individual factors (attitude, subjective norm, and behavioral control) and the environment interact to affect the provision of water to infants under six months of age in addition to breast milk. This research will help define an intervention, implement it and assess its impact on the exclusive breastfeeding (EBF) rate among this population.

Study Area
This study was conducted in Conakry and Kindia regions in the Republic of Guinea. Both regions contain 30% of Guinea's total population and encompass the bulk of urban households [24].
The Republic of Guinea is located on the Atlantic coast of the West Africa region, with about 13 million inhabitants [25]. With a human development index of 0.477, the country is in the "low human development" category [26]. The overall literacy rate is 32%, standing at 44% among men and 22% among women. The gross domestic product per capita was estimated at 962.84 US$ in 2019 [27].
In the Republic of Guinea, the mortality rate among children under five years has declined from 177/1,000 live births in 1999 [28] to 111/1,000 live births in 2018 [3]. During the same period, neonatal mortality decreased from 48/1,000 to 32/1,000 live births. The EBF rate increased from 11% in 1999 to 33% in 2018. No regional data are available on EBF practice.
This study is part of larger research investigating psychosocial and environmental factors associated with the provision of water in addition to breast milk by mothers to infants under six months of age; subsequently, to define, implement and assess an intervention to reduce the proportion of mothers implementing this practice.

Design and Sampling
This research has a quasi-experimental design. The sample size was determined to enable the detection of a ten-percentage point difference between proportions of mothers in the control and intervention groups who give water in addition to breast milk following the implementation of the intervention assuming an 80% power and 5% significance level and a 10% dropout rate [29]. The most recent data indicate that 34.5% of mothers give water to their children under six months of age [3]. The sample included 300 mothers of infants under six months of age, with 150 spread out in each study's arm, namely the control and the intervention group.
All health centers with their population coverage were listed in each region, and four were randomly selected, with two per region. In each health center, all mothers of infants under six months of age were invited to participate.
To be included in this study, the mother has to be aged 20 years or older and having given birth, through normal vaginal delivery, to a unique child with a birth weight greater than 2.5 kg. Whether or not the child was exclusively breastfed was not a selection criterion. The study excluded mothers of infants under six months of age who had severe medical (e.g., mental health issues), obstetric complications (c-section), babies with severe medical conditions (e.g., HIV/AIDS), or congenital malformations, and mothers aged below 20 years. Each mother's maternal and child health card was examined to check if the inclusion criteria were met.

Preparatory Work
As described previously [30], eight local enumerators were recruited and trained on the study methodology. The theoretical training was complemented by in-class and field practical exercises. During the training, questionnaires were translated into the three main local languages (Soussou, Maninka, and Pular) spoken in both regions and tested among a group of women to ensure its clarity.

Theoretical Framework
The extended model of the theory of planned behavior/TPB [23,31,32] was used to guide this research (Figure 1). According to the TPB, the intention of an individual is the main predictor of behavior. Based on the TPB, in our setting, the mother's intention to provide water in addition to breast milk would be determined by three constructs, a) her attitude toward the behavior, b) her perception of the subjective norm about the behavior, and c) her perceived control over the behavior [9]. In turn, the mother's beliefs ("behavioral belief") will determine the attitude about the behavior as well as the evaluation she makes about consequences ("evaluation of consequences") in adopting it or not [9]. The subjective norm is determined by the importance mother attaches to the opinion of people or groups of people around her ("normative beliefs") and by her "motivation to comply" with their opinion [9,33]. The perceived control is composed of the "control beliefs" and the "perceived behavioral control"; in other words, the mother's degree of control believes she can exercise over a given behavior.
Moreover, the perceived control might also directly predict the behavior [9,33]. Environmental factors are external social and physical characteristics that can influence the three constructs of the TPB (through moderation) as well as the transition from the intention to a concrete behavior [23]. They refer but are not limited to health centers, family, workplaces, community, maternity and paternity leave policies, childcare benefits and health insurance, sociodemographic characteristics.

Data Collection
The data collection was performed in December 2020. Data on EBF and the provision of water (the behavior under study) were collected using the infant and young child feeding module part of the Demographic and Health Survey questionnaires package used in the Republic of Guinea. Sociodemographic characteristics (household composition, namely age, gender and level of education of each member, housing characteristics, ownership of assets) were collected using an adapted version of the household questionnaire of the same package [3]. This data was collected through a face-to-face interview with the household head and/or the mother of each infant at home. The health card of each child was also checked to record their date of birth. Data on each theoretical framework's construct was collected using a validated and reliable questionnaire [30], developed based on qualitative information collected through focus-group discussions. This questionnaire includes 58 items (including one item to measure the intention) spread out, as follows for each construct: attitude, 20; subjective norm, 11; perceived control, 12; environment-related factors, 14 items. The questionnaire was administered to each mother during a face-to-face interview conducted by enumerators at home. Beforehand, the enumerator explained to each participating mother how to express their answers on the questionnaire items on a Likert scale before reading each item and responses' options. Each mother was invited to indicate with her finger or mark each respective scale's selected answer with a pen for every item.

Behavior Understudy
To assess if the child was provided water, the following questions of the infant and young child feeding module were used: a) if the child was breastfed yesterday (day or night), b) if plain water was offered to the child yesterday (day or night), and c) if any other liquid was offered to the child as it may indicate the provision of water under a different name/format such as bottle water. If the mother responded "yes" to the first question, "no" to question b, and "no" to question c, the mother was classified as not providing water to her baby and was assigned a score of zero. If the answer to the first question "a" was no or the answer to one of the questions "b" and "c" was a "yes", a score of one (1) was given.
The EBF status and whether the infant benefited from early initiation after birth were assessed using WHO guidelines [34].

Child, Mother, and Household Characteristics
Household, mother, and child socio-demographic data were entered directly into version 25.0 of IBM's SPSS. Descriptive analyses (chi-square tests) were performed to assess differences in proportions of mothers providing water to their infants between categories of the child, mother, and household characteristics.
A factor analysis using principal axis factoring was performed on the correlation matrix to define each household's socioeconomic score. Initially, 39 items on ownership of assets and housing conditions were used.
Items for which the frequency of distribution indicated that at least 90% of households were having or not the items or benefited from a particular service were removed. The final score considered 20 items (animals and farming land, radio, television, refrigerator/freezer, gas stove, table and chairs, wardrobe, couch, clocks, shelves, air conditioner, cell phone, motorcycle, bicycle, bank account or participate to community funds called "tontine", van), which explains 20,94% of the factor's total variance and composition of the socioeconomic score. The Kaiser-Meyer-Olkin Test, which assesses the data's suitability for the factor analysis, was 0.775, which is satisfactory [35].

Measurement of Psychosocial and Environmental Constructs
For each item of the questionnaire and each mother, a numeric value ranging from -2 (e.g., strongly disagree/unlikely/disapprove) to +2 (e.g., strongly agree/likely/approve) was assigned to each response on the Likert scale [32]. Thereafter, frequency distributions were performed to check for data completeness and accuracy.
Also, in preparation for subsequent analysis, scores of worded items so that the option to answer "in agreement" with a statement that did not reflect a positive attitude towards the behavior were reversed. For instance, for the question to the mother on her intention to provide water to the child, which was formulated as follows:" I intend to give water to my child under six months in addition to breast milk," responses options ranged from less likely (score of -2) to very likely (+2). For such items, responses were reversed or, in other words, a negative score (-2) was assigned to "very likely" while a score of +2 was allocated to the "very unlikely" response option.
Two types of measures were used to assess the extended TPB psychosocial constructs: indirect and direct, usually strongly associated [36,37]. According to Montaño and Kasprzyk [37], both measures are interesting if one focuses on intervention messages. Indirect measures help understand what "drives behaviors", while direct constructs are usually more associated with the intention. Investigating the relationship between each item under each construct and the indirect measure is also recommended [36]. Consequently, indirect measures were computed for the attitude by multiplying each mother's score on a behavioral belief concerning an outcome by the corresponding outcome (Supplementary material) [36,37]. For example, the mean score on the item "giving water help prevent constipation" (a behavioral belief concerning an outcome) and "for you, preventing your child from getting constipated" (the corresponding consequence of giving water) were multiplied. The same procedure was applied to other pairs of items used to assess the attitude construct (see Supplementary material) as well as to items of the subjective norm and the perceived control. After that, for each of the three constructs and each mother, the indirect measure was calculated by averaging all mean scores of paired items and individual items' raw scores. Besides, for each mother, direct measurements of each construct were calculated based on responses of items (attitude, the score of one unique item; subjective norm, mean of the score of two items; perceived control, mean score of three items). These direct measurements were differently worded from a semantic perspective. Instead, they asked the mother to rate behavior as useful / not useful, in agreement/disagreement, or as under her control of not [37]. For the environment construct, the mean score of all items was calculated for each mother.
Spearman correlations were performed to investigate bivariate relationships between a) categorical scores of individual items for each behavioral, normative and control beliefs and mean scores of indirect and direct measures for the attitude, the subjective norm and the perceived control constructs, b) mean scores of indirect and direct measures, c) categorical (only for the attitude given that there was only one item) or continuous scores (subjective norm and perceived control) of direct measures and the categorical (and reversed) score on the intention.
Path analyses with version 1.6 of MPlus 8 [38] were used to investigate pathways by which psychosocial and environmental factors influenced the provision of water in addition to breast milk among mothers at baseline. Figure 2 presents the hypothesized pathways with each external and environmental variable predicting the three directly measured constructs, as suggested by Godin [23]. Figure 3 shows the hypothesized pathways with external and environmental variables moderating the relationship between the intention and behavior as indicated by the same author.
The assessment of the final model fit was performed using the following indicators and criteria: a) chi-square statistic above 0.05, b) a root mean square error for approximation (RMSEA) < 0.08 with confidence intervals < 0.10, c) comparative fit index (CFI) and Tucker-Lewis index (TLI) > 0.90, and standardized root mean residual (SRMR) < 0.08. Insignificant pathways were removed using the Wald method while adding pathways was done with the Lagrange Multiplier method [39]. Theoretical considerations of possible pathways were examined before making changes to the model. For the final model to be accepted, all postulated pathways had to be statistically significant (p < 0.05). Standardized regression coefficients were computed for all remaining pathways to the size of each statistically significant path's relation.

Populations' Characteristics and Descriptive Analysis
Tables 1 and 2 depict the socio-demographic characteristics of infants, mothers, households, and differences in proportions of mothers providing water to their children for each characteristic. Children's mean age was 1.83 months (standard deviation / SD = 1.03), while their mother's average age was 24.8 years (SD = 4.9). About 50% of mothers were housewives, were from the Soussou ethnic group, and had some primary school education. Almost 60% received at least 4 antenatal care (ANC) visits in their last pregnancy. The majority (68%) gave birth in a public health facility and were assisted by a skilled birth attendant. More than half of the mothers (57%) reported giving their children water in addition to breast milk (result not shown).
Among women who practiced early initiation of breastfeeding and those who attended at least 4 ANC visits, the proportions of mothers giving water to their children were lower than women who did not initiate breastfeeding within the hour after birth (p = 0.022). The same findings were found for those whom a skilled birth attendant assisted at delivery. Moreover, the proportion of mothers giving water to their infants was higher among those who did not attend at least 4 ANC visits (p < 0.001, Table 1).
Tables 3a and 3b report frequency distributions on five-response scales and mean scores for each item of the questionnaire and mean scores for indirect and direct constructs. Overall, about 75% of mothers intended to provide water in addition to breast milk to their babies. The mean score on the intention was 0.86 out of a maximum of 2. Over 70% of mothers agree that implementing the behavior is necessary because water is the drink for the child or will avoid fatigue due to thirst or the throat becoming dry. The same proportion also believes that providing water in addition to breast milk will not cause abdominal pain. Over 80% of mothers consider it as important/very important to give water to prevent the child from being thirsty, having a fever, losing weight, or having abdominal pain and getting constipated. Around 70% of mothers considered giving water in addition to breast milk to their children as useful/very useful. Mean scores for the       indirect and direct measurements of the attitude construct were respectively 0.51 and 0.74.
About 40% of mothers believe that their mother-inlaw, the grandmother of the spouse, their grandmother, and their mother will approve/strongly approve if they give water in addition to breast milk to their babies. About a third of the mothers were motivated to comply or adhere to these persons' perceived opinions. Moreover, around 60% of mothers agree/strongly agree that the most influential people in their surroundings think that water should be given to their children. Mean scores of the indirect and direct measurements of the subjective norm construct were 0.49 and 0.59, respectively.
Regarding the construct on the perceived control, it was unlikely/very unlikely for 7 out of 10 mothers to not provide water if they do not have financial means, do not return to work, or do not leave their baby with someone else. For five and six out of 10, respectively, it was also unlikely/very unlikely that they would not provide water if they did not have to give medication or breast milk substitutes. For about 60% of them, it was likely/very likely that, if they wanted to, they could easily not provide water to their child in addition to breast milk. However, around 45% perceived it as difficult/very difficult to not giving water to their baby. The mean scores for the indirect and direct constructs on the perceived control were both around 0.
Environment-related items that may stop or facilitate mothers to provide water in addition to breast milk to their children are listed in Table 2. For more than 60%, it was unlikely/very unlikely that not having access to individual counseling or education group would stop them from providing water. The same findings were observed regarding giving birth in a health facility, being assisted by a skilled birth attendant at delivery, or receiving support to initiate breastfeeding after birth. On the other hand, more than 50% reported that it was likely/very likely that having the child exposed under the sun at the workplace or elsewhere may facilitate the provision of water. About 40% responded that it was likely/very likely that having received breast milk substitutes donation or being exposed to their advertising may facilitate the provision of water. For about 3 out of 10 mothers, it was likely/very likely that having a place to get water of good quality and financial means to buy it may facilitate the provision of water in addition to breast milk. The mean score of the environment construct was -0.34. Tables 4 to 6 show correlation results between items and indirect measurement as well as between indirect and direct measures for each psychosocial construct of the TPB. For the attitude, all individual items related to behavioral beliefs were correlated (p < 0.01) to the indirect measurement of the construct and to a lesser extent to its direct measurements, too ( Table 4). For the subjective norm beliefs, all individual items were correlated to the construct's direct measurement; except for one item, none was associated with the indirect measurement ( Table 5). As for the perceived control, all items assessing control beliefs were associated with the indirect measurement and, to a lesser extent, with the direct measure of the perceived control construct except one item ( Table 6). For each construct, indirect and direct measurements were correlated (p < 0.01). Besides, except for the indirect measurement of the perceived control, all constructs of the extended TPB, either direct     or indirect, were correlated to the intention of providing water in addition to breast milk among mothers ( Table  7). Correlation coefficients between measurements of direct and indirect constructs and the intention were respectively significant at p < 0.001 and p < 0.05.

Final Path Models
The results of the path analysis, including the standardized regression coefficients, are presented in scores also predict the attitude (β = 0.210, p = 0.001) and the subjective norm (β = 0.284, p < 0.001), but not on the perceived control. Having the mother practicing early breastfeeding initiation at birth positively predicted the perceived control score (β = 0.157, p = 0.017) and predicted a diminishing score of the subjective norm (β = 0.221, p = 0.003). The subjective norm predicted the attitude (β = 0.547, p < 0.001), while, in turn, the attitude could predict the perceived control score (β = 0.374, p < 0.001). As for the explained variance of each endogenous variable in the model, 13.2% of the subjective norm score was explained, 43.9% of the attitude score, 17.8% of the perceived control score, 59.2% of the intent score, and finally, 18.2% of the behavior.

DISCUSSION
This research used the extended theory of planned behavior to identify pathways in which individual factors (attitude, social norm, and perceived control) and the environment interact to influence the provision of water by mothers to their infants under six months of age in addition to breast milk.
Unfortunately, no regional data could be retrieved in the literature regarding the proportion of infants below six months of age receiving water in addition to breast milk. In our setting, findings indicated that most mothers (57%) reported giving water to their infants in addition to breast milk. This proportion is much higher than the national figure of 35% [3]. Moreover, the vast majority of mothers (about 75%) indicated that they intend to provide water in addition to breast milk to their babies. To our knowledge, no studies have investigated the mother's intention of providing water to their baby aged below six months in Africa. However, in India, Behera and Anil Kumar reported that 30% of pregnant women intended to exclusively breastfeed their baby without providing them with extra fluid or other foods [18]. To determine women's intention to practice EBF, the authors asked three questions. One of them was related to their intention to give water to their baby during EBF. Unfortunately, data on responses to this specific question were not presented in their paper.
Among our population, it seems that the intention was somewhat translated into actual behavior. In fact, the final path analysis model shows that the intention was associated with the behavior but that perceived control does not directly influence it. This result was expected as, according to the TPB, the intention is the strongest predictor of the behavior [9]. Although not explicitly related to the provision of water, our findings are consistent with previous studies that have assessed factors influencing EBF in different contexts and observed an association between the intention and the behavior [12,[14][15][16]20,40]. Moreover, in a metaanalysis and structural equation modeling on the efficacy of the TPB in predicting breastfeeding, the path estimates between breastfeeding intention and the behavior were 0.45, a value similar to ours (0.426). Identical to our context, results from the meta-analysis mentioned above revealed that the perceived control was not a direct and significant predictor of breastfeeding [41].
The final analysis also shows that all the constructs (subjective norm, attitude, and perceived control) influence the mothers' intention to provide water in addition to breast milk to their infants under six months of age. However, the subjective norm and attitude appear to have a more significant influence. Moreover, the subjective norm influenced the attitude, which in turn also impacted the perceived control positively. Our results concur with Guo et al.'s [41] findings from highincome countries, which show that all three constructs were significant predictors of breastfeeding (any). The authors observed a strong, though a two-way relationship as opposed to one-way association in our context, between the attitude with the subjective norm, the perceived control and the subjective norm as well as between the attitude and the perceived control. This result is an interesting finding that may be attributed to different contexts, high-vs. low-and middle-income countries, which certainly deserve further research. In any case, perceiving the subjective norm as favorable to water provision in our areas could lead to a mothers' attitude favorable to providing water. In turn, a positive attitude may lead them to be perceived as more capable of doing it.
Our findings are similar to previous studies investigating the relationship between women/mothers' attitude and their intention to initiate, pursue or implement any breastfeeding [12,15,16,18,[20][21][22]42]. In their meta-analysis, Guo et al. [41] found that attitude was the strongest predictor of mothers' breastfeeding intention. In our context, mothers' behavioral beliefs and the evaluation of consequences of not providing water contributed to a less favorable attitude regarding the behavior under study. Thus, behavior change communication strategies to improve EBF rates should address these beliefs and their consequences for which an important proportion of mothers agree and/or strongly agree with such as the importance of giving water because milk is the food and water is the drink or because water will avoid fatigue due to thirst or the throat becoming dry.
Furthermore, the results showed that the subjective norm was a strong and positive predictor of intention. These results indicate that the mother's perception of her immediate surroundings (her spouse, her mother, her mother-in-law, her sister) was supportive in giving water in addition to breast milk, which directly influenced her intention to provide water. The relationship between the subjective norm and breastfeeding has also been reported in previous studies [18,20,41]. Hence, to be efficient, interventions targeting mothers of infants below six months of age aiming to reduce water provision in addition to breast milk should also involve influential peoples of their immediate surroundings to ensure that they will be more supportive to the mother. In turn, this may lead to a change in the mother's normative beliefs and compliance with these beliefs, which will be more conducive to the non-provision of water.
Moreover, and as also observed in other settings [18,20,41], the perceived control affects the mother's intention to breastfeeding the provision or not of water in addition to breast milk. In our context, mothers could be supported to overcome what they perceived as barriers in not providing water. As such, making efforts to implement solutions (or what we have identified as factors that could prevent them from providing water in our study) that they have themselves identified could be a good start.
In research areas, the impact of the environment on the behavior was through mediation instead of a modulation role to the operationalization of the intention to the actual behavior as postulated in model B. The environment directly influenced mothers' attitudes and subjective norms. Thus, acting on specific elements of the environment such as improving access to counseling and group session on EBF supporting environment could contribute to a more positive mothers' attitude as well as leading to a perception of her subjective norm, which may be more favorable to the non-provision of water and ultimately, increases her intention not to provide water. Besides, the provision of support at birth for early initiation of breastfeeding has a positive and direct impact on the perceived control, as well as the subjective norm. Support for early breastfeeding initiation at birth has been associated with EBF practice in other contexts [43,44]. The positive relationship between early breastfeeding initiation and the subjective norm may help make breastfeeding the norm in society.
In accordance with Montaño and Kasprzyk's [37] reflections, in our context, direct construct measurements were more strongly associated with the intention than indirect measurements, and both were correlated. Besides, investigating the relationship between each item under each construct and the indirect measure has contributed to identifying specific content that drives behavior and could be delivered through behavior change strategies [36].
Findings from this research reiterate the need to create a supportive environment for behavior change to influence mothers' attitudes and hence, address specific beliefs and knowledge gaps. It also highlights the importance of targeting mothers with behavior change programs but also people in their surroundings who significantly influence their behavior if one wants to change their perception of the subjective norm and ensure that EBF will be the social norm. In Guinea, improving access to regular individual counseling and group education sessions on EBF during ANC visits and/or immunization as well as providing support to initiate breastfeeding at birth could shape the attitude, the subjective norm, and the perceived control to make them favorable to the non-provision of water in the first six months. This research has many strengths that need to be acknowledged. This is the first study that examines pathways through which individual and environmental factors influence water provision among mothers of infants under six months. Second, a valid and reliable measurement tool was used to identify interest behavior determinants and was initially theory-based. Third, our findings offer significant directions for improving the content of behavior change programs in Guinea and personalized them to enhance their efficacy. Our research also has some limitations. First, the results cannot be generalized to the entire mothers of infants below six months. Second, given the dichotomic measure of the behavior, this may have limited the path analysis's ability to explain the variance associated with the behavior.

CONCLUSIONS
This research suggests that the intention to give water can be thoroughly predicted by the mother's attitude, perceived control, and subjective norm scores, but predicting the final behavior is more complex than simply considering the intent. Yet, our findings can help tailor nutrition education programs to reduce the proportion of mothers providing water to their infants, and therefore improved breastfeeding rates.

DISCLOSURE STATEMENTS
The authors declare that they have no conflicts of interest.