Introduction

Best practices in infant and young child (IYCF) feeding are essential for improving the nutritional status of children and promoting healthy growth and development [1–3]. Yet both breastfeeding and complementary feeding practices remain suboptimal in developing countries, where the prevalence of exclusive breastfeeding is about 39% [4]. In Viet Nam, <20% of infants are exclusively breastfed in the first 6 months, and complementary foods are often introduced too early and tend to be of poor quality [5].

High-quality and timely IYCF counselling has been shown to be an effective strategy to improve both child feeding practices and the nutritional status of children [6–8]. Community-based initiatives using peer counsellors have been successful in supporting exclusive breastfeeding in several countries [9, 10]. Also, the delivery of clear educational messages on complementary feeding practices which emphasize the importance of hygienic, high-energy and appropriate home-prepared foods, coupled with the timely introduction of complementary foods, has been shown to lead to significant gains in the height and weight of children in low and middle-income countries [11].

Scaling up nutrition interventions to achieve adequate delivery, coverage, and utilization often requires integration with established delivery platforms, such as government health systems. However, health systems worldwide face challenges in improving service access and utilization [12], influenced by resource availability, service quality, cost, health beliefs and other user characteristics. As health systems increase their focus on nutrition to prevent poor health outcomes [13], further challenges may be seen in the delivery of preventive services that are not perceived as an immediate threat to illness [14] and require repeated use for sustained behavior change. Therefore, studies of the drivers of service utilization are needed to understand how to improve access and utilization of specific health services, particularly preventive services such as nutrition education.

Until recently in Viet Nam, government nutrition programs had not systematically integrated IYCF counselling services into routine health services [15, 16]. Health facilities lacked personnel to provide counselling services and adequate materials and infrastructure (e.g., a separate counselling room and sufficient guidelines and educational materials to support breastfeeding practices) [17]. In addition, health providers often had outdated information about optimal IYCF practices and lacked sufficient training and practical guidance on how to provide counselling to mothers. The quality, frequency, and follow-up of in-service training on IYCF was also poor [17].

To address these gaps, Alive & Thrive (A&T), an initiative aimed at improving suboptimal IYCF practices, applied the principles of social franchising to enhance IYCF counselling services in Viet Nam [18]. Incorporating elements of social franchising into government health facilities significantly improved the quality of nutrition counselling services, particularly the physical infrastructure (availability of counselling room and materials) and service delivery (increased knowledge by health staff of optimal breastfeeding practices, and improved technical content and interpersonal communication skills during counselling), resulting in high client satisfaction [19]. The services were incorporated into the functional and well-used government health system in Viet Nam [20], which provided a stable service-delivery environment. Utilization of the facility-based counseling services was lower than expected, however, one year after the launch of the first franchise [19, 21] as well as at endline (2 years later). This paper therefore examines the supply- and demand-side factors that influenced the utilization of IYCF counselling services in Viet Nam and highlights the key factors that, if strengthened, could markedly improve service utilization in this context.

Methods

Data sources and study population

Data used for this study were from household and health staff surveys which were conducted in the context of a cluster-randomized impact evaluation to compare the impact of two Alive & Thrive (A&T) intervention packages in Vietnam, i.e., an intensive package consisting of intensified interpersonal counseling and mass media compared to a non-intensive package consisting of standard nutrition counseling along with mass media [22]. The endline survey covered 4,052 households and 120 commune health staff in 40 communes in four provinces (Thai Nguyen, Thanh Hoa, Quang Ngai and Vinh Long) (Fig 1), which span the northern, central, and southern regions of Viet Nam and are geographically representative of the 15 provinces where A&T operates. All households with children under 5 years and their mothers in study communes were invited to participate in the survey. There were no exclusion criteria. Refusals to participate were not counted but were minimal. For the staff survey, 3 out of the 5–6 staff members in each commune were selected for interview. These included the commune doctor, the midwife, and the nurse in charge of nutrition activities.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Trial Profile.

https://doi.org/10.1371/journal.pone.0151358.g001

Data collection was carried out between June and August 2014, about 36 months after the launch of the first MTBT franchise. In addition to the household and health worker surveys (both using face-to-face interviews with structured questionnaires), structured observations of counseling sessions were also conducted.

This paper focuses on the factors influencing service utilization within MTBT franchises; therefore, analyses were restricted to 1,008 mother-child dyads (with a child up to 24 months of age) and 60 CHC staff in the MTBT areas only. Ethical approval was obtained from the Institutional Review Board of the Institute of Social and Medical Studies in Hanoi, Viet Nam and International Food Policy Research Institute, USA. Written informed consent was obtained from all mothers.

Conceptual framework for service utilization

The selection of factors potentially associated with service utilization was guided by the conceptual framework (Fig 2) adapted from the behavioral model of health services utilization developed by Andersen [23, 24]. This model has been used widely for empirical research on utilization of maternal health services [25–27], integrated community case management of childhood illness [28], and medical care [29]. The model proposed that use of health services is a function of characteristics of both individual user and the health system. In our study, we considered the user characteristics as demand-side factors with three components: predisposing characteristics, enabling factors, and need factors. Characteristics of the health care system are supply-side factors, which include the health service policy, service structure, and resources.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Conceptual framework of determinants of IYCF counseling services utilization.

¹Adapted from the behavioral model of health services utilization developed by Andersen [23, 24]^{. 2} Components related to health service policy and healthcare system management are assumed to be uniform across all MTBT franchises and thereby excluded from this framework. ³ Infrastructure/facility, equipment, and material resources are assumed to be uniform in all MTBT franchises.

https://doi.org/10.1371/journal.pone.0151358.g002

Results

Characteristics of the study sample

Mean age of mothers was 28 years, and the majority (83%) identified with the Kinh majority ethnicity (Table 1). Half of the mothers had at least nine years of education. The main occupations of mothers were farmers (41%), salaried employees (28%), and small traders (18%). Nearly half of the mothers received invitation cards to MTBT, 28% saw the A&T promotional television spots, and 13% saw an A&T billboard. Nearly three-fourths of the families lived within ten minutes of a MTBT center. While only 4% of children were born low birthweight, 7.7% currently were stunted and 2.5% were wasted, and 31% of children were sick during the last two weeks. Two-thirds of the families had one child under 5 years of age (62%).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Sample characteristics.

https://doi.org/10.1371/journal.pone.0151358.t001

MTBT service providers had a mean age of 41 years; 82% were female, and 62% had at least ten years of work experience (Table 1). Providers were highly confident on their job and experienced positive personal recognition and supportive supervision.

Patterns of MTBT service utilization

Overall, 45.1% of mothers never visited MTBT, 13.0% of mothers visited MTBT only once, 28.4% had repeated visits but did not achieve the minimum number of visits, and 13.5% achieved the recommended minimum number of visits. While repeat users increased significantly with child age, due at least in part to more opportunities for repeated visits, the proportion of children who achieved the minimum number of visits was lower for children 6–23.9 months of age compared to those <6 months (Fig 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Utilization of IYCF counseling services, by child age group.

https://doi.org/10.1371/journal.pone.0151358.g003

Demand-side factors associated with MTBT service utilization

Among the predisposing factors, maternal education (middle school and higher) was strongly associated with higher one-time visits to MTBT (PR: 3.5–3.9) and were more likely to achieve recommended number of visits (PR: 1.8–2.1), compared to mothers with primary school education or lower. Mothers of ethnic minority were close to twice more likely to visit the MTBT at the recommended minimum number of visits (PR: 1.73) compared to Kinh mothers (Table 2). Those who were farmers had 1.5 times higher probability of achieving recommended visits, compared to salaried workers. Maternal age was associated with a small but significant lower repeated use (PR: 0.98).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Associations between demand- and supply-side factors and categories of IYCF counseling services utilization (groups of MTBT users), compared to never use.

https://doi.org/10.1371/journal.pone.0151358.t002

Among the need factors, mothers with children 6–11.9 months and 12–23.9 months of age were, respectively, 1.5 and 1.6 times more likely to make repeated MTBT visits compared to those with younger children. This finding is probably a reflection of the longer duration of exposure to the program among older children compared to infants less than 6 months of age, which may increase their probability of attending more than once. These age differences were not observed for one-time users or achievers of the recommended minimum number of visits. Compared to mothers with non-wasted children, those who had a wasted child were 1.6 times more likely to have achieved minimum visits. For stunting, the opposite was true: mothers of a stunted child were less likely (PR: 0.5) to have achieved the recommended minimum visits compared to mothers of a non-stunted child. Mothers with a sick child (diarrhea or ARI) were less likely to have ever used the MTBT than mothers of healthy children, potentially due to seeking of treatment or services other than IYCF counselling during illness. Having more than one child under five years of age was also associated with 1.3 times higher of achieving the minimum visits.

Of all the demand (and supply side) factors studied, having received the MTBT invitation card was associated with the highest of MTBT service utilization. Mothers who received invitation cards had 3.0, 3.2 and 5.5 times higher probability of visiting once, repeating visits, and achieving the minimum visits, respectively. Other influential enabling factors included exposure to the promotional TV spots (PR: 1.2 of repeated visits; PR: 1.5 of achieving minimum visits) and to the promotional billboard (PR: 1.7 of one-time visit; PR: 1.3 of repeated visits; PR: 1.7 of achieving minimum visits). Household SES was not associated with use of MTBT, but distance traveled to reach the centers was a major barrier, with mothers traveling more than ten minutes being less likely to achieve the recommended minimum visits (PR: 0.58), and those traveling more than 15 minutes less likely to make repeated visits (PR: 0.47).

Population attributable risk estimation

Findings from our population attributable risk estimations indicate that if everyone had received an invitation card, an additional 6.1% of the population would have achieved the minimum number of visits compared to the current exposure level (44.8%, Table 1), holding other covariates constant (Fig 4). Similarly, an additional 3.2% and 7.5% of the population would have achieved the minimal number of visits if fully exposed to the promotional TV spot and billboard, respectively. Full exposure to the combination of all three demand generation variables would result in an additional 25.4% achieving the minimum number of visits. Good counseling skills was attributable to an additional 3.2% of the population achieving the minimum number of visits, and low work pressure accounts for 3.6%. Overall, an additional 49% of the population would have achieved the recommended minimum number of visits if the three demand creation approaches had reached all mothers and if good counseling skills and low work pressure had been the norm.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Population attributable risk of achieving the recommended minimum number of visits for select modifiable factors.

https://doi.org/10.1371/journal.pone.0151358.g004

Discussion

This study brings to light several factors that determine the utilization of IYCF counselling services at MTBT franchises in Viet Nam. Our findings on the demand- and supply-side factors generate insights on which factors, if strengthened, could generate the greatest increases in utilization of the IYCF counseling services offered through a social franchising model in the Viet Nam health system. Our study also contributes to the growing literature documenting the supply- and demand-side factors that influence service provision and use in a variety of contexts and for diverse types of services [38, 39].

Demand generation strategies (i.e., invitation cards to MTBT centers, billboards, and promotional television spots) were key enabling factors that stimulated utilization of IYCF counselling services for first time or repeated users, as well as for those who achieved the recommended number of visits. In particular, dissemination of invitation cards by village health workers was more effective than the other strategies in influencing not only the first visit but continued use of MTBT services. Thus, village health workers should be encouraged to continue this demand generation activity for improved and sustained service utilization. Also, exposure to all three MTBT demand generation activities had an additive effect on service utilization, with an additional 25.4% of the population potentially achieving the minimum number of visits if fully exposed to the three demand generation activities. This suggests that a combination of the demand generation activities could be used to increase MTBT service utilization.

While some studies suggest that lower SES may affect health services utilization due to resource constraints [40–43], this was not a barrier to accessing MTBT services as household SES was not associated with utilization of IYCF counselling services. This is likely because all preventive child health services in Viet Nam, including the MTBT services, are provided free of charge. Consistent with the results from other studies [43–45], we found that physical proximity to services played a role in their utilization. The likelihood of repeated visits to the MTBT centers significantly decreased when the proximity to the facility was greater than 15 minutes, and the likelihood of achieving the minimum number of visits further decreased when the distance to facility was greater than 10 minutes.

Predisposing factors, such as health beliefs, education and other user characteristics, have been found to determine the use of preventive and curative health services [25, 44, 46]. Consistent with findings from other studies on utilization of maternal health service [25, 26, 42, 43] and integrated community case management of childhood illness [28], we found that maternal education was strongly associated with higher MTBT service utilization, particularly one-time use and for achieving minimum number of visits. Mothers with a higher level of education may be more informed and have greater awareness of or capacity to understand health problems experienced by their children as well as the benefits of health care services, which may lead to more effective use of information and to optimal health-seeking behavior [14]. In our study, both ethnic minorities and farmers had higher odds of visiting MTBT. These groups are often marginalized or have less access to services, so it is possible that they were more willing to use MTBT services than those who are of majority ethnicity or engaged in other occupations, when invited or informed about the service. A study found that utilization of antenatal services gradually declined with increasing maternal age, possibly due to greater anxiety, concern, and receptive behavior among younger women [47]; this may have been the case also in our study, where higher maternal age was associated with slightly lower odds of repeated MTBT use. Older women may be less concerned about child feeding, or feel they already have the information they need, therefore less receptive to seeking IYCF counseling on a repeated basis.

Need factors may pose immediate or urgent reasons for seeking health services [48]. Thus, the lack of perceived need would pose a distinct challenge to high service utilization. For IYCF, the perception of need for services is less recognized than it is for illness. In our study, the perceived needs for counselling about child feeding practices were defined by low birthweight (<2500 grams), child stunting or wasting, and recent illness (diarrhea or ARI). Our results showed that mothers with wasted children were more likely to achieve the recommended minimum visits to MTBT, while mothers with stunted or sick children were less likely to achieve the minimum visits or to visit the MTBT even once, respectively. When children are sick, mothers likely seek health services other than nutrition counselling. Because perceived need often involves beliefs about susceptibility, consequences and intervention effectiveness, mothers may seek counselling services for wasting, which is more visible than stunting. It is also possible that wasted children show clear signs of picky eating behavior or anorexia, which could motivate mothers to seek help with child feeding. Although stunting is not always observable by the mother, it is unclear why mothers of stunted children were significantly less likely to continue using MTBT and achieve minimum visits than those whose children were not stunted.

As documented in the literature, we found that better counselling skills (providing appropriate technical information and demonstrating good interpersonal communication skills) was an important factor associated with any frequency of service utilization [8]. In previous studies other indicators of performance of health workers measured by knowledge, skills, incentives, motivation, and aptitude towards their duties were also found to have a strong influence on the use of services [49–51]. Our study, however, did not find an association between work performance indicators such as IYCF knowledge or personal recognition and service use. As expected, heath staff characteristics did not affect one-time use, but longer duration of work and greater work pressure was negatively associated with achieving minimum number of visits. It is possible that staff who have worked for ten or more years become lax in their work, less open to integrating new information and/or practices into their routine, or take on more responsibilities that lead to greater work-related stress, work pressure, or workload, which may compromise the quality of care and influence the quality of interaction with service users. Furthermore, high confidence and salary satisfaction were associated with more one-time use, but not with repeated use nor achieving minimum number of visits, which warrants a closer look at health staff performance within the current salary structure.

Our estimates suggest that an additional 49% of the population could achieve the recommended minimum number of MTBT visits if all three demand creation strategies (invitation cards, promotional spots and bill board) could reach all targeted mothers and if good counseling skills and lower work pressure could become the norm for all health staff. This would represent a 2.5-fold increase in the percentage of targeted women who would complete the minimum number of visits, which would empower them and help them adopt optimal IYCF practices.

Our study has some limitations. Data from cross-sectional household and health staff surveys were used to assess factors associated with counselling service utilization, thus we cannot determine causal relationships between them. For factors related to the need for services, our proxy measures may not reflect the actual needs as perceived by mothers. We did not directly assess the perceived need among mothers in our study, and our measures may not have covered mothers’ perceptions of other needs (e.g., perceptions of poor growth, limited appetite, insufficient breastmilk to fulfill child’s needs, or belief that child may be unhealthy or eating poorly). By applying a comprehensive framework of supply- and demand-side determinants of service utilization, however, we were able to examine various factors at the health system and client levels and identify critical factors influencing IYCF counselling service utilization and the recommended minimum number of visits for the service package.

Conclusions

Health systems are increasingly paying more attention to nutrition and incorporating preventive nutrition interventions into health services delivery. In this context, especially where health services are often facility-based, optimal use of nutrition interventions and services is critical to ensuring impact. It is, therefore, important to understand and address the factors influencing utilization in order to ensure that beneficiaries receive the full range of nutrition services and maximize their benefits. We have provided a comprehensive investigation of demand- and supply-side factors that influence utilization, focusing on the full range of utilization patterns–from one-time use to repeated use and ultimately achievement of the minimum package of services. Demand-generation strategies and counselling skills were the most important factors associated with any service use; strengthening these two demand -and supply-side factors could markedly improve service utilization with an additional 49% of the population achieving the recommended minimum number of MTBT visits. Given the continued efforts to improve the impact of IYCF interventions and to scale up nutrition through health systems globally, our study contributes robust empirical evidence on potential factors to optimizing facility-based counselling service utilization in Viet Nam. Further research is needed to identify and test methods for increasing and sustaining reach and use of similar services in other contexts.

Supporting Information

Acknowledgments

Data collection was contributed by the Institute for Social and Medical Sciences, Vietnam. Jean Baker (FHI 360) and Silvia Alyon (Save the Children) critically reviewed the manuscript.