Barriers and facilitators to growth monitoring and promotion in Nepal: Household, health worker and female community health volunteer perceptions

Abstract Growth monitoring and promotion (GMP) is both a service for diagnosing inadequate child growth in its earliest stages and a delivery platform for nutrition counselling. The widespread use of GMP services in developing countries has the potential to substantially reduce persistent child undernutrition through early diagnosis and by linking caregivers and their children to key health and nutrition services. However, researchers have questioned the effectiveness of GMP services, which are frequently undermined by underdeveloped health systems and inconsistent implementation. This analysis examined both supply‐ and demand‐side factors for GMP utility in Nepal from the perspectives of beneficiaries and service providers, particularly focusing on three components of GMP: growth assessment, analysis of growth status and counselling. The most common factors influencing GMP uptake included beneficiaries' perceptions of the relative importance of GMP and the knowledge and skill of frontline workers. Both providers and beneficiaries viewed GMP as a secondary health and nutrition activity and therefore less important than curative services. We found deficits in GMP‐related knowledge and skills among providers (i.e. health workers and female community health volunteers), as well as indications of poor training quality and coverage. Furthermore, we found variation in GMP utilization by maternal age, education and residency (alone, nuclear or extended), as well as household socio‐economic well‐being and rurality. This study is the first to assess factors influencing both beneficiaries and service providers for GMP utilization. Further research is needed to explore the implementation of improved GMP protocols and to evaluate facility‐level implementation barriers.


| INTRODUCTION
Nearly half of all deaths worldwide among children under 5 years of age are linked to undernutrition (Black et al., 2013). Nutritional wellbeing during this critical period has both immediate and lasting consequences on a child's physical and cognitive health, development and functioning (English, Peer, Honikman, Tugendhaft, & Hofman, 2017;Hossain et al., 2017;Martorell & Woodruff, 2017;Pietrobelli et al., 2017). To combat child undernutrition, many interventions target households in the 1,000-day period from gestation to the child's second birthday (Martorell & Woodruff, 2017;Schwarzenberg & Georgieff, 2018). Although interventions that support optimal nutrition beyond the first 2 years of life are important, damage sustained during the first 1,000 days is often irreversible (World Health Organization [WHO], 2013).
Interventions that target early diagnosis and corrective action for child undernutrition have been prioritized globally, particularly in lowand middle-income countries (LMICs) (Ashworth, Shrimpton, & Jamil, 2008). Growth monitoring and promotion (GMP), for example, is used in many LMICs to diagnose inadequate child growth in its earliest stages and in turn alter the child's growth trajectory through nutrition counselling and other health-promoting actions. According to WHO guidelines, GMP includes (1) the routine measurement of a child's weight and length/height; (2) the plotting of the child's measurements and comparison of the child's status to a standardized growth chart to assess growth adequacy; (3) growth-informed counselling; and, if necessary, (4) the undertaking of remedial, healthpromoting action (WHO, 2006(WHO, , 2008. When implemented correctly, GMP programmes have created linkages to key preventative and curative health services, increased mothers' knowledge of proper infant and young child feeding practices and provided the opportunity for early diagnosis and treatment of undernutrition (Adhikari, Khatri, Paudel, & Poudyal, 2017;Ashworth et al., 2008;Gyampoh, Otoo, & Aryeetey, 2014). Before the 1990s, GMP services were the subject of much enthusiasm, research and evaluation, but the attribution of GMP service utilization to positive changes in a child's growth status has long been the subject of debate (Ashworth et al., 2008;Garner, Panpanich, Logan, & Davies, 2000). Although GMP services are still in use in most LMICs, its efficacy is limited by numerous challenges including low service coverage, inadequate training of health workers and resulting measurement errors, incorrect interpretation of growth charts, and poor or nonexistent counselling (Ashworth et al., 2008;Bégin et al., 2019;de Onis et al., 2012;Feleke, Adole, & Bezabih, 2017;Laar, Marquis, Lartey, & Gray-Donald, 2018;WHO, 2006;WHO & UNICEF, 2009).
In Nepal, where the prevalence of underweight (27%), stunting (36%) and wasting (10%) remains high despite incredible progress over the last 20 years, GMP is a prioritized nutrition intervention (Cunningham, Headey, Singh, Karmacharya, & Rana, 2017;Ministry of Health, 2017). Nepal's Multi-Sector Nutrition Plan-II (MSNP-II) (2018)(2019)(2020)(2021)(2022) aims to address the complex causes of malnutrition by scaling up both nutrition-specific and nutrition-sensitive services and improving utilization of these services (Government of Nepal National Planning Commission [NPC], 2017). Nepal's MSNP-II frames GMP as a key platform for improving infant and young child nutrition and care and recommends that children under 2 years of age receive monthly GMP (NPC, 2017). Health workers are responsible for conducting GMP services-which focus on weight for age, not height-at local health facilities or at monthly primary health care outreach clinics (Child Health Division, 2016;de Onis et al., 2012). A cadre of more than 52,000 female community health volunteers (FCHVs), who serve as the first point of contact in communities across Nepal and refer people into the health system, support the implementation of GMP services (Ministry of Health and Population, 2019).
In this paper, we use quantitative and qualitative data to assess the first three of four components of GMP services-routine anthropometric measurements to assess growth, plotting and comparing child growth, and growth-informed promotive counselling-as noted by the WHO definition above. Specifically, for each of these three stages, we examine the current state of GMP service provision and utilization in Nepal and identify barriers and facilitators to optimal GMP service from both service provider and beneficiary perspectives.
We used Andersen's Behavioral Model of Health Service Use as a framework for our exploration of both contextual and individual factors that predispose and enable GMP service success in Nepal (Andersen, 1968(Andersen, , 1995(Andersen, , 2008. We hypothesize that routine GMP service use is minimal and that there are both individual and structural health system barriers contributing to gaps in coverage and delivery of high-quality GMP services. These analyses will help to fill not only gaps in research related to health services in Nepal but also global research gaps related to GMP failures and opportunities for improvement from both beneficiary and provider perspectives.

| Quantitative data collection and management
The quantitative data used are from a cross-sectional monitoring survey of Suaahara II, a United States Agency for International

Key messages
• Growth monitoring and promotion (GMP) is a preventative and promotive nutrition activity.
• Globally, there is limited evidence on the factors that affect the implementation and utilization of GMP from the sides of both providers and beneficiaries.
• In Nepal, GMP has the potential to alter the landscape of child undernutrition if systematic changes to countrylevel protocols are researched, designed and implemented. Development (USAID)-funded, multisectoral nutrition programme that aims to improve the health and nutrition status of mothers and children in 42 of Nepal's 77 districts. New ERA, a local survey firm, collected data from June 10 to September 10, 2017. Multistage cluster sampling and probability proportion to size (PPS) techniques were used to select the following: Suaahara II districts (n = 16), one rural and one urban municipality per district (n = 32), three wards per municipality (n = 96) and two clusters per ward (n = 192). For the final stage, 19 households with a child under 5 years were randomly selected from each cluster from a full list gathered by the survey firm (n = 3,648) (Suaahara II, 2018). The household survey collected key information on a variety of indicators, including household socioeconomic and demographic characteristics, nutrition-and healthrelated knowledge and practices, and utilization of Government of Nepal health and nutrition services. Additionally, one FCHV from each cluster (n = 192) and one health facility key informant from each ward (n = 96) were included in the survey. The FCHV and health facility key informant (preference for those in the highest level role of health facility in-charge, when available) questionnaires gathered data on socio-economic and demographic characteristics; exposure to training on key health and nutrition areas; perceptions of their work experience; exposure to Suaahara II platforms; and their detailed knowledge and skills related to counselling and following government protocols, including GMP (Suaahara II, 2018).
All data were collected electronically using Open Data Kit software on Android phones. Once collected and reviewed by a supervisor, the data were synced to a secure server. New ERA staff checked the quality and consistency of data and completed the first round of data cleaning and verification, as well as the translation of openended responses into English when necessary. Suaahara II staff further cleaned the data, including variable generation. The de-identified and cleaned data files were then used for this analysis. Of the 3,648 households surveyed, about half (n = 1,850) had a child under 2 years of age. These households were the focus of this study. Of the 96 health facilities surveyed, this analysis focuses only on the health posts (n = 91); two hospitals and three primary health care centres, which have different mandates and scopes of practice than health posts, were excluded to avoid the introduction of extreme heterogeneity.

| Qualitative data collection and management
In July 2018, a qualitative study was done to complement the prior quantitative survey's descriptive findings and enable deeper exploration of barriers and facilitators for participation in GMP services. Data were collected from mothers, frontline workers (health facility workers and FCHVs), and Suaahara II national and district staff by Square One, a local survey firm. A purposive sampling strategy was used to select districts representative of Nepal's three agroecological zones (i.e. Terai, hills and mountains)-Rupandehi, Bhojpur and Bajhang-as well as individuals with relevant knowledge and experience at national, district and community levels. There were 37 data collection points split equally across the three districts; data were collected through focus group discussions (FGDs) with Suaahara II staff, health workers and FCHVs, and 1,000-day mothers (one FGD with Suaahara II staff, one per district with health workers and one per district with mothers; n = 7) and in-depth interviews (IDIs) with 1,000-day mothers (10 IDIs per district; n = 30). During the FGDs, which ranged from five to 13 participants, one researcher facilitated the discussion while another researcher took notes. While most FGDs were conducted in Nepali, interviews were conducted in Awadhi and Bhojpuri in Rupandehi and similarly, Doteli language was used for interviews in Bajhang.
All qualitative data were digitally recorded, transcribed verbatim and translated to English from Nepali, Awadhi, Bhojpuri and Doteli by the local survey team. De-identified IDI and FGD transcripts were uploaded into Atlas.ti 8.2 for data management and analysis.

| Analyses
We used the quantitative data to summarize survey respondents' background information and GMP service uptake in the survey population. Potentially predisposing factors, such as mother's age, education, caste/ethnic group, religion, occupation, agroecological zone, residence, child age and child sex were explored at the bivariate level (Table 1). Age and level of education were constructed as continuous variables with education level referring to the total number of years of formal schooling received. Caste/ethnicity was categorized into three groups: socially advantaged (Brahmins/Chhetri), socially excluded (Dalit, Muslim and disadvantaged Janajati) and other groups (Gurung/Thakali, Newar, other non-Dalit Terai castes and others) (Aasland & Haug, 2011;Pandey, Dhakal, Karki, Poudel, & Pradhan, 2013). Household socio-economic status was measured using the Equity Tool, which generates quintiles based on ownership of key assets and quality of household structures (Metrics for Management, 2015). On the provider side, we summarized frontline workers' background information, formal training, and knowledge and skills related to GMP to describe the services available and contextualize the health care environment in which care is sought and provided ( Table 2). All analyses were conducted using Stata/IC 15.1 software.
The qualitative data were analysed using a content analysis approach to identify barriers to GMP service utilization and provision as well as potential solutions at the beneficiary and service provider levels. The transcripts were read repeatedly as part of the data familiarization process and then hand coded by the lead author who created a preliminary codebook containing key concepts and categories after reading a cross section of the interviews. Thematic codes from existing literature were identified and integrated into the codebook to ensure that both theory-based and emergent concepts were included.
The lead author then applied these codes to all interviews to consolidate and create more nuanced versions of the codes. These codes were used to compare responses across data points and were then gathered into several conceptual categories. Finally, selective coding was completed to generate results. The co-authors held regular meetings with the lead author to ensure that any questions or potential discrepancies were addressed. The lead author also consulted with Square One Research and Training, the research company who conducted the interviews, to ensure that she had correctly interpreted the transcripts.

| Ethical considerations
Ethical approval from the Nepal Health Research Council was received for both quantitative and qualitative studies. Participation in the study was voluntary, and written informed consent was obtained from each respondent prior to beginning each questionnaire and interview.

All data collection was approved by the Nepal Health Research
Council. Written informed consent was obtained from each quantitative survey respondent prior to beginning any interview, and verbal consent to continue the survey was obtained after the completion of each section in the questionnaire. Similarly, for the interviews and FGDs, written informed consent was obtained prior to any data collection (Suaahara II, 2019).

| Sample characteristics
The majority (60.0%) of surveyed households with children under 2 years belonged to a socially excluded caste/ethnic group (Table 1).

| Routine anthropometric measurement to assess growth
Nearly 90% of mothers with children under 2 reported having ever used GMP services (Table 1). There was no significant difference in GMP use by child sex (90.2% vs. 88.2% for male and female children) or age (11.3 months for those that ever received vs. 11.7 months for those who did not). Although nearly all reported that they had attended GMP at some point, only slightly more than one third of those who had ever attended had been to GMP in the 3 months prior to the survey. On the provider side, health workers reported that GMP provided an average of 19 days per month at their post (range: 1-30) ( could be trained to conduct GMP and come door-to-door then it will be easier for us to participate in GMP'. In the FGD in Bajhang, another mother noted that GMP is now done by FCHVs in her monthly Health Mothers' Group (HMG) meetings: In our HMG meeting, they [FCHVs] always take the measurements. There is a weighing machine as well ….
It wasn't there before, but from this year onwards, they have been measuring the weight of our children. The FCHV even asks us to bring our children to the meeting.

| Plotting and comparing child growth
Both health facility workers' and FCHVs' knowledge and skills related to GMP were assessed by presenting them with a hypothetical child's growth chart in which the child's measurements were in the 'red' area of the chart with a positive growth trend, indicating that the child had severe acute malnutrition but was improving. When presented with this sample chart, 65.9% of health facility workers and 78.6% of FCHVs incorrectly classified the child's status or did not know how to classify the status (Table 2) In all three districts, there were reports of frontline workers failing to record the child's measurements, which prevents monitoring of the child's growth trend. In some cases, mothers took it upon themselves to memorize their child's measurements, knowing that the measurements were not being recorded by frontline workers. Even when measurements were recorded and plotted on a growth chart, issues arose regarding the interpretation of a child's growth.

| Growth-informed counselling
Our survey results showed that among caregivers who had ever taken their child to GMP, 70.9% were not told about their child's growth (i.e. the change in the child's weight over time) during their last GMP (Table 1). Furthermore, among mothers whose child had ever received GMP, only 21.0% reported discussing child nutrition and 12.7% reported discussing child health with an implementing health frontline worker during their last GMP. Although most frontline workers ( When the interviewer asked what she liked most about the GMP services, one mother in her early 40s (Rupandehi; IDI) responded by detailing her conversations with an FCHV about child feeding practices: Everything was good. I like the way they speak to the mothers the most. The Female Community Health Volunteers were speaking politely with all mothers. I also like the way they were suggesting nutritional food that we must provide to our baby. They have also suggested that we eat food more regularly and give food regularly to our baby, as well.
Another mother in her mid-20s shared similar positive experi-

| DISCUSSION AND CONCLUSION
Comprehensive GMP services consist of several steps, including the weighing and measuring of a child, recording of measurements and interpretation of the growth chart, and provision of informed, nutrition-promotive counselling. Using both quantitative and qualitative data, we assessed these three components of GMP services, which each require skilled health service providers. For each stage, we identified both provider and beneficiary perspectives on the major barriers and facilitators to optimal GMP services in Nepal. We found that although awareness of GMP was high, routine utilization was low and both demand and supply-side constraints existed. Our findings were framed by Andersen's Behavioral Model of Health  (Garha, 2016;Overseas Development Institute, 2016;Roesler et al., 2018). Because GMP requires frequent use, proper monitoring and recording, and timely and appropriate nutrition counselling to produce clear results, the service must be readily available and accessible to ensure that frequent GMP use is feasible. As such, community-and home-based interventions may offer a stark advantage over facilitybased options, particularly for beneficiaries identified as prone to inconsistent attendance (Agbozo et al., 2018;George et al., 1993;Arole, 1998;Mayhew, Ickx, Stanekzai, Mashal, & Newbrander, 2014 This study provides insight into GMP implementation and utilization using quantitative and qualitative approaches and fills a gap in the literature by including both service provider and beneficiary perspectives. Although data on use ever were collected, data on recurrent GMP use would strengthen our understanding of the state of GMP service utilization in Nepal. Furthermore, although this study did not have data on the fourth component of GMP, this in-depth exploration of the first three components of GMP is unprecedented and critical to our understanding of service implementation and utilization. The qualitative IDIs and FGDs, although not generalizable to all of Nepal, enable an understanding of the variation of factors across the country given our sampling across agroecological zones and from east to west. As reducing persistent malnutrition is a high priority, further research should investigate the use of FCHVs and other community-based providers for strengthening GMP services in Nepal. Experimental implementation science studies at the district and subdistrict levels could also be done to test different implementation modalities for strengthening service provision for GMP and other health and nutrition services in various contexts.
Our findings show that future discussions regarding the merit of GMP in Nepal should focus on ways to increase both routine utilization and the consistent provision of quality GMP services for diverse communities across agroecological-and resource-variable settings.
Although GMP is largely ubiquitous in nature, it has become increasingly apparent that country-level applications of the intervention differ substantially in purpose, design, implementation and effectiveness (Bégin et al., 2019). As Nepal continues to strive for significant reductions in child undernutrition, there is a compelling need for effective and equitable GMP implementation. This paper indicates that frontline workers responsible for GMP must be trained and reoriented en masse about its importance and correct implementation, particularly regarding promotive nutrition counselling. In addition to training, ongoing supportive supervision is vital to effective delivery of services at the community level. Further research is needed to evaluate the salience of current GMP protocols in the context of Nepal's evolving health system and to research the causal relationship between individual and contextual factors and service utilization.

ACKNOWLEDGMENTS
The authors would like to thank all the respondents of the annual survey for providing valuable data and the New ERA and Square One Research and Training teams for collecting the data used in this analysis. The authors would also like to acknowledge the Suaahara II Monitoring, Evaluation, and Research team for their support. We would also like to thank Rolf Klemm, John Macom and Deependra Prasad Bhatt for their edits. The authors acknowledge the United States Agency for International Development (USAID) for providing support to conduct this study. This publication was prepared using data from

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.