mHealth - A Novel Approach to Improve Child Nutrition in India

Aims and Objectives

The study aims to develop, design and pilot a mHealth intervention to strengthen CF components of the ICDS programme. Specific objectives seek to:

1.1. Exploring caregivers, AWWs and other relevant ICDS personnel's views on the feasibility, acceptability and best means of using mobile phones to facilitate the process of delivering CF counselling within the ICDS programme.

1.2. Designing a context-specific, relevant and sustainable mHealth intervention to strengthen CF components of the ICDS programme based on existing evidence and views of caregivers, AWWs and other relevant ICDS personnel.

1.3. Pilot testing and refining the mHealth intervention.

Study design

It is anticipated that a mHealth intervention might potentially strengthen CF components of the ICDS programme. Accordingly, qualitative and quantitative strategies will be used to develop, design and pilot a mHealth intervention that endeavours to address the following three barriers within the ICDS programme: (i) inadequate feeding practices, (ii) not targeting children under 3 years of age, and (iii) poor AWW supervision and unmanageable workload.

Focus group discussions (FGDs) and in-depth interviews (IDIs) will be conducted in the study site with AWWs, other relevant ICDS personnel and mothers/caregivers of children aged 9-11 months to assess the feasibility, acceptability and best means of using mobile phones to facilitate the process of delivering CF counselling and improving AWW supervision and workload. Any changes in infant diet will be also be assessed using 24-hour dietary recalls. This infant age group has been selected since 9-11 month old infants will likely demonstrate greater nutritional benefit when compared to older children, and they have more established CF intakes than a younger age group (e.g. 6-8 months) that is more likely to be transitioning to eating complementary foods.

Socio-demographic questionnaires will also be administered to AWWs, ICDS personnel and mothers/caregivers for descriptive data. Findings will be presented in an intervention design workshop to develop the mHealth intervention, which will then be pilot tested and refined.

Mixed-methods research is recommended for exploring complex interventions. Accordingly, quantitative and qualitative methods will be used to address the study objectives. The study will use a wide range of respondents to gain a broader understanding of current behaviours, identify key influencers and understand the facilitators and barriers to behaviour change from different perspectives. Sample size will be determined by saturation sampling, that is, continuing investigating until no new information is uncovered.

Sampling strategy

AWWs, relevant ICDS personnel and mothers/caregivers of children aged 9-11 months will be purposively selected from the study area. Criterion for selection is that the participants are involved in nutrition counselling within the ICDS programme. Sampling will be purposive as the research team is aiming for inclusiveness, not representativeness. The number of FGDs and IDIs will depend on the number required to reach data saturation i.e., repeated and new material or themes do not emerge. The characteristics of AWWs and other relevant ICDS personnel in the FGDs and IDIs will reflect the diversity and breadth of potential AWWs and other relevant ICDS personnel involved in the promotion of nutrition messages. The characteristics of mothers/caregivers in the FGDs and IDIs will also reflect the diversity and breadth of potential mothers/caregivers receiving nutrition messages.

Methods

Objective 1.1: Assessing the feasibility, acceptability and best means of using mobile phones for CF counselling

FGD 1.1.1 and FGD 1.1.2 will take place with a trained facilitator and a person for transcription with fluency in local language(s). AWWs, relevant ICDS personnel and mothers/caregivers will answer questions on the feasibility, acceptability and best ways of using mobile phones to facilitate the process of delivering CF counselling within the ICDS programme. Open-ended questions will be asked, and then followed up by specific prompts which relate to the research question. FGDs will be recorded with digital recorders.

QQ 1.1.1, QQ 1.2.2 and QQ 1.2.3 will be translated into local languages as needed. Mothers/caregivers will answer an interviewer-administered questionnaire on socio-demographic information. Questions about mobile phones will also be asked by trained interviewers with fluency in local language(s) to inform Study 2. AWWs and other relevant ICDS personnel will answer self-administered questionnaires on mobile phone use, professional experience and socio-demographic information.

Objective 1.2: Designing a mHealth intervention to strengthen CF components of the ICDS programme

Data about potential usage of mobile phones within the ICDS system will be presented at an intervention development workshop where stakeholders and relevant experts will review the findings of the formative research and make decisions about the mHealth intervention's content and structure. Outputs will include:

• Behaviour change strategies
• mHealth intervention strategy and components, including who should be targeted and how
• Conceptual framework of the intervention

Intervention materials including manuals, protocols and creative briefs will be written as needed after the intervention design workshop. Relevant intervention components will also be tried out prior to the pilot intervention.

It should be noted that the outcome of this workshop is entirely evidence-driven; that is, no further details can be provided on the potential design of the mHealth intervention until data from FGDs with AWWs, other relevant ICDS personnel and mothers/caregivers have been collected and analysed.

Objective 1.3: Pilot testing and refining the mHealth intervention

The mHealth intervention will be pilot tested and refined with relevant ICDS personnel and mothers/caregivers in order to test all operational elements that were selected and developed in the intervention design workshop. This pilot study will purposively select participants - mothers/caregivers, AWWs and/or other relevant ICDS personnel depending on the intervention design - from the study site to pilot test the intervention.

AWWs will be trained on counselling messages prior to implementing the pilot test with mobile phones. The pilot test take place over 6 weeks with approximately 72 purposively selected participants. Any changes in infant diet will be measured by administering to mothers/caregivers in the pilot study before and after the pilot study. Process evaluation tools will also be used to measure context, fidelity to protocol, dose, and facilitators and barriers to implementation.

At the end of the pilot testing period, IDIs will be used with study participants to collect data on the following indicators:

• Feasibility (including frequency and timing of messages and/or phone calls)
• Acceptability
• Comprehensibility
• Tone of messages
• Barriers
• Facilitating factors
• Modifications

The number of IDIs will depend on the number required to reach data saturation i.e., repeated and new material or themes do not emerge. Participant characteristics in the IDIs will reflect the diversity and breadth of study participants recruited to test mobile phones in the pilot study.

QQs will also be used with study participants to collect socio-demographic information. Interviewer-administered QQs will be used with mothers/caregivers if they are part of the study due to low literacy rates in both the proposed study sites.

These data will be collected with the aim of describing the successes and challenges of implementing a mHealth intervention in this setting, and how it may potentially be used within the ICDS programme and other settings to deliver complementary feeding counselling. They will be used to further refine the mHealth intervention.

Detailed protocols including sampling strategies, QQs and forms to collect data will be finalised once a context-specific mHealth intervention has been developed in the intervention design workshop.

Data Analysis

Quantitative data analysis

Quantitative data collection methods in this study include QQs and 24-hour recalls. All quantitative data will be double-entered and data entry errors will be identified using range and consistency checks. After the data are cleaned, they will be transferred to STATA 11 software for analysis. Dietary data will be analysed in a specific diet program that has been developed using Microsoft Excel. Categories will be generated for continuous variables as needed. Data will be analysed and represented using numerical, graphical and tabular methods. In order to estimate any changes in infant diet when FBRs are tested with mothers/caregivers (Objective 1.3), paired t-tests will be used to identify differences in means with 95% CIs for all dietary outcomes. Differences will be considered significant at P< 0.05. Statistics will be used for descriptive purposes, i.e. the study population will be represented as proportions, means, etc. and dietary data from the pilot test will be represented in means and standard deviations.

Qualitative data analysis

The study will rely heavily on FGDs, IDIs and structured observations at various stages for insights into how to develop, implement, and evaluate the mHealth intervention. Qualitative analysis will be iterative, thus evolve alongside data collection. NVivo 8 software will be used to store and code all the FGD and IDI transcripts after they have been translated into English. Contents of FGD and IDI transcripts will identify the range of ICDS personnel and mothers/caregivers' reactions to FBRs and using mobile phones for nutrition counselling. A thematic approach will be used to code FDGs and IDIs, starting with initial free coding with an ultimate aim of categorizing themes into categories. In part, these codes will be based on predetermined themes emerging from the literature review about nutrition counselling interventions and the current state of the ICDS programme in India. This process will not use a grounded theory approach, in which coding is entirely inductive.11 However, all qualitative data will be coded allowing for emergence of new unanticipated themes which will influence the pre-conceptualised framework.

Triangulation of data is important to establish reliability of data. 12 Accordingly, data from all the FGDs and IDIs will be compared and checked against each other for inconsistencies and completeness. Doing so will help the student ascertain the most "correct" interpretation of data. All conclusions will be presented to key study staff for accuracy to further improve reliability.