mVAM for Nutrition Part I revolutionizing collection of nutrition - - PowerPoint PPT Presentation

mVAM for Nutrition Part I

revolutionizing collection of nutrition information

Learning Lab Data dive – Mobile for Nutrition

Mobile Vulnerability Analysis and Mapping (mVAM)

PR PROJ OJEC ECT T OVER ERVIE VIEW

Humanitarian decision making process

Data is stored in a database and analyzed by a ‘stats engine’

Mobile surveys

Databank Reports

Data is anonymized and cleaned

Results and data are shared as a global public good

Receive info on WFP

[2-way communication system]

Respondents are contacted

• n their mobile phones

Respondents contact WFP through their mobile phones

Record Feedback

[2-way communication system]

Photo: WFP / Lucia Casarin

IVR [Interactive Voice

Response calls]

SMS surveys

[Text messages]

Live calls

[Telephone operators]

mVAM for Nutrition

What is it?

Partnership

The Nutrition Division and mobile Vulnerability Analysis and Mapping unit at WFP

Innovation

Exploring innovative ways of collecting nutrition data using remote data collection methodologies

Today’s outline

 Why mVAM for Nutrition?  Kenya feasibility and validation study (with partners ICRAF)  Malawi pilot: data collection via SMS (MDD-W survey)  Your turn! SMS demo of collecting participants dietary data  Visualization of dietary information from Malawi study  Closing remarks and wrap-up  Questions & Answers

• Mobile collected information could help to
• provide early warning of deteriorating

nutrition situations

• support global efforts to strengthen

nutrition monitoring

• Nutrition data gap (GNR 2016)
• mVAM for Nutrition can support fill data gap

Why mVAM for Nutrition?

• Mobile phone access and ownership

increasing exponentially around the world Mobile data collection methodologies offer a quick and affordable way to collect data remotely

Convenience

Data collection in hard-to-reach and insecure areas

Cheaper

Cheaper large-scale data collection F2F $16 CATI $5

Faster

Real-time data Reduced time between data collection and information delivery

• mVAM for market and food security data
• can it work for nutrition?

Indicators: MDD-W & MAD

Internationally validated and corporate indicators of WFP

What does it measure?

Proxy to measure the micronutrient adequacy of

WRA at the population level

Proxy to measure the nutrient density of young children’s diet at the population level

Definition

The proportion of WRA who consume at least 5 out of 10 (core) food groups that make up the score. MAD: Minimum Dietary Diversity (MDD) + Minimum Meal Frequency (MMF) MDD: Consume at least 4 out of 7 (core) food groups MMF: Depends on how many months and if child is breastfed

Calculation

10 food groups 7 food groups (MDD) + 4 frequency questions (MMF)

Food groups

1. Grains, white roots and tubers, and plantains

• 2. Pulses (beans, peas and lentils)
• 3. Nuts and Seeds
• 4. Dairy
• 5. Meat, Poultry and Fish
• 6. Eggs
• 7. Dark green leafy vegetables
• 8. Other Vitamin A-rich fruits and vegetables
• 9. Other Vegetables
• 10. Other fruits

1. Grains, roots, tubers 2. Legumes, nuts 3. Dairy products (milk, yoghurt, cheese) 4. Flesh foods (meat, fish, poultry, organ meat) 5. Eggs 6. Vitamin A-rich fruits and vegetables 7. Other fruits and vegetables

Indicator Reporting

Required Stunting prevention programme Nutrition-sensitive programmes Highly recommended Micronutrient programmes MAM prevention programmes

Method

Open-based 24-hour recall

Kenya Case Study

Partner and study locations Kitui and Baringo county

Feasibility and validity of collecting data on MAD and MDD-W using Computer-Assisted Telephone Interviewing (CATI) A collaborative effort between WFP and ICRAF

Phase I: Formative Feasibility Study

Determine feasibility of using CATI methodology for collecting women and young children dietary data

Why? Document understanding of gender distribution of phones as how women use their phones remains largely unknown Objective

• Identify constraints and success factors in receiving

mobile surveys

• Understand cultural contexts and local diet patterns

Data collection:

• Women’s mobile phone usage patters
• Local diet of women and young children

Method:

• 17 Focus group discussions
• 16 in-depth surveys
• 22 key informant interviews

Location: 16 sub-locations in Kitui and Baringo

Phase I: Formative Feasibility Study

Determine feasibility of using CATI methodology for collecting women and young children dietary data

Results

Phone access and usage Access: high Ownership: high (60-90%) Sharing: inter- & intra household Mode: primarily calling Potential barriers Willingness: strong (phone and diet surveys) Trust: unknown numbers Gender constraints: husbands approval to participate Phone network coverage: some locations poor

Recommendations

Formative study to inform the design of CATI survey with women. Community sensitization to identify and address potential trust issues. Prior engagement with men/husbands - especially in areas where gender can be a barrier Community consultations to understand

• ptimal times and days to reach

respondents. Scheduling times for phone calls in advance

• in areas with limited phone network.

Account for the need to make multiple phone calls at different times of the day and

• n different days of the week.

Phase II: Mode Experiment

“Most rigorous test of this technology ever done”

Cost per survey F2F = $16 (16 enumerators x 2) CATI = $5 (8 operators) Mode experiment measured the accuracy of data collected on MAD and MDD-W using CATI versus traditional Face-to-Face interview

Experimental Design: Each indicator survey consists of three main groups. Treatment group 1 and Treatment group 2 randomize data collection mode (CATI and F2F) across the two sampling rounds. Control group 1 is a control for treatment effects, while Control group 2 in MDD-W is used to assess subpopulation bias.

Phase II: Mode Experiment

Test/Re-test design

CATI N (%) F2F N (%) Differences (CATI – F2F) Agreement (%) P-value MDD-W 208 (26.4%) 196 (24.9%) 2% 74.4% 0.44 MDD 225 (38.9%) 122 (21.1%) 18% 67% < 0.0001 MMF 409 (70.8%) 338 (58.5%) 12% 65.5% < 0.0001 MAD 171 (29.6%) 71 (12.3%) 17% 72% < 0.0001

Phase II: Mode Experiment

Proportion above and below the threshold score by mode

MDD-W via CATI compares well with F2F. Bigger differences are noted with MAD

Phase II: Mode Experiment

Change in MDD-W and MAD with CATI

For trend analysis, CATI can be used as a cost-effective method for collecting both MAD and MDD-W Point estimates for MAD need more research

mVAM for Nutrition Part 2

Malawi study and SMS data collection demo

Learning Lab Data dive – Mobile for Nutrition

Trial and Error

How we collected women's dietary information via SMS in Malawi

Method and Design

• Large-scale (national) feasibility testing
• Study site: Malawi
• Mode: SMS
• 5 rounds (Oct 2016 – April 2017)
• Indicator: Minimum Dietary Diversity

for Women (MDD-W)

• MDD-W: proxy for micronutrient intake of

women 15-49 years of age

• Near-real time data enabled optimization of

methodology throughout rounds

Lessons learned

• Using a mix of open-ended and list-based

questions to help respondents better understand;

• Keeping questions simple; in some cases splitting

questions to make it easier for the respondent to answer;

• Allowing respondents to take the survey in their

preferred language;

• Pre-stratifying and pre-targeting to ensure

representativeness;

• Post-calibrating to produce estimates that are

more comparable to face-to-face surveys.

• More information on:

http://mvam.org/2017/06/06/trial-and-error- how-we-found-a-way-to-monitor-nutrition- through-sms-in-malawi/

Trial and Error

How we collected women's dietary information via SMS in Malawi

MDD-W survey via SMS used in Malawi

Now it’s your turn to record nutrition information through SMS!

Demo: collecting MDD-W via SMS

Data visualization in Tableau

• WFPs corporate data visualization platform

Here

Malawi data visualizations

mVAM for Nutrition

Future opportunities

Future opportunities

• Further testing on MAD for point estimates
• Early warning

nutrition indicators for surveillance systems

• Research on other modes (SMS)
• Strengthen capacities & technical support

F2F training online learning to scale-up methodology

• Advanced statistical methods to make

adjustments for mode effects and sub-population bias.

• Management and visualisation of data

Visit the WFP mVAM blog for more information on the initiative ‘mVAM for Nutrition’

• http://mvam.org/2016/08/11/monitor-nutrition/
• http://mvam.org/2017/01/09/can-we-reach-rural-women-via-mobile-phone-kenya-case-study/
• http://mvam.org/2017/05/10/mvam-for-nutrition-findings-from-kenya/
• http://mvam.org/2017/06/06/trial-and-error-how-we-found-a-way-to-monitor-nutrition-through-sms-in-

malawi/

More information on mVAM for Nutrition

World Food Programme World Food Programme

Thank You!