Data collection

The online survey was created using Qualtrics software (Qualtrics, Provo, UT, USA), Version June 2018. The goal of our sampling was to get as many participants from the broader nutrition community as possible at all levels, and so we disseminated the survey through as many professional networks as possible. These included online nutrition-focused listservs with self-elected members (ie, Ag2Nut, Ag2Nut Ethiopia, CORE Group Nutrition, CORE Group General), institutional listservs (ie, WHO Nutritionlist and SUN), professional networks (ie, SUN, UNICEF, Bill & Melinda Gates Foundation, Johns Hopkins University), and Twitter. In this convenience sample, potential respondents were first contacted through an introductory email from the Data for Decisions to Expand Nutrition Transformation (DataDENT) study team, then were directed to a web-based portal in Qualtrics. Data were collected from July 16 to August 16, 2018.

Respondents were asked about their professional background, how they use data in their work, which nutrition indicators and data sources they accessed or used in the previous year as well as unmet data needs. Respondents were asked to identify which country-specific and consolidated data sources they accessed in the last 12 months. Consolidated data sources are those that present secondary data from multiple countries, typically using multiple data sources (eg, GNR, UNICEF State of the World’s Children Report, Scaling up Nutrition Monitoring, Evaluation, Accountability, and Learning (SUN MEAL)). Several questions allowed for multiple responses. Additional follow-up questions were asked about the data sources used by respondents that reported use of coverage data related to common interventions implemented in many countries, including routine growth monitoring for children, screening for acute malnutrition, treatment for severe acute malnutrition (SAM) or moderate acute malnutrition (MAM), vitamin A supplementation, iron folic acid (IFA) supplementation for women or adolescent girls, multiple micronutrient supplementation for women or adolescent girls, breastfeeding counseling, and complementary feeding counseling. The survey tool is available in Appendix S1 of the Online Supplementary Document.

Results were tabulated and compared across categories of geographic focus (single vs multi-country) and organization using Pearson χ-2 statistics calculated in Stata Version 14.0 (Stata Corporation, College Station, TX, USA). Respondents were not required to respond to all questions and therefore the sample size varies by question. The survey received a “Not Human Subjects Research” determination by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board.

Respondent characteristics

A total of 264 survey responses were received; 235 with responses providing information beyond the identifiers section. The majority of respondents were from NGO and research communities, with relatively few respondents working in government (Table 1). The sample was highly educated and experienced, with over 90% of respondents having a master’s or doctoral degree. Two-thirds of respondents made nutrition-related monitoring and evaluation decisions in their professional roles. Of the 84% of respondents who self-identified as a technical expert on nutrition related issues, the most common domains of expertise included infant and young child feeding (IYCF) (74%), child nutrition (71%), and maternal nutrition (59%).

Geographic work focus of respondents

The survey sample was evenly split between those who worked in a single country (49%) vs across multiple countries (51%). Among those who worked in a single country, half said their primary work focus was at the national level and half at the subnational level. Among those who worked across multiple countries, 40% worked at a global level, 25% at a regional level (eg, North Africa, Southeast Asia), 22% at a national level, and 14% at a subnational level (eg, state, district). Respondents were asked to list all the countries to which their work related in the last year, totaling 116 unique countries. By WHO region, the majority of respondents worked in at least one African country (63%), followed by South-East Asia (40%), the Americas (19%), the Eastern Mediterranean (17%), the Western Pacific (16%), and Europe (6%). The most frequently cited countries were India (n = 57), Ethiopia (n = 56), Kenya (n = 44), Bangladesh (n = 44), and Burkina Faso (n = 37).

Data for decision-making

When asked about their current role working with data, 72% reported that they make decisions using data analyzed by others, 63% consolidate or analyze data for decision making within their institution, and 49% consolidate or analyze data for external decision making. Decisions made using these data varied by single and multiple country focus. Respondents working in a single country were more likely to make decisions related to program administration compared to those who worked in multiple countries (25% vs 15%, P = 0.0499). Those with a multicountry focus were more likely than single-country focus respondents to make strategic program and policy planning decisions (49% vs 26%, P < 0.001) or high-level financing decisions (18% vs 7%, P = 0.014).

Access to country-specific and consolidated sources of data

Most respondents (81%) accessed at least one country-specific data source (eg, national household surveys, routine facility data systems, surveillance systems) in the last 12 months. Of those who accessed country-specific data in the last year, the DHS [8] were by far the most common (74%), followed by the MICS (42%) (Table 2). Respondents with a multi-country focus were more likely to access the DHS in the last year (85%) compared to those with a single-country focus (60%) (P < 0.001), a pattern that was even more apparent for the MICS (65% vs 16%, P < 0.001). Routine facility data sources such as the DHIS-2 and Health Management Information System (HMIS) were less accessed compared to household surveys, accessed only by about 30% of respondents, with no meaningful difference by respondents working at single vs multiple country levels. However, the DHIS-2 data was accessed more by single-country respondents working at a national level than those working at a subnational level (40% vs 21%, P = 0.048). The Service Provision Assessment (SPA), a health facility survey, was only accessed by one in ten respondents (11%).

Most respondents accessed at least one source of consolidated data in the last 12 months (75%), of which the most common was the GNR [9] (Table 3). The GNR was more likely to be accessed by data users with a multi-country focus (82%) than those with a single-country focus (66%) (P = 0.014). Multi-country focused respondents were also more likely to access the UNICEF State of the World’s Children Report, other UNICEF nutrition data sets, and Countdown to 2030 data products (all P < 0.05).

Access to specific types of indicators

When asked about types of data on nutritional status accessed in the previous 12 months, the most frequent responses were child anthropometric indicators, with the majority accessing data on child stunting and wasting (Figure 1). In comparison, fewer respondents accessed data on the nutritional status of other populations (adolescents, adults and school children). Data on IYCF practices, including breastfeeding and dietary diversity, were also accessed by most respondents (Figure 2). In contrast, less than a third of respondents accessed data related to the diets of women and adolescents (Figure 3).

Figure 1.  Access or use of nutritional status data (N = 235).

Figure 2.  Access or use of infant and young child feeding practices data (N = 235).

Figure 3.  Access or use of adult diet quality data (N = 235).

For intervention coverage and utilization indicators, more than half of respondents reported accessing data on coverage of breastfeeding counseling and complementary feeding counseling (Figure 4). Household surveys (69%) and administrative data (39%) were the most common sources reported by those who accessed breastfeeding counseling indicators (Table 4). The same two sources were reported at similar levels for complementary feeding counseling data. About half of those accessing breastfeeding or complementary feeding counseling coverage data wanted these data collected annually.

Figure 4.  Access or use of coverage or utilization data (N = 235).

Half of the respondents reported accessing data on IFA supplementation during pregnancy (Figure 4). The specific IFA indicators used or accessed varied: any IFA consumed (77%), minimum number of tablets consumed (58%), IFA purchased or received (49%).

About half of respondents reported accessing data on screening for acute malnutrition and/or treatment of SAM. More respondents reported accessing screening data from household surveys than administrative data. In contrast for SAM or MAM treatment, administrative data sources were most accessed (Table 4). For both screening and treatment, about half of respondents who reported accessing the indicators in the last year were satisfied with the current periodicity of available data. The most commonly accessed or used data on nutrition-sensitive interventions were data on water, sanitation and hygiene (Figure 5).

Figure 5.  Access or use of nutrition sensitive interventions or drivers data (N = 235).

Compared to respondents with a single-country focus, those with a multi-country focus were more likely to access indicators for under-5 wasting, overweight, and vitamin A deficiency, adolescent overweight, women of reproductive age stunting, underweight, and anemia, and gender inequality (all P < 0.05).

Challenges to accessing and using data

Several challenges to accessing and using nutrition data were reported by respondents. Most respondents reported that data were sometimes or frequently not available at the geographical disaggregation needed (82%), for the demographic group needed (77%), or were out-of-date (77%) (Figure 6). There were no significant differences between multi-country and single-country focus for reported challenges. Responses to an open-ended question about data gaps identified nutrition-sensitive intervention coverage, IYCF promotion and counseling coverage data as priorities.

Figure 6.  Challenges experienced in accessing and using nutrition data (N = 196).

Limitations

The generalizability of our survey findings to the broader global nutrition community is uncertain. The survey was disseminated broadly, using established listservs, institutional contacts, and Twitter. This study utilized convenience sampling; given our dissemination channels, those without reliable internet access and those who were not connected to the listservs or professional networks utilized may have been underrepresented. We had very few respondents from LMIC governments, a high priority user group. We also had very few respondents with less than a college education, perhaps suggesting under-representation of front-line workers. Further effort is needed to better understand how national and subnational government stakeholders access and use nutrition data.