Can the Nutrition Information System be ‘trusted’ to build on available data sources?

By Patrizia Fracassi

Patrizia holds an M.Sc. in Development Management. Over the past two years, she has consulted in Ethiopia for UNICEF and the World Bank. She previously worked for UNICEF Uganda as a Nutrition Specialist and for Italian NGOs, CESVI (Cooperazione e Sviluppo) and Ucodep (Unity and Cooperation for Development of Peoples, now Oxfam Italia) in Vietnam as Country Representative and Programme Manager. She also developed and manages the technical content for the website: www.motherchildnutrition.org.

Patrizia would like to acknowledge teammembers from the federal and regional Emergency Nutrition Coordination Units and from UNICEF Nutrition and Food Security Section for their continuous support, and especially government representatives at all levels for their willingness to share information and give their valuable insights.

The findings, interpretations and conclusions in this article are those of the author. They do not necessarily represent the views of UNICEF, its Executive Directors or the countries that they represent and should not be attributed to them.

In Ethiopia, the role of the national Nutrition Information System (NIS) has been clearly stated in the Ethiopian National Nutrition Programme (NNP). There are three constituent parts to this role. These provide a 'comprehensive' and holistic structure to NIS design: to support timely warning and adequate interventions at woreda and higher levels, to develop, manage and evaluate the NNP at all levels, and to inform other sectors like agriculture, water/sanitation and economic development. This comprehensive vision for the NIS is to inform understanding of the nutritional situation with respect to chronic and newly occurring problems, as well as the causes of these problems, and how these change over time in order to help in decision-making at all levels. However, while the NIS can effectively accommodate and be 'open' to an unlimited amount of data, the ability to trigger an effective and appropriate response requires that the information is timely, reliable and consistent. These conditions ultimately determine the basic parameters upon which the initial choice of information for the NIS is made. Put simply, all data should be trusted and continuously available, data should be triangulated to generate 'context-specific' and evidence-based information and there should be a clear process, agreed by all actors, to feed information into decision-making.

Unique data situation in Ethiopia

Ethiopia is in quite a unique position because, over the last thirty years, large amounts of data have been collected by the Early Warning System (EWS), including health and nutrition information. However, the nutrition information collected by the EWS (see Box 1) provides only scattered data - mostly alert signals based upon 'observable' degeneration. Data is collected directly from health workers at 'critical times' and without systematic comparison with what would be 'normal' for a given time of year. Nutritional assessments are required during these critical times to confirm 'emergencies' but the seasonality of these critical times creates a widespread, simultaneous demand for assessments, which rarely can be adequately met.

Child screening in Fedis woreda during Community Health Day

In recent years, targeting of surveys has been improved through increased use of routine data sources, at least to indicate where an assessment is most urgently needed. Nutrition data are now available and accessible on a monthly and quarterly basis at the lowest levels due in large measure to three programmes: The Community Based Nutrition programme (CBN), the Therapeutic Feeding Programme (TFP) and Community Health Days (CHD) (See Box 1). These routine systems are the monitoring backbone of the NNP, which - at least theoretically - can be combined to inform timely warning and be shared with other sectors. Similarly, a number of diseases are also currently being tracked on a weekly basis through the Public Health Emergency Management (PHEM) system (see Box 1). Thus, there is a very real potential for the EWS to systematically tap into specific data from existing health information sources and vice-versa. This will be most effective if a consensus is reached on key indicators, in particular for timely warning. The key question, ultimately, is whether decision-makers from all sectors are willing to exchange and use available routine data to inform their decisions and response.

Nutrition data management

While there are 'trust' issues on data quality and credibility, the administrative decentralisation and existence of a widespread health network, creates the rare opportunity to build capacities, accountability and transparency at lower levels like the woreda and the kebele.

Initial data collectors are volunteers and frontline health practitioners. Many report that data collection is an additional burden to their already crowded agenda. After the initial collection, data flows up through various levels via supervisors and health officials. However, little feedback is given through the system so that people directly involved have a limited sense of what is actually done with the information provided. The sheer volume of reports stored testifies to the regularity of data collection undertaken and the immediate priority that should be given to improve the 'efficiency' of the process. Currently, asking for nutrition information from a woreda official leads to a paper-chase given the amount of report forms collated. Where officials have been provided with a computer, data appears to have been regularly updated. Given the increased requirements for information management, it seems inevitable that woreda Health Offices will move from a paper-based system to a computerised one, allowing them to perform data quality checks that otherwise are time consuming and prone to mistakes if done manually. The implication here is that woreda level officials are mostly young, often computer-literate, professionals with degrees. Provision of adequate tools/software to practically manage information can help build their capacity to implement the system. If information is not properly valued at woreda level, where most data are collated and 'checked', then the task of quality assurance at higher levels is nigh-on impossible.

Added value of NIS: triangulation of data

Children with their caregivers on OTP day in Chalenqo Health Centre, Fedis woreda

What is 'new' in the NIS paradigm is the requirement for 'triangulation' to provide evidence-based information for decisionmaking. This implies that collected data are not interpreted in isolation but are brought together from different sources. The strength of triangulation is the 'contextualisation' of the data, meaning numbers and/or standardised observations are grounded in 'local knowledge'. Frontline practitioners in health-posts have access to nutrition and health information through regular contact with patients. With nutrition, for example, they are in the best position to judge if the deteriorating weight of a child during monthly growth monitoring or his/her admission in the Outpatient Therapeutic Programme (OTP) is linked to lack of food in the family or to other causes like illness, inappropriate feeding practices, etc. It is this 'proximity' that allows for the triangulation to be most helpful at community level whereby root causes of malnutrition can be identified. An example where this could be used is in chronically food insecure areas supported by the Productive Safety Net Programme (PSNP) where risk financing mechanisms exist to address new chronic or temporary food insecurity. By monitoring increases in underweight (as an early indicator) and OTP admissions (as a late indicator), frontline health practitioners, who are members of the Food Security Task Forces (FSTF), can play a crucial role in providing information for appeal processes. However, the credibility of their information will depend on their full understanding that risk financing mechanisms are only accessible when malnutrition is linked to food insecurity. Thus, triangulation of data at source is a kind of check by key people before information is fed into the decision-making processes or reported to higher levels.

Trust, accountability and transparency

Talking about data with Development Agents in Checheho kebele, Lay Gayint

In Ethiopia, in line with governmental decentralisation, woreda and kebele level administrations have been given increased power to analyse, assess and act on their own changing situation. They are therefore more responsible and accountable for both development and emergency response. Addressing the challenges of how information can feed into decision-making will ensure the credibility and sustainability of the NIS. At the moment, available data from routine sources are not adequately linked to information use. The main challenge for data utilisation at higher levels is that sources are not fully trusted while at lower levels there are limitations over capacities and mandate. While data quality assurance can be built into the system, especially by improving lower-level capacity, more emphasis needs to be given to the human aspect. 'Trust' cannot be built without attention to the role played by each stakeholder, starting with frontline practitioners. 'Accountability' cannot be acquired if there is no hand over of responsibility. 'Transparency' cannot be promoted without making response and feedback more visible.

The NIS in Ethiopia can be built upon coupling available data sources with adequate technical support provided throughout the health system. However, technical inputs are not enough to ensure its sustainability. A sense of 'value' is what motivates people and without it, the simple transmission of data to higher levels will not provide incentive to stakeholders for their input. 'Triangulation' is most effective at community level where individual data sets can be compared at source and understood within a given context. Frontline practitioners play a crucial role in building the credibility of the NIS but this can only come about with increased recognition of the role they play in informing decision-making. As the process of decentralisation continues within Ethiopia, important decisions to be taken at the lowest levels, risk financing mechanisms being an example, will require bringing together available data from different sources. This, in turn, will rely increasingly on key people accountable for informing this process at the frontline. Before trusting the Nutrition Information 'System', a vote of confidence should be given to empowering the information 'Source'. Credibility, after all, should always start with the people.

For more information, contact: Patrizia Fracassi, email: pat.fracassi@gmail.com

Box 1: A guide to key systems and programmes in Ethiopia implemented by the Government

Queuing during HD in Fedis Woreda

The Early Warning System (EWS) is implemented nationally under the Disaster Risk Management Food Security Sector (DRMFSS). Information are collected at kebele level by Development Agents and passed to the Woreda Early Warning Food Security Task Force. The Task Force analyses and interprets data and submits reports to the regional level and to the Woreda Council. Collected information include: rainfall condition, crop condition, migratory pests, input supply, livestock condition; human health condition, water access and availability, education related information, fast on-set disasters (flash and seasonal river floods, landslides, conflict, forest fire outbreak and livestock diseases outbreak and human epidemics), grain, livestock and other commodities prices, coping mechanisms, emergency relief responses. Information is fed into decision-making processes for classification of emergency affected areas/woredas and relief allocation. In Productive-Safety-Net Program woredas, EW information is used to help programme the newly established Contingency Funds. In addition, the DRMFSS is responsible for the implementation of bi-annual seasonal (belg and mehr) multi-sectoral assessments. Information is used to identify acutely food insecure woredas and estimate the number of affected population in need of relief.

The emergency nutrition surveys are coordinated by the regional and federal Emergency Nutrition Coordination Units (ENCU) under the DRMFSS. Surveys are conducted by non-governmental organisations (NGOs) and/or government multi-sectoral teams using the SMART methodology. ENCU is responsible for the technical quality through an approval process of the initial proposal and final findings. Request for an emergency nutrition survey should come or be endorsed by the woreda/regional authority based on EW information and/or results from the biannual assessments. Information is used to confirm 'emergencies' and mobilize response for relief and selective feeding programmes.

The Community Based Nutrition Programme (CBN), started in 2007, is now expanding in 150 woredas covering all kebeles. The programme is implemented by Volunteer Community Health Workers (vCHWs) under the supervision of the Health Extension Workers. It includes monthly Growth Monitoring and Promotion (GMP) of children under two years with referral of those who are severely underweight, not gaining weight for two months or with other health problems. Pregnant and Lactating Women (PLW) receive folic acid supplementation and are mobilised for antenatal care, safe delivery and other maternal services. Community Conversation (CC) is conducted monthly to assess child malnutrition, analyse causes and plan for action (Triple A Cycle Approach). Collected information includes: number of children under two years, number of weighed children, children with normal weight, underweight and severe underweight, number of trained and reporting vCHWs, number of implemented CC sessions and participants. Indicators are disaggregated by gender.

The Community Health Days (CHD) are implemented every three-months and expected to cover all 150 CBN woredas. MUAC screening is conducted for all children 6-59 months and for PLW. Those identified as malnourished are referred to therapeutic feeding programmes or supplementary feeding, where available. Every six months, children are additionally provided with Vitamin A and deworming. Collected information includes: targeted population (children 6-59 months), coverage of child and PLW MUAC screening, coverage of Vitamin A and deworming supplementation. Children and PLW are grouped on the basis of MUAC as follows: MUAC >12 cm, MUAC between 11-11.9 cm and MUAC <11 cm, PLW with MUAC <21 cm. Data are collected by the Health Extension Workers and volunteers and sent to the Woreda Health Office for further transmission to the zonal, regional and federal level. Where supplementary feeding is available, screening results are shared with the DRMFSS to establish the number of beneficiaries.

The Outpatient Therapeutic Feeding Programme (OTP), as part of the management of severe malnutrition, is now covering more than 6,000 sites (health centres and health posts). Collected information includes: number of in-charge beneficiaries, number of admitted beneficiaries (severely wasted, with oedema, transferred, etc.), number of discharged beneficiaries and performance indicators (cured, death, defaulter, etc.). Data are collected by the Health Extension Workers and health workers and sent to the Woreda Health Office (WorHO) for further transmission to the zonal, regional and federal level.

The Public Health Emergency Management (PHEM) system, coordinated by the Ethiopian Health and Nutrition Research Institute (EHNRI), is expected to be implemented on a nation-wide scale. Immediately reportable diseases include polio, anthrax, avian human influenza, cholera, guinea worm, measles, neonatal tetanus, pandemic influenza, rabies, severe acute respiratory syndrome, smallpox, viral hemorrhagic fever and yellow fever. Weekly reportable diseases are dysentery, malaria, relapsing fever, typhoid fever, typhus and severe acute malnutrition. The latter has been only recently added among the weekly reportable diseases. Health facilities have to inform the WorHO within thirty minutes after they see an immediately reportable disease. Further data transmission from woreda to zone/regions and from zone/region to EHNRI is expected to take place within an hour using any available communication mechanism. Health facilities (including health posts) report data from Monday to Sunday every Monday to WorHO. Data are expected to arrive at EHNRI every Thursday.