United Nations System
Standing Committee on Nutrition

One Day Pre-Session to the 28^th ACC-SCN Meeting in Nairobi

Date and Location: Sunday, April 1, 2001 (9:00)

Objective: To evaluate the first year of formal operation of the ANDI (African Nutritional Database Initiative) in terms of implications for setting up a common inter-agency Food Security and Nutritional Status database with world-wide coverage under the inter-agency FIVIMS (Food Insecurity and Vulnerability Information and Mapping Systems) initiative.

Invitees: Close associates of the ANDI project, members of the FIVIMS sub-group on the Common International Database, and any attendees to the 28^th ACC-SCN who are interested in the topic – all are welcome

Hotel and Other Arrangments: All participants are responsible for their own travel and hotel arrangements. The Secretariat of the ACC-SCN can provide assistance in making hotel arrangements -- contact at accscn@who.int

Let us know you are coming: In order to make sure we have a big enough room and to make any other needed arrangements (tea, etc.), please let the FIVIMS Secretariat know you plan on being at the workshop. Contact either or both: David.Wilcock@fao.org Jenny.Riches@fao.org

Background:

ANDI, with its development originally sponsored by the World Bank and Italian Cooperation during the period 1997-99, has involved the World Bank, FAO, UNESCO, and UNICEF providing data electronically into a common database managed by Dr. Dario Berardi in Rome. FAO has provided funding to the ANDI in its first year of formal operation and the World Bank will cover the second year of database maintenance. FAO sponsorship has been motivated by the desire to draw lessons from this experiment in sub-Saharan Africa, for a similar but larger database initiative focused on food insecurity, vulnerability and nutritional surveillance, under the sponsorship of the 25 plus members of the Inter-Agency Working Group (IAWG) on FIVIMS.

In 1999, the IAWG FIVIMS created a working group on Global FIVIMS with the main charge of investigating and making proposals concerning the creation of the Common International Database under FIVIMS. This working group met in October, 1999 at the World Bank, and proposed an ANDI-like database structure that would be called KIDS, or Key Indicator Database System. The report of this inter-agency working group is annexed to this proposal and was discussed an approved with minor modifications at the 4^th meeting of the IAWG FIVIMS in Bali, Indonesia in February, 2000. During the balance of 2000, most of the development effort within FIVIMS has been devoted to developing a prototype web-based KIDS software package. This has been done largely within a regional context with the development of Asia KIDS under funding from the Japanese government.

The purpose of the one day pre-session to the 28^th meeting of the ACC/SCN is to present and consolidate lessons learned in these different development efforts (ANDI, KIDS) and especially to focus on the key issues of the required inter-agency collaboration from the agencies that will be the major sources of data for the global database on food insecurity/nutritional surveillance. The ANDI website and the KIDS web-based software represent two experiments in ways of presenting the data from such a common database effort. The important issues that need finalisation are scope, geographical coverage, and degree of sub-national detail that should be in a common inter-agency food insecurity/nutrition database. In addition, the ANDI lessons, on how to work with different agencies and their databases, are vital to the future success to a global FIVIMS database.

Annex IAWG 4/8

FIVIMS (IAWG – Key Indicators Data System Group)
Don Mitchell, Kurt Vertucci, Naoki Minamiguchi, James Rowland, Hambisa Disassa and Jean-Charles Le Vallee
Washington D.C., October 18-20, 1999

Food insecurity and vulnerability remains a major problem in many countries and is the focus of FIVIMS activities. An internationally comparable database of key indicators can improve the ability of governments, civil society, international donors, and others to understand the causes of food insecurity and vulnerability, and prepare strategies and programs to deal with these problems. Such a system would make it easier to combine data from different agencies and examine causal relationships between demographic, economic, environmental, and other key indicator groups. The data needed for analysis already exists in individual UN agencies such as FAO, WHO, UNICEF and in international organizations such as the World Bank. However, much of the data is not easily available or is not available in a common format. Developing a system, which provides easy access to the best indicators and the tools for analysis, requires cooperation and coordination among data providers, and resources to develop the tools for analysis. The Key Indicators Data System (KIDS) is intended to provide such a database in an easily accessible and widely available form using the Internet and CD-ROM distribution under the FIVIMS activity.

2. OBJECTIVE OF KIDS

The objective of KIDS is to contribute to the better understanding of the causes and dimensions of food insecurity and vulnerability and the reduction of these conditions as mandated by the World Food Summit. It seeks to accomplish this by providing users with improved access to select reliable indicators and the tools to analyse these data.

The primary benefit of KIDS will be to allow easier and faster access to key indicators, which have been selected by knowledgeable experts to be useful in understanding food insecurity and vulnerability. In this way, it will improve the quality of research and analysis. Indicators will be selected to provide the best available data from the primary data provider. Data will be updated on a regular schedule. KIDS will also allow basic statistical analysis of the data and provide tools for graphing and mapping data. The site will also provide links to national and regional food security systems.

The primary users of KIDS are expected to be those individuals and groups at the national and international level working on food insecurity and vulnerability.

Users could include policy makers and researcher at government agencies, donor agencies, NGOs, universities, and the private sector. International organizations dealing with food insecurity and vulnerability such as FAO, World Bank, UNDP, WFP, UNCTAD, UNICEF, UNESCO WHO would also likely find the KIDS data base useful.

Concerns may arise from data providers about the design, development and implementation of KIDS. These concerns could include:

• loss of recognition as the data provider
• loss of control or supervision by sharing their data with other host providers

• misuse and/or mis-representation of data
• loss of revenue from sale of data
• time and resources required for providing data
• possible introduction of errors

While not all of these concerns can be fully alleviated, every effort will be made to address these concerns.

6. TECHNICAL ISSUES

Should the KIDS data be located in a common database at a single site or be selected through a common interface, accessing the data at the data provider’s site?

• Higher control on the quality of the data – Data sets distributed by Organizations externally generally go through a series of quality control checks prior to being shipped. Data accessed directly and automatically by a query engine against Organizations "live" database may bypass these quality checks.
• Improved ease and convenience for user – The user would be able to process one query to access and retrieve data from multiple agencies in a common format.
• Higher access and retrieval speed – Single query performed on single database; single site as opposed to multiple queries running on multiple sites.
• Higher control of service – The possibility of a query succeeding are greater if processed at one site rather than five or six different sites.
• Security – A common database approach would not require that the KIDS query engine needs to access machines within the data provider’s security firewalls.

• Considered by some data providers politically disadvantageous
• Central site must provide all system operational support.

• Administration of data and operations is spread out and part of normal agency activity
• Politically attractive to some agencies

• Slower speed - running multiple queries from multiple sites
• Less reliable – running queries from multiple sites
• Less Quality Control – interactive queries to live database, bypassing quality checks.
• Higher cost of development and maintenance of distributed access programs

KID’s will be able to allow for both methods of access to data, as well as variations of the two basic access methods, in order to satisfy the concerns of data providers. Although some of these access methods may be more awkward or time-consuming for users, this is preferred to not having access at all. A further discussion of some other suggestions for addressing data sharing concerns are elaborated in Section 8 of this report, entitled Ownership of Data.

There are also some concerns about the response time of an internet-based application in certain areas. This concern extends also to areas of no Internet connection. It is recommended that the KIDS technical architecture allow the Internet site and its data to be distributed also on CD-ROM. Obviously, data not provided to the KIDS Database, but only linked via the internet will not be available on a stand-alone CD-ROM product.

Each data provider has adopted their own country coding systems (e.g., ISO3, UNDP, FAOSTAT codes). The system must be able to import and export the data regardless of the data provider’s country coding system.

The system, centralised or decentralised, must be able to update the data in a timely fashion upon request of the data provider, because of a scheduled data update cycle or to correct an error.

KIDS database will be able to export data that can be directly ingested and displayed in KIMS (Key Indicator Mapping System). KIDS will serve as a source of updated data for KIM’s users.

The KIDS should allow provision for disclaimers for data provided by the collaborating institutions. This can be in the form of one general disclaimer for the KIDS site, or separate disclaimers for each of the data providers. Data will be provided in its original form, without modification. Acknowledgement and authorship of data source will be provided. Any products derived from the data will acknowledge the source of the original data, as well as the author of the derived product(s). KIDS will respect copyrighted information, and the site itself should be copyrighted against private resale

7.1 Metadata (meaning description of data or information on information)

Metadata must be given for each dataset or indicator provided by collaborating institutions. Source, authorship and use agreements are part of such metadata. This serves the purpose not only to better inform the end-user on the information but also to protect the providers on the use of this data.

The major sources of data will come from several United Nations agencies of which each is currently owner of their respective data. By sharing their data, the data also becomes part of KIDS and thus FIVIMS. Theoretically, each agency’s data is available to the public, and it will remain available to the public through FIVIMS. By gathering data from these agencies, and arranging the data in a certain format for food security research, (having put an effort in collecting and rearranging them, making it available with a better presentation, more user-friendly, mapping tools, etc.) FIVIMS gains part ownership in them as it has contributed to improving the data. But having shared this data with FIVIMS, some agencies might feel they have lost some control on the content or accuracy. Agreements on this issue will be necessary between the agencies (data providers), and FIVIMS to ensure that these providers are comfortable with such an arrangement. One suggested proposal was to give multiple ownership to all data providers to KIDS.

One method suggested to overcome the political reservations of the centralised approach, is to have all data providers to mirror any other data providers KIDS data on their own site.

Another approach, which extends the mirror site method, is allowing each participant to host the KIDS data in their own dissemination system either internally and/or externally.

Costs are divided by the data providers and the data disseminators.

Data provider costs should minimal. It is the cost of providing data that they are already providing to other "clients" to FIVIMS KIDS. It is expected that the data providers in-kind would contribute these minimal costs.

The costs of the data dissemination includes onetime and on-going costs covering the following areas:

Co-ordination costs
Development costs
Hardware/Software costs
Operational costs
Maintenance costs
Distribution costs – of KIDS products

These costs would need to be funded by FIVIMS.

In an attempt to recover some of the FIVIMS KIDS costs, there may be some recovery from the commercial market through the sale of KIMS CD-ROM version. In this case, all revenues would be put back into the operation and improvement of KIDS; any excess funds would be distributed equally among the data providers.

During the year 2000, the work related to the common Key Indicator's Database will focus on two major outputs:

1. A data exchange agreement with agencies providing KIDS data
2. A working prototype of the KIDS Database (Internet and stand-alone CD-ROM)

It is suggested that the World Bank take the lead role in developing the data exchange agreement together with WHO. The World Bank brings its experience with the development of the World Development Indicators publication and CD-ROM, which includes key indicators and data provided from several UN agencies (e.g. WHO, UNESCO, FAO), as well as its experience with the recent ANDI Nutritional Database initiative. WHO brings its experience and role in the data dissemination of major FIVIMS related health indicators.

For the development of the KIDS working prototype, it is suggested that FAO, the other co-chair of the KID’S sub-group, take the lead role working closely with WFP. This arrangement benefits from FAO's existing work plans for the development of a similar database and takes advantage of WFP's experience in data collection and dissemination of FIVIMS related data in the field. The proximity of the two Rome-based Organizations allows for easy and cost effective working relations.