population-based information, they do not provide data to evaluate the effectiveness of different service delivery programs that have recently been developed.
Unlike family planning, other sectors have not relied so exclusively on survey systems for evaluation purposes. Tsui emphasized the importance of learning from other sectors and expressed her appreciation to the participants who came from outside the family planning field. She expressed her hope that the meeting be an open dialogue among the various sectors so that everyone might learn something valuable from the others.
FACILITY INVENTORIES
Luis Rosero-Bixby began by citing several problems with traditional measurements of physical accessibility to clinics and suggested that data from geographic information systems (GIS) be used for improved measures and more precise analysis. Three problems with traditional measures are that they do not: (1) measure individuals' use of services in communities or areas other than those in which they live, (2) measure the extent to which different health facilities provide comparable services in the same geographic area, or (3) take the density of the population into consideration, that is, they do not recognize that, given a certain number of health facilities, individuals who live in more densely populated areas will have less access. Rosero-Bixby illustrated how, using GIS, researchers can prepare maps for an entire country, combining data on the geographic location of health clinics and population distribution, to overcome these shortcomings. Such maps, which could include
roads and geographic barriers, would show the location of health facilities relative to the density of the population and would make it easy for policy makers to locate health clinics that provide overlapping services in a given area and to discover areas in need of clinics. GIS would also show whether clients have greater access to clinics that are not the closest, because of road systems.
Joan Aron discussed the issues that researchers need to consider when deciding whether or not to use GIS. GIS are systems in which individuals can input, store, manipulate, analyze, and output spatial data. As such, one can use GIS to achieve a number of program-related goals: (1) data storage, (2) data management, (3) data querying, (4) modeling, and (5) decision making. Data for GIS come primarily from administrative records and survey data. She noted that a common misperception is that it is mandatory to use satellite imagery (remote sensing) data in GIS. Satellite imagery is only one type of data that may be used, and it presents several challenges: first, commercially available satellites do not constantly monitor the earth, so the needed data may not be readily available; second, satellite maps are very expensive; third, they are not easy to read and experts are needed to decipher the photos; and fourth, validation of data at the ground level must be undertaken. One key limitation to the success of GIS is the quality of the data entered into the system. Aron concluded by stating that, although GIS can be very useful, one needs to think carefully about whether it is appropriate for a particular project.
During the discussion, concerns were raised about whether data gathered through GIS would be more revealing than data already being collected. Although that would be difficult to determine, there was some optimism. Spatial variation in unmet need for contraception,