2009 Implementers Group Meeting Program Geographic Information Systems

Overview

This is page relates to the GIS session in breakout room 2. It was clear that there are many unique use cases and many different tools being used.

Action

We need to get together and consolidate in order to move towards aligning efforts and not duplicating work.

Working Group

Please sign up for this working group: GIS Working Group

Extremely Rough Notes

These notes may or may not get edited.

Using GIS:
	3 people, Peru, ?, ?

Kenya (Eldoret) collecting coordinates, but not using.

Moz: Are developing GIS system, wants to know best tools and best practices

Request for information about how different countries have delt with location in OPenMRs

More broadly: Integration with OpenMRS

Isaac:

	Analogy with mobile: Working on common integration platform

	Goal: try and develop similar intrgration methods with GIS

	What would module do: represent geographic information generally

	Interface with different mapping tools - interface layer

	Identify common database format

Carol:

	What to find a way to use GIS for public health

Cheryl:

	How can we use this for service delivery?

James:

	How Kenya is collecting data and what they're doing with it? (Use cases)

Isaac:

	Is GPS too narrow, how do you deal with location?

From Eldoret:

	Collecting GPS coordinate for home based counselling and testing. Coordinates are used 

	for following pregnant mothers not enrolled in ANC who may or may not be HIV+

	Used: For followup (not widely used) and to get additional information [90% success rate]

	Would like to use for epidemiolocal analysis

Cheryl:

	Using it for addresses basically (which would be easy in other countries)

Eldoret:

	Using PDA and Garmin Etrex GPS device

	PDA to computer synchronisation

	Challenge Pendragon is based on Access which is difficult to integrate with OpenMRS.

        There is an initiative going on to program form on devices (e.g G1 phones) that can    update OpenMRS.


Dave:
	Similar in terms of how data is captured

	Wrote interface for LIMA

	Use Garmin hendhelds

	Garmin publishes JS API

	Morning: Sync with coordinates (addresses and id's)

	Afternoon: Sync again and new obs get entered into OpenMRS 

	7000 TB cases

	Sputum gets genotyping and the plan is to link geographic data

	Lots of cluster analysis will be done

	Would like sms to be used to enter coordinates into OpenMRS database

Carol:

	What is the use for the data/tech after the study

Dave:
	Wrote OMRS interface for the study

	Goal: Identify certain gene knockouts

	Want to know what makes strains easy or difficult to transmit

Jeremy:

	Use GPS data to determine if patients are going to the right clinic

	Topographic analysis for ease of travel

	Sending mobile clinics

Aamir:

	(His work inspired Cheryl to pilot her GIS work)

	Use: Monitor field teams and supervisors

	$250 GPS enable nokia phones

	Users log in and store info

	Log into Google earth interface to manage staff

		see who's logged in (phone and ask if they're not on schedule)

	Completed forms have date/timestamp

	Small team to monitor large system/team

	Manage (monitor) vehicles (food basket delivery and other services)

	Also keep track of how long it takes from one place to another

	Monitoring allows nation system to be managed by one person in the CApital

	Patient CAre:
	~~~~~~~~~~~~

		TB and MDR TB patients are being mapped to their house

			limits lost to followup

			other health workers can find the house

			monitor that health worker sees patient twice a day

	Using chinese device with only one button and GPS which the patient presses when at home
	because not enough health workers to follow all patients home. $85 bought in bulk

	Allows high volumn of patients to central facility, and hand out devices and press when at home

	losses of device are acceptable

	Another use:

		Very ill patients could press other buttons (there are 5 buttons in total) if they fall over or something

		So it's a personal panic button with GPS location
	
	Another use:

		When they plot TB patients, they include census data

		They get a sense if they should have more patients in a certain area

		Help field teams to go in and meet with practitioners


	Workflow:

		Login form

		Surveilance form, Service form

		Logout form

	
	For MDR-TB module, can click on patient in google earth interface, patient info is displayed on a popup in real time

	Want patient alerts on the google earth interface so managers can sent health workers to certain areas

	Want to allow health workers to attach voice packets (via phone) linked to patient info on google earth. Also alerts.

	Busy building for monitoring, haven't had time to think about evaluation

	Esp interested because the system and 8 others have been budgeted for by Pakistan

Carols:

	All systesm will be good for research and interesting questions can be answered

	Need to sustain the systems

Aamir:

	Need to stop development and focus on integration

	Need a single stable download


	Problem: money availble after systems developed

Issac:

	How is OPenMRS talking to the systems

David:

	Thinks its directly at database level

Isaac:

	UShahiti(?) is open source mapping tool for crises

	Has some common developers with Baobab and OpenMRS

	MAy consider open source to make integration easier

	Solves cluster mapping problem (high concentrations can get lost)

Cheryl:

	Interesting the variety of use cases (SErvice, Research, Management)

WHO:

	Working with aggregated data

	Uses: service availablity mapping

	India, Zanibar, Tanzania: GIS to look at aggregated data

	Combine with monthly service data

	Basically service use mapping

	Look at something like maternal death, ask where you have birthing facilities

	working with DHIS and GIS for a while

	WHO Health Mapper, working closely with

		Tool that WHO has provided to support service availablility mapping, import sources
			
		Import your own data

	Nice to have mapping separated from data

	GUext is javascript mapping tool

	Used by managers, so needs to be simplified interface (giving overviews etc)

	Health Mapper 5 will be web based

	Using in Sierra Leone (not under Health Mapper name)

	No release date for Health Mapper 5

	launchpad had DHIS GIS project


Cheryl:

	Rwanda has use for mapping villiages

	Current standard for chief;s house to be centre of villiage

	Looked mapping individuals and villiages

		difficult for scaling up
		
		can't answer all research questions

	Have a few thousand villiages

	Want to attach villiages to address hierarchy

	Some patients will have two levels of address

	Need to work out of how to define locations

	Thinking about hosing coordinates in PostGIS

		have been collecting more than health information

		can use for analysis

	Want to develop and integration module


Carol:

	What are the different ways of mapping locations

	Wonders about different ontologies

	Public health implications

Rwanda:

	problem: what units can we use to map space

	have: distrit, cell, umadugadu(?)

	government changes categorisations of areas

	cost of mapping lowest categories is expensive, so getting centroids

	modelling and projections are being dones (e.g. project a umadugadu)

	want to collection other information

	want to empower local chief to monitor their own health

	problems: what the best way to present statistical data to local uneducated person

	let local people draw boundaries to their village

	enpower locals to become partners in management of their own health


Carol:

	Can't use name as concept because of cultural requirement (change name when someone dies)

	Noted the difference between people needs and tech needs

Cheryl:

	Mapped homes of HIV patients

	Worked with government and provided information to them

	Had good responses from local leaders

	People were exited about satellites

		Don't assume people won't be exited about learning about tech and involving comminuties

	Good bridge into community because you can show them maps etc.

?:

	Noted Rwanda has hierarchy of regional structures

	problem: combining h1n1 data from different places in the states

	data is reported differently

	noted that something about location that is not very satifying

	have developed nice ontology