Public Participation Health GIS: An Idea Who’s Time Has Come?

1. Public Participation Health GIS: An Idea Who’s Time Has Come? Russell S. Kirby, PhD, MS, FACE Dept of Maternal and Child Health School of Public Health University of Alabama at Birmingham

2. Objectives • Consider strategies for taking web-based access to mappable health data to the next level: public participation health GIS • Describe typologies for serving maps and atlases on the Internet • Identify methodological, administrative and political issues that must be overcome • Have some fun – with humor and grace!

3. From my former state: Wisconsin

5. And my current state: Alabama

7. Brief Summary for Those Who Are Knitting, Doing Crossword Puzzles, or Discerning the Geometric Pattern in the Carpeting • No vendor (hardware or software) or modeling approach has the hands-down best solution for serving public health maps on the Internet (and probably never will). • Must current applications fail to provide appropriate levels of end-user flexibility and functionality. • The technologies for true PPHGIS public health mapping applications exist – do we have the science and political will to make it happen?

8. THE TEN COMMANDMENTS OF PUBLIC HEALTH GIS With apologies to Mel Brooks, and thanks for editorial assistance to Elizabeth Kirby and for their insights to the following Internet contributors: Dabo Brantley, DRH, CDC Virginia Lee, ASTDR Dona Schneider, Rutgers University Russel Rickard, Colorado Department of Health and Environment Dianne Enright, NC State Department of Health Statistics Ravi Sharma, University of Pittsburgh R.S. Kirby, December 2005

9. The Ten Commandments of Public Health GIS Number 10 Thou shalt not expect thine health outcomes or disease states to respect administrative (block, census tract, ZIP Code, municipal) boundaries. But without collecting geography in thine data, thou hast nothing that can be mapped.

10. The Ten Commandments of Public Health GIS Number 9 Thou shalt not unknowingly commit spatial errors.

11. The Ten Commandments of Public Health GIS Number 8 Know thy purpose (in creating and using your public health GIS). Corollary: Thou shalt always be cognizant that the Scientific Method is not a built-in feature of any GIS software application.

12. The Ten Commandments of Public Health GIS Number 7 Thou shalt know and understand thine data* prior to bringing it into a GIS. *Editor’s note: this is a more specific example of the recently rediscovered 16th commandment, “Know thy data”

13. The Ten Commandments of Public Health GIS Number 6 Thou shalt remember that while thine maps are abstractions, the maps reflect the physical environment and are based on data representing actual events that occurred to individual people.

14. The Ten Commandments of Public Health GIS Number 5 Thou shalt protect individual records containing X,Y coordinates as thou wouldst protect records with individual identifiers, as both can reveal confidential information.

15. The Ten Commandments of Public Health GIS Number 4 Thou shalt not clutter thine health data maps with unnecessary layers and map elements (i.e. chartjunk), nor shalt thou ignore information necessary to interpret the patterns on your map.

16. The Ten Commandments of Public Health GIS Number 4 - addendum “The real art of cartography is knowing more what to leave out, more than what to put in.” - John Parsons

17. The Ten Commandments of Public Health GIS Number 3 Know thine Metadata.

18. The Ten Commandments of Public Health GIS Number 2 Thou shalt not assume that the default settings of your GIS software will generate useful and meaningful maps.

19. The Ten Commandments of Public Health GIS Number 1 Thou shalt show humility to others, and be gracious even unto those who thought it would take weeks to accomplish what thou hast done in a few hours.

20. Are there other Top Ten lists? • Yes – just point your browser to http://www.soph.uab.edu/kirbytop10 • Would you like to contribute to or suggest a topic for a future list? • Email me at rkirby@uab.edu • Current lists under development include: • The Ten Best Ways to Use Email Badly • The Ten Best Ways to Misuse Hospital Discharge Data, Part 2 • The long term plan? • Perhaps, a book with the working title: ‘The Practical Guide to Conducting Bad Health Research: A Book of Lists’

21. Topology of Data-Atlas-User Interfaces • Web interface as method of transmission only • Web interface for user to access pre-designed maps • Web interface for user to access database for generation of customized maps and tables • Web interface as point-of-service for integration of data from multiple geospatial data sources

22. Web Interface for Transmission Only • Utilize Web transactions to distribute static images and HTML files • User requests a map by submitting a request (clicking a button, entering a URL, etc) • GIS server prepares the images requested, sends to the Web server which then transmits to the user • User then accesses and manipulates the map with local GIS or mapping software • This method does not provide PPGIS, requires end-user to have knowledge of mapping software, and is limited in flexibility to customize the final map

23. Web Interface for User to Access Pre-Designed Maps • By far the most common current method for serving maps and atlases with health data currently • User connects to map or atlas website • User selects the desired map from a list, the map is then displayed • Some applications provide ‘customized’ variations, but in reality these are already prepared and stored for the user to access when selected • While this method provides end-user access to map content, the end-user has limited options and often cannot access or format the information as desired • Most of these applications are not PPGIS

24. Web Interface for User to Access Pre-Designed Maps (continued) • Many health examples: • http://www.cdc.gov/cvh/womensatlas/index.htm • http://www.cdc.gov/cvh/mensatlas/index.htm • http://www.cdc.gov/cvh/maps/strokeatlas/index.htm • http://www3.cancer.gov/atlas/ • http://www.dartmouthatlas.org/default.php • This is a growing list!

25. Web Interface for User to Access Database for Generation of Customized Maps and Tables • The application webpage acts as a portal for the user to define a request for a map or table • The web server posts this request to the GIS or database server, which formats the request and generates the desired results • The results are then sent to the user’s screen • Applications using this methodology vary widely in their design and flexibility

26. Web Interface for User to Access Database for Generation of Customized Maps and Tables(continued) • Some early applications (e.g. HIT the SPOT) gave the user broad ability to define data ranges, map format, access to a wide array of statistical variables • http://hitspot.utk.edu/hit/main/SPOT/frames/SPOT/index.htm • Others incorporate various decision rules that limit access to results under specified conditions • Most of these applications generate univariate or bivariate maps, but typically from a single data source

27. Web Interface for Point-of-Service for Integration of Data from Multiple Geospatial Data Sources • This may be the future of serving health data maps on the web. • Multiple spatially enabled public health databases would be simultaneously queried from a single end-user request, with the results transmitted to the desktop for further use. • Most local and state public health agencies are far from able to contemplate this type of application, although there are numerous corporate examples.

28. Web Interface for Point-of-Service for Integration of Data from Multiple Geospatial Data Sources (continued) • An early example is the Milwaukee COMPASS site: • http://www.milwaukee.gov/compass • Let me know if you have similar applications in your communities or states.

29. Public Participation GIS • Not a new concept – in the literature and GIS community since mid-1990s! But not part of the lexicon of public health or community assessment. • Generally PPGIS focuses on “the use and value of GIS by marginalized peoples and communities engaged in social change” (Sieber 2003). Sieber, R.E. (2003), Public participation geographic information systems across borders.The Canadian Geographer 47 (1), 50-61.

30. Applying PPGIS to Health Issues • The objectives of the community assessment initiative fit squarely within the framework of PPGIS. • However, to create truly PP-H-GIS will require transformational thinking beyond most current applications for serving maps and atlases on the web. • Some issues are the same:

31. Hardware-Software Issues • In developing map/atlas applications consideration must be given to: • Statistical server • GIS server • Web interface • GIS software • Decisions include: • Volume of traffic, number of users, level of interactive use

32. Hardware-Software Models • There are many vendors for both hardware and software. None is sufficiently better than all others to warrant its mentioning specifically. • A well-designed application may separate the database and GIS functions, and utilize a web server as the front end. Software might include a statistical package, a high-end RDBMS, any of several GIS packages, and perhaps some specialized spatial statistical applications. • The most important issue in designing an application to serve public health maps on the Internet is the capacity to meet all functions in a time-efficient manner.

33. Transformational Issues • Public participation implies freedom of choice as to what is important. • This means users need to have the capacity to initiate their own inquiries – and the PPHGIS must be designed to support this functionality. • End users shouldn’t feel constrained by the limitations of databases – the PPHGIS must be designed to handle obvious problems in background, things like: • Data elements of interest residing in two or more different databases • Aggregation of data into mappable units (by year, spatial unit, etc) • Records not linked at individual, household, ecological scales

35. Problems and Pitfalls Provincialism of program managers • Fear of unfamiliar cartographic methods • Concerns with confidentiality • Leads to over-restrictive decision rules for data release • Under-powered applications, particularly for background processing of data on demand • Fear of what might happen when the ‘public’ can see ‘our’ data • Data quality issues – there are no warts until someone sees them. But . . . The warts don’t heal until they are treated! (Beware of antibiotic resistant strains . . .) • Will current data integration initiatives evolve to the point where they can support the high-end PPHGIS?

36. Problems and Pitfalls • Observation: • None of these issues are new or novel observations (Kirby RS, Stat Med 1996) • What remains lacking is the broad vision of key leaders in public health and health data organizations to find the ways to overcome these obstacle

37. And finally, • Will the thinking of public health professionals evolve to the point that health data is considered within the framework of broadly based determinants of population health?

38. Health and Function CONCEPTUAL FRAMEWORK FOR POPULATION HEALTH Social Environment Physical Environment Genetic Endowment • Individual • Response • Behavior • Biology Health Care Disease Well-Being Prosperity Traditional Medical Model of Health Care Source: modified from Evans RG, Barer ML, Marmor TR, Eds, Why are some people healthy and others not? New York: Aldine de Gruyter, 1994

39. Ten Best Ways to Develop a Bad Public Participation Health GIS • Now it’s your turn: what should be included on this list?

41. Questions, thoughts or reflections after the conference? rkirby@uab.edu 205-934-2985