U.S. National Library of Medicine's “Haiti Earthquake People Locator”

Descriptive Material

Feb 2 Posting to Sahana Developers

As has been described here for a while (i.e., link from Haiti start page to original project concept), we have been rapid-prototyping a version of our Sahana-based “Lost Person Finder” system for the Haitian disaster. The new system, after expedited technical and legal work, is going live today at http://hepl.nlm.nih.gov/ as a “Haiti Earthquake People Locator”. In the current state, you can think of it as a public viewer for Google records (using our Notification Wall but made interactive), with some filters beyond name, and supplementary iPhone- or email-based input methods (with forwarding to Google). It was originally conceived as an aid for U.S.-based relatives of Haitians, but could well evolve to support specific field operations. It also contains a link to NLM-curated Haiti disaster resources.

A formal announcement will be out shortly, which will be promulgated more widely and posted on the Sahana wiki [see next item].

Beyond the borders of our project, thanks in particular to Nilushan Silva for spearheading the PHP PFIF code. - Glenn

Draft Long-Form Announcement of Feb 4 - the day the iPhone app became available.

National Library of Medicine Creates Haiti Earthquake People Locator

Interactive Web Site That Helps Reunite Family Members and Loved Ones

Imagine a group of worried relatives, in Florida, New York and various corners of Haiti, trying to find out whether their beloved grandmother in Port au Prince has been found and, if so, her condition. They can submit her photograph (and her name, age and other basics) electronically to a new database within seconds, in hopes that someone has news of her whereabouts and status.

Or imagine a doctor in Haiti treating a teenager for minor injuries in an area outside Port au Prince. The teen is expected to recover fully, but his home has been destroyed and he has lost several relatives in the earthquake. He has family in the Washington, DC area and in other parts of Haiti who are eager to find out where he is and how he is. (He does not own a cell phone.) The physician can take his photo with her digital camera, make notes on his name, age group and condition, and submit his photo and accompanying information to a new Web site and his family and friends can find out his status from any place, using a computer.

These scenarios aren’t the stuff of imagination. The National Library of Medicine (NLM), the world’s largest medical library and an arm of the National Institutes of Health (NIH), announces the release of the Haiti Earthquake People Locator (HEPL), a new system that can perform these functions and more, to assist in the reunification of family members and loved ones during the ongoing relief efforts in Haiti. HEPL consists of an interactive Web site that provides information about people who have been found in Haiti or who are still missing after the devastating earthquake of January 12, 2010. The site allows members of the public to search for people who have been located by medical staff and other relief workers in Haiti as well as to voluntarily post information about people who are still missing. Medical and relief personnel in the region can submit photographs and descriptive information about found or located people via computer, cell phone, or a specialized “Found in Haiti” iPhone application that was developed by the NLM. Using data from this app the site can even show the GPS coordinates of the spot where he was found. The site also has a meta-search engine to simultaneously search multiple sites setup by CNN, Google and the International Red Cross.

The HEPL system, developed by NLM’s Lister Hill National Center for Biomedical Communications, shares information with other person finder systems, including the one recommended by the US Department of State (Google’s Person Finder), to ensure that all searches operate across the largest possible set of matches. Based in part on the open-source Sahana disaster management system, HEPL’s technology was developed by the NLM as part of its R&D contribution to the Bethesda Hospitals Emergency Preparedness Partnership (BHEPP), a collaboration among the NLM, the NIH Clinical Center, the National Naval Medical Center and Suburban Hospital/Johns Hopkins Medicine, to improve emergency response capabilities in the National Capital Region, and also serve as a model for the nation.

Final Short-Form Announcement, Saturday, Feb 6, 2010

Main Features and Responsibilities (Feb 3)

The project is run by the Communications Engineering Branch (CEB), Lister Hill National Center for Biomedical Communications, NLM.

It was decided to attempt rapid prototyping of a customization of our software to assist to the Haitian earthquake disaster, including French and Haitian Creole translations. As background, the elements were:

  • Reporting of Missing/Found People by Email. This generalizes existing LPF parsing capabilities to less-structured language text. An aid worker or family member or self-reporter would enter minimal information on the subject line, namely, name, status information (missing, alive & well, injured, deceased), and auxillary information such as location. One (or in theory more) pictures could be attached. The result would be sent off to the HEPL web site, where it would be parsed and made accessible. G. Miernicki’s efforts were key here; G. Pearson arrange for multilingual keyword match strings.
  • iPhone Reporting Tool. As an example of one technology supportive of that approach, a iPhone-app-based version of TriagePic was created by B. Senevirathna and S. Antani. This would prompt for the desired information, and allow taking a photo with the cell phone’s camera.
  • Interactive Notification Wall. To assist relatives of Haitians, particularly those living outside Haiti, G. Miernicki built a version of LPF’s Notification Wall that has itself interactive-search capability. This would have some of the same purposes as LPF/Sahana “Search”, but when running will continuously display updated photos, name, and status about people who match user-specified text-search criteria. Thus, a relative (or aid worker) could leave this running continuously on a laptop. It became the main page of the “hepl“ site.
  • Data Gathering from Other Haitian Disaster Sites. We will attempt (ideally almost continuous) harvesting of photos and text from certain other appropriate missing-person sites. We will also strive to facilitate data sharing with other Haitian-response Sahana sites. As an early effort, NLM/CogSB's A. Kanduru provided screenscraping of CNN web sites. C. Cornwell and G. Miernicki, with Sahana programmer N. Silva, have been developing PFIF 1.1 and 1.2 interfaces to Google.
  • A meta-search engine, covering CNN, Google, and the International Red Cross, was spearheaded by NLM/CogSB's A. Bangalore.

G. Thoma, as CEB Branch Chief (and CogSB Acting Branch Chief), instituted and oversaw special project group meetings, and spearheaded the effort to address the managerial, procedural, and legal issues needed to bring the site public in an expedited way.

Within the NLM Office of the Director, D. Sharlip reviewed critical privacy regulations, and shepherded clearance requests with OMB. J. Sheehan offered critical policy review and coordination with NIH Project Management. T. Danielson provided administrative management, with policy review and coordination. NLM Deputy Director B. Humphreys presided over this effort.

This system was shown to the NLM Board of Regeants and, after NLM, NIH, and OMB approval, made public Feb 2.

Details

  • As outlined in our initial plan, this system does not expose the reporting method of the Lost Person Finder interface, but instead directs reporting to agglomerating sites. By a week after the earthquake, Google had become the consensus site for such agglomeration.
  • G. Pearson authored English text for four child pages. This was drafted and installed, with input from the LPF team, DIMRC (particularly C. Love), and legal-issue specialists. (Finalization continues throughout February). Translations of the main and child pages into Haitian Creole and French were organized. Main page translations were initially done during G. Pearson's visit to DC Crisis Camp by volunteers C. Bonnefil and L. Castera. Child page and other subsequent efforts prior to launch are by NCBI’s A. Neveol and NLM/OCCS’s D. Ordelus.
  • CEB's M. Chung created the HEPL and related logos.
  • CEB system administrators put up the site's public web server and database. L. Neve led this effort. M. Bopf helped with DNS and server issues. B. Neeriemer assisted regarding DNS, security, and privacy.
  • The HEPL Additional Resources page links to Haiti-specific material curated by NLM's Disaster Management Information Research Center, by C. Love and colleagues.
  • A screensaver called WebSavr, previously built by G. Pearson to support MARS, was considered for the Haitian effort, as a frame for hepl. It uses an IE 7 web browser object. G. Miernicki has just made hepl IE 7-capable. Problems remain, so WebSavr deployment is suppressed.
  • Possibilities for interacting with Twitter and photo-sharing sites like Flickr were explored by NLM/CSB's D. Bennett and S. Ward. Ideas for parsing approaches were considered by NLM/CogSB's G. Divito and A. Browne.
  • CEB's G. Thoma and DIMRC's S. Phillip presented the HEPL project to the Board of Regeants within the context of BHEPP projects, and paired with presentations of other BHEPP projects about Digital Paper (S. Liu, W. Ma) and Prescription Retrieval (K. Fung, C. MacDonald). M. Gill and S. Antani presented HEPL demos to Board subcommittees.
  • M. Gill took the lead in developing BoR slides, overview material for NIH management, and the formal announcement.
  • Cross-links from intranet and public NLM sites, and press notification, were coordinated by Office of the Director's K. Steely, K. Cravedi, and M. Modlin. Other team members also got the word out.

The broader Sahana community has put up instances of Sahana, and conferred with Haitian and NGO contacts about possible uses. We have assisted that effort (in addition to the PFIF work discussed above) by structuring data about Haitian populated-place names for potential import into Sahana or HEPL. G. Pearson retrieved Haitian place-name data from www.geonames.gov and put it into various derivative forms. This data source is not ideal, but a start. The derivation forms were in possible/speculative support for:

  • Reporting of the “origin” of missing people
  • Having an ontology of placenames and geolocations to help parse incoming emails.

This was done (with some manual processing) for the central and southern states of Haiti, where quake damage occurred. Derived files were posted to Google and made publicly available through the Sahana wiki.

Early PFIF 1.1 Mapping to Sahana/LPF

A Word document describing the initial mapping as of mid-February is here. (This document also summarizes the new fields introduced for PFIF 1.2) Work is on-going (in HEPL and CEPL contexts) on PFIF 1.2 import and export, and corresponding PHP classes, new database tables, and mappings to existing and new database entities.


Navigation
QR Code
QR Code haiti:hepl (generated for current page)