• The challenge of cultural assumptions embedded in standard health surveys (what is "jogging", exactly? how do you evaluate the health significance of "one glass of rice wine per day" in China vs. Japan?). Betsy Rolland of the Fred Hutchinson Cancer Research Center's Asia Cohort Consortium, a group tasked with wrangling and consolidating data from half a dozen Asian countries, is wrestling with these issues every day in her work.


• The impact of donor expectations on data collection. Health systems in developing world countries will, understandably enough, do whatever it takes to help donors determine the outcomes of their giving--but often doesn't translate into systematic, consistent data collection that best meets the needs of the entire system.


• The issues are not limited to the developing world. Finland has an excellent health data system, but many developed-world countries struggle with health data too--even in Seattle, a leader in advanced medical research and practices, getting good data from the University of Washington Medical Center to the Department of Public Health, just a few blocks away, is a non-trivial task.

• Jason Parker of the Gates Foundation told us about techniques Google is using to overcome issues of accuracy by dramatically increasing the quantity of data being analyzed. (We agreed that getting representation from commercial organizations like Google and Amazon at our Global Health Data meetings could well provide useful insights into the problems we wrestle with.)


• Sherrilynne Fuller of the UW's Center for Public Health Informatics observed that a typical attitude to these daunting data infrastructure problems is "let's deal with that tomorrow and do fun stuff now." To encourage investigation at the level of *system* rather than the isolated individual issues that get funders' attention, she proposes building a simplified conceptual model for health data that will help students, funders, and researchers--and the general public, too--gain greater insight into systemic data collection issues and solutions.


• David Lubinski of PATH, fresh from a series of recent trips to Africa and bringing a wealth of insights (for example, if a local clinic has drugs, the word spreads like wildfire and people come; no drugs, no people), is working on an exciting new health data management project that takes a "less is more" approach: identify a few key significant data points (for example, asking a clinic "how many patients were admitted?", "what were they admitted for?", "who's paying for their treatment?") and develop an end-to-end system that manages those key points all the way from collection through regional, national, and international aggregation to analysis.

This would massively improve what bibliophiles call book discovery. You're far more likely to hear about a book if a friend has highlighted a couple brilliant sentences in a Facebook update—and if you hear about it, you're far more likely to buy it in print. Yes, in print: The few authors who have experimented with giving away digital copies (mostly in sci-fi) have found that they end up selling more print copies, because their books are discovered by more people.

• Foremost for me, the result of the two-hour live chat with about 30 people was two or three new insights to challenge and refine my previous thinking -- which is great. Putting a substantive comment out to a bunch of interested people and getting their immediate feedback is provocative and useful. The fact that every Twitter comment is limited to 140 characters makes everyone cut to the main point quickly, which is nice. And because the conversation is fast moving, you don't really have to respond to comments that are less engaging!
• 
  The discussion quickly became multi-channeled in substance, but single-channeled on Twitter -- so you had to follow pretty closely to who was responding to which previous comment. I find the transcript less interesting than the real-time discussion was, when I was intently watching the comments and responding to those that caught my interest. With more people participating this might have gotten really confusing, though the moderator helped by marching us through a sequential set of discussion questions.
• 
  It was difficult to respond quickly enough -- periodically, by the time I formulated my 140 character response, the conversation had moved on. But unlike in-person and many other events, you don't have to worry about interrupting anyone else since it is multiple conversations in parallel, you do get to keep on expressing your line of thinking even while others go off on different tracks!
• 
  Some interesting questions and comments have continued to come in to Twitter after the formal discussion, so the Twitter format offers a way for people to continue the conversation.
• 
  The Twitter discussion was substantive, and generated interest in our work amongst participants. I also gained about 10 more personal Twitter followers, so far, and others seem to be finding the discussion stream. I have not found any blog or other coverage of the this discussion.
• 
  Technology: tricky despite my preparation. Twitter's search service, which many people use to stream the discussion, hiccupped at the exact time of the event. It all came back online a few minutes later. Nobody seemed to mind, but we don't know if we lost participants in the process. This highlights the ongoing tension between using 3rd-party web services which are often innovative and full featured and free (or nearly so), but out of your control, versus proprietary or your own tools which you might control more but innovate less and pay a bunch more money for.

• I am preparing three main talking points to anchor my comments, as I do with all press events.


• I wonder if this Twitter discussion will get picked up or elaborated by bloggers or others.


• I am spending more time than I anticipated tinkering to make sure I understand how the technology works, so I do not make a mistake in the middle of the discussions. Am learning a lot in the process.


• I am concerned (from ignorance!) about how my steam of Tweets on Wednesday will impact my other Twitter followers who aren't part of the discussion, since they'll see my comments coming through but out of context.


• I wonder who else will start following me on Twitter as a result – each time I Tweet about a new topic, additional people start following my stream.

Next Wednesday, I'm going to be speaking on a panel with Shannon and two other experts (Kae Dakin of Kae Dakin Consulting and Susan Blake of GRC Direct) on both online and offline fundraising strategies and tactics for global health organizations as part of the Global Health Council's annual Conference.  We're planning for a lively discussion and solid audience participation, so I'll likely come away with some other useful tips to share.