Is death the best measure, or the easiest?

Few things are more energizing than teaching a bright group of young people.

Last Sunday I led a session on “natural hazards and health: application of the problem solving method”.  Frankly all I know on natural hazards I learned while writing the paper for the Lancet.  So I was a bit nervous about getting up in front of a group of students and teaching them about the subject.  What if they had questions?!

But what I’ve realized, having taught a fair numbers of times now, is that the instructor has an incredible amount of power in directing the class in various directions.  Most classrooms have a “teacher as expert” feel, where knowledge and information flows one way.  Students are recipients of knowledge.  I prefer a more egalitarian classroom, where the goal is to draw on the collective knowledge and analysis of the room, and enable a discussion of critical thinking.  I’m most interested in showing the students that in 99% of cases, there is no easy “right” answer or decision.  Rather one is always faced with imperfect information, resource constraints, and trade-offs.

Case studies are a great tool because they let the students get into the shoes of various characters in a complex environment.  Students can analyze the situations and decisions, and through the help of a good teacher, see the pros and cons of multiple strategies.  I first began to appreciate this when I saw Michael Porter teaching at the business school.  At first he’d let the class paint a picture of why one option was the obvious right choice.  Then he’d let us convince him that in fact it was incredibly stupid.  And we’d do this for a few different strategies.  Students walked away not convinced of an answer, but stimulated in terms of understanding the complexities of decision making.

With this philosophy in mind, I approached the class with a few different questions—what are the major health concerns of floods and cyclones in Bangladesh?  We listed off a whole list.  Then we talked about how to measure the “success” of disaster management.  Bangladesh has seen a significant decline in deaths from natural hazards over the past 43 years, and therefore receives great international praise for its innovations in preparedness and response.  But what the students and I focused on was the “system”—how you measure “success” incentivizes certain behaviors.  For example, many of the consequences of floods are long term—disruption of economic livelihoods (jobs destroyed, crop cycles missed), schools out of session, the malnutrition and poverty that sets in more slowly—it won’t kill you immediately, but has huge consequences for long-term well being).  But because death is the measure that the media, global donors, and policy makers care most about, there’s little motivation or support for addressing the longer-term challenges.

Many have begun to talk about resilience as a better way of understanding natural hazards.  My former boss, Sandro Galea, gives a nice talk about this.  After the immediate aftermath of an event, are people: a.) worse off than they were before the event, b.) the same as they were before, or c.) better off than they were before?  Obviously b. or c. are preferable to a.  The students saw this immediately and had lots of good suggestions for metrics—comparing school enrollment, average income, number of unemployed, land productivity, etc from before the incident and afterwards (perhaps at 3, 6, and 12 months)

Of course the biggest problem with resilience measures is that they assume that the pre-hazard situation is well known.  Without knowing the school enrollment, unemployment numbers, and so forth before the event, there is no comparison point for what happens afterwards.  Developing the surveillance systems that track these numbers robustly is a prerequisite for moving towards better measures of disaster management.  After the Bhola cyclone in 1970, so little was known about the affected population (for example, the total population before the event), that death estimates ranged from 250,000-500,000.  Major improvements have been made, but even in events of the past decade, there have been many issues with determining post-disaster diarrheal disease outbreaks, because the normal incidence is not known.  Mortality is the “easy” thing to measure, but has significant limitations from a development and poverty reduction standpoint.

On the bright side: the new Bangladesh Bureau of Statistics site (thanks Manik for showing it to me!) is a great start at creating a baseline and making data available for disaster management and many other types of planning and evaluation.

This week I'll be teaching about tuberculosis in urban Bangladesh and the variety of control strategies employed by BRAC and others.