[Some of these posts will include topics that I'm working on. For such topics, I'd really like feedback. Now I know you don't think you have anything to contribute, but that's a lie. I want these to be understandable by everyone, both in vocabulary and concepts. So if there's something you don't understand please point it out. I have no other way to gauge how crazy my writing is becoming. So with that introduction, here is my first Blogged program idea.]
One of the problems in healthcare is that it is slow. If we ever want to find something out, we commission a study which takes a year or so. If we want to change something, it's like turning an aircraft carrier (which, I've been told, is difficult). In the US, it's a system which is quite disconnected and has little incentive to be fast or dynamic (but that's a whole different discussion). But, in starting from a very basic system, is there a way to speed things up?
The first thing that needs speeding is disease information. Collection has been so slow up to now, that only strategic information could be used. How many million people are infected with HIV? Is hypertension on the rise this decade? And when I say strategic, it's like counting the number of men in your army compared to your enemy's. Keeping to the military analogy, tactical is talking about things like how to defend a particular hill; we do no generally use tactical information in modern healthcare.
But in Kenya, disease moves much too fast. Three of the top culprits here are Respiratory Tract Infections, Malaria, and diarrhea. These three conditions don't even last a week (if you're lucky), so no existing system is nearly fast enough. What do we do?
Well it's actually quite easy (if you're Nuru). We ask people about the sickness in their family that week. And since they meet once per week anyways in their Nuru groups, it won't be that hard. And once we have this information, we can begin to do things like targeted interventions with our Nuru Healthcare Representatives (who we hope will become Community Health Workers; more on the details of this later).
Say there is a diarrhea outbreak. Say 10 people in Area A get diarrhea on Monday and Tuesday. On Wednesday, they would then report to the Health Rep during their Nuru meeting. So on Wednesday afternoon, we can equip the Health Reps with anti-diarrheal drugs to take to Area A on Thursday, and with soap to get to Area B next door to prevent diarrhea from spreading there. And it could be the same story with Malaria and respiratory infections.
Today it's diarrhea, but it could ultimately include many more things. To make it fast and efficient, we may make it phone-based (we have internet phones out here) and use online servers to process the data (i.e. a farmer could send a text message to find out what his Malaria risk was that week). We're planning to get GPS on all our farmers, so we could produce a map of our results, helping us to predict outbreaks. We've considered doing regular weighings which could help identify malnutrition, HIV, and TB. We could integrate screenings for other common diseases as well. We've even considered expanding the Disease Intelligence Network to the Poverty Intelligence Network and including all the information we have (farming yields, pests and weed growth, etc), making it a useful set of data for all programs.
The most important part of the Disease Intelligence Network is this: Information->Response. This is what we care about. There are about a million applications to this basic idea, but the unique thing about this is that we have connected information directly with a response, and both are fast enough to make a difference; the thing that makes DIN special is that it is tactical. And, God willing, tactical will mean lives saved.
The question of how do you get information out of the field quickly takes on several aspects. The more technology you use, typically the more expensive it is. The more manual and labor intensive you are the less expensive but typically the slower it is to receive information.
ReplyDeleteFor example, there are inferred video cameras at some international airports which help determine peoples’ body temperatures. If a person is running a temperature officials can funnel these people into a separate room to ask additional information; trying to determine if they have swine flu. Because this technology is completely automated they can monitor thousands of people a day as they file through the airport lines. This process is much faster than inserting a thermometer into the month of these people.
If these same cameras were to be positioned throughout a city where a good sampling of that populous were to frequent you would have the ability to determine the possibility of certain types of diseases based on knowing a percent of these people had a temperature. There may be other analytics that detected sweating, or pox marks, boils, or staggering. If the populous had a card swipe or active ID to let the system know who had this problem, immediate tracking could be accomplished maybe even before this person knew he had a problem. This information would be real time. However, because of the cost of this technology this may be cost prohibitive.
Let us now consider a combination of some technology and some labor intensive process and see if we can come up with something that may be doable. This time we use standard digital video cameras (or turn styles similar to those found at train stations) throughout the city. The populous can be educated to take their temperature manually when they are not feeling good. There may be other symptoms like diarrhea, or vomiting, or other physical information important to making diagnosis that you would want to collect. This time the village is trained to go through line 1 through 8 (you have 8 separate lines or turn styles to indicate the information you are collecting). Line 1 lets you know that person has a temperature, 2 indicates vomiting, 3 diarrhea, etc. This person may have to go through 2 or 3 lines if more than one symptom were present. There are analytics in the software that can count the number of people going through these lines. You now know real time how many people in each part of the city had certain symptoms.
Another possibility is using a questionnaire. If people are willing to walk through a line when they are sick, they may be willing to fill out a form with their name and address. Similar to the older “Scrantron” form we use to use to take test, the same centralized location we used to video tape the population, could have collection points where individuals scanned this information into a wireless or networked computer. This form may be more accurate because only one form could have the entire family information on it. 5 people had a temperature. Now you have even the address of this person so immediate plotting where this family lived and see how it compares with other issues in that city.
Self reporting is a great idea. Surveillence becomes a lot easier when ever set of eyes (or body temperatures) is recruited to help you. The technology which may be perfect for this is the cell phone; we can push forms out to these phones and have it automatically compile the data. Penetration is something like 40%, so it's not quite there yet.
ReplyDeleteThe population density is low, so getting people to go through a single location (or even set of locations) would be quite difficult.
Great ideas! I think this concept could really be pushed in other areas which are urbanized. I bet US public health and CDC officials would love to hear this pitch.
David