AMPATH's Home-Based Counseling and Testing Project (insert catchier title here)
76 views
Skip to first unread message
Zeshan Rajput
Dec 13, 2010, 12:21:32 PM12/13/10
to ict...@googlegroups.com
So I was challenged to come up with a funnier, yet informative post than our previous instruction on how to avoid Dark Wizards when conducting research. I'll try, but I won't be offended if you play a game of Angry Birds before coming back to this.
Here's an outline. Feel free to skip to what interests you - I'll be putting a lot in given the exceptional breadth of expertise on this list.
I. Personal Introduction
II. Why HCT?
III. So what is HCT?
IV. PDAs, tethered GPS devices, and Pendragon Forms
V. Android and ODK
VI. The (preliminary) Results
VII. Thoughts on what we have
VIII. Thoughts on where we're going
I. Personal Introduction
I'm Zeshan Rajput. Among the 10,000 hats I wear, the relevant ones here are that I'm a practicing internist and professor at Indiana University, a 2nd year Informatics fellow at the Regenstrief Institute, and "that guy" who's been trying to get OpenMRS to directly support asynchronous communication to patients and providers using methods such as email, twitter, and SMS. I'll be speaking about AMPATH's current incarnation of its Home-based Counseling and Testing Project (HCT) here, but feel free to ping me on and off the list if I can be of service in other capacities. I'm also around at some of the OpenMRS events, so just come up and introduce yourself!
II. Why HCT?
The short version of the story here, as I think you all may know it far better than I. 1.4 million Kenyans infected with HIV, and 80% aren't aware that they are infected [1,2]. Seventy percent of those suffering with HIV live in rural areas [1]. As previously stated, a supply and a demand problem here. There may not be a good supply of effective, high quality healthcare and people may not be aware, able, or willing to utilise resources [3]. The same goes for diabetes, hypertension, chronic cardiovascular disease, and tuberculosis [4,5].
There were several attempts to tackle the problem of undiagnosed or untreated HIV in Kenya before. Regional clinics and mobile testing areas did not adequately address the issue despite efforts to advertise their presence - you just can't wait for people to present for screening. The USAID-Academic Model Providing Access to Healthcare (AMPATH) Partnership in Eldoret [6] elected to choose a proactive solution - the surveillance of over two million people for HIV and other diseases in their own homes.
III. So what is HCT?
HCT is a community-surveillance program entrenched in those communities. We train people in those communities in the surveillance program, which includes HIV counselling and administering the test. It also includes counselling on mosquito netting, pre-natal healthcare, and other topics. See [7] for a more complete treatment.
IV. PDAs, tethered GPS devices, and Pendragon Forms
Our most recent incarnation utilized PDAs tethered to GPS devices. We recently presented results of using this system at MedInfo [7]. In summary, community health workers found the system faster, easier to use, and produced higher-quality data than a pen and paper system would. This is comparable to previous studies showing PDA-based data collection have less errors, be more complete, and require less cleaning at comparable costs [8]. The issues here were that we had a tethered device, so counselors had to maintain two devices plus cabling at all times. Also, with the pair of devices as well as licensing for Pendragon hardware cost us around $573 per counselor in the field.
V. Android and ODK
I suspect that, by now, y'all have heard enough on ODK and its capabilities. If not, I think Katrin just posted this link to the list from IEEE [9] which I'd highly recommend. We used HTC Dreams (a.k.a. T-Mobile's G1 in the US) as our device. Advantages - we were able to negotiate for refurbished G1s in the $300-400 range, though in retrospect my new prayer is 'may I never have to implement a scaling project on refurbished hardware again'! Also no more licensing fees - though in full disclosure open source is not free and we had an amazingly skilled programmer at our disposal. He had to introduce some customizations into both ODK and the XForms module of OpenMRS, but he was able to get things up and running. Continuing full disclosure, we have pretty significant expertise in OpenMRS. That all being said, alot of this has gotten easier recently, plus all of our modifications are freely available for your perusal [10].
VI. The (preliminary) Results
I say preliminary because we haven't finished surveying all 2 million people yet. We did just finish the Burnt Forest Division in Western Kenya using the Android devices, so I'll present a snapshot of that data. Please note - these numbers have not been officially published yet so please do not refer to them outside of this list until I get that done. Of the 18,850 households in the division, our counsellors were accepted into 96%. This translates into reaching 63,470 of the 65,763 we targeted for survey in this Division (total population - 70,009). Of these persons, 39,952 met criteria for HIV testing (which excludes previously known diagnosis). 99% were counselled regarding the risks and benefits of HIV testing, and 98% agreed to testing. 1.9% were found positive above and beyond what was previously diagnosed in the region. Of the 565 people we newly diagnosed with HIV, we were able to refer 491 for further treatment but only 168 presented for further care. Also of note, we identified 129 children of age <18 months as HIV positive; 34 of these presented for further treatment.
We also identified 993 pregnant women, of whom 45% were not receiving antenatal care and 2.1% were newly diagnosed with HIV. 85% of these women presented for treatment. We also got 46% of the women who weren't getting antenatal care to present for follow-up.
In general, our counsellors were very pleased with the change to the Android platform. Many recommended to continue using the Android devices. I'm still tabulating results from the usability survey we just completed, but the most common themes revolved around it being easy to carry and use, as well as being easier to synchronise with our servers. Problems revolved around high power consumption requiring a lot of charging and the longer time in collecting GPS coordinates. They also noted that the HTC Dreams were more fragile and sometimes lost data due to the application crashing.
VII. Thoughts on what we have
What we have is over 200 counsellors and Android devices in the field. Based on current estimates (or the last ones I saw), they are on schedule to complete surveillance of all 2 million people within the project's three year timeframe. We have multiple, nearly identical installs of OpenMRS where the counsellors are synchronising their devices. And we have a pretty substantial amount of data, though in a dozen mysql and Access databases.
We also have boxes of non-functional G1s (about 20% of the refurbished devices we received had defects that kept us from deploying them. This actually became so severe that we began testing all devices in the United States before sending them to Kenya, leading to about a 2 month delay between when we got a device from our distributors and when we got them into a CHW's hands in Kenya. And I'm not including the devices we've gotten back from the field). And boxes of PDAs and GPS devices that aren't being used right now, because they stored information in Access databases as opposed to OpenMRS' mysql tables.
VIII. Thoughts on where we're going
As a programmer: We're continuing to work with ODK to try and improve the app's stability. We're also working on integrating all the databases we have back into AMPATH's main installation of OpenMRS; this is requiring some pretty sophisticated patient-matching algorithms and hasn't been easy.
As the hardware-procurement guy: We're continuing to investigate new devices with the HCT Dream's end-of-life approaching (very excited for when the IDEOS become available for testing in Kenya!). That reminds me, I need to get my hands on one here. So many devices, so little time...
As a clinician: I'm thrilled to see another 2% of the population diagnosed. I'm also saddened to see that only a quarter or less are coming in for treatment after they find out. Need to figure out more about why these people aren't getting in - if we're lucky, SMS reminders to follow-up may be a substantial benefit and a decently low-hanging fruit once we get data into AMPATH's OpenMRS install and the messaging out of OpenMRS works. If we're not lucky, this is going to require community level resources to help get these people the care they need.
As a public health practitioner: This is probably the closest we'll ever get to true population statistics. It's a monumental undertaking to diagnose an additional 2% of the population (cost estimates for the PDA based solution were roughly $300,000 to survey the 2 million people - see [7] for a breakdown. They may be lower for the Android platform - I'm waiting to get a firm number of how many people our counsellors can see each day). Depending on how you look at it, that could also mean that the previous systems for diagnosing and treating HIV are adequate (5% was the previous population estimate for the area I talked about above). This may be a significant question of resources vs. gain to take up with the Kenyan Ministry of Health when we have more data.
So that's it. Sorry it ended up being a little drier than I had hoped for, but maybe it'll be informative and useful to y'all. As Neal and others know, I try not to be hard to find. Email at zrajput (at) regenstrief (dot) org or on this list. You can also send me a tweet at zrajput.
Have a great week!
-Zeshan
References
1) National AIDS and STI Control Programme, Ministry of Health, Kenya. July 2008. Kenya AIDS
Indicator Survey 2007: Preliminary Report. Nairobi, Kenya. Accessed at http://www.aidskenya.org/public_site/webroot/cache/article/file/Official_KAIS_Report_20091.pdf on 18 Oct 2010.
2) Towards Universal Access: Scaling up Priority HIV / AIDS Interventions in the Health Sector. Geneva: World Health Organization. 2008.
3) O’Donell O. Access to healthcare in developing countries: breaking down demand side barriers. Cad Saude Publica. 2007 Dec;23(12):2820-34.
4) The millenium development goals for health: rising to the challenges. Washington, DC: World Bank 2004.
5) Jones G, Steketee RW, Black RF, Bhutta ZA, Morris SS, Bellagio Child Survival Study G. How many child deaths can we prevent this year? Lancet. 2003 Jul 5;263(9377):65-71.
6) Inui TS, Nyandiko WM, Kimaiyo SN, Frankel RM, Muriuki T, Mamlin JJ, Einterz RM, Sidle JE. AMPATH: living proof that no one has to die from HIV. J Gen Intern Med. 2007;22(12):1745-50.
7) Were MC, Kariuki J, Chepng’eno V, Wandabwa M, Ndege S, Braitstein P, Wachira J, Kimaiyo SN, Mamlin B. Leapfrogging Paper-Based Records Using Handheld Technology: Experience from Western Kenya. Stud Health Technol Inform. 2010;160(Pt 1):525-9.
8) Galliher JM, Stewart TV, Pathak PK, Werner JJ, Dickingson LM, Hickner JM. Data collection outcomes comparing paper forms with PDA forms in an office-based patient survey. Ann Fam Med. 2008;6(2):154-60.
[9] Open Source Data Collection in the Developing World - How the Google Open Data Kit in saving lives in Africa. http://www.computer.org/portal/web/computingnow/1210/theme/computer.
[10] ampath-odk: AMPATH's implementation of Open Data Kit tools. http://code.google.com/p/ampath-odk/.
--
Zeshan A. Rajput, MD
Here's an outline. Feel free to skip to what interests you - I'll be putting a lot in given the exceptional breadth of expertise on this list.
I. Personal Introduction
II. Why HCT?
III. So what is HCT?
IV. PDAs, tethered GPS devices, and Pendragon Forms
V. Android and ODK
VI. The (preliminary) Results
VII. Thoughts on what we have
VIII. Thoughts on where we're going
I. Personal Introduction
I'm Zeshan Rajput. Among the 10,000 hats I wear, the relevant ones here are that I'm a practicing internist and professor at Indiana University, a 2nd year Informatics fellow at the Regenstrief Institute, and "that guy" who's been trying to get OpenMRS to directly support asynchronous communication to patients and providers using methods such as email, twitter, and SMS. I'll be speaking about AMPATH's current incarnation of its Home-based Counseling and Testing Project (HCT) here, but feel free to ping me on and off the list if I can be of service in other capacities. I'm also around at some of the OpenMRS events, so just come up and introduce yourself!
II. Why HCT?
The short version of the story here, as I think you all may know it far better than I. 1.4 million Kenyans infected with HIV, and 80% aren't aware that they are infected [1,2]. Seventy percent of those suffering with HIV live in rural areas [1]. As previously stated, a supply and a demand problem here. There may not be a good supply of effective, high quality healthcare and people may not be aware, able, or willing to utilise resources [3]. The same goes for diabetes, hypertension, chronic cardiovascular disease, and tuberculosis [4,5].
There were several attempts to tackle the problem of undiagnosed or untreated HIV in Kenya before. Regional clinics and mobile testing areas did not adequately address the issue despite efforts to advertise their presence - you just can't wait for people to present for screening. The USAID-Academic Model Providing Access to Healthcare (AMPATH) Partnership in Eldoret [6] elected to choose a proactive solution - the surveillance of over two million people for HIV and other diseases in their own homes.
III. So what is HCT?
HCT is a community-surveillance program entrenched in those communities. We train people in those communities in the surveillance program, which includes HIV counselling and administering the test. It also includes counselling on mosquito netting, pre-natal healthcare, and other topics. See [7] for a more complete treatment.
IV. PDAs, tethered GPS devices, and Pendragon Forms
Our most recent incarnation utilized PDAs tethered to GPS devices. We recently presented results of using this system at MedInfo [7]. In summary, community health workers found the system faster, easier to use, and produced higher-quality data than a pen and paper system would. This is comparable to previous studies showing PDA-based data collection have less errors, be more complete, and require less cleaning at comparable costs [8]. The issues here were that we had a tethered device, so counselors had to maintain two devices plus cabling at all times. Also, with the pair of devices as well as licensing for Pendragon hardware cost us around $573 per counselor in the field.
V. Android and ODK
I suspect that, by now, y'all have heard enough on ODK and its capabilities. If not, I think Katrin just posted this link to the list from IEEE [9] which I'd highly recommend. We used HTC Dreams (a.k.a. T-Mobile's G1 in the US) as our device. Advantages - we were able to negotiate for refurbished G1s in the $300-400 range, though in retrospect my new prayer is 'may I never have to implement a scaling project on refurbished hardware again'! Also no more licensing fees - though in full disclosure open source is not free and we had an amazingly skilled programmer at our disposal. He had to introduce some customizations into both ODK and the XForms module of OpenMRS, but he was able to get things up and running. Continuing full disclosure, we have pretty significant expertise in OpenMRS. That all being said, alot of this has gotten easier recently, plus all of our modifications are freely available for your perusal [10].
VI. The (preliminary) Results
I say preliminary because we haven't finished surveying all 2 million people yet. We did just finish the Burnt Forest Division in Western Kenya using the Android devices, so I'll present a snapshot of that data. Please note - these numbers have not been officially published yet so please do not refer to them outside of this list until I get that done. Of the 18,850 households in the division, our counsellors were accepted into 96%. This translates into reaching 63,470 of the 65,763 we targeted for survey in this Division (total population - 70,009). Of these persons, 39,952 met criteria for HIV testing (which excludes previously known diagnosis). 99% were counselled regarding the risks and benefits of HIV testing, and 98% agreed to testing. 1.9% were found positive above and beyond what was previously diagnosed in the region. Of the 565 people we newly diagnosed with HIV, we were able to refer 491 for further treatment but only 168 presented for further care. Also of note, we identified 129 children of age <18 months as HIV positive; 34 of these presented for further treatment.
We also identified 993 pregnant women, of whom 45% were not receiving antenatal care and 2.1% were newly diagnosed with HIV. 85% of these women presented for treatment. We also got 46% of the women who weren't getting antenatal care to present for follow-up.
In general, our counsellors were very pleased with the change to the Android platform. Many recommended to continue using the Android devices. I'm still tabulating results from the usability survey we just completed, but the most common themes revolved around it being easy to carry and use, as well as being easier to synchronise with our servers. Problems revolved around high power consumption requiring a lot of charging and the longer time in collecting GPS coordinates. They also noted that the HTC Dreams were more fragile and sometimes lost data due to the application crashing.
VII. Thoughts on what we have
What we have is over 200 counsellors and Android devices in the field. Based on current estimates (or the last ones I saw), they are on schedule to complete surveillance of all 2 million people within the project's three year timeframe. We have multiple, nearly identical installs of OpenMRS where the counsellors are synchronising their devices. And we have a pretty substantial amount of data, though in a dozen mysql and Access databases.
We also have boxes of non-functional G1s (about 20% of the refurbished devices we received had defects that kept us from deploying them. This actually became so severe that we began testing all devices in the United States before sending them to Kenya, leading to about a 2 month delay between when we got a device from our distributors and when we got them into a CHW's hands in Kenya. And I'm not including the devices we've gotten back from the field). And boxes of PDAs and GPS devices that aren't being used right now, because they stored information in Access databases as opposed to OpenMRS' mysql tables.
VIII. Thoughts on where we're going
As a programmer: We're continuing to work with ODK to try and improve the app's stability. We're also working on integrating all the databases we have back into AMPATH's main installation of OpenMRS; this is requiring some pretty sophisticated patient-matching algorithms and hasn't been easy.
As the hardware-procurement guy: We're continuing to investigate new devices with the HCT Dream's end-of-life approaching (very excited for when the IDEOS become available for testing in Kenya!). That reminds me, I need to get my hands on one here. So many devices, so little time...
As a clinician: I'm thrilled to see another 2% of the population diagnosed. I'm also saddened to see that only a quarter or less are coming in for treatment after they find out. Need to figure out more about why these people aren't getting in - if we're lucky, SMS reminders to follow-up may be a substantial benefit and a decently low-hanging fruit once we get data into AMPATH's OpenMRS install and the messaging out of OpenMRS works. If we're not lucky, this is going to require community level resources to help get these people the care they need.
As a public health practitioner: This is probably the closest we'll ever get to true population statistics. It's a monumental undertaking to diagnose an additional 2% of the population (cost estimates for the PDA based solution were roughly $300,000 to survey the 2 million people - see [7] for a breakdown. They may be lower for the Android platform - I'm waiting to get a firm number of how many people our counsellors can see each day). Depending on how you look at it, that could also mean that the previous systems for diagnosing and treating HIV are adequate (5% was the previous population estimate for the area I talked about above). This may be a significant question of resources vs. gain to take up with the Kenyan Ministry of Health when we have more data.
So that's it. Sorry it ended up being a little drier than I had hoped for, but maybe it'll be informative and useful to y'all. As Neal and others know, I try not to be hard to find. Email at zrajput (at) regenstrief (dot) org or on this list. You can also send me a tweet at zrajput.
Have a great week!
-Zeshan
References
1) National AIDS and STI Control Programme, Ministry of Health, Kenya. July 2008. Kenya AIDS
Indicator Survey 2007: Preliminary Report. Nairobi, Kenya. Accessed at http://www.aidskenya.org/public_site/webroot/cache/article/file/Official_KAIS_Report_20091.pdf on 18 Oct 2010.
2) Towards Universal Access: Scaling up Priority HIV / AIDS Interventions in the Health Sector. Geneva: World Health Organization. 2008.
3) O’Donell O. Access to healthcare in developing countries: breaking down demand side barriers. Cad Saude Publica. 2007 Dec;23(12):2820-34.
4) The millenium development goals for health: rising to the challenges. Washington, DC: World Bank 2004.
5) Jones G, Steketee RW, Black RF, Bhutta ZA, Morris SS, Bellagio Child Survival Study G. How many child deaths can we prevent this year? Lancet. 2003 Jul 5;263(9377):65-71.
6) Inui TS, Nyandiko WM, Kimaiyo SN, Frankel RM, Muriuki T, Mamlin JJ, Einterz RM, Sidle JE. AMPATH: living proof that no one has to die from HIV. J Gen Intern Med. 2007;22(12):1745-50.
7) Were MC, Kariuki J, Chepng’eno V, Wandabwa M, Ndege S, Braitstein P, Wachira J, Kimaiyo SN, Mamlin B. Leapfrogging Paper-Based Records Using Handheld Technology: Experience from Western Kenya. Stud Health Technol Inform. 2010;160(Pt 1):525-9.
8) Galliher JM, Stewart TV, Pathak PK, Werner JJ, Dickingson LM, Hickner JM. Data collection outcomes comparing paper forms with PDA forms in an office-based patient survey. Ann Fam Med. 2008;6(2):154-60.
[9] Open Source Data Collection in the Developing World - How the Google Open Data Kit in saving lives in Africa. http://www.computer.org/portal/web/computingnow/1210/theme/computer.
[10] ampath-odk: AMPATH's implementation of Open Data Kit tools. http://code.google.com/p/ampath-odk/.
--
Zeshan A. Rajput, MD
Chris Moses
Dec 13, 2010, 1:23:34 PM12/13/10
to ict...@googlegroups.com, zra...@regenstrief.org
Hi Zeshan,
Great post! And thanks for including references. Coming from another Android-based team, I would love to hear about cost breakdowns of deploying the phones on such a scale.
Cheers,
Chris Moses
twitter: @sanamobile
--
You received this message because you are subscribed to the Google Groups "ict4chw" group.
To post to this group, send email to ict...@googlegroups.com.
To unsubscribe from this group, send email to ict4chw+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/ict4chw?hl=en.
Paul
Dec 14, 2010, 12:16:21 PM12/14/10
to ict...@googlegroups.com, zra...@regenstrief.org
Good quick summary. I think this paper needs something near the sentence "We used HTC Dreams (a.k.a. T-Mobile's G1 in the US) as our device." [to record a survey asking about...] or [record patient data...]. In the above I can't tell what kinds of things were being asked. One or two sentences would bring the gap to explain what was being surveyed.
thanks
-Paul
thanks
-Paul
neal lesh
Dec 14, 2010, 5:33:32 PM12/14/10
to ict...@googlegroups.com, zra...@regenstrief.org
Dear Zeshan,
Thanks for this great post! One of the things I really appreciate about this group is the strong showing across a range of styles and perspectives we see here, and I think you’ve added a new style and perspective to the list. (In other words, there is no need to beat Dark Wizards at their own game.)
It was really instructive to see all these numbers, and I imagine you generated all kinds of new data. Were you aware of previous estimates for what percentage of people who find out they are HIV+ present for care or how quickly they do it? How do you know whether or not people presented at a clinic after being identified as HIV+ by HCT? I’m guessing you tracked them by giving them an ID but I wonder if it was common to enroll without using it, in which case your 168 might be a substantial under estimate?
And are you saying that previously it was thought that 5% of this population was HIV+ but did not know it and you found out that it was in fact 2%? I can see how this reduced the cost effectiveness of the HCT, but it also seems a strong example of how better surveillance is needed to know what’s really going on. Is that a fair interpretation?
In the current scheme, do the counselors return to the households ever and does or should the software help with that (either with reminders or by using GPS for navigation, for example)? I wouldn’t have guessed that 96% of the people were home when you visited? Did you achieve those results with many return visits?
Is the current use of the Androids is to improve data collection compared to paper? Have you considered using it as a job aid? As an example building off some earlier posts on this list, you could consider incorporating videos into the counseling process to encourage people to seek care at the clinics.
Thanks again and as always, just answer the ones you want to.
take care,
neal
Zeshan Rajput
Dec 15, 2010, 10:02:42 AM12/15/10
to neal lesh, ict...@googlegroups.com
Excellent questions, Neal! Replies inline below.
-Zeshan
-Zeshan
Were you aware of previous estimates for what percentage of people who find out they are HIV+ present for care or how quickly they do it?
I'm personally not aware of previous estimates for this population. The data I personally have access to is in a care setting, so anyone diagnosed at the care setting is already in care (and gets followed up by our excellent patient outreach group, which isn't currently following patients identified by HCT).
How do you know whether or not people presented at a clinic after being identified as HIV+ by HCT?
Your guess is exactly right - each patient identified through HCT is given an identifier, which is also written on a card and given to the patient. It's possible that some may not have the IDs nor mention they were identified in HCT during their intake to AMPATH, but my suspicion of both being true is low. These will be identified during the patient matching into AMPATH's OpenMRS installation, so I can let you know in a bit whether my suspicion holds true or not.
And are you saying that previously it was thought that 5% of this population was HIV+ but did not know it and you found out that it was in fact 2%? I can see how this reduced the cost effectiveness of the HCT, but it also seems a strong example of how better surveillance is needed to know what’s really going on. Is that a fair interpretation?
No, I misspoke on that. We excluded anyone already known to be HIV positive, which was the previously known 5%. We found an additional 2%. I need to look at the data to make this interpretation, but I suspect that true population prevalence in this area is actually closer to 7%. The point I was making is that HCT is a $300k cost to identify an additional 2% of the population as HIV positive if you take this as the only positive outcome of HCT (there are others, specifically in identifying sick individuals, TB positive patients, and mothers without prenatal care).
In the current scheme, do the counselors return to the households ever and does or should the software help with that (either with reminders or by using GPS for navigation, for example)? I wouldn’t have guessed that 96% of the people were home when you visited? Did you achieve those results with many return visits?
I don't have the answer to this on hand. In our ODK implementation we have a mechanism to allow counselors to note if a family wasn't home, requested a follow-up visit, and so on. I just don't have that data, as I don't believe it traverses to the OpenMRS installations (which is what I had pooled for this data). So I suspect you're correct as to the number of return visits, but I can't corroborate that yet.
Is the current use of the Androids is to improve data collection compared to paper? Have you considered using it as a job aid? As an example building off some earlier posts on this list, you could consider incorporating videos into the counseling process to encourage people to seek care at the clinics.
The current use is to improve data collection. We haven't considered it as a job aid yet, though I'm aware of the fantastic material that's been posted on this list. The first step is to try and identify why people aren't following up. If it's forgetfulness, then an SMS reminder may be the answer. If it's a lack of understanding, then perhaps improved counselor training or incorporating videos would help. I suspect that this is multifactorial, in which case we will likely target the highest factors, try to correct these, and see if this helps. But first, since this is preliminary data, the next step is to verify our findings in a second district. We're well on the way there, and perhaps I'll have an opportunity to update everyone here again.
From: ict...@googlegroups.com [mailto:ict...@googlegroups.com] On Behalf Of Chris Moses
Sent: Monday, December 13, 2010 1:24 PM
To: ict...@googlegroups.com; zra...@regenstrief.org
Subject: Re: [ict4chw] AMPATH's Home-Based Counseling and Testing Project (insert catchier title here)
Hi Zeshan,
Reply all
Reply to author
Forward
0 new messages