Health IT

come on nathan, even resource limited settings have an infinite supply of numbers; even the poorest of people have names, and even people with the same name have distinguishing characteristics (often mother's name is used); that is not your problem. the problem is your clinic is not keeping ANY information on the patients it sees. even in the lowest literacy areas there are copyists, or you can use a one-write system. it's clear your doctors don't see much value in a longitudinal patient record, and until they do neither OpenMRS nor any other EMR system can succeed.

Roger makes a number of good suggestions but let's be clear here, it is not easy to do this well in the sort of environment that you are working in Nathan. In fact it is not that well solved even in many rich countries as they may know the ID but in the US the records are often still stuck in the other hospital... There is a great benefit in making progress though, if you had 90% correctly identified they will get better care, and the ease of registering patients, finding their records and reducing the number of paper record should motivate the staff to use the system more. Another caveat though, you will need to get the system (paper, human and electronic) running smoothly and build up the number of registered patients before the benefits will be clear. Some projects like Baobab in Malawi have had good results with barcoded ID cards or labels attached to health passports. There are some projects that have reported success using finger prints, but I would definitely start with the simpler approach and consider finger prints if that isn't enough.
Regards

Hamish

Nathan,
We have been working in a similar environment and have found that biometrics has worked to our advantage.
We work for the cause of TB. As you know, lack of oversight and inconsistent patient monitoring is one of the reasons patients stop taking their meds. This led us to develop such a system and Operation ASHA teamed up with Microsoft Research and Innovators in Health to launch the next phase of tuberculosis treatment: eDOTS. Building on the WHO’s effective DOTS strategy (Directly Observed Therapy), Operation ASHA combined biometric tracking and SMS to ensure that our health workers quickly respond when a patient misses a dose.
The system is simple: A commercial laptop, fingerprint reader and cellphone. Each time a patient visits an OpASHA center, the visit is confirmed by a fingerprint scanner. At the end of the day, the eDOTS terminal sends an attendance report to a central Electronic Medical Record using SMS. If the record shows that a patient has missed a dose, a separate text message informs the patient’s health worker that immediate action is needed. The counselor has 48 hours to visit the patient and deliver the medicines.
Let me know if you need ...

Nathan,
We have been working in a similar environment and have found that biometrics has worked to our advantage.
We work for the cause of TB. As you know, lack of oversight and inconsistent patient monitoring is one of the reasons patients stop taking their meds. This led us to develop such a system and Operation ASHA teamed up with Microsoft Research and Innovators in Health to launch the next phase of tuberculosis treatment: eDOTS. Building on the WHO’s effective DOTS strategy (Directly Observed Therapy), Operation ASHA combined biometric tracking and SMS to ensure that our health workers quickly respond when a patient misses a dose.
The system is simple: A commercial laptop, fingerprint reader and cellphone. Each time a patient visits an OpASHA center, the visit is confirmed by a fingerprint scanner. At the end of the day, the eDOTS terminal sends an attendance report to a central Electronic Medical Record using SMS. If the record shows that a patient has missed a dose, a separate text message informs the patient’s health worker that immediate action is needed. The counselor has 48 hours to visit the patient and deliver the medicines.
Let me know if you need more information on such a system.
Prateek

Hello Nathan,
We also face the same issue here in Swaziland and much of it has to do with
the Ministry of Health not adopting/agreeing upon a consistent strategy
with respect to national IDs as well as the prevalence of same names across
a group of individuals and the lack of enforcement of needing a birth
certificate. Additionally, there are a lot of politics and concern about
the ID (normally used for Department of Home Affairs) being used for
medical reasons and distrust around that.

Is there a group that stores a set of IDs you can work with as a starter if
there is a certain amount of coverage? Do you have a patient ID system in
the works by chance? Realize how difficult and political this can be to
start getting set up (we are in talks right now to strategize creating one
at the clinic level) but if you have this you could capture a list of
defining attributes for which to identify someone and perhaps do a fuzzy
match on the back end? I have also heard about GIS technology which
provides coordinates for home addresses but that is usually linked to head
of household ...

Hello Nathan,
We also face the same issue here in Swaziland and much of it has to do with
the Ministry of Health not adopting/agreeing upon a consistent strategy
with respect to national IDs as well as the prevalence of same names across
a group of individuals and the lack of enforcement of needing a birth
certificate. Additionally, there are a lot of politics and concern about
the ID (normally used for Department of Home Affairs) being used for
medical reasons and distrust around that.

Is there a group that stores a set of IDs you can work with as a starter if
there is a certain amount of coverage? Do you have a patient ID system in
the works by chance? Realize how difficult and political this can be to
start getting set up (we are in talks right now to strategize creating one
at the clinic level) but if you have this you could capture a list of
defining attributes for which to identify someone and perhaps do a fuzzy
match on the back end? I have also heard about GIS technology which
provides coordinates for home addresses but that is usually linked to head
of household. A potential starting point?

If you like, I'm also happy to have a separate conversation with you as we
are working on this very same challenge as we speak. Let me know.
Anne

Nathan, are you following the work of Nandan Nilekani and UIDAI (Unique Identity Authority of India)? This is the ex-head of Infosys, now working for the Indian government on a project called Aadhaar with UIDAI, rolling out unique IDs for millions, based on biometrics. They are running into glitches, of course, people whose fingerprints are worn away from years of labor, people who don't know when they were born, but they are making progress. Perhaps the Aadhaar system can be matched to interoperate with the prevalent EMR system you want to use.
A link: http://uidai.gov.in/index.php?option=com_content&view=article&id=145&Itemid=2

Dear  Prateek

Thisis good information and utility culd be extended to other areas which require revisit like immunization but subject to fact that patienst / family owns a mobile phone
anil

 
<font style="background-color:rgb(128, 192, 255);" face="arial narrow"><font size="1" face="verdana">IMMUNIZE EVERY CHILD FOR HEALTHY FUTURE,</font> </font><font style="background-color:rgb(255, 255, 64);">PREVENT AIDS ,-PRACTICE SAFE SEX</font> ,<font style="background-color:rgb(255, 128, 255);" face="courier">EAT HEALTHY FOODS SLEEP WELL</font> , <font style="background-color:rgb(216, 163, 164);" face="arial narrow">EXERCISE DAILY</font> , <font style="background-color:rgb(128, 255, 0);" face="comic sans ms">MEDITATE FOR HEALTHY LIVING</font>

Hi Nathan,

We have faced similar challenges here in Pakistan and largely solved the
problem using cell phone contact numbers as a patient identifier. It turns
out that even the poorest families now own a cell phone here. If that is
the case in your locality, this might be a pretty quick and easy way to
find records. Granted, several people in a family could be attached to the
same phone number, but once you have narrowed it down to that level, it
should be not much of a problem using name, gender and other details.

Another major hospital I recently visited has employed a camera to take
pictures of patients at registration which then gets attached to their
records and is another simple and easy method of ensuring identification.
Finger printing or other advanced biometrics could probably work, but I
usually find simpler methods work best in low resource settings.
Introducing complex tools would probably be more challenging because of the
learning curve involved for the staff.

Best.

Usman Raza
www.linkedin.com/in/uraza
Healthcare & Everything Else<http://healthcareandeverythingelse.blogspot.com>
chs.prime.edu.pk

Prateek,
This sounds really interesting.

What fingerprint scanner and software do you use? and did you need to
make any customization to your system to have the fingerprint be read,
or did it input it as a text string or something?

Joaquín
___________________________________________________________________
Gerente de Desarrollo, eHealth Systems
Research Fellow, Escuela de Medicina de Harvard
Moderador, GHDOnline.org

Mobile phones may be used, in combination with a simple PIN, to establish very robust 2-factor authentication of individuals. The premise is that the phone's SIM provides a globally unique ID (something you HAVE) which can be coupled with the PIN (something you KNOW). This also works in cases where the phone is shared -- as long as each person has their own PIN.

As a best practice, the PIN should be keyed using an interactive voice response type of system (IVR) or some other alternate to SMS. This is because SMS leaves the PIN codes in the txt outbox (which does not protect the PINs between multiple people sharing the same phone).

A sample workflow could be:

1. client pings ID service (via SMS or "beep")
2. ID service calls client; voice prompt to "enter your PIN and press #"
3. client enters PIN

At this point, the workflow has authenticated the client and could do a number of useful things including providing patient-centric education, results or alerts. If some manner of workflow identifier was sent as part of step 1, the workflow could even retrieve and send information to the care provider's mobile or network-connected device.

In areas ...

Mobile phones may be used, in combination with a simple PIN, to establish very robust 2-factor authentication of individuals. The premise is that the phone's SIM provides a globally unique ID (something you HAVE) which can be coupled with the PIN (something you KNOW). This also works in cases where the phone is shared -- as long as each person has their own PIN.

As a best practice, the PIN should be keyed using an interactive voice response type of system (IVR) or some other alternate to SMS. This is because SMS leaves the PIN codes in the txt outbox (which does not protect the PINs between multiple people sharing the same phone).

A sample workflow could be:

1. client pings ID service (via SMS or "beep")
2. ID service calls client; voice prompt to "enter your PIN and press #"
3. client enters PIN

At this point, the workflow has authenticated the client and could do a number of useful things including providing patient-centric education, results or alerts. If some manner of workflow identifier was sent as part of step 1, the workflow could even retrieve and send information to the care provider's mobile or network-connected device.

In areas where there is NOT strong mobile phone adoption, a "community phone" at the clinic could be employed. This is the shared phone idea taken to a wider extreme and potentially being used to authenticate a large group of individuals, each of which has their own PIN. Such systems may be rapidly and inexpensively deployed at a low capital cost; the run costs will depend on the mobile carrier's price structure.

Such systems are used, today, in some regions to support mBanking. Has anyone tried using similar authentication schemes for healthcare?

Derek.

Hi Derek,

We had a lot of trouble using PINs with patients in rural Uganda (see the abstract/paper link below). Our success rates improved a lot with just-in-time training. That is, we set up an automated system in which our staff were notified of PIN and other end user-related problems by SMS, so that they could assist the end user at the time he/she was having problems. PINs make work better in other settings in which people are more familiar with them, but I would do some focus groups or other such testing before assuming they will function well.

Best,
Jessica


AIDS Behav. 2010 Dec;14(6):1294-301.
Challenges in using mobile phones for collection of antiretroviral therapy adherence data in a resource-limited setting.
Haberer JE, Kiwanuka J, Nansera D, Wilson IB, Bangsberg DR.
Source
Department of General Internal Medicine, Massachusetts General Hospital, Boston, MA, USA.
Abstract
Frequent antiretroviral therapy adherence monitoring could detect incomplete adherence before viral rebound develops and thus potentially prevent treatment failure. Mobile phone technologies make frequent, brief adherence interviews possible in resource-limited settings; however, feasibility and acceptability are unknown. Interactive voice response (IVR) and short message service (SMS) text messaging ...

Hi Derek,

We had a lot of trouble using PINs with patients in rural Uganda (see the abstract/paper link below). Our success rates improved a lot with just-in-time training. That is, we set up an automated system in which our staff were notified of PIN and other end user-related problems by SMS, so that they could assist the end user at the time he/she was having problems. PINs make work better in other settings in which people are more familiar with them, but I would do some focus groups or other such testing before assuming they will function well.

Best,
Jessica


AIDS Behav. 2010 Dec;14(6):1294-301.
Challenges in using mobile phones for collection of antiretroviral therapy adherence data in a resource-limited setting.
Haberer JE, Kiwanuka J, Nansera D, Wilson IB, Bangsberg DR.
Source
Department of General Internal Medicine, Massachusetts General Hospital, Boston, MA, USA.
Abstract
Frequent antiretroviral therapy adherence monitoring could detect incomplete adherence before viral rebound develops and thus potentially prevent treatment failure. Mobile phone technologies make frequent, brief adherence interviews possible in resource-limited settings; however, feasibility and acceptability are unknown. Interactive voice response (IVR) and short message service (SMS) text messaging were used to collect adherence data from 19 caregivers of HIV-infected children in Uganda. IVR calls or SMS quantifying missed doses were sent in the local language once weekly for 3-4 weeks. Qualitative interviews were conducted to assess participant impressions of the technologies. Participant interest and participation rates were high; however, weekly completion rates for adherence queries were low (0-33%), most commonly due to misunderstanding of personal identification numbers. Despite near ubiquity of mobile phone technology in resource-limited settings, individual level collection of healthcare data presents challenges. Further research is needed for effective training and incentive methods.

Free full article available here: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2975780/?tool=pubmed

@Derek - An alternative to SMS is USSD but the 'startup costs' in terms of setting up the operator agreements to access that sort of interaction may make it unfeasible for one-off, pilot, or smaller-than-national programs. Which is a pity because it leads to chicken/egg situations (can't design good systems because of infrastructure setup / don't set up infrastructure due to lack of demand from good designs)

@Jessica -((Thanks for the paper! the testimonials are great; and its awesome that it got published with all these sort of lessons)) for large- scale support of mHealth activities we have worked with the operators' help desks so that basic questions about some 'apps' could be answered by the commercial helpdesk (the same you would call if, say, you have trouble adding credit to your phone). But this is just first-tier support and predicated on a good design up-front - where you test the peer-to-peer training, usability, etc. High call rates would have them growling.. But this did not seem a requirement for the Uganda research.

I wonder what a good participatory interaction design session around the issues of identification, confidentiality, etc that PINs try to address might yield if one does not ...

@Derek - An alternative to SMS is USSD but the 'startup costs' in terms of setting up the operator agreements to access that sort of interaction may make it unfeasible for one-off, pilot, or smaller-than-national programs. Which is a pity because it leads to chicken/egg situations (can't design good systems because of infrastructure setup / don't set up infrastructure due to lack of demand from good designs)

@Jessica -((Thanks for the paper! the testimonials are great; and its awesome that it got published with all these sort of lessons)) for large- scale support of mHealth activities we have worked with the operators' help desks so that basic questions about some 'apps' could be answered by the commercial helpdesk (the same you would call if, say, you have trouble adding credit to your phone). But this is just first-tier support and predicated on a good design up-front - where you test the peer-to-peer training, usability, etc. High call rates would have them growling.. But this did not seem a requirement for the Uganda research.

I wonder what a good participatory interaction design session around the issues of identification, confidentiality, etc that PINs try to address might yield if one does not start with the assumption of a PIN number as the solution. The principle of 'something you know, something you have' may still come through in unexpected ways!
__________________________
Eduardo Jezierski, InSTEDD, CTO
skype: eduardojezierski
twitter: @edjez

Thank you Jessica and Ed. Very helpful comments and, Jessica, thanks so much for the link to the paper.

-Derek.

Anil/ Nathan,

The setup of such a system includes commercially available small laptop, a fingerprint reader device, a simple mobile phone (we use Nokia which costs less than USD25) and a data cable to connect this mobile phone with the laptop.

Each time a patient visits an OpASHA center, the visit is confirmed by a fingerprint scanner. At the end of the day, the eDOTS terminal sends an attendance report to a central EMR using SMS. We rely on SMS network rather than internet as its availability in rural areas is a concern.

Patient DOES NOT need to own a mobile phone.

Prateek,

To followup on Joaquin's comment, could you provide more details on the software and how it works?

For the purposes of the discussion, I would like to refer our colleague to http://www.unaids.org/en/media/unaids/contentassets/documents/dataanalysis/20...

This document is about Developing and Using Individual Identifiers for the Provision of Health Services including HIV, and it is from the
Proceedings of a Workshop, 24–26 February 2009.

I concur with Hamish's observations. And wish to add that health data privacy policy environment in some countries restrict the use of Unique Ids for patient records. In addition, the jurisdictional boundaries and lack of harmonized policies within in sub national level in some countries makes it difficult to use Unique IDs. Perhaps we need a global conference to discuss the topic in greater detail and lead towards a policy-enabled environment that includes all concerns and still provide an opportunity for Unique IDs adoption (since technically generating Unique IDs is not a technical challenge nor the center of the problem).

Thanks, Ramesh,

Sorry I have been out of the discussion for awhile. At the Millennium Villages Project (MVP), we are using OpenMRS with a base-30 character set (excluding any letter that looks like a number) and a Luhn Mod 30 identifier validation scheme as a check digit. We use this for both OpenMRS-entered patients as well as for the ChildCount+ CHW information system (mobile phone entry).

In cases where there are multiple competing jurisdictions, it means that there has to be an enterprise master patient index (eMPI) set up. Obviously, patient privacy through obfuscation is not a good thing for the patient so we had better use other methods for ensuring patient privacy :)

Best,
Andy
----------------------------------
Andrew S. Kanter, MD MPH
Asst. Prof. Clinical Biomedical Informatics
and Clinical Epidemiology
622 W. 168th Street, VC5
New York, NY 10032
Email:
Office: +1 212 305 4842
Cell: +1 646 469 2421
Skype: akanter-ippnw

Hi Andree Kanter in my country to avoid problems with the index patient as
a policy we have taken is working as the patient's unique history id = ID
number in Ecuador call it identity card
is the only way to have a single patient record anywhere in the country.
We are working on schedular all over the country as quickly as possible
especially is this hubicando ID card centers for newborns.
I would like to know a little more like the mobile project you're working
me very interested in that topic

Hi everyone,

Thank you for this excellent exchange. We've summarized it in a Discussion Brief and added links to more resources. Discussion briefs are peer-reviewed and accessible online in the discussion to signed-in members.

Let us know what you think, and keep on adding more resources to this discussion and inviting colleagues to chime in and share how they do patient identification.

Best, Sophie

Nathan, the problem you are encountering as you correctly said is global. In Kenya, the MoH is working very closely with the department of vital registration (registration of deaths an births) in developing a unique identification system for Kenyan citizens, and the plan is to use this platform to extend to the various sectors that need such a service, evidently the health sector is one of them. This effort is really more for the long haul as you know how long any gov projects take to bear fruits. In the meantime, I know various partners in the health sector have implemented very novel unique identification mechanisms, and this includes the largest OpenMRS installation globally, @ AMPATH (http://www.ampathkenya.org/). My input into this is that just like several people mentioned about cell phones, I think the success of any unique identification system is only as good as how much the people value it. So for instance, you will find illiterate people have memorized their cell phone numbers because they know the value of their cell phone, and so using a combination of whatever method that works, I think we cannot circumvent that factor of especially where the environment is not ...

Nathan, the problem you are encountering as you correctly said is global. In Kenya, the MoH is working very closely with the department of vital registration (registration of deaths an births) in developing a unique identification system for Kenyan citizens, and the plan is to use this platform to extend to the various sectors that need such a service, evidently the health sector is one of them. This effort is really more for the long haul as you know how long any gov projects take to bear fruits. In the meantime, I know various partners in the health sector have implemented very novel unique identification mechanisms, and this includes the largest OpenMRS installation globally, @ AMPATH (http://www.ampathkenya.org/). My input into this is that just like several people mentioned about cell phones, I think the success of any unique identification system is only as good as how much the people value it. So for instance, you will find illiterate people have memorized their cell phone numbers because they know the value of their cell phone, and so using a combination of whatever method that works, I think we cannot circumvent that factor of especially where the environment is not mature enough to support advanced technology that would bear that burden so that patients don't have to memorize PINs, and other codes.

hi there,
i have really enjoyed this discussion to the fullest.i can see how technology is being alined in healthcare service delivery.
in masaka Uganda,an ART care program,we have seriously embarked capturing patients fone contacts,geographical locations,patients photo-shots(small size) and droping a fone call if they miss their appointments.but this in early stages,it seems to be yielding results,as patients are feeling a sense of belonging and being cared for.