Anything is possible in terms of technical feasibility. I would not focus on the technical issues at the moment.
The issue is really about what research question are you hoping to answer with this approach? What added value does spatial information provide?
Can you provide a study design that includes a research question that can be tested using a case-control intervention?
And what granularity of information do you require? Is knowing the person's place of residence sufficient or are you attempting to track travel? Is the focus on physical location or social interaction?
Is this proposed as a general approach for many interventions or only for one or more targeted interventions?
You would need very strong guarantees of security and privacy. Participants would need to provide informed consent by explicitly opting-in to such a study.
At minimum, the plan for managing any individually-identifiable spatial data should be very clearly described in an IRB application that goes through a formal review process with the appropriate stakeholders, including representatives from the patient privacy advocacy community.
Steps need to be taken to ensure such information does not fall into the wrong hands -- such a breach would be considered a human rights violation.
Sorry for my late response, I was at the OpenMRS implementers meeting in Cape Town and then came home to Eid Holidays.
Mikhail brought up some important points regarding security. When you use GPS information, you definitely want to access it through a secure log-in to specified users only (in our organization access is typically restricted to 2 or 3 people heading the project), and keep patient identifiers at a minimum when you display data.
I just wanted to share with you examples of how we have been using GPS in our projects in Pakistan to help us monitor our research and program field sites, our field workers, and allow us to respond to problems in the field quickly and efficiently.
Below are links to a couple of videos that demonstrate the use of technology in different projects:
Sorry for my late response, I was at the OpenMRS implementers meeting in Cape Town and then came home to Eid Holidays.
Mikhail brought up some important points regarding security. When you use GPS information, you definitely want to access it through a secure log-in to specified users only (in our organization access is typically restricted to 2 or 3 people heading the project), and keep patient identifiers at a minimum when you display data.
I just wanted to share with you examples of how we have been using GPS in our projects in Pakistan to help us monitor our research and program field sites, our field workers, and allow us to respond to problems in the field quickly and efficiently.
Below are links to a couple of videos that demonstrate the use of technology in different projects:
Lastly our Pneumonia Surveillance study - here we aren't using GPS, but we are using RFID for identification to help us follow study participants for 2 years.
http://www.youtube.com/watch?v=6nGxO2Zqe9g
If you'd like more information on any of these, feel free to contact me on
Best,
--
Julia Irani
Director, Center for Health Informatics
Interactive Research & Development (IRD)
Karachi, Pakistan
I will soon be using GPS in a study of PrEP adherence in Uganda. We will be conducting home visits and will use GPS devices to help reliably locate participants' homes. I agree with the earlier comments that confidentiality and informed consent are important; however, I believe they can be used responsibly and will be very valuable in our study setting where few roads or paths are labelled.
I'd be happy to share our experiences after collecting data, which should start next month.
Regards,
Jessica Haberer
Research Scientist, Harvard Initiative for Global Health
Hi Jessica,We've got a huge challenge in our program in Nigeria when it
comes to locating our patients' homes.
In this setting, very few roads are mapped as well.
Please can you tell me more about the use of GPS in tracking these patients
in these kinds of settings?
Any information or information sources will be highly appreciated.
I agree with you. In addition to that it will be a good experience in Africa (Uganda) given the situation of several african countries whereby HIV program is surving due to shifting of tasks by doctors to low caders.
Shortage of medical staff explaning their workload cannot sustain such policy in several countries. Even trained Family physician are failing to applu home visits as required in their principles. Lets wait and see What we will learn from the study and probably implications on others programs run by those participants in their various facilities.
I would like to introduce you to Julius Wasubire. He is a colleague from
Kampala who has quite a bit of experience with GPS from his studies at
University. He will also be in charge of using GPS to track participants
homes in our upcoming pre-exposure prophylaxis study. I am introducing him
to GHDonline, and he said he would be happy to talk with you about GPS. His
email is .
I am also looking into another contact for you in Mbarara, Uganda.
We have used GPS tracking extensively for adherence and retention research projects in Kenya. As for using the device to track and trace patients, it is OK. I think the CDC has used GPS to map homes for surveys and it is useful for this type of activity. Frankly, I think the act of mapping a home using GPS is more helpful than anything. This requires the community health worker who is doing the mapping to be careful in determining where the house is located. We always have the worker document the house location as well -- such as nearby landmarks -- to verify the GPS data. All of this activity seems to help cement the location of the house in the mind of the worker so that follow-up doesn't necessarily require the GPS. Feel free to e-mail me at if you have more questions.
I have been following up with the discussion "Effectiveness of GPS for
following study participants". I must say it is a very interesting
one. I have however, wondered the feasibility of this in practical
terms. I understand you are conducting a study which you must have
gotten patients' consent and ethical clearance for. However, i am of
the opinion that the patients are less likely to default since the are
aware of the study (or is it a blind study?).
Thinking through the field experience of which i have a wealth of and
also from the research i did in the UK on determinants of
non-completion of TB treatment, i found out that most of the patients
that default are low income earners (who most times do not have jobs
at all) who are most likely to be highly mobile.
Therefore, what strategies do you have to ensure that the mobility of
these patients is reduced without introducing bias to your study?
**Chinenye, I would also love to know your thoughts as regards to my
concerns since we have a common location with, i presume, common
challenge.
It might be helpful for the community to hear exactly how you are using the
GPS. For instance, could you address these questions?
1. Which devices are you using?
2. Do you map the whole route or key geographic areas?
3. Do you store the maps on the devices (which typically have limited
capacity) or use a desktop computer for storage and download maps as needed?
4. Which staff use the devices and how much/what type of training did they
need?
5. Who did the training?
Hi Ifunanya,
Thanks for your insight in the discussion.
I agree with you that study participants who have given their consent and
are probably remunerated in some way for participating in the study, are
unlikely to default,
however, apart from reducing Loss-to-follow-up in research studies,
tracking of patients is very useful in programs where non-adherence and its
consequences (resistant strains, etc) result from defaulting clinic and
pharmacy refill appointments;
So, I'm actually looking beyond research.
Again in looking at determinants of non-completion of TB patients,
mobility and socio-economic status may turn out to be some of the
determinants you're looking for,'
except you control for them ab initio.
Controlling for mobility might mean recruiting only stable patients.
I think that as long as you make your eligibility criteria and study design
clear, that won't be a problem.
The caveat is that, your results might lack the generalizability it ought to
have had.
I would not exclude such patients but will track them for ethical reasons.
However in the analysis you can look at associations btw socio-economic
status, mobility, and non-completion of treatment.
You may also look at sub-groups of pts who moved or defaulted during the
study ...
Hi Ifunanya,
Thanks for your insight in the discussion.
I agree with you that study participants who have given their consent and
are probably remunerated in some way for participating in the study, are
unlikely to default,
however, apart from reducing Loss-to-follow-up in research studies,
tracking of patients is very useful in programs where non-adherence and its
consequences (resistant strains, etc) result from defaulting clinic and
pharmacy refill appointments;
So, I'm actually looking beyond research.
Again in looking at determinants of non-completion of TB patients,
mobility and socio-economic status may turn out to be some of the
determinants you're looking for,'
except you control for them ab initio.
Controlling for mobility might mean recruiting only stable patients.
I think that as long as you make your eligibility criteria and study design
clear, that won't be a problem.
The caveat is that, your results might lack the generalizability it ought to
have had.
I would not exclude such patients but will track them for ethical reasons.
However in the analysis you can look at associations btw socio-economic
status, mobility, and non-completion of treatment.
You may also look at sub-groups of pts who moved or defaulted during the
study and compare the associations with those who weren't mobile.
Thank you for your contributions to this discussion.
Chinenye, I agree with your comments.
Ifunanya, we are using GPS solely for tracking directions and the locations of study participants’ homes in studies of ART adherence (which have full IRB approval and participant consent). I don’t think the GPS influences study retention rates, although we are not studying it. I think our use of frequent contact and incentives (e.g. sugar or soap at home visits) are more influential. We do limit participation to within 20 to 30 km of our site due to logistical reasons (e.g. fuel and staff time needed to reach the homes), but that radius does not exclude many participants in the recruitment pool.
You raise some very interesting points about loss to follow up and patient tracking. I’d like to refer you to a recent article by one of my colleagues, Elvin Geng. The paper, Understanding Reasons for and Outcomes of Patients Lost to Follow-Up in Antiretroviral Therapy Programs in Africa Through a Sampling-Based Approach (JAIDS epub), refers to the clinic in Mbarara, Uganda, where many of the participants in our studies are seen for routine clinical care. Geng ...
Thank you for your contributions to this discussion.
Chinenye, I agree with your comments.
Ifunanya, we are using GPS solely for tracking directions and the locations of study participants’ homes in studies of ART adherence (which have full IRB approval and participant consent). I don’t think the GPS influences study retention rates, although we are not studying it. I think our use of frequent contact and incentives (e.g. sugar or soap at home visits) are more influential. We do limit participation to within 20 to 30 km of our site due to logistical reasons (e.g. fuel and staff time needed to reach the homes), but that radius does not exclude many participants in the recruitment pool.
You raise some very interesting points about loss to follow up and patient tracking. I’d like to refer you to a recent article by one of my colleagues, Elvin Geng. The paper, Understanding Reasons for and Outcomes of Patients Lost to Follow-Up in Antiretroviral Therapy Programs in Africa Through a Sampling-Based Approach (JAIDS epub), refers to the clinic in Mbarara, Uganda, where many of the participants in our studies are seen for routine clinical care. Geng et al employed a human “tracker”, without GPS, to do the follow up work and was quite successful. I would be happy to bring him into the discussion if you are interested in learning more about his methods.
Mikhail Elias
Anything is possible in terms of technical feasibility. I would not focus on the technical issues at the moment.
The issue is really about what research question are you hoping to answer with this approach? What added value does spatial information provide?
Can you provide a study design that includes a research question that can be tested using a case-control intervention?
And what granularity of information do you require? Is knowing the person's place of residence sufficient or are you attempting to track travel? Is the focus on physical location or social interaction?
Is this proposed as a general approach for many interventions or only for one or more targeted interventions?
You would need very strong guarantees of security and privacy. Participants would need to provide informed consent by explicitly opting-in to such a study.
At minimum, the plan for managing any individually-identifiable spatial data should be very clearly described in an IRB application that goes through a formal review process with the appropriate stakeholders, including representatives from the patient privacy advocacy community.
Steps need to be taken to ensure such information does not fall into the wrong hands -- such a breach would be considered a human rights violation.
1:30 PM, 14 Sep 2009 | Permalink
Julia Irani
Hi Chorongo Salee,
expand commentSorry for my late response, I was at the OpenMRS implementers meeting in Cape Town and then came home to Eid Holidays.
Mikhail brought up some important points regarding security. When you use GPS information, you definitely want to access it through a secure log-in to specified users only (in our organization access is typically restricted to 2 or 3 people heading the project), and keep patient identifiers at a minimum when you display data.
I just wanted to share with you examples of how we have been using GPS in our projects in Pakistan to help us monitor our research and program field sites, our field workers, and allow us to respond to problems in the field quickly and efficiently.
Below are links to a couple of videos that demonstrate the use of technology in different projects:
1. MDR-TB Program:
http://www.youtube.com/watch?v=U7RoBIO1xaU
2.Google Earth Data Visualization
http://www.youtube.com/watch?v=v-3lqG3hSYM
3. National Dog bite and Rabies Surveillance in Pakistan
http://www.youtube.com/watch?v=NaD5DIc5RUc
Lastly our Pneumonia Surveillance study - here we aren't using GPS, but we are using RFID for identification to help us ...
3:41 AM, 2 Oct 2009 | Permalink
Jessica Haberer, MD, MS
Hi All,
I will soon be using GPS in a study of PrEP adherence in Uganda. We will be conducting home visits and will use GPS devices to help reliably locate participants' homes. I agree with the earlier comments that confidentiality and informed consent are important; however, I believe they can be used responsibly and will be very valuable in our study setting where few roads or paths are labelled.
I'd be happy to share our experiences after collecting data, which should start next month.
Regards,
Jessica Haberer
Research Scientist, Harvard Initiative for Global Health
8:54 PM, 7 Oct 2009 | Permalink
Chinenye Ugoji
Hi Jessica,We've got a huge challenge in our program in Nigeria when it
comes to locating our patients' homes.
In this setting, very few roads are mapped as well.
Please can you tell me more about the use of GPS in tracking these patients
in these kinds of settings?
Any information or information sources will be highly appreciated.
Thanks,
Chinenye
7:17 AM, 8 Oct 2009 | Permalink
Prosper Lutala
Dear Chinenye;
I agree with you. In addition to that it will be a good experience in Africa (Uganda) given the situation of several african countries whereby HIV program is surving due to shifting of tasks by doctors to low caders.
Shortage of medical staff explaning their workload cannot sustain such policy in several countries. Even trained Family physician are failing to applu home visits as required in their principles. Lets wait and see What we will learn from the study and probably implications on others programs run by those participants in their various facilities.
Prosper Lutala
3:38 PM, 8 Oct 2009 | Permalink
Jessica Haberer, MD, MS
Dear Chinenye,
I would like to introduce you to Julius Wasubire. He is a colleague from
Kampala who has quite a bit of experience with GPS from his studies at
University. He will also be in charge of using GPS to track participants
homes in our upcoming pre-exposure prophylaxis study. I am introducing him
to GHDonline, and he said he would be happy to talk with you about GPS. His
email is .
I am also looking into another contact for you in Mbarara, Uganda.
Regards,
Jessica
11:49 PM, 14 Oct 2009 | Permalink
Michael Chung
Dear Chinenye,
We have used GPS tracking extensively for adherence and retention research projects in Kenya. As for using the device to track and trace patients, it is OK. I think the CDC has used GPS to map homes for surveys and it is useful for this type of activity. Frankly, I think the act of mapping a home using GPS is more helpful than anything. This requires the community health worker who is doing the mapping to be careful in determining where the house is located. We always have the worker document the house location as well -- such as nearby landmarks -- to verify the GPS data. All of this activity seems to help cement the location of the house in the mind of the worker so that follow-up doesn't necessarily require the GPS. Feel free to e-mail me at if you have more questions.
Michael
1:46 AM, 16 Oct 2009 | Permalink
Chinenye Ugoji
Thanks Jessica. That would be really helpful.
Chinenye
2:41 AM, 16 Oct 2009 | Permalink
Ifunanya Igbojekwe
Dear Jessica,
attn: Chinenye,
I have been following up with the discussion "Effectiveness of GPS for
following study participants". I must say it is a very interesting
one. I have however, wondered the feasibility of this in practical
terms. I understand you are conducting a study which you must have
gotten patients' consent and ethical clearance for. However, i am of
the opinion that the patients are less likely to default since the are
aware of the study (or is it a blind study?).
Thinking through the field experience of which i have a wealth of and
also from the research i did in the UK on determinants of
non-completion of TB treatment, i found out that most of the patients
that default are low income earners (who most times do not have jobs
at all) who are most likely to be highly mobile.
Therefore, what strategies do you have to ensure that the mobility of
these patients is reduced without introducing bias to your study?
**Chinenye, I would also love to know your thoughts as regards to my
concerns since we have a common location with, i presume, common
challenge.
10:34 AM, 16 Oct 2009 | Permalink
Jessica Haberer, MD, MS
Hi Michael,
It might be helpful for the community to hear exactly how you are using the
GPS. For instance, could you address these questions?
1. Which devices are you using?
2. Do you map the whole route or key geographic areas?
3. Do you store the maps on the devices (which typically have limited
capacity) or use a desktop computer for storage and download maps as needed?
4. Which staff use the devices and how much/what type of training did they
need?
5. Who did the training?
Thank you for your time and comments,
Jessica
1:39 PM, 16 Oct 2009 | Permalink
Chinenye Ugoji
Hi Ifunanya,
expand commentThanks for your insight in the discussion.
I agree with you that study participants who have given their consent and
are probably remunerated in some way for participating in the study, are
unlikely to default,
however, apart from reducing Loss-to-follow-up in research studies,
tracking of patients is very useful in programs where non-adherence and its
consequences (resistant strains, etc) result from defaulting clinic and
pharmacy refill appointments;
So, I'm actually looking beyond research.
Again in looking at determinants of non-completion of TB patients,
mobility and socio-economic status may turn out to be some of the
determinants you're looking for,'
except you control for them ab initio.
Controlling for mobility might mean recruiting only stable patients.
I think that as long as you make your eligibility criteria and study design
clear, that won't be a problem.
The caveat is that, your results might lack the generalizability it ought to
have had.
I would not exclude such patients but will track them for ethical reasons.
However in the analysis you can look at associations btw socio-economic
status, mobility, and non-completion of treatment.
You may also look at sub-groups of pts who moved or defaulted during the
study ...
5:55 PM, 18 Oct 2009 | Permalink
Jessica Haberer, MD, MS
Dear Ifunanya and Chinenye,
expand commentThank you for your contributions to this discussion.
Chinenye, I agree with your comments.
Ifunanya, we are using GPS solely for tracking directions and the locations of study participants’ homes in studies of ART adherence (which have full IRB approval and participant consent). I don’t think the GPS influences study retention rates, although we are not studying it. I think our use of frequent contact and incentives (e.g. sugar or soap at home visits) are more influential. We do limit participation to within 20 to 30 km of our site due to logistical reasons (e.g. fuel and staff time needed to reach the homes), but that radius does not exclude many participants in the recruitment pool.
You raise some very interesting points about loss to follow up and patient tracking. I’d like to refer you to a recent article by one of my colleagues, Elvin Geng. The paper, Understanding Reasons for and Outcomes of Patients Lost to Follow-Up in Antiretroviral Therapy Programs in Africa Through a Sampling-Based Approach (JAIDS epub), refers to the clinic in Mbarara, Uganda, where many of the participants in our studies are seen for routine clinical care. Geng ...
2:37 PM, 21 Oct 2009 | Permalink