ChildCount.org

Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away.

Antoine de Saint-Exupery on Engineering

French writer (1900 – 1944)

When I arrived at the Sauri Cluster, one of fourteen clusters of villages across sub-Saharan Africa that are part of Columbia University’s Millennium Villages Project (MVP – www.millenniumvillages.org), I had some ideas for programs aimed at patient tracing to aid Prevention of Mother To Child Transmission (PMTCT) of HIV, having carried out a similar research project in Malawi. While the team I was a part of wrote up our findings, I focused specifically on our recommendations for a patient tracing software that could be developed for mobile phones. These recommendations for Malawi were based on one major factor in the failure of PMTCT programs: a high rate of defaulters accessing the services provided, due to lack of follow-up by community health workers (CHWs). We had some good ideas about how mobile phones could be used to drastically lower the rate of loss to follow up (the technical term for defaulters). However, when the programmers found that creating a persistent database of patients for tracing was too daunting, given the timeframe of our project in Malawi, the tracing system was drastically scaled back.

The opportunity came to try and implement a mobile-based patient tracing program for PMTCT in Kenya with the Millennium Villages Project in the Sauri cluster. The timing was perfect: in September 2009, UNAIDS and MVP signed a memorandum of understanding to establish MTCT-Free Zones within the MVP sites. It reflected a shared commitment to bring together MVP’s multi-sectoral and science-based development and primary health care strategy with UNAIDS’ expertise in community and family centered PMTCT, along with greater involvement of people living with HIV. In the “MTCT-free zones,” coordinated application of rights-based and evidence-informed best-practice in PMTCT could be amplified by community engagement and support, to provide access to high quality services that meet WHO guidelines and to overcome social barriers to the uptake and continuation of PMTCT regimens. Mother-to-child transmission of HIV can be virtually eliminated through a four-prong strategy implemented simultaneously:

Prong 1: Preventing women of child-bearing age from acquiring HIV infection;

Prong 2: Preventing unintended pregnancies among women living with HIV;

Prong 3: Preventing HIV transmission from women living with HIV to their children;

Prong 4: Provision of care, treatment and support to mothers living with HIV, their children and their families.

In low resource settings, implementation of this comprehensive response faces a host of well-documented obstacles, particularly due to low levels of coverage with existing reproductive health services, including antenatal care and health services for newborns. In addition, in many settings, systems for tracking mothers and providing them and their families with adequate support throughout the continuum of care are poorly developed. With this timely focus on PMTCT services across all MVP sites, and the added support of UNAIDS, the implementation of a mobile-based tracing software for PMTCT was both needed and well-timed.

The only problem was that the software which had been written for the earlier project in Malawi wouldn’t work for a comprehensive patient tracing system. While in New York, I talked to the ChildCount+ team about these challenges, and found that by leveraging the existing CC+ platform, we had the benefit of a well-tested persistent patient registration system, which was the missing vital component of the previous project, as well as forms and training materials which had already been developed. With the underlying framework already in place, we could focus our efforts on building the module for PMTCT.

I will first explain the design process and then the final product.

The Design Process

When I arrived in Sauri in mid-June for my first day of field research, I had in mind a system where the mothers could be enrolled in a certain treatment program. This would have a set schedule for the follow-up appointments, and the mother would be assigned to a particular treatment program at the initial antenatal visit based on her CD4 count. The initial idea is in the diagram below.

I decided that the best way to figure out what was feasible was to visit each health center and talk to the stakeholders: the CHWs themselves, the nurses who saw the pregnant women and the under 5s, and the patients. They were the ones who know what the problems are and what the solutions should be. They were also the only ones who could tell me if the program that we had thought up while sitting at a desk in New York was a good idea, or a bad idea. I stressed this whenever I talked to people, that the program as it existed was only an idea and that I was counting on them. If it was a bad idea, please tell me and we could change any part of it. At first they laughed, but after they realized that they had the opportunity to influence the program, they were very open about changes that needed to be made and had no trouble telling me where the ideas I had arrived with were not compatible with the reality on the ground.

When I visited the health centers, I found that while in an ideal world, the idea that someone could be put on a set schedule and that would just work was nice, it just wasn’t feasible. Some health centers had all of the HIV+ pregnant women coming in on the same day of the month; sometimes, when someone would be travelling, the nurse would shift an appointment; and, if there were less than 4 months left before a woman gave birth, they would come in more often in order to get in all of their antenatal appointments. Having the schedule set on CD4 count was also problematic as the CD4 count wasn’t available during the initial antenatal appointment. Also, the Maternal and Childhood Health Nurses said it was a bad idea. So, with that, the idea of a set schedule was scrapped for a more flexible system of setting appointments manually.

We knew that we had to strike a balance between collecting enough information to make the impact that we wanted and overwhelming the nurses and CHWs with extra work. So, we focused on only collecting information that would be used to initiate a corresponding action. For example, we included the estimated date of delivery for each pregnant woman in order to have their CHW visit them 3 weeks in advance of their due date to go over their personal birth plan with them. We decided to collect data in a chronological order, having one form which contained information which could be collected before the initial antenatal visit filled out by the CHW in the home of the woman, then another form for the nurse to fill at the initial antenatal visit, and finally one to be filled on follow up visits. These were the only forms that collected any data, but we had several additional forms for procedure, like one for CHWs to fill when they reminded people of their appointment, and one for the clinic front desk to fill when someone came for a clinic visit. The new workflow is pasted below. You don’t need to know much about the forms other than that there were too many of them. This would have required 3 new forms at the clinic, and for CHWs to carry and fill out two more forms on their household visits; this was in addition to those that they already had to fill out and had indicated to me in a focus group that were taking far too much time. The only pre-existing form in this workflow was form A, and new forms in the workflow have been circled.

I held a focus group with CHWs who expressed concerns that any new module would lead to a lot more work, and that they were already filling out forms which could take hours when visiting a new home. They asked me to be sensitive to the workload they already had and to add as little as possible. At this point I met with Ben Nemser, Andy Kanter, and Yanis Ben Amor from the Earth Institute who challenged me to find ways to reduce this excessive form-filling. (When you are too close to a project you tend to think of all the reasons you came up with a form in the first place, I encourage all project managers to seek out fresh eyes to help pare down unnecessary steps.) We managed to cut out forms R and V all together by utilizing the household visit form (already being filled out by CHWS), and which has a referral section to replace the reminder. This was great, because the CHWs were already making a household visit to fill out the form at least monthly, so we were also not adding any extra visits to the household unless the person did not attend their appointment. We replaced form V by counting a clinic visit when there is a new appointment made (which makes perfect sense, as there will be a new appointment following any visit). We also shifted the form P (now FP) to be filled out at the front desk of the Maternal and Childhood Health unit by the CHW on duty. This reduced the number of new forms for the CHW to carry and fill out to zero and kept the number of new forms in the clinic at three, however, now all the forms contained information and none were kept for only the sake of procedure. In addition, we were able to identify information which CHWS had been collecting manually on the household visit form that could be replaced by the system, reducing the length of the form by 50% and saving CHWs a significant amount of time filling out forms each month.

Below is the final workflow.

When presenting this new module to CHWs, they appreciated very much that the attempt was made to not add any extra household visits or forms to fill and there was very little resistance to the implementation.

They felt as though their concerns were heard and the feedback that they had given during the planning stages of the project had been taken seriously. When presenting the forms to the nurses and clinical officers in the field, we received a nearly unanimous response: this program will greatly reduce the number of defaulters from care (loss to follow up) and that it should ensure that all children in the cluster receive their immunizations on time.

The PMTCT Module

A community health worker (CHW) visits every pregnant woman and every child 18 months and below with an upcoming appointment to remind them when to go to the clinic and makes a follow up visit if they don’t attend. We are specifically targeting one of the biggest challenges to PMTCT programs: loss-to-follow up, which are those who begin the PMTCT program but end up not attending their clinic appointments. Our goal with the reminder/tracing system is to eliminate those who default casually (forgetting an appointment, etc), and to identify those who are consciously making a decision not to attend so we can follow up with them. The issue of default is very sensitive in PMTCT. The pregnant women are given just enough medication to last until the next appointment, and if the drug regimen is adhered to for the entire pregnancy there is an extremely good chance that the baby will be born without HIV. Without any preventive measure around 25% to 40% of babies born to infected mothers will receive the virus from their mothers. Of these infected babies, without treatment, one-third will have died before their first birthday from an HIV-related cause, and half before their second birthday. With patient tracing, pregnant women get treatment, preventing children from contracting HIV and any children who are born HIV+ are put on the treatment plan that they need as soon as possible.

The CHW does not know the patient’s HIV status at any point. To avoid the problems of stigma against people living with HIV, all pregnant women and children are followed. The messages sent to the CHW are exactly the same if the patient is HIV+ or – and all patients are followed up in the same manner. Because of this, we not only protect the patient’s confidentiality but we also experience positive side effects, which I like to call “collateral benefits.” These include the increased number of women attending at least the recommended 4 antenatal clinic visits and more children receiving all of their vaccinations at the appropriate time (as all vaccinations are complete by 18 months).

We begin the process by collecting some demographic information on form FP (Facility: Pregnancy) at the front desk of the Maternal and Childhood Health Unit of the Health Center for each new pregnancy. On the first visit to the MCH unit, this information is collected by the community health worker stationed there.

Once this is complete, they go into the nurse’s office for the initial antenatal visit. Form FP 2 (Facility: Pregnancy 2) is then filled out.

This information is used for 2 separate interventions: the first is to remind women of their follow up antenatal appointments, and the second is to prompt CHWS to go over personal birth plans with pregnant women, which we hope will raise the proportion of births facilitated by a skilled birth attendant and the proportion of births which take place at a facility.

Three weeks before the Expected Date of Delivery, the CHW assigned to this pregnant woman will receive a text reminder to visit the woman and go over her personal birth plan with her; this includes things like where she will deliver and how she expects to travel there. Again, we hope that this will increase the number of births by skilled birth attendants (usually this means in the health facility).

3 days before the Date of Next Appointment, the CHW assigned to this woman will receive a text message saying “NAME has an appointment at the health center on DATE. Please visit the client and refer to the clinic for their appointment,” then, the CHW will visit the client in her home (as already happens at least once a month) and fill out form C (Consultation form) (already in use). The last field of form C is a referral, which serves as the record that the CHW has reminded/referred the client to the health center for her appointment.

If the client returns to the health center for her appointment, form FP3 (Facility: Pregnancy 3) is filled out by the MCH nurse.

As you can see, the follow-up visit form is very simple: just the next appointment. If the client doesn’t come back for their appointment for two weekdays after the date of appointment, the CHW receives another text message stating that “NAME has not attended their appointment on DATE. Please visit them and advise them to visit the health center as soon as possible.” If two weekdays pass from the date of this reminder, the client is added to a list of defaulters.

When a newborn child is registered in the system, any outstanding appointments for the mother are removed, and a new appointment is made for as close to the child’s 6 week birthday as possible. The same patient tracing/defaulter system is in place for the child; when the child reaches 18 months and graduates out of the following program the system automatically removes them from the reminder cycle.

There are additional features to the program, such as an automated way to inform patients that test results have been returned and the ability for the lab to send results electronically. All of the forms can be entered in via either mobile phone or online form. The HIV status of all clients is kept confidential and the CHW is never aware of the details of a client’s status or other medical history.

There is also a simple module built on the same ideas for tracing Tuberculosis patients, ensuring that they complete their course of treatment.

Nine health facilities in the Sauri Millennium Village Cluster have been trained on this module, the last being completed in early September. So far 39 pregnant women have been enrolled in the following program with more being added each day.

I’d like to thank the Childcount+ team in New York for challenging my ideas and helping them get to their best possible versions, Yanis for taking on the never enviable task of trying to get all the resources in place and meeting with me week after week to iron out any and all challenges, Dickson and Moses in Kisumu and Kevin in Nice for all the help getting the software running and putting up with my constant changes, James in New York for helping me see that the program needed to expand from its origins, Jessica and the Millennium Villages Project team in Kisumu for all their support, Komolo for showing me around to all the health centers and using his amazing network of contacts to get me meetings with everyone, working with me on the trainings and being an all around indispensible right hand man (and for making sure there was mandazi and tea), the health care workers and pregnant women who took time to talk to me and answer my never-ending questions and last, but certainly not least all the Community Health Workers, without your hard work, none of this is possible. We will keep the community updated on the results of the project, stay tuned, this is not the last you’ll hear about this.