The instability following the 2007 Kenyan presidential election is an example of a countrywide disruption that affected all aspects of life including housing, transportation, education, commerce and healthcare. In this paper, we have shown that the AMPATH system maintained a high level of functioning with minimal gaps in service delivery. Prompt implementation of a response system allowed most patients’ uninterrupted access to ART or rapid return to care. We identify lessons learned based on the experience with the AMPATH response and recognize system factors that facilitated the success of that response. We believe that the lessons learned during the Kenyan post-election crisis may be valuable to other ART programs located in regions of the world prone to instability due to political or natural crises.
Lesson 1: A rapid organizational response is critical
Minimizing disruptions to healthcare systems during times of crisis requires forethought, planning and organization. In analyzing AMPATH’s response to the crisis a key factor in the success of the response was the organization’s ability to rapidly establish a multidisciplinary response team. As such, one recommendation arising from our experience is that a Crisis Response Team should be established in advance as part of the structure of large HIV care and treatment programs located in potentially unstable regions. Médecins Sans Frontières (MSF) similarly designates staff to prepare for instability and create an Emergency Preparedness Plan. The Response Team should include representatives from program leadership as well as key departments (i.e. pharmacy, social work, psychosocial, nutrition and outreach). Members should be aware of their designated roles and responsibilities and alternative representation should also be designated. Priorities for patient care should be clearly defined in advance and strategies for restructuring care delivery to compensate for staff shortages, increased patient load, potential supply disruptions and curtailed record keeping should be addressed.
At the time of a crisis, this team should be immediately assembled (in person, electronically, or via phone) and tasked with the rapid assessment of the situation to determine the impact of the crisis on the safety and stability of the communities surrounding the healthcare system, the structural and functional integrity of the clinic(s) and the extent to which displacement has affected staff and patients. These assessments should be used to tailor the team’s response to the crisis, which might include the use of multi-disciplinary field teams to address patient needs, public service announcements (address the identified needs of the specific population)[19–21, 26], activation of an information hotline (utilizing numbers from each major phone carrier in order to minimize problems with patient accessibility), activation of community networks, and establishment of alternative care delivery sites. The developed response mechanism, including key personnel and procedures, should be reviewed and evaluated on a regular basis.
Lesson 2: Clinic autonomy and self-sufficiency is indispensable
The ability of the AMPATH clinics to function autonomously in the early period of the crisis was key to the success of the systems response because of disruptions in transport, communications, and issues related to safety. This finding supports the recommendation that within a multi-clinic system the capacity for some autonomous operations, for at least limited time periods, should be developed in the individual clinics. This would include ensuring the staff members are cross trained to take additional or alternative roles[19–21], maintenance of at least one month stock of supplies and medications, and ensuring that emergency funds are available at the clinic to meet unforeseen operating costs. The optimal quantity of medications and supplies to be stocked requires local assessment of patient volume, storage capacity, anticipated duration of disruption(s), and the program’s capacity for alternative means for drug/supply delivery. MSF has made similar recommendations for maintaining increased drug supplies at clinics and distributing extended supplies of patient medications when instability is anticipated as was done by AMPATH practitioners, particularly in Burnt Forest, at the time of the 2007 Kenyan elections and holidays[19, 20]. In times of unanticipated crisis, MSF had recommended the rapid distribution of pre-prepared “runaway stock” (1–3 months of ART and OI prophylaxis) to patients. In addition, clinics should be given license to adjust patient flow and documentation in order to accommodate reduced staff numbers, and in some cases, increased patient numbers. Streamlining of patient visits and the minimization of paperwork as used by both AMPATH and MSF can prevent disruptions in patients’ ART at the reasonable expense of limited data collection.
Lesson3: Patient knowledge of their HIV status and treatment regimen is necessary
Promoting patient self-sufficiency in seeking care is also important. Patients should receive ongoing education to ensure that they are familiar with the status of their HIV disease (CD4 count, viral load, history of OIs) and know the component drugs of their ART regimen. MSF has utilized patient “passports” that include previous and current ART regimens and any known side effects. AMPATH currently prints a “Clinical Summary” for health providers at each patient visit that shows the patient’s current ART regimen and OI medications, current and past problem list and a chronological listing of the patients weights, hemoglobin counts, CD4 counts, viral loads and SGPTs. As this information is already available, it could be provided to patients at each scheduled visit with instructions to keep it in a safe place such that the patient could present it to other providers if needed in the future.
Lesson 4: Patient and community networks are essential
At Burnt Forest the strong ties between the community, including both patients and other community members, and the clinic was found to be invaluable to ensuring that patients received uninterrupted supplies of ART. The recommendation based on this finding is that clinics/ART treatment programs develop significant relationships with their surrounding communities. This could include establishing patient contact lists and facilitating self-forming patient groups to aid communication between the clinic and the community of patients both for the enhancement of routine care and to facilitate rapid communication during times of disruption. Clinics/programs utilizing CHWs or expert patients will want to consider how to incorporate these key members of the community into their crisis response plan. As emphasized previously by MSF, each clinic/program must project an image of neutrality in its healthcare delivery in order to maintain the support of the community, especially when there is discord among community members as was seen in Burnt Forest.
Lesson 5: Ensuring staff security and access to basic needs is fundamental
The Kenyan post election violence impacted the full cadre of staff working in the AMPATH clinics. Staff members experienced displacement, food insecurity, transportation issues, and security concerns. By addressing these issues AMPATH facilitated the return of staff to functional positions within the system. For individuals in targeted ethnic groups in high risk areas, that included relocation to positions outside their primary clinic. For others it required finding alternative housing, providing funds to replace damaged or stolen property, or to buy food. This finding supports the recommendation that programs make arrangements for rapidly assessing staff needs and then provide with a plan to meet those needs either internally or by establishing agreements with organizations whose missions can provide for those needs.
Lesson 6: The ability to rapidly identify and track all patients requiring ART is key
At Burnt Forest, in particular, many patients were unable to access the clinic due to security issues and/or displacement. The AMPATH electronic medical record (EMR) system was a critical tool in identifying patients who had not returned for their scheduled appointments and providing a list of patients that were at risk for ART disruptions. This finding leads to the recommendation that programs should have the capacity to rapidly generate lists of patients who have not returned for care within a specified period of time. In the case of ART care systems/clinics with EMRs this means that the programming to generate such lists should be developed prior to a potential crisis and be ready to be utilized when needed. In the case of programs without EMRs, paper based systems for identifying patients at risk for ART disruption should be developed and tested. However, generation of such lists is only useful if the program maintains up to date contact information on all patients. Contact information should include cell phone number(s) if available as well as the patient’s complete address. In countries where street addresses are uncommon a detailed description of how to get to a patient’s house including a map should be maintained in the patient’s file. Decentralization of EMR capacity to local clinic sites would improve the ability for individual clinics to track patients, verify treatment regimens, and monitor treatment locally. AMPATH is currently working to decentralize EMR operations to all clinics.
Some clinic systems have designated staff (outreach workers, trackers, etc.) and systems for locating patients who have not returned to clinic within a specified time frame. These individuals should be included in a crisis response team and provide insight into the development of plans to contact patients at times of social disruptions. Clinics without this capacity should identify and train staff and develop systems for providing outreach to patients who are at risk of ART disruption. This may include the use of patient networks as described above.
Although the majority of patients were able to return to care in a timely manner, this case study is limited in its ability to quantify the actual proportion of patients that had interruptions in care. In order to determine this number, all patients actively receiving ART prior to the elections would need to be known and then compared against a list of all patients returning to care within a specified amount of time following the elections. However, patients’ scheduled return dates were not consistently recorded in their charts and may have varied during this time due to the holidays and upcoming elections. Patients may have returned to clinic before or after their scheduled date with an unknown interruption in ART. Furthermore, due to the disruptions in some clinics, only pharmacy logs were kept to record ART distribution with no formal encounter form filled to log a patient’s visit. Furthermore, preparations for the holidays and elections affected the average number of days that drugs were given for patients at the time of the election. Drug refills for patients normally attending the most affected clinics is difficult to determine as patients were able to receive refills at other clinic sites, ancillary clinics, non-AMPATH clinics and IDP camps.
In the two months following the election specific documentation of the location and status of 50% of the patients on ART at Burnt Forest was recorded. Individual patient tracking was then discontinued and specific patient level data is not available to allow for a true assessment of how many of the remaining patients were lost to follow-up due to the crisis. An analysis of aggregate data from AMRS of the patient cohort starting ART prior to the crisis was done. It showed that during the two years immediately following the crisis the rate of patients lost to follow-up at Burnt Forest was greater than AMPATH in its entirety (where the majority of patients were not affected by the crisis) by 3-4% but equalized by 2010.
This case report highlights the response of the AMPATH HIV clinic system to the crisis following the 2007 Kenyan elections. We acknowledge that the experiences and responses documented here may not be applicable to all settings or crisis situations.