Between July 2016 and June 2019, 3094 patients were registered in the programme (188 with DM only, 2473 with HTN only, 433 with both DM and HTN). We describe below our experience of setting up the programme along the lines of our conceptual framework of enablers.
1. Decentralization
Currently, according to the Zimbabwean national policy, diagnosis of DM and HTN is under the responsibility of doctors at hospital level, while refills of selected NCD medications can be given at PHC level. The programme did not aim for coverage of the entire health district, but rather to introduce DM and HTN care at a selection of 11 PHC sites in line with available resources. The site selection process was done together by MOH and MSF. First, we excluded clinics in urban settings and hospitals other than the referral sites. Then, using a quantitative and qualitative evaluation process, rural clinics were scored and ranked according to a set of inclusion and exclusion criteria described in table 3.
Table 3: Criteria for site selection in Chipinge District
Inclusion criteria
|
|
Exclusion criteria
|
Larger size of population served
|
Number of nurses < 2 per clinic
|
Larger size of ART cohort
|
Driving time from district capital to health facility > 90 minutes (one-way)
|
Higher number of documented cases of NCDs
|
High turnover of human resources
|
Higher number of nurses/clinics
|
Proximity to other possible selected clinics
|
Stronger recommendation of MOH
|
Receiving support from other NGOs
|
At the end of 2018, two PHC sites struggling to adopt the services were dropped from the pilot, while two with satisfactory results but a very remote location were handed over to MOH .
2. Integration
In the Zimbabwean ART programme, some health facilities have vertical services exclusively for HIV positive patients, while in other sites care for HIV patients is integrated within the day-to-day general outpatient department (OPD). Thus in our programme, DM and HTN were either managed in an integrated chronic care clinic (ICCC) where patients with HIV and/or NCDs receive care in the same service, or merged with the general OPD [18]. Our aim was to search for the ‘best fit’ model for service integration for each facility. Regardless of which model was used, when PLHIV had DM or HTN, files would be merged and appointment dates synchronized so that both conditions could be treated at a single visit.
Two out of 9 PHC facilities with existing separate HIV care added and merged DM and HTN management. In 7 out of 9 health facilities the general OPD was the underlying platform for integration. This yielded acceptable results to both staff and patients without compromising existing services. One of these facilities subsequently attempted to move HIV and NCD patients away from the OPD into an ICCC. However, this stretched pre-existing human resource shortages and the coping threshold was reached when the cohort increased in size forcing them to revert to integrating HIV and NCDs into the general OPD.
At the two secondary care facilities, HIV and OPD services operated independently in separate departments at the inception of the pilot. Our initial approach was to introduce DM and HTN services within the general OPD for HIV negative patients, while separately supporting the HIV department to diagnose and manage DM and HTN among their existing ART cohort. However, at SPMH an overwhelming number of DM and HTN cases began to compromise OPD services. Given the smooth functioning of the HIV department, we decided that this would be a better site to host all DM and HTN care. By building additional clinic infrastructure, adding human resources and optimizing the organization of existing services within the HIV department, we established a well flourishing integrated chronic care clinic (ICCC). CDH continued to operate in two different sites according to HIV status, due to HR challenges and infrastructural constraints. Organisation of DM and HTN care in this hospital did not achieve the same level of service delivery as at SPMH.
3. Simplification
Learning from the ART programme which used simplified clinical guidelines and treatment algorithms to ensure safe use at PHC, we developed standard operation procedures (SOPs) for clinical management, adopted from evidence-based international guidelines, MSF guidelines and experience from projects elsewhere in SSA, aligned with the national medicine formulary [19]. Context adaptation was a dynamic process involving regular reviews with technical support from internal MSF and external specialists, and feedback from the clinicians using the protocols and patients.
An example of simplification is with regard to treatment targets for HTN. Initial targets were looser for elderly patients (age > 65 years) in order to reduce the risk of iatrogenic hypotension. Subsequently, as we observed poor target achievements overall with clinical inertia being a possible significant factor, we simplified protocols by choosing a single target of 140/90 for all age-groups.
4. Mentorship and Task-shifting
We adopted a step-wise on-the-job mentoring approach which emphasized knowledge, practical skills and behaviour. The mentors were provided by MSF in three categories: nurses, pharmacy technicians and doctors. The nurses were qualified Registered General Nurses (RGN, three-year diploma) trained in mentorship, with practical experience in mentoring HIV care. The pharmacy technicians had a four-year diploma training in dispensing and pharmacy management. The doctors were general practitioners with significant experience in chronic disease management. All mentors underwent training in DM and HTN care using context-adapted guidelines and on-the job training by MSF supervisors. Mentees from the MOH comprised of RGNs, Primary Care Nurses (18 months training), nurse aides and primary counsellors (lay cadres with ordinary level training and a 6-month training on counselling HIV patients). A mentoring curriculum on organisation of services and DM and HTN care was developed and linked to an evaluation grid to score competencies (the competency dashboard). Mentees would receive an initial theoretical training (one day), then, according to an established schedule, two mentoring teams visited the health facilities every one to two weeks. This on-the-job mentoring cycle was intended for three months, after which mentees would graduate and provide DM and HTN care with periodic on-site and off-site decision support (task-sharing). In certain instances, an extra nurse would be added to the mentoring team to free the MOH nurse undergoing mentorship from her/his usual tasks. Mentoring involved (a) on-site group meetings with all MOH staff with case discussions and lectures on related topics, (b) side-by-side clinical decision support and/or counselling, (c) practical demonstrations of efficient service organisation, including organisation of patient flow and spacing of appointments, and pharmacy management practices. Review meetings were organised in clusters twice a year to analyse performance for a group of health facilities, and to exchange experience.
This mentorship approach allowed a small, mobile team of mentors to set up services in multiple sites simultaneously, thereby accelerating the provision of DM and HTN services and standardising practices across eleven sites. Three sites (one hospital and two PHC clinics) with the highest potential were given a more intensive mentorship schedule and developed as model sites. When sufficient technical capacity was built, patient appointments were spread throughout the week, rather than being clustered on the day of the mentorship visits, and mentorship time was decreased.
7 out of 9 PHC facilities and one hospital achieved sufficient competency-dashboard scores to diagnose, initiate treatment and monitor DM and HTN patients. However, even when dashboard indicators were achieved, the desired knowledge and competencies were attained over a longer period than expected. Challenges for mentors were long travelling time which decreased the daily mentorship coverage or high DM and HTN patient volume on arrival which limited the mentorship time. As noted above, two very remote sites were handed over earlier than planned after having achieved basic skills, in order to free up mentoring time. Due to general human resource (HR) shortages, MOH faced difficulties in freeing up dedicated core staff for regular mentoring, preventing some nurses from completing a full mentoring cycle. To address the problem of staff shortages and high workload, we hired additional staff on short contracts as a temporary solution. We felt a major reason for failure to achieve mentorship targets was lack of clear programme leadership and poor staff ownership of their additional duties. We noted that when good leaders emerged and staff were motivated and willing to be mentored, positive results were achieved, and conversely.
5. Affordable medicines
To standardise treatment, medications were chosen, titrated upwards and different classes added stepwise in consecutive consultations based on clinical and laboratory results. Choice of medications was a compromise between effectiveness, availability, affordability and user friendliness informed by the WHO essential drug list [20], the Zimbabwe national formulary [19] and international guidelines [21, 22]. Before the start of the pilot, NCD medication supply to health facilities was inadequate and inconsistent, with frequent stock outs. Patients received less than a month’s supply and had to come more frequently to top up their prescriptions, or even had to buy medications privately. Patients reported taking suboptimal dosages, or not taking treatment at all until it became available. Thus for this programme, medicines and laboratory reagents were largely provided by MSF, and based on MSF principles, they were given for free to all patients registered. The reliable medication supply allowed us to offer three-monthly medication refills for stable patients. A component of advocacy was embedded in the pilot to lobby at facility, district, provincial, national and international levels for resource mobilization for DM/HTN medicines.
However, providing free medications in our pilot proved a pull factor attracting patients from neighbouring districts and provinces. The worsening economic situation in late 2018 further exacerbated this problem as patients who were previously able to afford to buy their own medications were no longer able to do so and turned to MSF-supported clinics for subsidised drugs. This meant that the MoH was not able to increase its responsibility for medication supply, maintaining dependence on MSF.
6. Equipment and quality assured laboratory support
Our programme took advantage of existing MOH staff, infrastructure and equipment. Existing conventional laboratory-based biochemistry machines were used to measure creatinine, blood glucose and HbA1C at hospital level. Despite reinforcement, the existing sample transport system proved to be inefficient as sample transport time was as long as two to three days from the farthest clinics to the hospital laboratory, interfering with the quality of the samples. The laboratories became overloaded and the result turnaround time was long. We then opted for handheld glucometers for diagnosis of DM, and in 2018 introduced point of care (POC) machines for HbA1C (Fine Care®) and creatinine (Novastart®) measurement. All HbA1C testing platforms were enrolled in a monthly External Quality Assurance (EQA) scheme. Table below chronicles the simplification process of the SOPs for diagnosis of DM.
Table 4: Simplification process of SOPs for the diagnosis of DM
|
2016
|
2017
|
2018
|
2019
|
Diagnostic approach
At site level
|
No diagnostic devices used at site level
|
FBS/RBS on hand-held glucometer
|
FBS/RBS on hand-held glucometer
|
FBS/RBS on hand-held glucometer
And HbA1c on Point of care platform to confirm
|
Conventional laboratory (hospital based)
|
Lab-based FBG
|
Confirm with lab-based FBG
|
Not applicable
|
Not applicable
|
Lab-based HbA1c to confirm
|
Lab-based HbA1c for final diagnosis
|
Lab-based A1c to confirm
|
Not applicable
|
Rationale
|
Dependency on sample transport (ST) and stability
Lab overwhelmed
Delays in result reception
|
Same dependency on ST
Lab workload improved
Still delays in result reception
|
Same dependency on ST
Lab workload improved.
Still delays in result reception
Need for motivated HR to follow quality control procedures for HbA1c
|
Increased autonomy at PHC
Faster decision making
|
We found that the use of conventional laboratory machines for HbA1c needed motivated human resources to follow the quality control procedures required. Performance improved after the introduction of POC devices. DM monitoring with HbA1c is a fairly new concept in SSA where there may be additional sources of error due to haemoglobinopathies or malaria [9], or high HIV prevalence [23, 24]. However, HbA1c measurement was a game-changer against the inconvenience of repeated blood sugar measurements and the poor correlation of these with good glycaemic control.
7. Patient empowerment
This concept involved enabling patients to acquire the knowledge and skills to understand and take responsibility for their own health. Individual and group counselling sessions for DM and HTN emphasised knowledge about glycaemic and blood pressure (BP) control. An active decision not to prioritise defaulter tracing was taken as resources were scarce and we considered there was no public health danger, contrary to contagious diseases such as HIV or TB. In the long run, we aim to differentiate services according to the needs of specific patient subgroups i.e. differentiated service delivery (DSDs) models. At the time of writing this report, DSDs are emerging at community and health facility level.
8. Dedicated monitoring and evaluation (M&E)
Before the programme started, there were no individual patient files for NCD patients.
Health facilities used improvised registers, which did not allow recording of information on follow up and treatment outcomes. There was likely substantial under-diagnosing and underreporting of NCDs.
We designed medical records inspired by the ART patient files and provided them at all sites. The files accommodated both identification numbers for HIV and DM/HTN to enable health workers to identify patients with co-existing conditions and to synchronize appointments. A set of indicators for monitoring and evaluation, following the standard cohort approach used in HIV/TB control programmes, were defined to measure service provision, case-enrolment, follow up, treatment results and retention in care. Data from patient clinical records were entered onsite or offsite into an electronic database by trained data encoders and were analysed during quarterly review meetings with mentees and managers for use in programmatic decision making. Further support for data evaluation was provided intermittently by MSF technical referents.
With specific and detailed patient records in place, nurses were able to provide improved longitudinal follow up for patients as well as quantifying DM and HTN service demand and medication needs for the programme. In some sites incomplete data was a challenge, and this increased as the cohort sizes grew. A simplified chronic patient card with only key parameters for clinical decision-making was therefore implemented in late 2018. We observed that minimizing the number of variables and storing the patient records close to the consultation area increased completeness of data.
9. Referral system
The focus of this programme was on empowering nurses to manage DM and HTN, thereby minimising referrals to secondary care. Where the management of complex cases exceeded the limits of the care provider, context-adapted criteria were developed to identify these patients in a timely way for consultation by a medical doctor on or off site.