The results are presented according to each RE-AIM domain and subdomain (Table 1). These have been somewhat reordered to facilitate logical presentation of the results.
Reach
According to the household access survey, it was estimated that over one fifth of adult Syrians in Irbid governorate had at least one self-reported relevant NCD (21.8% of 8041 surveyed adults aged 18 or over); applying this proportion to UNHCR-derived figures, 60,041 adult Syrian refugees >18 years resident in Irbid governorate in 2017 had an MSF-targeted NCD, of whom 5.9% (n=3531) were reached by the MSF programme (35).
Retrospective data from 5045 patients enrolled during the study period (3664 Syrians, 1365 Jordanians and 16 refugees of other origins) were analysed. The Irbid cohort represented a group of older patients [mean 54.7 years (SD 15.7)] with multi-morbidity and relatively high rates of self-reported disability (9.9%). The majority (59.8%) were women and 71% (n=3582) had two or more target NCD conditions, with hypertension (60.4%), type 2 diabetes (53.1%), cardiovascular disease (25.9%), hypothyroidism (7.6%) and asthma (7.0%) the most commonly treated conditions (Supplementary material S4). These findings are consistent with the household access survey, which reported a similar prevalence of target NCDs and higher prevalence of HTN and diabetes among women. However, this cohort had greater rates of NCD multi-morbidity compared to the adults with self-reported NCDs in the household survey (71% vs. 44.7%). NCD risk factors levels were high at enrolment with obesity levels of 62.6%, self-reported smoking rates of 22.7% and 37.2% self-reporting low or zero regular physical activity (Supplementary material S4).
By the end of the reporting period, only 24 patients had received a formal diagnosis of a comorbid mental health condition, while 154 patients attended individual counselling sessions and 66 group sessions were held in 2016, when recording began. The total number of patients reached by the MHPSS service was not captured by routine data, while qualitative data highlighted staffs’ perceptions that an enormous burden of MHPSS needs was going undetected at the time of the study.
Adoption/acceptance
Accessibility/ acceptability: Our qualitative data showed that most patients considered the MSF clinics physically accessible, in terms of distance and convenience, availing of public, taxi or private transport. Syrian respondents, in particular, reported carefully balancing stretched household finances, spending on aspects of the service they valued and procuring preferred medications from multiple sources. They prioritised expenditure on transport costs for medical consultations over those for MHPSS, health education or laboratory visits. Patients generally appreciated the appointment system (which minimised long waits and prevented the perceived favouritism they experienced in the MOH system) and the SMS (short message service) reminders, but most perceived it as rigid, and inaccessible outside of prescribed appointment times. Staff strongly encouraged appointment adherence, achieving a 90% adherence rate.
Qualitative data confirmed that Syrian community members had limited access to alternative, affordable primary level NCD services in north Jordan. The MOH and a limited number of other NGOs provided such care but patients incurred co-payments or travel costs to attend, which many reported as unaffordable. They coped by selectively forgoing medications, sharing with family or neighbours or purchasing from private pharmacies:
Syrian patient: “If there is a family that can’t bring medicine, we collect pills from here and here, so people help each other ... because there is extra. So people give each other. I know a kid who takes insulin…I give people. I’m forced to help people.”
Staff perceived that most Jordanian patients, who made up 27.1% of the cohort, did not meet vulnerability inclusion criteria since they could access alternative services via national or military insurance. This was the case for all interviewed Jordanian patients.
Respondents perceived the referral pathway for NCD complications or other conditions, overseen by UNHCR and their local implementing partner, Jordan Health Aid Society (JHAS), as opaque, inaccessible, unaffordable and inadequate. UNHCR funded a limited number of referrals to MOH secondary or tertiary services via JHAS, who had a gatekeeper role, and whose decision-making process was seen as inconsistent and lacking clear criteria. MSF clinical staff also reported frustration at the lack of information returned to them post-referral. Stakeholders concurred regarding the complex and unsatisfactory nature of this referral pathway. MSF had successfully brokered agreements with other NGOs to meet some referral needs, e.g. interventional cardiology free-of-cost to patients, but this was often for a defined project period only and was limited by short-term funding cycles. MSF and other stakeholders’ proposed solution was to encourage other international actors to provide funding and services to fill referral gaps.
Acceptance: Our qualitative data also showed that the programme was highly acceptable to patients, staff and stakeholders. Patients reported they received good quality care in a caring and respectful environment. They valued the free medications, regular laboratory and vital sign testing most highly. A female patient stated:
“(MSF is) honestly caring about the patient, caring about his appointments even the medication availability. We have never come here and told us that the medication is not available. Their performance is great.”
Multiple stakeholders believed the programme addressed an important health need among both Syrians and Jordanians and relieved a significant burden on the MOH.
Adoption/participation: There was good community awareness of the programme and a waiting list of over 200 patients to be enrolled. The main barrier to patient participation was the MSF-imposed cap on cohort size. Clinical staff were mainly Jordanian medical and paramedical university graduates, many with previous NGO experience. They were committed, proud to work for MSF and derived satisfaction from observing patients’ improvements.
“I learned here how to see others problems… the disaster they are coming from…how we work here like a team or a family for the benefit of the patients; how you can give to the people…without taking, with nothing in return.” Clinical staff member.
There was low turnover among clinical cadres other than non-specialist doctors, who tended to resign after gaining several months’ experience with MSF to pursue specialist training. This turnover was considered problematic by clinical supervisors, staff and patients who valued continuity of care. Some staff were dissatisfied with the perceived lack of promotion opportunities or job security (given the limited duration of MSF programmes), high workload and six-day working week.
The MSF NCD guideline was largely acceptable and usable by medical staff, who had adopted it particularly as a means to negotiate patient demands. In 2017, the guideline covered most clinical scenarios that doctors encountered, but it had limitations, including inadequate programmatic guidance (e.g. setting the limits as to what “stepped-up” primary care means i.e. “what components are included…that is not clear”; and predicting the numbers needing referral for screening and/or management of complications e.g. laser eye treatment for diabetic patients), perceived promotion of poly-pharmacy (since each condition was treated in a vertical manner) and there was limited guidance on complex, multi-morbid patients with renal impairment or frailty who may have benefited from “de-prescribing”. In addition, clinical supervisors, who were generally of non-Jordanian origin, commented that some Jordanian doctors’ felt the guideline limited their autonomy and offered “second-class” care since it recommended older, generic medications. Jordanian doctors also commented on the limited user-friendliness of the paper-based guideline and their preference for a digital application that could be accessed on their smart phones during consultations.
Patients were largely unaware of either the MHPSS or HLO services provided by MSF. Patient reluctance to attend and an initial distrust from the medical team in the MHPSS service influenced low referral rates, which were partially addressed through multidisciplinary staff training sessions, widening of referral rights to nurses and the introduction of psycho-education sessions aimed at patients in the waiting room.
Effectiveness
The retrospective analysis of routine data allowed exploration of clinical and quality indicators.
Clinical indicators: Among 4729 adult patients meeting our inclusion criteria (i.e. diagnosed with hypertension and/or Type II diabetes and enrolled during the study period), 2912 (61.6%) had hypertension and 2546 (53.8%) had type II diabetes, while 1530 (32.4%) had a dual diagnosis. From the programme perspective, BP decreased among patients with hypertension by 6.86 mmHg since the programme began, from a mean of 137.2 mmHg (95%CI: 134.7 to 139.7) to 131.2 mmHg (95%CI: 129.8 to 132.6) at 6 months, while mean fasting blood glucose similarly decreased from 173.2 mg/dL to 165 mg/dL after the first year of operation. HbA1c control improved markedly during the programme’s life from a mean of 8.7% in month one to a mean consistently below the target of 8% and more than 60% of patients achieving control, after the first six months. From the patient perspective, the proportion achieving blood pressure control improved from a baseline of 63% to over 70% by month 6 post enrolment/new diagnosis. Among diabetic patients, there was a marked improvement in FBS level from mean 187.5 mg/dL (95%CI: 184.0 to 190.9) at enrolment/new diagnosis to 160.7 (95%CI: 155.7 to 165.7) by six months and over 70% of patients with DM II achieved FBS targets by month four. These results and loss to follow up are elaborated on in our companion paper (28).
Quality indicators: Additional clinical outcome indicators and process indicators are presented in Table 2. At each health education session patients were asked to categorise their exercise level as active, inactive, moderately active, and moderately inactive but exercise was not otherwise quantified. Similarly, we could not determine whether smoking behaviour had changed since it was not quantified and patients’ self-reported smoking behaviour change was only recorded relative to their previous visit. Attainment of certain process indicators was suboptimal e.g. statin prescribing, CVD risk scoring and performance of annual urinary protein testing in diabetic patients.
Perceived effectiveness: Qualitative data confirmed that both staff and patients perceived the programme as effective, while staff observed greater improvements and commitment among Jordanians versus Syrian patients. Patients reported feeling better after attending the programme, linking this both to physical improvements and to the reduced financial burden and worry around obtaining their medications.
Jordanian patient: “(Since coming to) the clinic to be honest, I feel relieved and comfortable since the first day I came here, I felt the difference in my disease, before I used to take pills for diabetes and hypertension but nothing changed.”
Implementation
Fidelity of programme delivery: The indicators derived from our analysis of routine data that are related to fidelity of programme implementation are presented in Table 2. Our qualitative data highlighted the impact of war and the refugee experience was the key challenge to implementing and maintaining effective NCD care for the Syrian refugee population. This theme was explored in detail in our linked paper (27). Syrian patients’ social suffering had profound implications for their ability to engage with the programme in terms of medication adherence, dietary and lifestyle advice, and affordability of access:
“The hypertension goes high not all the time but when I get sad and remember my sons in Syria and they tell me what happens with them I keep crying and crying then my hypertension goes high or goes down I don’t know then I take a hypertension pill to settle down whenever I read some news about them,” Syrian patient.
Space, patient transport costs and limited patient engagement were barriers to implementation of group MHPSS sessions. Clinicians and their clinical supervisors described a didactic and knowledge-based approach to individual patient education rather than the preferred patient-centred approach. Additional challenges encountered by the team around healthy living and behaviour change included: cultural dietary and exercise norms (high fat, high salt diet and low habituation to exercise for health or leisure) and acceptance of smoking (especially in men), the obesogenic environment (with highly available calorie dense foods and lack of exercise options) and most patients’ perception and expectation that medications would provide the necessary solutions.
Adaptations: The programme adapted dynamically to patient and programmatic needs. Staff learned to adapt health education messages to patients’ literacy and education levels and their financial means and to involve family members as informal treatment supporters. Other essential components and adaptations to the programme included the initial introduction of MHPSS in response to significant identified mental health needs among Syrian patients; the introduction of the HLO social work role to address social and protection needs (although this was reportedly underutilised); and the expansion and reorientation of the MPHSS service to provide ad hoc psycho-education sessions, a targeted group ‘living well’ programme and peer-support groups. However, the team reported a lack of good quality onward referral options for patients requiring prescription of psychotropic medications or psychiatric input and management staff planned to train one family medicine specialist to address this need. The team also established essential referral pathways where possible and introduced and expanded the home visit service (to encompass a wider radius and more staff). Similarly, clearer admission criteria related to patient vulnerability were piloted and operationalized.
Costs: The total annual financial cost of the NCD programme from the provider perspective increased annually in parallel with greater patient volume, increased service complexity and the addition of specialist staff. It increased by 52% from INT$ 4,206,481 in 2015 to INT$ 6,400,611 in 2016 and by a further 5% to INT$ 6,739,438 in 2017. Per patient per year cost increased 23% from INT$ 1,424 (2015) to 1,751 (2016), and by 9% to 1,904 (2017), while cost per consultation increased from INT$ 209 to 253 (2015-2017). The major cost drivers were human resources (accounting for 38.9%-42.6% of total annual costs) and medications (34.8-43.2%). The costs are reported in detail in a related paper (26).
Maintenance
Individual level: The majority of patients enrolled during the study period (N= 5045) were retained in care for over six months (85.2%); one third of enrolled patients exited (including 12.5% cumulative loss to follow up and 2.6% deaths) (Table 3). Over half of adherence survey participants (N=300; 50.4%) were prescribed between four and six MSF-provided medications, while almost a quarter (24%) were prescribed over seven (Supplementary material S4B). The majority (60.4%) also took medications from another source. Most patients (89%; N=300) had very high self-reported medication adherence scores. While the majority of individual interview participants (especially Syrians) declared themselves “very committed” to taking medications, several described stopping, taking intermittently or sharing medications with those in need. Qualitative data confirmed that patients’ medication adherence and behaviour change was facilitated by support from family and MSF staff.
Staff perceived that excellent patient-staff rapport; positive experiences of supervision, support and training; and good teamwork with colleagues assisted them with programme implementation. Challenges to maintenance from the staff perspective included: Syrian patients’ war-related trauma, poor mental health and social suffering, as well as their poverty, lower education levels and perceived greater medical complexity compared to Jordanians. Staff and patients both emphasised the negative impact of mental distress on adherence to medications and healthy living advice:
“As I was hearing the stories I thought…this man’s problem is not that he’s smoking too much. His problem is that he … experienced sexual violence, physical violence in prison in Syria… these two are linked.” Clinical staff.
In addition, patients (of both nationalities) tended to visit multiple concurrent providers, which also complicated care delivery.
Organisational level: Qualitative data highlighted the importance placed by MSF on providing a good quality service that fulfilled its humanitarian remit. Multiple respondents emphasised the difficulties the programme encountered around the lack of adequate referral pathways:
“The credibility of any service often depends on its ability to refer upwards, doesn’t it? That is just as true for people with angina and coronary artery disease (as it is) for mental health,”
Management staff.
Proposed task shifting of stable patients’ medical reviews to nurses had occurred in a very limited manner because of lack of clarity generally on clinical activity and patient flow, lack of clear eligibility criteria, and reported resistance from patients and medical staff. In our related paper, we explored potential cost efficiencies that may be realised from reorganisation of medical consultation workflow, identifying the most important factors as frequency of review and proportion of patients categorised as stable, and therefore suitable for nurse review or longer doctor review intervals (26).
Contextual challenges to programme maintenance related to operating within the Government of Jordan regulatory environment, including: the requirement that medications must be locally purchased rather than imported; the lack of single focal point or set of regulations governing NGOs; and significant bureaucratic delays. Finally, there was perceived tension between the MSF team’s desire to continually improve and add complexity to the programme and the need to consider long-term planning for the programme with a potential future handover. Several management staff discussed the need to engage with the MOH as the likely handover partner but pointed to the gulf between the current MSF and MOH models of NCD care.
Management staff referred to the internal debate within MSF as to whether NCDs fall under the remit of a humanitarian medical organisation:
“An NCD Programme is a relatively recent departure for the MSF and it is getting very close to the dividing line between humanitarian and development aid. So, what actually is MSF’s direction here, I think, partly is driven by the general sense of the humanitarian community that NCDs are an epidemic and need to be dealt with, but I am not sure we have … view of how this should be managed...” Management staff.
Several of the interviewed MSF management staff, both at project and country coordination level, questioned the sustainability of the current Irbid model. They characterised MSF’s usual approach as providing a relatively short-term solution to a health care gap identified in a vulnerable population, with the eventual aim to either hand over to other actors, to ensure adequate, alternative services are available or to close down if a specific crisis has passed. Several suggested that, in designing future NCD chronic interventions, MSF should engage more closely with the pre-existing health system in order to facilitate a future exit strategy, while one advocated for MSF to build on the HIV service model, by maximising task sharing and decentralisation of care to community level. However, experienced management staff also discussed the MSF rationale for maintaining a vertical programme in Irbid, given that it served as an opportunity to “learn by doing” and to understand the essential components required for NCD care. Participants also acknowledged that operating in the context of a middle-income country with established systems, regulations and policies required a different type of engagement and negotiation with authorities compared with other contexts where MSF has traditionally worked.