AMAFRICA, a patient-navigator program for accompanying lymphoma patients during chemotherapy in Ivory Coast: a prospective randomized study

The treatment of malignant lymphoma (ML) in low-middle income countries (LMICs) faces many obstacles. Indeed, delayed diagnosis, refusal or abandonment of therapy, suboptimal treatment adherence, and inadequate supportive therapy, contribute among other parameters to poor outcome.

These obstacles are themselves related to adverse factors such as limitation of universal health care, private or familial financial constraints, low number of specialists, limited access to imaging (e.g. CT scan) and to routine laboratory practices, including biopsy specimen analysis. Transportation difficulties or interference with traditional medicine may also occasionally play a role.

Historically, the Patient Navigator (PN) was promoted in the late 80s by Dr. Freeman in New York City, to help underprivileged patients suffering from breast cancers at the earliest phase of their health care trajectory [1]. From the start, the PN has been proposed as an accompanying procedure based on coordinated interactions between the oncologist and the patient through phone calls at the patient’s home or visits by volunteers or nurses during the initial phase of the patient’s trajectory. PN was found to shorten the delay between screening and initiation of therapy as well as increase patient reassurance [1]. In the US, PN was found to be efficient for reducing health care disparities and it is now supported by the Patient Protection and Affordability Act [2]. Surprisingly, PN received little attention in LMIC with some exceptions, such as in Brazil [3].

PN has been applied through different variants, among which the AMA (Ambulatory Medical Assistance) procedure implemented at the Hematology Department of Toulouse University Medical Center (France) for lymphoma patients treated with chemotherapy [4]. AMA consists of a systematic weekly phone call to the patient’s home by a specialized nurse (nurse navigator/NN) who collects all information concerning drug-induced toxicities. Based on an algorithm, the signs are considered as insignificant, manageable by the nurse, or requiring immediate intervention of the oncologist. AMA was found to be feasible and highly efficient for detecting and managing complications during chemotherapy. AMA benefited from overwhelming support from patients and caregivers. AMA also appeared to be efficient in its triage function, with considerable time saving for medical staff. More recently, a randomized study demonstrated that AMA could increase patient adherence in patients treated with chemotherapy for chronic lymphocytic leukemia [5].

Based on these findings and considerations, we concluded that AMA was a simple and relatively inexpensive procedure that could be applied to LMIC patients and that had the potential to efficiently reduce refusal or abandonment of therapy or to improve observance in treated patients.

For this reason, we designed a randomized study comparing the standard survey procedure and AMA (here designated as AMAFRICA) in a cohort of 100 lymphoma patients treated with chemotherapy in the Clinical Hematology Department of Abidjan University Medical Center (Ivory Coast). This study was in part supported by Pierre Fabre Foundation, a not-for-profit charitable organization involved in health care in French-speaking LMICs, especially in South Asia and Africa.

This study investigated the impact of AMAFRICA procedure, a patient navigator patient program, on the management of patients treated for ML in Ivory Coast. This randomized study showed a significant impact of AMAFRICA with the rate of refusal and abandonment. However, response rate and overall survival were unaffected.

AMAFRICA was derived from AMA, a patient navigator variant used in France for the management of chemotherapy intercourse for ML patients. According to this procedure, a specialized nurse (“nurse navigator”/NN) performed a systematic outgoing phone call at the patient’s home, collected all information and served as coordinator between the oncology unit and the patient [4]. AMA is now used as a standard in numerous French institutions, including ours. Nearly 3000 patients have been enrolled. More recently, AMA was used for monitoring lymphoma patient survivors [9]. Previous studies have suggested that the AMA procedure could improve the quality of clinical management, including observance, safety, comfort and reassurance [4,5]. We reasoned that such a procedure could be useful in LMIC to prevent refusal or abandonment of therapy, two major limitations of cancer care in underserved countries.

Ivory Coast is one of the West African French-speaking countries with 25 million inhabitants, with nearly one third living in Abidjan, the capital city. Ivory Coast has a gross domestic product (GDP) of about 1662 USD per capita (3130 PPP/Purchasing Power Parity) and a human development index (HDI) of 0.474 (182nd position in the world). Despite a fast-growing economy, Ivory Coast meets health financing criteria of low-income countries by limitation of universal health care and large contributions of out-of-pocket financing [10]. Thus, whereas total health spending per capita was PPP 172 USD, 51% of this amount was paid out-of-pocket by households, (www.​africanstrategie​s4health.​org) and less than 10% of the Ivorian population has adequate health coverage. Despite these limitations, health care demand remains high. For example, in 2018, the hematology department of Abidjan University Medical center managed 1142 patients, among them 159 cases of ML.

The rate of treatment refusal and abandonment is very high (60%). Similar rates of treatment refusal have been also described in other countries such as India [11] and Kenya [12]. In a meta-analysis, Gupta and coworkers made an inventory of 83 studies in various LMIC (mainly in Asia), and found a global refusal/abandonment rate of 54% [13]. However, most of these studies were retrospective. We believe that the rate we found is more accurate to reflect health care access in the real African world. This rate is largely explained by financial constraints.

In this context, encouraging patients to be treated is challenging. Indeed, a large majority of patients who abandoned (a total of 17%), left after only one cycle, again mostly for financial reasons. The cost of each CHOP cycle has been estimated at 376 USD. Thus, the cost of complete CHOP therapy represents about 70% of the mean annual PPP income, while spending more than 10% of total expenditure on out-of-pocket health care costs is considered by WHO as a threshold of so-called “catastrophic health spending” [10]. Despite some occasional support (including family generosity), this situation poses a threat in terms of dispossession and misery. Therefore, any plan of treatment should be considered cautiously, based on patient financial resources.

Financial considerations are the main reason put forward by the patients to justify refusal and abandonment. Family opposition is often also based on costs. However, our study revealed other reasons such as transportation, already documented in other LMIC [14], or cultural reasons such as fatalism or interference with traditional medicine, also documented elsewhere [12]. All these factors also play a role in delay of clinical management and the high incidence of advanced forms of diseases.

Reports dealing with treatment efficacy in LMIC cancer patients based on intent-to-treat are uncommon. Our study shows poor response rate and survival based on intent-to-treat in lymphoma patients, compared to developed countries. For example, the CHOP regimen yielded about 50% CR rate and RCHOP (the standard in Europe and in North America) about 70% or even more, depending on the initial risk factors. These results should be compared with the 16% CR rate of our cohort. The combination of delayed diagnosis, treatment refusal or abandonment, poor adherence to treatment and death under therapy converge to yield a low rate of response, even if adherent patients displayed an encouraging 50% CR rate. Another implication of the high rate of refusal and abandonment we described herein is that future prospective clinical trials should take into account such limitation for calculating the number of patients needed to be recruited in order to avoid not conclusive study.

The North/South cooperation built for this study was also instructive for pathological review. Indeed, the provisional diagnosis provided by the local pathologist was changed in about 60% after reanalysis by an expert using IHC. Such high rate of discordance could be judged as alarming. Since the referral (here our pathologist colleague working in Abidjan) used HE staining without IHC, a high discordance rate was anticipated, especially for classifying small B-cell and T cell lymphoma. As a comparison, the rate of discordance in France between referral pathologist and expert was 20%, nearly half of cases corresponding to B-cell derived NHL subtype reclassification [6]. there is no doubt that pathological analysis should be improved in Ivory Coast. In our study, specimen reanalysis was found to be feasible for a reasonable cost (about 90 USD for air transportation). However, this process is not satisfactory. In the near future, IHC should be implemented in Abidjan and in anticipation, one of our Ivorian pathologist colleagues has received specific training in IHC procedure and interpretation. Telepathology represents also an attractive approach [15].

Finally, this prospective study confirms that treatment refusal and abandonment are major issues for treatment of ML in LMIC. Some parameters should be considered such as education, communication with patients and families, access to medical staff, motivation and logistics. However, the main obstacles are financial. It is very likely that the limitation of universal health care plays a role even if it is still debated [16]. Our study may have implications for health care policy applied to ML management. Indeed, the financing of “essential medicine” [17] like CHOP for the whole population appeared more critical that the introduction of new drugs, such as targeted therapies, including under their generic forms. At the least, further studies are needed to measure cost efficacy studies, for example through DALY calculation [18]. This approach has been already developed in Brazil and Malawi in the context of pediatric malignancies [19]. Finally, since they are feasible, PN-derived procedures like AMA could improve patient management in other less aggressive clinical settings such as in infectious or degenerative chronic diseases, as well as in sickle cell anemia.