Global estimation of anti-malarial drug effectiveness for the treatment of uncomplicated Plasmodium falciparum malaria 1991–2019

Giulia Rathmes, Susan F. Rumisha, Tim Lucas, Katherine A. Twohig, André Python, Huong Lan Thi Nguyen, A. Nandi, Suzanne Keddie, Emma L. Collins, Jennifer Rozier, Harry S. Gibson, Elisabeth G. Chestnutt, Katherine E. Battle, Georgina Humphreys, Punam Amratia, Rohan Arambepola, Amelia Bertozzi-Villa, Penelope A. Hancock, Justin Millar, Tasmin L. Symons, Samir Bhatt, Ewan Cameron, Philippe J. Guérin, Peter W. Gething, Daniel J. WeissMalaria 2020 · DOI

Abstract

This study provides evidence that health system performance, drug quality and patient adherence influence the effectiveness of anti-malarials used in treating uncomplicated falciparum malaria. These results provide guidance to countries' treatment practises and are critical inputs for malaria prevalence and incidence models used to estimate national level malaria burden.

Introduction

In the global effort to control and eliminate malaria, policy decisions are frequently guided by drug efficacy—the performance of a medication under the controlled, ideal conditions of a clinical trial. However, the real-world effectiveness (AmE) of these drugs is often significantly lower due to "implementation gaps." These gaps arise from factors such as patient non-adherence to dosage regimens, the prevalence of substandard or falsified drugs, and systemic weaknesses in healthcare infrastructure.

Until recently, there has been a lack of standardized, high-resolution global data on how these real-world factors degrade drug performance over time and space. This research presents the first global maps of anti-malarial drug effectiveness from 1991 to 2019, bridging the gap between clinical trials and routine care.

Methods

The researchers developed a Bayesian spatiotemporal modeling framework to estimate effectiveness at a 5 km × 5 km pixel level. The methodology involved several key components:

  1. Data Standardization: Raw trial data from the WorldWide Anti-malarial Resistance Network (WWARN) was re-analyzed using a consistent "Intention-To-Treat" (ITT) approach to ensure comparability across 232 trials involving over 86,000 individuals.
  2. Covariate Integration: The model incorporated 69 environmental and socio-economic variables from the Institute for Health Metrics and Evaluation (IHME) and the Malaria Atlas Project (MAP), including healthcare access, maternal education, and proximity to urban centers.
  3. Effectiveness Adjustments: To translate trial efficacy into real-world effectiveness, the authors applied evidence-based "penalties": a 19.1% reduction for poor drug quality and a 5.0% reduction for imperfect patient adherence.
  4. Computational Inference: Integrated Nested Laplace Approximations (R-INLA) were used to manage the intensive computational demands of global modeling.

Key Results

The study reveals a significant "efficacy-effectiveness gap." Artemisinin-based combination therapies (ACTs) were found to be approximately 29.6% less effective in routine care than in clinical trials. For non-artemisinin drugs, this gap widened to 36%.

Global maps of predicted artemisinin-based drug effectiveness across three time periods: 1991–2000, 2011–2015, and 2016–2019.

Global effectiveness of ACTs saw a modest increase from 67.4% (1991–2000) to 71.8% (2016–2019). In contrast, non-ACT drugs stagnated at 52–55% effectiveness over the 25-year period. Regional analysis showed:

  • Sub-Saharan Africa: Improvements were noted after 2016, though central regions continue to face low effectiveness due to political instability and healthcare disruption.
  • Southeast Asia: Effectiveness has declined in parts of Thailand, Myanmar, and Cambodia, driven by emerging biological resistance to both artemisinin and partner drugs.
  • Socio-Economic Drivers: Strong predictors of effectiveness included health system strength (proxied by skilled birth attendant prevalence) and maternal education, highlighting that infrastructure is as critical as biology in malaria treatment.

Discussion and Future Directions

While providing a groundbreaking global view, the study acknowledges several limitations. Data sparsity in Central and South America increases uncertainty in those regions. Furthermore, the application of uniform penalties for drug quality and adherence does not yet account for local variations between urban and rural settings.

Future research should aim to incorporate patient-level pharmacokinetic data and routine surveillance data from national health systems. Refining these "top-down" estimates with "bottom-up" local data will be essential for improving the accuracy of malaria burden models and informing national treatment policies.

Summary updated

Abstract

This study provides evidence that health system performance, drug quality and patient adherence influence the effectiveness of anti-malarials used in treating uncomplicated falciparum malaria. These results provide guidance to countries' treatment practises and are critical inputs for malaria prevalence and incidence models used to estimate national level malaria burden.
BibTeX
@inproceedings{rathmes2020global,
  author = {Giulia Rathmes and Susan F. Rumisha and Tim Lucas and Katherine A. Twohig and André Python and Huong Lan Thi Nguyen and A. Nandi and Suzanne Keddie and Emma L. Collins and Jennifer Rozier and Harry S. Gibson and Elisabeth G. Chestnutt and Katherine E. Battle and Georgina Humphreys and Punam Amratia and Rohan Arambepola and Amelia Bertozzi-Villa and Penelope A. Hancock and Justin Millar and Tasmin L. Symons and Samir Bhatt and Ewan Cameron and Philippe J. Guérin and Peter W. Gething and Daniel J. Weiss},
  title = {{Global estimation of anti-malarial drug effectiveness for the treatment of uncomplicated Plasmodium falciparum malaria 1991–2019}},
  booktitle = {Malaria},
  year = {2020},
  doi = {10.1186/s12936-020-03446-8},
}
Figure