Dr James Watson

Research Area: Bioinformatics & Stats (inc. Modelling and Computational Biology)
Technology Exchange: Computational biology and Medical statistics
Scientific Themes: Tropical Medicine & Global Health and Clinical Trials & Epidemiology
Keywords: Bayesian statistics, pharmacokinetics and pharmacodynamics, Exploratory subgroup analysis and G6PD deficiency
Predicted optimal ascending primaquine regimen for G6PD deficiency. Left panel: the predicted mean fall in haemoglobin over time under an idealised and a more pratical regimen. Right panel: the ascending dose regimen (daily total mg doses).

Predicted optimal ascending primaquine regimen for G6PD deficiency. Left panel: the predicted mean ...

The dynamics of bunching and the relationship with the MIC of a blood stage antimalarial drug. A hypothetical drug with 1st order elimination (top left panel) and the response-concentration curve (top right panel). The fate of daily reinfections emerging from the liver (bottom left) and the corresponding distribution of patent infections (bottom right).

The dynamics of bunching and the relationship with the MIC of a blood stage antimalarial drug. A ...

My main interest is characterizing the biology of relapse in vivax malaria via the analysis of genetic and epidemiological data. Alongside this, I'm interested in developing statistical models to better understand the pharmacology of antimalarial drugs. I have a particular interest in G6PD deficiency in the context of radical curative drugs for vivax malaria. I am working on developing a new primaquine regimen which would be safe in G6PD deficiency. My other interests are in Bayesian statistics and exploratory subgroup analysis.

Name Department Institution Country
Professor Chris Holmes Wellcome Trust Centre for Human Genetics Oxford University, Henry Wellcome Building of Genomic Medicine United Kingdom
Professor Sir Nicholas J White FRS Tropical Medicine Oxford University, Bangkok Thailand
Professor Joel Tarning Tropical Medicine Oxford University, Bangkok Thailand
Dr Aimee Taylor Harvard School of Public Health United States
White NJ, Watson J, Ashley EA. 2017. Split dosing of artemisinins does not improve antimalarial therapeutic efficacy. Sci Rep, 7 (1), pp. 12132. | Show Abstract | Read more

It has been suggested recently, based on pharmacokinetic-pharmacodynamic modelling exercises, that twice daily dosing of artemisinins increases malaria parasite killing and so could "dramatically enhance and restore drug effectiveness" in artemisinin resistant P. falciparum malaria infections. It was recommended that split dosing should be incorporated into all artemisinin combination regimen designs. To explain why parasite clearance rates were not faster with split dose regimens it was concluded that splenic malaria parasite clearance capacity was readily exceeded, resulting in the accumulation of dead parasites in the circulation, that parasite clearance was therefore an unreliable measure of drug efficacy, and instead that human immunity is the primary determinant of clearance rates. To test these various hypotheses we performed a logistic meta-regression analysis of cure rates from all falciparum malaria treatment trials (n = 40) with monotherapy arms containing artemisinin or a derivative (76 arms). There was no evidence that split dosing enhanced cure rates.

van der Pluijm RW, Watson J, Woodrow CJ. 2017. Antimalarial Resistance Unlikely To Explain U.K. Artemether-Lumefantrine Failures. Antimicrob Agents Chemother, 61 (7), pp. e00721-17-e00721-17. | Read more

Watson J, Taylor WR, Menard D, Kheng S, White NJ. 2017. Modelling primaquine-induced haemolysis in G6PD deficiency. Elife, 6 | Show Abstract | Read more

Primaquine is the only drug available to prevent relapse in vivax malaria. The main adverse effect of primaquine is erythrocyte age and dose-dependent acute haemolytic anaemia in individuals with glucose-6-phosphate dehydrogenase deficiency (G6PDd). As testing for G6PDd is often unavailable, this limits the use of primaquine for radical cure. A compartmental model of the dynamics of red blood cell production and destruction was designed to characterise primaquine-induced haemolysis using a holistic Bayesian analysis of all published data and was used to predict a safer alternative to the currently recommended once weekly 0.75 mg/kg regimen for G6PDd. The model suggests that a step-wise increase in daily administered primaquine dose would be relatively safe in G6PDd. If this is confirmed, then were this regimen to be recommended for radical cure patients would not require testing for G6PDd in areas where G6PDd Viangchan or milder variants are prevalent.

Watson J, Nieto-Barajas L, Holmes C. 2017. Characterizing variation of nonparametric random probability measures using the Kullback-Leibler divergence STATISTICS, 51 (3), pp. 558-571. | Show Abstract | Read more

© 2016 Informa UK Limited, trading as Taylor & Francis Group. This work characterizes the dispersion of some popular random probability measures, including the bootstrap, the Bayesian bootstrap, and the Pólya tree prior. This dispersion is measured in terms of the variation of the Kullback–Leibler divergence of a random draw from the process to that of its baseline centring measure. By providing a quantitative expression of this dispersion around the baseline distribution, our work provides insight for comparing different parameterizations of the models and for the setting of prior parameters in applied Bayesian settings. This highlights some limitations of the existing canonical choice of parameter settings in the Pólya tree process.

Watson J, Holmes C. 2016. Rejoinder: Approximate Models and Robust Decisions STATISTICAL SCIENCE, 31 (4), pp. 516-520. | Read more

Watson J, Holmes C. 2016. Approximate Models and Robust Decisions STATISTICAL SCIENCE, 31 (4), pp. 465-489. | Show Abstract | Read more

Decisions based partly or solely on predictions from probabilistic models may be sensitive to model misspecification. Statisticians are taught from an early stage that "all models are wrong, but some are useful" however, little formal guidance exists on how to assess the impact of model approximation on decision making, or how to proceed when optimal actions appear sensitive to model fidelity. This article presents an overview of recent developments across different disciplines to address this. We review diagnostic techniques, including graphical approaches and summary statistics, to help highlight decisions made through minimised expected loss that are sensitive to model misspecification. We then consider formal methods for decision making under model misspecification by quantifying stability of optimal actions to perturbations to the model within a neighbourhood of model space. This neighbourhood is defined in either one of two ways. First, in a strong sense via an information (Kullback-Leibler) divergence around the approximating model. Second, using a Bayesian nonparametric model (prior) centred on the approximating model, in order to "average out" over possible misspecifications. This is presented in the context of recent work in the robust control, macroeconomics and financial mathematics literature. We adopt a Bayesian approach throughout although the presentation is agnostic to this position.

White NJ, Watson J, Ashley EA. 2017. Split dosing of artemisinins does not improve antimalarial therapeutic efficacy. Sci Rep, 7 (1), pp. 12132. | Show Abstract | Read more

It has been suggested recently, based on pharmacokinetic-pharmacodynamic modelling exercises, that twice daily dosing of artemisinins increases malaria parasite killing and so could "dramatically enhance and restore drug effectiveness" in artemisinin resistant P. falciparum malaria infections. It was recommended that split dosing should be incorporated into all artemisinin combination regimen designs. To explain why parasite clearance rates were not faster with split dose regimens it was concluded that splenic malaria parasite clearance capacity was readily exceeded, resulting in the accumulation of dead parasites in the circulation, that parasite clearance was therefore an unreliable measure of drug efficacy, and instead that human immunity is the primary determinant of clearance rates. To test these various hypotheses we performed a logistic meta-regression analysis of cure rates from all falciparum malaria treatment trials (n = 40) with monotherapy arms containing artemisinin or a derivative (76 arms). There was no evidence that split dosing enhanced cure rates.

Watson J, Taylor WR, Menard D, Kheng S, White NJ. 2017. Modelling primaquine-induced haemolysis in G6PD deficiency. Elife, 6 | Show Abstract | Read more

Primaquine is the only drug available to prevent relapse in vivax malaria. The main adverse effect of primaquine is erythrocyte age and dose-dependent acute haemolytic anaemia in individuals with glucose-6-phosphate dehydrogenase deficiency (G6PDd). As testing for G6PDd is often unavailable, this limits the use of primaquine for radical cure. A compartmental model of the dynamics of red blood cell production and destruction was designed to characterise primaquine-induced haemolysis using a holistic Bayesian analysis of all published data and was used to predict a safer alternative to the currently recommended once weekly 0.75 mg/kg regimen for G6PDd. The model suggests that a step-wise increase in daily administered primaquine dose would be relatively safe in G6PDd. If this is confirmed, then were this regimen to be recommended for radical cure patients would not require testing for G6PDd in areas where G6PDd Viangchan or milder variants are prevalent.

Watson J, Holmes C. 2016. Approximate Models and Robust Decisions STATISTICAL SCIENCE, 31 (4), pp. 465-489. | Show Abstract | Read more

Decisions based partly or solely on predictions from probabilistic models may be sensitive to model misspecification. Statisticians are taught from an early stage that "all models are wrong, but some are useful" however, little formal guidance exists on how to assess the impact of model approximation on decision making, or how to proceed when optimal actions appear sensitive to model fidelity. This article presents an overview of recent developments across different disciplines to address this. We review diagnostic techniques, including graphical approaches and summary statistics, to help highlight decisions made through minimised expected loss that are sensitive to model misspecification. We then consider formal methods for decision making under model misspecification by quantifying stability of optimal actions to perturbations to the model within a neighbourhood of model space. This neighbourhood is defined in either one of two ways. First, in a strong sense via an information (Kullback-Leibler) divergence around the approximating model. Second, using a Bayesian nonparametric model (prior) centred on the approximating model, in order to "average out" over possible misspecifications. This is presented in the context of recent work in the robust control, macroeconomics and financial mathematics literature. We adopt a Bayesian approach throughout although the presentation is agnostic to this position.

3411