Uptake of direct oral anticoagulants in primary care: an ecological and economic study

Figures

• 

  Figure 1.. Variation in the relative uptake of DOACs across CCGs, January to March 2017 DOAC = direct oral anticoagulant; CCG = clinical commissioning groups. Anticoagulants include dabigatran, rivaroxaban, apixaban, and edoxaban (DOACs) and warfarin. Percentage of patients dispensed anticoagulants given a DOAC calculation = (Total number of patients dispensed a DOAC/Total number of patients dispensed any anticoagulant) multiplied by 100

• 

  Figure 2.. Average (mean) quantity of anticoagulants (DOACs and warfarin) dispensed by a general practice in England, by year quarter. DOAC = direct oral anticoagulant. A: Average (mean) quantity of DOACs (dabigatran, rivaroxaban, apixaban, and edoxaban) dispensed per practice. B: Average (mean) quantity of all anticoagulants and warfarin dispensed per practice. C: Average (mean) relative uptake of DOACs per practice (total quantity of DOACs/total quantity of anticoagulants multiplied by 100) The grey shaded areas represent 95% confidence intervals (mean ±1.96*SD)

• 

  Figure 3.. Rates of ischaemic stroke, intracranial bleeding and gastrointestinal bleeding, per 1000 population, by year quarter. GI = gastrointestinal. Ischaemic stroke defined as any I63 ICD-10 code occurring in the primary diagnostic position for any care episode within the first 7 days of admission. Intracranial bleeding (ICD-10 codes I160 to I62) and gastrointestinal bleeding (K25.0/2/4/6-K28.0/2/4/6, K92.0–2) recorded as the primary diagnosis for any episode of care. Denominator data were based on practice population sizes obtained from quarterly reports of patient numbers to NHS Digital, and where missing, annual figures recorded as part of the English primary care pay-for-performance scheme, the Quality and Outcomes Framework.

• 

  Figure 4.. Average (mean) total cost of anticoagulants and associated outcomes, per 1000 patients. DOAC = direct oral anticoagulant; CCG = clinical commissioning group. A: Average (mean) total cost per 1000 patients stratified by percentage anticoagulants a DOAC; January to March 2017. B: Average (mean) total cost per 1000 patients stratified by CCG; January to March 2017. C: Average (mean) total cost per 1000 patients stratified by year quarter. Total cost includes cost of anticoagulants (dabigatran, rivaroxaban, apixaban, edoxaban, and warfarin), management costs of warfarin, and costs associated with a hospital admission for ischaemic stroke, and intracranial and gastrointestinal bleed (see Table S1). The grey shaded areas in Figures A and C represent 95% confidence intervals (mean ±1.96). Figure A restricted to practices within the 95% range of outcome (10%–50% anticoagulants dispensed a DOAC). Analysis excludes practices with practice population <50 (n = 1).

• 

  Figure 5.. Marginal effect of total cost per 1000 patients, by change in relative uptake of DOACs, with CCG distribution as of January to March, 2017. DOAC = direct oral anticoagulant; CCG = clinical commissioning group. Generalised linear models (GLMs) were used to estimate the costs associated with a change in relative uptake of DOACs, whereby the logarithm of the conditional expectation of the total cost was estimated, and the relationship of the mean to variance in the outcome data were assessed and modelled (models presented in Table S4). CCG was included as a random effect to account for clustering and robust standard errors were used to allow for potential misspecification of the link and family function. Marginal effects represent a change in the outcome for an increase in the exposure, keeping all other covariates at their observed levels, and averaged over all patients. Models were adjusted for demographic, clinical and practice factors CCG distribution reflects the average DOAC uptake in general practices within a CCG. NHS Digital did not have information available before 1 April 2013 so the final models are restricted to after this date.