| Univariable analysis | Full model | |||
|---|---|---|---|---|
| OR (95% CI) | P value | OR (95% CI) | P value | |
| Patient demographics | ||||
| % malesa | 0.98 (0.97 to 0.98) | <0.001 | 1.00 (0.99 to 1.00) | 0.708 |
| % aged ≥65 yearsa | 0.99 (0.99 to 1.00) | <0.001 | 0.98 (0.97 to 0.99) | <0.001 |
| Ratio 85:65 | 1.05 (1.05 to 1.06) | <0.001 | 1.05 (1.05 to 1.06) | <0.001 |
| Practice population size | 1.00 (1.00 to 1.01) | <0.001 | 1.00 (1.00 to 1.00) | 0.915 |
| Disease prevalence | ||||
| AF | 1.00 (1.00 to 1.01) | <0.001 | 1.02 (1.01 to 1.03) | 0.001 |
| CHD | 0.99 (0.98 to 0.99) | <0.001 | 1.00 (0.99 to 1.01) | 0.946 |
| CKD | 0.99 (0.99 to 0.99) | <0.001 | 1.00 (1.00 to 1.00) | 0.694 |
| DM | 0.93 (0.93 to 0.94) | <0.001 | 0.94 (0.93 to 0.94) | <0.001 |
| GP Patient Survey | ||||
| Experience | 1.01 (1.01 to 1.02) | <0.001 | 1.00 (1.00 to 1.01) | 0.628 |
| Preferred GP | 1.01 (1.01 to 1.01) | <0.001 | 1.01 (1.01 to 1.01) | <0.001 |
| Trust | 1.01 (1.00 to 1.01) | <0.001 | 1.00 (0.99 to 1.00) | 0.087 |
| Total QOF scoreb | 1.00 (1.00 to 1.01) | 0.060 | 1.00 (1.00 to 1.00) | 0.771 |
| CCG variance | ||||
| 95% mid-rangec | 1.97 (1.74 to 2.28) | <0.001 | 1.99 (1.76 to 2.31) | <0.001 |
Multilevel logistic regression models with a binomial distribution were used, with CCG included as a random effect to account for clustering. All measures were standardised (using sample mean values and SDs), and OR represent the odds of a patient dispensed an anticoagulant (DOAC and warfarin) given a DOAC, for a unit increase in the exposure of interest. DOACs include dabigatran, rivaroxaban, apixaban, and edoxaban
aDenominator data used was practice-list size in January–March 2017. bTotal QOF score is total QOF points achieved as a percentage of all achievable points (max 559) for 2015–2016. cCalculated from the variance of the random effect (σ2) and is given by e2 ×1.96×σ and represents the odds ratio comparing a practice at the 2.5th percentile of the distribution of practices to one at the 97.5th percentile