Table 2. Variation in reported invasive versus conservative NSTEMI management between hospitals which offer PCI services all the time in England (n=51) and according to eGFR range at NSTEMI admission).
| n with reported invasive cardiac management (row %) | Row total | Model 1* | Model 2† | Model 3‡ | |||
|---|---|---|---|---|---|---|---|
| OR (95% CI) | OR (95% CI) | OR (95% CI) | APP (95% CI) | ||||
| STEMI | Missing | 5728 (97.3) | 5890 | – | 0.51 (0.25 to 1.05) | 0.31 (0.14 to 0.67) | 93 (91 to 95) |
| eGFR range 1 | 32 837 (99.0) | 33 179 | – | 2.57 (2.06 to 3.2) | 0.89 (0.75 to 1.05) | 97 (96 to 98) | |
| eGFR range 2 | 32 401 (97.4) | 33 280 | – | 1 (reference) | 1 (reference) | 97 (96 to 98) | |
| eGFR range 3a | 6778 (93.2) | 7273 | – | 0.37 (0.32 to 0.42) | 0.58 (0.51 to 0.67 | 96 (94 to 97) | |
| eGFR range 3b | 3134 (89.1) | 3519 | – | 0.21 (0.17 to 0.25) | 0.42 (0.36 to 0.50) | 95 (93 to 96) | |
| eGFR range 4 | 886 (81.5) | 1087 | – | 0.11 (0.09 to 0.13) | 0.24 (0.2 to 0.29) | 92 (89 to 94) | |
| eGFR range 5 | 163 (81.1) | 201 | – | 0.11 (0.07 to 0.17) | 0.17 (0.11 to 0.27) | 89 (85 to 93) | |
| Coded renal failure | 1721 (86.0) | 2002 | – | 0.17 (0.14 to 0.2) | 0.32 (0.26 to 0.39) | 93 (91 to 95) | |
| Rho (intracluster coefficient) | – | – | 0.27 (0.17 to 0.41) | 0.24 (0.17 to 0.42) | 0.29 (0.18 to 0.43) | – | |
| NSTEMI | Missing | 2397 (86.0) | 2788 | – | 0.86 (0.55 to 1.34) | 0.62 (0.43 to 0.89) | 80 (76 to 84) |
| eGFR range 1 | 14 882 (93.6) | 15 898 | – | 2.7 (2.17 to 3.35) | 0.85 (0.75 to 0.96) | 83 (80 to 87) | |
| eGFR range 2 | 20 110 (84.5) | 23 788 | – | 1 (reference) | 1 (reference) | 85 (82 to 88) | |
| eGFR range 3a | 4909 (72.7) | 6756 | – | 0.48 (0.43 to 0.53) | 0.75 (0.68 to 0.83) | 82 (79 to 86) | |
| eGFR range 3b | 2169 (59.4) | 3651 | – | 0.26 (0.22 to 0.3) | 0.48 (0.43 to 0.53) | 77 (73 to 81) | |
| eGFR range 4 | 520 (44.0) | 1180 | – | 0.13 (0.1 to 0.16) | 0.25 (0.20 to 0.31) | 68 (63 to 73) | |
| eGFR range 5 | 147 (51.6) | 285 | – | 0.17 (0.13 to 0.24) | 0.19 (0.13 to 0.26) | 64 (57 to 70) | |
| Coded renal failure | 2757 (62.7) | 4394 | – | 0.26 (0.22 to 0.31) | 0.4 (0.34 to 0.45) | 75 (70 to 79) | |
| Rho (intracluster coefficient) | – | – | 0.26 (0.16 to 0.39) | 0.28 (0.17 to 0.42) | 0.28 (0.17 to 0.43) | – | |
To understand the variation in cardiac management across hospitals, we compared three logistic regression models which incrementally accounted for variation across hospitals using centre as a random effect (model 1), additionally eGFR range at AMI admission as a fixed effect (model 2) and additionally several other potential confounders of the association between eGFR range and AMI management strategy (model 3). Of interest is the intracluster coefficient (rho) across these three models. The intracluster coefficient is calculated by dividing the between-cluster variability and the sum of the within-cluster and between-cluster variabilities, meaning it describes the proportion of the variation in the outcome (invasive vs conservative cardiac management) explained by the centre-level variation, after accounting for any other fixed effects included in the model (models 2 and 3).
eGFR ranges (mL/min/1.73 m2): range 1 (≥90), range 2 (60–89), range 3a (45–59), range 3b (30–44), range 4 (15–29) and range 5 (0–14).
Logistic regression model with cardiology centre as random effect (no fixed effect independent variables).
Logistic regression model with cardiology centre as random effect and eGFR range as fixed effect.
Logistic regression model with cardiology centre as random effect, eGFR ranges, sex, age, admission year, ethnicity, comorbidities (previous MI, angina, hypertension, hypercholesterolaemia, peripheral vascular disease, COPD, heart failure, type 2 diabetes), co-prescriptions (RASi, beta-blocker, statin).
AMI, acute myocardial infarction; APP, adjusted predicted percent; COPD, chronic obstructive pulmonary disease; eGFR, estimated glomerular filtration rate; MI, myocardial infarction; NSTEMI, non-ST-elevation myocardial infarction; OR, odds ratio; PCI, percutaneous coronary intervention; RASi, renin-angiotensin system inhibitors; STEMI, ST-elevation myocardial infarction.