Primary care-based multifaceted, interdisciplinary medical educational intervention for patients with systolic heart failure: lessons learned from a cluster randomised controlled trial

Table 4 Heart failure hospital admissions and total mortality during seven-month trial follow-up, selected prescription rates for patients at seven-month trial follow-up, mean NT-proBNP-values at baseline and seven-month follow-up and primary care activity data during seven-month trial follow-up

Outcome	Time	Intervention group		Control group
			No		No
						Adjusted odds ratio* (95% CI); p**-value
Death (any cause)	Seven months	9	91	4	77	2.0 (0.6 to 7.1); 0.27
Patients admitted to hospital due to heart failure, no. (%)	Seven months	23 (27.1)	85	16 (23.5)	68	1.2 (0.5 to 2.6); 0.67
No. of heart failure hospital admissions	Seven months	31	85	34	68	0.8 (0.3 to 2.1); 0.63
Heart failure hospital admission or death	Seven months	29	89	18	70	1.4 (0.7 to 2.9); 0.35
No. (%) of drugs			85		68	adjusted odds ratio* (95% CI); p-value
ACE inhibitor or A2RA	Seven months	78 (91.8)		61 (89.7)		0.9 (0.2 to 3.4); 0.87
β-blocker	Seven months	68 (80)		57 (83.8)		0.7 (0.2 to 2.7); 0.58
ACE inhibitor/A2RA and β-blocker	Seven months	65 (76.5)		53 (77.9)		0.8 (0.2 to 2.9); 0.78
Spirononolactone/Eplerenone	Seven months	36 (42.4)		16 (23.5)		4.0 (1.2 to 13.0); 0.02
Mean NT-proBNP- values (pg/ml) (SD)						Adjusted mean difference** (95% CI); p-value
Crude	Baseline	2462.5 (2821.5)	87	2732.2 (5793.5)	69
Transformed*****		4.3 (1.0)		4.1 (1.1)
Crude	Seven months	2031.6 (3575.4)	71	1411.7 (2218.1)	57	0.17 (-0.04 to 0.39); 0.11
Transformed*****		3.9 (1.1)		3.6 (1.1)
Mean practice attendances (SD)	Seven months	24.0 (16.0)	80	21.6 (15.3)	67	3.9 (-2.9 to 10.7); 0.25
Mean referrals to cardiologist (SD)	Seven months	2.2 (2.5)	85	2.2 (1.9)	68	-0.03 (-0.8 to 0.7); 0.93

*Poisson regression model comparing number of outcomes between groups
**Adjusted for recruitment site (cluster), baseline value, New York Heart Association class, age and gender
***According to generalised linear mixed effect models with a logit link using SAS proc glimmix to compare the number of outcomes between groups
****Based on analysis of covariance comparing results at seven–month follow-up
***** Due to a skewed distribution of NT-proBNP, logarithmic transformation was performed. Approximation to a normal density was achieved using the formula t = 2*log₁₀(B+10)-2, where B denotes the raw NT-proBNP value[39]

ISSN: 1745-6215