Multidrug and optimal heart failure therapy prescribing in older general practice populations: a clinical data linkage study.

OBJECTIVE: To investigate multidrug therapy in the cardiovascular disease (CVD) population and whether it was associated with suboptimal drug prescribing in heart failure (HF). DESIGN: A population-based cross-sectional clinical data linkage study. SETTING: The clinical database populations were registered with three general practices in North Staffordshire that are part of a research network. PARTICIPANTS: 3155 patients aged 50 years and over were selected on the basis of a CVD-related prescription and a CVD consultation code applied to their electronic medical record in a 2-year time period. All available diagnostic data were linked to all drugs prescribed data during this time period. Two study groups were: (1) HF and (2) non-HF CVD (reference group). EXPOSURE: A standard drug formulary system was used to define four multidrug count categories based on the number of different British National Formulary drug chapters prescribed at the same time. PRIMARY AND SECONDARY OUTCOME MEASURES: Optimal HF therapy was defined as the prescribing of ACE inhibitor (ACEi) or a combination of ACEi and β-blocker in the 2-year time window. An additional three specific CVD drug categories that are indicated in HF were also measured. RESULTS: The HF group, compared with the reference group, had higher non-CVD multidrug therapy (26% with 7 or more counts compared with 14% in the non-HF CVD reference group). For the first-choice optimal drug treatment for HF with ACEi (64%) or ACEi and β-blocker combined therapy (23%), the multidrug-adjusted associations between the HF group and the reference group were OR 3.89; 95% CI 2.8 to 5.5 and 1.99; 1.4 to 2.9, respectively. These estimates were not influenced by adjustment for sociodemographic factors and multidrug counts. CONCLUSIONS: Multidrug therapy prescribing is much higher in the HF group than in a comparable CVD group but did not influence optimal drug prescribing.