Objectives Understanding non-adherence to guidelines in patients with co-morbidities by supplementing quantitative data through patient-centred qualitative research. It is hypothesised that clinical constraints and patient-related factors explain the vast proportion of non-adherence.

Design Mixed-method case study.

Setting Primary and secondary acute care hospital, Department for Internal Medicine.

Participants All consecutive patients having chronic heart failure (CHF) being hospitalised within a 2-year period.

Results Quantitative drug prescribing analysis in 348 patients with CHF confirms moderate guideline adherence: the guideline adherence index (GAI) corresponds to 0.7 for the three most important drug classes indicated for CHF and to 0.6 for all five recommended drug classes. Corrections with regard to the most important clinical contraindication (renal insufficiency) for these drugs raise the GAI to 0.8 in both categories. Semistructured interviews in 50 consecutive patients show relevant reasons for non-adherence in half of the remaining patients with non-adherence to guidelines and raise the adjusted GAI to 0.9. Up to 75% of de jure non-adherence can thus be explained by clinical constraints.

Conclusion Quantitative data analysis of treatment regimens in patients with CHF is an inaccurate method for measuring guideline adherence. Combining quantitative prescribing data with semistructured interviews shows a 90% match concerning guideline adherence compared to an only 60% match based on quantitative data alone. Thus, neither quality nor economical assessments of the treatment strategy in patients with chronic diseases should be solely based on quantitative analysis. Understanding non-adherence is crucial for defining and improving quality of care.