This study has three main findings: Firstly, patients with coronary heart disease from European general practices perceive that the quality structured chronic care and counselling is not optimal. During this research the mean overall PACIC score was calculated as 2.84 (maximum = 5). During previous research, the PACIC instrument has been applied to a wide range of chronic conditions and populations [38, 39] including individuals suffering from diabetes , coronary heart disease (CHD) , osteoarthritis  and mental health , with overall scores reported being between 2.49  and 3.80. Compared to diabetes care , in the present study goal setting and follow-up support activities were less often provided during CHD care. Furthermore, key elements of the CCM, namely assisting patients with self management and arranging follow-up support, were provided significantly less often, indicating possible quality deficiencies and areas for quality improvement, particularly in relation to continuity of care .
The second and third findings are related to factors associated with patients’ assessments of receiving structured chronic care and counselling. At the patient level, being male, having more frequent practice contacts and having fewer other conditions were associated with higher PACIC scores. Other studies reported higher PACIC scores for younger patients  or did not demonstrate significant associations with patients’ characteristics [16, 19]. We applied the PACIC instrument to routine primary care settings in various European health care systems. Within these settings, patients who had been included in quality assessment previously scored significantly higher. This finding resonates with previous research, which demonstrated that patients who were enrolled in interventions and were cared for more intensively scored higher, especially when interventions were tailored to special elements of the CCM. For example, patients participating in disease-management programs have been reported as having higher PACIC scores than those receiving usual care . Furthermore, PACIC scores of patients participating in patient-centred case-management interventions can be improved from baseline to post-interventional measurement . These findings are consistent with the necessity to align evaluation research with care improvement strategies [42, 43].
At the practice level, there was a positive association between patients’ evaluations of the quality of care they received and quality scores reflecting quality-management and cardiovascular-care processes of general practices. However, the variance proportion at this level was less than the variance proportion at patient or country level. The variance caused by country specific factors can only be explained marginally in this study, as explanatory variables at the country level were not available. Although the variance between practices relating to the patients’ evaluations was relatively small, practice quality scores explained a significant portion of these variance. It had been presumed that larger practices with more full time equivalent (FTE) GPs would provide good quality of care . However, the number of FTE GPs was not associated with patients’ evaluations during this study.
There is not one standardised way in which to measure quality of care, and each method has strengths and limitations. Ideally, it would be desirable that patients’ evaluation of their experiences with health care were congruent to other quality measures of practice care, to provide feedback to health care providers. Previous research has demonstrated few associations between objective quality measures and patients’ satisfaction [32–34, 45]. Wensing et al. stated that patients may assess care differently from recommended care strategies . Furthermore, it has been argued that patients value humanistic and affective items (e.g. staff’s friendliness) more than items concerning organization and governance , and that patient assessments depend on the personal relationship to the practice team where trust, loyalty and positive regard may influence the assessment [34, 47]. In addition, several processes and structures of care are outside the direct experience or observation of patients, and not all patients are capable of understanding the risks and benefits of clinical choices [35, 48]. The quality of providers’ performance may not be reflected in patients’ perceptions if general satisfaction scores that summarize the assessment of different health care processes into one global score are used . The lack of positive association between quality measurement from the patients’ perspective and other quality measures of practice performance may be due to different underlying theoretical constructs of various instruments being used to assess patient perspectives [42, 49]. Patient satisfaction may reflect the relationship between patient and practice team, which is dependent on patient and practice characteristics, whereas patient experiences with care may focus on organisational and procedural aspects of care . The PACIC instrument predominantly questions the receipt of specific clinical processes of care. The underlying theoretical construct of the PACIC is the CCM; a patient centred care approach that is proactive, planned and includes goal setting, problem-solving and follow-up support. Therefore, positive association of patients’ perspectives with practice quality measures can be explained, as the quality indicators of practice organisation and chronic care used during this study could be related to these proactive CCM elements . As patient experience is an important component of quality of care, and patient involvement is central to achieving good outcomes, future research is necessary to specify the constructs of patient assessment instruments that are required to obtain valid and reliable patient judgements of health care processes, particularly if patient experience is used within pay for performance systems .
Strengths and limitations
The EPA cardio study is one of the largest international studies concerning management of cardiovascular care in European primary care . To eliminate different health care effects, countries with strong primary health care systems (UK, the Netherlands) and countries characterized by a weaker primary care orientation  were included, and multilevel modelling was used to adjust for these differences. We used validated patient measures and assessed practice quality indicators through well-proven means [24, 25]. All measures were pilot tested before being included in this study .
However, in Germany, Austria and Switzerland it was difficult to enrol 36 practices per country, as intended in the study protocol. As the PACIC instrument was only validated in three languages, countries such as France and Slovenia were excluded, decreasing the number of eligible patients. As we used an observational design, it was not possible to demonstrate a causal relationship between included characteristics or measures and patients’ perceived quality of care.