We found significant gaps in recommended care for this high risk CVD patient group, particularly for prevention activities such as smoking cessation and weight management. A reduction in the smoking rate of high risk CVD patients could significantly impact mortality
[20–22]. Despite the existence of community help programs and public health campaigns, we found low rates for counseling patients to quit smoking, smoking cessation medication prescribing, and referrals to community programs, findings that are consistent with studies conducted across Canada, Europe and the United States
[23–27]. This is a missed opportunity for prevention as primary care physicians are seen as a credible source of information, and a number of studies have shown that counselling in combination with a nicotine replacement therapy can double ones chances of successfully quitting
. Lack of time, competing demands, lack of reimbursement, and perceived patient resistance are barriers to care
. Improvements in organizing care at the level of the practice may reduce these barriers
, along with patient self-management approaches such as motivational counselling
, as well as greater linkages between the primary care practice and the community.
Our findings also highlight the significant ongoing gap in diabetes management within primary care practices in Eastern Ontario. Optimizing care and achieving clinical targets could reduce mortality as people with diabetes are at high risk of suffering a cardiovascular event
[30–33]. These findings are consistent with other international findings, as a number of studies have demonstrated the challenges associated with treating patients with diabetes to target levels
Furthermore, the results of this study are consistent with other findings in Canada and Ontario which have demonstrated high levels of adherence to guidelines associated with hypertension
[37–39]. The adherence rates seen in this and other Canadian studies have been higher than those in the United States and other developed nations
[40, 41]. This marked improvement in hypertension management is likely due in part to local community programs, greater awareness and the establishment of the Canadian Hypertension Education Program (CHEP), an initiative developed in the late nineties to support primary care providers and patients by providing them with guidelines and recommendations for managing and preventing hypertension
We found a spectrum in uptake patterns for new evidence in primary care. For example, although the benefits of measuring eGFR levels in diagnosing and managing chronic kidney disease were not published until 2004
, uptake of this guideline in primary care has been rapid at least in part to the automatic reporting of eGFR by laboratories when serum creatinine is requested. This is apparent as 84% of patients with diabetes had undergone eGFR testing in this study. Similar uptake patterns have been seen internationally as well
[44, 45]. In contrast, despite evidence that waistline is a strong predictor of cardiovascular related morbidity and mortality and has been recommended since 1998
, only one in ten patients had a measure done
[47, 48]. Low levels of uptake have also been documented in other countries
[49, 50]. One study reported that family physicians cited lack of time, extra workload, opportunity costs, and concerns about the acceptability of this manoeuvre as barriers to uptake
Despite high adherence by providers to recommended guidelines for patients with hypertension, dyslipidemia and coronary artery disease, a high proportion of patients still did not meet clinical target levels, an observation that is likely due to several factors. For example, low patient compliance to medications could have impacted poor control rates, however, we are unable to confirm this from chart audit data alone. Also, changing guideline targets such as the recent changes made to the Canadian Lipid Guidelines in 2006, which decreased the target LDL level from 2.5 mmol/L to the more stringent 2.0 mmol/L may have also impacted the rates
. Furthermore, some experts have suggested that the choice of medication and a lack of dose titration are two potential reasons for poor LDL control rates
, while a recent study suggested that 38% of high risk patients would be unable to reach an LDL target of <2.0 mmol/L even when using a maximum dose statin monotherapy
. As well, this difficulty in management could potentially be due in part to clinical inertia - resistance of a health care provider to intensify therapy when indicated – as there could be a need to change drug therapies, doses or initiate counselling to change lifestyle habits (e.g., diet, exercise, etc.). Alternately, the poor control rates may simply highlight the complexities of controlling LDL and blood pressure levels in high risk multimorbid patients.
Translating research evidence into practice is challenging as the research which underpins clinical guidelines is often obtained from studies that exclude patients with multimorbidities like the ones examined in this study
[54–56]. As such, a cogent case can be made that guidelines may not easily apply to patients with multimorbidities. For example, if a multimorbid patient is newly diagnosed with hypertension and is already taking multiple medications, it may be less appropriate for a primary care physician to prescribe an additional drug. This is a concern when interpreting papers that look at guideline adherence, as most simply report on whether a treatment was delivered or not, even when it may be inappropriate to do so. In our study, as we wanted to capture whether appropriate care was delivered, we recorded physicians as adhering to a guideline if the specific care manoeuvre was performed, recommended or considered, regardless of whether the patient followed the recommendation such as compliance to prescriptions. This could also explain some of the gaps between processes and clinical outcomes in our results.
Although the participants in this study represent diverse family practices, they all voluntarily participated in IDOCC resulting in a potential selection bias. We found differences in physician sex, training background and remuneration model between IDOCC participants and both non-participants and provincial averages. This bias may impact the generalizability of our findings, as practices that tend to take part in quality improvement initiatives such as IDOCC, are likely highly motivated and are higher performing than provincial averages. As such, we anticipate that the gaps in care observed in this study are likely greater amongst the entire practice population in Eastern Ontario. Also, we were only able to present clinical test results for those who had screenings during the chart audit year, and thus, we are unable to comment on control rates for those who did not have any tests done. We presented summary indicators of the quality of CVD care using baseline data from three groups of primary care practices enrolled over three distinct steps. Although these data were collected over a three year period (2007–2009), we did not explore the presence of any trends in quality over time, because trends would have likely been confounded with observed differences among the groups of practices allocated to the different steps in the stepped wedge design
. Lastly, this study had a measurement bias. As with all studies relying on chart audits, we could only assess guideline adherence through implication - activities performed but not charted would not have been captured by our methodology. This is less of a problem when adherence is measured by the performance and interpretation of a pathology test, than when dependent on documentation of clinical activities (such as providing advice to smokers to quit). Another limitation of using medical chart data was that we were unable to determine whether patients actually complied with medical drug prescriptions that were given to them by their provider, a factor that may have played a role in patients being unable to reach clinical targets. Notwithstanding the above limitations, chart audits are a rich source of information and remain the standard for capturing process of care data, as alternative methods of data collection at the practice are expensive and not feasible for large trials such as this.