Optimizing Audit and Feedback for Primary Care

• Primary Process Outcome: A composite score counting the number of appropriate processes achieved for patients within appropriate time-frames, including laboratory testing, blood pressure measurements and active prescriptions [ Time Frame: 24 months ] [ Designated as safety issue: No ]
• Primary Disease-Related Outcomes: mean BP and mean LDL level. [ Time Frame: 24 months ] [ Designated as safety issue: No ]

• Prescriptions: percent of patients on insulin, beta blockers, ace-inhibitors or angiotensin-receptor blockers, aspirin, statins. [ Time Frame: 24 months ] [ Designated as safety issue: No ]
• Subgroup Analysis: 1) Compare effect of intervention for patients with only diabetes, only coronary artery disease, or both. 2) Compare outcomes for those with high vs low baseline performance. [ Time Frame: 24 months ] [ Designated as safety issue: No ]
• proportion of patients who are meeting targets recommended in guidelines for glucose, blood pressure and LDL cholesterol [ Time Frame: 24 months ] [ Designated as safety issue: No ]
• Efficacy of intervention [ Time Frame: 24 months ] [ Designated as safety issue: No ]
  The direct efficacy of the goal-setting and action-plan worksheet will be assessed by 1) stratifying results by physicians who properly completed the worksheet and those who did not and 2) determining if improved outcomes were more often associated with clinical topics that providers specified in the worksheet.

• Prescriptions: percent of patients on insulin, beta blockers, ace-inhibitors or angiotensin-receptor blockers, aspirin, statins. [ Time Frame: 12 months ] [ Designated as safety issue: No ]
• Clinical parameters: mean blood pressure, mean low density lipoprotein (LDL), mean A1C. [ Time Frame: 12 months ] [ Designated as safety issue: No ]

There remains a large gap between ideal and actual care provided to patients with chronic diseases. Performance feedback reports are often used as a foundation for quality improvement interventions. There have been hundreds of trials investigating the use of feedback reports; the important question to ask now is not whether performance feedback reports can help to improve quality of care, but how to optimize feedback interventions to accomplish that goal. The purpose of this study is to test whether a theory-based intervention added to feedback reports sent to primary care providers can result in improved outcomes for patients with chronic disease.

There remains a large gap between ideal and actual care provided to patients with chronic diseases such as type-2 diabetes (DM2) and coronary artery disease (CAD), making them a common focus for translational research. Quality improvement strategies such as audit and feedback, which can increase adoption and adherence to guidelines, may be a good choice for such problems. Audit and feedback is felt to be effective because it may overcome physicians' limited ability to accurately self-assess. A recent Cochrane review concluded that it may increase appropriate clinical practice by 5 to 10 percentage points. However, the authors noted great variability in the results of the trials. It is thought that optimal design and delivery of feedback will more consistently lead to improved results, but few studies have tested different designs of feedback. A more recent meta-analysis of feedback found that interventions incorporating theoretical principles from psychology in the design were more effective than intuitive feedback designs. The author of that study has urged the creation of more 'actionable' feedback. There is both theoretical and empirical reason to believe that feedback will be more effective if the recipients set goals and develop action plans.

A group of family physicians across Ontario have previously signed data sharing agreements with ICES allowing their electronic medical records to be audited. These physicians will be invited to participate in this trial. Participants will receive performance feedback reports summarizing the percentage of their DM2 and CAD patients who are meeting evidence-based targets for quality of care.

Participating physicians will be split into two groups at the level of their practice to reduce risk of contamination. Minimization software will be used to ensure balance at baseline in the primary outcomes between the intervention and control group. New potential participants that have signed data sharing agreements at ICES will be invited to join the trial for a maximum of six months after the first practices are allocated.

The intervention group will receive in addition to the feedback reports a theoretically informed worksheet meant to facilitate goal-setting and the development of action plans. This will occur every six months for two years. Outcomes will be analyzed after 24 months.

• Diabetes Mellitus, Type 2
• Coronary Artery Disease

• No Intervention: Feedback report only
  This arm will receive performance feedback reports but no worksheet to facilitate goal-setting and action plans.
• Experimental: Goal-Setting Worksheet
  This arm will receive a theory-informed worksheet to facilitate the development of goals and action plans in response to the performance feedback reports.
  Intervention: Other: Worksheet to facilitate goal-setting and action plans

Inclusion Criteria:

• family physicians in Ontario who have signed data sharing agreements with the CCORT-ICES EMR team

Exclusion Criteria:

• specializing in an area (such as palliative care) where general guidelines do not apply

• Institute for Clinical Evaluative Sciences
• Canadian Institutes of Health Research (CIHR)