Automated clinical reminders for primary care providers in the care of CKD: a small cluster-randomized controlled trial

Abstract

Background: Primary care physicians (PCPs) care for most non-dialysis-dependent patients with chronic kidney disease (CKD). Studies suggest that PCPs may deliver suboptimal CKD care. One means to improve PCP treatment of CKD is clinical decision support systems (CDSSs).

Study design: Cluster-randomized controlled trial.

Setting & participants: 30 PCPs in a university-based outpatient general internal medicine practice and their 248 patients with moderate to advanced CKD who had not been referred to a nephrologist.

Intervention: 2 CKD educational sessions were held for PCPs in both arms. The 15 intervention-arm PCPs also received real-time automated electronic medical record alerts for patients with estimated glomerular filtration rates <45 mL/min/1.73 m(2) recommending renal referral and urine albumin quantification if not done within the prior year.

Outcomes: Primary outcome was referral to a nephrologist; secondary outcomes were albuminuria/proteinuria assessment, CKD documentation, optimal blood pressure (ie, <130/80 mm Hg), and use of renoprotective medications.

Results: The intervention and control arms did not differ in renal referrals (9.7% vs 16.5%, respectively; between-group difference, -6.8%; 95% CI, -15.5% to 1.8%; P = 0.1) or proteinuria assessments (39.3% vs 30.1%, respectively; between-group difference, 9.2%; 95% CI, -2.7% to 21.1%; P = 0.1). For intervention and control patients without a baseline proteinuria assessment, 27.7% versus 16.3%, respectively, had one at follow-up (P = 0.06). After controlling for clustering, these findings were largely unchanged and no significant differences were apparent between groups.

Limitations: Small single-center university-based practice, use of a passive CDSS that required PCPs to trigger the electronic order set.

Conclusions: PCPs were willing to partake in a randomized trial of a CDSS to improve outpatient CKD care. Although CDSSs may have potential, larger studies are needed to further explore how best to deploy them to enhance CKD care.

Trial registration: ClinicalTrials.gov NCT00688285.

Figure 1

Sample clinical decision support alert

Figure 2

Physician enrollment GIM general internal medicine

Figure 3

Process of care outcomes at study completion, showing percentage and 95% confidence intervals. ^†P = 0.1, *P = 0.008 ACEI angiotensin converting enzyme inhibitor, ARB angiotensin receptor blocker, CKD chronic kidney disease, BP blood pressure, NSAID non-steroidal anti-inflammatory drug.

Figure 4

Intervention group odds of receiving specified care at study completion.* *Analysis is restricted to patients in each arm who did not meet the optimal specified care parameter at baseline assessment. CI confidence interval, ACEI angiotensin converting enzyme inhibitor, ARB angiotensin receptor blocker, CKD chronic kidney disease, BP blood pressure, NSAID non-steroidal anti-inflammatory drug