Implementation of a pragmatic randomized trial of screening for chronic kidney disease to improve care among non-diabetic hypertensive veterans

Following enrollment, PCPs randomized to usual care continue to manage the patients as usual. No systematic screening for CKD among non-diabetics is performed at SFVAMC. Referral to the PACT pharmacist remains an option, as this is part of usual care within the PACT model. Triple-marker screening for CKD is clinically available, and providers in the usual care arm are able to order these lab tests if deemed appropriate.

We designed each intervention design to be feasible and to follow the usual workflow. For included patients, we identify their upcoming appointments using CDW data on a weekly basis. Study staff place orders for a serum creatinine, serum cystatin C and urinary albumin to creatinine ratio. All tests are performed locally at the clinical lab. Serum and urine creatinine are measured by the Jaffe method, cystatin C is measured using the particle-enhanced cystatin C assay from Gentian [26], and the urine albumin by nephelometry. Tests are ordered prior to the next scheduled visit with their assigned PCP. The study staff ensures that the PCP receives the test results, along with the CKD stage (or no CKD), and a summary of KDIGO guidelines for BP management appropriate for the specific CKD stage in the form of an electronic note which the PCP has to co-sign. These electronic notes also include a reminder to counsel on NSAID use, and the use of ACE/ARB in persons with albuminuria. (Additional file 1) All participating patients in the intervention arm receive notification of test results by mail. All persons with screen-detected CKD and their PCPs will receive educational materials available from the National Kidney Disease Education Program (NKDEP) as a web link on the note or letter they receive by mail.

In addition to the screening and education as above, providers randomized to this arm will have the option to refer patients with screen-detected CKD to be co-managed by a clinical pharmacist. The clinical pharmacists at the VA hold doctoral degrees (PharmD) and they are active members of the care teams, as described above. The pharmacists already provide consultation for BP management at SFVAMC when referred by PCP. At SFVAMC, pharmacists are authorized to initiate therapy, adjust doses, monitor complications and order labs tests, under the supervision of the attending physician. Specific to this study, when the triple marker screen identifies CKD, research staff will notify providers through a note in the EHR that pharmacist co-management is available for their patient and they can refer to the pharmacist through the electronic consult request for CKD and hypertension management and education. Physicians in the other arms may still refer to pharmacy for other indications, but they are not notified of the availability of this service. Upon receipt of referral, the pharmacist schedules a series of appointments aimed at education on CKD and its complications, as well as education on avoiding NSAID use. The visit also focuses on BP control and initiation or up-titration of ACE/ARB therapy in persons with albuminuria. At each visit, the pharmacist measures and records two seated BP readings, and will perform medication inventory and review adherence. Education regarding diagnosis of CKD is based on previously designed materials provided by NKDEP that include lifestyle modification. Management follows evidence-based algorithms with the goal of achieving BP <140/90 mmHg and using ACE or ARBs for persons with albuminuria. Lower BP goals can be requested by the PCP. Initially, pharmacist follow-up visits are every 2 weeks, and once a patient’s BP is controlled on2 or more consecutive visits, follow-up ends. The pharmacist orders a chemistry panel to monitor for hyperkalemia or rising creatinine if ACE/ARB is initiated or uptitrated. If a patient is found to have eGFR <30 ml/min/1.73 m², were commend nephrology referral for pre-dialysis care. At the end of 12 months, the pharmacist will stop following patients. To verify the fidelity of the educational intervention, the pharmacist presents several randomly chosen cases to the nephrologist involved in algorithm design. The pharmacists collaborated on creating the CKD visit algorithm along with study MDs (Additional file 2).

We identified 2,293 patients at the SFVAMC who met initial criteria. A total of 114 were either assigned to the provider who did not consent or to a study physician (M.S) and were thus excluded. Providers excluded an additional 138 patients after reviewing their patient list. We identified 36 patients with an EF < 40%, but 9 of these patients had already been excluded by their PCP, and 2 persons who had died since the data pull, and who had not previously been excluded by PCP. Since mailing of the information letters to 2012 patients, we received 141 telephone calls. Of these, 73 called to opt out. The rest of the calls reflected interest in the program and wanting to participate or alerting us to upcoming appointments. We have had only 1 phone call from a veteran stating that receiving the information letter caused great anxiety. We excluded a total of 193 persons who opted-out or had no available address and no other contact information. A total of 1,819 patients are included in this trial (Fig. 2).

The EHR administrative data accurately identified persons for inclusion and exclusion in this trial. Among 50 charts reviewed, one person had been living at the nursing home and had incomplete information to assess all the comorbidities. Of the remaining 49 charts, 45 (92%) had confirmed hypertension by the reviewers, and agreement for diabetes status was 100%. Agreement of EHR and physician review is shown in Table 2. While we found good agreement for coronary artery disease, we found less agreement for cerebrovascular disease and congestive heart failure, where chart review was more likely to identify these diagnoses that the EHR.

Overall, the median age (interquartile range) of 1, 819 included participants was 68 years (61–72), and only 8 were female. Approximately 16% are Black, and 5% Hispanic. We have found that 18% of participants have active prescription for NSAIDS, and 35% were on ACE/ARB at study start. We found that the majority of participants had a prior serum creatinine tested, and about 50% of participants had a prior urinary dipstick result in the chart. Less than 10% of persons had an albumin to creatinine ratio (ACR) tested prior to randomization. Characteristics of participants included in the study vs. those were excluded are presented in Table 3. Overall, excluded patients were similar in age, demographics, and comorbidities. Excluded persons were less likely to be using an anti-hypertensive medication, were somewhat less likely to have prior testing for proteinuria, and had a higher prevalence of drug use disorders. In the period between data pull and randomization, a total of N = 34 patients had new diagnoses of diabetes, and N = 11 had a new diagnosis of CKD among participant included in the study, but these were relatively evenly distributed by arm.