Twenty-First Century Solutions to Increase Medication Optimization and Safety in Kidney Transplant Patients

• kidney transplantation
• medication optimization
• medication safety
• pharmacists

In 2019, 7% of adult kidney transplant recipients experienced acute rejection within 1 year of transplant; a higher percentage (9%) was experienced by younger recipients (18–34 years old) (1). Optimizing immunosuppressant and other medications by improving adherence and preventing medication errors in the post-transplant period are likely key factors in reducing graft loss and hospitalization (2,3). Assuring that immunosuppressive therapy is safe, is effective, and can be taken by the patient as intended in the context of their other conditions and medications is crucial to reducing immunosuppression adverse events and poor health outcomes. There is a need for innovative, cost-effective solutions to reduce medication therapy problems in patients with kidney transplants to reduce graft loss, reduce hospitalizations, and improve survival.

In this issue of CJASN, Gonzales et al. (4) evaluated the effects of a pharmacist-led digital health intervention in patients with kidney transplants compared with usual care in a 1-year, randomized controlled, parallel-arm, semiblinded study (n=136) on medication errors, adverse events, and hospitalizations. Adult kidney transplant recipients who were between 6 and 36 months post-transplant were eligible. Control and intervention patients received usual care, including pharmacist care during the transplant hospitalization and for 6 months post-transplant. Patients in the intervention group received supplemental clinical pharmacist–led medication therapy monitoring and management between 6 and 36 months post-transplant utilizing a smartphone application (app) compared with the usual care group. The app allowed for two-way communication between the pharmacist and transplant recipient and captured home-based (BP, blood glucose, medication administration, and medication adverse events) and electronic health record (EHR; tacrolimus drug levels, health care appointments, and medication refill data from SureScripts) data (5). Data were compiled in a Health Insurance Portability and Accountability Act-compliant web-based portal to allow pharmacists to identify high-risk patients and determine clinical interventions and frequency of follow-up telepharmacy visits. The app allowed patients to easily call the transplant center and pharmacy; to send email alerts to the pharmacist when a medication changed; and to document new medications, health care visits, and side effects or transmit BP or blood glucose information from Bluetooth home monitoring devices. The patient’s medication list was automatically updated when changes were made in EHR. No patient-level data were stored on the device. Information on several medication safety risk factors (tacrolimus level variability, systolic BP, blood glucose, missed appointments, self-reported medication adherence, and patient symptoms) was collected. The pharmacist was alerted when a safety threshold was surpassed based on their standard operating procedure manual risk factor scoring system. This information guided frequency of telepharmacy visits. The study measured the number and severity of medication errors and adverse events as well as hospitalizations (4).

Gonzales et al. (4) showed that medication errors were reduced by more than half in the intervention versus control group. There were differences between groups in some important baseline characteristics that could have affected results, so they adjusted for some (but not all) of these factors in multivariable models. They demonstrated a significant 61% reduction in the incident risk rate of medication errors over the 12-month period. They also showed that the intervention group had a significant 45% lower incident risk of grades 3–5 adverse events. A grade 3 adverse event is a severe or medically significant adverse event that is not immediately life threatening, but may result in a hospitalization or prolonged hospitalization. A grade 4 adverse event is a life-threatening event where urgent intervention is indicated, and grade 5 is death related to the adverse event. Unadjusted hospitalizations per patient-year were also fewer (1.08 versus 0.65). After adjustment, the intervention showed a significant 54% reduction in hospitalization as compared with control. The primary causes of hospitalization were infection, AKI, and cardiovascular- or gastrointestinal-related conditions, which mirrored the types of adverse events reported.

Gonzales et al. (4) have demonstrated that a multifaceted pharmacist-led intervention including an app interfaced with the patients’ EHR and providing a patient and pharmacist portal for two-way communication significantly reduced medication errors, adverse events, and hospitalizations in adult patients with kidney transplants. This study moves beyond other investigations, which have shown varying results on medication adherence, BP control, and postmyocardial infarction outcomes using apps that only engage patients unilaterally (7,8). This study adds to a growing understanding that improving important patient outcomes, such as medication adherence (9), as well as reducing hospitalizations in patients with transplants requires a multifaceted approach and two-way communication between patients and health providers that promote trusted therapeutic relationships.

This study had several strengths. It was a well-designed study; the study team had generated preliminary data and patient/provider experiences that supported the study design, sample size determination, and design of and functionality of the app. Only two patients (intervention group) discontinued the study, and they were included in the intent-to-treat analysis. Blinded researchers determined medication errors, adverse effects, and severity independent of patient care pharmacists. Importantly, they evaluated process (number of medication errors and adverse events), intermediate outcome (reduction in medication errors and grades 3–5 adverse events), and health outcome (hospitalization) metrics. Linking changes in intermediate outcomes to health outcome metrics is critical.

There were limitations to the generalizability of this single-center study. Of 774 adult transplant recipients assessed for eligibility, 501 were excluded; the most common reason for ineligibility was not meeting the post-transplant study window of 6–36 months (specific numbers are not given) (4). Of 273 eligible patients, 137 declined to participate—a portion of those felt uncomfortable with technology after training, but numbers were not given, calling into question how easy the app was to use. Patients were required to self-report medication adherence through the app; given the number of medications that patients with kidney transplants are prescribed, this could be a time-consuming endeavor. Integration of other digital health technology to track adherence (e.g., smart pill bottles or containers) may increase objectivity in measuring adherence and reduce patient load (9). The authors pointed out that they did not use educational text messaging or other “attention control” tools in control patients and did not capture patient education level. Individual components of the intervention (mobile health app, pharmacist-led care, and increased patient attention) could not be separately evaluated to determine their direct effect on study outcomes.

The small sample size contributed to significant clinical differences in study participants’ baseline characteristics, which may have affected study results. There were several characteristics suggesting that the intervention group recipients or donors were healthier at baseline (4). Compared with controls, fewer intervention subjects had diabetes (28% versus 52%), intervention subjects were more likely to be on peritoneal dialysis than hemodialysis, donor age was younger, Kidney Donor Risk Index was lower, and recipients were less likely to be cytomegalovirus positive at transplant. However, higher percentages of intervention participants were more likely to experience delayed graft function and have donor-positive, recipient-negative cytomegalovirus serostatus. Some but not all of these characteristics were adjusted for in multivariable models. Other exploratory end points that investigators were going to evaluate included acute rejection, infections, graft function, and graft failure, but they were not reported. The study was underpowered for these outcomes, but it would be important to evaluate trends to inform future research.

The authors note that an economic analysis is pending, which will include inpatient, outpatient, device, and data plan costs as well as personnel costs to train patients to use the app and pharmacist time. Demonstration of cost-effective data-driven pharmacist care utilizing digital health services will be instrumental in enhancing value-based care in patients with kidney diseases. The Centers for Medicare and Medicaid Services recently launched a new value-based Kidney Care Choice Model that incentivizes kidney transplantation, including a bonus for each beneficiary who receives a kidney transplant (10). The bonus (maximum of $15,000) is paid out over 3 years if the transplant remains successful. It would also encourage a more standardized approach of pharmacist care post-transplant. Transplant centers have multidisciplinary transplant teams, including pharmacists, but the roles and responsibilities of pharmacists vary between transplant centers for pre- and postkidney transplant care.

Additional data from this study would be helpful for other transplant centers to implement and integrate similar digital technology and pharmacist services in transplant programs. Details on workflow transformation and implementation issues may be of benefit for institutions looking to engage in similar digital health services in transplant care. These details could include time metrics on activities related to intervention or time savings from this new process. Further information on patient use experiences, such as performing and uploading their BP and blood glucose reading and personal medication data, related to the app would be useful. Information on how their organization or pharmacy department created buy-in to engage practitioners and patients in this app platform and new service would also be helpful. Finally, many will be interested to know if this app will be available free of charge or on the open market.

Overall, this was a well-designed and well-conducted study that leveraged a novel digital health technology to engage patients in their own self-care management and integrate pharmacist expertise in addressing medication therapy problems to reduce medication errors, adverse events, and hospitalizations in adult patients 6–36 months postkidney transplantation. We are eagerly awaiting the economic analysis that determines the return on investment for incorporation of these new approaches into value-based care of these patients.

• Copyright © 2021 by the American Society of Nephrology