Clinical Decision Support Tools for Reduced and Changing Kidney Function

Kidney disease can alter the pharmacokinetics and pharmacodynamics of numerous medications. Failure to dose adjust medications during kidney disease can lead to compromised drug effectiveness and safety (1). Approximately two thirds of medications are excreted extensively by the kidneys and may require dose adjustments to address the reduced drug clearance in kidney impairment (2,3). Renal dosing is complex and involves identifying a patient with decreased or changing kidney function, recognizing an ordered medication that requires dose adjustment, and selecting an individualized dose on the basis of patient- and treatment-related factors (4). Clinical decision support systems (CDSS) can optimize this process through automation.

In this issue of Kidney360, Sonoda et al. provided compelling evidence of the benefits of CDSS for identifying and managing medications in patients with kidney dysfunction (5). In their investigation, an inpatient prescription checking system was implemented for four renally eliminated medications associated with a risk for adverse drug events. The CDSS intervention added the text [Renal] before the medication name and provided the patient’s creatinine clearance and serum creatinine on the printed inpatient prescription for pharmacist review. To streamline identification of appropriate dose adjustments, pharmacists were also provided with a printed sheet listing renal dose recommendations for the medications of interest. The authors found that the new CDSS and resultant pharmacist actions significantly reduced the dosage error rates from 26% to 3% (P<0.001). These findings are of particular interest because they indicate that passive interventions that do not require extensive information technology (IT) resources for implementation can increase clinician awareness of the need for renal dose adjustments and meaningfully reduce error rates.

Resource availability is an important consideration with any renal dosing workflow. The methodology described by Sonoda et al. required pharmacists to perform complex dosing evaluations for each medication order in the study. Although successful in this evaluation of four select drugs, when expanded to all candidate renally eliminated medications, such a workflow may be prohibitive. This challenge is further exacerbated when considering the myriad of factors that a pharmacist must evaluate when reviewing medication orders. Providing nontargeted information could become overwhelming amid the many other order- and patient-specific factors a pharmacist must evaluate before dispensing a medication. For example, one of the drugs targeted in the Sonoda et al. article was pilsicainide, an antiarrhythmic that functions through sodium channel inhibition. In addition to evaluating for renal dose adjustments, a pharmacist would also be responsible for evaluating drug appropriateness, contraindications, electrolyte levels, electrocardiogram monitoring, and drug interactions that could lead to dangerous dysrhythmias. If the dose selected was indeed appropriate for kidney function, adding the “[Renal]” designation as proposed along with the printout would lead to additional review without benefit. Rather, an efficient renal dosing intervention in this situation would only flag the pharmacist when an issue needs to be addressed before medication verification, allowing for greater efficiency.

Although the benefits of renal dosing CDSS have been well established in the literature, there is significant heterogeneity among previous investigations in the level of automation utilized. Some systems provide a general flag to promote clinician review for the potential need to dose adjust renally (5–8). Others provide individualized dosing recommendations within the alert (9). The most sophisticated tools recommend a replacement order that is preemptively adjusted for level of kidney function (10). There are benefits and drawbacks to each kind of CDSS. Electronic systems of greater complexity require more IT resources, whereas those of lower complexity often require greater clinician resources.

Our institution has implemented a suite of renal dosing CDSS (Figure 1). To adjust medication doses appropriately at the point of entry, a sidebar directly within the order entry screen displays renal dosing thresholds for the medication of interest, along with the most recent serum creatinine, creatinine clearance, and dialysis modality, if ongoing. In select cases, the default dose recommended at the point of order entry is prepopulated on the basis of the patient’s kidney function. Each drug ordered by a provider is verified by a pharmacist. At the point of verification, the pharmacist has access to similar objective data about the patient’s kidney parameters and the suggested dose adjustment thresholds. During therapy, rules are used to screen a medication dose for appropriateness on the basis of the indication, the patient’s kidney function, dialysis use, age, and weight. In cases of potential dose errors, the pharmacist receives an alert and recommendation for an alternative strategy. Dynamic kidney function (both worsening and improving) is detected with a clinical score to identify patients in need of emergent review (11). A suite of CDSS tools such as this has great potential to improve the effectiveness and safety of renally dosed medications. Providers are engaged in dose optimization at the point of order entry. Pharmacists are presented with support at order verification and during iterative review. The IT resources needed to deploy this intervention are more extensive than those required by Sonoda et al.; however, our approach is comparatively less clinically burdensome, which may be more scalable to a large portfolio of medications and dynamic clinical contexts.

Figure 1.

Visualization of patient monitoring tools. (A) An example of the order entry screen displaying renal dosing thresholds, most recent serum creatinine, creatinine clearance, dialysis modality, and predefaulted dosing selections on the basis of kidney function. (B) The information pharmacists review during order verification. Pharmacists have access to nearly identical dosing information. (C) The tools used for ongoing monitoring. Patients with dosing errors are identified on the pharmacist Clinical Monitoring List. Expanding the alert allows the pharmacist to view the current dosing of the medication and the recommended dosing adjustment. Functionality shown adapted from © Epic Systems Corporation.

The article by Sonoda et al. highlighted the clear importance of CDSS in improving the safe and effective use of renally dosed medications. The proposed intervention is straightforward, does not include complex logic, requires limited IT resources, and clearly decreased dosage errors. Future efforts should continue to build upon this work to limit potential undue cognitive burden incurred by clinicians with nontargeted CDSS. Renal dosing recommendations are excellent candidate drugs for automation. Our institution has successfully implemented frameworks that assist with initial dose selection and regimen tailoring throughout therapy. Regardless of the explicit strategy used, CDSS clearly offers a mechanism to reduce errors associated with renally dosed medications.

Disclosures

E.F. Barreto reports research funding from the Agency for Healthcare Research and Qualit and the National Institute of Allergy and Infectious Diseases; honoraria from Vifor Pharma; and an advisory or leadership role for FAST Biomedical (advisory board, paid as needed for consulting services) and Wolters Kluwer (paid as needed for consulting services). The remaining author has nothing to disclose.

Funding

E.F. Barreto received funding from the National Institute of Allergy and Infectious Diseases (K23AI143882).

Author Contributions

D.J. Schreier wrote the original draft of the manuscript, and both authors were responsible for the conceptualization and reviewed and edited the manuscript.