Evaluation of Pharmacist Renal Dose Adjustments and Planning for Future Evaluations of Pharmacist Services

Abstract

Background: Clinical pharmacy services are known to improve patient outcomes. Pharmacists contribute to patient care in the acute care setting in multiple ways, including providing advice and information to patients and the health care team, performing medication histories to prevent waste and support medication adherence, analyzing the cost-effectiveness of medications, and ensuring patient safety through patient monitoring and medication review. Specific clinical pharmacist services include managing intravenous to oral medication adjustments, renal dose adjustments, and performing pharmacokinetic dosing of medications, among others. Many of these clinical services are performed daily but are not evaluated for clinical quality or compliance with policies. Evaluating these clinical services may provide a multitude of benefits to pharmacy departments, health systems, and patients. Methods: The purpose of this study was to evaluate pharmacist use and percent compliance of a renal dose adjustment policy upon initial order verification and discharge. This was completed through retrospective chart review to determine if dose adjustments were made appropriately and descriptive statistics were used to establish pharmacist compliance. Those orders that were inappropriately adjusted were analyzed for trends that could lead to possible policy improvements or pharmacist education opportunities. The completed evaluation also led to the development of an evaluation system that can be utilized to routinely assess clinical pharmacist services. Conclusions: The results of this study are being used to develop and support future clinical service evaluations, inspire process improvements, and improve patient outcomes and pharmacist accountability.

Introduction

Clinical pharmacy services are known to improve patient outcomes. ¹ Pharmacists contribute to patient care in the acute care setting in many ways, including providing medication information and recommendations to the health care team and patients, performing medication reconciliation and gathering medication histories, analyzing the cost-effectiveness of medications, and ensuring patient safety through medication monitoring and review. One study in Northern Ireland found that comprehensive medication management services in hospitals led to reductions in patient length of stay and readmission rates. ² In the United States, studies have also demonstrated a reduction in medication errors within hospitals that have increased pharmacy services including drug information services, drug protocol management, and medical round participation among others. ³

When assessing the role of clinical pharmacy, it can be very hard to determine quality, as patient outcomes are influenced by numerous members of the interdisciplinary team. ⁴ A 2001 study that obtained data from US national databases established similar findings, noting that shorter lengths of stay were associated with pharmacist-provided drug protocol management for unknown reasons, but speculating that this could be due to pharmacy services providing quality control for patient therapies. ⁵ They also discovered that 17 different clinical pharmacy services showed improvements in length of stay, drug costs, total cost of care, and mortality rates. ⁵

Various pharmacy services, such as adjusting medication doses and routes, performing pharmacokinetic dosing, and ensuring cost-effectiveness for patients and health systems, are routinely completed but are not often evaluated for pharmacist compliance with policies. Evaluating these clinical services may provide a multitude of benefits to pharmacy departments, health systems, and patients. Assessing compliance may determine if policies and services are optimized to improve workflow and patient care. Evaluations may spark quality improvement projects for departments to undertake, and they may also provide evidence of benefits for a health system, whether in positive impact to patient outcomes, cost savings, or improved interdisciplinary teamwork.

This study was conducted at a single health system, where the acute care pharmacy department utilizes policies across 4 acute care hospital sites (3 tertiary and 1 long-term acute care). The pharmacy department does not currently have a system in place to evaluate pharmacist compliance with the services the department provides, even though the services are performed daily. The purpose of this study is to perform a compliance evaluation of a single policy, the renal dose adjustment policy.

The renal dose adjustment policy, a portion of the Adult Medication Dose Adjustment Policy, contains a list of automatic adjustments for 48 different medications. It allows for pharmacists to adjust medication dosages based on a patient’s renal function and other criteria without contacting the prescriber. The policy was developed by the pharmacy department and approved through the system pharmacy and therapeutics committee, a team of pharmacists, physicians, and nurses. Manufacturer recommendations and resources such as American Hospital Formulary Service (AHFS) Drug Information and the National Institute of Diabetes and Digestive and Kidney Diseases were used to develop the policy. Newly hired pharmacists are trained on use of the policy, and competency is validated through use of an online learning management system.

Pharmacists utilize the renal dose adjustment policy daily to improve patient care and prevent adverse drug events or other effects from incorrect medication dosage, including readjusting doses in patients with fluctuating renal function. Despite its frequent utilization, compliance with the policy is not routinely assessed and has not been officially evaluated. This study evaluates pharmacist compliance with the renal dose adjustment policy, establishing a baseline for a clinical service that is utilized frequently within the acute care setting.

Adjusting medication doses for renal function is arguably one of the most important interventions a clinical pharmacist can make. According to the National Institute of Diabetes and Digestive and Kidney Diseases, approximately 14% of the population has chronic kidney disease. ⁶ The high prevalence has remained steady since the early 2000s, and the risks associated with chronic kidney disease, including cardiovascular disease, hyperlipidemia, anemia, and metabolic bone disease, are well established.^6,7 It is also widely believed that pharmacists can make a difference in patient outcomes by performing renal dose adjustments for patients with chronic kidney disease; however, there is not much existing literature regarding the quality of these adjustments. ⁸ A study by Altunbas et al from 2016 evaluated the appropriateness of renal dose adjustments made in a single hospital, examining heart failure patients upon hospital discharge and found that 12.6% of eligible prescriptions were inappropriately prescribed based on the estimated glomerular filtration rate (eGFR). ⁹ Sheen and Choi looked at rates of medication “overdoses,” defined as medications that should have been given at a lower dose based on a patient’s renal function. They found an overall overdose rate of 5.3% over 4 years; however, the role of the pharmacist in renal dosing at this institution was not discussed. ¹⁰ Hassan et al found a decrease in physician noncompliance with dosing guidelines from 53% to 27.5% with the addition of a pharmacist attending medical rounds and offering dose recommendations. ¹¹ A 2019 study from Lebanon also examined the appropriateness of medication doses for renal dysfunction ordered by prescribers, finding high rates of inappropriate doses, and further emphasizing the need for pharmacist involvement to improve patient care and prevent prescribing errors. ¹²

An additional aspect to evaluating the compliance of renal dose adjustments performed by pharmacists during hospitalization is ensuring these adjustments are appropriately carried over to a patient’s discharge orders. Michaelsen et al explored the importance of performing medication reconciliation upon discharge from the hospital and discovered that a large portion of discrepancies relate to a prescription’s route, dose, or frequency. ¹³ This is an important consideration for many clinical service evaluations; if interventions are not carried out at discharge, the impact on patient outcomes may be significantly less.

A secondary purpose of this study is to create an evaluation system that can be utilized to routinely assess the compliance of clinical pharmacy services. Creating this system, however, would prove difficult without first performing an analysis of a specific policy or service, as the feasibility and scope of the various evaluations is relatively unknown. Completing an evaluation of the renal dose adjustment policy provides the basis to create this system. The goal for this system is to include a stepwise approach that can be applied to evaluate a variety of services. This would also include a timeline for initial evaluations and define a frequency for repeat evaluations.

Methods

The examination of pharmacist renal dose adjustments was completed through a retrospective chart review. The scope of this chart review included 4 acute care community-based hospitals within a tertiary health system and spanned from January 1 to June 30, 2018. For each medication listed in the renal dose adjustment policy, data were extracted by the medications ordered from the electronic medical record (EMR). The medication order may have originated at any point during the patient’s stay. Medication orders were included for patients at least 18 years of age with a serum creatinine of greater than 0.6 mg/dL across all patient care areas. This serum creatinine value was chosen to collect all potential creatinine clearance values that can result in a dose adjustment, as a variety of other patient characteristics (age, height, and weight) also affect creatinine clearance calculations.

Medication orders were excluded for patients who did not have a serum creatinine value recorded within the prior 72 hours; the renal dose adjustment policy uses this timeframe as the acceptable limit for a serum creatinine to be used in evaluating patients without fluctuating renal function. Patients who did not have a body weight recorded were also excluded as a creatinine clearance value could not be calculated at the time of medication ordering. In addition, medication orders prescribed as dispense as written (DAW) or placed by an infectious disease service provider or a nephrology service provider were excluded. The renal dose adjustment policy does not allow for automatic dose adjustment of orders entered by these specialty providers; pharmacists must contact the provider to obtain their consent before changing the order.

The data extraction included patient demographics such as age, actual body weight, gender, and race. Information about the patient’s hospital stay, including the length of stay, patient care unit, diagnoses, and most recent serum creatinine value, was also included, as was information specific to the medication order; medication name, dose, frequency, ordering date, discontinuation date, ordering provider, provider specialty, and verifying pharmacist. Unfortunately, the data extraction tool was not able to extract the EMR calculated creatinine clearance values, so a creatinine clearance was calculated from the pulled patient demographics using the Cockcroft-Gault equation. ¹⁴ The renal dose adjustment policy dictates use of this equation by pharmacists to approximate patients’ renal function when a measured creatinine clearance is not available; it is also the equation used in the EMR.

For each medication, the initial data set was filtered to include only orders eligible for renal dose adjustment based on the estimated creatinine clearance. Orders were included if the estimated creatinine clearance was below the creatinine clearance cutoff for each medication listed in the renal dose adjustment policy, with a 10% or 3-mL/min buffer added to the cutoff (whichever was less). This buffer allowed for inclusion of orders that may have fallen into an area of pharmacist clinical discretion for adjustments. If the dose adjustment per policy only involved adjusting the dosing interval, then orders with a frequency of “once” were removed from the list of eligible orders, as pharmacists would not have intervened on those orders. The total number of medication orders after applying these filters and all inclusion and exclusion criteria are reported as the number of orders eligible for renal dose adjustment in Table 1.

Table 1.

Medications Evaluated From the Adult Medication Dose Adjustment Policy.

Medication	Total number of orders	Number of orders randomly sampled	Inappropriately high dose	Inappropriate dose reduction	Percent compliance
Acyclovir	41	10	0	0	100
Allopurinol	199	20	0	0	100
Amantadine	5	5	2	0	60
Amoxicillin	7	7	1	0	85.7
Amoxicillin/clavulanate	55	10	3	0	70
Ampicillin	8	8	3	0	62.5
Ampicillin/sulbactam	0	—	—	—	—
Aztreonam	27	10	0	0	100
Cefazolin	73	10	7	0	30
Cefdinir	25	10	1	0	90
Cefepime	31	10	1	0	90
Cefotaxime	2	2	0	0	100
Cefpodoxime	0	—	—	—	—
Ceftaroline	0	—	—	—	—
Ceftazidime	7	7	1	0	85.7
Ceftazidime/avibactam	0	—	—	—	—
Cefuroxime intravenous	0	—	—	—	—
Cephalexin	95	10	1	1	80
Ciprofloxacin	62	10	2	0	80
Clarithromycin	1	1	0	0	100
Daptomycin	4	4	0	0	100
Demeclocycline	0	—	—	—	—
Enoxaparin	115	12	0	0	100
Ertapenem	0	—	—	—	—
Famotidine	500	50	11	0	78
Fluconazole	41	10	0	0	100
Fondaparinux	4	4	1	0	75
Ganciclovir	1	1	0	0	100
Ketorolac	57	10	0	1	90
Levofloxacin	381	30	1	0	96.7
Magnesium hydroxide	30	10	6	0	40
Meropenem	142	14	3	0	78.6
Metoclopramide	43	10	2	0	80
Nitrofurantoin	1	1	0	0	100
Oseltamivir	403	40	2	1	92.5
Penicillin G	0	—	—	—	—
Peramivir	0	—	—	—	—
Piperacillin/tazobactam	465	40	3	0	92.5
Posaconazole	1	1	0	0	100
Rifampin	3	3	0	0	100
Sitagliptin	28	10	0	2	80
Sulfamethoxazole/trimethoprim	13	10	1	0	90
Tetracycline	0	—	—	—	—
Tramadol	102	10	7	0	30
Valacyclovir	29	10	3	0	70
Valganciclovir	1	1	0	0	100
Voriconazole	3	3	0	0	100
Zoledronic acid	11	10	0	1	90

Of these resulting medication orders, a random sample of orders were further examined to determine the clinical appropriateness of the medication dose and/or frequency. For any medication that had less than 10 orders, all orders were included in the random sample; for those that had greater than 10 but fewer than 100 orders, 10 orders were included in the random sample; for all medications that had more than 100 orders, the random sample included approximately 10% of the orders. This is shown in Figure 1. A random number generator was used to select the orders to be included.

Figure 1.

Number of medications included in the random sample based on the total number of medication orders.

Once the sample was obtained, chart review was performed and included verifying whether the ordered medication was continued at hospital discharge, and if so, the medication name, dose, frequency, and the patient’s most recent serum creatinine value prior to discharge were obtained to determine dose appropriateness. Pharmacists’ daily duties include reviewing provider-completed discharge medication reconciliation and attempting to resolve any discrepancies prior to a patient’s discharge; the chart review also reviewed documentation from this process if applicable. In determining clinical appropriateness of orders that may have fallen into an area of pharmacist clinical discretion, chart review included reading pertinent order notes, pharmacist interventions, progress notes, trends in renal function, or other aspects of the patient charts to provide clarity on the clinical decision-making process of the verifying pharmacist.

The random samples were analyzed, and descriptive statistics were used to determine pharmacist compliance with the renal dose adjustment policy. Those orders that were inappropriately dose adjusted were analyzed for trends in certain drug classes, medication doses or frequencies, and specific providers or provider groups.

Results

Of the 48 medications analyzed from the Adult Medication Dose Adjustment policy, there were 10 medications that did not have any eligible orders for renal dose adjustment. An additional 14 of the medications had less than 10 eligible orders meeting inclusion criteria. These medications that were not commonly ordered often had a higher percent compliance with appropriate dose adjustment than those medications that were ordered more frequently. Tables 2 and 3 illustrate which randomly sampled medications had greater than or equal to 80% compliance (target chosen based on internal benchmark set by the leadership team) and which medications were less than 80%, respectively. All medications that had less than 30 total orders in the 6-month time period had at least 60% compliance with policy outlined dose adjustments in the random sample, as shown in Table 4. Only 4 of these 20 medications had compliance of less than 80%.

Table 2.

Medications With Compliance Greater Than or Equal to 80%.

Medication	Total number of orders	Number of orders randomly sampled	Percent compliance
Acyclovir	41	10	100
Allopurinol	199	20	100
Amoxicillin	7	7	85.7
Aztreonam	27	10	100
Cefdinir	25	10	90
Cefepime	31	10	90
Cefotaxime	2	2	100
Ceftazidime	7	7	85.7
Cephalexin	95	10	80
Ciprofloxacin	62	10	80
Clarithromycin	1	1	100
Daptomycin	4	4	100
Enoxaparin	115	12	100
Fluconazole	41	10	100
Ganciclovir	1	1	100
Ketorolac	57	10	90
Levofloxacin	381	30	96.7
Metoclopramide	43	10	80
Nitrofurantoin	1	1	100
Oseltamivir	403	40	92.5
Piperacillin/tazobactam	465	40	92.5
Posaconazole	1	1	100
Rifampin	3	3	100
Sitagliptin	28	10	80
Sulfamethoxazole/trimethoprim	13	10	90
Valganciclovir	1	1	100
Voriconazole	3	3	100
Zoledronic acid	11	10	90

Table 3.

Medications With Compliance Less Than 80%.

Medication	Total number of orders	Number of orders randomly sampled	Percent compliance
Amantadine	5	5	60
Amoxicillin/clavulanate	55	10	70
Ampicillin	8	8	62.5
Cefazolin	73	10	30
Famotidine	500	50	78
Fondaparinux	4	4	75
Magnesium hydroxide	30	10	40
Meropenem	142	14	78.6
Tramadol	102	10	30
Valacyclovir	29	10	70

Table 4.

Percent Compliance for Medications With Less Than 30 Orders.

Medication	Total number of orders	Number of orders randomly sampled	Percent compliance
Amantadine	5	5	60
Amoxicillin	7	7	85.7
Ampicillin	8	8	62.5
Aztreonam	27	10	100
Cefdinir	25	10	90
Cefotaxime	2	2	100
Ceftazidime	7	7	85.7
Clarithromycin	1	1	100
Daptomycin	4	4	100
Fondaparinux	4	4	75
Ganciclovir	1	1	100
Nitrofurantoin	1	1	100
Posaconazole	1	1	100
Rifampin	3	3	100
Sitagliptin	28	10	80
Sulfamethoxazole/trimethoprim	13	10	90
Valacyclovir	29	10	70
Valganciclovir	1	1	100
Voriconazole	3	3	100
Zoledronic acid	11	10	90

The evaluation of renal dose adjustments upon transitions of care at discharge found few pertinent findings from the random sample chart review. Some exceptions will be mentioned in the discussion.

Discussion

When examining pharmacist compliance with application of the renal dose adjustment policy, a wide range of compliance was found depending on the medication. Infrequently ordered medications often had better compliance. This contradicts the authors’ hypothesis of the study that infrequently ordered medications would not be appropriately adjusted as pharmacists may not consciously think of renally adjusting those medications. This is possibly due to an unfamiliarity causing pharmacists to look up potential dose adjustments and, therefore, rendering a higher number of renal adjustments.

This study is not without limitations. Some orders eligible for renal dose adjustment may have been missed during initial filtering of the data due to differences in how the estimated creatinine clearance was calculated in the EMR versus in Microsoft Excel. The EMR rounds the serum creatinine value up to 0.8 mg/dL in patients aged 65 and older. It also adapts which body weight (actual, ideal, or adjusted) is used in the Cockcroft-Gault equation based on set parameters (adjusted weight is used for patients whose body weight is 20% greater than the ideal body weight); the Microsoft Excel calculation used only actual body weight and did not round serum creatinine. In addition, hemodialysis patients could not be distinguished from others based on the pulled patient values; hemodialysis patients with calculated creatinine clearance values above dose adjustment cutoffs may have been incorrectly counted as appropriately dosed. These discrepancies would only have affected the total number of orders eligible for renal dose adjustment, as all orders included in the random sampling had estimated creatinine clearance values confirmed through comparison with the EMR calculated creatinine clearance during chart review.

During data collection, eligible orders were filtered to remove infectious disease and nephrology providers as outlined in the methods. Furthermore, providers who may have been working for these specialty services, such as nurse practitioners or medical residents, were not excluded because their assigned specialty was listed as another service area. This means orders that would not have been eligible for dose adjustment per policy may have been included in the data pool; however, the random sampling of orders did not reveal any such cases. Despite these limitations, the evaluation of the renal dose adjustment policy provided insight into opportunities for further pharmacist and provider education, as well as policy improvement opportunities.

One of the medications with the lowest percent compliance was cefazolin at 30%. Of the 10 orders randomly sampled, only 3 were appropriately dose adjusted for renal function. Six of the 7 orders not adjusted were ordered from a postoperative order set, in which cefazolin is prescribed every 8 hours for 2 doses to start 8 hours after administration of the preoperative antibiotic dose. For a creatinine clearance below 35 mL/min, the renal dose adjustment policy calls for only 1 postoperative dose given 12 hours after administration of the preoperative dose. Whether it is due to an increased comfort level with orders originating from order sets or pharmacists just not recognizing the need for adjustment, this is an opportunity for pharmacist education. A process improvement opportunity exists to increase visibility within the cefazolin orders in these order sets. Informing pharmacists within the verification step to check creatinine clearance for adjustment and auto populating the most recent serum creatinine in the order have the potential to increase compliance with this dose adjustment.

Famotidine, the most frequently ordered medication, had a percent compliance of 78%. This medication also often originates from order sets, but there did not seem to be an association with order set use and inappropriate doses. The creatinine clearance cutoff per policy for famotidine dose adjustment is 50 mL/min. A possible reason for the lower compliance rate is that since this is different from many medications with cutoffs of 30 or 60 mL/min, pharmacists are not recognizing the need for adjustment. This is also a pharmacist education opportunity, as the risk associated with famotidine doses that are too high has the potential to cause harm. Dose-related confusion, delirium, hallucinations, agitation, seizures, and lethargy have been reported in literature. Accurate and timely appropriate dose adjustments may prevent these from occurring.

For orders evaluated in the random sample, magnesium hydroxide had a 40% compliance with policy outlined dose adjustments. This medication is contraindicated at creatinine clearance values of less than 30 mL/min, and pharmacists should discontinue the order per policy. The 40% compliance rate from the random sample includes orders for patients with borderline creatinine clearance or other clinical factors (such as a single dose for symptom relief in a hospice patient) that may have been considered appropriate at the time of verification. Many of the magnesium hydroxide orders also stem from order sets; however, the orders already contain language guiding the pharmacists to evaluate the renal function for appropriateness.

Tramadol, at a 30% policy compliance rate, appears to have a poorer rate than it may actually have. The renal dose adjustment policy for tramadol specifies a dose interval adjustment, as well as a daily dose limit. Seven of the 10 orders sampled were inappropriate doses based on the policy dose adjustments; however, only 1 of these orders would actually have been over the specific daily limit. Reenforcing the importance of interval adjustments may decrease the number of orders where the dose was inappropriately adjusted in lieu of the interval.

Other opportunities for improvement of the renal dose adjustment policy were investigated during this study. One of the notable discrepancies within the policy is that cefuroxime intravenous dosing is included for automatic adjustment; however, cefuroxime oral dosing is not. Inclusion of oral dosing within the policy could potentially prevent unnecessary and time-consuming conversations between pharmacists and providers.

An additional process measure that may increase compliance is to evaluate the dose adjustments in the policy and those recommended by drug manufacturers or other sources. Ketorolac is a medication where the policy differs from other sources; the manufacturer and policy recommend decreases to 15 mg every 6 hours. In this study, a large number of 7.5 mg doses were ordered, and while this could be considered an unnecessary dose reduction, there is evidence to support this decreased dose in certain patients. ¹⁵ Updating the policy to reflect current practice may lead to higher compliance rates for select medications.

When examining transitions of care, almost all orders were appropriately dosed upon discharge from the hospital and no trends were seen in inappropriate doses. Notable exceptions included an inappropriately high dose of ciprofloxacin for the patient’s renal function, and an inappropriately low dose of oseltamivir that was continued despite improvement in the patient’s renal function prior to discharge.

Overall, there are many potential opportunities to improve the utilization of the adult renal dose adjustment policy including pharmacist education and policy improvements. This review established the baseline percent compliance for utilization of the policy. Improvements that were found will encourage increased compliance for those medications which were lagging. To accomplish the secondary purpose of this study, upon completion, the evaluation process was examined for effectiveness, ease, timeframe, and adjustability for other policy and service evaluations. From this, a standard system was developed that incorporates basic steps of the evaluation process as well as recommendations for how frequently different policies/services should be assessed, what time frames may be good for assessment, and what resources can be used throughout the process. Table 5 illustrates a possible format for documenting various evaluations performed including this policy examination as an example. The basic evaluation steps illustrated through this study are shown in Figure 2; these will be used for future evaluations of other policies. Resources developed through initial evaluations, such as the data extraction tool built in this study, will be saved along with the evaluation documentation to be used as references for these future evaluations.

Table 5.

Example Pharmacy Policy Evaluation Record.

Service/policy	Period for evaluation	Evaluation frequency	First completed	Last completed	Last time period evaluated
Renal dose adjustment	6 months	Every 2 years	June 30, 2019	June 30, 2019	January 1, 2018 to June 30, 2018
Warfarin dosing	3 months	Every 2 years
Vancomycin dosing	3 months	Every 2 years
Aminoglycoside dosing	1 year	Every 3 years
Intravenous to Oral	6 months	Every 3 years
Formulary Interchange	6 months	Every 3 years

Figure 2.

Algorithm for evaluating pharmacy policies.

Conclusions

When assessing the role of clinical pharmacy, it can be very hard to determine quality, as patient outcomes are influenced by numerous members of interdisciplinary teams. Determining the impact of pharmacy services alone is difficult to separate out from the influence of other disciplines. Evaluating clinical pharmacy services to assess compliance can help to determine if policies are being optimized. When policies are enhanced to allow pharmacists to practice to the best of their abilities, improvements can be seen in patient outcomes, cost savings, and in improved interdisciplinary teamwork. The evaluation of the renal dose adjustment policy not only illustrated possible policy improvements and education opportunities but also assisted in planning for future evaluation of pharmacist clinical services.