Abstract
(1) Background: This follow-up retrospective analysis used electronic medical record (EMR) data from a health system to identify patients and medications prescribed in accordance with Clinical Pharmacogenetics Implementation Consortium (CPIC) guidelines. (2) Methods: This analysis included EMR data from a clinical research data warehouse encompassing 928,291 patients seen at an academic medical center between 2020 and 2024. The study evaluated 75 commercially available medications linked to 52 evidence-based CPIC pharmacogenomic (PGx) guidelines. (3) Results: Of the 928,291 patient encounters, 709,673 medication orders were recorded, with 416,621 patients (44.8%) prescribed at least 1 of the 75 CPIC-associated medications. This compares with 845,518 patients who had an encounter in 2015–2019 with 590,526 medication orders, and 335,849 (56.9%) patients had medication orders represented by CPIC-associated medications. One to three CPIC-associated medications accounted for 76.6% of patients in 2020–2024 compared to 75.6% in 2015–2019. (4) Conclusions: The findings demonstrate that the proportion of patients prescribed a CPIC-actionable medication remained just under half of those evaluated within a single institution’s EMR. About three-quarters of patients over the ten-year period had between one to three CPIC-associated medications identified, and the top five classes of medications remained the same in the two periods. This understanding of patient volume may help organizations as they begin to assess the implementation of PGx services.
1. Introduction
Pharmacogenetics (PGx) examines how genetic variations among individuals influence their responses to medications. The goal of PGx is to enable clinicians to choose the most effective therapies and dosing strategies while minimizing the risk of adverse drug effects [1,2,3,4]. As a significant building block to precision medicine, this includes pharmacogenomics.
Pharmacogenomic (PGx) factors can influence over half of the of prescriptions in the U.S., and nearly everyone carries at least one high-risk genetic variant that can affect drug response [3,5,6,7]. PGx testing is often considered for reasons such as toxicity, side effects, lack of efficacy, family history, or pre-emptive treatment [8]. The U.S. Food and Drug Administration (FDA) includes pharmacogenomic labeling for more than 200 medications to support safer prescribing, with some labels incorporating “Black Box” warnings for potentially life-threatening risks [9,10,11]. Examples of medications with actionable PGx information include warfarin and codeine, where labels indicate whether genetic testing is recommended before use [12].
The international group of scientific and clinical experts that compose the Clinical Pharmacogenetics Implementation Consortium (CPIC) offers evidence-based guidelines that help clinicians interpret pharmacogenomic test results and apply them to clinically actionable prescribing decisions for well-characterized gene–drug pairs. When the CPIC recommends pharmacogenomic testing, this is after an evidence review by a panel of scientific experts that shows there is substantial published scientific evidence indicating testing should be conducted to improve treatment outcomes and/or reduce harm to patients [5]. These guidelines focus on the most common and clinically actionable gene–drug interactions and are intended for implementation across diverse clinical practice settings. Patient-specific diplotypes are interpreted to generate corresponding clinical phenotypes and drug dosing recommendations [5]. The CPIC categorizes evidence into four levels (A–D), with level “A” representing strong or moderate evidence supporting the use of genetic information to guide prescribing decisions. In contrast, level “D” denotes limited published evidence, unclear clinical actionability, minimal mechanistic rationale, predominantly weak data, or substantial conflicting findings. As of 2025, the CPIC had published 29 pharmacogenomic guidelines encompassing 35 genes and 168 medications across diverse therapeutic categories [5].
Pharmacogenetic (PGx) testing can be performed either reactively or pre-emptively. In the U.S., most PGx tests are reactive, ordered by a single prescriber after a patient experiences an adverse drug effect or fails to respond to therapy. These tests often focus on a single gene–drug pair, limiting future usefulness for other prescribers. In contrast, pre-emptive PGx testing is conducted before a medication is selected to help prevent adverse effects and improve treatment efficacy [6,13,14]. Because a patient’s genotype remains stable over time, pre-emptive results can guide future prescribing as health conditions and medication needs evolve, as well as when prescribers change over the lifetime of a patient. As costs decline and more guidelines develop, pre-emptive PGx testing is a more cost-effective strategy, especially when performed early in life, providing valuable information throughout a patient’s lifetime [3,6,13].
Building on our previous retrospective analysis of EMR data from 2015 to 2019 at a tertiary integrated health system, we examined records from 2020 to 2024 to identify patterns in medication prescribing based on CPIC guidelines. During the 2015–2019 period, 845,518 patients had at least one clinical encounter, resulting in 590,526 medication orders [15]. Of these patients, 335,849 (56.9%) received one or more medications included in the 52 CPIC guidelines. These prescriptions were issued by 2803 providers and encompassed 42 commercially drugs, highlighting the scope of PGx-relevant prescribing within the health system [15].
Identifying the scope of patients who may benefit from pharmacogenomic testing can support the development of proactive testing strategies, targeted physician resources, and pharmacist-led workflows to improve clinical outcomes if a PGx program is adopted. The release of updated CPIC guidelines, along with new medications and gene recommendations since 2019, motivated this follow-up retrospective analysis using EMR data from the same tertiary integrated health system. The primary aim was to assess if and how prescribing patterns have shifted in response to the updated CPIC guidelines. Conducting two internal analyses over a ten-year span provides valuable insight for medical center decision makers in evaluating the potential scope and necessity of implementing PGx services.
2. Materials and Methods
Deidentified electronic medical record (EMR) data from the academic medical center were consolidated within the Clinical Data Warehouse. The academic medical center is an adult Level I Trauma Center, located in Milwaukee, Wisconsin. At the time of data extraction there were more than 2000 physicians with over 40,000 inpatient admissions and approximately 1.5 million outpatient visits across 45 health centers and clinics [16].
We evaluated prescribing records and demographic characteristics for 928,291 patients with at least one encounter in the five-year period from 2020 to 2024. Patient records accessed included both inpatient and outpatient encounters. The drug formulary was cross-referenced with RxNorm at the ingredient level to extract all medication orders corresponding to 75 drugs meeting CPIC level “A” or PharmGKB level “A1” evidence criteria.
We characterized “new” patients to evaluate the frequency and complexity of PGx drug ordering among individuals presumed naïve to PGx testing at the start of the extraction period. A “new” patient was defined as having no prior PGx drug history and at least one encounter occurring one year before their first PGx drug order.
As performed previously, the frequency and duration of drug orders were analyzed using an “ordering episode” heuristic [15]. An episode was defined as all orders for the same drug occurring within a year and a half of each other. This approach grouped recurring annual or more frequent orders into a single episode. Because computational date calculations are precise but real-world scheduling can vary, we applied an additional 0.5-year buffer to account for this variability. For patients receiving multiple PGx medications, we determined an “ordering interval,” defined as the number of days between the first prescription of the initial PGx drug and the earliest prescription of the final PGx drug during the study period. This was a consistent approach from the 2015–2019 study [15].
3. Results
Demographic Comparison of Patients with Clinical Encounters, Medication Prescriptions, and CPIC-Listed Drug Orders, 2020–2024: Table 1 provides a summary of the demographic characteristics of 928,291 patient charts analyzed from the EMR over the five-year period from 2020 to 2024. Females were more represented (54.5%) than males (45.5%). Based on these 928,291 encounters, there were 709,673 medication orders processed. A total of 416,621 (44.8%) patients had medication orders represented by the 75 CPIC drugs.
Table 1.
Demographics of patients with clinical encounters, medication orders, and CPIC-listed drug orders for 2020–2024.
3.1. Comparison of CPIC Drug Ordering Characteristics 2020–2024
As shown in Table 2, CPIC drug ordering characteristics are organized according to CPIC guidelines, with counts based on unique patients. Ondansetron was the most frequently prescribed medication (255,153 orders in 61.0% of all CPIC patients) with a 26.9% change from 2015 to 2019 [15]. This was followed by non-steroidal anti-inflammatory drugs (NSAIDs) (ibuprofen, meloxicam, celecoxib, piroxicam, and flurbiprofen), statins (rosuvastatin, pitavastatin, fluvastatin, simvastatin, lovastatin, atorvastatin, and pravastatin), opioids (hydrocodone, tramadol, and codeine), proton pump inhibitors (PPIs) (pantoprazole, dexlansoprazole, lansoprazole, and omeprazole), and selective serotonin reuptake inhibitors (SSRIs) (escitalopram, paroxetine, venlafaxine, fluvoxamine, sertraline, vortioxetine, and citalopram). Of note, some of these medications, including ondansetron or NSAIDs, may have been prescribed within an order set or with directions to be taken as needed, which does not guarantee that the patient used the medication. However, with the medication being ordered, this allows for the use of the medication and therefore shows the potential benefit of PGx testing.
Table 2.
CPIC drug ordering characteristics for 2015–2019 [15] compared to 2020–2024, grouped by CPIC guidelines and patients.
Table 2 shows the percent change between medication prescribed as per CPIC guidelines between the 2015–2019 and 2020–2024 groups. This highlights the prescribing trends during this time period. The medication with the greatest percentage increase in prescribing was atomoxetine, prescribed only 809 times during the earlier timeframe compared to 1854 times during the later (129.2% increase).
3.2. Combinations of CPIC Drugs Ordered 2020–2024
The demographic characteristics of the 2020–2024 cohort in Table 3 that used a CPIC drug are described, with the majority being female (57.1%) and White (74.5%), followed by other races of Black (17.1%) and Asian (2.3%). The most common ethnicity identified was non-Hispanic (93.6%).
Table 3.
Comparison of combinations of CPIC drugs ordered in 2015–2019 [15] vs. 2020–2024.
Of the 75 total drugs with CPIC guidelines assessed in the data set, only 63 had orders identified in the EHR data. Overall, of these 63 medications ordered, 57 were ordered singly at least once (the number in the first row of Table 3), and 6 were only ever ordered in combination, hence the apparent discrepancy between 63 drugs but only 57 distinct single drug orders in Table 3.
As shown in Table 3, from 2020 to 2024, 156,578 patients (37.4%) were prescribed at least one CPIC medication, encompassing 57 distinct drug combinations. Among these, 100,396 patients (24.0%) were prescribed two CPIC medications, and 63,363 patients (15.2%) were prescribed three CPIC medications. In comparison, during 2015 to 2019, 133,380 patients (39.7%) were prescribed at least one CPIC medication with the same 57 drug combinations; 84,334 patients (25.1%) were prescribed two CPIC medications, and 51,357 patients (15.3%) were prescribed three CPIC medications.
4. Discussion
The 928,291 patients identified between the 2020–2024 extraction period was more than what was observed for the 2015–2019 timeframe (845,518). The results from this study show that 416,621 (44.8%) were prescribed a CPIC-actionable medication within the academic medical center between the years 2020 and 2024.
A total of 335,849 patient records were reviewed from 2015 to 2019. Among these, 133,380 patients (39.7%) were prescribed at least one CPIC medication, representing 47 distinct drug combinations. Approximately one-quarter of patients (84,334, 25.1%) received two CPIC medications, and 51,357 (15.3%) received three, accounting for 80.1% of patients prescribed between one and three CPIC medications during this period [15]. In comparison, the 2020–2024 analysis identified 418,144 unique patients, of whom 156,578 (37.4%) were prescribed at least one CPIC medication, covering 57 distinct drug combinations. Additionally, 100,396 patients (24.0%) were prescribed two CPIC medications, and 63,363 patients (15.2%) were prescribed three CPIC medications. Overall, three-quarters (76.6%) of patients were prescribed between one and three CPIC medications during this period, 2020–2024. Patients who were prescribed three or less CPIC drugs had the same pattern in both collection periods: 80.1% (2015–2019) and 76.6% (2020–2024) [15]. Given that a significant number of patients were prescribed three or less CPIC drugs, it would help a healthcare system project potential volume of patients if PGx services were to be offered.
The number of medications recommended for PGx testing and monitoring continues to increase due, in part, to the continued growth of research evidence [5]. This is, in turn, reflected in the growing list of drug–gene pairs added by the scientific community to CPIC guidelines, which did increase over the specified time periods observed in these two studies. This growth reflects advancements in PGx research, improvements in knowledge of drug–gene interactions, and broader adoption of precision medicine approaches across different specialties and health systems [17]. As prescribing patterns change and more medications are added to CPIC guidelines, it is crucial that pharmacists and providers are equipped to manage this added complexity.
Prescribing patterns of certain medications were different between the different timeframes, with increasing and decreasing frequencies. While this study did not identify the discipline of the prescribers, it would be reasonable to assume that the 19 medical specialties identified previously, with the highest frequencies observed from the departments of anesthesiology, oncology, internal medicine, and family practice, would be similar, as the medical center and EHR remained the same [15]. This is important if an organization determines it wants to offer PGx services and in what departments or disciplines it is most likely to prescribe medications identified by the CPIC.
Paradoxically, while the number of CPIC-guided medications has increased, this is not always reflective in clinical practice and thus the data. Some older medications that have CPIC guidelines may be prescribed less due to advances in other therapeutic categories and recommended clinical guidelines. Warfarin is an example of this with the gene VCORC1, and other anticoagulant agents can be substituted that do not fall under CPIC guidelines.
The clinical significance of PGx testing lies in its ability to use evidence-based guidelines, such as those from the CPIC, to translate patient-specific genotypes into actionable phenotypes. This process enables clinicians to optimize medication selection, avoid harmful drug–gene interactions, and adjust dosing appropriately [2,7]. PGx-guided prescribing can enhance patient safety, improve therapeutic outcomes, and, in some cases, reduce overall healthcare costs. Integrating PGx data into EMRs is critical for enabling timely decision support for clinicians and ensuring seamless implementation in routine practice [13].
With multiple prescribers encountered in the routine care of patients, having a single discipline overseeing PGx services makes sense from a resource and workflow standpoint, as well as based on economic perspectives for health systems. Pharmacists are well positioned to lead PGx services through precision medicine efforts as well as contribute to individual’s health outcomes, given their unique pharmacotherapy knowledge related to a wide array of medications.
This study has limitations. It included patients and prescribers from a single academic medical center, so regional and national prescribing trends were not assessed. The analysis did not determine whether any PGx tests had been ordered, as this was beyond the scope of the project. Patients included in this analysis may have been assessed in the 2015–2019 cohort; however, each encounter was different so it is important to continue to emphasize the utility of genetic testing throughout a patient’s lifetime [15].
Additionally, all comparisons in Table 2 relied on raw prescribing counts for distinct patients and did not adjust for changes in total prescriptions or patient volume between the periods 2015–219 and 2020–2024, limiting the interpretation between the two time periods. However, the top five classes of medications remained the same in the two periods (ondansetron, statins, proton-pump inhibitors, codeine, and NSAIDs). Such information is useful when assessing the implementation of a PGx service.
Another limitation is that some drugs are prescribed less, in part due to the lack of knowledge associated with pharmacogenomics and education provided to prescribers on the risk of certain classes of medications. For example, in the codeine-related products, a reduction of 21.3% was observed in codeine, tramadol, and hydrocodone at this institution. This reduction in these medications may, in part, be explained by external and internal guidelines on prescribing analgesics for acute pain management.
Also, it is recognized that between 2020 and 2024, the COVID-19 pandemic led to a change in healthcare utilization. Clinicians needed to manage both acute infections and complications in patients with many comorbidities, leading to increased prescription and over-the-counter medications used. The use of medications with pharmacogenomic considerations became increasingly more prominent as CPIC-guided drugs—such as analgesics (opioid and non-opioid) and antidepressants—were frequently used to treat underlying conditions or symptoms that coexisted with COVID-19 [18,19]. The analysis of the impact of the pandemic was beyond the scope of this study.
5. Conclusions
Given the consistency observed where about three-quarters of patients over the ten-year period had between one and three CPIC drugs prescribed, as well as the common medication classes identified, leveraging EMR data provides an effective means to quantify patients and the medication involved who may be eligible for PGx testing. This analysis and the findings may help healthcare organizations as they begin to assess the implementation of PGx services as more precision medicine platforms are adopted.
Author Contributions
Conceptualization, G.E.M.III and U.B.; methodology, G.E.M.III and U.B.; software, G.E.M.III; validation, G.E.M.III and U.B.; formal analysis, G.E.M.III and M.M.; investigation, G.E.M.III and U.B.; resources, G.E.M.III; data curation, G.E.M.III, M.M., and U.B.; writing, M.M., G.E.M.III, and U.B., original draft preparation, M.M., G.E.M.III, and U.B.; writing—review and editing, M.M., G.E.M.III, and U.B.; visualization, M.M. and G.E.M.III; supervision, G.E.M.III; project administration, G.E.M.III and U.B.; funding acquisition, G.E.M.III. All authors have read and agreed to the published version of the manuscript.
Funding
This research received no external funding.
Institutional Review Board Statement
The data were extracted from the fully deidentified version of the Clinical Research Data Warehouse and were therefore not subject to an IRB review or specific protocol by the Institutional Review Boards at the Medical College of Wisconsin, Milwaukee, USA.
Informed Consent Statement
Not applicable.
Data Availability Statement
The data presented in this study are available on request from the corresponding author due to privacy or ethical restrictions.
Acknowledgments
The Froedtert & the Medical College of Wisconsin (F&MCW) regional health network is a partnership between Froedtert Health and the Medical College of Wisconsin (MCW), operating eastern Wisconsin’s only academic medical center and adult Level I Trauma Center at Froedtert Hospital, Milwaukee. At the time of this work, the F&MCW health network included 10 hospital locations, more than 2000 physicians, and more than 45 health centers and clinics. Outpatient visits were nearly 1.5 million, inpatient admissions were 40,056, and visits to network physicians exceeded 1.1 million [16]. Alexander Stoddard, formerly of MCW’s Clinical & Translational Science Institute, provided the data extractions and analyses. During the preparation of this manuscript/study, the authors used M365 Copilot for the purposes of language clarification. The authors have reviewed and edited the output and take full responsibility for the content of this publication.
Conflicts of Interest
UB is the Founder, CEO, and holds equity in RPRD Diagnostics LLC. The remaining authors declare no conflicts of interest.
Abbreviations
The following abbreviations are used in this manuscript:
| EMR | Electronic Medical Record |
| CPIC | Clinical Pharmacogenetics Implementation Consortium |
| PGx | Pharmacogenomics |
References
- Wang, L.; McLeod, H.L.; Weinshilboum, R.M. Genomics and drug response. N. Engl. J. Med. 2011, 364, 1144–1153. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Hawke, R.L. Chapter 1 Developing Perspectives on Pharmacogenomics. In Pharmacogenomics: An Introduction and Clinical Perspective; Bertino, J.S., Jr., DeVane, C., Fuhr, U., Kashuba, A.D., Ma, J.D., Eds.; McGraw-Hill: New York, NY, USA, 2013. [Google Scholar]
- Saulsberry, L.; Jameson, J.C.; Gibbons, R.D.; Dolan, M.E.; Olopade, O.I.; O’Donnell, P.H. A National Study Among Diverse US Populations of Exposure to Prescription Medications with Evidence-Based Pharmacogenomic Information. Clin. Pharmacol. Ther. 2025, 117, 1793–1802. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Lefaivre, A.; Litinski, V.; Vandenhurk, M. Pharmacogenetic Testing May Improve Drug Treatments and Shorten Disability Leaves. Benefits Q 2017, 33, 43–49. [Google Scholar] [PubMed]
- Clinical Pharmacogenetics Implementation Consortium (CPIC). [Database]. Available online: https://cpicpgx.org/guidelines/ (accessed on 10 December 2025).
- Dunnenberger, H.M.; Crews, K.R.; Hoffman, J.M.; Caudle, K.E.; Broeckel, U.; Howard, S.C.; Hunkler, R.J.; Klein, T.E.; Evans, W.E.; Relling, M.V. Preemptive clinical pharmacogenetics implementation: Current programs in US medical centers. Annu. Rev. Pharmacol. Toxicol. 2015, 55, 89–106. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Ji, Y.; Skierka, J.M.; Blommel, J.H.; Moore, B.E.; VanCuyk, D.L.; Bruflat, J.K.; Peterson, L.M.; Veldhuizen, T.L.; Fadra, N.; Peterson, S.E.; et al. Preemptive Pharmacogenomic Testing for Precision Medicine. J. Mol. Diagn. 2016, 18, 438–446. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Manzi, S. PGx Experience at Boston Children’s Hospital. [Webinar]; Boston Children’s Hospital: Boston, MA, USA, 2019; Available online: https://www.rprdx.com/webinar-pgx_bch/ (accessed on 10 December 2025).
- Pharmacogenomics Testing: Harnessing the Power of Genetic Data for Personalized Medicine. Translational Software: [Whitepaper]; 2019; pp. 3–7. Available online: https://pharmaceutical.report/whitepapers/pharmacogenomics-testing-harnessing-the-power-of-genetic-data-for-personalized-medicine (accessed on 8 November 2019).
- National Institute of General Medical Sciences. Pharmacogenomics; U.S. Department of Health and Human Services: Washington, DC, USA, 2025. Available online: https://www.nih.gov/about-nih/nih-turning-discovery-into-health/promise-precision-medicine/pharmacogenomics (accessed on 10 December 2025).
- What is ClinPGx? [Database]. Available online: https://www.clinpgx.org/whatIsPharmgkb/prescribing (accessed on 5 December 2025).
- Center for Drug Evaluation and Research. Table of Pharmacogenomic Biomarkers; U.S. Food and Drug Administration: Silver Spring, MD, USA, 2019. Available online: https://www.fda.gov/drugs/science-and-research-drugs/table-pharmacogenomic-biomarkers-drug-labeling (accessed on 20 November 2025).
- Keeling, N.J.; Rosenthal, M.M.; West-Strum, D.; Patel, A.S.; Haidar, C.E.; Hoffman, J.M. Preemptive pharmacogenetic testing: Exploring the knowledge and perspectives of US payers. Genet. Med. 2019, 21, 1224. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Verbeurgt, P.; Mamiya, T.; Oesterheld, J. How common are drug and gene interactions? Prevalence in a sample of 1143 patients with CYP2C9, CYP2C19 and CYP2D6 genotyping. Pharmacogenomics 2014, 15, 655–665. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- MacKinnon, G.E., III; Mills, M.; Stoddard, A.; Urrutia, R.A.; Broeckel, U. An EMR-Based Approach to Determine Frequency, Prescribing Pattern, and Characteristics of Patients Receiving Drugs with Pharmacogenomic Guidelines. Pharmacy 2023, 11, 178. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- About Our System. Froedtert & the Medical College of Wisconsin. Available online: https://www.froedtert.com/about (accessed on 20 November 2025).
- Ingelman-Sundberg, M.; Nebert, D.W.; Lauschke, V.M. Emerging trends in pharmacogenomics: From common variant associations toward comprehensive genomic profiling. Hum. Genom. 2023, 17, 105. [Google Scholar] [CrossRef] [Scilit] [PubMed] [PubMed Central]
- Martinez-Cengotitabengoa, M.; Sanchez-Martinez, M.; Sanchez-Martinez, A.; Long-Martinez, D.; Dunford, D.; Revuelta, P.; Echevarria, E.; Calvo, B. New Antidepressant Prescriptions Before, During, and After the COVID-19 Pandemic: Sex and Age Differences in a Population-Based Ecological Study. Healthcare 2025, 13, 502. [Google Scholar] [CrossRef] [Scilit] [PubMed] [PubMed Central]
- Rikin, S.; Perez, H.R.; Zhang, C.; Khalid, L.; Groeger, J.; Deng, Y.; Starrels, J.L. Changes in Outpatient Opioid Prescribing During the COVID-19 Pandemic: An Interrupted Time Series Analysis. J. Prim. Care Community Health 2022, 13, 21501319221076926. [Google Scholar] [CrossRef] [Scilit] [PubMed] [PubMed Central]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2026 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license.