Trial of Preemptive Pharmacogenetic Testing in Underserved Patients (TOPP UP): Study Framework and Baseline Characteristics of Randomized Controlled Trial Participants

ABSTRACT

Research in pharmacogenetics has advanced over the years, but its broad clinical implementation remains limited. While medically underserved patients may particularly benefit from preemptive pharmacogenetic testing, its clinical impact in this population has not been well explored. The aim of the Trial of Preemptive Pharmacogenetic Testing in Underserved Patients is to assess the feasibility of clinically implementing preemptive pharmacogenetic testing and determine its effect on patient medication satisfaction in medically underserved patients. The Trial of Preemptive Pharmacogenetic Testing in Underserved Patients is a randomized, controlled, pragmatic clinical trial conducted across four clinics in North Central and Northeast Florida. Primary care patients from clinics predominantly caring for medically underserved populations are randomized to receive either pharmacogenetic panel testing to support standard of care dose optimization or standard of care without pharmacogenetic testing. The primary effectiveness outcome will be the patient‐reported Treatment Satisfaction Questionnaire for Medication score. The primary implementation outcome will be prescriber acceptance rates of pharmacogenetic clinical decision support. Patient‐reported outcomes are collected through surveys at baseline, 6 and 12 months after enrollment. This manuscript presents an overview of the study design and provides the enrolled participants' baseline characteristics, which demonstrate recruitment of a diverse, medically underserved population from the primary care setting.

Study Highlights

• What is the current knowledge on the topic?

  • ○
    There are only a few clinical trials to date that analyze the feasibility of PGx testing implementation, and even fewer in patients who are medically underserved.
• What question did this study address?

  • ○
    This study aims to examine the effect of preemptive pharmacogenetic testing in medically underserved populations, who have often been underrepresented in prior pharmacogenetic research. Specifically, it aims to evaluate the feasibility of implementing pharmacogenetic testing in these populations and to assess patient‐reported outcomes such as treatment satisfaction, while considering socioeconomic factors influencing these measures. The goal is to support expanded access to pharmacogenetic testing for underserved groups.
• What does this study add to our knowledge?

  • ○
    The study successfully recruited a medically underserved patient population for a preemptive PGx testing clinical trial, as demonstrated by participants‘ baseline demographics and social determinants of health, thereby providing a strong framework for future feasibility and patient satisfaction analyses.
• How might this change clinical pharmacology or translational science?

  • ○
    This study advances clinical translational science by demonstrating successful recruitment of a medically underserved population for preemptive PGx testing, providing proof‐of‐concept that such trials can be conducted in real‐world, diverse patient groups and offering a framework for broader implementation.

1. Introduction

Genetic polymorphisms can play a crucial role in patient responses to medications, influencing both therapeutic effectiveness and adverse events [1, 2, 3, 4, 5]. Research over the past several decades has identified numerous genetic mutations that contribute to this variability [1, 2, 3, 4, 5]. Several studies have shown that pharmacogenetic (PGx) testing can improve medication effectiveness and reduce the risk of medication‐related adverse effects [6, 7]. However, despite this, the clinical use of PGx testing outside the field of cancer remains limited [8, 9, 10].

Most experts in the field of PGx agree that clinical implementation will eventually evolve to a preemptive testing model, in which patients are genotyped for multiple PGx variants at one time, and results are integrated into their electronic health record (EHR) for future use [9, 11]. Preemptive testing offers potential for greater clinical utility as a patient's genotype is known before a medication is prescribed, enabling more informed medication selection than reactive testing performed after a prescribing decision is indicated. In addition, since preemptive testing is not time‐sensitive, batch processing of samples can decrease genotyping cost on a per‐patient basis [12].

Patient populations who particularly benefit from preemptive PGx testing may include those who are medically underserved, such as patients with low income, racial minorities, and patients living in rural areas [13, 14]. These patients often have fewer interactions with healthcare providers and thus may have more difficulty obtaining optimized pharmacotherapy through the traditional trial‐and‐error prescribing approach [15]. Our previous research suggests that patients with lower access to healthcare, which includes medically underserved patients, are prescribed drugs with PGx guidelines available at an increased rate compared to other patients [15]. However, little is known about implementing preemptive PGx testing in these settings. Available insights primarily stem from pilot studies, stakeholder interviews, and case studies, highlighting implementation barriers such as cost, EHR reporting standardization, and the need for efficient patient and provider education, with limited information on actual clinical implementation in this population [16, 17, 18]. A previous pilot study we conducted suggested that the implementation of preemptive PGx testing in the primary care setting may be feasible but was limited by a small sample size and the lack of a comparator group [19].

The Trial of Preemptive Pharmacogenetic Testing in Underserved Patients (TOPP UP) aims to assess the feasibility of implementing preemptive PGx testing in medically underserved patients as well as its effect on patient medication satisfaction. Herein, we present the study design, recruitment process, and baseline characteristics of the trial participants.

2. Materials and Methods

2.1. Study Design

TOPP UP (NCT05141019; https://clinicaltrials.gov/study/NCT05141019; date of registration: November 19, 2021) is a prospective, randomized, open‐label, pragmatic, hybrid clinical trial evaluating the effect of preemptive PGx testing on patient medication satisfaction. The trial is also assessing the feasibility of clinically implementing preemptive PGx testing in medically underserved populations. The study is being conducted across four University of Florida Health System (UF Health) primary care clinics located in Gainesville, Old Town, and Jacksonville, Florida, United States. These clinics primarily serve medically underserved patients and are geographically diverse: two urban clinics in Jacksonville, one suburban clinic in Gainesville, and one rural clinic in Old Town. The study was approved by the University of Florida Institutional Review Board (IRB #202101316), and all participants are required to provide informed written consent prior to participating. The study was conducted in accordance with the Consolidated Standards of Reporting Trials (CONSORT) guidelines (Table S1) [20].

2.2. Participants

Participants eligible for this study were established patients in the aforementioned clinics aged 18 years or older who met the following criteria: (1) had at least three active prescription medications documented in their medical record, (2) had a diagnosed condition treatable with a drug subject to PGx guidelines (Tables 1 and 3) had a documented history of a recent medication change (within the prior 8 months). Individuals with a documented history of any of the following: prior clinical PGx testing, chronic dialysis, allogeneic stem cell transplant, liver transplant, a life expectancy of less than 12 months, or an inability to consent due to an underlying physical or psychiatric condition were excluded from the trial.

TABLE 1.

Diagnoses required for TOPP UP inclusion and their associated PGx medications.

Diagnosis	Associated drug or drug class
Generalized depression	SSRI, SNRI, or TCA
Anxiety	SSRI, SNRI, or TCA
Inflammatory bowel disease	Azathioprine or mercaptopurine
Gastroesophageal reflux disease	PPI
Erosive esophagitis	PPI
Gastric ulcer	PPI
Acute coronary syndrome	Antiplatelet or beta‐blocker or statin
Chronic pain	Opioid or NSAID
Fibromyalgia	Opioid or NSAID
Rheumatoid arthritis	NSAID
Osteoarthritis	Opioid or NSAID
Dyslipidemia	Statin
Heart failure	Beta‐blocker
Deep vein thrombosis	Anticoagulant
Atrial fibrillation	Anticoagulant
Stroke	Antiplatelet

Abbreviations: NSAID, nonsteroidal anti‐inflammatory drug; PPI, proton pump inhibitor; SNRI, selective serotonin norepinephrine reuptake inhibitor; SSRI, selective serotonin reuptake inhibitor; TCA, Tricyclic Antidepressant.

TABLE 3.

Comparison of Descriptors of Medical Underservice Between Enrolled Study Participants and Florida and the United States.

	Sample	Statewide (Florida)	National (US)
Highest Educational Level a
Bachelor's degree or higher	89 (17.0%)	5,947,429 (32.6%)	87,909,516 (33.5%)
Health Insurance b
Medicaid coverage alone or in combination	180 (29.1%)	3,954,513 (17.8%)	70,327,938 (21.3%)
None	12 (1.9%)	2,381,397 (10.7%)	26,169,652 (7.9%)
Poverty Level c
At or Below Poverty Level	174 (39.6%)	2,729,519 (12.3%)	40,763,043 (12.5%)
Geocoded Indices d
Social Deprivation Index (SDI) Percentile	74.0 (52.0–84.3)	52.0 (29.0–75.0)	50.0 (25.0–75.0)
Social Vulnerability Metric (SVM) Percentile	77.9 (43.1–92.9)	50.1 (26.6–70.6)	40.8 (16.5–66.5)
Community Assets and Relative Rurality (CARR) Index	0.30 (0.28–0.32)	0.27 (0.25–0.29)	0.28 (0.25–0.31)

^a Highest Education Level: Sample is restricted to adults 18 years of age and older, consistent with US Census Bureau 2023 methodology, to ensure demographic comparability with the recruited population [21].

^b Insurance: Statewide and national estimates include the civilian noninstitutionalized population, as defined by the 2023 American Community Survey (ACS) [22, 23].

^c Poverty Level: Percentages are based on individuals who responded to the survey question covering income level. Statewide and national data were obtained from Poverty in States and Metropolitan Areas: 2023 [24].

^d Geocoded Indices: Weighted median values for the Social Vulnerability Metric (SVM), Social Deprivation Index (SDI), and Community Assets and Relative Rurality (CARR) are reported as estimates, as specific score values were unavailable for certain Zip Code Tabulation Areas (ZCTAs). ZCTAs/ZIP codes without available values were excluded from the index calculations.

Patients on clinic schedules were screened for eligibility through EHR review by a member of the study team. Identified eligible individuals were subsequently recruited by study staff via multiple methods: referral by clinic providers or staff during appointments, dissemination of IRB‐approved advertisements and brochures, or mailing of an IRB‐approved letter. Potential participants expressing interest were approached by a study coordinator who reviewed the IRB‐approved Informed Consent Form with them. Participants consenting to participate provided written informed consent prior to enrollment in the trial.

2.3. Interventions

Participants were cluster‐randomized by enrollment clinic into one of two groups using a 1:1 allocation. Allocation concealment was maintained through the electronic data capture system, and study personnel responsible for enrollment and assignment did not have access to the random allocation sequence. The preemptive PGx testing group receives preemptive PGx testing with results entered into participants' EHR and consult note provided within 1 month of enrollment. The usual care group does not receive PGx testing until after their participation in the trial concludes. Participants have study visits at enrollment and at 6 months and 12 months after enrollment. REDCap electronic data capture tools are utilized to manage the randomization process and to collect and store data during the trial [25].

Buccal swab (or occasionally blood if buccal swab failed or if preferred by participant) samples were collected from all participants at enrollment for DNA isolation. Samples from the preemptive PGx testing group were processed promptly after collection, and PGx results were entered into the EHR. In contrast, samples from the usual care group have been collected, stored, and then genotyped only after the participant has completed all trial activities. If initial genotyping failed and an additional sample was necessary, patients were mailed a collection kit to provide a repeat buccal swab sample or asked for the collection of a blood sample for re‐testing.

PGx testing is performed at the CAP/CLIA‐certified UF Health Molecular Pathology Laboratory. The primary assay used is the GatorPGx Plus panel, a validated test covering 14 genes/gene regions and 62 variants relevant to major ancestral groups [12]. Due to technical issues with the equipment used for GatorPGx Plus testing, the previous generation, clinically validated GatorPGx panel, covering 8 genes and 32 variants [26], was used for samples genotyped between October 21, 2022, and March 22, 2024. After the technical issues were resolved, testing reverted to the GatorPGx Plus panel for the remainder of the trial.

For results return, the PGx results are first automatically imported into the EHR by the Pathology Laboratory. Based on the participant's PGx results, if a relevant drug is prescribed, automated clinical decision support alerts built into the UF Health EHR display to healthcare providers for gene‐drug interactions [27, 28, 29]. In addition, a PGx‐specialized pharmacist from the UF Health Precision Medicine Program interprets the results and provides a personalized consult note with PGx‐based pharmacotherapy recommendations, which is routed to both the patient's primary care and ordering providers (if they differ). Finally, participants are provided with their PGx results (by mail or in person) on a laminated, credit card‐sized card, along with a letter explaining its purpose and intended use for sharing PGx information with healthcare providers, particularly those outside the UF Health system. Participants are contacted approximately 10 days after their result card is mailed to confirm receipt of the card and answer any related participant questions.

2.4. Outcomes

TOPP UP has two prespecified primary outcomes. The primary effectiveness outcome is change in patient treatment satisfaction over the 12‐month follow‐up period, as measured by the global satisfaction score in the validated Treatment Satisfaction Questionnaire for Medication (TSQM) score version 1.4 [30]. The TSQM includes 4 domains: effectiveness, side effects, convenience of treatment, and global satisfaction [30]. As the TSQM was developed and validated for considering a single medication, questions in the instrument were nominally adapted for consideration of a participant's full medication regimen. The primary implementation outcome is the healthcare provider acceptance rate of PGx recommendations provided by EHR‐integrated clinical decision support when applicable medications are ordered within the 12‐month follow‐up period.

Predefined secondary effectiveness outcomes include comparison of the number of medication changes (related to the inclusion diagnosis/diagnoses) between the preemptive PGx testing group and the usual care group within the 12‐month follow‐up. Medication changes are defined as additions, removals, or dose adjustments of PGx medications or medications within the same drug class. In addition to the overall analysis, a subgroup analysis will be conducted among patients who actually received a PGx‐relevant medication. In addition, changes in the other TSQM domains [30], in medication adherence measured via Patient‐Reported Outcomes Measurement Information Systems (PROMIS) Medication Adherence Scale (PMAS) [31], and in quality of life measured via the Health‐Related Quality of Life‐4 (HRQOL‐4) [32] survey will be assessed. Additional secondary feasibility outcomes include comparisons between preemptive PGx testing and usual care groups with regard to the total number of healthcare encounters within the follow‐up period, PGx test turnaround time, and the amount of time participants report spending discussing medications with their healthcare providers. Participant self‐reported use of the PGx result card will also be evaluated. Potential harms were systematically assessed through physician safety assessments reviewed during quarterly study monitoring meetings.

2.5. Data Collection

Surveys are completed by participants in‐person, by phone, or electronically in REDCap at three time points—at the time of enrollment, at 6 months and 12 months after enrollment—and include the TSQM, the HRQOL‐4, the PMAS, and medical history [30, 31, 32]. Patient demographics such as age, gender, race, ethnicity, and ZIP code were collected at baseline via survey to characterize the study population. Socioeconomic data, including education level, employment status (including work in healthcare or biomedical fields), income, health insurance coverage, medical literacy, and household size, were gathered by the study team to identify potential barriers to the feasibility and patient satisfaction regarding preemptive PGx testing. To assess the feasibility and sustainability of the intervention, participants in the preemptive PGx testing group are asked implementation‐specific questions including satisfaction with the PGx testing process, understanding of their PGx results, utilization of the results card, and perceived usefulness of the results card in facilitating communication and medication management. Surveys are to be completed within a ±1‐month window from the designated follow‐up date at 6‐ and 12‐month post‐enrollment. Participant medical diagnoses, medication lists, and healthcare utilization were collected from the EHR. Any information that a participant is unable to provide is also obtained from the EHR, if possible.

Sociodemographic variables such as educational attainment, health insurance coverage, and poverty levels which are often utilized as descriptors for medically underserved populations were compared to state of Florida and national measures, using data from the 2023 one‐year US Census American Community Survey, as 2023 represented the midpoint year for study recruitment [13, 14, 21, 22, 23, 24, 33, 34, 35]. Poverty status was also determined using the 2023 US Department of Health and Human Services poverty thresholds [34]. Individual thresholds for poverty were calculated utilizing participant‐reported household size. Participants reporting household income below the 100% poverty threshold were classified as being in poverty [35]. As income was collected as a range, when a reported range overlapped a calculated poverty threshold, the midpoint of the range was utilized to determine if the threshold was met.

To characterize the recruited population and assist with estimating socioeconomic status and rurality, three geocoded indices were computed: the Social Deprivation Index (SDI), the Social Vulnerability Metric (SVM), and the Community Assets and Relative Rurality (CARR) score [36, 37, 38]. Participants' ZIP codes were cross‐walked to ZIP Code Tabulation Areas (ZCTAs) for mapping to the SDI. The SDI is a composite measure of area‐level social deprivation calculated utilizing seven household and demographic measures from the 2019 American Community Survey [36]. Higher SDI percentile scores indicate greater levels of social deprivation. A more recently derived and more robust measure, SVM, is a neighborhood‐level composite social determinants of health (SDOH) metric derived utilizing a model of 24 measures from the 2018 Agency for Healthcare Research and Quality SDOH data [37]. Higher SVM percentile scores indicate greater social vulnerability. Finally, the CARR index is a composite measure of rurality that incorporates 2019 measures of remoteness, population density, and physical infrastructure [38]. The ZIP code‐based CARR index applies the census block group‐defined CARR measure to ZIP code boundaries using area‐weighted averages, rescaled from 0 to 1 [38]. Higher values (closer to 1) of the CARR index correspond to increased rurality. For comparison purposes, population‐weighted estimates of SDI, SVM, and CARR were calculated using frequency weighting both statewide for Florida as well as nationwide. Using SDI as an example, this weighting was accomplished such that each ZCTA's index score was weighted by its population; for instance, a ZCTA with a population of 10 would contribute its SDI score as 10 identical observations in the dataset. For the statewide metric, only ZIP codes and ZCTAs in the state of Florida were included, and for the nationwide estimate, ZIP codes and ZCTAs in the United States were utilized. Population data for SVM and SDI were taken directly from their respective datasets. For CARR, ZIP code population estimates were derived from 2020 Census data, as the original dataset did not provide population counts. These statewide and nationwide calculated values remain estimates as SDI, SVM, and CARR are individually not available for all geographical areas.

2.6. Statistical Analysis

The TOPP UP trial was powered to detect a 10% (7.1‐point; SD = 22.6 points) difference in the global satisfaction domain of the TSQM with 90% power at a two‐sided alpha level of 0.05. Accounting for a maximum projected attrition rate of 21% over the 12‐month follow‐up period, a total sample size of 542 was needed to maintain adequate statistical power for the primary outcome analysis.

Descriptive statistics were used to summarize participant demographics and clinical characteristics. Categorical variables are presented as counts and percentages. Continuous variables are reported as means and standard deviations for normally distributed data or as medians and interquartile ranges (IQRs) for non‐normally distributed data. For the geocoded indices (SVM, SDI, and CARR), population‐weighted medians and IQRs were calculated. All statistical analyses were performed using R software (version 4.0.1).

3. Results

Between August 8, 2022, and August 13, 2024, a total of 3951 patients were screened for eligibility, and 542 eligible individuals consented to participate in the trial (Figure 1). Of these, 535 participants underwent randomization. After randomization, 10 individuals were excluded from the trial after withdrawing or being identified as screening failures, leading to a sample of 525 individuals being included at the time of writing.

FIGURE 1.

TOPP UP Participant Flow Through Screening, Randomization, and Baseline Analysis. EHR, electronic health record; PGx, pharmacogenetic. *Counts of ineligible individuals represent accurate totals; however, stratified counts are not mutually exclusive, as some participants met more than one exclusion criterion, failed to meet multiple inclusion criteria, or both.

Baseline characteristics for the 525 active participants are presented in Table 2. The median age (IQR) is 60 (49–67) years, with the largest age group (29.7%) between 61 and 70 years old. The cohort is predominantly female (66.3%) and racially diverse, with 31.2% identifying as Black/African American, and 5.1% reporting Hispanic or Latino ethnicity. Among the sample, 24% of individuals report being employed, while more than 60% report being retired or disabled. The majority of participants have public insurance coverage, including Medicare (31.5%) and Medicaid (29.1%), with 27.3% having commercial/private insurance and only 1.9% uninsured. Enrolled participants reside across North Central and Northeast Florida (Figure S1).

TABLE 2.

Participant demographics at enrollment.

Characteristics a	N = 525
Recruitment site
UF Health Family Medicine Main Street	365 (69.5%)
UF Health Total Care Clinic	91 (17.3%)
UF Health Family Medicine Old Town	68 (13.0%)
UF Health Family Medicine and Pediatrics‐ Elizabeth G. Means Center	1 (0.2%)
Age, years (IQR)	60 (49–67)
Sex
Female	348 (66.3%)
Male	176 (33.5%)
Prefer not to answer	1 (0.2%)
Race
White/European	344 (65.5%)
Black/African	164 (31.2%)
Asian	3 (0.6%)
Native American	2 (0.4%)
Other or Mixed Race	12 (2.3%)
Ethnicity
Hispanic or Latino	27 (5.1%)
Not Hispanic or Latino	496 (94.5%)
Prefer not to Answer/Unsure	2 (0.4%)
Health Insurance b
Commercial/Private	169 (27.3%)
Medicare coverage alone or in combination	195 (31.5%)
Medicaid coverage alone or in combination	180 (29.1%)
Other government provided insurance	63 (10.2%)
None	10 (1.9%)
Employment Status
Retired	171 (32.6%)
Disabled	159 (30.3%)
Employed	126 (24%)
Homemaker	17 (3.2%)
Seeking work	14 (2.7%)
Other/Unknown/Prefer not to answer	38 (7.2%)

Abbreviations: IQR, interquartile range; UF, University of Florida.

^a Categorical variables are presented as counts and percentages, and age is presented as median (IQR).

^b As patients could have more than one type of health insurance, percentage total exceeds 100%.

The TOPP UP population exhibits distinct characteristics compared to statewide and national averages (Table 3). The proportion of TOPP UP trial participants with an educational attainment of a bachelor's degree or higher (17.0%) is approximately half that of both the statewide and national percentages (32.6% and 33.5%, respectively). Compared with state and national figures (17.8% and 21.3%, respectively), a higher proportion of TOPP UP participants have Medicaid insurance (29.1%). Additionally, a higher proportion of enrolled individuals are at or below the poverty level (39.6%) compared with available state and national data (12.3% and 12.5%, respectively).

The trial population also exhibits higher social vulnerability and deprivation compared to the state and national population (Table 3). The median SDI for TOPP UP participants is 74.0 (IQR 52.0–84.3), higher than the state median (52.0 [IQR 29.0‐75.0]) and higher than the national median (50.0 [IQR 25.0–75.0]), indicating greater area‐level social deprivation. Similarly, the weighted median SVM percentile is 77.9 (IQR 43.1–92.9) in TOPP UP, exceeding the corresponding state and national medians of 50.1 (IQR 26.6–70.6) and 40.8 (IQR 16.5–66.5), respectively, suggesting that participants, at the median, live in ZIP codes more socially vulnerable than over three‐quarters of US communities. In contrast, the median CARR score among participants is 0.30 (IQR 0.28–0.32), comparable to state and national medians of 0.27 (IQR 0.25–0.29) and 0.28 (IQR 0.25–0.31), respectively.

4. Discussion

Our baseline data indicate that TOPP UP successfully recruited a large and demographically diverse cohort representing medically underserved patients across urban, suburban, and rural primary care settings. Baseline analyses indicate that participants experience greater socioeconomic vulnerability and social deprivation compared with statewide and national populations. These findings support the feasibility of enrolling medically underserved patients in a pragmatic randomized trial of preemptive PGx testing, setting the stage for evaluating both implementation and effectiveness outcomes.

The TOPP UP cohort demonstrated strong indicators of medical underservice. Compared to state and national estimates, participants reported lower levels of educational attainment, a higher proportion living at or below the federal poverty threshold, and higher Medicaid reliance [21, 22, 23, 24, 34, 35]. Indices of social vulnerability and deprivation further reinforce the recruitment of economically and socially vulnerable populations. TOPP UP participants had higher SDI and SVM percentiles than the Florida and national population medians, indicating residence in areas with higher socioeconomic disadvantages [36, 37]. In contrast, CARR scores (a rural‐specific metric) for participants were comparable to state and national levels, suggesting that geographic remoteness was not a main driver of vulnerability in this cohort, but that participants faced marked deprivation independent of rural status [38]. The SDI primarily measures socioeconomic deprivation using household and demographic variables, whereas the SVM captures a broader array of social determinants, including healthcare access, housing conditions, and community structure. The close alignment between both indices in TOPP UP supports the conclusion that multidimensional socioeconomic vulnerability is a consistent feature among recruited participants. This successful enrollment of a consistently vulnerable cohort establishes the foundation needed to evaluate preemptive PGx testing in this underserved population. Despite the proven benefit of preemptive PGx testing in reducing risk of adverse drug effects and improving therapeutic effectivness [6, 39, 40, 41], access to this new health technology remains limited in medically underserved populations, a group in which its effectiveness has yet to be evaluated through randomized controlled trials. The inverse equity hypothesis states that new healthcare technologies, such as PGx testing, tend to favor patients of higher socioeconomic status over medically underserved patients, who are often the last to receive access [42, 43]. Meanwhile, patients with lower healthcare access may benefit the most from preemptive PGx testing since they are prescribed medications with PGx guidelines at higher rates than populations with higher healthcare access [15]. TOPP UP will help address this critical gap by analyzing effectiveness and implementation metrics of preemptive PGx testing in a medically underserved population.

Regarding demographic diversity, 31.2% of participants identified as Black or African American, which is higher than both the Florida and national averages reported by the 2023 American Community Survey (14.9% and 12.1%, respectively) [44]. Hispanic or Latino participants comprised 5.1% of the sample, below both the Florida and national averages (27.4% and 19.4%, respectively), but consistent with the demographics of North Central and Northeast Florida (4%–12% across participating counties) [44]. This diversity addresses a gap in previous PGx studies, which have predominantly enrolled participants of European or Asian ancestry, limiting the generalizability of PGx evidence to other populations [45].

TOPP UP also fills another critical gap, the implementation of preemptive PGx testing in community‐based primary care settings. Previous preemptive PGx implementation trials or programs have largely included specialty or tertiary care environments [6, 46, 47]. Primary care offers a more convenient venue for preemptive PGx testing, as it alleviates pressures related to urgent result return and facilitates ongoing access to results for prescribers managing patients long term; however, additional evidence on clinical utility, cost‐effectiveness, and sustainable implementation integration is likely needed to support widespread adoption [19, 48]. Also, the pragmatic design of TOPP UP permits clinicians to exercise their judgment in response to PGx‐informed recommendations, providing a real‐world assessment of provider uptake and clinical utility within primary care settings serving medically underserved patients. Ultimately, we anticipate that the results of this trial will support future clinical trials assessing the effect of preemptive PGx testing on more concrete clinical outcomes.

Among the strengths of this study are its relatively large sample size and diverse participant composition, which have the potential to enhance the generalizability of the findings. The integration of participant questionnaires with EHR data further strengthens the dataset, enabling comprehensive capture of clinical, demographic, and socioeconomic information. Moreover, although conducted as a randomized controlled trial, the pragmatic nature of the trial allowed healthcare providers to exercise clinical judgment in accepting or denying recommendations from PGx‐specialized pharmacists, more closely mimicking real‐world clinical practice. However, several limitations should be considered. First, the open‐label nature of the trial introduces the possibility of bias in patient‐reported outcomes. However, all PGx results were entered into patients' medical records, making blinding impossible. Conducting the trial within a single health system may limit generalizability beyond North Central and Northeast Florida, where demographic and socioeconomic characteristics differ from those of other US regions. Comparisons to American Community Survey data and ZIP code‐based indices are only estimates, as American Community Survey estimates have inherent sampling and estimation uncertainties, and area‐level measures may not accurately reflect individual‐level socioeconomic conditions. Furthermore, differences in variable definitions and data collection methods between our study population and state or national datasets may somewhat affect the comparability of demographic and socioeconomic measures. Nevertheless, the inclusion of clinics spanning urban, suburban, and rural settings, serving a high proportion of medically underserved patients, enhances the relevance and applicability of the findings to this population.

In conclusion, the TOPP UP trial is designed to assess the feasibility of preemptive PGx testing in medically underserved patients, a population historically underrepresented in PGx research. Baseline data demonstrate successful recruitment of a representative cohort, laying the foundation for future analyses of implementation and effectiveness outcomes.

Author Contributions

A.M.M., C.L., I.F., B.E.G., J.D.D. wrote the manuscript; B.E.G., R.G.S., L.H.C., J.D.D. designed the research; A.M.M., B.E.G., I.F., E.N.E., M.L.N., E.J.C., K.A.N., R.E.J., G.H., J.Z.M. performed the research; A.M.M., I.F., B.E.G., C.L. analyzed the data.

Funding

Research reported in this publication was supported by the National Human Genome Research Institute of the National Institutes of Health under Award Number R01HG011800. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1: Supporting Information.