Workforce readiness for pharmacogenomics and key elements for sustainment within the Veterans Health Administration

Abstract

Aim:

Understanding barriers and facilitators to pharmacogenomics (PGx) implementation and how to structure a clinical program with the Veterans Health Administration (VA).

Materials & methods:

Healthcare provider (HCP) survey at 20 VA facilities assessing PGx knowledge/acceptance and qualitative interviews to understand how best to design and sustain a national program.

Results:

186 (12% response rate) surveyed believed PGx informs drug efficacy (74.7%) and adverse events (71.0%). Low confidence in knowledge (43.0%) and ability to implement (35.4–43.5%). 23 (60.5% response rate) interviewees supported a nationally program to oversee VA education, consultation and IT resources. Prescribing HCPs should be directing local activities.

Conclusion:

HCPs recognize PGx value but are not prepared to implement. Healthcare systems should build system-wide programs for implementation education and support.

There is increasing evidence for the impact of genetic variation on drug metabolism, transport, targets, immune response and kinetics (i.e., pharmacogenomics). Drug–gene associations have been identified that impact treatment for many common and serious medical conditions including depression, cardiovascular disease and cancer [1–4]. Use of reactive pharmacogenomic (PGx) testing can lead to a 30% decrease in adverse drug events as compared with usual care [5].

Clinical guidelines from the Clinical Pharmacogenetics Implementation Consortium (CPIC) and the Dutch Pharmacogenetics Working Group (DPWG) advise healthcare providers (HCPs) on how to use PGx information to inform prescribing [6–8]. Guidelines with level A or B evidence have now been developed for over 130 drug–gene pairs.

The Veteran Health Administration (VA), with 170 healthcare systems and 9 million Veterans receiving care, is the largest integrated healthcare system in the USA. One out of two veterans receiving care at VA have been prescribed a medication potentially impacted by PGx in the past 4 years [9]. Based on PGx variant prevalence and veteran demographics, it is predicted that 99% of Veterans carry at least one actionable pharmacogenetic variant. Many adverse drug events, tracked by the VA, are impacted by PGx [10].

Given the potentially significant impact, the VA began piloting, pre-emptive PGx implementation at selected VA healthcare systems in 2019 through the Pharmacogenomics Testing for Veterans (PHASER) program. The PHASER program used an implementation facilitation model which involved a comprehensive program designed by a national team and led by local clinical site champions that includes extensive education efforts for HCP as well as patients and significant adaptations of the electronic medical record (EMR) to allow PGx ordering, return of results and clinical decision support (CDS) through best practice alerts [11]. Since its inception, PHASER has been implemented at 71 VA healthcare systems and PGx testing has been completed on over 35,000 patients.

There are numerous factors that must be considered when preparing to put clinical evidence such as PGx into practice. Implementation science is the study of the methods used and the barriers and facilitators to implementation of evidence-based practices in clinical care [12]. Use of implementation science frameworks can help guide program development and provide a systematic way to evaluate observed barriers and facilitators to success. Iterative evaluations of the PHASER program using implementation science methods have been completed over the course of PHASER roll out so that the program could be improved as it expanded [13].

Based on the initial successes of PHASER as well as approval and funding from institutional leadership to form a permanent National Pharmacogenomics Program, VA sought to assess readiness of the broader VA healthcare system for enterprise wide PGx implementation, adoption and explore how best to structure the program for sustainability. We designed a survey for HCPs to assess PGx knowledge and key implementation constructs relevant to PGx implementation (e.g. relative advantage, self-efficacy) to understand VA HCPs readiness for PGx implementation. We also completed qualitative interviews with key national and local stakeholders and PGx subject matter experts (SMEs) on how best to design and sustain a national PGx program. While the structure of the VHA and its funding model will potentially require different implementation strategies than traditional US healthcare systems, lessons on workforce knowledge and acceptance have the potential to apply more broadly.

Materials & methods

This was a quality improvement evaluation to understand the knowledge, interest and acceptance of PGx among VA clinicians and leaders. The evaluation included a survey sent to all HCPs at a representative sample of VA facilities and qualitative interviews of key SMEs and stakeholders involved in the PHASER program. As a quality improvement evaluation, no IRB approval was required. All activities were conducted according to the principles of the Declaration of Helsinki. A VA contractor team, Titan Alpha LLC, was used for distribution of the survey, qualitative interview completion and data analysis. Survey development and interview guides were created under the direction of VA employees and academic collaborators.

Setting & participants

Of VA's 170 facilities, 22 were chosen to be surveyed that represented a diversity of sizes, geographic locations and complexity levels. A mix of PHASER and non-PHASER sites were included. See Appendix 1 for participating sites. For the survey, all clinical providers (MDs, DOs PharmDs, Advanced practice providers [APPs]) at selected VA facilities were invited to participate. For the qualitative interviews, SMEs were identified by PHASER leadership as individuals who might have insight into future program development. SME stakeholders were chosen from the following representative VA programs: pathology and laboratory medicine, pharmacy, oncology, cardiology, primary care, mental health, genomic medicine, neurology, National Center for Ethics in HealthCare and the Program Evaluation and Resource Center.

Methods & measures

VA email addresses for all clinical providers at facilities selected for the survey were identified using a centralized database. The survey was distributed by VA email and reminders were sent five-times at regular intervals between May and July 2022.

The survey was comprised of four sections: demographics, knowledge and beliefs about pharmacogenomics, current use of and interest in pharmacogenomics in practice, and preferences for pharmacogenomics practice integration. (See Appendix 2 for full survey.) Constructs from which questions were taken or adapted are shown in Table 1. The acceptability of intervention measure (AIM), intervention appropriateness measure (IAM) and feasibility of intervention measure (FIM) measures was used to measure intermediate implementation outcomes of acceptability, appropriateness and feasibility, constructs defined by Proctor and colleagues [14,15]. Each construct is measured using four questions with a response scale of 1 (strongly disagree) to 5 (strongly agree). An average is taken of responses to the four questions to determine a mean score for each outcome.

Table 1. . Survey constructs and measures.

Construct	Validated/published measure	Ref.
Demographics	NIH categories where possible
Knowledge and beliefs about pharmacogenomics
• Acceptability	Acceptability of intervention measure (AIM)	[14]
• Appropriateness	Intervention appropriateness measure (IAM)	[14]
• Feasibility	Feasibility of intervention measure (FIM)	[14]
• Relative advantage	Rogers' adoption questions: relative advantage; other published survey instruments	[16,17]
• Relative priority	IGNITE survey	[18]
• Self-efficacy	Other published survey instruments	[17]
Current use of and interest in pharmacogenomics in practice	Previously published survey instruments	[17,19]
Preferences for pharmacogenomics integration	Created content

For the qualitative analysis, a semi-structured interview guide was developed to understand SME knowledge and perceptions of PGx and views on the development of a national VA program. See Appendix 3 for interview guide. The interview guide covered potential benefits, barriers to implementation, the role of various stakeholders, education required and what would be necessary for sustainment of a national VA-funded PGx program, including VA policies and restrictions.

Interviewees were recruited by email after being identified by PHASER leadership as individuals who might have insight into program development and needs. Interviews were completed between April and July 2022. A primary interviewer, either a physician or a genomic medicine expert, led each interview, following the interview guide, and a second interviewer assisted by asking clarifying questions as needed. A note taker was present during the interviews to capture meeting content. Interviewed individuals agreed to participate in this QI process and to recording the interview for analysis purposes. Participants are anonymized for all analyses. Recordings of the individual interviews, along with transcripts and notes from the interviews, were stored in a secure folder.

Analytic plan

Quantitative analysis

Descriptive statistics summarize respondent demographics, knowledge and beliefs about PGx, reasons for not ordering PGx and roles and responsibility for PGx implementation steps. Logistic regression was used to analyze how responses to knowledge and belief questions were associated with: respondent demographics (age, years since training completion), provider type (MD/DO, PharmD/RPH, APP), clinical area and whether they were from a PHASER site. Linear regression was used to analyze the AIM, IAM and FIM measures [14]. Current use of and interest in PGx were also analyzed using logistic regression with the same covariates.

Qualitative analysis

Qualitative interviews were coded to evaluate for key themes across SMEs. All completed interviews were reviewed. Transcripts and recordings were compared with interview notes to ensure accuracy and clarity. A rapid analysis of emerging themes was completed for the first five interviews to provide early insight into the results and give an opportunity to shape future interviews based on gaps in knowledge. Two coders coded all interviews. The VA team was routinely debriefed on these themes and given the opportunity to provide feedback by the contractor. The coding structure was developed a priori from the research questions. The interview guide was expanded on throughout the interview and qualitative analysis process. It underwent multiple iterations of team review and consensus.

The qualitative software Dedoose was used for analysis, and two coders completed the coding [20]. The inter-rater reliability was evaluated to ensure the consistency of code application across interview records. As subsequent interviews were conducted, the list of emerging themes was updated and expanded, and the coding structure was supplemented with new codes and information.

Results

Quantitative analysis

Demographics

Of the 1551 HCPs that the survey was sent to, 186 (12%) responded. Of those invited, responses were received from 18.4% of pharmacists, 13.7% of APPs and 10.4% of physicians. Site response rates ranged from 6.8 to 20.4% (Appendix 1). Response rates did not differ between PHASER and non-PHASER sites. Demographics of survey respondents are shown in Table 2.

Table 2. . Demographics of survey respondents.

Demographic	N (%)
Female	97 (52.7)
Race
• Asian	29 (15.8)
• Black	6 (3.3)
• White	120 (65.2)
• Other	3 (1.6)
• Decline to answer	26 (14.1)
• Hispanic/Latino	9 (4.9)
Provider type
• MD/DO	104 (56.5)
• PharmD/RPH	26 (14.1)
• APP	50 (27.2)
• Other (DPM, RN)	3 (1.6)
• Decline to answer	1 (0.5)
Provider specialty
• Cardiology	9 (4.9)
• Hematology/oncology	12 (6.5)
• Medicine	34 (18.5)
• Primary care/hospitalist	54 (29.4)
• Psychiatry	42 (22.8)
• Surgery	30 (16.3)
• Other	2 (1.1)
Years since completing last training program
• <10 years	39 (21.2)
• 10–19 years	70 (38.0)
• 20–29 years	45 (24.5)
• 30+ years	24 (13.0)
• Decline to answer	6 (3.3)

DPMH: Doctor of podiatric medicine; DO: Doctor of osteopathy; MD: Medical doctor; PharmD: Doctor of pharmacy; RN: Registered nurse; RPH: Registered pharmacist.

Knowledge & beliefs

Respondents largely agreed/strongly agreed that PGx is or will become an important tool to improve drug efficacy (74.7%) and predict the risk of adverse events (71.0%) (Table 3). Despite these beliefs, only 43.0% felt confident in their knowledge of PGx or well informed about PGx (39.8%). PharmDs were more likely to feel well informed (53.8%) as compared with MD/DO (42.3%) and APPs (28%; p = 0.03). Likewise, confidence varied by provider specialty with some specialties being more likely than others to agree or strongly agree that they were confident in their knowledge (hematology/oncology 66.6%, cardiology 55.5%, psychiatry 54.7%, other medicine specialties 42.8%, primary care 33.3%, surgery 16.7%; p = 0.047). Few respondents thought that PGx was too difficult for providers to understand (12.9%) but those further out from training were more likely to think so (<10 years out 2.6%, 10–19 years out 8.6%, 20+ years out 23.2%; p = 0.04). Time since training did not affect providers' views of patients' ability to understand PGx results; however, clinical area did with cardiology and hematology/oncology being more likely to believe that PGx was too difficult for patients to understand (cardiology 44.4%, hematology/oncology 41.6%, surgery 30%, other medicine specialties 28.6%, primary care 24.1%, psychiatry 21.4%; p = 0.006).

Table 3. . Knowledge and beliefs of survey respondents.

Statement	N (%)
	Strongly agree	Agree	Neither agree nor disagree	Disagree	Strongly disagree
PGx testing is or will become a valuable tool to improve drug efficacy	72 (38.7)	67 (36.0)	34 (18.3)	10 (5.4)	1 (0.5)
PGx testing is or will become a valuable tool to predict risk of adverse events	69 (37.1)	63 (33.9)	45 (24.2)	6 (3.2)	1 (0.5)
I feel confident in my knowledge about the influence of genetics on medication therapy	21 (11.3)	59 (31.7)	38 (20.4)	46 (24.7)	20 (10.8)
I feel well informed about the role of PGx testing in therapeutic decision-making	20 (10.8)	54 (29)	34 (18.3)	49 (26.3)	27 (14.5)
PGx is too difficult for providers to understand	2 (1.1)	22 (11.8)	59 (31.7)	50 (26.9)	51 (27.4)
PGx is too difficult for patients to understand	12 (6.5)	38 (20.4)	56 (30.1)	57 (30.6)	21 (11.3)
In general, incorporating PGx would lead to better patient outcomes than not using it (relative advantage)	62 (33.7)	60 (32.6)	57 (31.0)	4 (2.2)	1 (0.5)
PGx would be more effective at guiding medication choice than our current process (relative advantage)	66 (35.9)	53 (28.8)	52 (28.3)	11 (6.0)	2 (1.1)
Using PGx would improve the overall quality of care for patients (relative advantage)	67 (36.4)	55 (29.9)	54 (29.4)	7 (3.8)	1 (0.5)
PGx is better than our current methods for selecting medications (relative advantage)	53 (28.8)	39 (21.2)	81 (44.0)	10 (5.4)	1 (0.5)
The information generated by PGx testing is important for patient care (relative priority)	60 (32.6)	62 (33.7)	57 (31.0)	5 (2.7)	0
I am confident in my ability to: (self-efficacy)
• Identify patients appropriate for PGx testing	24 (13.0)	47 (25.5)	54 (29.4)	37 (20.1)	22 (12.0)
• Counsel a patient on risks and benefits of PGx testing	30 (16.3)	42 (22.8)	50 (27.2)	36 (19.6)	26 (14.1)
• Interpret a PGx test	20 (10.9)	45 (24.5)	58 (31.5)	29 (15.8)	32 (17.4)
• Use PGx results to guide medication decisions	28 (15.2)	52 (28.3)	54 (29.3)	26 (14.1)	24 (13.0)

PGx: Pharmacogenomics.

For the AIM, IAM and FIM measures [14] of PGx use in clinical practice (score range 1–5), perceptions of acceptability was 3.9 (SD: 0.9), appropriateness was 3.8 (SD: 0.9) and feasibility was 3.7 (SD: 0.8). There was no significant difference in responses based on demographics or other characteristics listed above. HCPs generally rated the relative advantage and relative priority of PGx highly (50–66.3 and 66.3% respectively agree/strongly agree). HCPs had mixed confidence in their ability to perform steps necessary to integrate PGx testing into their practice with confidence expressed by 38.5% for identifying appropriate patients, 39.1% for counselling patients on the risks and benefits of PGx testing, 35.4% for interpreting a PGx report and 43.5% for using PGx results to guide medication decisions. Responses to all of the above did not vary based on respondent demographics or being at PHASER versus non-PHASER sites.

Current use & interest in pharmacogenomics

Most HCPs had not ordered a PGx test in the past year (N = 160, 87%), but those at PHASER sites were more likely to have placed an order (PHASER N = 18, 19.6% vs non-PHASER N = 6, 6.5%; p = 0.009). Statistical likelihood to order was not impacted by HCP demographics, years since completing training or clinical practice area. Most common therapeutic areas that testing was ordered for were mental health (N = 12, 6.5%) and cardiology (N = 8, 4.3%). The two most common reasons for not ordering (answers not mutually exclusive) were ‘did not know what test to order’ (N = 90, 48.9%) and ‘did not know how to order' (N = 83, 45.1%). Responses were similar between PHASER and non-PHASER sites (Table 4).

Table 4. . Reasons for not ordering pharmacogenomics test stratified by PHASER site variable.

Response†	N (%)
	Phaser (N = 92)	Non-phaser (N = 92)
Do not know what test to order	47 (51.1)	44 (47.8)
Do not know how to order	40 (43.5)	44 (47.8)
Do not know what to do with the information	1 (1.1)	6 (6.5)
Uncertain about clinical value of the test	33 (35.7)	23 (25)
Not applicable for my patient population	12 (13)	10 (10.1)
Privacy concerns	4 (0.04)	0
Patient declined testing	0	0
Not available at my facility	6 (6.5)	15 (16.3)
Other	2 (2.2)	8 (8.7)

^†Multiple answers accepted.

Despite lack of ordering of PGx, HCPs indicated strong interest in more training and education in PGx (N = 158, 85.9%). Top subjects of interest focused on how to practically apply PGx to clinical care (selecting appropriate patients for testing, 73.9%; PGx test interpretation, 72.3%; prescribing recommendations, 64.1%.)

Practice integration preference

When asked if informed consent should be required when ordering a PGx test, 52.7% (N = 97) of HCPs agreed and 28.3% (N = 52) were unsure. 35 (19%) did not think consent would be necessary. The majority felt it was acceptable when ordering a PGx test related to one medication, to test a panel of genes impacting multiple medications (N = 132, 71.7%) and that an EMR template for documenting the indication for testing should be required (N = 101, 54.9%). A minority (N = 43, 23.4%) felt that an electronic consult for approval to complete PGx testing should be required. HCPs were asked who should perform the various tasks associated with PGx testing. Responsibilities were recommended to be distributed predominantly between the provider ordering the PGx test and the provider prescribing the impacted medication. There was support for local staff being involved in the consenting process (N = 84, 45.7%) and monitoring for future medication changes impacted by PGx results (N = 56, 30.4%) (Table 5).

Table 5. . Who should be primarily responsible for tasks associated with pharmacogenomics ordering and use?

Task	N (%)
	Local staff (RN, PharmD)	Non-local staff (RN, PharmD)	Impacted medication prescriber	Provider who ordered pharmacogenomic test	Automated algorithm	No explanation required	Other/none of the above
Ordering testing	47 (25.5)	2 (1.1)	131 (71.2)	–	–	–	4 (2.2)
Consenting patient for testing	84 (45.7)	7 (3.8)	78 (42.4)	–	–	–	15 (8.2)
Reviewing PGx test results impact on current medications	42 (22.8)	5 (2.7)	65 (35.3)	34 (18.5)	33 (17.9)	–	5 (2.7)
Notifying prescriber of an actionable drug–gene interaction	52 (28.3)	6 (3.3)	–	64 (34.8)	55 (29.9)	–	7 (3.8)
Adjusting medications according to PGx test results	28 (15.2)	2 (1.1)	103 (56.0)	46 (25.0)	–	–	5 (2.7)
Explaining PGx test results to patients	41 (22.3)	4 (2.2)	58 (31.5)	72 (39.1)	–	1 (0.5)	8 (4.4)
Monitoring patients for future medications impacted by PGx results or changes in guidelines on PGx interpretation	56 (30.4)	8 (4.4)	43 (23.4)	28 (15.2)	43 (23.4)	–	6 (3.3)

PGx: Pharmacogenomics; RN: Registered nurse; PharmD: Doctor of pharmacy.

Qualitative analysis

Of 38 SMEs and clinical providers invited to interview, 23 (60.5%) individuals agreed to participate. Specialty area of those invited and who participated are shown in Figure 1. Key themes and findings are summarized in Table 6 and described further below.

Figure 1. . Interviewees invited and who participated by specialty area.

Table 6. . Summary of key qualitative themes and findings.

Theme	Findings	Question domain
Program development considerations	• Participants generally preferred a national-level program to mitigate concerns around test access and program resourcing • Participants noted that a national program could benefit VA by reducing adverse medication reactions and the costs associated with such events • As part of program development, participants indicated a need for clear legal and process guidelines to ensure data security and governance and address provider concerns related to liability • Actionable program metrics are key for messaging related to outcomes and ensuring equitable access to testing services, according to participants	• Program structure, implementation and sustainment • Perceived benefits and challenges • Policy and/or restrictions (national vs regional) • Health equity
Program structure and processes	• Interview participants supported the creation of a head program office tasked with overseeing implementation and daily operations • Most participants indicated that prescribers should be responsible for ordering tests and presenting actionable results to the patient • Participants recommended some level of consultation between the provider, the pharmacy and the lab to interpret test results, especially for instances with actionable results	• Program structure, implementation and sustainment • Program participation commitments/obligations • Provider role and education
Patient experience	• Most participants concluded that increased drug efficacy and a lower risk of side effects would result in better patient health outcomes • Patient concerns about data privacy may be allayed with targeted education that informs the PGx testing process and intended use • Stakeholder opinion varied about what patient population should be tested, but most agreed PGx testing should not be mandatory	• Patient role and education • Health equity • Clinical validity and utility • Perceived benefits and challenges • Specific medications (or classes) important for PGx testing
Provider Experience	• For providers, participants reported increased confidence in prescribing medications, enhanced therapeutic efficacy, and fewer adverse drug reactions for patients as the key benefits of PGx testing • If barriers related to the demand for further evidence and administrative bandwidth are improved, resistance to adoption and utilization of PGx testing may be reduced • Point-of-care training and education, workflow processes, and consistent engagement would increase provider buy-in, along with providing clinical decision support	• Clinical validity and utility • Perceived benefits and challenges • Program structure, implementation and sustainment • Provider role and education
Program Sustainment	• Long-term program sustainment would be promoted by support from leadership, messaging related to program metrics, and provider engagement to improve awareness and uptake of the program • Critical components of provider engagement include coordinating with the informatics team to ensure simplistic electronic health record test ordering and result display and incorporating PGx program data into existing clinical decision support mechanisms	• Program structure, implementation and sustainment • Program participation commitments/obligations • Perceived benefits and challenges • Provider role and education

PGx: Pharmacogenomics.

Should a PGx program be implemented, interview participants indicated that it would benefit VA. The two main benefits described were the potential for cost savings and the opportunity to improve the healthcare services available to the Veteran population.

“Strategic, developed deployment of genetic testing would save lives and potentially reduce cost of care for Veterans, particularly for those that have the specific variants.” (Informatics).

Program development considerations

Participants wondered how, without a national-level program, different levels of site uptake, provider training and awareness, and results returned via EHR would affect the clinical utility of the tests and existing disparities. They said that having a national program would mitigate differences in implementation which might result if healthcare facilities were using their own budget to support the program. One participant also noted that many of their patients are in rural areas, and without the benefit of the interconnected VA health system, these patients might have less access to healthcare services in general and sub-specialty services in particular.

“I would hope it would be a national program and not come out of the local lab's budget because I think that is another way to worsen disparities that already exist.” (Pathology and Lab Medicine Service).

Interviewees felt there needed to be one overarching PGx team or office to serve as a knowledge holder, policy developer and implementation driver. Participants consistently raised their frustrations with not knowing where to go with their questions or not having clear direction from a single unified source. This team should be adequately staffed with PGx SMEs and funded to support the administrative, logistical, educational and data analytics tasks inherent to program implementation.

Program structure & processes

In alignment with survey respondents, participants felt a Veteran's HCP should be the test-ordering entity as opposed to a more centralized ordering process although they did not indicate a particular provider specialty. Interviewees indicated that providers should be responsible for ordering based on their relationship with the patient, their oversight of and insight into the patient's medical record and history, and their likelihood of prescribing medications. Some said that pharmacists could also order the test based on their experience with medication dosing and contraindications.

“PGx is a rapidly evolving field, something a lot of clinicians aren't going to be up to date on for the latest literature in this area. PGx is a place where pharmacist can play a particular role and have a strong presence. There are many more clinical pharmacists in VA than cardiologists, by order of magnitude. I like idea of pharmacists being experts in this particular setting. In VA there is such a presence and availability of pharmacists and I think this fits their skillset perfectly. It could be either provider or pharmacist ordering the test, but in most cases a well-trained pharmacist will be more familiar with this topic than most providers” (Cardiologist).

When asked about interpreting results, providers said there should be some level of care team consultation and discussion with those involved in patient decision-making (provider, pharmacy, lab). Some suggested that the pharmacy should interpret the results. One participant noted that it would be important for providers to be involved in the process because:

“There's the medical experience piece that has to be brought into the picture because it isn't a 1–1, but that's why medicine is an art; it is not flipping a switch and the lightbulb comes on.” (Genomic Medicine).

However, interviewees indicated that the decision-making process may depend on the level of detail provided in the results. If the results are well described, the prescriber should be able to act on the findings; if they are not, additional discussion or consultation with pharmacy could be warranted.

Regarding presenting results to the patient, there was strong support for this being the responsibility of the HCP. However, participants acknowledged that varying levels of provider familiarity with PGx could be a barrier to their comfort in relaying results. Participants said that a pharmacist could also present results based on established patient relationships or whether medication changes are indicated.

Patient experience

Interviewees were divided on what types of patients should be tested. No participant stated that testing should be mandatory for patients at this time. Some expressed that all patients should receive pre-emptive PGx testing. This opinion was influenced by factors including the relatively low cost of testing. Participants noted that if testing was inexpensive and had minimal risk, it would provide a valuable reference in the patient's future. Interviewees with this perspective also noted that the timely return of testing affected their opinion. Timing was an important factor for mental healthcare because treatment choice often cannot be delayed for patients with an immediate need for medication, and current testing timelines require around two weeks for determination. Timing concerns also applied to cardiology care and the use of PGx testing for prescribing blood thinners and anti-platelet agents. Some expressed that testing should be limited to specific groups of patients who would be prescribed medications with known drug–gene interactions. One participant noted that most of the VA population does not receive medications with known drug–gene interactions, so testing should be limited to those who would be prescribed a medication with a known significant PGx interaction. Interviewees also discussed the potential for bias in patient testing. Several participants noted that if specific groups were tested, bias may be introduced with the targeting of specific racial groups that are more inclined to carry genes with known drug interactions.

While all participants believed education would be an important component for patients, like survey respondents, participants varied on the degree of consent that would need to be obtained for PGx testing. Many believed verbal consent and education with the patient, with notation of this consent and education in the patient chart or treatment plan, would be sufficient. Two participants noted that consent should not be required for the patient in any form because it is part of the treatment plan. One of these participants stated that “consent should not be required as providers don't consent patients for treatment guidelines.” (National Center for Ethics in Healthcare).

Provider experience

PGx testing gives providers the opportunity to combine precision medicine results with their own clinical expertise. Participants noted that PGx testing would provide a level of confidence to providers by easily identifying patients who would not tolerate certain therapies and improving clinical decision-making to select an alternative therapy. One participant noted that PGx testing would allow primary care providers to be more engaged in medication management by knowing appropriate guidance for dosing and medication selection without relying on a sub-specialist.

When asked about potential challenges providers may have when adopting PGx testing into clinical practice, some voiced administrative bandwidth concerns related to staff burnout and a general lack of time and resources. This burden creates resistance for providers who are overwhelmed with change.

“We are trying to retain and repair our workforce, so implementing a new program may be difficult and we may see a lot of resistance.” (Pharmacy).

It was noted that addressing issues of bandwidth and burnout could increase uptake and utilization. Ultimately, the goal to improve health outcomes in patient care is something the majority of interviewees agreed on.

Identifying and communicating the program's benefits to providers is key to their uptake. Half of the participants referenced the need to see the clinical impact before buy-in could be established. These participants believe that the burden of proof required to implement PGx has yet to be met but that it has promise.

“There are different domains of PGx testing and different clinical practice areas, so from any individual provider, the question is what is the benefit to my patients and my patient population? That's the burden of proof needed prior to implementing… You need to see evidence that it's effective and beneficial. Other objections could be cost to patient and cost to practice. Time cost for having conversations, but if the benefit is shown then it is worth the time cost. First and most important is knowing that it is beneficial to patients.” (Psychiatrist).

Most participants emphasized the importance of point-of-care training and education. Recommendations included in-person training, virtual workshops or webinars, and online training modules. Stakeholders specifically noted that point-of-care education on how to use test results was important. An interviewee suggested that a workflow of what to do with the results when they are abnormal would be helpful for clinical decision-making. Consistent engagement would also address risk concerns up front and normalize PGx testing into the culture. One participant said, “If there is more education, people will feel comfortable. If they feel more comfortable, they will order it.” (HCP) Regardless of the type of training and education selected, it is generally agreed that the most crucial factor is well written, carefully curated test results that are simple to understand.

“It's the same as a red, yellow, green approach-don't give these [medications], you might give these, and these are safe to give. Having the free will and value that the individual who is looking at the results of the test and deciding on medication they're giving, it's their responsibility to finally give what they think is best and what they're most experienced with.” (Pathology and Lab Medicine Service).

Program sustainment

As mentioned earlier, administrative burden is a large obstacle to the willingness of a provider to engage with new initiatives, and several participants specifically identified administrative burden as a challenge to PGx program uptake.

“This is more work, more data, and more noise. There are more clicks on the computer-inbox is getting even fuller with more data. You're not getting paid to do this, but you just have more work.” (HCP).

Reducing the administrative burden on providers is of the utmost importance to adoption and sustainability. Participants most commonly noted engagement of the informatics team and the integration of PGx testing into the technology that providers are already using daily as an important consideration. Providers want to order the test with one click, they want the results to be shown in an easily digestible manner where all other test results are displayed, and they want the information to be integrated throughout their CDS tools.

“Decision support tools are key. Those tools that come through the electronic medical record that help inform providers of the meaning of results are absolutely critical.” (Office of Mental Health and Suicide Prevention).

Providers described their frustration with not having an easy way to share or access test results. Others described a need for point-of-prescription alerting of linked test result information that could affect prescribing and dosing. In general, most stakeholders agreed that if the data were available and applicable, they would use it. However, if the data are not easily accessible, or the program does not seem like it will have a lasting presence, providers are not interested in investing their time and effort.

“The biggest barrier is informatics. Currently, there's not a good way to share results among providers. It's an informatic problem, it's not the test [or] getting people tested but how do we manage a test result that stays with the patient for years. Point of prescription notification [are] not well shared among providers. If it's in a lab package five years ago, I'm never going to see it.” (HCP).

Discussion

Through this quality improvement evaluation, we have found that HCPs see the potential value that PGx offers to them and their patients and that they feel PGx is acceptable, appropriate and feasible for their practice. Yet the majority do not feel knowledgeable about the science of PGx or how to use it in therapeutic decision making. However, HCPs recognize this limitation and are interested in learning more so that they can incorporate PGx into their practice. These findings both represent the gap in HCP education but also the opportunity to capitalize on an audience eager to practice precision medicine.

Stakeholder interviews demonstrated overwhelming support for PGx integration and a strong perception of clinical value and cost–effectiveness. Concerns regarding equity were identified and ways to mitigate that risk through centralized programming were noted. Having a national program was also seen as a way to provide consistent and comprehensive EMR tools needed to support PGx ordering, results integration, and CDS for HCPs. SMEs and HCPs both supported the prescribing clinician as the driver of PGx integration at the patient level. Yet they also felt that having robust and easily accessible resources through pharmacy would be important for the program to succeed.

There are limitations to conclusions drawn from this evaluation. While the VA sites surveyed were representative of the larger VA system, there is the potential for responder bias. HCPs most interested in PGx or those most strongly opposed may be more likely to have responded to the survey. Survey response rate was low at 12% though HCPs are known to have low response rates as compared with other groups [21,22]. While the rate was lower than average for an HCP survey, the knowledge gained from those responses still has value as VA plans for the future. The qualitative interviews were done with SMEs pre-selected by the PHASER program office and therefore views of key personnel not identified by the PHASER program would have been overlooked. Additionally, application of these findings to healthcare systems beyond the VHA are limited by differences in finance models (i.e. insurance reimbursement model vs federally-funding program) and limits of PGx (and other) data visibility between healthcare systems as compared with VHA's nationally integrated electronic medical record.

In 2023, VA began the process of using the information gathered through this evaluation and the prior experience in PHASER to design a national PGx program which will be nationally funded but locally supported. This National Pharmacogenomics Program (NPP) launched in 2023 and will serve as a centralized program with both prescriber and pharmacist SMEs to support the national implementation of PGx across VA. The learnings from this study have been crucial for NPP to recognize the gaps in education for clinicians and the important infrastructure, funding, and roles and responsibilities that need to be considered to make a national roll-out of PGx effective. The full scope of the NPP is still under development but early goals include funding the training and integration of local, regional, and national PGx-trained pharmacists that can support all 170 VA healthcare systems in PGx education, consultation, population health management and implementation. Given pharmacists' increasing clinical role in healthcare as clinicians that can order laboratory testing and, in some cases, modify and prescribe medications, pharmacists are in a unique position to support PGx implementation and unburden other HCPs both within VA and in other healthcare systems as well [23]. Providers responded that medication prescribers should retain primary responsibility for ordering tests and making medication adjustments but that local staff can play a supportive role. The NPP interpreted these sentiments to create a population health management approach where PGx-trained pharmacists will be able to provide clinical review for important drug–gene interactions through dashboards that integrate not only PGx data but also, comorbidities, concomitant medications, patient specific variables (i.e., organ function) and prior drug history into a consolidated medication plan that can be communicated with the prescriber and patient. Newly trained PGx pharmacists will be educators of not only pharmacy peers but also physicians and APPs using a variety of methods as outlined in this study's findings, including point-of-care educational approaches well-established at VA through Academic Detailing Services [24].

Many interviewees endorsed the need for specific indications for PGx testing. The PHASER program allowed for ‘pre-emptive’ testing without a specific indication or medication given the donation of PGx testing to the VA. However, based on these findings and a transition to VA-funded PGx testing, the NPP is programmatically shifting from a pre-emptive approach for PGx testing to a reactive approach to PGx testing. To support this approach the NPP has established a variety of subject matter expert (SME) workgroups comprised of physicians and pharmacists to evaluate the evidence for clinical utility for a given drug–gene interaction and make recommendations for testing in specific patient populations. SME workgroups are supported by the NPP and liaise with national stakeholders on their work products. Once patient populations are identified, SMEs develop specific workflows to integrate PGx testing, pharmacist review and communication of actionable findings, criteria for population health management dashboards, and provider facing educational materials that can be used by field-based PGx pharmacists with relevant providers. To date the NPP has formed SMEs in cardiology, mental health, primary care, oncology, neurology and rheumatology, with others to follow.

It is thought that this more targeted approach to testing will lend itself to more efficient opportunities to support provider education. This will highlight drug–gene pairs that will be most impactful to veteran care and are most relevant to a given provider's subspecialty. For example, a targeted approach to testing in cardiology would focus on acute coronary syndrome populations that require stenting and a P2Y12 inhibitor. Detailed information on when to order the PGx test and how to utilize the results is then offered to providers who practice in the area this drug gene pair would be applicable to. Further, academic detailing services can be deployed to support the individual provider within their given subspecialty and at the appropriate level of expertise given the provider's background and experience with PGx. By providing dedicated resources for PGx through formation of the NPP, VA continues to further innovative practice and grow precision medicine's impact for the benefit of Veterans.

Conclusion

HCPs recognize the value of PGx in clinical care but do not feel prepared to implement without significant resource development and education efforts. Healthcare systems should aim to build system-wide PGx programs to support the provider education, CDS and information technology needs of PGx implementation.

Executive summary.

• Increasing evidence for the clinical utility of pharmacogenomics in clinical care but there are still significant barriers to widespread adoption and utilization.

• This was a quality improvement project to evaluate Veteran Health Affairs (VHA) health care providers (HCPs) views on pharmacogenomics and readiness to adopt. HCPs from representative VHA facilities were surveyed using established implementation constructs. Qualitative interviews were completed with subject matter experts and key stakeholders on the barriers and facilitators to VHA enterprise-wide implementation and how an ideal program would be structured.

• The majority of HCPs believed that pharmacogenomics was important to consider in medication prescribing but they did not feel that they presently had the knowledge or skillset to integrate it into their clinical practice. They were interested in further practical education on pharmacogenomic implementation.

• Subject matter experts and stakeholders believed that a nationally led program for pharmacogenomic implementation across the VHA would be important to providing equitable care to all Veterans with equal access for all VHA facilities to funding, HCP education, and implementation tools and support.

• VHA is currently developing the National Pharmacogenomics Program to provide testing to all Veterans. Findings from this quality improvement study have informed how the program is being structured and what supports are being established to address the needs to HCPs in delivering high quality pharmacogenomic informed care to Veterans.