2. Purpose of this review

Each stakeholder has a unique perspective on the role of PGx testing, although all are similarly challenged with demonstrating or appraising its cost-to-benefit value. Coverage by insurers is a critical step in achieving widespread adoption of PGx testing, and the current and future landscape relative to reimbursement is a focus throughout this review. In the following sections, trends in PGx testing utilization and uptake by payers in real-world practice are discussed, the role of pharmacoeconomics in assessing cost-effectiveness is highlighted using a case study in psychiatric care, and several issues that will affect adoption of PGx testing in the US over the next few years are reviewed. The purpose of this review is to increase the understanding of the current state of PGx testing, discuss key factors impacting its acceptance, and to foster additional research and evidence generation in the field.

3. Current real-world PGx-testing practice

Over 140 US Food and Drug Administration (FDA)-approved drugs have PGx information in their labelling (FDA, 2014 b ) that refers to pharmacodynamic relationships, such as identification of a specific gene that correlates with a drug mechanism of action, and/or pharmacokinetic mechanisms that describe drug metabolism effects on efficacy and ADRs. For a limited number of drugs, labelling mandates PGx testing and specific actions are taken based on the PGx information (examples, primarily highlighting the area of oncology, are provided in Table 1(a)). Where PGx testing has been identified as necessary for the safe and effective use of the corresponding therapy, companion diagnostic tests have been approved (premarket approval) or cleared (equivalent to marketed device) by the FDA (FDA, 2014 a ) (examples provided in Table 1(a)). In these situations, the cost for testing indicated by the FDA is generally reimbursed by most insurance plans (Hresko & Haga, Reference Hresko and Haga2012; Graf et al., Reference Graf, Needham, Teed and Brown2013). This is well illustrated in the field of oncology. Since many cancers are not viewed as a single disease, but rather as a group of several subtypes, each with a distinct molecular signature, identifying the genomes of the malignancy and of the patient can aid in effective and safe treatment.

Table 1. Examples of drugs with PGx biomarker and use, labelling status, companion PGx test, and payer coverage.

Other drugs have labelling with PGx specific guidelines for a patient subtype but don't have a required companion PGx diagnostic. For example, ivacaftor is approved for cystic fibrosis therapy specifically for patients with the G551D mutation, which occurs in 4% of patients (Clancy et al., Reference Clancy, Johnson, Yee, McDonagh, Caudle, Klein, Cannavo and Giacomini2014). The majority of cystic fibrosis patients have a mutation panel performed at the time of diagnosis as an existing standard of care; labelling guidelines require an FDA approved mutation test in cases where the genotype is not known (Vertex, 2014). In either case, costs are typically reimbursed (Hresko & Haga, Reference Hresko and Haga2012; Graf et al., Reference Graf, Needham, Teed and Brown2013). Examples of other drugs with recommended, but not mandated, PGx guidelines are shown in Table 1(b). While these examples are not comprehensive, they illustrate the diversity of drugs and therapeutic areas where the FDA has approved or cleared PGx tests, but the companion testing is not required prior to prescribing. Coverage for testing in these cases varies widely by insurers.

However, PGx-specific labelling for many drugs is informational only, and testing is neither mandated nor recommended prior to prescribing. Table 1(c) provides examples of two such cases (rovustatin and warfarin). In these cases, PGx testing has generally not been endorsed by expert committees since the supporting evidence has been judged statistically insufficient, and consequently, insurers will not pay for it (Hresko & Haga, Reference Hresko and Haga2012; Graf et al., Reference Graf, Needham, Teed and Brown2013). This is particularly relevant for tests related to genotyping for metabolizing enzyme polymorphisms for use with drugs having narrow therapeutic ratios and/or serious toxic effects. Figure 1 illustrates the growth of PGx testing in this area between 2011 and 2013 using data available from a private practitioner medical claims database (Symphony Health Solutions). FDA approval or clearance is not a requirement, so many are laboratory developed tests (i.e. offered and used within a single laboratory) that do not undergo FDA review. However, there is currently new FDA draft guidance in review that may change the regulatory oversight of laboratory developed tests (FDA, 2015 b ). For acceptance and reimbursement of these tests by payers, comparisons with standard methods of assessing therapeutic drug safety and efficacy are needed, as well as subsequent demonstration of value. Since factors other than genetic polymorphisms (including race, gender, epigenetic factors and lifestyle choices) can have a significant effect on drug pharmacokinetics and pharmacodynamics, it remains unclear whether PGx testing would eliminate the need for simultaneous use of other methods of therapeutic drug monitoring.

Fig. 1. Increase in PGx testing from 2011–2013. Figure shows the change in unique patient counts for selected assays from 2011 through 2013. On 1 January 2012, Medicare requested that claims for Molecular Pathology Procedures reflect both the existing CPT ‘stacked’ test codes that are required for payment, and the new single CPT test code. Patient counts are based on CMS 1500 claims and are courtesy of Symphony Health Solutions private practitioner medical claims database. CPT, Current Procedural Terminology.

In the area of therapeutic drug monitoring a recent example with potentially broad applications has been the use of PGx testing to guide warfarin anticoagulation therapy for the prevention of thromboembolic events. Two genes have been identified that play a role in outcomes related to warfarin therapy: CYP2C9 is a gene for an enzyme that is primarily responsible for the metabolism of warfarin, and VKORC1 codes for the warfarin drug target. The current warfarin label acknowledges that dose requirements are influenced by CYP2C9 and VKORC1 and states that genotypic information, when available, can assist in selecting the starting dose, but does not recommend routine PGx testing for determining initial or maintenance doses. Several diagnostic tests are cleared by the FDA for monitoring of warfarin anticoagulation (FDA, 2015 a ). However, the lack of standardization of these tests in detecting the relevant CYP2C9 variants, lack of understanding of the role of specific VKORC haplotypes and varying result turnaround times (1 to 8 hours) have raised skepticism regarding their decision-making utility (Rosove & Grody, Reference Rosove and Grody2009). Randomized controlled clinical trials examining use of PGx-guided warfarin dosing compared with current clinical data monitoring approaches have shown inconsistent results in stabilization of treatment responses, reducing the number of necessary office visits or decreasing risk of bleeding events. Consequently, consensus guidelines (e.g. American College of Chest Physicians, American College of Medical Genetics), do not advise use of genetic testing to guide warfarin dosing (Flockhart et al., Reference Flockhart, O'Kane, Williams, Watson, Gage, Gandolfi, King, Lyon, Nussbaum, Schulman and Veenstra2008; Guyatt et al., Reference Guyatt, Akl, Crowther, Gutterman and Schuunemann2012). As a result, most payers (including the Centers for Medicare and Medicaid Services (CMS)) are reluctant to reimburse for testing for PGx-guided warfarin therapy. However, it is unclear what role patient compliance with treatment contributes to the inconsistent efficacy and health outcomes, and studies addressing non-genomic factors' influence on treatment outcomes are critical to justify modification to standards of care and reimbursement decisions.

The drug clopidogrel, an anti-platelet therapy, is another example where genotype-specific labelling guidelines do not meet the indeterminate threshold for standard of care testing. CYP2C19 genotyping can determine normal, poor, intermediate, rapid and ultra-rapid metabolizers of clopidogrel in order to identify patients who may benefit from non-standard, ‘recommended’ dosing or have better responses to an alternative drug. In 2010, the FDA added a boxed warning to clopidogrel labelling alerting patients and health care professionals that the drug can be less effective in people who cannot metabolize the drug to convert it to its active form, and modified the warning to include guidance on data suggesting adverse cardiovascular event risks for acute coronary syndrome (ACS) patients with the CYP2C19 genotype following a percutaneous coronary intervention (PCI) (Bristol-Myers Squibb, 2013 b ). However, the boxed warning does not mandate genetic testing, is not specific on the exact patient profile to benefit from genetic testing, and is vague on alternative treatment approaches in poor metabolizers. Consensus guidelines issued by the American Heart Association (AHA) and American College of Cardiology (ACC) express concerns about the lack of outcomes data documenting the benefit of routine genotyping and concluded the available literature did not support CYP2C19 genotyping for all patients being prescribed clopidogrel (Holmes et al., Reference Holmes, Dehmer, Kaul, Leifer, O'Gara and Stein2010). AHA and ACC guidelines to date support CYP2C19 genotyping in ACS patients at high risk for poor outcomes after PCI or in patients receiving antiplatelet therapy for whom the genotype poses potential risk for reduced antiplatelet efficacy (Scott et al., Reference Scott, Sangkuhl, Stein, Hulot, Mega, Roden, Klein, Sabatine, Johnson and Shuldiner2013). Genotyping for one CYP2C19 polymorphism in patients prescribed clopidogrel is reimbursed by some payers (Hresko & Haga, Reference Hresko and Haga2012; Graf et al., Reference Graf, Needham, Teed and Brown2013).

Genetic testing has also been considered to mitigate the risk of serious adverse events with statins, which are generally safe and well tolerated and among the most widely prescribed medications (e.g. almost 50% of adults 65 years and older have statin prescriptions (CDC, 2013). Myopathy is the most common side effect and can result in life-threatening rhabdomyolysis with muscle damage and acute renal injury. Although mild myalgias may not result in any physical harm, they threaten patient adherence to statin therapy. The incidence of statin-associated myopathy varies between 1 and 5% in clinical trials, with a higher incidence observed in clinical practice (Thompson et al., Reference Thompson, Clarkson and Karas2003). There is a heritable component to the risk for statin-induced myopathy and the role for PGx testing in guiding statin pharmacotherapy is evolving, particularly for the solute carrier organic anion transporter family, member 1B1 (SLCO1B1) gene. Routine PGx testing for statin therapy is not recommended by current guidelines (Ramsey et al., Reference Ramsey, Johnson, Caudle, Haidar, Voora, Wilke, Maxwell, McLeod, Krauss, Roden, Feng, Cooper-DeHoff, Gong, Klein, Wadelius and Niemi2014; Talameh & Kitzmiller, Reference Talameh and Kitzmiller2014) nor covered by most payers. Additional studies are needed to determine clinical utility and cost-effectiveness of PGx testing in patient populations given factors including the type of statin (as pharmacokinetic profiles for each statin are unique), the dose and concomitant use of other drugs (Sorich et al., Reference Sorich, Wiese, O'Shea and Pekarsky2013; Talameh & Kitzmiller, Reference Talameh and Kitzmiller2014).

Inclusion of informational PGx tests by the FDA on the revised labels of many drugs without clear guidance on dosing recommendation and/or therapeutic alternatives by PGx result has led to confusion among providers, who are eager for guidance around an emerging technology with which they are largely unfamiliar. Professional organizations are attempting to educate and guide physicians on drug selection and dosing for drugs with PGx informational labelling. For example, the Clinical Pharmacogenetics Implementation Consortium has issued dosing guidelines taking into consideration patient genotype for warfarin, opioids, abacavir and other drugs with and without PGx informational labelling (CPIC, 2015). However, although these organizations can provide some level of guidance, challenges remain as there are limited nationally and internationally accepted standards around dosing decisions. Continued education of health care providers regarding the impact of metabolizer status and frequency of variants in a given population is needed to fully leverage PGx testing and information.

4. Pharmacoeconomic analyses to demonstrate value: case study of PGx-guided psychiatric intervention

The use of PGx testing needs to exhibit clinical, economic and quality of life benefits; demonstrating one or all of these values is critical to adoption among stakeholders. In addition to reducing side effects and ADRs, goals of PGx testing include lowering costs and improving patient compliance. When PGx testing results guide treatment decisions, positive treatment responses will be more quickly attained, which in turn, in theory, reduces the indirect and direct costs associated with failed treatment trials and ADRs. As long as the PGx testing costs and practitioner time for interpretation are lower than these indirect and direct costs, the net result is cost savings. In addition, patient quality of life improves when treatment failure and ADRs are reduced. While indirect costs are important considerations for providers and patients, payers are most influenced by evidence of direct impact, thus research has a multifaceted challenge of proving value on several levels. One well-established approach for determining value is to assess both costs and outcomes within pharmacoeconomic analyses. There have been relatively few published cost-effectiveness analyses of PGx interventions. The following example demonstrates how PGx-guided psychiatric intervention was associated with increased compliance with therapy and direct cost savings (Fagerness et al., Reference Fagerness, Fonseca, Hess, Scott, Gardner, Koffler, Fava, Perlis, Brennan and Lombard2014). While cost-effectiveness data are not the only information needed to influence acceptance of PGx testing by stakeholders, payers are particularly interested in these data, and the ability to use real-world claims and clinical databases is an efficient method to demonstrate utility in the absence of randomized clinical trials.

Known pharmacodynamic and pharmacokinetic variations exist that impact responses to common psychiatric medications (Burroughs et al., Reference Burroughs, Maxey and Levy2002). An estimated two-thirds of patients with depression will not respond adequately to first line treatment and more than one-third of patients will become treatment resistant (Kessler et al., Reference Kessler, Berglund, Demler, Jin, Koretz, Merikangas, Rush, Walters and Wang2003; Souery et al., Reference Souery, Papakostas and Trivedi2006; Warden et al., Reference Warden, Rush, Trivedi, Fava and Wisniewski2007). Genetic variations affect efficacy and tolerability of psychotropics as a result of variable drug disposition, metabolism and transport (Evans & McLeod, Reference Evans and McLeod2003; Kao et al., Reference Kao, Fang, Zhao and Kuo2011; Porcelli et al., Reference Porcelli, Drago, Fabbri, Gibiino, Calati and Serretti2011; Salloum et al., Reference Salloum, McCarthy, Leckband and Kelsoe2014). Many antidepressants and antipsychotics show differences in plasma drug levels as a result of cytochrome P450 polymorphishms (e.g. CYP2D6 and CYP2C19) (Kirchheiner et al., Reference Kirchheiner, Nickchen, Bauer, Wong, Licinio, Roots and Brockmoller2004; Tansey et al., Reference Tansey, Guipponi, Hu, Domenici, Lewis, Malafosse, Wendland, Lewis, McGuffin and Uher2013). The Genecept™ Assay (Genomind) is a laboratory validated test used to analyse variations in both pharmacodynamic and pharmacokinetic genes associated with treatment response, side effects, metabolism, tolerability and overall efficacy of psychiatric medications (Kato & Serretti, Reference Kato and Serretti2010; Lencz et al., Reference Lencz, Robinson, Napolitano, Sevy, Kane, Goldman and Malhotra2010; Zhang et al., Reference Zhang, Lencz and Malhotra2010; Psychiatric GWAS Consortium Bipolar Disorder Working Group, 2011; Bhat et al., Reference Bhat, Dao, Terrillion, Arad, Smith, Soldatov and Gould2012; Porcelli et al., Reference Porcelli, Fabbri and Serretti2012). Despite its commercial use for over 4 years in more than 30 000 patients, the Genecept Assay is considered ‘experimental’ by many insurers. Typical to challenges in this field, while this assay has accumulated multiple levels of data arguably supporting its utility, the evidence to overcome this label is rigorous, but paradoxically not well defined.

One such piece of evidence is a retrospective, observational study that assessed patient claims data to measure direct health costs as a result of patient and clinician access to genetic information during psychiatric treatment selection guided by the Genecept Assay (Fagerness et al., Reference Fagerness, Fonseca, Hess, Scott, Gardner, Koffler, Fava, Perlis, Brennan and Lombard2014). Genetic test results and pharmacy and medical claims were integrated at the patient level. Propensity score matching (PSM) was used to identify case and control patients to reduce potential sources of bias (Baser, Reference Baser2006; Austin, Reference Austin2010; Luo et al., Reference Luo, Gardiner and Bradley2010; Austin, Reference Austin2011). PSM adjusted for covariates of patient demographics, payer type, medical data and treating practitioner's specialty, and identified 111 case and 222 matched control patients from pools of over 1000 case and 100 000 control patients. Patient's highest adherence levels pre-index vs. post-index were compared (index = date when genetic test results were available for tests and similar calendar date for controls). An observed increase in adherence to therapy in cases was associated with a significant decrease in overall costs associated with outpatient activity. Although pharmacy costs for cases increased relative to controls, likely due to consistent medication fills concomitant with improved medication adherence, the overall use of medical services decreased, and the relative cost savings over a 4-month follow-up period was 9·5%, or $562 per patient (Fagerness et al., Reference Fagerness, Fonseca, Hess, Scott, Gardner, Koffler, Fava, Perlis, Brennan and Lombard2014).

While randomized controlled trials remain the gold standard for clinical investigation, this example demonstrates how retrospective observational studies can use real-world clinical data to assess whether PGx-guided therapy can improve patient outcomes. This approach can assist in building the clinical and economic evidence base for PGx testing in order to provide information to stakeholders. To be effective in influencing uptake of PGx testing, health care providers will benefit from increasing awareness of the data on cost-effectiveness and economic utility of PGx.

5. Acceptance of PGx-guided therapy by insurers

6. Summary and conclusions

The adoption of PGx-guided therapy faces commercial, economic, educational and ethical barriers to integration into clinical practice and acceptance by practitioners, patients and payers. Acceptance of PGx testing varies considerably by therapeutic area. In some areas, such as oncology and cardiovascular disease, PGx testing is already utilized for selecting appropriate patients and/or establishing treatment and dosing guidelines, while in other areas such as antiplatelet therapy, the PGx approach has been mainly used for the identification, validation and development of new meaningful biomarkers, and is considered experimental by most stakeholders. The acceleration of adoption of precision medicine in general, and for PGx testing in particular, will be determined by how quickly robust evidence can be accumulated that shows a return on investment for payers in terms of real dollars, for clinicians in terms of patient clinical responses, and for patients in terms of economic, health and quality of life outcomes.