Original Article

The Pharmacogenomics Journal (2016) 16, 129–136; doi:10.1038/tpj.2015.39; published online 19 May 2015

Cost-effectiveness of one-time genetic testing to minimize lifetime adverse drug reactions

O Alagoz1, D Durham2 and K Kasirajan3

  1. 1Department of Industrial & Systems Engineering, University of Wisconsin-Madison, Madison, WI, USA
  2. 2Department of Psychiatry, Sage Neuroscience Center, Albuquerque, NM, USA
  3. 3Ally Clinical Diagnostics, Dallas, TX, USA

Correspondence: Dr K Kasirajan, 150 Alamo Springs Drive, Alamo, CA 94507, USA. Email: kk@allyclinicaldiagnostics.com or k.kasirajan@aol.com

Received 14 October 2014; Revised 12 February 2015; Accepted 7 April 2015
Advance online publication 19 May 2015

Top

Abstract

We evaluated the cost-effectiveness of one-time pharmacogenomic testing for preventing adverse drug reactions (ADRs) over a patient’s lifetime. We developed a Markov-based Monte Carlo microsimulation model to represent the ADR events in the lifetime of each patient. The base-case considered a 40-year-old patient. We measured health outcomes in life years (LYs) and quality-adjusted LYs (QALYs) and estimated costs using 2013 US$. In the base-case, one-time genetic testing had an incremental cost-effectiveness ratio (ICER) of $43165 (95% confidence interval (CI) is ($42769,$43561)) per additional LY and $53680 per additional QALY (95% CI is ($53182,$54179)), hence under the base-case one-time genetic testing is cost-effective. The ICER values were most sensitive to the average probability of death due to ADR, reduction in ADR rate due to genetic testing, mean ADR rate and cost of genetic testing.

Top

Introduction

Medication prescription is very common in both inpatient and outpatient settings. Due to the high volume of prescribed medications, safety is of critical concern in addition to the beneficial effects of pharmacotherapy.1 Recent data suggest that adverse drug reactions (ADRs) are a major cause of disability and death. In addition, even when medications cause no harm they are far too often ineffective.2 Pharmacogenomics is a promising area that has the potential to significantly improve healthcare outcomes by tailoring pharmacotherapy to individual patients. In particular, there is evidence that ADRs, which lead to >100000 deaths annually in the US, can be significantly reduced via pharmacogenomics.3

As a result, pharmacogenomics is rapidly gaining popularity to optimize drug delivery. As policymakers and providers attempt to prioritize high quality healthcare, they are confronted with a dearth of level I data on the use of personalized medicine, specifically, little data exists on the healthcare resource expenditures relative to possible medical benefit when pharmacogenomics testing is routinely used to help minimize ADRs. Such an analysis would be extremely valuable and necessary in setting priorities when choices must be made in the face of limited resources.

There are three main types of economic evaluations in healthcare: cost-effectiveness, cost-utility and cost-benefit analyses. The valuation of costs in all three types is made in monetary units, whereas they differ in the way the health outcomes are identified and valued.4 In a cost-benefit analysis, health outcomes are also measured in monetary units. In a cost-effectiveness analysis, health outcomes are measured using a single clinical effect of interest such as life-years gained, number of ADRs prevented and so on. On the other hand, in a cost-utility analysis, health outcomes are measured in single or multiple effects such as quality-adjusted life year (QALY), which is a composite measure of the quantity and quality of life. Therefore, cost-utility analysis may provide a better appreciation of the overall health benefits, harms and costs of laboratory tests in the diagnostic decision making process and the induced health outcomes. While cost-utility analysis is a broader form of analysis than cost-effectiveness analysis, many authors prefer not to make a distinction between the two types due to their similarity and use the two terms interchangeably.4, 5

The purpose of this study is to determine if pharmacogenomic testing is cost effective to minimize lifetime ADRs for a given age group. While pharmacogenomics testing is becoming more popular for helping to select treatment for a particular individual, there have been few studies considering the cost-effectiveness of pharmacogenomics testing over a patient’s lifetime in relation to several prescribed medications. All cost-effectiveness studies in the literature focus on patients who already are considered for treatment for a particular disease such as certain psychiatric illnesses, smoking cessation therapy and cardiovascular diseases.6, 7, 8 On the other hand, to the best of our knowledge, this work is the first to study the cost-effectiveness of one-time genetic testing over a patient’s lifetime.

Top

Materials and methods

A Markov-based Monte Carlo simulation model was built to estimate the incremental cost-effectiveness of one-time genetic testing compared with no genetic testing.9 The base-case scenario considered a 40-year-old patient and simulated his/her lifetime outcomes. During sensitivity analyses, patients at the age groups of 50, 65, 60 and 75 were also considered. Only direct medical costs were taken into consideration to avoid excessive confounding variables. Costs and benefits were discounted using a 3% discount rate as recommended by the cost-effectiveness panel.5

Estimates include LYs, QALYs, costs (in 2013 USD) and calculated incremental cost-effectiveness ratio (ICER) as cost per LY gained and cost per QALY gained. QALYs are similar to LYs with the significant difference being QALYs consider morbidity and quality of life effects on patients.5 In calculating QALYs, one needs to assign a utility score between 0 (representing death) and 1 (representing perfect health state) to the current health state and then adjust LYs using that value. For example, suppose a patient with a disease assigns a utility score of 0.8 for his/her disease status. Assuming that this patient lives in that particular health state for 1 year, his/her QALYs during that time is equal to 0.8 QALYs.

Markov model

A Markovian modeling approach was used to address time-varying events (Figure 1). The cycle time of the Markov model is 1 year, that is, the transitions occur on an annual basis. The model works as follows (assuming that we consider 40-year-old patients): A 40-year-old patient enters the simulation model. If the patient is not offered genetic testing, at every year the patient may experience an ADR that leads the patient to visit emergency department (ED) or outpatient clinic (OC) with a probability estimated from the literature. The ADR may lead to no major change in health, but these ADRs and their potential side effects are ignored in this analysis, hence we provide a conservative estimate of the ICER of genotyping. Once the patient visits ED/OC due to an ADR, given certain probabilities the patient may be hospitalized or may die. The simulation ends once the patient dies or reaches age 100.

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Decision tree representing the Markov model for genetic testing problem for a target age group. This figure shows the conceptual model used in this study. In the figure, the square represents decision nodes, the circles represent chance nodes (random events), the reverse triangles represent the outcomes/end points and the node with ‘M’ label shows the Markov nodes. A patient entering the simulation model is either offered a genetic testing or not. Then, at every age, the patient may experience an ADR that leads the patient to visit emergency department (ED) or outpatient clinic (OC) with a probability estimated from the literature. Once the patient visits ED/OC due to an ADR, given certain probabilities the patient may be hospitalized or die. ADR, adverse drug reaction.

Full figure and legend (211K)Download PowerPoint slide (287 KB)

Parameters in Table 1 were then used to populate this model. Among these parameters, three of them are particularly noteworthy: mean ADR rate (that is, annual probability of an ADR leading to ED/OC visit), rate of reduction in ADR rate with genetic testing and probability of death due to an ADR. As there is significant variability in the literature for these three parameters, an extensive sensitivity analysis was conducted.


Main sources for estimating mean ADE rate are the study by Bourgeois et al.,10 which provided an 11-year analysis of the nation-wide ADE rates in outpatient settings, and the study by Budnitz et al.,11 which described the frequency and characteristics of ADRs that lead to ED visits in the United States. The study by Bourgeois et al.10 considered both the ED and OC visits and therefore is more comprehensive. The study by Bourgeois et al.10 reported that 1.55% of the population visited an ED or OC annually due to an adverse drug event. The study by Budnitz et al.11 found that 0.24% (95% confidence interval (CI), 0.17–0.3%) of the population needed an ED visit because of an ADR. Two studies also provide age-specific estimates of ADR rates.11, 12 We used these age-specific estimates to calculate an age-specific ADR for 5-year age groups.11, 12 Given the mean ADR rate of 1.55%, we found the following age-specific ADR probabilities: for 40–44 years age group, the mean ADR rate is 1.27%, 45–49 is 1.27%, 50–54 is 1.40%, 55–59 is 0.52%, 60–64 is 1.65%, 65–69 is 2.16%, 70–74 is 2.66%, 75–79 is 3.68%, 80–84 is 4.31%, 85–89 is 4.95%, 90–94 is 4.95% and 95–99 is 4.95%.

The most challenging parameter to estimate was the rate of reduction in ADR rates due to genetic testing. For this purpose, we first used the study by Budnitz et al.11 to identify the sources of ADRs and their share in total ADRs such as warfarin (33%), insulin (14%), oral antiplatelet agents (13%), oral hypoglycemic agents (11%), opioid analgesics (5%) and others (24%). We then found various reported reduction rates in ADRs due to genetic testing. In particular, there was conflicting reports for the reduction rate in the use of warfarin due to genetic testing. While one study13 reported a 32% reduction in hospitalization, a recent article14 reported that there was no difference in anticoagulation control when genotyping was used. For oral antiplatelet agents, a 30% reduction in ADRs due to genetic testing was reported.15 Unfortunately, no other study directly reported ADR reductions with genetic testing. Under the most conservative assumption based on this data, the reduction occurs only in oral antiplatelet agents and no reduction occurs in others in which case the estimated rate of reduction due to genetic testing would be equal to 4%. On the other hand, under the largest reduction assumption that genotyping reduces ADRs in all drugs similar to oral antiplatelet agents, the estimated rate of reduction due to genetic testing would be 30%. Therefore, we chose the middle ground by assuming that the reduction in ADR is the average of 4 and 30%, and is equal to 17%.

The third important parameter is the probability of death due to ADR. The study by Lazarou et al.3 is the most widely quoted work on the incidence and mortality of ADRs for inpatients and outpatients. It should be noted that the incidence of ADRs and ADR deaths in the US has increased since 1998. This study estimated that in 1994 there were a total of 106000 deaths related to ADR and 4986000 total ADRs in the US. Using these estimates, the probability of death from an ADR is equal to 106000/4986000=0.0213—the figure used for our analysis. There is evidence that the probability of ADR-related death changes by age as reflected by more recent data from the Centers for Disease Control and Prevention (CDC).16 Specifically, this data logically points out that it is more likely for an older patient to die due to ADR than a younger patient. The 1999–2010 reports of ADR-related deaths at CDC were used to estimate the probability of death by age group for our analysis.16 As the study by Sarkar et al.17 noted, ‘the number of ADRs reported in death certificates may be falsely low because the terminal physiologic event may have been recorded as the cause of death,’ therefore we did not use the actual numbers in the CDC data, but used them to estimate age-specific odds ratios of death due to ADR. Using CDC data, with the mean probability of death due to ADR being 0.0213, we found the following age-specific probability of ADR deaths: for the 40–44 years age group, the probability of ADR-related death is 0.044552, 45–49 is 0.056999, 50–54 is 0.055462, 55–59 is 0.066211, 60–64 is 0.084245, 65–69 is 0.085929, 70–74 is 0.075963, 75–79 is 0.082042, 80–84 is 0.101135, 85–89 is 0.090054, 90–94 is 0.1354 and 95–99 is 0.106425.

Validation

Our model included data from various sources, hence, it was important to validate the data. For this purpose, we first calculated the expected number of certain outcomes using our parameter estimations then compared these estimates to those reported by the literature. More specifically, using our model, we estimated the number of ADR-related visits to ED/OC in a year, number of ADR-related ED visits in a year, number of ADR-related hospitalizations in a year, number of ADR-related deaths in a year and average cost of an ADR, and compared each of these estimates to those reported in the literature.

Top

Results

Validation results

Using our parameter estimates, there would be 4936750 ADR-related visits to ED/OC in 2014. It was reported that there were a total of 5348912 ADR-related visits to ED/OC in 2005, making our estimate conservative.10 Similarly, the study by Sarkar et al.18 estimated that there were 4.5 million annual ADR-related visits between 2005 and 2007, which is again consistent with our estimate.

In addition, using our parameter estimates, there would be a total of 1036078 visits to ED in 2014, whereas the study by Bourgeois et al.10 estimated that there would be 791082 ED visits annually between 1995 and 2005. On the other hand, another study reported that 28% of all outpatient ADR visits were to an ED, and they estimated that this corresponds to 1260000 ED visits a year, therefore our estimates are within the acceptable range.18

Using our estimates, there would be a total of 161243 ADR-related hospitalizations in 2014, whereas the study by Budnitz et al.10 estimated that there would be 107468 hospitalizations annually between 1995 and 2005. Considering that ADR rates have increased over time, we believe our estimates are within the acceptable range.

We also found that using our estimates, there would be a total of 105153 outpatient ADR-related deaths in 2014, whereas the study by Lazarou et al.3 estimated that there were 43000 outpatient ADR-related deaths in 1994. Using 1994 population figures, our outpatient ADR-related death estimate would be equivalent to 86872, that is, twice the estimate reported in the study by Lazarou et al.3

In terms of costs, we estimated that the average cost of each ADR-related ED/OC visit increases with age with ~$2600 for 40-year olds and $3750 for 84-year olds. A study by Bond et al.19 reported that the average cost of an ADR was $2401 per patient. Hence, the cost parameters of this study are consistent with our estimates.

Base-case results

The results are presented in Table 2. For a cohort of 40-year olds, the ICER of genetic testing versus no testing is $43165 (95% confidence interval (CI) is ($42769, $43561)) per LY and $53680 per QALY (95% CI is ($53182, $54179)). For every 1000 genetically tested 40-year olds, 95 projected ADR-related ED/OC visits, 6 projected ADR-related hospitalizations, and 3 projected ADR-related deaths would be prevented by genetic testing over lifetime. As presented in Table 2, the ICER increases by age. For example, for 65-year olds, the ICER rises to $61465 per QALYs.


Sensitivity analysis

Due to a lack of more current data on several parameters a one-way sensitivity analysis was conducted on the four most influential parameters: cost of genetic testing, mean ADR rate, rate of reduction in ADR rate with genetic testing and probability of death due to an ADR. The results of one-way sensitivity analysis are presented in Figures 2, 3, 4, 5. One-way sensitivity analyses show how the ICER changes with different parameter values. For example, as Figure 2 shows, if the cost of genetic testing is less than ~$2000, the ICER of genetic testing is <$100000 per QALY, and therefore is cost-effective. Conversely, if the cost of genetic testing is over $2000, it is no longer cost-effective to do genetic testing. Figure 6 includes a tornado diagram showing the summary of one-way sensitivity analysis on several parameters. A tornado diagram shows the ICER of genetic testing when the lowest and highest values of a parameter are used in the model. If the range of ICER is high for extreme values of a parameter, it indicates that the ICER is very sensitive to that parameter. As presented in Figure 6, the following parameters affect the results of cost-effectiveness analysis with respect to their effects on ICER: average probability of death due to ADR, reduction in ADR rate due to genetic testing, mean ADR rate, cost of genetic testing and discount rate. Hence, the most influential parameter affecting ICER is average probability of death due to ADR. On the other hand, we find that the cost of an OC visit, cost of an ED visit and QOL loss due to an ED/OC visit have almost no effect on the ICER.

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

ICER per QALY versus cost of genetic testing. This figure shows the one-way sensitivity analysis, where the x-axis represents the cost of genetic testing and the y-axis represents the ICER per QALY gained by genetic testing so the graph shows the change in ICERs as a function of cost of genetic testing. ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life year.

Full figure and legend (37K)Download PowerPoint slide (174 KB)

Figure 3.
Figure 3 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

ICER per QALY versus mean ADR rate. This figure shows the one-way sensitivity analysis, where the x-axis represents the cost of mean ADR rate and the y-axis represents the ICER per QALY gained by genetic testing so the graph shows the change in ICERs as a function of the mean ADR rate. ADR, adverse drug reaction; ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life year.

Full figure and legend (55K)Download PowerPoint slide (214 KB)

Figure 4.
Figure 4 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

ICER per QALY versus rate of reduction in ADR rate due to genetic testing. This figure shows the one-way sensitivity analysis, where the x-axis represents the rate of reduction in ADRs due to genetic testing and the y-axis represents the ICER per QALY gained by genetic testing so the graph shows the change in ICERs as a function of the rate of reduction in ADRs due to genetic testing. ADR, adverse drug reaction; ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life year.

Full figure and legend (46K)Download PowerPoint slide (183 KB)

Figure 5.
Figure 5 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

ICER per QALY versus probability of death given that the patient visited ED/OC due to an ADR. This figure shows the one-way sensitivity analysis, where the x-axis represents the probability of death and the y-axis represents the ICER per QALY gained by genetic testing so the graph shows the change in ICERs as a function of the probability of death. ADR, adverse drug reaction; ED, emergency department; ICER, incremental cost-effectiveness ratio; OC, outpatient clinic; QALY, quality-adjusted life year.

Full figure and legend (43K)Download PowerPoint slide (182 KB)

Figure 6.
Figure 6 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Tornado diagram. This figure summarizes the results of one-way sensitivity analysis. In this figure, the x-axis represents the ICER per QALY gained by genetic testing and the x-axis lists the parameters that were changed as part of one-way sensitivity analysis that were ordered with respect to their effect on ICER. ADR, adverse drug reaction; ED, emergency department; ICER, incremental cost-effectiveness ratio; OC, outpatient clinic; QALY, quality-adjusted life year.

Full figure and legend (59K)Download PowerPoint slide (216 KB)

Since there are multiple parameters that are used for this cost analysis with a possible wide range as reported in the literature, three two-way sensitivity analyses were conducted. The results of these analyses show the combinations of the values for two parameters at which ICER exceeds the accepted cost-effectiveness threshold of $100000 per additional QALY. Figures 7, 8, 9 show the results of the three two-way sensitivity analyses: average probability of death due to ADR versus the rate of reduction in ADR rate due to genetic testing, average probability of death due to ADR versus the cost of genetic testing and rate of reduction in ADR rate due to genetic testing versus cost of genetic testing. For instance, Figure 7 shows that if the rate of reduction in ADRs due to genetic testing is 10%, as long as the probability of death due to ADR is >0.02, the ICER is <$100000 per additional QALY, and therefore is cost-effective. In Figure 7, any values of these two parameters higher than the plotted line results in an ICER <$100000 per QALY.

Figure 7.
Figure 7 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Two-way sensitivity analysis for the average probability of death due to ADR versus rate of reduction in ADR rate due to genetic testing. This figure shows the two-way sensitivity analysis, where the x-axis represents the rate of reduction due to genetic testing and the y-axis represents the probability of death. The values above the line represent the combinations of these two parameters for which the ICER value is <$100000 per QALY gained. ADR, adverse drug reaction; ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life year.

Full figure and legend (47K)Download PowerPoint slide (185 KB)

Figure 8.
Figure 8 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Two-way sensitivity analysis for the average probability of death due to ADR versus the cost of genetic testing. This figure shows the two-way sensitivity analysis, where the x-axis represents the probability of death and the y-axis represents the cost of genetic testing. The values below (south-east) the line represent the combinations of these two parameters for which the ICER value is <$100000 per QALY gained. ADR, adverse drug reaction; ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life year.

Full figure and legend (47K)Download PowerPoint slide (186 KB)

Figure 9.
Figure 9 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Two-way sensitivity analysis for the rate of reduction in ADR rate due to genetic testing versus the cost of genetic testing. This figure shows the two-way sensitivity analysis, where the x-axis represents the rate of reduction due to genetic testing and the y-axis represents the cost of genetic testing. The values below (south-east) the line represent the combinations of these two parameters for which the ICER value is <$100000 per QALY gained. ADR, adverse drug reaction; ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life year.

Full figure and legend (47K)Download PowerPoint slide (184 KB)

Top

Discussion

Cost-effectiveness analysis must be clearly differentiated from risk-benefit analysis to emphasize the value of cost-effective medical decision making. While data is available comparing the potentially undesirable outcomes and side effects of the traditional approach for selecting and prescribing medications to the benefits of routine use of pharmacogenomics testing (risk-benefit analysis), little data exists on cost-effectiveness of genetic testing for drug delivery. Hence, our analysis is an effort to understand how the health outcome benefits from pharmacogenomics testing are comparatively greater than the cost of genetic testing. Data currently exist on the benefits of personalized medicine for drug delivery. There are over 120 drugs that mention or recommend genetic testing in the package insert. Many of these are Food and Drug Administration (FDA) black box warnings recommending routine genetic testing to avoid side effects or to reduce the risk of treatment failure. We believe that if pharmacogenomic testing had data showing its routine use to measurably reduce costs, it would already have been used as a routine tool in clinical practice. It is very likely this lack of financial data supporting its cost-effectiveness has delayed the adoption of pharmacogenomics testing throughout the wider medical community.

The study by Bond and Raehl19 reported that the annual cost of adverse drug events for the Centers for Medicaid and Medicare Services (CMS) is ~$516 million. However, CMS officials will not formally consider cost-effectiveness when determining coverage to avoid the accusation that CMS will ration care for older Americans.20 Few studies exist on the cost savings of pharmacogenomics testing. Our analysis of the literature revealed that the sample size of most studies was often too small to generate this valuable data. For example, it was reported that treating psychiatric patients who were categorized as ultrarapid or poor metabolizers of CYP2D6 costs an average of $4000 to $6000 more per year than treating patients who were extensive metabolizers when these patients were prescribed drugs that are metabolized by the CYP2D6 enzyme. This translated to an estimated $112000 to $168000 per year in added healthcare expenses at this psychiatric hospital directly related to this single gene polymorphism.21 The ideal study design would be a randomized study comparing patients who receive standard therapy (that is, traditional ‘trial and error’ selection of mediation) to patients where the treatment selection is guided by pharmacogenomic testing. Randomized studies, however, with large patient groups may not even be possible in the current clinical environment largely controlled by managed care companies due to the time and cost required to run these mega trials. Thus, there is a need to improvise using clinical decision modeling, as we have done, until large randomized clinical trials can be designed and completed.

Our Markov analysis implies that if we genetically test a group of 40-year olds and follow them until death, the ICER of genetic testing verses no testing would be $53680 per additional QALY. In cost-effectiveness studies, any new intervention/program that costs <$100000 per additional QALY is typically assumed to be cost-effective.22 The smaller the ICER, the more cost-effective genetic testing is. Therefore, under the base-case scenario, genetic testing is cost-effective. It is important to note that as the ICER increases, it implies that genetic testing is less cost-effective. Interestingly, the ICER increases with age. This may be due to the fact that more ADRs could be prevented over a lifetime with earlier testing, making genetic testing more cost effective if performed at an earlier age. It is possible that with continued accumulation of data on the benefits of pharmacogenomics, this technology may be justified in pediatric or even neonatal patients.

The major shortcomings of our approach are related to the estimation of the parameters used for our analysis. In particular, there is no consensus in the literature regarding the values of three important parameters of the model (mean ADR rate, rate of reduction in ADR rate with genetic testing and probability of death due to an ADR). For example, we found that using our estimates, there would be a total of 105153 outpatient ADR-related deaths in 2014, whereas this estimate appears to be significantly higher than the estimate reported by the only major study estimating the ADR-related deaths in 1998.3 On the other hand, considering that ADR rates increased dramatically since 1994, our estimate appears to be acceptable. A study by FDA reported that there were 98518 deaths in 2011 due to ADRs.23 While our estimate appears to be higher than their estimate, it was reported that between just 2010 and 2011 ADR-related deaths have increased by almost 25%. This suggests underreporting of ADR-related deaths in the FDA report, as they depend on voluntary reporting of ADRs. Those critical of cost-effectiveness models have often cited the hidden biases of authors and study sponsors, resulting in discretionary cost-model building and cherry picking during data selection in these analyses.24 To overcome this limitations posed by such conflicting information in the literature, we conducted extensive sensitivity analyses and reported the cost-effectiveness results for a range of values for these parameters.

In summary, we developed a Markov model to evaluate the cost-effectiveness of one-time genetic testing for asymptomatic patients. We found that genetic testing is cost-effective under most scenarios. Further prospective randomized studies are necessary for conclusive evidence on cost-effectiveness ratio. In the end, pharmacogenomics is not simply to be viewed as a cost-containment tool but, rather a method to bring additional value to effective medication prescription. It may or may not save money, but would certainly represent good value for the dollars spent.

Top

Conflict of interest

The authors declare no conflict of interest.

Top

References

  1. Classen DC, Pestotnik SL, Evans RS, Lloyd JF, Burke JP. Adverse drug events in hospitalized patientsExcess length of stay, extra costs, and attributable mortality. JAMA 1997; 277: 301–306. | Article | PubMed | ISI | CAS |
  2. De Smet PA, Denneboom W, Kramers C, Grol R. A composite screening tool for medication reviews of outpatients. Drugs Aging 2007; 24: 733–760. | Article | PubMed |
  3. Lazarou J, Pomeranz BH, Corey PN. Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies. JAMA 1998; 279: 1200–1205. | Article | PubMed | ISI | CAS |
  4. Drummond MF, O'Brien BJ, Stoddart GL, Torrance W. Methods for the Economic Evaluation of Health Care Programmes. Oxford University Press: New York, NY, USA, 1998.
  5. Gold MR, Siegel J, Russell L, Weinstein M. Cost-effectiveness in Health and Medicine: Report of the Panel on Cost-effectiveness in Health and Medicine. Oxford University Press: New York, NY, USA, 1996.
  6. Heitjan D, Asch D, Ray R, Rukstalis M, Patterson F, Lerman C. Cost-effectiveness of pharmacogenetic testing to tailor smoking-cessation treatment. Pharmacogenomics J 2008; 8: 391–399. | Article | PubMed |
  7. Perlis RH, Patrick A, Smoller JW, Wang PS. When is pharmacogenetic testing for antidepressant response ready for the clinic? a cost-effectiveness analysis based on data from the STAR* D Study. Neuropsychopharmacology 2009; 34: 2227–2236. | Article | PubMed | ISI |
  8. Reese ES, Daniel Mullins C, Beitelshees AL, Onukwugha E. Cost-effectiveness of cytochrome p450 2c19 genotype screening for selection of antiplatelet therapy with clopidogrel or prasugrel. Pharmacotherapy 2012; 32: 323–332. | Article | PubMed |
  9. Alagoz O, Hsu H, Schaefer AJ, Roberts MS. Markov decision processes: a tool for sequential decision making under uncertainty. Med Decis Making 2009; 30: 474–483. | Article | PubMed |
  10. Bourgeois FT, Shannon MW, Valim C, Mandl KD. Adverse drug events in the outpatient setting: an 11-year national analysis. Pharmacoepidemiol Drug Saf 2010; 19: 901–910. | Article | PubMed |
  11. Budnitz DS, Pollock DA, Weidenbach KN, Mendelsohn AB, Schroeder TJ, Annest JL. National surveillance of emergency department visits for outpatient adverse drug events. JAMA 2006; 296: 1858–1866. | Article | PubMed | ISI | CAS |
  12. Budnitz DS, Lovegrove MC, Shehab N, Richards CL. Emergency hospitalizations for adverse drug events in older Americans. N Engl J Med 2011; 365: 2002–2012. | Article | PubMed | ISI | CAS |
  13. Epstein RS, Moyer TP, Aubert RE, Kane DJ, Xia F, Verbrugge RR et al. Warfarin genotyping reduces hospitalization rates results from the MM-WES (Medco-Mayo Warfarin Effectiveness Study). J Am Coll Cardiol 2010; 55: 2804–2812. | Article | PubMed | ISI | CAS |
  14. Kimmel SE, French B, Kasner SE, Johnson JA, Anderson JL, Gage BF et al. A pharmacogenetic versus a clinical algorithm for warfarin dosing. New England Journal of Medicine 2013; 369: 2283–2293. | Article | PubMed |
  15. Damani SB, Topol EJ. The case for routine genotyping in dual-antiplatelet therapy. J Am Coll Cardiol 2010; 56: 109–111. | Article | PubMed | ISI |
  16. Cause of Death 1999-2010 on CDC WONDER Online Database, released 2012. Data are from the Multiple Cause of Death Files, 1999-2010, as compiled from data provided by the 57 vital statistics jurisdictions through the Vital Statistics Cooperative Program. Available at http://wonder.cdc.gov/ucd-icd10.html. 2014.
  17. Chyka PA. How many deaths occur annually from adverse drug reactions in the United States? Am J Med 2000; 109: 122–130. | Article | PubMed | CAS |
  18. Sarkar U, López A, Maselli JH, Gonzales R. Adverse drug events in US adult ambulatory medical care. Health Serv Res 2011; 46: 1517–1533. | Article | PubMed |
  19. Bond C, Raehl CL. Adverse drug reactions in United States hospitals. Pharmacotherapy 2006; 26: 601–608. | Article | PubMed |
  20. Neumann PJ. Why don’t Americans use cost-effectiveness analysis. Am J Manag Care 2004; 10: 308–312. | PubMed |
  21. Chou WH, Yan F-X, de Leon J, Barnhill J, Rogers T, Cronin M et al. Extension of a pilot study: impact from the cytochrome P450 2D6 polymorphism on outcome and costs associated with severe mental illness. J Clinical Psychopharmacol 2000; 20: 246–251. | Article |
  22. World Health Organization (2014) Table: Threshold values for intervention cost-effectiveness by Region. World Health Organization. Available at http://www.who.int/choice/costs/CER_levels/en/index.html. Accessed 7 November 2013.
  23. Food and Drug Administration (2014). FDA Adverse Events Reporting System (FAERS) Patient Outcomes by Year 2014. Available at http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/Surveillance/AdverseDrugEffects/ucm070461.htm.
  24. Rennie D, Luft HS. Pharmacoeconomic analyses: making them transparent, making them credible. JAMA 2000; 283: 2158–2160. | Article | PubMed | ISI | CAS |
  25. Wu C, Bell CM, Wodchis WP. Incidence and economic burden of adverse drug reactions among elderly patients in Ontario emergency departments. Drug Safety 2012; 35: 769–781. | Article | PubMed |
  26. Pfuntner A, Wier LM, Steiner C. Statistical brief #146 by Agency for Health Care Research and Quality: Costs for Hospital Stays in the United States, 2010. Agency for Health Care Policy and Research, Rockville, MD, USA, 2013. Available at: http://www.hcup-us.ahrq.gov/reports/statbriefs/sb124.pdf.
  27. Shaw JW, Johnson JA, Coons SJ. US valuation of the EQ-5D health states: development and testing of the D1 valuation model. Med Care 2005; 43: 203–220. | Article | PubMed | ISI |
  28. Centers for Disease Control Prevention (2013). CDC/NCHS National Hospital Discharge Survey, 2010. Number and rate of discharges from short-stay hospitals and days of care, with average length of stay and standard error, by selected first-listed diagnostic categories: United States, 2010. Available at http://www.cdc.gov/nchs/data/nhsr/nhsr029.pdf.
  29. Agency for Healthcare Research and Quality (2011). Emergency Room Services-Mean and Median Expenses per Person With Expense and Distribution of Expenses by Source of Payment: United States. Medical Expenditure Panel Survey Household Component Data 2011. Available at: http://www.meps.ahrq.gov/mepsweb/surveycomp/household.jsp.
Top

Acknowledgements

This research is funded by Renaissance Rx.