Development and use of active clinical decision support for preemptive pharmacogenomics

From Clinfowiki
Jump to: navigation, search

Prescribing medication is changing and with the help of CDS and EHR there will be more and more "gene-based drug prescribing". [1]

First Review

Background

Many institutions have CDS as part of their EHR; however, few institutions perform pharmacogenetics testing and use this information to prescribe medication. This institution is looking at patients who have been genetically tested and have their results stored in the EHR. Alerts are generated based on the medication prescribed and whether the medication is a high-risk medication for the patient. If considered a high-risk medication, the CDS alerts the physician and prompts the physician to consider prescribing a medication with fewer risks or one that is potentially more efficacious. The CDS may also recommend decreasing or increasing the dose based on the patient's ability to metabolize the drug.

Methods

Using the St. Judes PG4KDS protocol, samples were collected and approximately 225 results were placed in the EHR in order to make recommendations based on the gene-drug pair. Recommendations are based on CPIC guidelines. Once determined they are then placed into the EHR system as such pairs. The CDS is activated when a physician orders a medication that could potentially be high risk based on the pharmocogenetic test results.[1] This warning can serve different purposes: it can suggest a different medication type altogether or perhaps suggest a change in the standard dose.

Results

During an 18month period of information collection and new alerts implemented it was ultimately found that the alerts were not being ignored and the correct medication and dosage was being prescribed. It was found that approximately "95% patients who had a post-test alert at the time of the first prescription received the appropriate change in therapy as guided by the on-screen alert."[1]

Conclusion

Implementing a CDS system that delivers real time alerts based on the pharmocgenetic testing results in a higher incidence of appropriate mediation administration. This has helped physicians to be proactive when prescribing medication from dosing to perhaps even changing the course of therapy.

Comments

This is new information for me and this is exciting in the way of how medicine is continuously evolving. It is a privilege to see how medicine and modern day technology making advancements in patient therapy and treatment. As always, it makes me wonder where will we be in the next 5-10 years!


Second review

Introduction

This paper describes the development, implementation and evaluation of preemptive gene-based prescribing information into a clinical decision support system CDS using a commercially available electronic health record EHR. This study was performed to assess the feasibility of implementing a pharmacogenetic clinical decision support system at the point of care.

Background

Pharmacogenetics uses genotype and phenotype data to predict optimal patient-specific pharmaceutical therapy based on the cytochrome P450 metabolic enzymes. Implementing such a system is extremely relevant since more than half of patients seen in the primary care setting are prescribed medications affected by the cytochrome P450 system. Pharmacogenetic test results are unique in that they are static over a patient’s lifetime. As such, this data should be retained indefinitely within a patient’s record for reference throughout the patient’s life.

Methods

The study was performed at St. Jude Children's Research Hospital. The protocol included genetic testing for 225 genes. All phenotypes with clear drug-gene interaction information were entered into the EHR. Ideally this information is collected and stored in a patient’s EHR before high-risk medications are prescribed. More phenotype-drug correlations are added as clinical recommendations become available. Each phenotype is entered as either routine or high risk, where high risk is defined as a required change of therapy including a different drug or a different dose. The CDS rules were written to correlate specific CYP450 genotypes to phenotypic diagnoses and subsequently trigger messages or interruptive alerts based on medications prescribed. The system established at St. Jude’s included pre-test advice and specific post-test warnings and recommendations. Pre-test (pre-genetic testing) advice was provided when pharmacogenetic phenotypes were not in the patient’s record but a high-risk medication was being prescribed. The message included a recommendation to order pharmacogenetic studies before prescribing the high-risk medication. Post-test (post-genetic testing) warnings for high-risk drugs with specific recommendations based on the patient’s phenotype interrupted the medication order and required an override by both the physician and the pharmacist. The alerts, warnings, and other messages were crafted by the pharmacogenetics team and given final approval by the Pharmacogenetics Oversight Committee. High-risk phenotypes are listed as discrete entries in the problem list and these entries serve as the trigger for the post-test alerts. St. Jude’s developed a customized nomenclature for the phenotypes because SNOMED and ICD lacked codes. The system maintained logs of all alerts and how the alert was handled. Charts were also analyzed to see if dose modification recommendations were followed.

Results

To prevent alert fatigue, only the high risk phenotype-drug pairings were set to interruptive alerts; others were sent via email for physician edification. CDS logic was written carefully to cover very specific scenarios to assure the relevance of each alert. The committee considered limiting the number of alerts to one per patient per specified time period, but the physicians found the alerts to be highly effective and valuable and opted not to apply limits to the number of alerts. When the 18 months of alert data were combined for two of the high-risk drugs, 95% of the alerts were followed as recommended.

Discussion

A multidisciplinary oversight committee consisting of physicians, pharmacists, geneticist, pathologists, and clinical informaticians was established to plan and oversee the pharmacogenetic alerts. This committee reported to the Pharmacy and Therapeutics Committee which reports to the Medical Executive Committee. New phenotype-gene pairs are added to the EHR based on recommendations from CPIC and further exploration by St. Jude’s oversight committees. This study demonstrated a successful rollout of a pharmacogenetic alert system using core components of a commercially available EHR, and therefore shows potential to be transferred to other facilities. The system was also shown to be extensible, as additional phenotype-drug pairs were added as knowledge became available.

Conclusion

This study demonstrated the feasibility of implementing a pharmacogenetic clinical decision support system to avoid prescribing or to modify doses of prescribed medications that are contraindicated based on the patient’s phenotypic diagnoses.

Comments

This study showed that following best practices for the design and implementation of CDS can be effective and well-received. Best practices included the use of a multispecialty oversight committee, rigorous attention to programming, deliberate policies to reduce alert fatigue, and the use of both active (interruptive) alerts for high risk drugs and passive information available for lower risk drugs.

References

  1. 1.0 1.1 1.2 Bell GC, Crews KR, Wilkinson MR, Haidar CE, Hicks JK, Baker DK, Kornegary NM, Wenjian Y, Cross SJ, Howard SC, Freimuth RR, Evans WE, Broeckel U, Relling MV, Hoffman JM, 26 August 2013, Development and use of active clinical decision support for preemptive pharmacogenomics, Journal of the American Medical Informatics Association, 2014,21, 93-99. http://jamia.oxfordjournals.org/content/21/e1/e93