We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.HC

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Computer Science > Human-Computer Interaction

Title: Design Principles and Clinician Preferences for Pharmacogenomic Clinical Decision Support Alerts

Abstract: OBJECTIVE: To better understand clinician needs and preferences for the display of pharmacogenomic (PGx) information in clinical decision support (CDS) tools.
MATERIALS AND METHODS: We developed a semi-structured interview to collect feedback and preferences in six key areas of PGx CDS design, from clinicians who had prior experience with live PGx CDS tools. Eight clinicians from Northwestern Medicine's (NM) General Internal Medicine clinic participated in the study.
RESULTS: Clinicians expressed preference for interruptive pop-up alerts during order entry, brief descriptions of relevant drug-gene interactions, and a clear and specific recommended alternative course of action when a medication is contraindicated. They did not wish to see detailed genetic data, preferring phenotypic information predicted from the genotype. Nor did they wish to be interrupted when genetic test results do not indicate a change in treatment plan. Clinicians reported little familiarity with Clinical Pharmacogenetic Implementation Consortium prescribing recommendations but reported trusting recommendations of their professional societies and resources like UpToDate. Analysis of unstructured comments concurred with structured results, indicating a general uncertainty among participants around how to interpret and apply PGx information in practice.
DISCUSSION: Results point to several underlying principles that can inform future PGx CDS alert designs: Be Specific and Actionable; Be Brief; Display Phenotypes not Genotypes; Rely on Sources Clinicians Already Trust; and, Be Adaptable to Learning Effects.
CONCLUSION: This study is part of a broader socio-technical design approach to PGx CDS design underway at NM and provides a baseline for future PGx CDS development. Designs based on these results have the potential to improve clinician education and adherence levels, and to improve patient outcomes.
Subjects: Human-Computer Interaction (cs.HC); Genomics (q-bio.GN)
Cite as: arXiv:2002.00044 [cs.HC]
  (or arXiv:2002.00044v1 [cs.HC] for this version)

Submission history

From: Timothy Herr PhD [view email]
[v1] Fri, 31 Jan 2020 20:18:13 GMT (422kb)

Link back to: arXiv, form interface, contact.