Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
Safe Opioid Prescription: A SMART on FHIR Approach to
Clinical Decision Support
Shyamashree Sinha1,2, Mark Jensen1,3, Sarah Mullin1, Peter L Elkin1
1. Department of Biomedical Informatics, University at Buffalo State University of New York
2. Department of Anesthesiology, University at Buffalo State University of New York
3. Department of Philosophy, University at Buffalo State University of New York
Abstract
Background
Prescription opioid pain medication overuse, misuse and abuse have been a significant contributing factor
in the opioid epidemic. The rising death rates from opioid overdose have caused healthcare practitioners
and researchers to work on optimizing pain therapy and limiting the prescriptions for pain medications.
The state of New York has implemented a prescription drug monitoring program(PDMP), amended public
health law to limit the prescription of opioids for acute pain and utilized the resources of the state and
county health departments to help in curbing this epidemic. The recent publication of guidelines for
prescription opioids from CDC [1] and ASIPP (American Society of Interventional pain practitioners) have
independently reviewed literature and found good evidence of limiting opioid prescription for acute and
chronic non cancer pain. [2]
Method
Over the last decade, advanced technology has increased the complexity of electronic health records
systems leading to the development of Clinical Decision Support Systems (CDSS) to aid the work flow of
healthcare providers. There are several systematic reviews on the effectiveness and utility of CDSSs. A
common consensus is that commercially and locally developed CDSS are effective in improving patient
measures while actual workload improvement and efficient cost-cutting measure are not significantly
improved by CDSS. Patient provider involvement in developing CDSS is a determinant of its success and
utilization rates. [7] Therefore, a plug and play form of CDSS which can be implemented from an external
platform through secure channels would be more effective.
Design
The Health Level Seven’s (HL7) open licensed interoperability standard Fast Health Interoperability
Resources (FHIR) has a platform, Substitutable Medical Applications and Reusable Technologies (SMART)
for CDSS app development by a third party. [3] We adopted these open source standard to plan to develop
an app for accessible and efficient implementation of the recently published guidelines for management
of pain with prescription opioid medications.
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
Introduction
The mortality rate due to Opioid overdose has been consistently rising over the last decade,
reaching epidemic proportions. New York State, including Western New York, has had a steady
rise in opioid related deaths from 2014 to 2015: natural and semisynthetic opioid overdose death
rose by 13.3%, methadone related deaths rose by 9.1%, and synthetic opioid other than methadone
related deaths rose by 135.7% [1]. There is evidence that overuse, misuse. and abuse of
prescription opioid pain medication is one of the major contributing factors in the rising death toll.
This reasoning led to the publication of the recent CDC guidelines for prescription [2].
The population health problem due to prescription opioid use, abuse and dependence has been
progressively worsening over the last two decades. Clinicians are faced with the challenge of
treating pain adequately to improve quality of life while trying to prevent the potential of overuse,
abuse and dependence among patients who are being treated by prescription opioid medications.
Many studies have shown narcotic pain medications are known to decrease pain threshold and
increase the need for pain medications [3].
The need for clinical decision support for pain control and monitoring of patients’ medication has
led to many attempts at developing tools based on established guidelines. The ASSIP recently
published a two-part guideline on pain control which gives a stepwise approach to pain medication
management and adequate pain control. This guideline states opioid pain medication therapy
should only be limited to patients with intractable chronic pain who have shown demonstrable
improvement with therapy [4].
Recently, the Centers for Disease Control and Prevention published a guideline for chronic opioid
prescription based on a systematic review of current literature in 2016 [2]. These practice
guidelines provide clinicians with a useful tool for making decisions of optimizing the use of
prescription pain medications. However, the attempts at implementation of these guidelines in the
form of applications have had limited outcomes that have been discussed later.
Aim
The goals for this CDSS tool are to achieve proper monitoring of prescription drugs, patients’ medication
list and potential interactive medications, surveillance for abuse/ misuse, patient involvement in
alternative therapy, reporting problems and obtaining adequate pain control. The final step would be to
implement the system in clinical practice and to monitor usage rates and measure the outcomes of
successful or unsuccessful implementations.
DOI: 10.5210/ojphi.v9i2.8034
Copyright ©2017 the author(s)
This is an Open Access article. Authors own copyright of their articles appearing in the Online Journal of Public Health Informatics.
Readers may copy articles without permission of the copyright owner(s), as long as the author and OJPHI are acknowledged in the copy
and the copy is used for educational, not-for-profit purposes.
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
Clinical Decision Support Systems (CDSS) based on Electronic Health Records have been
evolving for more than a decade. A paper published by the federal health IT unit outlines the need,
importance and available technology behind the evolving CDSS tools [5].
Based on their recommendations a stepwise process of designing a CDSS would be to:
1. Identify and quantify a ‘high-priority gap’ in clinical quality between current
outcomes and a stated goal. ‘High priority’ is the need for quality of care
improvement aligns with the strategic goals and treatment outcomes that the health
care organization is aiming for.
2. Map the workflow by which the clinical care pertaining to the target issue is
delivered.
3. Design a future state workflow that includes the best ideas for closing the gap and
eliminates as much waste as possible.
4. Identify the information necessary to support the future state workflow. Then,
design, test, and perfect CDS interventions using the Four Rights- Right Info, Right
Person, Right Medium and Right Format.
Our goal is the creation of an application (app) based decision support tool that is high priority in
terms of population health needs, minimizes the consumption of time for the provider, and is more
interactive for the patient. In addition, this app could be utilized as an aide in implementation of
the goals of precision medicine, focusing on patient-centered care.
Therefore, in line with these goals, we propose the creation of the application OPIacutepain that
follows the opioid prescribing guidelines set forth by the CDC to aide in the efficient and accurate
prescription of opioid medication for acute pain.
1. Background
1.1 Justification for implementation of a CDSS for prescription of Opioid pain
medication
The state of New York has recently implemented an amendment of the NY public health law
(NYPHL 3331) limiting the number of days of prescription for opioid pain medication to not more
than seven days for any acute non-cancer/non-terminal illness pain [6]. Due to the rising death
rates from opioid overdose, the CDC's prescription opioid pain medication guidelines are aimed at
curbing opioid dependence, abuse and misuse of pain medications. The patients on opioid pain
medications for chronic pain are required to be monitored by Prescription Drug Monitoring
Programs (PDMPs) and regular urine screening. Applying the proper CDSS implementation to
prescription opioid pain medications, the five steps are as follows [2]:
· High Priority gap in the clinicians’ prescription of pain medication for acute pain
conditions. The amendment of state law and CDC guidelines all aim at cutting
down on that acute pain prescriptions.
· The work flow of prescription of pain medication starts as follows-
➢ with the patient being aware of the other options for pain control
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
➢ then, deciding upon a treatment goal and working towards fulfilment
➢ followed by an evidence or guideline based approach to prescribing pain
medications
· Design a CDSS based on the latest technology. In order to be effective, the CDSS for
opioids should be interactive and less invasive. Which means it can be used by the
providers on demand and would be able to pull the right information for the patient
in order to aid the clinician with decision making
· The SMART on FIHR app would be written based on the current guidelines from
CDC and ASIPP.
1.2 Review of Current Opioid-related Apps and CDS systems
The number of attempts at developing CDSS has resulted in an aversion to yet another system.
The clinicians and healthcare providers look at it as more of a deterrent in their work flow than
any help [7] [5]. However, different technological approaches may be able to help in finding an
answer to the complexity of electronic health records. The need for development of newer CDSS
were initially based on national stimuli for EHR adoption, Meaningful Use requirements, Medicare
Access and CHIP Re Authorization Act(MACRA) and Merit-based Incentive Payment
System(MIPS). The rise of precision medicine, genomics and dynamic changes in the approach to
medical care increased the complexity of care needed more emphasis on coordination of care. The
increasing cost of medical care, patient empowerment/awareness and technological push for App
culture are also drivers for development of newer CDSS [7] [5].
There are many opioid conversion calculators and morphine equivalency calculators on the market.
However, each one has issues as a point of care tool, whether a hard to use interface, requiring the
internet to download guidelines, having no citations for evidence based assessment, or requiring
many plug in values that could be automatically taken from the electronic health record (EHR) or
electronic medical record (EMR). Especially when it comes to conversion, many of the apps are
not reliable. A recent study assessed the output variability and professional medical involvement
in the authorship of 23 different apps [8]. 52% of the apps had no stated medical professional
involvement and 48% of the apps provided references for opioid conversion ratios [8].
The CDC recently released on app summarizing the 2016 guidelines. CDC Opioid Prescribing
Guideline 2016 contains a morphine equivalency calculator, brief synopses of recommendations
from the guidelines, glossary and a section on how to perform motivational interviewing with pain
patients. The app includes links to full sections of the CDC guideline and references, but requires
an internet connection. The MME calculator, which allows for multiple drugs to be calculated for
a total MME/day, is built in and does not require internet connection and provides an alert when
MME/day≥90. The alert suggests referring to a specialist and scheduling reassessment at least
every 3 months, as well as suggesting recommendations through proper links in the app [9].
The pH Medical Opioid Converter App by Philip Eagan from the iTunes app store, use 2016 CDC
guideline for Prescribing Opioids for Chronic Non-Cancerous Pain patients. The app includes an
easy to use morphine equivalence calculator and an opioid to opioid calculator, using generic or
trade names. Although the app has a simple to use interface and is based on current
recommendations, it is the only one that comes at a cost ($1.99), it does not have an Android
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
counterpart, and very little information can be found about the author. In addition, like the CDC
app, the guidelines and links are not built into the app and therefore, an internet connection is
required [10].
The Safe Opioids app was developed with support from the Substance Abuse and Mental Health
Services Administration through the Prescribers' Clinical Support System for Opioid Therapies. It
includes four categories: Evaluation, Manage, Discontinue, and other online Resources. Under
Evaluation, there is information on the assessment of pain, linking to an outside the app pdf, the
Opioid Risk Tool, and links to drug prescription monitoring programs. Under manage, there is
information on opioid management including links to medical guidelines on clinical use in
treatment of chronic pain (e.g.: VA/DoD Clinical Practice Guidelines) from 2010. There is also a
list of common side effects and advise for talking to patients about opioid abuse. There are also
links to a tool for assessing depression. The app provides citations, showing how they are using
evidence based medicine for decisions, however, the app is outdated and the citations have not
been updated since 2010 [11].
OpioidCalc NYC, developed by The New York City Department of Health and Mental Hygiene
(DOHMH), calculates MME based on the guidelines developed with the help of the CDC that
outline key principles of safe and judicious prescribing practices, including high dosage defined
as MME/day ≥100. The app provides citations for evidence-based practice decisions. The app
allows users to quickly and easily calculate the total daily MME a patient is taking, based on type
of opioid analgesic, strength, and quantity. Multiple types of opioid analgesics can be included in
the total daily MME calculation. An alert is displayed for total daily MME greater than or equal
to 100, indicating an increased risk for overdose. This alert also suggests reassessing the patient’s
pain status and treatment plan and provides a link to the DOHMH opioid-prescribing guidelines
for additional information. It does not include I.V. formulations of medications and does not align
with current CDC guidelines [12].
Opioid pain medication documentation and surveillance system (PODS) from electronic health
records is one of the early attempts at medication management and patient education efforts. Like
many other apps included in the electronic health records it was time intensive and cumbersome
due to the fact that it was not interactive with the rest of the ATHENA EHR-Opioid Therapy for
chronic non-cancer pain pop-up [13]. There were many good features in this system like patient
safety features and some early warnings. The pop up in the EHR had a patient identifier on graphic
user interphase, caution window where the patients risk factors against prescribing pain medication
were mentioned, the treatment options, data tables, treatment checklist and information for
researchers would appear in separate sections of the window. The dosage calculator and other
options were present in a dropdown section.
The usability testing of this CDSS showed the lack of provider education, confusion in dosing
calculations and titration schedules, access to relevant patient information, provider discontinuity,
documentation, and access to validated assessment tools. Clinicians reported the time constraints
in completing each prescribing decision and effective pain management based on guidelines. The
figure below shows the pop up window.
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
Figure 1. GUI Source: https://www.ncbi.nlm.nih.gov/books/NBK43756/figure/advances-
michel_92.f1/?report=objectonly
[Int J Med Inform. 2015 Apr;84(4):248-62. doi: 10.1016/j.ijmedinf.2015.01.004.
Epub 2015 Jan 17.]
2. Proposed Solution
The SMART on FHIR platform for interactive and integrated CDSS development is an open
source resource which will be used to implement CDC guidelines with an aim to cut down the rate
of opioid prescription for acute pain. The third party app development process aids the wide scale
utilization of the app in FHIR compatible EHR environment. This is a process of using cell phone
technology in building a more dynamic (can be used by healthcare providers and patients through
different portals) CDSS tool that is user friendly and does not have to be built in as a part of the
EHR system.
The SMART app will be independently placed in an APP library from where it can be accessed
by the clinician while making the decision for treatment of pain.
Figure 2. explains the flow of the platform.
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
Figure 2. Flow diagram for OPIacutepain a SMART on FHIR approach for prescription opioids
for acute pain
The CDC guidelines published in 2016 [2] were aimed at providing a decision making tool for
clinicians, nurses and other healthcare providers to aid in the treatment of chronic and acute pain.
The following chart (Figure 3) shows the different categories of pain the patients may present with
and how the clinician may make their decision.
Acute pain (first) Acute pain (prior
history)
Acute on chronic pain Chronic pain on opioid
Inform physician of
treatment goals for
pain(Recommendation
2)
Inform physician of
treatment goals for
pain(Recommendation
2)
Inform physician of
treatment goals for
pain(Recommendation
2)
Inform physician of
treatment goals for
pain(Recommendation
2)
Non pharmacologic
(rec 1)
Assess Baseline pain
and function- PEG
assessment scale (see
Appendix A)
Assess Baseline pain
and function- PEG
assessment scale (see
Appendix A)
Assess Baseline pain
and function- PEG
assessment scale (see
Appendix A)
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
Alerts: history of
mental illness, Alcohol
or other substance
abuse, smoking
Alerts: history of
mental illness, Alcohol
or other substance
abuse, smoking
Alerts: history of
mental illness, Alcohol
or other substance
abuse, smoking
Alerts: history of
mental illness, Alcohol
or other substance
abuse, smoking
Figure 3. Categories of Pain and Recommendations from CDC guidelines
The characteristics of the SMART Application that we propose are as follows:
The Application is aimed at providing interactive EHR based guidelines for acute pain patients
where non pharmacologic, non-opioid pain medications are tried first followed by a short term
opioid pain medication therapy.
The potential for fatal overdose may be there if the patient is on Benzodiazepines while being on
opioid pain medication. Concurrent benzodiazepine usage may lead to fatalities in 31%–61% of
the total overdose deaths [1]. This could be prevented by the interactive application.
In addition, the initial daily dose of opioids should not exceed 50 MME/day and could go up to a
maximum of 90MME per day. For dosages from 50 to 100 MME/day, risk for overdose increased
by a factor of 1.9 to 4.6 compared to dosages of 1 to 20 MME/day. Greater than 100 MME/day
increased by a factor of 2 to 8.9 compared to 1 to 20 MME/day [1]. Risk for serious harm on higher
dosage of opioid outweighs the treatment benefits.
The app will include a medication list of all the opioids a patient is currently using, and if there are
opioids present, the conversion chart to calculate total MME/day produced by the CDC will be
used to automatically calculate the total for the clinician (Figure 4).
Figure 4. CDC Calculating MME Conversion Factor chart
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
Finally, the app will check if a urine drug screen (UDS) has been ordered. The CDC guidelines
state, that a UDS provides the ability to “identify patients who might be at higher risk for opioid
overdose or opioid use disorder, and help determine which patients will benefit from greater
caution and increased monitoring or interventions when risk factors are present [1]. Depending on
the outcome of medications and observations, recommendations will be provided that include
counseling, doing the PEG-assessment, ordering a UDS, not prescribing opioid medications,
tapering of opioid medications, and checking the state prescription drug monitoring program
(PDMP) data. An outline of the workflow of the proposed app is included in Figure 5.
3. Methods
3.1 FHIR
FHIR® (Fast Healthcare Interoperability Resources) is a next generation standards framework that
builds on Health Level Seven International (HL7), an ANSI-accredited standard developing
organization which provides standards for the exchange, integration, sharing, and retrieval of
electronic health information supporting clinical practice and management. According to the FHIR
standards, it “leverages existing logical and theoretical models to provide a consistent, easy to
implement, and rigorous mechanism for exchanging data between healthcare applications”. [14]
In order to assure alignment with the current HL7 standards and interoperability, FHIR has built-
in mechanisms for traceability to the HL7 RIM [14].
The basic building blocks of FHIR are called resources which share a set of characteristics,
including a definition, a common set of metadata, and a human readable part [14-16]. Table 1
provides the list of resources and their elements that are of interest to the application. Following
Meaningful Use Stage 2 criteria, resources rely on ontologies and terminologies, specifically
SNOMED-CT, LOINC, RxNorm [14]. Therefore, our CDSS is ontology and terminology driven.
Currently, FHIR version 1.0.2 is supported by Epic, Cerner, and Allscripts Professional, three
prominent electronic health record (EHR) systems [7]. For instance, according to the Open Epic
website, Epic’s integration works with FHIR 1.0.2 (DSTU2). With this specification, Epic supports
retrieving data for most of the top-level resources such as “patient”, “observation”, “procedure”,
“medication.” Since these are primarily the resources needed for our proposed app, this app could
seemingly be used any system equipped with Epic, Cerner, or Allscripts professional and can
easily be accessed through their web services and other methods, such as relational databases.
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
Table 1. FHIR resources used for OPIacutepain
Resource Type: Patient
Definition:
The Patient Resource provides general demographic information about a
person receiving health care services from a specific organization.
Elements Definition Use in App
Patient.name.given First name
Used to Make Sure you are querying correct
patient
Patient.name.family Last name
Used to Make Sure you are querying correct
patient
Patient.birthdate birthdate
Used to Make Sure you are querying correct
patient
Patient.gender gender
Used to Make Sure you are querying correct
patient
Resource Type: MedicationOrder
Definition:
An order for both supply of the medication and the instructions for
administration of the medication to a patient.
Elements Definition Use in App
MedicationOrder.Status
Lifecycle of prescription: want
active not draft
MedicationOrder.dateEnded
Date when prescription was
stopped Used to filter in only active/current prescriptions
MedicationOrder.patient
A link to a resource representing
the person to whom the
medication will be given. Medication for specific patient
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
MedicationOrder.medicationCodeableConcept
Identifies the medication being
administered. Uses RxNorm codes to identify medications
MedicationOrder.reasonCodeableConcept Reason for writing prescription
Can check for opioids used for chronic pain
(SNOMED-CT code)
MedicationOrder.dosageInstruction.dose[x] Gives dose and units (ie 5 ml) Used for MME calculator
MedicationOrder.dosageInstruction.timing
Gives how many times daily
(timing can be coded by
http://hl7.org/fhir/timing-
abbreviation) Used for MME calculator
Resource Type: Diagnostic Report
Definition:
Findings and interpretations of diagnostic tests performed on patients. The
report includes clinical context such as requesting and provider information
Elements Definition Use in App
DiagnosticReport.code
Codes that describe Diagnostic
reports
Urinary Drug Screen (using LOINC code)
ordered
DiagnosticReport.result References Observation
Resource Type: Observation
Definition:
Simple name/value pair assertions for laboratory data and other results such as vital
signs, imaging results, and social history.
Elements Definition Use in App
Observation. Patient
A link to a resource representing
the patient to whose lab values
they are.
Observation. Code
Codes identifying names of
simple observations; Use
LOINC or SNOMED codes Use LOINC codes for UDS results
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
Observation. Date
Date range into which the
observation falls
Recent Urinary Drug Screen or screen from
history
Observation. Encounter Identifies the medication being
administered. Uses RxNorm codes to identify medications
Observation. Value[x]
Information determined as a
result of making the
observation .value Quantity
gives a value; .Code able
Concept gives a code; .value
String gives a string output like
“Positive” or “Negative”
Can check for Positive and negative results from
UDS
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
3.2 SMART on FHIR
SMART on FHIR is an open access standard based technology platform that creates apps that can
run across different healthcare systems using the FHIR standards. With this interoperable system,
patients, clinicians and healthcare providers can easily and efficiently access a library of apps to
improve care based on personalized and precision medicine standards [8].
Unlike other CDS systems, SMART on FHIR focuses on creating an open and interoperable based
technology platform, using FHIR standards, that allows vendor independent third-party apps to
run securely using EHR data [8]. A publicly accessible app gallery links to dozens of clinical
applications built on this platform that have been providing care at healthcare institutions such as
Boston Children’s Hospital and Duke Medicine. SMART on FHIR platform has even been
expanded to include genomic data standards to unify how genomic variant data are accessed from
multiple sequencing systems [8,13].
SMART on FHIR has multiple components which include adaption of FHIR standards,
authorization and authentication. Authorization is done using OAuth2, a Web standard for
authorization whose key function is to enable an end user to approve a SMART app to access an
EHR. Authentication is accomplished through OpenID Connect, a Web standard for
authentication. It defines an OAuth2-based protocol allowing end users to sign into apps using
external identity providers. The SMART on FHIR system, a health IT system that has implemented
all of these components, can then run against a SMART on FHIR app. SMART on FHIR profiles
require data to be coded using Meaningful Use terminologies and express constraints such as
terminology restrictions, element cardinality restrictions, data type choice restrictions, and
hierarchical structuring of resources [8]. Source code and examples of implementation are publicly
available [3]. In addition, sandboxes, secure virtual testing environments that mimic a live EHR,
are available through outlets such as Cerner's open developer [7,16].
3.3 Ontologies and Terminologies
FHIR is designed to work with over 40 standardized terminologies and ontologies widely used in
the biomedical informatics community [17]. See table 2. This project uses SNOMED-CT, RxNorm
and LOINC. SNOMED CT is a standardized vocabulary for clinical terms used by physicians and
other health care providers for the electronic exchange of medically-relevant health information
[18]. In addition to providing thesaurus-capability for interlinking other coding systems and
terminologies, SNOMED-CT provides a taxonomy, which is encoded in an ontological format.
Elkin et. al [19] used the Mayo Clinic Vocabulary Server (MCVS) to successfully map free text
clinical concepts to SNOMED -CT codes with a positive predictive value of 99.8 percent. RxNorm
provides a standardized coding system for drugs, linking RxNorm identifiers to multiple other drug
vocabularies [20]. RxNorm also provides some semantic structure by separating drug formulations
from ingredients and separating brand names from clinical names, and so on. However, there is
substantial overlap amongst codes, with no taxonomic way of grouping together the functional
characteristics of drug families, such as the benzodiazepines. LOINC is a common reference
terminology for clinical and laboratory measurements, assays, patient information, and so on [21].
However, unlike RxNorm, LOINC does not have any taxonomic structure encoded in way that
allows machine reasoning as is possible with an ontological implementation. Neither terminology
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
links codes via formalized relations that support moving between levels of granularity in any
systematic way.
This is problematic for our project due to the large of number of potential codes we need to search
on discover the prescription of opioids. Over 13,000 codes for various forms of opioids alone, not
including benzodiazepines, and related drugs like muscle relaxants. There exists no subset of codes
that would map to the entire value set constructed of drugs with the general opioid characteristics
for potential of abuse. Enumerating all the potential codes is possible, even could be done
automatically with the right scripting tools. However, it is not an extensible or easily updated
methodology. If this application were to be effectively deployed in a clinical environment, a much
more robust way of coding the resource requests would be required. At a minimum, the high-level
logic of the decision tree would need to decouple form the particular value sets the enumerate all
the possible drug and procedure codes
Table 2: List of standard terminologies available in FIHR
Name of Standard Name of Standard
SNOMED CT IETF language
RxNorm
Mime Types Multipurpose Internet Mail
Extensions
LOINC
Medical Device Codes defined in ISO
11073-10101
UCUM: UnitsOfMeasure DICOM Code Definitions
NCI Metathesaurus Health Canada Drug Identification Number
AMA CPT codes NUCC Provider Taxonomy
NDF-RT National Drug File – Reference
Terminology
HGNC: Human Gene Nomenclature
Committee
Unique Ingredient Identifier (UNII) ENSEMBL reference sequence identifiers
NDC/NHRIC Codes
REFSEQ: National Center for Biotechnology
Information (NCBI)
CVX (Vaccine Administered) ClinVar
ISO 2 letter Country Codes Sequence Ontology
NUBC code system for Patient Discharge
Status
HGVS: Human Genome Variation Society
RadLex
DBSNP: Single Nucleotide Polymorphism
database
ICD-9 & ICD-10
COSMIC: Catalogue Of Somatic Mutations
In Cancer
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
ICPC International Classification of Primary
Care
LRG: Locus Reference Genomic Sequences
ICF Classification of Functioning, Disability
and Health
OMIM: Online Mendelian Inheritance in
Man
Version 2 tables PubMed
A HL7 v3 code system
PHARMGKB: Pharmacogenomics
Knowledge Base
Anatomical Therapeutic Chemical
Classification System
ClinicalTrials.gov
Anatomical Therapeutic Chemical
Classification System
European Bioinformatics Institute
FHIR allows for the use of Concept Maps that allow for the linking of one or more codes (taken
here to be concepts) to one or more codes in another system. The mappings use equivalence
properties, similar to the kinds of synonymy relations found in thesauri, such as narrower, wider,
specializes. In addition to mapping between concepts (codes), FIHR allows for the mapping
between one or more concepts and a value set. This would allow for the trigger for query on
Medication Order to use a generic set of codes to map to a much broader set of enumerated
RxNorm codes for all relevant opioids. However, it still doesn’t alleviate the problems surrounding
the maintenance and generation of that list as the guidelines or coding systems change.
3.4 Proposed Development
A hybrid HTML5 app is currently being built using the SMART JavaScript client library, an open
library designed to assist with calling the FHIR API and handling the SMART on FHIR
authorization and authentication workflow [4]. A HTML5 app was chosen for cross-platform use
and to facilitate the app within the EHR environment by running it through a browser widget.
Native iOS and Android applications may be developed in the future.
In order to access the FHIR resources in the EHR through the EHR's web services, the app will
use the following process. When a clinician wants to use OPIacutepain to assess the risk of
prescribing, the EHR redirects to the SMART launch URI, implemented in the file launch.html,
then redirects to the FHIR authorization server, and then after a successful authentication, redirects
to the file index.html. The FHIR authorization server can then be accessed using the fhir-client.js
file and calling FHIR.oauth2.authorize in the launch.html file with the client id. The authorization
code is then exchanged for an access token to the authorization server using the FHIR client. When
Index.html is invoked, the SMART application will have the ability to request FHIR resources
from the EHR for the patient to run the SMART application.
The FHIR JavaScript client also facilitates calling and searching code able concepts through the
function by Codes to find our medication list through coded terms. Finally, the visual form of the
application can be called from draw Visualization, a built in function that uses the identification
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
placeholders defined in index.html. This file contains the format for the output of the app. The
displayed output for the clinician within the app will contain the MME/day risk calculator when it
can be calculated based on listed medications, whether or not a previous UDS was ordered and the
result, a table of medications for benzodiazepines and opioid medications, and patient
identification details, including gender, date of birth, and condition. Finally, the appropriate patient
specific recommendation based on the guidelines presented by the CDC and dependent on the
outcomes from the medication lists, urine drug testing results, and MME/day calculator will be
displayed for the clinician within the app.
Usability testing will be done following the creation of the app to ensure that it adds to the
workflow process of the clinician and is designed efficiently and productively for clinicians who
frequently prescribe opioid medications.
4. Discussion and Conclusion
Prescription opioid pain medications are a big contributor to the opioid epidemic. Attempts have
been made in the past to create CDSS tools within the EHR to aid clinicians in prescribing and
managing patients with chronic pain. The effectiveness of CDSS depend on how well they are
integrated and how easy it is to access them.
The SMART on FHIR platform addresses both ease of access while maintaining safety and
security of sensitive patient information through an authorization process. It also gives feedback
based on personalized patient information. This type of third-party CDSS may be able to improve
quality of care through precision medicine. In addition, there is no user training required as the
app is able to collect the back ground information automatically and provide it on demand through
a secure authentication process. The FHIR interoperability standards makes this usable by all
EHRs that are compatible with HL7 FHIR standards.
5. Limitations
This OPIpain SMART on FHIR app is an open source interoperability standard based third party
app that has not been tested and is a work in progress. The template that it is built on is new and
has not received much feedback on its usability and effectivity. The outcome measure of these
apps would take longer. Taking this background into consideration the following aspects would
have to be taken into account in further developing and implementing this app.
1. The accuracy of the EHR-specific logic to transform the specific data structures to
corresponding FHIR resources and with SMART specific profiles so that it is able
to pull up the proper information for this on demand app.
2. The SMART on FHIR apps that have been used so far have not been able to validate
the data that they have been pulling through the app. It is a work in progress and
can only be evaluated on implementation.
3. Semantic parsing of terminologies through FHIR would be based on the concept
mapping which relies on how good the terminologies and codes are.
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
The interventions for regulating and monitoring prescription pain medications and optimizing pain
management cannot be delayed till there are better methods of concept mapping. Hence it would
only be as good as the current mapping standards.
References
1. Rudd RA. 2016. Increases in drug and opioid-involved overdose deaths—united states, 2010–
2015. MMWR Morbidity and Mortality Weekly Report. 65. 2017 Feb;20(2S):S3-S92.
PubMed PMID: 28226332.
2. Dowell D, Haegerich TM, Chou R. 2016. CDC guideline for prescribing opioids for chronic
pain—united states, 2016. JAMA. 315(15), 1624-45. PubMed
https://doi.org/10.1001/jama.2016.1464
3. Mandel JC, Kreda DA, Mandl KD, Kohane IS, Ramoni RB. 2016. SMART on FHIR: A
standards-based, interoperable apps platform for electronic health records. J Am Med Inform
Assoc. 23(5), 899-908. Epub Feb 2016. doi:https://doi.org/10.1093/jamia/ocv189. PubMed
4. Manchikanti L,et. al. 2017 Responsible,Safe, and Effective Prescription of Opioids for
Chronic Non-Cancer Pain: American Society of Interventional Pain Physicians (ASIPP)
Guidelines. Pain Physician.
5. Federal Health IT website: CDSS guidelines:
https://www.healthit.gov/sites/default/files/clinical-decision-support-0913.pdf
6. New York Consolidated Laws, Public Health Law - PBH § 3331. Scheduled substances
administering and dispensing by practitioners - http://codes.findlaw.com/ny/public-health-
law/pbh-sect-3331.html
7. Lesselroth BJ, Yang J, McConnachie J, et al; Addressing the sociotechnical drivers of quality
improvement: a case study of post-operative DVT prophylaxis computerised decision
support
8. Haffey F, Brady RR, Maxwell S. 2013. A comparison of the reliability of smartphone apps
for opioid conversion. Drug Saf. 36(2), 111-17. PubMed https://doi.org/10.1007/s40264-
013-0015-0
9. Guideline Resources. CDC Opioid Guideline Mobile App;
https://www.cdc.gov/drugoverdose/prescribing/app.html
10. pH – Medical Opioid Converter by Philip Eagan; https://itunes.apple.com/us/app/ph-
medical-opioid-converter/id1082147868?mt=8
https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=26977696&dopt=Abstract
https://doi.org/10.1001/jama.2016.1464
https://doi.org/10.1093/jamia/ocv189
https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=26911829&dopt=Abstract
https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=23322549&dopt=Abstract
https://doi.org/10.1007/s40264-013-0015-0
https://doi.org/10.1007/s40264-013-0015-0
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
11. Safe Opioids app aims to prescribe narcotics more appropriately with patients; by Douglas
Maurer, DO/MPH/FAAFP | February 1, 2016
https://www.imedicalapps.com/2016/02/safe-opioids-prescribing-narcotics- patients/
12. OpioidCalcNYC App.
https://play.google.com/store/apps/details?id=gov.nyc.health.MME&hl=en
13. Wilsey BL, Fishman SM, Casamalhuapa C, Gupta A. 2009. Documenting and improving
opioid treatment: the Prescription Opioid Documentation and Surveillance (PODS) System.
Pain Med. 10(5), 866-77. Epub Jul 2009. doi:https://doi.org/10.1111/j.1526-
4637.2009.00652.x. PubMed
14. Using FHIR to develop a healthcare mobile application. Wireless Mobile Communication
and Healthcare (Mobihealth), 2014 EAI 4th International Conference on; 2014: IEEE.
15. FHIR. 2017. HL7.org: https://www.hl7.org/fhir/overview.html.
16. Millenium. 2017. CernerCode: Cerner code: http://fhir.cerner.com/millennium/dstu2/
17. Alterovitz G, Warner J, Zhang P, Chen Y, Ullman-Cullere M, et al. 2015. Smart on fhir
genomics: Facilitating standardized clinico-genomic apps. J Am Med Inform Assoc. •••,
ocv045. PubMed https://doi.org/10.1093/jamia/ocv045
18. SNOMED IHTSDO. (2017) https://www.nlm.nih.gov/healthit/snomedct/index.html
19. Elkin PL, Brown SH, Husser CS, Bauer BA, Wahner-Roedler D, et al. 2006. Evaluation of
the Content Coverage of SNOMED CT: Ability of SNOMED Clinical Terms to Represent
Clinical Problem Lists. Mayo Clin Proc. 81(6), 741-48. PubMed
https://doi.org/10.4065/81.6.741
20. RxNorm NLM. UMLS (2017) https://www.nlm.nih.gov/research/umls/rxnorm/
21. 1998. Huff SM etal. (1998) Development of the Logical Observation Identifier Names and
Codes (LOINC) vocabulary. J Am Med Inform Assoc. 5(3), 276-92. PubMed
https://doi.org/10.1136/jamia.1998.0050276
22. Manchikanti L,et. al. 2017 Responsible,Safe, and Effective Prescription of Opioids for
Chronic Non-Cancer Pain: American Society of Interventional Pain Physicians (ASIPP)
Guidelines. Pain Physician.
Acknowledgements/Grants
This work has been supported in part by an NIH NCATS Clinical and Translational Science
Award (UL1TR001412-01) and an NIH T32grant (T32 GM099607) from the Department of
Anesthesiology.
https://doi.org/10.1111/j.1526-4637.2009.00652.x
https://doi.org/10.1111/j.1526-4637.2009.00652.x
https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=19594846&dopt=Abstract
https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=26198304&dopt=Abstract
https://doi.org/10.1093/jamia/ocv045
https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=16770974&dopt=Abstract
https://doi.org/10.4065/81.6.741
https://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=9609498&dopt=Abstract
https://doi.org/10.1136/jamia.1998.0050276
Safe Opioid Prescription: A SMART on FHIR Approach to Clinical Decision Support
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 9(2):e193, 2017
OJPHI
Appendix 1: OPIacutepain: flow diagram