This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Medical Informatics, is properly cited. The complete bibliographic information, a link to the original publication on http://medinform.jmir.org/, as well as this copyright and license information must be included.
Patients commonly transition between health care settings, requiring care providers to transfer medication utilization information. Yet, information sharing about adverse drug events (ADEs) remains nonstandardized.
The objective of our study was to describe a minimum required dataset for clinicians to document and communicate ADEs to support clinical decision making and improve patient safety.
We used mixed-methods analysis to design a minimum required dataset for ADE documentation and communication. First, we completed a systematic review of the existing ADE reporting systems. After synthesizing reporting concepts and data fields, we conducted fieldwork to inform the design of a preliminary reporting form. We presented this information to clinician end-user groups to establish a recommended dataset. Finally, we pilot-tested and refined the dataset in a paper-based format.
We evaluated a total of 1782 unique data fields identified in our systematic review that describe the reporter, patient, ADE, and suspect and concomitant drugs. Of these, clinicians requested that 26 data fields be integrated into the dataset. Avoiding the need to report information already available electronically, reliance on prospective rather than retrospective causality assessments, and omitting fields deemed irrelevant to clinical care were key considerations.
By attending to the information needs of clinicians, we developed a standardized dataset for adverse drug event reporting. This dataset can be used to support communication between care providers and integrated into electronic systems to improve patient safety. If anonymized, these standardized data may be used for enhanced pharmacovigilance and research activities.
Patients commonly transition between health care locations and care providers. Yet, electronic medical records containing critical information about a patient’s medical care are usually confined to one health care sector within a geographic location (eg, a hospital) or to a group of care providers who share a common office or the same specialty (eg, a family physician group practice) [
Despite significant progress in this area, information sharing about adverse drug events (ADEs) remains inadequate, even though these are the leading cause of emergency department visits and hospitalizations [
Electronic systems could enable more accurate and complete documentation of medication safety risks and play a pivotal role in electronically communicating this information to other care providers to close this gap, but they are presently underutilized for this purpose [
Our objective was to develop a set of standardized data fields that clinicians could use to document and share information about ADEs from the point-of-care to address the information needs of clinicians and the limitations of existing systems. A secondary objective was to understand how electronic ADE documentation could be integrated into clinical activities to minimize the burden of documentation while improving reporting consistency, accuracy, and quantity.
This was a mixed-methods study completed in British Columbia (BC), Canada, between March 2014 and April 2016 using a phased approached. As we have previously published our research methodology [
In the first phase, we completed a systematic review of the existing ADE reporting systems [
The University of British Columbia Research Ethics Board reviewed and approved the study protocol. Workshop participants provided implied consent, and care providers observed during workplace observations and pilot testing provided verbal consent. Consolidated criteria for reporting qualitative research informed the reporting of study findings.
We began our work by completing a qualitative systematic review to synthesize data fields from existing ADE reporting systems worldwide [
After identifying ADE reporting concepts and data fields, we imported them into the visual thinking software Inspiration 9.2 (
In order to understand the limitations of and existing means of documenting ADEs in clinical practice, we completed qualitative observations of care providers. Trained research assistants (CB, DP, SSS, and MW) observed clinical pharmacists and physicians in emergency departments and wards in 4- to 8-hour shifts at various times of the day and days of the week. We recruited a convenience sample of participants via email, word of mouth, and personal connections of care providers on the research team. Research assistants observed the process of patient care, which included clinicians managing patients’ medications and occasionally investigating, documenting, and treating ADEs. We sought to understand the real-world clinical experiences related to ADEs, recognizing that retrospective accounts of ADEs may gloss over important characteristics, challenges, and work activities. We aimed to produce nuanced accounts of clinicians’ workflows—patterns in their activities and interactions with patients and other clinicians and artifacts in the care setting (eg, electronic medical records, paper charts, forms). We used the findings from our observations to inform design decisions, particularly in relation to the implementation of the set of data fields. Two research assistants (DP and SSS) coded the field notes from the observations using qualitative data analysis software (NVivo 11). Following initial inductive coding, the team met regularly to discuss emergent results and finalize a coding structure.
To obtain feedback on our preliminary ADE data field set, we created a preliminary reporting form using Microsoft Visual Basic for Applications to resemble a screenshot from a computer (
We presented this preliminary form to groups of clinicians in workshops scheduled during lunchtime rounds for clinicians practicing in hospital settings and scheduled in the evening for clinicians practicing outside of hospitals. We targeted groups of hospital- and community-based pharmacists, emergency department physicians, general practitioners, and hospitalists as these individuals commonly diagnose, treat, or follow up patients with acute ADEs. We recruited prospective participants using posters, email invitations, and in-person conversations with colleagues. The sessions were led or co-led by a practicing physician (CMH) or clinical pharmacist (KB) on the team and were attended by research assistants (DP and SSS) who took field notes during the sessions. We informed participants that our principal goal was to design a novel system to document and report ADEs and to obtain their feedback on our preliminary form. We presented ADE cases that we had observed in prior qualitative fieldwork to facilitate the discussion. We asked participants to identify information about the event that they felt was, or was not, required and how and where the information should be documented within the form. We asked participants to contemplate the required information needs from the perspectives of someone needing to document the information as well as receiving the report in order to balance the need to minimize documentation while ensuring that the required information was available.
Sample screenshot from a preliminary adverse drug event (ADE) reporting form created using Microsoft Visual Basic for Applications.
Sample screenshot from a preliminary adverse drug event (ADE) reporting form created using Microsoft Visual Basic for Applications.
Following each workshop, our team revised the data fields to incorporate feedback, producing a refined data field set for the next meeting or group. We maintained a log of changes that were made to the form, including the rationale for each change. Two research assistants (DP and SSS) coded the field notes from the workshops using the same approach as the qualitative observations. We considered the form to be a draft data field set when no novel suggestions or concerns were raised.
A research assistant (AC) piloted the form in two clinical settings to test it for content and functionality prior to its planned computerization [
We identified 108 active ADE reporting systems worldwide through our systematic review containing 1782 unique data fields [
Data fields deemed relevant and necessary for adverse drug event reporting by clinicians.
Data Field | Format | Description or Value Set | |
Date of birth | Alpha or Numeric | Autopopulate from EMRa or other electronic system, DD-MMM YYYY (eg, 01-FEB 1997) | |
Gender | Value set | Autopopulate from EMR or other electronic system (Male or Female or Other or Unknown) as per FHIRb | |
Name | Alpha | Autopopulate from EMR or other electronic system | |
Personal Health Number | Numeric | Autopopulate from EMR or other electronic system | |
Name | Alpha | Autopopulate from EMR login, or entered by clinician | |
Practitioner role | Value set | Autopopulate from EMR login, or entered by clinician (Doctor or Nurse or Pharmacist) as per FHIR | |
Hospital name and department | Alpha | Autopopulate from EMR, abbreviated | |
Date of report | Alpha or Numeric | Autopopulate from EMR, DD-MMM YYYY (eg, 01-FEB 1997) | |
ADE type | Value set | (Adverse drug reaction, Allergy, Incorrect drug, Subtherapeutic dose, Supratherapeutic dose, Treatment failure, Drug withdrawal, Drug interaction, Nonadherence, Other) derived from results of 4 prior prospective studies [ |
|
Symptom caused or exacerbated by ADE | Value set | Predictive entry from MedDRAd Preferred Terms | |
Diagnosis caused or exacerbated by ADE | Value set | Predictive entry from MedDRA Preferred Terms | |
Relevant tests or lab data (include dates) | Free text | Option for clinician to import from EMR (ideal) or enter manually | |
Outcome caused by ADE | Value set | (Death, Permanent disability, Exacerbated pre-existing condition, Congenital anomaly, Hospitalization, Emergency Department visit, Other, Unknown) derived from Health Canada Adverse Drug Reaction reporting standards, amended to reflect qualitative results)—not mutually exclusive | |
What happened after dechallenge or treatment? | Value set | (Resolved, Recovering, Ongoing, Resolved with Sequelae, Fatal, Unknown) as per FHIR | |
Level of certainty that the adverse event was caused by the suspect drug(s) | Value set | (Certain, Probably or Likely, Possible, Unlikely, Conditional or Classified, Unassessable or Unclassifiable, Refute) as per FHIR | |
Suspect drug actions | Value set | (Discontinue, Change dose, No change) | |
Add new medication | Multiple fields (suspect drug or product name, dose, route, frequency, other information)—see Health Product data fields below | ||
Treatment Status | Value set | (Ordered, Recommended, Received) | |
Suspect drug or product name(s) | Value set | Option to select from patient's medication list in EMR (ideal) or predictive entry from Canadian Clinical Drug Dataset combined with the Licensed National Health Products Database. Drugs included in the provincial formulary prioritized in search results. Drugs must also be searchable using a DINe or NPNf. Multiple products may be entered as suspect drugs for the same event. | |
Dose taken or received | Alpha or Numeric or Special | Manual entry | |
Dose unit | Value set | (g, mg, mcg, IUg, Units) | |
Route of administration | Value set | (Oral, SCh, IMi, IVj, Topical) | |
Frequency taken or received | Alpha or Numeric or Special | Manual entry | |
Indication for drug | Value set | Prescription indication for use subset developed by Canada Health Infoway [ |
|
Other dosing information | Free text | Manual entry | |
Additional information (important details or context, timelines, follow-up) | Free text | For clinicians to provide additional information about any of the above, specify follow-up. |
aEMR: electronic medical record.
bFHIR: Fast Healthcare Interoperability Resources.
cADE: adverse drug event.
dMedDRA: Medical Dictionary for Regulatory Activities.
eDIN: Drug Identification Number.
fNPN: Natural Product Number.
gIU: International Unit.
hSC: subcutaneous.
iIM: intramuscular.
jIV: intravenous.
Clinicians discussed the tradeoffs of various data formats, noting that structured documentation eliminated confusing shorthand and led to more succinct, comprehensible reports and analyzable data. However, they also noted that they were unwilling to use forms where the data formats forced them to enter inaccurate or incomplete information. Clinicians expressed frustration with value sets that were incomplete or where only one option could be selected when several were relevant (eg, if they had to choose a single symptom or ADE type that did not accurately reflect their patient’s situation or the use of an allergy field for documenting a drug-disease state interaction). Many health outcomes are not mutually exclusive. Therefore, clinicians are able to select more than one health outcome from the list (see “Outcome caused by ADE” field in
Clinicians highlighted the importance of knowing whether an ADE was treated, and if so, how (see “ADE Treatment Information” fields in
Clinicians pointed out that a chief concern surrounding ADE documentation was that causality is often uncertain. They needed to be able to indicate their level of certainty regarding the causality of an ADE (see “Level of certainty” field in
Clinicians pointed out that a chief concern surrounding ADE documentation was that causality is often uncertain. They needed to be able to indicate their level of certainty regarding the causality of an ADE (see “Level of certainty” field in
Data fields from the existing adverse drug event reporting forms that clinicians felt should be omitted.
Data field | Justification for excluding | |
Height or weight | Future providers can obtain this information from patients or their records. Height and weight may be relevant to dosing, but are not essential for assessing most ADEsa and patients’ medication regimen. | |
Medical history or concomitant disease states | Burdensome to enter, especially for complex patients. Future providers can often obtain this information from patients or their records. | |
Phone or mailing address or email | Burdensome to enter. If future providers have the reporter’s name, role, and institution, they will likely be able to find the contact information online. | |
Reaction start or end date, duration | Can be difficult to pinpoint (eg, delirium). Free text description of timelines is more accurate and in line with clinician charting practices. | |
Severity or seriousness | Even with standardized definitions, severity or seriousness assessments are often subjective, differ across contexts, and are prone to error. This information may be better communicated via other fields such as the patient’s outcome (eg, was the patient hospitalized?), their treatment (did the ADE require treatment? Was the drug discontinued?), symptom or diagnosis (eg, anaphylactic reaction or upset stomach), lab data (eg, low sodium of 115 or 125), and dechallenge results (resolved or worsening). | |
Rechallenge information | Often unavailable at the point-of-care, or impractical, unethical, or harmful to re-expose the patient intentionally. | |
Prescribed dose or frequency | The dose prescribed is less relevant than the dose that the patient actually took or received in relation to the ADE. Prescription information can be accessed elsewhere if needed. | |
Product strength | The dosage taken by the patient is more important. Given the product name or DINb or NPNc, product strength can usually be obtained. | |
Source (eg, pharmacy, grocery store, internet, other) | Generally not essential for assessing the ADE and the patient’s medication regimen. | |
Product start or end date, duration | Can be difficult to accurately collect (must rely on patient memory or prescription records that may be unavailable or inaccurate). Free text description of timelines is more accurate and in line with clinician charting practices. | |
Manufacturer | Not essential for assessing the ADE and the patient’s medication regimen. | |
Batch or lot # | Burdensome to gather; will often require tracing to pharmacy. Very rarely essential for assessing the ADE and the patient’s medication regimen. | |
Concomitant health products | Providers should be able to enter multiple suspect drugs, but a complete account of the patient’s medication regimen is burdensome to enter, especially for complex patients. Future providers can usually obtain other medication information from the patient, their records, or by linkage to electronic medication dispensing information depending on the jurisdiction where care is provided. |
aADE: adverse drug event.
bDIN: Drug Identification Number.
cNPN: Natural Product Number.
Throughout our work, clinicians stressed that duplicate documentation of work was a problem with existing ADE reporting forms, which took time away from patient care activities. They argued in favor of a reporting form that was integrated into the local electronic medical record and could be prepopulated with reporter information (associated with their user account), patient information (associated with the patient’s file), and possibly drug and dosing information (associated with the patient’s medication history).
Discussions with clinicians emphasized striking a balance between too little and too much information. Clinicians felt that ADE documentation should be comprehensive enough to be clinically useful and not require future providers to seek out further information (eg, a documented allergy without enough information can complicate clinical decision making). At the same time, clinicians noted that in complex cases, they might be overwhelmed with the amount of information needed to keep track of a suspect ADE, until such time as a definitive ADE diagnosis could be made. Clinical utility, simplicity, convenience, and, to a lesser degree, signal generation were central considerations for clinicians when refining the set of data fields.
When observing clinical pharmacists pilot-tested the preliminary ADE reporting form containing the data field set developed by us, they felt that its length and level of detail were appropriate. They provided few important suggestions to abbreviate the dataset further. For example, they noted that the “Date of Last Dispense” field was irrelevant to clinical care and could be omitted and that “Follow-up Items” could be noted under “Additional Comments” [
Our objective was to describe a set of data fields for clinicians to document and communicate ADEs from the point-of-care to support clinical decision making and improve patient safety. We were able to take a large number of nonstandardized data fields currently in use by ADE reporting systems internationally and condense them to one standardized dataset, while mapping some required fields to internationally recognized standards. In this process, we had to make exclusions and tradeoffs. While not all participating clinicians agreed on every field, our iterative process engaged different types of end users and was far more inclusive than is customary in information technology design in health care.
We recognize that the omission of regulatory fields may be controversial. We have taken this approach from the perspective that clinical tools need to be designed foremost to enhance the immediate delivery of care. Incomplete and nonstandardized information sharing about ADEs across health care sectors and between care providers puts patients at risk [
First, given the complexity of the ADEs observed by our team, the immediate clinician’s assessment is likely more reliable than a retrospective, at-a-distance assessment. In addition, clinicians preferred to provide causality data from the point-of-care as this assessment was felt to be crucial for informing subsequent clinical decisions.
Second, it was clear that clinicians regarded data fields used to support retrospective information gathering for regulatory agencies as burdensome. To obtain information related to manufacturer quality control issues, such as batch and lot number, clinicians often must attempt to trace the drug back to the pharmacy, a time-consuming activity that is irrelevant for most ADEs. Similarly, fields gathering information already contained in the electronic format prior to the ADE assessment, such as concomitant therapies or product start and end dates, require clinicians to duplicate the entry of information that exists elsewhere. If regulatory assessments require this data, it may be more effective to establish links to complementary datasets (such as prescribing information in a jurisdictional drug information system) than to request that clinicians provide it. We may, simply by easing documentation burden, see an increase in ADE reporting, which would contribute to improved data compared with conventional systems that most clinicians reported never having accessed.
As health systems internationally struggle to motivate providers to report ADEs and new electronic infrastructures are established to improve health information sharing across settings through e-prescribing or drug information systems, our results are timely. New electronic systems offer the potential to streamline information gathering and data entry and consolidate the multiple forms, platforms, information sources, and medication ordering tools that are necessary for clinical work. However, in practice, these systems have the potential to increase documentation burden on clinicians, cause unexpected errors, and desensitize clinicians to alerts due to overflagging and alert fatigue [
While the selection of standardized data fields alone cannot guarantee the generation of high-quality ADE reports, a clinician-informed design is more likely to result in relevant data. Implementation strategies for this dataset should continue to seek input from clinicians to facilitate uptake and adoption and ensure end-user engagement and adaptation to local contexts. If implemented with attention to clinical workflow, standardized and clinically relevant data fields may yield more accurate and complete information that can inform clinical care and improve patient safety while providing higher-quality representative data for surveillance and research activities.
At the time of publication, our team has programmed this set of data fields into an electronic app, called
Existing electronic systems allow clinicians to document ADEs, but are nonstandardized and provide limited information that can be shared across health settings and between providers. The structured and standardized set of data fields presented by us are intended to meet the needs of frontline clinicians while enabling a standardized, unambiguous communication between care providers and electronic systems to increase care quality and safety. If implemented, the minimum required data fields have the potential to address the informational discontinuity and reduce ADEs while improving the available health data for pharmacosurveillance and research purposes as a by-product of safer care.
adverse drug event
British Columbia
Drug Identification Number
electronic medical record
Fast Healthcare Interoperability Resources
intramuscular
International Unit
intravenous
Medical Dictionary for Regulatory Activities
Natural Product Number
subcutaneous
This study would not have been possible without the support of many frontline care providers who allowed us to observe their work and who provided us with insightful comments to enrich our understanding of the issues under study. This research was sponsored by the Canadian Institutes of Health Research, Partnership for Health System Improvement Grant (No. 293546), the Michael Smith Foundation for Health Research (No. PJ HSP 00002), Vancouver Coastal Health, the BC Patient Safety & Quality Council, and Health Canada. During the time of this study, CMH was funded through a CIHR New Investigator grant (No. 201109ND1-261895-157349).
None declared.
Excerpt of adverse drug event (ADE) concepts and data fields found during our systematic review using Inspiration 9.2.