Article Text

Download PDFPDF
Impact of an electronic consultant system on hypersensitivity reactions to iodinated radiocontrast media: an observational study
  1. Min-Suk Yang1,2,3,
  2. Sang-Il Choi4,
  3. Woo-Jung Song2,3,
  4. Sae-Hoon Kim2,3,5,
  5. Hye-Ryun Kang2,3,
  6. Heung-Woo Park2,3,
  7. Sang-Heon Cho2,3,
  8. Kyung-Up Min2,3,
  9. Jae-Hyoung Kim4,
  10. Yoon-Seok Chang2,3,5
  1. 1Department of Internal Medicine, SMG-SNU Boramae Medical Center, Seoul, Korea
  2. 2Institute of Allergy and Clinical Immunology, Seoul National University Medical Research Center, Seoul, Korea
  3. 3Department of Internal Medicine, Seoul National University College of Medicine, Seoul, Korea
  4. 4Department of Radiology, Seoul National University Bundang Hospital, Seongnam-si, Korea
  5. 5Department of Internal Medicine, Seoul National University Bundang Hospital, Seongnam-si, Korea
  1. Correspondence to Dr Yoon-Seok Chang, Department of Internal Medicine, Seoul National University Bundang Hospital, 166 Gumi-ro, Bundang-gu, Seongnam 463-707, Korea; addchang{at}snu.ac.kr

Abstract

Objective To evaluate the impact of an electronic consultation support system on the incidence of hypersensitivity reactions to iodinated radiocontrast media (RCM).

Materials and methods A retrospective observational study was conducted before and after the introduction of the consultation support system. The 1-year study period was divided into two 6-month periods: before and after 1 December 2012 (baseline and intervention periods, respectively), which was when our consultation support system was introduced. Data from examinations were collected retrospectively from the hospital information centre and problem reporting sheets in the radiology department. The primary outcome was the incidence of RCM reactions before and after the introduction of the consultation support system. Generalised estimating equations were used to account for the correlation between the same patients measured on multiple occasions.

Results There were 317/20 179 (1.6%) and 186/19 873 (0.9%) hypersensitivity reactions during the baseline and intervention periods, respectively. The consultation support system significantly decreased the odds of the occurrence of a RCM reaction (OR=0.59, 95% CI 0.49 to 0.71, p<0.001) compared with baseline. There was also a twofold increase in the premedication rate after initiation of the consultation support system (OR=2.05, 95% CI 1.16 to 3.65 p=0.01). However, there was no significant difference in the recurrence rate between the periods (OR=0.97, 95% CI 0.58 to 1.65, p<0.93).

Conclusions The introduction of the consultation support system reduced the incidence of hypersensitivity reactions to RCM and increased the use of premedication in patients with known hypersensitivity to RCM.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Background and significance

Iodinated radiocontrast media (RCM) are some of the most widely used medical compounds worldwide1 and are one of the major causes of drug hypersensitivity in Korea.2 ,3 RCM hypersensitivity can be divided into immediate reactions, which occur within 1 h after exposure, and delayed reactions, which occur from 1 h to 7 days after exposure. The severity of immediate reactions varies from mild (eg, simple cutaneous symptoms such as itching and urticaria), moderate (eg, mild systemic symptoms including abdominal, respiratory and cardiovascular symptoms) to severe (eg, life-threatening anaphylactic shock).4 ,5 Mild immediate reactions occur in 0.7–3.1% of patients receiving RCM, while severe reactions occur in 0.02–0.04%.6 Fatal reactions occur in 1–3 patients per 100 000 RCM administrations.6 Risk factors for immediate reaction to RCM include a previous history of RCM reaction; asthma; or history of any allergic reaction. The most important of these is a previous history of RCM hypersensitivity.7–10 Identification of ‘at-risk patients’ is important as there is evidence that premedication with steroids and H1 antihistamines for patients with a history of moderate and severe RCM reaction prevents recurrent reactions.11–13 Indeed, failure to obtain an adequate history of previous RCM reaction is a recognised ‘avoidable error’ that can lead to recurrence of RCM reaction.14 It is essential that all departments delivering iodinated RCM should have systems designed to minimise RCM reactions and this should include 100% identification of patients at risk.

One approach to identifying patients at risk is through the use of electronic medical records (EMRs) that include alerts for adverse drug reactions. These allow the development of computerised clinical decision support systems (CDSSs) with software-based algorithms set up to link clinical decisions with information already in the EMR and thus alert physicians to patients who are at risk.15 ,16 CDSSs are used to improve many aspects of clinical performance, including prescribing, diagnosis and preventative care.17–19 They offer an approach to the development of a system to prevent RCM hypersensitivity in patients at risk. However, CDSSs related to the drug allergy alert have been found to be less effective when the alerts are too complex as this encourages ‘alert fatigue’ with over-riding of computer-generated alerts.20 ,21 The factors that encourage drug allergy alert over-riding include incorrect information on the patient's allergy provided by the EMR and lack of knowledge about possible alternatives.20 ,22 Thus we set up a simple CDSS system for prevention of hypersensitivity reaction to RCM to minimise any tendency to ‘over-ride’ generated alerts (box 1). Our system can be described as an allergy specialist-centred drug allergy management, which helps physicians consult an allergy specialist who reviews the accuracy of the history of the patient's allergies and provides alternative ways of managing patients with RCM hypersensitivity. The strengths of this programme are its ease of use as well as immediate response, reliability and efficiency for clinicians.

Box 1

Electronic medical record (EMR) and consultation support system (figure 1)

In our EMR system, medically important information on patients, such as drug allergies including radiocontrast media (RCM) hypersensitivity, food allergies and communicable diseases (eg, viral hepatitis and HIV infection), is recorded in an “alert” window (figure 1B) for all patients in their own default pages (figure 1A).

Once a staff member checks the RCM alert in a patient's alert window, an ‘alarm’ window pops up in the EMR when the patient is prescribed any radiological study using RCM in our consultation support system. The alarm window indicates that this patient has had an RCM reaction; it provides a button to automatically produce a consultation request to an allergist when clicked. The alarm window also provides the standard protocol for the prevention of RCM reactions. The standard RCM hypersensitivity prevention protocol recommends intravenous injection of chlorpheniramine 4 mg 1 h before RCM injection for patients with previous mild reactions and prednisolone 50 mg orally three times (13, 7 and 1 h) before RCM injection with intravenous chlorpheniramine 4 mg injection 1 h before RCM injection for patients with more severe previous reactions.

If the patient has no record of reaction in his or her alert window, then another pop-up window appears, asking the clinician to obtain the medical history of RCM reactions from the patient when they prescribe RCM-based imaging. If the clinician clicks ‘yes’, meaning that the patient has experienced RCM hypersensitivity reactions, then the same alarm window appears and the clinician proceeds with a consultation with the allergist, as outlined above (figure 1C).

In this study we evaluated the impact of this CDSS on the incidence of RCM reactions (figure 1).

Figure 1

Consultation support system. *Note: this web page was originally written in Korean and subsequently translated into English for publication. EMR, electronic medical record; RCM, radiocontrast media.

Materials and methods

An observational study was conducted from June 2011 to May 2012 at Seoul National University Bundang Hospital (SNUBH, box 2). This 1-year study period was divided into two 6-month periods: before and after 1 December 2012, which was when our consultation support system was introduced. The first 6 months was designated the baseline period and the second 6 months, the intervention period. The occurrence and recurrence rates of RCM reactions as well as the frequency of premedication prescription were compared between the two periods. Data from examinations were collected retrospectively from the hospital information centre and problem reporting sheet (see ‘Data collection’ below).

Box 2

Seoul National University Bundang Hospital (SNUBH) and the department of radiology

SNUBH is a teaching hospital with 906 beds and ∼50 000 outpatients visiting each day. It uses a fully computerised electronic medical record (EMR) system. SNUBH opened in 2003 as a fully digital hospital, successfully launching a new hospital information system named “BESTCare” (Bundang Hospital Electronic System for Total Care).18 In October 2010, SNUBH was accredited as having a stage 7 comprehensive EMR system according to the Healthcare Information and Management Systems Society (HIMSS) Analytics Electronic Medical Record Adoption Model (EMRAM); this was the first stage 7 accreditation outside North America.23 HIMSS Analytics EMRAM evaluates the state of information technology implementation of a hospital according to eight levels (0–7) of EMR capability in areas ranging from limited ancillary department systems to a paperless EMR environment.

The department of radiology at SNUBH provides more than 1 million diagnostic and interventional services annually. They have seven CT machines including one 16-, three 64- and three 256-multidetector CT machines. There are 53 radiologists (27 professors, 11 clinical fellows and 15 residents), 4 nurses and 22 radiographers in the department.

Examinations prescribed during the baseline period and executed during the intervention period were excluded. CT scans that covered different parts of the body of a patient and performed on the same day were counted as single examination. All other examinations performed during the study period were included.

Data collection

Data from the hospital information centre

We collected data on all CT scans performed during the 1-year study period from the SNUBH medical information centre. This dataset included patient identification number, sex, age, the type of scan, the date it was ordered, the date it was performed and the amount of RCM used.

Data from problem reporting sheets

Trained attending nurses in the department of radiology completed problem reporting sheets for patients who had problems with any radiological examination. The problem reporting sheet included all previous and current problems related to radiological examinations, such as low glomerular filtration rate and metformin use. Thus, all the patients who underwent RCM-enhanced CT scans were primarily screened for previous RCM reactions by trained nurses. The problem reporting sheets of patients with a medical history of RCM hypersensitivity included demographic data, and details of their hypersensitivity reaction were described.

Inclusion and exclusion criteria

An allergist reviewed problem reporting sheets to extract data on RCM hypersensitivity. We subsequently identified examinations with RCM hypersensitivity from the lists of all the CT scans (from the hospital information centre) and entered the data on the RCM hypersensitivity (from problem reporting sheets). The severity of the reactions was classified as mild, moderate or severe. Mild reactions included urticaria, flushing, pruritus, chills and rhinitis symptoms such as rhinorrhoea and nasal obstruction, as well as conjunctival injection. Moderate reactions included angioedema, dizziness, dyspnoea and gastrointestinal symptoms. Severe reaction included anaphylaxis, loss of consciousness and desaturation.

Isolated nausea and emesis were excluded, because we could not differentiate pharmacological toxicity from hypersensitivity reactions. Reports on delayed hypersensitivity reactions were excluded because they were dependent solely on the patient's voluntary reports.

Data analysis

Simple data description, mean and SD for continuous variables and counts and proportions for categorical variables, is used to summarise the data. The main variable of interest was the proportion of patients with RCM hypersensitivity and whether this was different before and after the introduction of the consultation support system. Although the units of analysis were individual radiological examinations about 20% of patients had examinations in both measurement periods. To account for the correlation between the same patients measured on two occasions generalised estimating equations (GEEs) were used with a binary logit link function and using an autoregressive (1) correlation structure to account for the correlation between repeated measurements on the same patients. Other variables of interest were differences in the proportions of patients who had premedication, the severity of reactions that did occur and recurrence rates of RCM hypersensitivity in those who had previously documented RCM hypersensitivity; all between the two time periods. A covariate adjusted analysis was also done; with age, sex, type of RCM used and the severity of previous RCM reactions as adjustment variables. IBM SPSS V.20.0 was used.

Ethical approval

This study was approved by the institutional review board of SNUBH (IRB No B-1307-212-122). The requirement to obtain informed consent was waived because of the retrospective study design. All patients’ records and information were anonymised and de-identified before analysis.

Results

Demographics

Examinations

A total of 64 400 RCM-enhanced CT scans were performed in the 1-year study period. In the baseline and intervention periods, 20 179/36 092 (55.9%) and 19 873/21 368 (93.0%) CT scans met the inclusion criteria, respectively (figure 2). The drop-out rate in the baseline period was high, because 14 430/36 092 (40.0%) examinations were prescribed in the baseline period and performed in the intervention period. Iohexol was the most frequently used RCM, followed by iomeprol and iopromide. The mean amounts of RCM used in the baseline and intervention periods were 137.1±41.2 and 138.0±41.6 mL, respectively (table 1).

Table 1

Demographic characteristics of patients

Figure 2

Study flow diagram. RCM, radiocontrast media. *The study period consisted of the baseline and intervention periods. **Because the date of prescription and the CT scan could be different, some patients were prescribed CT scans during the baseline period and received the scan during the intervention period. To eliminate the effect on the awareness of healthcare workers, we excluded such cases. †CT scans that covered different body parts of a patient and performed on the same day were counted as a single examination. ‡Includes the low glomerular filtration rate, use of metformin, etc.

Patients

Mean patient ages in the baseline and intervention periods were 56.6±16.5 and 57.1±16.0 years, respectively. In both periods, 54.8% of the patients were men and 45.2% were women. Furthermore, 20.5% of patients were exposed to RCM two or more times during the study period.

Reported reactions to RCM

There were 3254 reports in the problem reporting sheets. Reviewing the reports disclosed 1010 reports (563 and 447 in the baseline and intervention periods, respectively) of RCM hypersensitivity. Among these, 15 records on delayed hypersensitivity were excluded. The remaining 995 reports included reports on previous RCM reactions that the attending nurse had already known about when the problem reporting sheet was completed (ie, a previous reaction), events that occurred when the problem reporting sheet was completed (ie, present reaction) or both: 555 (268 with previous and 317 with present reactions) in the baseline period and 440 (283 with previous and 186 with present reactions) in the intervention period (figure 2).

Present RCM reactions

There were 317/20 179 (1.6%) and 186/19 873 (0.9%) hypersensitivity reactions in the baseline and intervention periods, respectively. Of these, 284/317 (89.6%) were mild, 27/317 (8.5%) were moderate and 6/317 (1.9%) were severe reactions in the baseline period and 156/186 (83.9%) were mild, 27/186 (14.5%) were moderate and 3/186 (1.6%) were severe reactions in the intervention period. GEE analysis indicated that the consultation support system significantly decreased the odds of the occurrence of an RCM reaction (OR=0.59, 95% CI 0.49 to 0.71, p<0.001) in the intervention period compared with that in the baseline period. Age, sex, the type of RCM used and the severity of previous reactions were included as covariates; except for iomeprol versus iodixanol, all were significantly different between the baseline and intervention periods (table 2).

Table 2

Univariate and multivariate analyses of the occurrence of radiocontrast media hypersensitivity

Patients with previous RCM reactions

There were 268/20 179 (1.3%) and 283/19 873 (1.4%) examinations for patients with previous RCM reactions in the baseline and intervention periods, respectively. Of these, 258/268 (96.3%) were mild, 9/268 (3.4%) were moderate and 1/268 (0.4%) were severe previous reactions in the baseline period and 269/283 (95.1%) were mild, 5/283 (1.8%) were moderate and 9/283 (3.2%) were severe previous reactions in the intervention period. In the baseline and intervention periods, 30/268 (11.2%) and 29/283 (10.2%) examinations resulted in the recurrence of RCM reactions, respectively. GEE analysis indicated that there was no significant difference in the recurrence rate after the introduction of the consultation support system (OR=0.97, 95% CI 0.58 to 1.65, p<0.93). In addition, age, sex and the type of RCM did not differ significantly between the two periods (table 3).

Table 3

Univariate and multivariate analyses of the recurrence rate of patients with previous RCM reactions

Premedication of patients with previous RCM reactions

In the baseline period, 233/268 (86.9%) patients with previous RCM reactions were pretreated: 228 (97.9%) and 5 (2.1%) had mild and moderate previous reactions, respectively. In the intervention period, 264/283 (93.3%) patients with previous RCM reactions were pretreated: 254 (96.2%), 3 (1.1%) and 7 (2.7%) had mild, moderate and severe previous reactions, respectively. GEE analysis indicated that the premedication rate increased twofold after the introduction of the consultation support system (OR=2.05, 95% CI 1.16 to 3.65 p=0.01). Age and sex did not significantly influence the premedication rate (table 4).

Table 4

Univariate and multivariate analyses of the premedication rate of patients with previous RCM reactions

In the baseline period, 23/233 (9.9%) recurrent hypersensitivity reactions were observed despite premedication. Two of these reactions were moderate and two others were severe; before these incidents, the patients had only experienced mild reactions. The other 19 reactions were of the same severity as the patients’ previous reactions, including 17 mild and 2 moderate. In the intervention period, 28/264 (10.6%) hypersensitivity reactions occurred despite premedication. Three of these reactions were moderate and occurred in patients with previously mild reactions; the other 25 were of the same severity as the patients’ previous reactions, including 22 mild, 2 moderate and 1 severe reaction.

Discussion

The introduction of a simple consultation support system decreased the incidence of hypersensitivity reactions to RCM. In addition, more patients with a history of hypersensitivity to RCM were identified and given premedication.

A recent meta-analysis has shown that CDSSs in the form of computer reminders achieve a median improvement in target process adherence of 4.2% (IQR 0.8–18.8%). However, the improvement of a clinical endpoint was less obvious than that of process adherence: the median absolute improvement in a clinical endpoint, which was mainly reducing blood pressure, was only 2.5% (IQR 1.3–4.2%).24 The target process in our study was prescribing premedication to at-risk patients, and the clinical outcome was reducing the incidence of RCM reactions. Our system successfully and substantially improved both target processes and clinical outcome. The system may have been successful, because clinicians were aware of the specific features of RCM hypersensitivity itself but were unfamiliar with the proper management of patients with a history of RCM reactions. This might explain the similar success reported by other Korean researchers for the prevention of RCM reactions.25

Our system seemed to improve clinicians’ level of awareness of RCM hypersensitivity reactions. However, our system did not affect the recurrence rate of patients with previous RCM reactions; that is to say, the abovementioned effect of our system involved patients without a previous RCM reaction. This is an unexpected result, because the risk of RCM hypersensitivity for patients without a history was unpredictable. We expected that our system would improve clinicians’ performance in prescribing premedication, which would consequently lead to better patient outcomes.26 This unexpected result can be explained by increased awareness of the clinicians; that is to say, more thorough interviews with patients might have disclosed a history of RCM reactions that had gone unrecognised. Improving awareness among medical and paramedical personnel was one of the ancillary goals of this programme.

This study has a few limitations. First, we failed to detect a reduction in the recurrence rate. As there were few recurrent RCM reactions (23/233 (9.9%) and 28/264 (10.6%) in the baseline and intervention periods, respectively), it is possible that we cannot confirm the effectiveness of our system for reducing the recurrence rate. As the efficacy of premedication in the prevention of RCM reactions remains to be established,27 the limitation of premedication might explain the lack of improvement in the recurrence rate. Second, we did not identify known risk factors for RCM reactions, such as atopy and concomitant use of β-blocking agents.28 Because this study adopted an observational ‘before and after’ design, the problem of a secular trend could arise29; that is to say, increased awareness of the general population of RCM reactions might have influenced the results but could not be adjusted for. In addition, the baseline period included summer and autumn, while the intervention period included winter and spring; this is a possible source of bias, because seasonal variation might influence the incidence of RCM reactions.30 Thus, further studies with longer baseline and intervention periods are required to rule out this possibility.

In conclusion, the introduction of a simple consultation support system decreased the incidence of hypersensitivity reactions to RCM through consultations with allergy specialists and increased awareness among medical and paramedical personnel of RCM hypersensitivity.

Main messages

  • Our consultation support system effectively decreased the incidence of radiocontrast media hypersensitivity.

  • Our system increased the prescription rate of premedication for patients with previous radiocontrast media reactions.

  • Our system appears to have increased the awareness of medical and paramedical personnel of radiocontrast media hypersensitivity.

Current research questions

  • Will the effect of the consultation support system on radiocontrast media (RCM) hypersensitivity demonstrated in this study be reproduced in other centres or countries?

  • Will the recurrence rate decrease in a larger study?

  • Is premedication effective for preventing recurrence in patients with a history of RCM hypersensitivity?

Key references

  • Park CS, Kim TB, Kim SL, et al. The use of an electronic medical record system for mandatory reporting of drug hypersensitivity reactions has been shown to improve the management of patients in the university hospital in Korea. Pharmacoepidemiol Drug Saf 2008;17:919–25.

  • Segal AJ, Bush WH Jr. Avoidable errors in dealing with anaphylactoid reactions to iodinated contrast media. Invest Radiol 2011;46:147–51.

  • Shojania KG, Jennings A, Mayhew A, et al. The effects of on-screen, point of care computer reminders on processes and outcomes of care. Cochrane Database Syst Rev 2009;(3):CD001096.

  • Bae YJ, Hwang YW, Yoon SY, et al. The effectiveness of automatic recommending system for premedication in reducing recurrent radiocontrast media hypersensitivity reactions. PLoS One 2013;8:e66014.

References

Footnotes

  • Contributors M-SY wrote the statistical analysis plan, cleaned and analysed the data and drafted the paper. S-IC, W-JS, S-HK and J-HK designed data collection tools, cleaned and analysed the data. H-RK, H-WP, S-HC and K-UM monitored data collection and revised the draft paper. Y-SC designed data collection tools, monitored data collection and revised the draft paper. He is guarantor.

  • Competing interests None.

  • Ethics approval The institutional review board of Seoul National University Bundang Hospital.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement Datasets are available, from the corresponding author via email, to any investigators who are interested in these datasets.