Article Text

Download PDFPDF

Sustained reductions in emergency department laboratory test orders: impact of a simple intervention
  1. Kevin H Chu1,2,
  2. Amol S Wagholikar3,
  3. Jaimi H Greenslade1,2,
  4. John A O'Dwyer3,
  5. Anthony F Brown1,2
  1. 1Department of Emergency Medicine, Royal Brisbane and Women's Hospital, Brisbane, Queensland, Australia
  2. 2Burns, Trauma and Critical Care Research Centre, School of Medicine, University of Queensland, Brisbane, Queensland, Australia
  3. 3Australian e-Health Research Centre, CSIRO ICT Centre, Brisbane, Queensland, Australia
  1. Correspondence to Dr Kevin Chu, Department of Emergency Medicine, Royal Brisbane and Women's Hospital, Butterfield Street, Herston, QLD 4029, Australia; k.chu{at}uq.edu.au

Abstract

Objectives To determine whether a pathology request form allowing interns and residents to order only a limited range of laboratory blood tests prior to consultation with a registrar or consultant can reduce test ordering in an emergency department (ED).

Methods A prospective before-and-after study in an adult tertiary-referral teaching hospital ED was conducted. A pathology request form with a limited list of permissible tests was implemented for use by junior medical officers. Tests for patients 16 years and older presenting in a 20-week pre-intervention period from 19 January 2009 were compared with those in a corresponding 20-week post-intervention period from 18 January 2010. Main outcome measures were the number and cost of blood tests ordered.

Results 24 652 and 25 576 presentations were analysed in the pre- and post-intervention periods, respectively. The mean number of blood tests ordered per 100 ED presentations fell by 19% from 172 in the pre- to 140 in the post-intervention period (p=0.001). The mean cost of blood tests ordered per 100 ED presentations fell by 17% from $A3177 in the pre- to $A2633 in the post-intervention period (p=0.001). There were falls in the number of coagulation profiles (11.1 vs 4.8/100 patients), C-reactive protein (5.6 vs 2.7/100 patients), erythrocyte sedimentation rate (2.5 vs 1.3/100 patients) and thyroid function tests (2.2 vs 1.6/100 patients).

Conclusions Pathology request forms limiting tests that an intern and resident may order prior to consultation with a registrar or consultant are an effective low maintenance method for reducing laboratory test ordering in the ED that is sustainable over 12 months.

  • ACCIDENT & EMERGENCY MEDICINE
  • PATHOLOGY
  • HEALTH SERVICES ADMINISTRATION & MANAGEMENT
View Full Text

Statistics from Altmetric.com

Introduction

The ordering of laboratory blood tests is an integral part of the diagnostic investigation of emergency department (ED) patients. A focused history and physical examination, while vital, usually needs to be supplemented by laboratory tests and radiological investigations. Blood tests are, however, frequently found to be ordered excessively in an indiscriminate manner to help  reach a final diagnosis. While this problem is not new, it is a recurring one.1 ,2

The international literature over 20 years overwhelmingly reports that tests are ordered excessively.3–10 There are negative consequences from the overuse of blood tests.4 New tests such as D-dimer and B-type natriuretic peptide (BNP) have become available in the last decade. An initial enthusiasm in their uptake has led to excessive ordering and consequent, often unnecessary, further diagnostic testing including medical imaging for frequent false positive results. This has necessitated reconsideration of the indications for their use.11 ,12

Various interventions for reducing test ordering have been reported and summarised.4 ,9 ,13 ,14 Successful interventions have included multifaceted strategies combining several modalities. However, sustaining a long-term reduction in test ordering is challenging. Commonly employed interventions such as education, protocols and feedback, require ongoing monitoring and are resource intensive and problematic in departments where staff turnover is high. Successful implementation using a low maintenance strategy is necessary to sustain a long-term reduction in excessive test ordering.

The objective of this study was to determine whether a pathology request form designed to limit the range of laboratory blood tests which an intern or resident may order without first consulting a registrar or consultant, can result in sustainable long-term reductions in laboratory blood test ordering in an ED where junior medical staff turnover is high.

Methods

Design and setting

The study took place within a large ED in Brisbane, Australia (see box 1). A prospective before-and-after quasi-experimental design was used. Pre-intervention data were collected for 20 weeks from 19 January 2009 over two complete intern-resident rotations. The study intervention commenced on 8 June 2009 at the start of a new intern-resident rotation. Data were collected for a further 52 weeks from the commencement of the intervention. Comparison was made on the post-intervention data that were collected during the 20 weeks from 18 January 2010; the pre-intervention data were previously collected at the same time of year.

Box 1

Study setting

The study was performed in the emergency department of a metropolitan public adult tertiary-referral teaching hospital in Australia. The department includes a 16-bed short-stay (observation) ward and has over 70 000 attendances per year. It is staffed by at least two emergency medicine consultants 16 h a day, at least two emergency medicine training registrars 24 h a day, and six to seven interns and residents at any given time. The interns and residents rotate through the department five times a year.

Intervention

A pathology request form was designed and implemented to limit the range of laboratory blood tests that could be ordered by interns and residents without first consulting a registrar or consultant. Permissible tests (see box 2) on this form were ordered by ticking a check box. Several frequently requested tests that were thought to be ordered excessively (see box 2) were specifically omitted from the form. However, any test could still be ordered following consultation with a registrar or consultant.

Box 2

Laboratory blood tests that may be ordered by interns and residents on the pathology request form

Interns and residents may order tests from the following list without first needing to consult with a registrar or consultant:

  • Full blood count

  • Electrolytes liver function tests (including electrolytes, glucose, liver function tests, calcium and urate)

  • Troponin

  • Lipase

  • International normalised ratio*

  • Human chorionic gonadotropin

  • Paracetamol

  • Lactate†

  • Blood culture

Tests deliberately omitted from the above list:

  • Coagulation profile (including international normalised ratio and activated partial thromboplastin time)

  • Erythrocyte sedimentation rate

  • C-reactive protein

  • Thyroid function tests

  • D-dimer

  • B-type natriuretic peptide

*If the patient is taking warfarin.

†Lactate was included because of our department's research interest in sepsis.

Interns and residents were instructed on the use of this form as part of their orientation to the ED rotation (see box 3). The paper request form tool was designed to be a low maintenance strategy to produce sustainable reductions in test ordering. No other specific educational activity or feedback was provided.

Box 3

Implementation of the use of pre-printed pathology forms

The pathology form was designed to be self-explanatory with tick boxes for permissible tests and a statement on the form that additional tests require registrar or consultant authorisation. Information about the use of the pathology form was incorporated into a printed orientation booklet provided to interns and residents at the start of their emergency department rotation. This booklet (along with that of other specialties) was available electronically on the hospital intranet. Instructions were also provided informally at the bedside during the interns’ and residents’ first emergency department shift.

Data collection

Clinical and laboratory data were sourced electronically from the ED information system (EDIS) and the hospital pathology databases, respectively. The two databases were merged for this study using the Health Data Integration tool developed by the Australian e-Health Research Centre.15 ED attendances were identified from EDIS. Patients were excluded if they were 15 years of age or younger, did not wait to be seen or were dead on arrival. A test was attributed to a patient if it was requested between ED arrival and departure time. When the patient was admitted to the ED short-stay ward under the care of an emergency medicine consultant, the departure time was the ED short-stay ward departure time.

Outcome measures

The primary outcome was the number of laboratory blood tests ordered. For the purpose of this study, a single laboratory test is one where a state pathology public price is assigned. The secondary outcome was the cost of laboratory blood tests ordered. Costs (in Australian dollars) were derived from the 2011–12 state pathology public price schedule.

Data analysis

It was estimated that the study intervention would achieve a 20% reduction in laboratory blood test ordering. Since coagulation profile was considered one of the most overused tests, the sample size was estimated based on a 20% reduction in coagulation studies from 10 to 8 studies per 100 ED attendances (α 0.05, β 0.10). Post-intervention data were collected beyond the required sample size (4300 in each of the pre- and post-intervention periods) over 52 weeks in order to examine trends over a 1-year period.

Study data were summarised graphically and numerically using standard measures for categorical and continuous variables. The 20-week pre-intervention outcome measures in 2009 were compared with the 20-week post-intervention measures over the same months in 2010 to investigate changes in laboratory blood test ordering following the implementation of the intervention. The comparison was made using Mann–Whitney U tests. Statistical analysis was performed using SAS (SAS Institute, Cary, North Carolina, USA).

Results

There were 26 885 and 27 356 presentations to the ED in the pre- and post-intervention periods, respectively. After the exclusion of patients who were aged 15 years or younger, did not wait to be seen or were dead on arrival, 24 652 and 25 576 presentations were included in the analysis.

Baseline data

Patient demographics and disposition were similar over the two periods (see table 1). Patients averaged around 41 years of age, 35% arrived by ambulance and just under 30% were admitted.

Table 1

Patient characteristics and disposition in the pre- and post-intervention periods

Full blood count (FBC), electrolytes liver function tests which included glucose, calcium and urate (ELFT), coagulation profile which included international normalised ratio and activated partial thromboplastin time (COAG), troponin (TN) and lipase accounted for two thirds of the blood tests ordered in the pre-intervention period (see table 2).

Table 2

Most frequently ordered laboratory blood tests in the 20-week pre-intervention period

Changes in test ordering

The mean number of laboratory blood tests ordered per 100 ED presentations fell by 19% from 172 in the pre- to 140 in the post-intervention period (p=0.001). The mean cost of laboratory blood tests ordered per 100 ED presentations fell by 17% from $A3177 in the pre- to $A2633 in the post-intervention period (p=0.001). There were falls in the number and cost of COAG, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), thyroid function tests (TFT) and BNP (see table 3 and figure 1).

Table 3

Comparison of laboratory blood tests ordering in the pre- and post-intervention periods

Figure 1

Number of laboratory blood tests per 100 emergency department presentations over 72 weeks. For the analysis, the pre- and post-interventional periods were weeks 1–20 and weeks 53–72, respectively. The intervention (see box 3) began in week 21 (arrow) at the start of a new intern-resident rotation. Each rotation was of 10–12 weeks’ duration. (A) Full blood count (FBC); (B) electrolytes liver function tests (ELFT); (C) coagulation profile (COAG) and international normalised ratio (INR); the light shaded area represents the coagulation profile (INR and activated partial thromboplastin time) and the dark shaded area represents the INR; (D) C-reactive protein (CRP); (E) erythrocyte sedimentation rate (ESR); (F) thyroid function tests (TFT); (G) troponin; (H) D-dimer; and (I) B-type natriuretic peptide (BNP).

COAG fell by over a half. International normalised ratio (INR) orders rose as they replaced COAG orders in patients taking warfarin. However, overall COAG and INR testing combined fell by just under 50% (figure 1). Other tests that significantly rose were lipase and human chorionic gonadotropin.

Discussion

The use of a pathology request form designed to limit the range of blood tests which an intern or resident may order without first consulting a registrar or consultant can result in sustainable long-term reductions in test ordering in an ED where junior medical staff turnover is high. Our intervention has addressed a need within the healthcare system to curtail the unsustainable growth in pathology ordering accompanying the growth in clinical activity.16 The simplicity and apparent sustainability are the strengths of our interventions that add to the multifaceted strategies reported in the literature.

Numerous interventions for reducing test ordering have been reported. Stuart et al described an educational programme together with audit of/feedback concerning a test ordering protocol.2 Their 30–40 min educational programme was delivered to groups of three or four ED medical officers. They reported a 33% and 32% per patient reduction in haematology and biochemistry costs, respectively, 12 months after the intervention. By comparison, our strategy showed a more modest 17% reduction in costs. The study by Stuart, unlike our study, did not report on the baseline number of tests ordered per patient, thus making direct comparison difficult.

Similarly, efforts by the Rational Investigation Ordering (RIO) collaborative in 2002 were able to reduce pathology tests by 25%.17 The collaborative used a ‘traffic light system’ where tests were grouped by colours, indicating the level of authority required for the requested tests. However, 2 years after the implementation of this intervention, the number of pathology tests requested had risen, necessitating further education. Our strategy, unlike that of Stuart et al or RIO, was designed to be low maintenance without requiring ongoing education or feedback.

This study has several limitations. First, our strategy was implemented in a large metropolitan adult teaching hospital ED, which may limit its generalisability. This weakness is, however, also a strength. In an environment where junior staff turnover is high, a low maintenance strategy is required to ensure the sustainability of any reduction in test ordering.

Second, this study did not examine the appropriateness of test ordering or morbidity and mortality following our intervention. Quality of care relating to patient outcome from ordering or not ordering a test is difficult to quantify.13 ,18 ,19 Occasional media reports on the adverse consequences of not ordering a test such as CRP20 may add to medicolegal fears of inadequate test ordering. Our strategy, however, permitted the ordering of any blood test when indicated following consultation with more senior staff.

Third, cost savings in test reduction may come at the expense of the consultants’ time. In our ED, all interns and residents are required to discuss the management of their patients with a registrar or consultant, so any additional expense of the consultants’ time would be marginal. This practice provides not only good clinical care but also teaching opportunities for junior staff.

Fourth, feedback from interns, residents as well as registrars, was not specifically sought for this project. Anecdotally, the restriction on test ordering was thought to be practical rather than cumbersome, because it articulated what was expected of junior staff. Finally, this study did not specifically examine changes in blood test ordering elsewhere in the hospital as a result of the ED strategy to control test ordering.

Ten blood tests accounted for more than 80% of the number and costs of our laboratory blood tests. Moreover, FBC and ELFT accounted for 40% of the total. Even through low cost, these high volume investigations are known to contribute significantly to overall healthcare costs.13 ,21

The six tests specifically omitted from the permitted list (see box 2) accounted for only a minor portion of the tests ordered (13%) and of the total cost (12%). During the 20-week post-intervention period, almost every other test including TN showed a significant reduction. We believe that the re-designed pathology form not only restricted the ordering of targeted tests, but also brought more awareness to whether a test should be ordered or not for all patients presenting to the ED. This awareness was reinforced by the hospital's campaign to reduce pathology ordering throughout the institute, particularly the repeat daily ordering of previously normal tests. The hospital campaign has been successful and the ED has not been observed to shift pathology costs to inpatient departments.

In conclusion, the results of our study show that use of a pathology request form limiting the tests that an intern or resident may order prior to consultation with a registrar or consultant is an effective method to control laboratory blood test ordering in a teaching hospital ED, and appears sustainable over 12 months.

Main messages

  • Pathology request forms allowing interns and residents to order only a limited range of laboratory blood tests prior to consultation can reduce test ordering.

  • The intervention was simple to implement and maintain.

  • The reduction in test ordering was sustained over 12 months.

Current research questions

  • How should the appropriateness/inappropriateness of ordering or not ordering a test be measured?

  • How can different interventions to reduce test ordering in different clinical settings be compared?

  • How can various interventions to reduce test ordering across the healthcare system be combined?

Key references

  • Alonso-Cerezo MC, Martin JS, Garcia Montes MA, et al. Appropriate utilization of clinical laboratory tests. Clin Chem Lab Med 2009;47:1461–5.

  • Schattner A. The unbearable lightness of diagnostic testing: time to contain inappropriate test ordering. Postgrad Med J 2008;84:618–21.

  • Miyakis S, Karamanof G, Liontos M, et al. Factors contributing to inappropriate ordering of tests in an academic medical department and the effect of an educational feedback strategy. Postgrad Med J 2006;82:823–9.

  • Lewandrowski K. Managing utilization of new diagnostic tests. Clinical leadership & management review 2003;17:318–24.

References

View Abstract

Footnotes

  • Contributors KC conceived the study. All contributed to the design, data collection and analysis, and writing of the manuscript.

  • Funding The Queensland Emergency Medicine Research Foundation funded this study (QEMRF-PROJ-2009-016).

  • Competing interests None.

  • Ethics approval The Royal Brisbane and Women's Hospital, Human Research Ethics Committee approved this study.

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

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.