Article Text
PDF
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
A considerable fraction of laboratory testing done in high income countries is estimated to be wasteful use of healthcare ressources. The potential for reducing testing volume has previously been documented in the short term in limited patient populations.
WHAT THIS STUDY ADDS
This study documents the feasability af of limiting excess biochemical testing in a diverse internal medicine patient population and sustaining this for a prolonged period through organisational awareness, educating junior physicians in the rational use of biochemical testing and addressing workflow issues related to the ordering of tests.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTISE OR POLICY
The employed strategies might inspire other hospitals to implement similar programs to reduce unneccesary biochemical testing.
Introduction
Problem
At Gødstrup Hospital, the former West Jutland Regional Hospital, a Danish secondary and tertiary centre with at the time two principal hospital sites in Herning and Holstebro, the Department of Medicine had wards in both locations with 108 beds combined in 2018. Seventy two of these beds were in wards at the Holstebro site, where the yearly number of admittances rose from 6751 to 7655 (+13%) between 2015 and 2018, while the average length of hospital stay decreased from 4.2 days to 3.7 days. Meanwhile, the number of biochemistry tests ordered increased in the same timespan by 18% (figure 1).
Number of biochemical tests at the Holstebro site. Annual number of test results from the Department of Clinical Biochemistry to the Holstebro site from 2014 to 2018. The number includes all the tests performed or facilitated from the Department of Clinical Biochemistry and includes blood, urine and serious fluid samples analysed in the laboratory, point-of-care-tests, blood gas analyses performed locally and ECGs.
This discrepancy—together with observations by some of the authors in routine practice of test being done without likeliness of them impacting on patient treatment and care—raised the question of potential overtesting. Budgetary constraints in the Department of Clinical Biochemistry further added to the decision in January 2019 to form a multidisciplinary group with participation from both the Department of Clinical Biochemistry and the Department of Medicine to address the issue within the framework of the hospital’s improvement programme, which is based on the ‘Model of Improvement’ as developed by the Associates in Process Improvement. The group operated independently and was authorised by department managements to implement changes in workflow and biochemistry profiles as it saw fit at the internal medicine wards at the Holstebro site, where the department has 72 beds in specialised wards for haematology, nephrology and pulmonology as well as geriatric beds. Initially, no intervention was planned at the Herning site, where there are 36 beds for infectiology, gastroenterology, endocrinology as well as geriatric patients. The Herning site could serve as a quasi-control group for the intervention.
Available knowledge
Ageing populations in large parts of Europe and the rest of the Western world as well as medical innovation have led to a steady increase in healthcare usage and associated costs over the past decades.1–3 At the same time, treatment regimens for many conditions have evolved, reducing the average length of inpatient care considerably.4 In Denmark, whose public, universal healthcare system is tax-funded and controlled by regional political elected bodies, this trend has been augmented by alterations in the organisation of the healthcare sector, centralising hospital services and creating more capabilities for continued treatment, care and rehabilitation closer to the patients’ home with either outpatient management or in a care facility in the municipality.
The number of available hospital beds within internal medicine departments countrywide has shrunk over the last decade from roughly 7500 beds in 2009 to less than 5000 beds in 2019. This happened despite an ageing population, making short patient turn-over times a necessity and laboratory tests serve as an integral part of the accelerated diagnosing and monitoring process as well as being used in assessing, when patients can safely be discharged from the hospital.5
On the other hand, there are estimates internationally of considerable wasteful spending in healthcare, including the use of laboratory tests, although the data supporting the accuracy of these estimates is limited.6–8 Observed weekend testing practices at a large district hospital in Staffordshire, UK have been audited, with 47.5% of all phlebotomies reported likely unnecessary and 60.8% of requested assays graded probably unnecessary or unnecessary when judged by medically qualified chemical pathologists.9 Reports by clinicians too have found substantially wasteful practices when auditing the use of specific biochemistry tests in more select patient populations.10–12 Reports on physicians ordering test for a multitude of other reasons than patient care,13 14 even ones perceived as unnecessary by themselves when ordered also support the notion of widespread unnecessary biochemistry testing.15 A survey conducted by Laegevidenskalige Selskaber, the Danish umbrella organisation for Scientific Medical Societies, among the medical scientific societies as well as patient interest groups in May 2020 revealed the widespread conviction, that unnecessary testing is performed in the Danish healthcare system as well.16
While inappropriate laboratory testing is a misuse of the finite resources in public healthcare systems it also raises the ethical questions of patient harm through unnecessary phlebotomy, contributing to hospital-acquired anaemia, and false-positive results, including the risk of derailing the diagnostic process through abnormal results unrelated to the patient’s complaints. On the other hand, patient care and safety could potentially suffer, if indicated tests were omitted due to restrictive policies. Therefore, our approach in this quality improvement project (QIP) focused on optimising the likelihood of biochemistry testing being chosen wisely.
Rationale
We theorised two major drivers for possible inappropriate biochemistry testing: (a) insufficient knowledge about the biochemistry tests ordered and their rational use by the ordering physician and (b) workflow related to requesting biochemistry.
Was based on supervising and informally interviewing house officers. This revealed poor knowledge of the precise content of the different biochemistry profiles and the rational use of biochemistry tests as well as ordering being driven by habit or organisational expectations (perceived by the junior physicians) of daily blood tests. It also put a spotlight on redundant biochemistry profiles with almost identical names but slightly differing content between wards, potentially leading to unintended tests being ordered as physician-staff circulate between wards.
Was assessed by an anonymous questionnaire to nurses involved in rounds, who according to local routine in the majority of cases put in the request for blood tests after rounds. This revealed several issues, but a major finding of concern was that in a quarter of the reported cases no decision on future biochemistry testing was clear to the nurse after rounds—a gap to a large extent filled by the nurse ordering according to her/his own judgement based on local routine, despite the fact, that ordering biochemistry testing is an intervention in principle requiring medical approval from the physician.
Specific aim
The multidisciplinary group set its own goal: to reduce the number of biochemistry tests ordered on inpatients at the Holstebro site, excluding point-of-care-testing (POCT, eg, blood glucose, urine dipstick, blood gas analysis) as well as microbiology tests, by 5% by the end of the year 2019. The goal represented a rough estimate of the perceived realisable reduction potential agreed on in the group.
Methods
Interventions
In an effort to improve junior doctor knowledge, all biochemistry profiles were re-evaluated, removing redundancy, standardising them across all wards and securing relatable profile names in a multitiered system based on organ systems in addition to a broad screening profile intended for use at admittance. These new profiles were implemented in the end of June 2019 and displayed on posters in all ward offices and distributed on pocket cards, with some rather rarely indicated and/or costly profiles being highlighted as requiring a senior physician concurring in the indication for these tests.
Educational sessions on rational use of biochemistry analyses (discussing indication, typical pre-analytical errors, variability, minimal retesting interval and pitfalls in interpretation of results) were established starting in February 2020 targeting the most junior physicians about 1 month into their employment. These sessions together with recurrent presentation of this QIP and its preliminary results during physician meetings also served to assure the junior physicians about the acceptability of not ordering biochemistry daily without clinical indication and to heighten the awareness for incorporating the rational use of biochemistry testing into routine supervision of junior physicians.
Workflow and communication between physician and nursing staff during rounds was articulated, emphasising the physician’s responsibility for planning and conveying the future need for biochemistry testing, including a special focus on the need for testing during the weekend, and a checklist was implemented.
A change in routine towards physicians themselves entering the request for biochemistry testing was tested in a selected subgroup in a PDSA-fashion (Plan, Do, Study, Act), but not chosen for obligatory general implementation. The reason for this were time concerns, resistance in the group of physicians as well as considerations about the role of informing about biochemistry testing plans in the communication between professional groups.
To minimise unnecessary retesting, default Electronic Medical Records lab sheet settings were adjusted to biochemistry tests from the last 12 months being displayed instead of the last 30 days.
These interventions were implemented gradually as marked in figure 2.
Number of biochemical tests and interventions in quality improvement project. Timeline showing number of biochemical test before, during and after the quality improvement project (QIP). Yellow arrows with numbers indicate milestone interventions in the project. (1) Project startup—multidisciplinary group formed; (2) Electronic Medical Records lab sheet setting adjusted; (3) questionnaire about workflow and communication; (4) Plan-Do-Study-Act (PDSA): physicians entering requests for biochemistry testing; (5) new biochemistry profiles implemented including posters and pocket cards; (6) checklist implemented; (7) educational sessions on rational use on biochemistry analyses established. Horizontal lines indicate median number of biochemical tests before (red), during (green) and after cessation (blue) of the QIP. The only intervention at the Herning site was that new biochemistry profiles were implemented.
Measurements
Effects were studied by a monthly monitoring of the number of performed biochemistry tests excluding POCT (eg, blood glucose, urine dipstick, blood gas analysis) according to the laboratory information management system (Labka II) with data for year 2018 serving as baseline. Results were adjusted for activity as expressed by patient bed occupancy (hours) and total caseload as registered in the hospital’s business intelligence database.
Potential adverse effects were studied using the same data sources, monitoring average length of hospital stay, delayed discharge, re-admittance and ordering of add-on testing of drawn samples or late blood draws, potentially disrupting workflow in both ward and lab. A qualitative survey among laboratory staff was performed to detect changes in perceived workload by demand for additional biochemistry tests done on blood drawn earlier.
In addition, a unanonymised qualitative chart audit was performed on 20 consecutive patients with pneumonia as principal diagnosis independently by three voluntary senior registrars to detect whether indicated test were being omitted at a time point, when use of biochemistry analyses seemed to have stabilised on a new level.
Finally, we estimated the resulting savings using data on the observed aggregate number of ordered individual biochemistry tests if ordered more than 10 times in the period between 1 October 2019 and 28 February 2020 and comparing them with the corresponding data 1 year earlier. The intention was to analyse a 6-month period after the activity adjusted use of biochemistry tests seemed to have stabilised on a new level, but March 2020 had to be ruled out as an outlier due to the SARS-CoV-2 pandemic. The aggregated number of tests for each period was multiplied with the laboratory’s marginal test-price on an individual test basis for tests done in-house and multiplied by the agreed quota-price for tests done in other laboratories. The observed difference in cost was then activity adjusted and extrapolated to estimate yearly differences.
Results
Over the course of spring 2019 we observed a gradual decrease in biochemistry tests ordered starting even before any interventions were implemented. The decrease was accentuated in the autumn 2019 after the implementation of new biochemistry profiles, workflow interventions and continued awareness through the assessments and PDSA cycles made.
By the end of 2019 median monthly biochemistry tests had decreased by 13% when adjusted for ‘bed hours’ and 21% when adjusted for total caseload. During 2020, the reduction appears to be at least sustained, especially when disregarding March 2020, which was particularly impacted by the SARS-CoV-2 (COVID-19) pandemic (figure 3A,B). The apparent continued reduction at the Holstebro site during the pandemic has to be interpreted with caution in relation to this QIP, as patients with symptoms compatible with COVID-19 would be admitted to the Herning site first and to considerable extend transferred to the Holstebro site afterwards if COVID-19 could be ruled out, skewing some of the testing done at admittance toward the Herning site.
Number of biochemical tests per inpatient at the intervention site in Holstebro (A) and the control site Herning (B). Horizontal lines indicate median number of biochemical tests per inpatient before (red), during (green) and after cessation (blue) of the quality improvement project (QIP). At the Herning site no QIP was performed and therefore only a red line is presented. From April 2020 and forward data from the Herning site are highly influenced by the COVID-19 pandemic and data collection from this site stopped in March 2020.
The pre-pandemic reduction at the Holstebro site was not paralleled by any decrease in the use of biochemistry tests at the ‘control’ site in Herning, where there was no intervention initially. Instead, the steady rise in overall testing activity seen in both sites in previous years continued, until the department management in October 2019 decided to implement the re-evaluated biochemistry profiles from the Holstebro site there, without this having any readily identifiable impact on the activity adjusted testing volume on its own (figure 3B). Data beyond February 2020 are not shown for the Herning site, as the pandemic influenced the site far more the remaining year than the Holstebro site due to its infectiology specialty, obscuring any interpretation in relation to this QIP. Note that the activity adjusted data from the two sites is not comparable in absolute terms, partly due to differences in case mix, but mainly due to different organisation, as the initial biochemistry screening on admittance is performed in the emergency department at the Herning site.
The mean length of hospital stay was not impacted adversely by this QIP (figure 4A). The same was true for the percentage of readmittances and hospital discharges before noon (a parameter measured in Denmark as part of agreements with the surrounding municipalities, designed to ease their planning of continuous care after discharge) (figure 4B,C). As the number of individual biochemistry requisitions did not rise as a consequence either, we suggest the number of phlebotomies endured by patients is unlikely to have risen due to the intervention (figure 4D).
Possible adverse effects of the quality improvement project (QIP): figure shows average length of stay in days (A), readmission rate (B), discharges before noon in (C) and percentage number of biochemical requisitions (D). Horizontal lines indicate median number of biochemical tests before (red), during (green) and after cessation (blue) of the QIP.
In a qualitative survey the laboratory staff did not report an increased workload by demand for additional biochemistry tests done on blood drawn earlier, nor was there any emerging new pattern in which tests were ordered this way.
The chart audit performed in August 2019 found no cases of more than one auditor considering an indicated biochemistry test missing (ie, no cases of concordance between the three independent auditors), and the most frequent comment was the potential to further limit biochemistry testing.
The observed reductions in tests included in the analysis (detailed in online supplemental file 1) between the two time periods analysed in depth correspond to yearly savings of approximately Kr320 000 (−15%) at the Holstebro site. The estimate does not include utensils to draw blood, salary for the phlebotomist that makes the blood draw or salary for clinical staff that enter the requisition and interpret the result—nor does it include any overhead, as charged by laboratories in non-integrated healthcare systems/commercial laboratories.
Supplemental material
Discussion
This QIP significantly reduced the number of ordered biochemistry tests—even considerably more so than we initially set as our goal—without overtly affecting patient safety or patient turnover adversely.
It is not possible to make strong suggestions on which parts of the various interventions contributed most to this result, but the observed reductions even before we implemented tangible changes suggest, that awareness among staff on the organisational focus on the subject as witnessed by the establishment of this group made a considerable contribution in itself. One can only speculate whether the awareness triggered a renewed ability to rely less on habit and to question local requesting culture when ordering biochemistry testing; or whether it provided the organisational backing for clinicians necessary to break with defensive testing practices due to the perceived risk of harsh regulatory sanctions for medical error prominent in Denmark in recent years; or which other of the multitude of drivers for ordering inappropriate or unnecessary laboratory testing mentioned in the literature were influenced.13–15
The employed interventions echo several of the recommendations made in the literature for demand management by laboratories.17 18 One being the joint group between departments acting as both a link from the laboratory to the clinical specialty and a forum in which to agree on appropriate test usage. The re-evaluation and display of profiles quasi-targeting the laboratory request form in a time of digitalised requesting and employing staff-grade specific limitations being a second. The establishing of educational activities to support and sustain rational ordering practices being a third and the reporting of the QIP results at physician meetings acting as a feedback on test volume a fourth. We would very much have liked to include guidance during electronic requesting on minimal retesting intervals, but our Electronic Health Record system so far does not support this.
The results in our view further support targeting behavioural factors when trying to limit inappropriate laboratory testing, especially when targeting more than one factor.19 While many of the more recent reports of reductions in laboratory testing due to interventions targeting behavioural factors in a hospital setting focused on a specific test or panel in a rather specific patient demographic or covered a short timeframe, we report a reduction with a broad scope intervention and in a more diverse patient population, that was sustained for more than a year.10–12 20 We intend to continue the educational sessions and have made the group linking laboratory and the department of internal medicine permanent, although assembling with a lower frequency. Time will tell whether this will be sufficient to sustain the observed reductions in the long term.
It is very likely given the observed differences in relative reductions between individual wards (data not shown), that local routines, case mix and patient demographics had a significant impact on the observed effect size of the employed interventions. We tried to reassess the workflow-interaction between nurses and physicians regarding future laboratory tests in mid-2020. Unfortunately, results were not possible to subject to meaningful interpretation as the response rate of the repeated questionnaire was critically low, suggesting that staff engagement in further advancing the project might be faltering, possibly in light of the ongoing pandemic, despite stable reductions in laboratory test demand.
While the workflow related issues together with differences in case-mix and healthcare system architecture represent obvious limitations to the generalisability of our findings, it is striking, that the effect size reported by Thakkar et al after interventions primarily targeting awareness through education of staff was of comparable magnitude and so to us suggest a potential for a significant effect of raising awareness and improving continuous education in this regard in other settings than our local department as well.
We estimate cost savings of 15% on the analysis of biochemical test. As the estimate is based on the marginal cost, the actually realised savings are likely to be even greater, especially in a non-integrated healthcare system. At the same time, the estimate does not include the expected reduction in time spent by clinical staff entering the requisition for biochemistry tests as well as signing off on, interpreting and conveying the results to patients. We did not find it feasible within this QIP to estimate such savings on salary. However, reducing time spent for sampling and interpretation of blood and urine samples can create time for other clinical activities.
Conclusion
This QIP met its declared goal of reducing biochemistry testing, confirming the potential to reduce unnecessary testing through awareness, education and workflow interventions.
Unnecessary or inappropriate testing seems widespread and of considerable volume according to the literature. Budgetary constraints in healthcare will undoubtedly intensify in the coming years under the combined stress of ageing populations and new costly treatment options. This project might serve as inspiration for similar approaches to limit wasteful and harmful ordering practices.
One might speculate whether an effort to promote more rational use of laboratory testing—being one of the most frequent investigations done on patients—also might facilitate a similar change in ordering practices for more expensive or potentially more harmful investigations (eg, involving radiation exposure) through daily training in questioning the impact on treatment of routine investigations. Future projects should consider reporting such measures as well.
While there is a need for medical professionals to collaborate to meet the educational needs to promote rational use of resources and to address flaws in institutional culture facilitating and perpetuating wasteful practices, the sustained success of such efforts is also dependent on the removal of external incentives driving physicians to ordering rather more than less, whether they are of financial nature, fear of regulatory sanctions or litigation. Reversing the trend towards over testing in the long run will therefore probably also require persuading and working with patient interest groups and advocacies to facilitate political reforms.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information. The data is available upon request.
Ethics statements
Patient consent for publication
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Twitter @TobiasREberlein
Contributors All authors contributed to the development and implementation of the QIP. MSM and PN were responsible for data collection and presentation. All authors contributed to data interpretation. TRE wrote the first draft of this manuscript. TRE, JK, LL, MSM, PN and FHM collaborated in the editing of the manuscript. TRE, LL and FHM are guarantors of the manuscript.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.