Abstract
Objective
To evaluate the impact of a compulsory pop-up form on the ordering pattern of proBNP blood tests by general practitioners in the Capital Region of Denmark.
Design
A follow-up study comparing the average number of proBNP tests ordered before and after the implementation of an intervention.
Setting and subjects
From 2016 to 2021, the average number of proBNP tests increased by over 300%. In March 2022, a compulsory pop-up form was introduced in the electronic request system (WebReq), requiring general practitioners to select one of three indications for ordering proBNP, as recommended by the Danish Society of Cardiology. The study included 528 general practitioners in the Capital Region of Denmark, with data available from January 2020 to 2023, encompassing 83,576 proBNP results from 56,645 patients.
Main outcome measure
The average number of proBNP tests ordered per month and the median level of proBNP before and after the intervention.
Results
Following the intervention, the average number of proBNP tests per month decreased by 60% over a 22-month follow-up period. The highest reduction was seen among general practitioners who previously ordered the most tests (≥3 per 1000 biochemical tests). In this group, the median proBNP level increased from 12.1 pmol/L before the intervention to 13.5 pmol/L after the intervention (p < 0.0001).
Conclusions
This study demonstrates a significant decrease in the number of proBNP requests from general practitioners in the Capital Region of Denmark after the introduction of a pop-up form in the requisition system containing the current guidelines.
Keywords: proBNP, laboratory testing, evidence-based interventions, general practitioner, healthcare cost
Introduction
Healthcare costs are rising worldwide, including costs for laboratory tests. While laboratory costs typically constitute less than 5% of the total hospital budget, the laboratory tests have an extensive impact, influencing nearly 60–70% of all medical decisions [1]. The true cost of blood testing can be considerably higher if it includes subsequent downstream testing and procedures, unnecessary hospitalizations, prolongation of hospitalizations, prescription of medical treatments, etc. caused by randomly abnormal blood tests [1]. The clinical consequences of frequent phlebotomy can also result in hospital-acquired anaemia [1]. Initiatives aimed at minimizing the number of laboratory tests can enhance patient satisfaction and lower costs without negative effects on patient outcomes [2,3]. Therefore, different interventions have been implemented to reduce the nu mber of laboratory tests. These interventions are categorized as weak, moderate, or strong based on their effectiveness in reducing inappropriate testing [4]. Educational methods, like referring to evidence-based guidelines, are considered weak tools that should be combined with other interventions. The moderate and strong tools are typically integrated into the laboratory, hospital, and/or general practitioners’ (GPs’) information systems. During the requesting process, the implementation of several limitations is possible, such as minimum retesting intervals, a complete removal of unnecessary tests, requesting algorithms and reflex testing. Additionally, they involve designing request entries with decision support and incorporating pop-ups [4].
The pro-B-type natriuretic peptide (proBNP) is a natriuretic peptide synthesized and secreted by ventricular cardiomyocytes from the myocardial wall. proBNP is constantly released in the blood and the main stimulus for increased level of proBNP is myocardial stress [5–7]. It is a commonly used biomarker in heart failure (HF) for diagnosis and prognostication [8]. In 2021, the European Society of Cardiology (ESC) established a set of guidelines for the diagnosis of heart failure, including the use of proBNP as a biomarker. The presence of elevated proBNP and other evidence of structural heart disease (e.g. echocardiographic measures of LV filling or increased left atrial size) make the diagnosis likely, but are not mandatory for the diagnosis if there is certainty regarding the measurement of LVEF [9]. Furthermore, studies have shown that elevated levels of proBNP could be confounded by patient-related factors, which can result in inaccuracy in the diagnosis of heart failure [10].
In the Capital Region of Denmark, GPs have been able to order proBNP blood tests through an electronic request system used by GPs (WebReq) for several years without any restrictions. The indication has included suspected heart failure, but clear guidelines for the use of proBNP in general practice have previously been lacking. In recent years, the trend has been towards requesting echocardiography for all patients with clinical suspicion of heart failure, which has narrowed the indications for proBNP requests. In January 2021, the Danish Society of Cardiology released a position paper on proBNP in heart failure, which included recommendations for the use of proBNP in general practice among other topics [11].
From 7 March 2022, a compulsory pop-up form was implemented in WebReq. This initiative was introduced because of the large increase in the number of proBNP requisitions from GPs in the Capital Region in recent years which could probably not be explained by a corresponding increase in the number of patients with heart failure.
The aim of the present study is to evaluate the effect of this compulsory pop-up form.
Materials and methods
Data and patients
ProBNP data spanning from 1 January 2016 to 31 December 2023 was extracted from the Laboratory Information System (Labka II, CSC Denmark) from the laboratories in the Capital Region of Denmark. Patients were included from all GPs in the Capital Region of Denmark. The total number of analyses per GP, as well as data on the demographics of the associated patients were provided by CEK (Central unit of Quality assurance, Capital Region of Denmark). In total, 528 GPs with available data from 1 January 2020 to 31 December 2023 with a minimum of 1000 requested biochemical analyses (of any kind) per year were included. The study included 83,576 proBNP results from 56,645 patients.
Measurement of proBNP
The analysis of proBNP (NPU26811) was performed in seven different laboratories in the Capital Region of Denmark. All the laboratories participated in external quality control schemes to confirm the reliability and accuracy of the assays over time. However, due to the different methods employed by the laboratories, their patient median values of proBNP ranged from 11.3 pmol/L to 18.4 pmol/L with an overall median of 13.8 pmol/L. The proBNP values from the individual laboratories were adjusted accordingly. The different measurement platforms used by the laboratories and corresponding median proBNP levels are listed in Table 1.
Table 1.
Characteristics of the included laboratories before intervention (2020 and 2021).
| Lab | Median proBNP (pmol/L) | Analysis platform | Number of measurements |
|---|---|---|---|
| 1 | 12.5 | Cobas 8000 (Roche) | 13385 |
| 2 | 14.0 | Cobas 8000 (Roche) | 321 |
| 3 | 12.6 | Cobas 8000 (Roche) | 5871 |
| 4 | 11.5 | Cobas 8000 (Roche) | 2043 |
| 5 | 18.4 | Vista (Siemens) | 5820 |
| 6 | 11.3 | Atellica (Siemens) | 372 |
| 7 | 16.1 | Atellica (Siemens) | 5417 |
The same analysis platforms were used after the intervention (2022 and 2023).
Intervention and indications for proBNP measurement
Based on the recommendation from the Danish Society of Cardiology [11], the intervention was implemented in WebReq from 7 March 2022. The pop-up form included a checkbox field with the indications, requiring GPs to mark if the indication was met when ordering a proBNP test. Furthermore, a link to the guideline from the Danish Society of Cardiology about the indications for requesting proBNP in general practice was included in the pop-up form. These indications are: (1) suspicion of heart failure without definite clinical high-risk markers for cardiac dysfunction, (2) in the monitoring of heart failure patients, proBNP measurement may be considered in patients with known LVEF ≤ 40% if the patient has clinical deterioration or (3) in patients with known LVEF ≤ 40% if the patient is not receiving the latest recommended treatment according to national treatment guidelines.
Statistical analysis
Continuous variables are shown as medians with the 25th and the 75th percentiles. Categorical values are shown as numbers (percent). Differences across groups were tested using Kruskal-Wallis tests and chi-square tests respectively. Trends over time were analyzed using linear regression analysis with 95% confidence intervals for the predicted values. Finally, based on the age specific proBNP reference levels form the national guidelines [11], we used a chi-square test to compare the proportion of patients before and after the intervention, who were ruled-out to have heart failure. Likewise, the proportion of patients with elevated proBNP levels were also compared.
All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC). For all the statistical analyses, p < 0.05 was considered statistically significant.
Ethics
Data from the laboratory information system was anonymized before analysis. In Denmark, no approval from the Ethics Committee is required for this type of register-based study.
Results
Table 2 illustrates the basic characteristics of the included GPs and patients in 2020 and 2021. The groups are based on the number of proBNP analyses each GP requested per 1000 unspecific analyses. The majority of GPs (n = 377) requested less than 1 proBNP analysis per 1000 analyses during 2020 and 2021, group 1, while 115 GPs ordered more than or equal to one and less than three proBNP analyses, group 2, and a smaller group 3 of 36 GPs requested more than or equal to three. The number of proBNP measurements per GP varied considerably from 0 to 21.4 per 1000 requested analyses overall during the period from 1 January 2020 to 31 December 2021.
Table 2.
Characteristics of the included GPs before intervention (2020 and 2021).
| Group 1 | Group 2 | Group 3 | ||
|---|---|---|---|---|
| ProBNP per 1000 analyses in 2020 and 2021 | n < 1 | 1 < = n < 3 | n > = 3 | p-Value |
| Number of GPs | 377 | 115 | 36 | – |
| Patients per GP | 2691 | 2862 | 2013 | – |
| Patient age per GP | 40.0 | 40.3 | 39.7 | 0.7 |
| Total number of proBNP measurements | 7206 | 10,662 | 15,898 | – |
| proBNP (pmol/L) | 16.2 (6.7; 52.6) | 15.5 (6.0; 46.0) | 12.0 (4.5; 30.2) | <0.0001 |
| Age at proBNP measurement | 71.4 (59.0; 79.3) | 69.4 (56.9; 78.7) | 67.1 (54.5; 76.6) | <0.0001 |
| Sex (females/males, n(%)) | 3894 (54.0)/3312 (46.0) | 5579 (52.3)/5083 (47.7) | 8490 (53.4)/7408 (46.6) | 0.06 |
Continuous variables are shown as medians with the 25th and the 75th percentiles. Categorical values are shown as numbers (percent).
Differences across groups were tested using Kruskal-Wallis tests and chi-square tests respectively.
There were no differences between the median age or the number of patients associated with each practice group. The patients of group 3, who ordered the most tests, had the lowest levels of proBNP and were approximately 4 years younger than for group one when the test was taken.
Figure 1 shows the trend in the number of proBNP analyses over time before and after the intervention was implemented. Before the intervention, the average number of proBNP measurements increased constantly. We observed a marked decrease of 60% in the number of proBNP analyses immediately after the implementation of the intervention.
Figure 1.
Time course of proBNP measurements (linear regression with 95% confidence intervals) before and after the intervention.
Figure 2 shows the number of requested proBNP analyses per 1000 analyses from 2020 to 2023. There was a decrease in the number of proBNP analyses in all three groups from 2021 to 2022 when the intervention was implemented. The decrease was more pronounced in Group 3 and Group 2 compared to Group 1, although the overall trend is decreasing in all three groups. The most significant reduction occurred in group 3. On average, the GPs in Group 3 ordered around 6 proBNP analyses per 1000 analyses in 2021, decreasing to around 2 and 1 in 2022 and 2023 respectively (Figure 2). The intervention has thus reduced the wide dispersion between the three groups of GPs. Table 3 shows the median levels of proBNP before and after the intervention in the three groups. The change was not significant for group 1. However, group 2 and 3 both saw significant increases in the level of proBNP.
Figure 2.
Requests of proBNP measurements per 1000 analyses from 2020 to 2023 for group one to three. Group one requested less than one proBNP analyses per 1000 analyses. Group two requested more than or equal to one and less than three proBNP analyses. Group 3 requested more than or equal to three analyses. Median levels and interquartile ranges are shown. Differences across groups were tested using Kruskal-Wallis tests.
Table 3.
Before intervention covers the period from 1 January 2020 to 6 March 2022.
| Group 1 | Group 2 | Group 3 | |
|---|---|---|---|
| proBNP (pmol/L) | n < 1 | 1 < = n < 3 | n > = 3 |
| Before intervention | 16.0 (6.4; 52.8) | 15.4 (5.7; 46.2) | 12.1 (4.3; 30.1) |
| After intervention | 16.7 (6.9; 55.3) | 18.1 (7.2; 59.6) | 13.5 (4.3; 34.2) |
| p-Value | 0.09 | <0.0001 | <0.0001 |
After intervention covers the period from 7 March 2022 to 31 December 2023.
Values shown are medians with the 25th and the 75th percentiles.
The differences before and after intervention were tested using Mann–Whitney U tests.
Finally, based on the median proBNP levels and the national guidelines [11], the proportion of patients where heart failure was ruled out decreased significantly from 61.2% before the intervention to 57.8% after the intervention. Likewise, the proportion of patients with elevated proBNP levels increased significantly from 16.6% to 19.5%.
Discussion
Summary
The present study investigates the effect of introducing a pop-up form requiring GPs to specify the indication for the request of proBNP measurements. Following the implementation of this intervention, a notable decrease in the number of proBNP measurements was observed. As the guidelines from the Danish Society of Cardiology were released more than a year prior to the implementation of the intervention, it is likely that the intervention itself is the main reason for the reduction in the number of proBNP requests. We further observed a significant decrease of proBNP requests in the group of GPs who requested more than 3 proBNP tests, which could indicate that the intervention did reduce unnecessary measurements on healthy individuals. Supporting this point, we also found that, the proportion of patients, where the diagnosis of heart failure could be ruled out decreased, while the proportion of patients with elevated proBNP levels increased after the intervention. Specifically, the two groups requesting the highest number of proBNP measurements (group 2 and 3) both exhibited significant increases in the level of proBNP after the intervention, further suggesting that more relevant patients were assessed. However, in a clinical perspective, the increase is relatively small, suggesting there is still room for a more targeted patient selection for this blood test.
Prior to the implementation, we noted a significant increase in the number of requests of proBNP combined with large differences between GPs in the number of proBNP requests. However, the intervention has reduced the wide dispersion between the three groups of GPs.
Comparison with existing literature
Our findings corroborate previous research on the evolving patterns of test ordering in primary healthcare [1,4], and extends the understanding of reducing laboratory test orders by introducing a pop-up form. A Danish study investigated the reduction of vitamin D analyses of the GPs after implementing a pop-up form [4]. From 1 January 2017, the pop-up form was introduced to GPs, who had to state the indication for measuring vitamin D in the Capital Region of Denmark. This intervention led to a 25% reduction in the number of vitamin D requests from GPs [4]. Another study described the development in test numbers of the most frequently requested tests and simulated the effect of introducing minimal retesting intervals [12]. Increases in requests were observed in both hospitals and GPs, and these increases could not be accounted for by an increase in population size and aging population, suggesting a possible inappropriate increase in patient monitoring. The simulated effect of applying minimal retesting intervals led to a significant reduction in tests; however, the effect showed only minimal reductions in the number of tests for GPs [12]. Minimal retesting intervals might therefore not be the best type of intervention to reduce the number of requested analyses in general practice.
Studies suggest that in order to enhance the utilization of laboratory tests, a combination of various interventions is more probable to succeed [13,14]. The most successful and long-lasting interventions are multi-faceted and have included a combination of education, feedback and audit, and administrative changes [15]. It has been suggested, that the interventions are matched to the guidelines, culture and information technology infrastructure [15]. Therefore, the effect and proper utilization of laboratory tests can be enhanced by combining different interventions. In our study, we found a notable effect of an intervention, which combined the strong tool of implementing a pop-up form in the GPs’ requisition system with the weaker tool of referring to an evidence-based guideline. While it is well known that although electronic health record systems use alerts to help prevent medical errors, clinicians override many of these alerts due to desensitization from constant exposure (alert fatigue) [16,17]. Therefore, the laboratories must continuously prioritize the information load and only expose the most necessary pop-up forms to the GP’s. Pausing the pop-up forms after a certain period of use might also be helpful. Furthermore, this type of intervention could probably benefit from involving general practitioners in its design and implementation.
Strengths and limitations
The primary strengths of the study lie in its large sample size and its primary healthcare setting, where initial suspicion of structural heart disease often arises. Another notable strength is the alignment of proBNP testing indications in the pop-up form with national recommendations. The reductions in proBNP requests following the intervention supports concerns regarding potential overuse of proBNP testing before the intervention was implemented.
The observational design of the study covers the GPs patterns in requesting proBNP analyses before and after intervention. However, causality cannot be inferred with certainty from the observed results. Furthermore, some GPs could already have consulted the paper from the Danish Society of Cardiology prior to the implementation of the intervention, which could potentially overestimate the effect of the pop-up intervention.
Finally, this study examined the changes in the requisition patterns of proBNP. Therefore, assumptions regarding clinical effects, e.g. changes in the number of patients diagnosed with heart failure before and after the intervention, should be interpreted with caution. One could argue that there is a risk of underdiagnosing heart failure by introducing the pop-up form and thereby reducing the number of requested proBNP analyses. However, this aspect is beyond the scope of the present study and could be examined in future studies.
Conclusion
This study demonstrates a significant decrease in the number of proBNP requests from general practitioners in the Capital Region of Denmark after the introduction of a pop-up form in the requisition system containing the current guidelines. The effect of this type of intervention could probably be enhanced by involving general practitioners in its design and implementation.
Disclosure statement
No potential conflict of interest was reported by the author(s).
References
- 1.Eaton KP, Levy K, Soong C, et al. Evidence-based guidelines to eliminate repetitive laboratory testing. JAMA Intern Med. 2017;177(12):1833–1839. doi: 10.1001/jamainternmed.2017.5152. [DOI] [PubMed] [Google Scholar]
- 2.Hjortsø CJS, Møller MH, Perner A, et al. Routine versus on-demand blood sampling in critically ill patients: a systematic review. Crit Care Med. 2023;51(6):717–730. doi: 10.1097/CCM.0000000000005852. [DOI] [PubMed] [Google Scholar]
- 3.Zhi M, Ding EL, Theisen-Toupal J, et al. The landscape of inappropriate laboratory testing: a 15-year meta-analysis. PLoS One. 2013;8(11):e78962. doi: 10.1371/journal.pone.0078962. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Munk JK, Bathum L, Jørgensen HL, et al. A compulsory pop-up form reduces the number of vitamin D requests from general practitioners by 25 percent. Scand J Prim Health Care. 2020;38(3):308–314. doi: 10.1080/02813432.2020.1794399. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.de Lemos JA, McGuire DK, Drazner MH.. B-type natriuretic peptide in cardiovascular disease. Lancet. 2003;362(9380):316–322. doi: 10.1016/S0140-6736(03)13976-1. [DOI] [PubMed] [Google Scholar]
- 6.Cowie MR, Jourdain P, Maisel A, et al. Clinical applications of B-type natriuretic peptide (BNP) testing. Eur Heart J. 2003;24(19):1710–1718. doi: 10.1016/s0195-668x(03)00476-7. [DOI] [PubMed] [Google Scholar]
- 7.Weber M, Hamm C.. Role of B-type natriuretic peptide (BNP) and NT-proBNP in clinical routine. Heart. 2006;92(6):843–849. doi: 10.1136/hrt.2005.071233. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Januzzi JL, Tan X, Yang L, et al. N-terminal pro-B-type natriuretic peptide testing patterns in patients with heart failure with reduced ejection fraction. ESC Heart Fail. 2022;9(1):87–99. doi: 10.1002/ehf2.13749. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.McDonagh TA, Metra M, et al. 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure: developed by the Task Force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC). With the special contribution of the Heart Failure Association (HFA) of the ESC. Eur J Heart Fail. 2022;24(1):4–131. [DOI] [PubMed] [Google Scholar]
- 10.Nishikimi T, Nakagawa Y.. Potential pitfalls when interpreting plasma BNP levels in heart failure practice. J Cardiol. 2021;78(4):269–274. doi: 10.1016/j.jjcc.2021.05.003. [DOI] [PubMed] [Google Scholar]
- 11.Dansk Cardiologisk Selskab . 2024. [updated 2021 Jan 2021; cited Jan 4]. https://www.cardio.dk/brain-natriuretic-peptide-bnp-ved-hjertesvigt-2021.
- 12.Munk JK, Hansen MF, Buhl H, et al. The 10 most frequently requested blood tests in the Capital Region of Denmark, 2010–2019 and simulated effect of minimal retesting intervals. Clin Biochem. 2022;100:55–59. doi: 10.1016/j.clinbiochem.2021.11.002. [DOI] [PubMed] [Google Scholar]
- 13.Thomas RE, Vaska M, Naugler C, et al. Interventions at the laboratory level to reduce laboratory test ordering by family physicians: systematic review. Clin Biochem. 2015;48(18):1358–1365. doi: 10.1016/j.clinbiochem.2015.09.014. [DOI] [PubMed] [Google Scholar]
- 14.Jackson BR. Managing laboratory test use: principles and tools. Clin Lab Med. 2007;27(4):733–748, v. doi: 10.1016/j.cll.2007.07.009. [DOI] [PubMed] [Google Scholar]
- 15.Yeh DD. A clinician’s perspective on laboratory utilization management. Clin Chim Acta. 2014;427:145–150. doi: 10.1016/j.cca.2013.09.023. [DOI] [PubMed] [Google Scholar]
- 16.Murphy DR, Reis B, Sittig DF, et al. Notifications received by primary care practitioners in electronic health records: a taxonomy and time analysis. Am J Med. 2012;125(2):209.e1-7–209.e7. doi: 10.1016/j.amjmed.2011.07.029. [DOI] [PubMed] [Google Scholar]
- 17.McCoy AB, Waitman LR, Lewis JB, et al. A framework for evaluating the appropriateness of clinical decision support alerts and responses. J Am Med Inform Assoc. 2012;19(3):346–352. doi: 10.1136/amiajnl-2011-000185. [DOI] [PMC free article] [PubMed] [Google Scholar]