Breaking the guidelines: how financial unawareness fuels guideline deviations and inefficient DVT diagnostics

Monitoring of diagnostic efficiency

Low-value care (LVC), offering minimal to no benefit to the patients while incurring excess costs, is a global challenge and is estimated to cost the US healthcare system between 76 and 102 billion USD yearly [1]. Overuse of tests is a common form of LVC [2] and can be driven by financial incentives, medico-legal concerns, and availability/turn-around time. The de-implementation of LVC is essential to reducing resource waste and thus indirectly improving patient and population health [3]. Clinical guidelines, algorithms, evidence-based recommendations, and indices [4] that quantify and monitor hospital-level LVC can optimize patient care.

Diagnostic considerations for deep vein thrombosis

The current study examines the diagnostic evaluation for suspected deep vein thrombaosis (DVT) to analyze diagnostic considerations of healthcare providers (HCPs). DVT represents a critical medical concern due to the risk of progressing to pulmonary embolism. DVT is a relatively frequent provisional diagnosis, and data indicate a rising incidence of DVT cases [5], 6]. Current guidelines advocate that suspected DVT evaluations should include a pretest probability assessment, with the Wells score being the most widely used tool [7], 8]. In the low probability group, D-dimer is the suggested initial test; if normal (<500 ng/mL), no further testing is necessary. However, elevated D-dimer levels prompt the performance of duplex ultrasound of the affected lower extremity [7]. The high probability group is recommended to undergo immediate ultrasound examination without prior D-dimer testing [7]. In the moderate probability group, diverse clinical perspectives exist regarding initiating D-dimer testing or directly proceeding to ultrasound examination [7].

Healthcare and hospital financial models

Healthcare financing varies globally; however most industrialized countries operate some form of publicly funded healthcare with universal coverage. In Hungary, healthcare predominantly operates under a national health insurance model, wherein most hospitals are state-funded, and patients are nationally insured and receive healthcare services free of charge at the point of use. The National Health Insurance Fund of Hungary (Hungarian abbreviation: NEAK) reimburses hospitals based on procedures performed. A “performance unit” score is assigned to each procedure (e.g., diagnostic tests). Reimbursement is based on a flat conversion rate of the performance units. This Fund aims to promote cost-efficient operations by allocating a monthly allowed budget of “performance units”. Exceeding allocated budgets leads to unreimbursed services, encouraging providers to prioritize lower-cost diagnostics.

Objectives

At the Department of Laboratory Medicine of the University of Debrecen Clinical Center (UDCC), a decline in the D-dimer test orders for suspected DVT has been observed over the past decade. Given the prominent role of D-dimer in DVT diagnosis, our objective was to explore the diagnostic practices for suspected deep vein thrombosis at the Emergency Department of the UDCC and compare that to gold-standard guidelines. Rather than proposing new guidelines, we aimed to assess factors contributing to deviation from existing recommendations.

Data sources

We used hospital-level data from the University of Debrecen Clinical Center. The clinical data warehouse was accessed via Microsoft Azure Databricks PaaS (Platform as a Service), a cloud-based solution for processing, storing, sharing, analyzing, and modeling datasets. We used Spark SQL queries and rule-based text processing techniques to generate data for subsequent statistical analysis.

Defining and finding DVT-suspect cases in medical records

We tried various filter queries to identify all patients who presented with suspected DVT at the ED. It was not possible to identify cases where test results were negative using discharge codes because discharge codes do not include diagnoses that were ruled out during the visit. To capture all patients with suspected DVT, we analyzed the presenting complaints, past medical history, physical examination results, and discharge summary in free text format. We found that medical phrases and discharge language were relatively consistent, which aided patient identification with classic data mining library of regular expressions (regex).

Our final filters included admission date (January 1, 2007–December 31, 2020), department code, the presence of the Hungarian terms for “DVT” or “deep vein thrombosis” in the discharge summary in combination with the phrases: “was ruled out”, “was proved”, “was confirmed” or “was excluded”. The chosen expression reflects terminology commonly used in everyday clinical practice. We excluded those cases where neither D-dimer nor lower extremity duplex ultrasound test results were available despite being mentioned in the discharge summary.

Information on the identified patient records, including admission date, chief complaints, medical status, D-dimer results, lower extremity duplex ultrasound results, and discharge summary, were exported for all cases. We curated the exported cases by thoroughly reviewing chief complaints and discharge summaries, ensuring the relevance of all included cases for the study.

The extracted case numbers matched clinical expectations. The good sensitivity of patient identification is further supported by literature data indicating an annual incidence of 1–2 deep vein thrombosis (DVT) cases per 1,000 population [9], 10]. Given the 200,000–250,000 population in the ED catchment area, the annual DVT count was expected to range from 200 to 500. Our method identified 2–5 times this number, which is reasonable considering that a portion of patients initially suspected of having DVT were ultimately not diagnosed with the condition. The specificity of patient identification was considered 100 %, due to highly targeted filters.

Reimbursement rates

All diagnostic tests are assigned performance points by the NEAK, which we retrieved from the official NEAK website [11], 12]. NEAK determines the reimbursement valuation of the performance points [13]. Reimbursement based on NEAK’s valuation, may not reflect actual costs incurred by the hospital. Hungarian Forint (HUF) was converted to US Dollar (USD) at the average exchange rate of 308 HUF/USD based on the average conversion rate between January 1, 2020, and December 31, 2020 [14].

Survey

We conducted an anonymous survey among the current healthcare providers to explore factors influencing their diagnostic decision-making. All physicians, including residents and specialists, had the opportunity to fill out the survey during their daytime shifts, voluntarily. Of the 20 attending specialists and 17 resident physicians employed at the Emergency Department of the University of Debrecen, 12 (60 %) specialists and 7 (41 %) residents participated. 40 % of the responders were resident physicians, 60 % were specialists. 35 % of the responders had specializations in emergency medicine; the rest had other specializations or did not have any (i.e., resident physicians). Responders’ years of experience in emergency care are shown in Figure 1.

Figure 1:

Responders’ years of experience in emergency care.

Statistical analysis

Anonymized data from health records were exported in.csv and further analyzed in Microsoft Excel (Microsoft) and GraphPad Prism version 10.2 (GraphPad Software). The data were tested for normal distribution with the D’Agostino & Pearson test. Those that passed the normality test were further examined with the Pearson correlation test. For those that were non-Gaussian, a non-parametric Spearman correlation test was performed. All of these were done using GraphPad Prism.

Cost calculation at the UDCC

We calculated the reimbursements based on the performance points and the reimbursement rates defined by NEAK. Lower extremity venous duplex ultrasound was associated with 1,296 points, and the D-dimer laboratory test was associated with 3,543 points in 2020 [11], 12], 15].

We observed that the performance points allocated for the duplex ultrasound and D-dimer remained unchanged throughout the examined period between 2007 and 2020 (Figure 3). To ascertain whether these assignments were outliers or reflective of the standard practice of stagnant performance point levels, we checked other commonly requested diagnostic procedures. We found that performance point levels were virtually always constant, suggesting that performance points assigned to the diagnostic tests do not change.

Figure 3:

Reimbursement rates and performance points assigned to selected diagnostic tests by NEAK. Performance points remained constant between 2007 and 2020 for all the diagnostic tests. There was one universal reimbursement rate for all diagnostic tests, and that increased over time between 2007 and 2020 (shading under USD cents/points only serves better visibility of that particular function). NEAK, National Health Insurance Fund of Hungary; USD, United States Dollar.

However, reimbursement valuation of performance points increased gradually over time in a stepped method, from 0.44 USD cents/point in January 2007 by 45 % to 0.64 USD cents/point in January 2020 (Figures 3 and 4) [13].

Figure 4:

Real and “accounted” costs of D-dimer and duplex ultrasound tests in the United States and Hungary. The remarkably low Hungarian state reimbursement value for the duplex ultrasound is in red. HU, Hungary; US, lower extremity duplex ultrasound; USA, United States of America; USD, United States Dollar.

As the cumulative inflation rate during this period was comparable to this, 54.45 % [16], we used the latest reimbursement values to compare the costs of diagnostic tests to private healthcare costs. We collected the current prices (as of 2024) of diagnostic tests in a private setting (Figure 4), as the pricing history of private services is not publicly available. Via online searches, we collected costs from 19 private facilities for D-dimer diagnostic tests and from 39 sites for lower extremity duplex ultrasound. In the private setting, the costs of the examined diagnostic tests exhibited notable disparities. Costs of D-dimer tests in private healthcare in Hungary were similar to the state reimbursement rates (24.59 USD compared to 22.67 USD); however, the price of the duplex ultrasound was nearly 10 times higher in the private setting compared to the state reimbursement value, with an average of 74 USD vs. 8.29 USD.

To compare Hungarian pricing ratios, we also collected pricing and reimbursement data from the USA. Medicare reimbursement was 10.52 USD for a D-dimer and 195.87 USD for a lower extremity duplex ultrasound. Prices in a private setting were slightly higher but proportional; the average price of a D-dimer test at 174 healthcare sites was reported to be 27 USD [17], while the unilateral lower extremity duplex ultrasound averaged 238 USD [18] at 146 sites across the country. While ultrasound assessments were significantly more expensive than D-dimer tests in the U.S. Medicare and private and in the Hungarian private setting, the Hungarian state reimbursement scheme is striking: ultrasound examination prices constituted one-third that of a D-dimer test reimbursed by the state, altogether approximately 30 times less than that observed in the USA.

Calculation of actual costs of diagnostic tests

Upon comparing the state-funded reimbursement rates for duplex ultrasound to prices in private settings in Hungary in the USA, they appeared outdated and did not accurately reflect the actual costs. We actual costs for this labor-intensive diagnostic test at the UDCC, performed by radiologists with administrative support. Publicly available data on salaries [19], [20], [21], [22], accessory pricing from online price comparison platforms [23], 24], energy prices from the service provider’s official price list [25] and equipment depreciation with maintenance intervals and replacement periods based on manufacturers’ recommendations and previous research studies [26], [27], [28], [29] were used. We used an average examination time of 14–15 min, factoring in preparation and breaks [30].

We modeled three cost scenarios: best-case, worst-case, and the average (Table 1). Expenses included wages, accessories, energy consumption, and depreciation. In the best-case scenario, lower wages (junior physicians) and prices were used with a shorter examination duration, reduced accessory usage, and an extended device lifecycle. In the worst-case scenario, higher wages and prices were used with long examination duration, increased accessory consumption, and a shorter device lifecycle.

Actual costs ranged between 11.24 USD and 44.51 USD, averaging 25.16 USD. The primary cost driver was wages, led by radiologists (7.2 USD to 33 USD per test), assistants (2.1 USD to 5.8 USD per test), followed by hospital attendants’ wages for helping with the moving and transportation of the patient, at 1.4 USD to 3.3 USD. Wages accounted for approximately 96 % of all expenses.

Accessories contributed ∼3.3 % (e.g., gel, gloves, paper), and depreciation of equipment added 0.81–1.47 %. Negligible factors like energy costs and minor furniture depreciation were excluded.

Study among ED healthcare providers

We surveyed Emergency Department healthcare providers to explore factors influencing their diagnostic decision-making. When evaluating suspected deep vein thrombosis, 75 % of respondents reported ordering Duplex ultrasound as their initial test, 15 % ordered both tests concurrently, and 10 % preferred starting with a D-dimer test.

Time efficiency was cited by 75 % of physicians at least once as a key reason for not prioritizing D-dimer testing (Figure 5), despite 60 % reporting similar turnaround times for D-dimer and ultrasound (Figure 6). This preference appears tied to the belief that starting with ultrasound avoids additional tests after a positive D-dimer. All attending specialists (100 %) strongly agreed, and all residents (100 %) agreed or strongly agreed that elevated D-dimer levels can have multiple underlying causes.

Figure 5:

Responses to the survey question “what influences your preferred diagnostic approach in suspected DVT?”

Figure 6:

Responses to the survey question “which test results do you get sooner (within shorter turnaround time)?”

Cost considerations were directly mentioned by only 10 % of respondents, though 45 % believed D-dimer costs were comparable to or higher than duplex ultrasound (58 % among specialists) (Figure 7). Additionally, just 15 % (3 respondents) cited “workplace requirements” as a factor influencing their decisions, and management confirmed that no official policy mandates ultrasound use. These findings suggest that the local diagnostic practices have evolved into informal customs that some providers feel obligated to follow.

Figure 7:

Responses to the survey question “which test costs more money to your department?”