Cost-effectiveness of follow-up algorithms for chronic thromboembolic pulmonary hypertension in pulmonary embolism survivors

Abstract

Introduction: Achieving an early diagnosis of chronic thromboembolic pulmonary hypertension (CTEPH) in pulmonary embolism (PE) survivors results in better quality of life and survival. Importantly, dedicated follow-up strategies to achieve an earlier CTEPH diagnosis involve costs that were not explicitly incorporated in the models assessing their cost-effectiveness. We performed an economic evaluation of 11 distinct PE follow-up algorithms to determine which should be preferred.

Materials and methods: 11 different PE follow-up algorithms and one hypothetical scenario without a dedicated CTEPH follow-up algorithm were included in a Markov model. Diagnostic accuracy of consecutive tests was estimated from patient-level data of the InShape II study (n=424). The lifelong costs per CTEPH patient were compared and related to quality-adjusted life-years (QALYs) for each scenario.

Results: Compared to not performing dedicated follow-up, the integrated follow-up algorithms are associated with an estimated increase of 0.89-1.2 QALYs against an incremental cost-effectiveness ratio (ICER) of EUR 25 700-46 300 per QALY per CTEPH patient. When comparing different algorithms with each other, the maximum differences were 0.27 QALYs and EUR 27 600. The most cost-effective algorithm was the InShape IV algorithm, with an ICER of EUR 26 700 per QALY compared to the next best algorithm.

Conclusion: Subjecting all PE survivors to any of the currently established dedicated follow-up algorithms to detect CTEPH is cost-effective and preferred above not performing a dedicated follow-up, evaluated against the Dutch acceptability threshold of EUR 50 000 per QALY. The model can be used to identify the locally preferred algorithm from an economical point-of-view within local logistical possibilities.

FIGURE 1

The 11 scenarios evaluating different follow-up algorithms to detect chronic thromboembolic pulmonary hypertension (CTEPH) in acute pulmonary emoblism (PE) patients. a) Algorithms identified using a literature search; b) hypothetical algorithms in which diagnostic tests are performed in all acute PE patients; and c) hypothetical algorithms in which diagnostic tests are only performed in patients that remain symptomatic during follow-up. A green line presents a “positive result” and a red line presents a “negative result” of the diagnostic test. All algorithms start ∼3 months after an index PE diagnosis. Assumptions made in the evaluation of the algorithms: if available, we used the echocardiogram performed as a part of the InShape II algorithm to calculate diagnostic accuracy of the echocardiogram. For patients who were not subjected to an echocardiogram as part of the InShape II algorithm, we used the 2-year follow-up echocardiogram as a surrogate outcome. The algorithms by Lewczuk et al. [15] (A3) used repeating measurement of diagnostic tests at different time-points in case of a negative test result in the preceding test. The InShape II study did not perform all diagnostic tests in all patients at different time-points. We therefore modified the algorithm of Lewczuk et al. [15] to fit a one time-point approach which could be evaluated using the InShape II cohort. We had no data on the sensitivity and specificity of ventilation/perfusion (VQ) scans or on echocardiography in patients with a positive VQ scan in the InShape II study. Taking CTEPH incidence after acute PE [8] and the rate of positive scans after acute PE into account (Cimini et al. [26]), we made the assumption that the false-positive rate for VQ scan performed ∼3 months after acute PE diagnosed would be 36%, meaning that 36% of the non-CTEPH patients were assumed to have persistent perfusion defects on VQ scan after 3 months [37, 38]. We had no data on the false-positive rate for echocardiography in patients with persistent perfusion defects on VQ scan. Prevalence of an estimated pulmonary artery pressure of >30 mmHg is estimated to be 25–48% [27]. Therefore, we assumed that 37.5% of the non-CTEPH patients with an abnormal VQ scan would have an intermediate–high risk of pulmonary hypertension on echocardiography. CPET: cardiopulmonary exercise test; CTPA: computed tomography pulmonary angiogram; ESC: European Society of Cardiology; NT-proBNP: N-terminal pro-brain natriuretic peptide; RHC: right heart catheterisation; RVD: right ventricular dysfunction; TTE: trans-thoracic echocardiogram.

FIGURE 2

Flowchart of dedicated follow-up, (additional) diagnostic and treatment costs of each scenario. a) Scenario 0: no structural follow-up for chronic thromboembolic pulmonary hypertension (CTEPH). b) Scenarios in which a dedicated follow-up algorithm is applied. As right heart catheterisation (RHC) is the golden standard for CTEPH diagnosis, there were considered to be no false-positive CTEPH diagnoses. However, a false-negative result from a diagnostic test prior to RHC is possible; therefore a temporarily false-negative CTEPH diagnosis is included in our model (b). PE: pulmonary embolism; VQ: ventilation/perfusion.

FIGURE 3

Lifelong healthcare costs per chronic thromboembolic pulmonary hypertension patient (in euros) for the 12 scenarios. RHC: right heart catheterisation; VQ: ventilation/perfusion; TTE: trans-thoracic echocardiogram; ESC: European Society of Cardiology.

FIGURE 4

Total lifelong healthcare costs and quality-adjusted life-years (QALYs) per chronic thromboembolic pulmonary hypertension (CTEPH) patient. A scenario is dominated if there is another algorithm that is at least as good on both costs and QALYs, and strictly better on at least one of costs and QALYs. A scenario is weakly dominated if it is dominated by a mixture of two other algorithms. The scenarios that are not (weakly) dominated by others are potentially cost-effective and together form the efficient frontier (black line). 0: no dedicated CTEPH follow-up; A1: European Society of Cardiology; A2: InShape II; A3: Lewczuk et al. [15]; A4: Held et al. [16]; A5: InShape IV; B1: trans-thoracic echocardiogram (TTE)-ventilation/perfusion scan (VQ)-right heart catheterisation (RHC); B2: VQ-TTE-RHC; B3: VQ-RHC; C1: symptom-TTE-VQ-RHC; C2: symptom-VQ-TTE-RHC; C3: symptom-VQ-RHC.