Eligible patients met the following inclusion criteria: an age between 18–90 years, New York Heart Association (NYHA) functional class II to III, and were scheduled for elective heart valve surgery. Exclusion criteria ruled out individuals with cardiovascular instability, pre-surgical pulmonary infection or severe atelectasis, infective endocarditis, aortic aneurysm, aortic dissection, or previous prehabilitation interventions within the preceding 8 weeks.

Both groups received the usual preoperative education regarding the surgical procedure, pain management, and the use of drains and catheters, and were delivered by a doctor and a cardiac nurse. An informative booklet was also provided to participants.

Patients in the intervention group were administered an additional three-day prehabilitation program, comprising inspiratory muscle training, aerobic muscle training, and specific education (TIME). The TIME program included 12 sessions, totaling 240 minutes.

The inspiratory muscle training (IMT) component involved six 20-minute sessions (twice daily) using a resistance IMT device (Digi IMT X1, XEEX Co., Ltd., Xiamen, China) set at 30% of maximal inspiratory pressure. A 20-minute specific education session, supervised by a physiotherapist, was provided daily following one of the IMT sessions, either in the morning or afternoon, for three days. The specific education was designed to ensure that patients fully understood key concepts, including the prevention of PPCs, deep breathing and coughing techniques, and self-stretching exercises. It also highlighted the importance of preoperative physical activity, reduced daytime bed rest, early mobilization, and self-directed breathing exercises for ICU and cardiac ward stays.

Aerobic muscle training consisted of daily 20-minute walking sessions in the hospital corridor, with intensity adjusted to 60% of heart rate reserve based on the 6-minute walk test at admission.

The TIME program was implemented by a multidisciplinary team including doctors, cardiac nurses, and physiotherapists. The physiotherapist was the primary program lead, while the cardiac nurse facilitated coordination and communication. A doctor supervised patient safety throughout the program.

Patients underwent follow-up within the first 14 days following their surgery, during which the length of ICU stay, any rehospitalizations, and postoperative complications were carefully recorded. The evaluation of PPC incidence was conducted in accordance with the criteria from the Kroenke et al. study [27], where PPCs were diagnosed based on symptoms, classified into four levels. A diagnosis of PPCs was made when patients exhibited either two or more symptoms at grade 2 or at least one symptom at grade 3 or 4 (Supplementary File 1).

Representing the duration of time a patient is in a specific health state, QALYs were estimated by quantifying the area under the curve. A score of zero indicates death, while a score of one represents perfect health [28]. A model was used to project the patient’s QALYs one year after surgery, assuming a two-stage postoperative health trajectory [24]. The first stage, covering the initial 4 weeks post-surgery, was characterized by surgical shock. The second stage, from postoperative week 4 to 52 [29], involved a gradual return to the pre-operative health state. The health state at week 52 was estimated based on baseline levels at hospital admission. The 12-month postoperative QALYs were calculated as the sum of these two stages. Health utility values were derived from the EuroQol five-dimensional questionnaire (EQ-5D) [30], a validated tool for assessing health-related quality of life (HRQoL) in valve surgery patients [31].

All costs were presented in Chinese Yuan (CNY) for the year 2022 (the average exchange rate: 1 US dollar = 6.73 CNY). Since we only audited the costs during the patient’s hospital stay, the discount rate was not utilized.

In the context of China’s healthcare system, auditing the specific costs of each element of the preoperative TIME program presented numerous challenges. For instance, materials for booklets were not directly accounted for, and calculating salaries based on working hours for nurses, doctors, or physiotherapists was complex. Additionally, patients were not required to purchase expensive equipment like IMT devices or monitoring equipment for aerobic training. Instead, these were provided by the hospital and were not included in the cost calculation. The cost of the TIME program was based on session rates: 44 CNY per IMT session, 60 CNY per aerobic muscle training session, and 44 CNY per education session. With 12 sessions in total (6 IMT, 3 aerobic training, and 3 education), the total charge for the TIME program was 600 CNY per participant. Usual care costs were not audited in either group as they followed routine medical pathways.

For postoperative hospital costs, we referenced a study by Boden et al. [24] conducted in Australia and New Zealand. These costs included ICU and cardiac ward stays, respiratory support, chest imaging, electrocardiogram (ECG) monitoring, nursing care, laboratory examinations, and antibiotics. However, there were notable differences in the audited items between China and Australia. For example, in China, medical visit costs were included in hospital ward charges, while nursing care and ECG monitoring were calculated separately. Additionally, antibiotic costs were based on the amount of the specific types of antibiotics that were used, rather than the number of days as in Australia. In this study, we audited only the total costs of antibiotics and pathology due to the diversity of types and tests.

All costs were retrospectively audited by a blinded assessor through the Hospital Information System after patient discharge. All medical interventions performed during hospitalization were decided by their blinded supervising physicians; the physiotherapist was not involved in any medical decision-making.

The cost-effectiveness of the TIME program in preventing PPCs and improving QALYs was evaluated from a hospital perspective using the incremental cost-effectiveness ratio (ICER). The ICER was calculated by dividing the incremental cost ($\mathrm{\Delta }$C) — the net mean hospital cost difference per participant between the TIME and control groups — by the incremental effects ($\mathrm{\Delta }$E), which were the differences in PPCs rates 14 days post-surgery and QALYs 12 months post-surgery.

Next, ICER was utilized to assess the cost-effectiveness of the TIME program in the prevention of PPCs and the enhancement of QALYs from a hospital perspective. The incremental cost ($\mathrm{\Delta }$C) was the difference of net mean hospital costs per participant between the TIME group and the control group. Meanwhile, the incremental effect ($\mathrm{\Delta }$E) was defined as the difference in PPCs rates 14 days post-surgery and QALYs 12 months following surgery. The ICER was calculated by dividing incremental cost by incremental effects (ICER = $\mathrm{\Delta }$C/$\mathrm{\Delta }$E) [32].

Bootstrapping was used to address the large heterogeneity in the postoperative hospitalization costs. Bootstrapping referred to running the economic and clinical outcomes of the original trial through a mathematical model that uses the variance of the original data to simulate hypothetical outcomes after thousands of simulations of the trial. This approach provides a broader representation of potential cost-effective outcomes in a population, rather than relying solely on the specific trial sample. The cost-effectiveness ratios of these simulations were then plotted on a cost-effectiveness plane (Fig. 2a). The plane was divided into four quadrants, with the northeast quadrant representing more costly and more effective; the southeast quadrant representing less costly and more effective (dominance); the southwest quadrant representing less costly and less effective; and the northwest quadrant representing more costly and less effective (dominated) [33].

Fig. 2.

Evaluation of cost-effectiveness for the preoperative TIME program vs. usual care in PPC prevention. Panel (a) of Fig. 2 presents a cost-effectiveness plot comparing the preoperative TIME program with usual care in the prevention of PPCs. Points are colored to distinguish between the patient cohorts: blue for the overall patient population, red for those undergoing TAVR, and green for patients receiving open-heart surgery. The majority of blue points in the southeast quadrant indicate the TIME program’s tendency to be more effective and less costly for the overall cohort. Panel (b) shows the cost-effectiveness acceptability curve for each patient subgroup. The blue line represents the overall patient cohort with a high probability of cost-effectiveness, while the red line indicates a near-certain cost-effectiveness for TAVR patients. The green line illustrates the cost-effectiveness for open-heart surgery patients, which is lower than TAVR but still substantial. These curves represent the probability that the TIME program is cost-effective at different willingness-to-pay thresholds. TIME, three-day of inspiratory muscle training, aerobic muscle training, and specific education; PPCs, postoperative pulmonary complications; TAVR, transcatheter aortic valve replacement; ICER, incremental cost-effectiveness ratio; CNY, Chines Yuan.

To address healthcare providers’ uncertainty about the value of new treatments that are more costly but improve clinical outcomes, a cost-effectiveness acceptability curve (CEAC) was utilized. The CEAC, derived from the joint density of incremental costs and effects, represents the likelihood of the intervention being cost-effective under various payment thresholds [33].

IBM SPSS version 24.0 (IBM Corp., Armonk, NY, USA) was used for data analysis in the RCT. All continuous data were presented as mean (SD) and normality was assessed through the Kolmogorov-Smirnov test. The prevalence of PPCs and pneumonia were examined by the $\chi$² test. Student’s t test was used to examine the difference of costs between groups. All probability values were two-sided, with values 0.05 considered statistically significant. Because of heterogeneity in medical resource usage, differences in costs and effectiveness (PPCs reduction and QALYs) between groups, along with associated 95% confidence intervals, were analyzed using bootstrapping resampling. A total of 5000 paired ICER estimates were bootstrapped and presented on a cost-effectiveness plane. These estimates were further transformed into net benefits to generate the CEAC.

A total of 165 patients were randomized into two groups: the preoperative TIME (intervention group, n = 82) and standard care (control group, n = 83). The detailed baseline characteristics are presented in Table 1. In the TIME group, 45 (55%) were male, with 46 (56%) receiving transcatheter aortic valve replacement (TAVR). Similarly, in the usual care group, 44 (53%) were male, with 51 (61%) undergoing TAVR.

Detailed outcomes of the randomized controlled trial are presented in Table 2. In the intervention group, the incidence of PPCs was 31.7% (26/82), significantly lower than the 53.0% (44/83) observed in control group, demonstrating statistical significance (odds ratio [OR] 0.41; 95% confidence interval [CI] 0.22–0.78). However, there were no significant differences between groups in terms of the converted EQ-5D scores (0.926 versus 0.921, 95% CI –0.019 to 0.033) 4 weeks after surgery or the estimated QALYs (0.909 versus 0.898, 95% CI –0.013 to 0.034) 12 months after surgery.

The detailed costs and use of postoperative hospital resources were presented in Table 3. The mean cost reduction in the intervention group was 1131 CNY (95% CI –2403 to 140) below the control group, with total costs for the intervention group amounting to 8484 CNY (3207) and costs for the usual care reaching 9615 CNY (4869), with no significant difference between the treatments. However, significant cost savings were observed in the intervention group for specific resources: antibiotics (339 versus 667 CNY, 95% CI –605 to –51), nursing (1021 versus 1200 CNY, 95% CI –330 to –28), and ECG monitoring (685 versus 929 CNY, 95% CI –421 to –67), while no significant difference was observed in pathology costs (3825 vs. 4449 CNY, 95% CI –1299 to 51).

In Fig. 2a, the majority of data points (blue) were positioned in the southeast quadrant, illustrating the TIME program’s dual benefits of increased effectiveness and lower costs in preventing PPCs, with a 92.6% probability of cost-effectiveness compared to usual care, as highlighted by the blue line in Fig. 2b. The distribution of data points for TAVR (red) and Open-heart surgery (green) patients shows a similar pattern. The CEAC indicated a 99% probability of cost-effectiveness in TAVR patients and 80% in Open-heart surgery patients (Fig. 2b), showcasing the substantial economic and clinical benefits of the TIME program across different surgical approaches.

In Fig. 3a, it is evident that the scatter points are largely concentrated in the southeast and southwest quadrants, reflecting the uncertainty in QALYs improvement, yet indicating a probable cost reduction associated with the TIME program. The CEAC (Fig. 3b) suggests a 93% probability of the TIME program being cost-effective for QALYs improvement. Subgroup analysis showed a 99% probability of cost-effectiveness in TAVR patients and 82% chance in patients undergoing open-heart surgery, assuming cost neutrality. Due to the uncertainty for improving QALYs, the probability of cost-effectiveness diminishes as the willingness-to-pay values increase (Fig. 3b). Overall, the hospital cost savings was the primary factor driving cost-effectiveness.

Fig. 3.

Evaluation of cost-effectiveness for the preoperative TIME program vs. usual care for QALY enhancement. Panel (a) offers a cost-utility plot that compares the preoperative TIME program with usual care in improving QALYs, with scatter points distributed mainly in the southeast and southwest quadrants. This distribution suggests potential cost savings despite uncertain QALY gains, with blue representing the complete patient set. Panel (b) displays the cost-effectiveness acceptability curve for the TIME program, where the blue line signals a 93% probability of cost-effectiveness for the overall group. For subgroup analysis, red denotes TAVR patients, with a 99% probability of cost-effectiveness, and green represents open-heart surgery patients, indicating an 82% probability. These findings underscore the economic and health-related benefits of the TIME program across different surgical interventions. TIME, three-day of inspiratory muscle training, aerobic muscle training, and specific education; TAVR, transcatheter aortic valve replacement; QALYs, quality adjusted life years; ICER, incremental cost-effectiveness ratio; CNY, Chinese Yuan.