Ji-Hao Shi, Liang Luo, Xiao-Li Chen, Yi-Peng Pan, Zhou Zhang, Hao Fang, Ying Chen, Wen-Dong Chen, Qian Cao

Abstract

Key Words: Infliximab; Crohn’s disease; Maintenance therapy; Cost-effectiveness; Outcomes; Direct medical costs

INTRODUCTION

Crohn’s disease (CD) is a chronic disease that can affect any part of the gastrointestinal tract[1]. Even though the causes of CD have not yet been fully clarified, current research has proven that tumor necrosis factor-α (TNF-α) is the proinflammatory cytokine enhancing leukocyte migration, activating leukocytes, inducing acute-phase reactants and metalloproteinases, and inhibiting apoptosis of inflammatory cells in CD patients[2-4]. As the first developed TNF-α blocker, infliximab has been shown to be effective across the spectrum of CD, including refractory luminal CD, steroiddependent CD, and refractory fistulizing CD. The ACCENT 1 trial demonstrated the clinical benefits of infliximab used as a maintenance therapy. In this trial, there was a significantly higher clinical remission rate, a higher mucosal healing rate, and a lower hospitalization rate associated with 1-year infliximab maintenance therapy (IMT) when compared to placebo[5].

Using a conservative estimate, CD affects at least 200000 patients across China[6]. Similar to CD patients in western countries, Chinese CD patients are relatively young, and their quality of life and social function are significantly impaired by CD. Even though infliximab was approved to treat CD in China shortly after its launch in 2005, access to infliximab in Chinese CD patients was highly limited due to the lack of reimbursement coverage. Thus, anti-inflammatory drugs and immunosuppressants are still the mainstay treatment for CD in China, and the limited clinical benefits and side effects associated with these drugs are still the main concern regarding the utilization of these drugs for moderate to severe Crohn’s disease (MS-CD). The purpose of this study was to clarify the real-world health outcomes, medical costs, and cost-effectiveness associated with IMT for MS-CD in Chinese patients and generate real-world evidence to support reimbursement decision making related to the treatments for MS-CD in China.

MATERIALS AND METHODS

This study consisted of a real-world study and cost-effectiveness analysis comparing IMT and conventional maintenance therapy (CMT) for health outcomes and direct medical costs in a clinical cohort of MS-CD patients. Research ethics approval for this study was obtained from Sir Run Run Shaw Hospital, Hangzhou, China.

Real-world study

This study identified MS-CD patients who visited the Inflammatory Bowel Disease clinic of Sir Run Run Shaw Hospital in two time windows: January 1, 2014 to December 31, 2014 and July 1, 2017, to June 30, 2018. This study included all patients with a diagnosis of MS-CD who received maintenance therapy in Sir Run Run Shaw Hospital. To minimize the risk of selection bias, this study only excluded patients with insufficient information for data analysis. Hospital medical records associated with the included patients during the 1-year observation time period were reviewed to extract patient demographics including age, gender, body mass index, socio-economic status (employment, residence, and marital status),lifestyle (smoking and drinking), disease site, history of CD-related surgery, CD-related complications, extraintestinal manifestations, and comorbidities. The prescription records associated with the included patients during the 1-year observation period were the data source for the therapy pattern. The documented telephone follow-up questionnaires of the identified patients from the time window from July 1, 2017 to June 30, 2018were the data sources to assess disease activity using Harvey-Bradshaw Index and rate quality of life on a0 to 100 scale (0 indicated the worst health status, and 100 indicated the best health). The measured disease activity and quality of life associated with the followed-up patients were used to develop the prediction formulas from the multiple linear regression analyses that used patient characteristics and treatment pattern as independent variables. The developed prediction formulas for disease activity and utility for quality of life were used to estimate the disease activity and quality of life associated with the identified patients from the time window between January 1, 2014 and December 31, 2014. The billing records associated with the included patients’ outpatient clinic visits and hospitalizations in Sir Run Run Shaw Hospital during the 1-year observation period were used to extract the health resources utilization (outpatient visits, hospital admissions, and hospital stay length) and direct medical costs.

This study stratified the included patients into two groups for the data analysis. The included patients receiving infliximab-contained maintenance therapy were assigned into the IMT group. The other included patients receiving maintenance therapy without containing infliximab were assigned into CMT group. The patient baseline characteristics associated with the two groups were summarized using descriptive statistical methods. Studentttest, chi square test, and Wilcoxon rank sum test were used to compare the two groups for their patient characteristics and measured outcomes, which included disease remission, quality of life, health resources utilization, and direct medical costs over 1-year observation time. To adjust the potential confounding effects associated with patient baseline characteristics,this study conducted multivariable conventional regression analyses, including logistic regression analysis, linear regression analysis, Poisson regression analysis, betabinomial regression analysis, and generalized linear regression analysis, with adjustment of patient baseline characteristics to compare IMTvsCMT for disease remission (defined as Harvey-Bradshaw Index score < 5)[7], utility for quality of life, health resources utilization, and direct medical costs. The statistical significance in these analyses was defined as the two-sidedPvalue less than 0.05.

Cost-effectiveness analysis

A decision-analytic model was constructed to simulate health outcomes and direct medical costs associated with two model maintenance therapy scenarios: IMTvsCMT. For each model scenario, the decision-analytic model used a Markov model design to simulate treatment cycles between induction therapy and maintenance therapy for treatment response, disease relapse, surgery, and mortality associated with MS-CD patients. The decision analytic model defined the induction therapy as any treatments used with the goal to achieve disease remission in the MS-CD patients who were relapsed from maintenance therapies, which were regularly given to patients to maintain disease remission after induction therapy. The introduction therapies used in real-world study cohort, including steroids (55.2%), infliximab monotherapy (14.3%), a combination of infliximab and immunosuppressant (9.7%) or enteral nutrition (11.7%), and enteral nutrition monotherapy (9.1%),were applied to the decision analytic model to simulate the distribution of introduction therapies in the model cohort. The identified maintenance therapies in the IMT group and CMT group from the realworld study were used to simulate the distributions of IMT and CMT in the model cohort. The administration ofinfliximab as introduction therapy and maintenance therapy in the real-world study cohort was based on theproduct monograph of infliximab for MS-CD (5 mg/kg administrated at 0, 2, and 6 wk for introduction therapy, subsequent administration using the same treatment dosage every 8 wk for maintenance therapy).The constructed decision-analytic model allowed patients to receive repeatedly induction therapy and maintenance therapy after disease relapse. The model also assumed that the surgical treatment for the complications only occurred in patients with active CD. The post-surgery patients entered another treatment cycle including induction and maintenance therapy until the occurrence of another surgical treatment in the model.

The decision-analytic model took into account the mortality associated with surgical treatment, disease remission, and active disease to estimate the survival rate associated with each model cycle. The cycle length of the Markov models in the decision-analytic model was 3 mo to align with the patients’ regular follow-up frequency. The simulation time horizon in the decision-analytic model was set to lifetime to estimate overall survival, quality-adjusted life years (QALY), cumulative risk of CD-related surgery, and reimbursed medical costs under the reimbursement policy for CD patients in Zhejiang province (annual co-payment: ￥20000; reimbursement percentage: 80%; annual reimbursement cap:￥40000). The structure of the decision-analytic model is illustrated inFigure 1.

A literature review was conducted to estimate model variables for the treatment response associated with induction therapy[8-13], risk of disease relapse associated with maintenance therapy[14], surgery risk associated with active CD, perioperative mortality associated with surgery, and the hazard ratio of mortality associated with active CD relative to age and gender-matched general population[15-30]. Meta-analysis was used as the main approach to synthesize the identified evidence from the literature review. The constructed decision analytic model was used to conduct costeffectiveness analysis, which included base-case analysis, one-way sensitivity analyses, and probabilistic sensitivity analysis (PSA). The point estimates of QALY gains and lifetime reimbursed medical costs from the base-case analysis were used to calculate the incremental cost-effectiveness ratio (ICER) per gained QALY associated with IMT relative to CMT. One-way sensitivity analyses assessed the change of ICER associated with IMT by varying each model variable within its 95% confidence interval (CI) or ± 25% of its baseline value. PSA was conducted using a Monte Carlo simulation method to run 10000 iterations of cost-effectiveness analyses based on the distributions of model variables (beta distributions for probability and utility variables; gamma distributions for cost variables). The cost-effectiveness proportion associated with IMT relative to CMT was calculated under the willingness-to-pay (WTP) of one, two, and three times of the 2018 Chinese gross domestic goods per capita (GDPPC) (￥64644 or $9769)[31], respectively.

RESULTS

The real-world study initially identified 593 MS-CD patients. Of the identified MS-CD patients managed in the study hospital, 393 patients received maintenance therapy for MS-CD. After further exclusion of 4 patients without sufficient information for data analysis, this study eventually included 389 patients to conduct the data analysis. Of the included 389 patients, 259 patients received IMT, including the combination of infliximab and immunosuppressant (38.2%), infliximab monotherapy (44.6%), the combination of infliximab and enteral nutrition (12.4%), and the combination of infliximab, immunosuppressant, and enteral nutrition (4.8%). The other 130 patients received CMT that included immunosuppressant (61.7%), 5-aminosalicylates (25.8%), enteral nutrition (7.0%), the combination of immunosuppressant and enteral nutrition (3.1%), and the combination of immunosuppressant, 5-aminosalicylates, and/or enteral nutrition (2.4%).

Figure 1 Structure of the decision analytic model assessing the cost-effectiveness of infliximab maintenance therapy relative to conventional maintenance therapy for moderate to severe Crohn’s disease in China. CMT: Conventional maintenance therapy; IMT: Infliximab maintenance therapy; MS-CD: Moderate to severe Crohn’s disease.

Real-world outcomes associated with IMT and CMT

The comparisons of the patient baseline characteristics associated with the two study groups for IMTvsCMT identified significantly younger age (34.1 ± 10.9 yearsvs37.2 ± 11.2 years,P= 0.004), lower rate of married patients (56.0%vs67.7%,P= 0.026), higher unemployment rate (6.9%vs1.5%,P= 0.023), higher proportion of disease site at ileocolon (52.5%vs33.8%,P< 0.001), higher proportions of historical complications for anus fistula (36.3%vs14.6%,P< 0.001), perianal abscess (27.4%vs13.8%,P= 0.003), and intestinal fistula (10.4%vs3.1%,P= 0.012); and lower proportions of comorbidities including chronic hepatitis B (3.1%vs8.5%,P= 0.020), gastroenteritis (1.5%vs4.6%,P= 0.071), and kidney diseases (0.4%vs3.8%,P= 0.009) in the IMT group. The patient baseline characteristics associated with the two study groups are summarized in Table 1.

The unadjusted comparisons of the measured clinical outcomes, health resources utilization, and hospital costs associated with the two created study groups for IMTvsCMT from the included 389 patients are summarized inTable 2. The multivariate regression analyses with the adjustment of patient demographics, social economic status, disease site at diagnosis, history of CD-related complications, history of CDrelated surgery, and extraintestinal manifestation at baseline confirmed that IMT was associated with significantly higher disease remission chance [odds ratio (OR): 4.060, 95%CI: 1.643 to 10.753,P= 0.003], lower risk of developing any new complications (OR: 0.527, 95%CI: 0.323 to 0.858,P= 0.010), and higher utility value for quality of life (coefficient: 0.822, 95%CI: 0.218 to 1.426,P= 0.008) than CMT; IMT was associated with significantly lower outpatient clinic visits (coefficient: -0.564, 95%CI: -0.703 to -0.425) and shorter hospital stay length related to active disease management (coefficient: -4.725, 95%CI: -7.112 to -2.337,P< 0.001) than CMT; and IMT was associated with significantly lower outpatient costs (coefficient: -1.248, 95%CI: -1.651 to -0.850) and total medical costs related to active disease management (coefficient: -0.378, 95%CI: -0.659 to -0.101,P= 0.008) than CMT. The results of multivariate regression analyses are summarized by clinical outcomes (Table 3), health resources utilizations (Table 4), and medical costs (Table 5).

Cost-effectiveness of IMT relative to CMT

Based on the model variables that are summarized inTable 6, the comparisons of the point estimates of the model outputs associated with two model scenarios in the base case analysis without discounting the measured outcomes estimated that the IMT model scenario was associated with the increase of overall survival by 2.871 years (43.815 yearsvs40.944 years), QALY by 2.476 years (33.365 QALYvs30.889 QALY), and reimbursed medical costs by ￥96201 (￥ 469958vs￥373757). The cumulative CDrelated surgery risk associated with the IMT model scenario was reduced by 39.7%. The discounted point estimates of QALYs and reimbursed medical costs associated with the two model scenarios in the base-case analysis estimated that the ICERassociated with the IMT scenario relative to the CMT scenario was ￥55260, 85.5% of the 2018 Chinese GDPPC. The results of the base case analysis before and after discounting are summarized inTable 7.

Table 1 Patient characteristics associated with the included moderate to severe Crohn’s disease patients in real-world study

Lung nodes 3.9%2.3%0.421 Gastroenteritis 1.5%4.6%0.071 aP < 0.05.bP < 0.01. BMI: Body mass index; CMT: Conventional maintenance therapy; IMT: Infliximab maintenance therapy; SD: Standard deviation.

One-way sensitivity analyses indicated that the cost-effectiveness of IMT could be more attractive, indicated by the reduction of ICER over ￥20000, as shown by increasing the following model variables: Quality of life associated with disease remission, relapse risk associated with CMT, treatment discontinuation risk associated with IMT, relapse risk after treatment discontinuation, and hazard ratio of mortality associated with active disease relative to the general population. The cost-effectiveness of IMT was less attractive, indicated by the increase of ICER over ￥20000, when increasing the following model variables: Treatment response of induction therapy with enteral nutrition, steroids plus 5-aminosalicylates, and steroids alone, distribution of induction therapy using steroids, disease relapse risk after the discontinuation of IMT, quality of life associated with active disease, and unit price of infliximab. The impacts of these key model variables on the cost-effectiveness of IMT relative to CMT are illustrated inFigure 2.

The 10000 generated ICER values associated with IMT from the Monte Carlo simulations were ranked to identify the median ICER (￥68512) and its 95% credible interval (￥-238869 to ￥601293). The cost-effectiveness proportions of IMT relative to CMT under the WTP of one, two, and three times of the 2018 Chinese GDPPC were 47.6%, 74.7%, and 86.4%, respectively.

DISCUSSION

This study observed that IMT was highly effective in a real-world setting by achieving a 94.6% disease remission rate. This is much higher than the reported disease remission rate of infliximab in randomized trials[9,32,33], which reported about 60% disease remission rate associated with 1-year infliximab monotherapy in MS-CD patients. Because the IMT group consisted of approximately 60% of patients receiving a combination of infliximab and immunosuppressants or enteral nutrition in the realworld study, the MT containing infliximab and traditional treatments could be more effective in MS-CD patients. The superior treatment effects associated with the combination of infliximab and immunosuppressants for MS-CD have been proven in randomized clinical trials. However, the reported disease remission rate in these randomized trials was not as high as what was observed in this study. Because the treatment efficacies of infliximab were mainly assessed in randomized clinical trials conducted in western countries, the observed treatment effects of infliximab in this study might suggest that patient ethnicity might play a role in the treatment effects of infliximab. This speculation was supported by another retrospective study that observed nearly the same disease remission rate (97.1%) associated with 1-year treatment with infliximab for MS-CD in Korean patients[34]. Additionally, male gender was found to predict better treatment response of infliximab[35], and the high male proportion in the patient cohort in this study could further increase the disease remission rate. Thus, IMT could gain more clinical benefits and have more attractive cost-effectiveness in Chinese MS-CD patients.

Similar to previous studies reporting reduced health resource utilization associated with infliximab in CD patients, this study confirmed that the high disease remission rate associated with IMT reduced health resources utilization related to active disease management during the 1-year observation period. According to the multivariate regression analysis, IMT significantly reduced outpatient clinic visits and hospital stay days related to active disease management. These impacts on health resources utilization could save direct medical costs and partially offset the high drug acquisition costs of infliximab. However, the drug acquisition costs of infliximab were much higher than conventional medications used for CMT. The overall direct medical costs associated with IMT was about four times of the direct medical costs associated with CMT. Because the cost-effectiveness was assessed by ICER, which is the ratiobetween the difference in lifetime medical costs and difference in QALY associated with IMT and CMT, the drug acquisition costs of infliximab were likely to be the main driving factor for the cost-effectiveness of IMT for MS-CD in China.

Table 2 Unadjusted comparisons of the measured clinical outcomes, health resources utilization, and direct medical costs associated with infliximab maintenance therapy and conventional maintenance therapy during 1-year observation time

This study constructed a comprehensive decision-analytic model that fully accounted for the induction and maintenance treatment cycles, surgery related to developed complications, and mortality risk related to disease status and surgery tosimulate lifetime health outcomes and reimbursed medical costs associated with IMT and CMT. This study leveraged the generated evidence from the real-world study for the estimation of the model variables to maximize the generalizability of the costeffectiveness analysis. Consistent with real-world studies with long-term follow-ups and the cost-effectiveness analyses assessing IMT for MS-CD in high-income countries[36], the constructed decision-analytic model confirmed that IMT could gain more clinical and health benefits than CMT by increasing overall survival, increasing QALY, and reducing the risk of surgery for CD-related complications. Additionally, IMT was only associated with a modest increase of reimbursed medical costs under current Chinese reimbursement policy. In this case, the cost-effectiveness of IMT relative to CMT for MS-CD in China was highly attractive by having the ICER value less than the 2018 Chinese GDPPC. This result also suggested that the reimbursement coverage in Chinese patients was unlikely to substantially reduce the out-of-pocket costs associated with the disease management. Thus, the affordability of IMT could be still a significant barrier for patient access to infliximab even with reimbursement support. Since the cost-effectiveness of IMT was highly sensitive to the price of infliximab, it might be beneficial to use our constructed decision-analytic model to identify further the appropriate price of infliximab and reimbursement policy to improve patient access to IMT.

Table 3 Summary of the multivariable regression analyses for clinical outcomes in the included moderate to severe Crohn’s disease patients

aP < 0.05.bP < 0.01. BMI: Body mass index; CD: Crohn’s disease; CI: Confidence interval; CMT: Conventional maintenance therapy; IMT: Infliximab maintenance therapy; OR: Odds ratio; SD: Standard deviation.

This study conducted one-way sensitivity analysis and probability sensitivity analysis to assess the impact of uncertainty associated with the model variables on the cost-effectiveness of IMT relative to CMT in MS-CD patients. The one-way sensitivity analyses clearly demonstrated that quality of life, measured as utility in our study, associated with disease remission and active disease, could substantially change the ICER due to their wide 95%CIs. Thus, the validity of the utility associated with disease remission and active disease in our study was critical for the robustness of our costeffectiveness analysis. Because the estimated utilities for disease remission and active disease were highly comparable as previously reported results of a meta-analysis[37]based on 17 studies (utility for disease remission: 0.829vs0.840; utility for active disease: 0.743vs0.753), the utility variables in cost-effectiveness analysis should have sufficient external validity. Our PSA took into account overall uncertainty associated with utility variables and also other model variable to estimate the distribution of thecost-effectiveness of IMT relative to CMT under the 10000 Monte Carlo simulations. Our base-case analysis indicated that IMT was highly cost-effective by having an ICER less than 2018 Chinese GDPPC (85.5%). Our PSA estimated that 47.6% of simulated ICERs less than 2018 Chinese GDPPC. Thus, base-case analysis was likely to overestimate the cost-effectiveness of IMT. As the cost-effectiveness proportion associated with IMT relative to CMT was 86.4% under the recommended costeffectiveness threshold, both base case analysis and PSA supported the attractive costeffectiveness of IMT in Chinese MS-CD patients.

Table 4 Summary of the multivariable regression analyses for health resources utilization in the included moderate to severe Crohn’s disease patients

aP < 0.05.bP < 0.01. BMI: Body mass index; CD: Crohn’s disease; CI: Confidence interval; CMT: Conventional maintenance therapy; IMT: Infliximab maintenance therapy; NA: Not available; NS: Not significant.

Except infliximab, the other launched TNF-alpha inhibitors, such as etanercept and adalimumab, were launched in China as well. However, the approved indications of etanercept and adalimumab did not include MS-CD when this study was conducted. The other biologics indicated for MS-CD, including vedolizumab and ustekinumab, were recently launched in China. Thus, our cost-effectiveness analysis did not includethese biologic treatments. Even though the maintenance therapy with these newly approved biologics were reported to have a higher disease remission rate than IMT, the higher acquisition costs associated with these biologics could make their costeffectiveness relative to IMT unlikely attractive in MS-CD patients. Thus, the newly approved biologics are mainly recommended in the second-line treatment setting after the failure with infliximab treatment.

Table 5 Summary of the multivariable regression analyses for direct medical costs in the included moderate to severe Crohn’s disease patients

aP < 0.05.bP < 0.01. BMI: Body mass index; CD: Crohn’s disease; CI: Confidence interval; CMT: Conventional maintenance therapy; IMT: Infliximab maintenance therapy; NA: Not available; NS: Not significant.

Even though the cost-effectiveness analysis based on the real-world data minimized the uncertainty and variability associated with the model variables, the real-world observation period was only 1 year, which was not sufficiently long to assess the impact of IMT on long-term clinical outcomes, such as the development of complications, surgeries, and mortality. The predictions of these long-term clinical outcomes in the cost-effectiveness analysis were based on literature evidence. Thus, the generalizability of the cost-effectiveness analysis needs further improvement by future real-world studies assessing these long-term outcomes associated with IMT in Chinese patients with MS-CD. Another main limitation in this study was the small sample size of the study cohort from one tertiary care hospital. The study cohort might not be large enough to represent fully the MS-CD patients across China. As the incidence rate of CD in China was as low as 0.46/1000000[38], it is challengeable to identify a large cohort of MS-CD patients from a single center. However, our study cohort had comparable patient baseline characteristics as the Chinese MS-CD patientsin other observational studies[39,40].Similar to the Chinese MS-CD patients in previously published observational studies, our study cohort was characterized by younger age, more male patients, higher proportion with disease site at colon, and one-third patients with history of surgery for CD-related complications.

Table 6 The summary of the main model variables in the decision analytic model assessing the cost-effectiveness of infliximab maintenance therapy relative to conventional maintenance therapy for moderate to severe Crohn’s disease in China

CONCLUSION

In summary, this study confirmed that IMT was superior to CMT regarding disease remission rate, quality of life, and health resources utilization in real-world Chinese patients with MS-CD. The extremely high disease remission rate associated with IMT suggested that Chinese patients might have better treatment response to infliximab. Based on the generated real-world evidence, the cost-effectiveness of IMT relative to CMT was highly attractive as IMT cost the Chinese public health insurance payers less than the 2018 Chinese GDPPC to gain one QALY in Chinese MS-CD patients.

Table 7 Summary of the results of undiscounted and discounted point estimations of measured outcomes in base case analysis comparing infliximab maintenance therapy vs conventional maintenance therapy in the constructed decision analytic model

Figure 2 Impact of key model variables on the cost-effectiveness of infliximab maintenance therapy relative to conventional maintenance therapy for moderate to severe Crohn’s disease in one-way sensitivity analyses. CD: Crohn’s disease; CMT: Conventional maintenance therapy; ICER: Incremental cost-effectiveness ratio; IMT: Infliximab maintenance therapy.

ARTICLE HIGHLIGHTS

Research conclusions

IMT was confirmed to be superior to CMT in Chinese real-world MS-CD patients. With the overall uncertainty associated with clinical effectiveness, quality of life, and direct medical costs associated with IMT and CMT in Chinese MS-CD patients, the cost-effectiveness of IMT relative to CMT was attractive from the perspective of Chinese health care payers.

Research perspectives

This study only followed up a relatively small cohort with MS-CD patients from a single treatment center. The generalizability associated with generated evidence in this study needs confirmation by future studies with large sample size of patients enrolled from more treatment centers. Additionally, this study followed up MS-CD patients for only 1 year. Future studies are needed to follow up patients longer to assess the impact of IMT on long-term clinical outcomes, which should include survival outcomes and CD-related to surgery and complications.

ACKNOWLEDGEMENTS

We want to thank Professor Krahn M from THETA Collaborative at the University of Toronto to help with reviewing the statistical methods and proofreading the manuscript of this study.