Research Report Quality Assessment Results

1. Title

2. Summary table

3. Introduction

4. Sample characteristics

5. Location

6. Research perspective

7. Comparison

8. Research period

9. discount

10. Effective selection

11b. Measure treatment effectiveness based on total… 11a. Measure treatment effectiveness based on research…

12. Measure and calculate the effectiveness of choices 13b. Estimate resources and costs based on the model…

13a. Estimate resources and costs based on…

14. Monetary value, time value and conversion

15. Model selection

16. Assumptions

17. Analysis method

18. Research parameters

19. incremental costs and benefits 20b. 20a. Uncertainty: economic assessment…

20a. Uncertainty: economic valuation based on…

21. Heterogeneity

22. opinions, research limitations, generalizations and…

23. Funding sources

24. Conflict of interest

Have

Are not

0 50 100

Figure 1.11. Results of research report quality assessment

Similarly, based on the decision analysis model quality assessment table consisting of 57 questions developed by Philips et al. in 2004 [119], the results showed that the quality of decision analysis models was quite high, from 57-67%, in which, the study of author Rochau [125], [126] and the study of author Li N.

[101] met the criteria for a higher quality model than M.Romero [127]. In 15 sub-items with 57 evaluation questions, the average number of questions answered “yes” in 4 studies was 35/57. Some items such as structural origin, model type, model running time, disease status and model cycle were fully recorded in all studies, while the item assessing uncertainty when building the model was not met by all studies and many evaluation questions could not find information (Figure 1.12).

Question D2A.3 D2A.2 D2A.1

D2.1 D1.5 D1.4 D1.3 D1.2 D1.1 S9.1 S8.1 S7.2 S7.1 S6.1 S5.3 S5.2 S5.1 S4.2 S4.1 S3. 3 S3.2 S3.1 S2.4 S2.3 S2.2 S2.1 S1.3 S1.2 S1.1

Question C2.3 C2.2

C2.1

C1.1 D4D.2 D4D.1 D4C.1 D4B.1 D4A.1 D4.2 D4.1 D3.5 D3.4 D3.3 D3.2 D3.1 D2D.3 D2D.2 D2D.1 D2C. 3 D2C.2 D2C.1 D2B.5 D2B.4 D2B.3 D2B.2 D2B.1 D2A.4

0	50	100	0	50	100

Maybe you are interested!

Yes No No information available

Figure 1.12. Results of decision model quality assessment

Chapter 2. RESEARCH OBJECTS AND METHODS

2.1 Research subjects

Research subjects:

- Intervention protocol: cost and effectiveness of NL (Tasigna®) regimen 300mg x 2 times/day for first-line treatment of BCMDT patients,

- Comparative regimen: cost and effectiveness of IM regimen (Glivec®) dose 400mg/day for first-line treatment of BCMDT patients with/without the addition of NL (Tasigna®) dose 300mg x 2 times/day after BCMDT patients are resistant or intolerant to IM.

Observation objects: BCMDT medical records; list of outpatient medical examination and treatment costs form number BV01 and list of inpatient medical examination and treatment costs form number BV02.

Data collection location: National Institute of Hematology and Blood Transfusion and Ho Chi Minh City Blood Transfusion and Hematology Hospital.

Project implementation time: October 2017 – December 2019.

2.2. Research method

2.2.1. Research design

The study used a cost-effectiveness assessment method using modeling to compare the cost-effectiveness of IM and NL in the first-line treatment of BCMDT. The model parameters were determined using a systematic review and meta-analysis method to collect secondary data from published studies, and a cross-sectional descriptive method was used to collect primary data on treatment costs and treatment effectiveness for each disease state in the model. In addition, a sensitivity analysis method was used to analyze the uncertainty in the study results.

The steps and methods of conducting the research are presented in Figure 2.1.

Objective 1. Model Setup

Characteristic

Model parameters

 Model structure

 Model input population characteristics

 Cycle and model run time

Probability of disease state transition in tissue

image

Treatment costs for each condition

disease

Treatment effectiveness of each disease state

 System overview

 Meta-analysis

 Cross-sectional description

 System overview

Determine the scope of the study

 Study population: BCMDT patients aged 18 years and older in chronic, accelerated, and acute phases.

 Comparator drugs: IM and NL in first-line treatment

 Analytical perspective: Society and

paying agency

 Model selection: Markov

 Effective choice: QALY

 Analysis index: CER, ICER

Objective 2. Cost-effectiveness analysis

Manure

base area

Comparison with other studies

Uncertainty analysis

 Analyze costs, effectiveness, CER, ICER of each regimen.

 Compare with payment threshold

 One-way sensitivity analysis

 Multidimensional sensitivity analysis

 Probabilistic sensitivity analysis

 System overview

 Convert ICERs based on CPI index

Figure 2.1. Research steps and methods

2.2.2. Determine the scope of the research

2.2.2.1. Study population

The Markov model was used to analyze a hypothetical population of newly diagnosed patients with CKD aged 18 years and older who were receiving first-line treatment with IM or NL drugs.

2.2.2.2. Interventions evaluated

The NL regimen (Tasigna®) at a dose of 600 mg/day is proposed to replace the IM regimen (Glivec®) at a dose of 400 mg/day in the first-line treatment of BCMDT patients, so that the cost-effectiveness of these two regimens is compared. In case of drug resistance or intolerance, chemotherapy or bone marrow transplantation will be indicated, without dose escalation or treatment with another TKI.

In addition, the NL step 1 regimen was also compared with the IM step 1 regimen with additional NL treatment at a dose of 600 mg/day if the patient was resistant or intolerant to IM. Similarly, the patient was assigned to chemotherapy or bone marrow transplantation if the drug treatment failed.

Imatinib

Chemotherapy or bone marrow transplant

Imatinib

Nilotinib

Chemotherapy or bone marrow transplant

The comparison protocols are presented in Figure 2.2 Protocol 1.

Protocol 2.

Nilotinib

Chemotherapy or bone marrow transplant

Protocol 3.

Figure 2. 2. Evaluated regimens

2.2.2.3. Research perspective

In order to have a comprehensive view of the economic burden of treatment from the perspective of the whole society, as well as to provide adequate information and authentic evidence in supporting managers to consider making payment policies and developing new treatment guidelines, the study was conducted from the perspective of society and the perspective of the payer agency.

2.2.3. Model characteristics

2.2.3.1. Model structure

Due to the limitations of RCT studies, a Markov decision analysis model was developed to evaluate the cost-effectiveness of the regimens. Following the model building guidelines and the results of the systematic review conducted in section

1.4 [58], the Markov decision model for patients with BCMDT in the study will include three basic disease states corresponding to three stages of disease progression: chronic stage, accelerated stage and acute transition stage. In the chronic stage, patients will include different health states corresponding to the current treatment direction in Vietnam. After being built, the model will be consulted with clinical experts at VHHTMTW and HCM City Hospital of ENT to assess its suitability in the context of Vietnam.

2.2.3.2. Patient characteristics at model initiation

Determining the characteristics of patients entering treatment or entering the model is important because it is related to the natural survival probability by age and sex in the model. The input patient characteristics including mean age and sex ratio were determined by literature review combined with meta-analysis to find the pooled result of mean age and sex ratio.

Selected documents included in the synthesis include: epidemiological studies, results from clinical studies, cost studies, and treatment effectiveness studies related to chronic stage CKD using previous TKI in Vietnam.

For meta-analysis, the pooled results from the random-effect model were used if the p value of the Chi-square test was less than 0.05 or the heterogeneity between studies (I 2 ) was statistically significant (p=0.000). Otherwise, the pooled results from the fixed-effect model were used for analysis.

2.2.3.3. Cycle, model run time and discount rate

Based on the Guidelines for the Treatment of BCMDT [12], [26], the topic chooses the transition period between disease states in the cost-effectiveness analysis model as one month corresponding to the time the patient comes for a follow-up appointment. Because BCMDT is a chronic disease, the model analysis time chosen is the entire patient's life cycle to cover the entire treatment process.

A 3% discount rate based on WHO recommendation [52] was applied to calculate the change in value of both cost and effectiveness along the analysis period based on the formula:

In which: C is the current cost at the present time

C future cost at a specific time in the future r : discount rate (3%)

n: is the time distance from the present moment to a specific moment in the future (number of years)

2.2.4. Model input parameters

2.2.4.1. Parameters of disease state transition probability

Once the structure of the decision analysis model is determined, transition probabilities between disease states are collected using a literature review of clinical studies. Ideally, data should be collected from clinical trials or longitudinal studies of CKD patients in Vietnam; especially when CKD is associated with genetic mutations, the efficacy and safety of TKIs may vary depending on the population [62], [104], [110]; therefore, Vietnamese-specific data are needed. In cases where no relevant clinical studies are available, these probabilities will be borrowed from studies in other countries where the study population has similar demographic characteristics to Vietnam; in addition, systematic reviews and meta-analyses with high confidence levels based on the evidence pyramid are preferred.

The systematic review was conducted on the Pubmed electronic database and databases in Vietnam including: libraries of universities (Hanoi Medical University, Hanoi University of Pharmacy), the library of the Ho Chi Minh City Blood Transfusion and Hematology Hospital and on medical document library websites. The search was conducted based on the following keywords: chronic myeloid leukemia, chronic granulocytic leukemia, imatinib, nilotinib, tyrosine kinase. The study inclusion and exclusion criteria included:

- Selection criteria: longitudinal clinical studies with long-term follow-up and large sample size, with full-text, written in English or Vietnamese and providing treatment effects consistent with the disease state transformation direction of the topic.

- Exclusion criteria: clinical studies on pediatric subjects, cost estimation studies, journal letters and treatment guidelines.

After collection, the transition frequency will be recalculated to match the model period through the conversion formula from rate to probability as follows [58]:

p = 1 – e -rt

In which: p: probability; r: ratio; t is the time interval in which the event occurs.

In the analytical model, in addition to the cause of death due to the disease (assumed to occur only in the transition state), the patient also has the risk of death in any disease state due to the probability of natural death according to age and sex. This probability of death will be collected from the latest updated life table of Vietnam [156].

2.2.4.2. Parameters of treatment costs in each disease state

Cost estimation perspective

The study was conducted from a societal perspective and a payer perspective, and therefore includes direct medical costs from a payer perspective, direct medical costs from a societal perspective, and direct non-medical and indirect costs. Depending on the perspective, the total treatment costs are estimated as follows:

Paying agency's perspective: TCPĐT = CTTYT BHYT

Comment