Cost-utility analysis of combination medical therapies in chronic coronary syndrome: a comparative study using real-world and patient-level data from Iran

Abstract

Objectives

The main objective was to evaluate the cost-effectiveness of various medical therapy combinations in managing chronic coronary syndrome (CCS) in Iran, based on real-world and patient-level data.

Design

A cost-utility analysis employing a Markov model was conducted using data from a retrospective cohort study.

Setting

The study was conducted in the healthcare setting of Iran, focusing on primary and secondary care.

Participants

Patients with CCS were included in the study. Numbers entering and completing the study were reported, with clear definitions of selection, entry and exclusion criteria.

Interventions

All combinations of recommended medical therapies for CCS were permitted. Ultimately, taking into account the sample size and study power, a comparison was made between the combination therapy of β-blockers (BB), long-acting nitroglycerin (LAN), aspirin (ASA) and statin versus the group receiving only BB, ASA and statin.

Primary and secondary outcome measures

The primary outcome measure was the incremental cost-effectiveness ratio, along with an initial evaluation of disability-adjusted life-years (DALYs) and costs related to the interventions.

Results

The BB/LAN/ASA/statin combination was cost-saving and effective, averting 0.02 DALYs and saving $172 compared with BB/ASA/statin. This combination was cost-effective in over 97% of the probabilistic sensitivity analysis results.

Conclusions

Incorporating LAN into the combination therapy of BB, ASA and statin is cost-effective in Iran. This finding provides evidence for policymakers on resource allocation in low-income countries.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The study uses a Markov model for cost-utility analysis, providing a robust framework for evaluating long-term outcomes.

• Real-world and patient-level data were used, enhancing the study’s external validity.

• Deterministic and probabilistic sensitivity analyses were conducted, ensuring the reliability and robustness of the results.

• The study faced limitations, including the unavailability of information regarding patients' lifestyles and non-insured medications.

• Medication adherence was estimated based on claims data, which may not reflect actual medication use.

Introduction: background and objectives

Cardiovascular diseases (CVDs) have remained the leading cause of death in developed countries for the past century. However, the most pressing concern regarding these diseases is in lower-middle-income countries. This concern is due to the rise in the rate of death and cardiovascular (CV) attacks, which can be attributed to population growth, ageing, Western lifestyle and an increase in smoking.¹

Non-communicable diseases account for over 76% of Iran’s disease burden.^{2 3} Moreover, a report from the WHO in 2016 revealed that 46% of deaths in Iran were linked to CVDs.²

The most prevalent form of CVD is ischaemic heart disease (IHD)⁴ and is the leading cause of disability-adjusted life-years (DALYs) in Iran and worldwide.⁵ The consequences of IHD on society and the economy are significant. In recent years, the health sector costs associated with IHD have risen. Therefore, it is essential to identify cost-effective treatments to help decision-makers choose options with better cost-effectiveness profiles.

The most common form of IHD is chronic coronary syndrome (CCS), primarily characterised by stable angina.⁶ To manage disease symptoms in CCS, optimal medical therapy (OMT) or revascularisation can be used.⁶ Revascularisation includes percutaneous coronary intervention (PCI) or coronary artery bypass graft (CABG). The primary objectives of CCS treatment are to eliminate angina and minimise the likelihood of future health risks, particularly myocardial infarction (MI) and premature death.⁷

Studies show that OMT can be cost-effective for most patients with CCS and non-inferior to revascularisation strategies regarding the survival benefit.⁷,⁹ In addition, revascularisation strategies, PCI and CABG, might be applicable in patients with objectively proven severe ischaemia or intractable chest pain.^{10 11}

According to a 2014 study, CADs impose a substantial economic burden on the Iranian healthcare system, primarily due to treatment costs, particularly angioplasty. In contrast, pharmaceutical expenses accounted for only 1.15% of the total costs.¹²

According to CCS management guidelines, several pharmacotherapy options/combinations can benefit plaque stabilisation, symptom relief and secondary prevention.¹³,¹⁵ Also, multimedicine therapy offers a wide range of medical therapy (MT) options/combinations. So, it is crucial to pay attention to the effects and consequences of multimedicine therapies.

Considering the chronic nature of the disease and multiple possible MT options/combinations, economic evaluation of different MT combinations in managing CCS in the real world was essential. Therefore, we used real-world and patient-level data from two referral hospitals in this study to help decision-makers choose the most cost-effective alternative. Notably, all steps in this study were according to the Consolidated Health Economic Evaluation Reporting Standards 2022.¹⁶

Methods

Study population, setting and location

This real-world and patient-level cost-utility analysis draws on data from a retrospective cohort study of patients with CCS in Iran. The design of the cohort study has been detailed in a previous publication elsewhere.¹⁷ Final data were collected from 10 961 angiography records from two major referral heart centres in Iran in 2014. Following the application of specific eligibility criteria, a retrospective cohort of 505 eligible patients with CCS was identified, and electronic health records were monitored from January 2014 to March 2020. The main inclusion criteria for that study encompassed patients with CCS aged between 45 and 65 years who were discharged with MT as their primary treatment strategy. Patients diagnosed with acute coronary syndrome or unstable haemodynamic, confirmed congenital heart disease, underlying comorbidities such as overt valvular heart disease or peripheral vascular disease or with a history of or were candidates for revascularisation (ie, PCI or CABG) were excluded from the study. Further details, including a patient inclusion flow diagram of the cohort study and additional information, are provided in online supplemental figure S1.

Based on the referenced cohort, the mean age of patients was 56.4 (5.8) years, with 63% being male (online supplemental table S1 and figure S2). The distribution of patients by the severity of CAD was as follows: 83.56% with single-vessel disease, 15.05% with double-vessel disease and 1.39% with three-vessel disease (online supplemental table S2). Patients were referred from all 31 Iranian provinces. Additionally, these patients presented with a range of risk factors, including hypertension (52.1%), dyslipidaemia (45.5%), obesity (33.5%), cigarette smoking (31.4%) and a familial history of coronary artery disease (CAD) (22.1%) (online supplemental figure S3).

Comparators

Concerning the real-world method of the study, all possible combinations of MT for CCS were allowed to be included in our study. Based on the performed retrospective cohort study, three MT combinations with the most frequent prescription (online supplemental figure S4) were compared in the economic evaluation phase, including (1) beta-blockers (BB) + β-blockers (ASA) + statin; (2) ASA + statin and (3) BB + long-acting nitroglycerin (LAN) + ASA + statin.

Perspective, time horizon and discount rate

This study was conducted based on the player’s perspective. The study time horizon was considered to be 6 years, which was in line with the 6-year follow-up of the reference cohort study. All included costs and utilities were annually discounted by 7.2% and 3%, respectively.

Selection, measurement and valuation of outcomes

According to the natural history of CCS and the results of the conducted systematic review and meta-analysis,¹⁸ MI, PCI, CABG and mortality rates were selected as the primary health outcomes. Moreover, DALYs averted were used to pool these outcomes. DALY is an index that combines and summarises the time lost due to premature death and the duration of life in less-than-optimal conditions, generally called disability. DALY for each year of premature death is equal to 1, and for each year living with disability is between 0 and 1. Disability scores for discrete health states were extracted from published studies.¹⁹,²² Initially, the number of DALYs associated with each MT combination in the follow-up period was calculated. Then, the DALYs averted by each MT combination were measured by head-to-head subtracting the DALYs between MT combinations.

Measurement and valuation of resources and costs

All direct medical resources and costs related to medicines, monitoring, unplanned revascularisations and adverse effects were measured and captured. Moreover, all the expenses included in the analysis were converted into US$ using the 2021 Iranian purchasing power parity.

Rationale and description of the model

A Markov model with five distinct health states—stable CCS, post-MI, post-PCI, post-CABG and death—was used to simulate patient outcomes. All patients start the model in a stable CCS state. Patients could experience PCI or CABG without having an MI, and similarly, they could experience an MI without undergoing PCI or CABG. Depending on the treatment they receive, patients can transition to post-MI, post-PCI and post-CABG, or remain in the stable CCS state for the next cycle. In all health states, including the stable CCS state, patients face the probability of death. The probability of death in the CCS state follows the natural mortality rates in Iran, while probabilities in other health states were derived from survival analysis of our CCS patient cohort study in Iran. Additionally, patients in the post-MI state faced the probability of experiencing PCI or CABG in each cycle (figure 1). The core rationality of adapting this five-state Markov model was the survival results of the time-trend data from our retrospective cohort study, and the natural history of CCS.²³

Figure 1. The model structure. CABG, coronary artery bypass graft; MI, myocardial infarction; PCI, percutaneous coronary intervention.

Analytics and assumptions

According to the results of the conducted cohort, it was assumed that patients do not transfer between post-CABG and post-PCI health states. Also, based on Iranian national health accounts, we assumed that 80% of patients get their health services from the public sector and the remaining 20% from the private sector.

This study used Microsoft Excel and IBM SPSS Statistics V.26 for the cohort study analyses. The cost-utility analysis was performed using TreeAge Pro Healthcare V.2022. Additionally, statistical significance was set to p-values <0.05.

Characterising heterogeneity

We conducted an analysis of variance (ANOVA) to find heterogeneity between the baseline health characteristics of the cohort study’s included patients. Only three combinations with homogeneous patients were included in the Economic Evaluation (EE) (online supplemental table S3).

Characterising distributional effects

All costs, utilities, probabilities and other essential model inputs were entered into the model with appropriate distributions (online supplemental table S4).

Characterising uncertainty

Deterministic sensitivity analysis and probabilistic sensitivity analysis (PSA) were performed to evaluate the robustness of the model. First, we did a tornado analysis to find critical parameters in the model. Finally, results were reported based on PSA.

Patient and public involvement

As part of this study, a policy brief resulting from the research findings was presented to health sector policymakers. This engagement aimed to involve key stakeholders in understanding and using the study results to inform policy decisions. Although patients and the broader public were not directly involved in the study design or conduct, the dissemination of results through the policy brief ensures that the findings reach and potentially benefit a wider audience.

Results

Study parameters

All transition probabilities in the model were calculated using the following formula,²⁴ based on survival analysis of patients in the retrospective cohort study (online supplemental tables S5–S7).

$${\displaystyle P(t)=1-e^{\wedge }(-rt)}$$

where P represents probability, e represents Euler’s number (2.71828), r represents rate and t represents time.

The disability weights used in the model varied from 0.033 (0.008) for stable CCS to 0.22 (0.005) for the post-CABG state. Moreover, we applied 0.005 (0.003) and 0.02 (0.005) disability weights for PCI and CABG procedures, respectively (table 1).

Table 1. Disability weights of discrete health states.

Health state	Disability weight mean (SD)	Reference
CCS	0.033 (0.008)	Salomon et al21
MI	0.074 (0.0143)	Salomon et al21
Post-PCI	0.191 (0.0196)	Brandão et alandde Noordhout et al19 20
Post-CABG	0.22 (0.0126)	Brandão et alandde Noordhout et al19 20
PCI procedure	0.005 (0.003)	Weintraub et al22
CABG procedure	0.02 (0.005)	Weintraub et al22
Death	1.00	Salomon et al21

CABG, coronary artery bypass grafting; CCS, chronic coronary syndrome; MI, myocardial infarction; PCI, percutaneous coronary intervention

Finally, all relevant costs were included from the Iran market and Iran health services tariff V.2021 (table 2).

Table 2. Cost parameters in the model.

Cost item	Value (US$)	SD
CABG procedure	8059.58	3486.53
PCI Procedure	3982.26	1220.54
CCU hospitalisation	158.28	26.05
Cardiac rehabilitation	18.14	4.59
Visit	8.03	2.83
Statin monitoring	5.16	1.22
Clopidogrel	0.20	0.13
Long-acting nitroglycerin	0.06	0.01
Statin	0.05	0.01
β-Blocker	0.05	0.05
Calcium channel blocker	0.04	0.01
ASA	0.03	0.01

ASAaspirinCABG, coronary artery bypass grafting; CCU, cardiac care unit; PCI, percutaneous coronary intervention

Summary of main results

The base-case model illustrated that the BB/LAN/ASA/S combination, on average, is associated with US$890 and 0.376 DALY during the 6-year follow-up period. Considering that the BB/LAN/ASA/S combination is associated with fewer costs and DALYs compared with the BB/ASA/S and ASA/S combinations, the BB/LAN/ASA/S combination is dominant (online supplemental figure S5). In this case, due to the dominance of one of the alternatives, there is no need to compare the calculated incremental cost-effectiveness ratio (ICER) with the willingness to pay (WTP) threshold (table 3).

Table 3. Cost-effectiveness ranking.

Dominance	Combination	Cost (US$)	DALY	Incremental cost (US$)	DALY averted	ICER (US$)
Absolutely dominated	BB/ASA/S	1062	0.395	Ref	Ref	Ref
Absolutely dominated	ASA/S	1255	0.411	193	−0.016	−12,063
Dominance	BB/LAN/ASA/S	890	0.376	−172	0.019	−9,053

ASAaspirinBB, β-blocker; DALY, disability-adjusted life-year; LAN, long-acting nitroglycerin; S, statin

Effect of uncertainty

All the parameters used in the model were entered into the tornado analysis. By removing the non-influential parameters on the certainty of the model in a stepwise manner, the most critical parameters remained. However, the study model was not sensitive to the variation of remaining parameters in the range of ±20% (online supplemental figure S6).

Based on the Monte Carlo approach, the model’s PSA result showed that the BB/LAN/ASA/S combination is cost-effective in over 96% of the resulting ICERs compared with the BB/ASA/S combination (figure 2).

Figure 2. Incremental cost-effectiveness scatterplot; comparing the β-blocker/long-acting nitroglycerin/aspirin/statin (BB/LAN/ASA/S) combination against the BB/ASA/S combination. DALY, disability-adjusted life-year.

Discussion

The main objective of this study was to analyse the cost-effectiveness of MT combinations for CCS in Iran. As far as we know, this is the first economic evaluation study to compare potential CCS MT combinations, including medicines for disease and comorbidity control and medicines to prevent secondary outcomes.

Initially, a retrospective cohort study was conducted to gather all relevant data from real-world practice at the patient level in Iran. We discovered that three MT combinations—(1) BB/ASA/S, (2) ASA/S and (3) BB/LAN/ASA/S—were prescribed to over 50% of patients during the 6-year follow-up period. It is crucial to note that the conducted ANOVA demonstrated no significant statistical variation among the three MT combinations concerning the patient’s baseline risk factors (online supplemental table S3).

Over 6 years, the average total costs per patient for MT combinations 1–3 were $1062, $1255 and $890. Moreover, the calculated DALY for combinations 1–3 during this period were 0.395, 0.411 and 0.376, respectively. Considering the cost savings and lower DALYs associated with the BB/LAN/ASA/S combination, it was determined to be the dominant MT combination. Furthermore, a PSA with 1000 resamples indicated that the BB/LAN/ASA/S combination is cost-effective in more than 96% of ICERs, given a WTP threshold of US$8682 in Iran.

While revascularisation may improve outcomes for patients with ST-elevation myocardial infarction (STEMI) and multivessel CAD, recent meta-analyses suggest that there may not be significant differences in health outcomes for patients with CCS between MT therapy, PCI and CABG.^{18 25 26} Previous economic studies have shown that MT is a cost-effective alternative to revascularisation for patients eligible for both treatments.^{9 27} However, it is vital to consider revascularisation as a complementary strategy for patients with CCS.²⁸ As Fearon and colleagues discovered, PCI of coronary lesions with reduced fractional flow reserve leads to improved outcomes. It is economically attractive compared with the best MT among patients with stable CAD and abnormal fractional flow reserve.²⁹ Moreover, Kodera and colleagues, in a cost-effectiveness analysis, found that PCI is cost-effective compared with MT in patients with STEMI in Japan.³⁰ However, our study in Iran discovered that revascularisation is the costliest item, with an average cost of $3982 (with an SD of $1221) for each PCI procedure and $8060 (with an SD of $3487) for each CABG procedure. We found that unplanned revascularisation resulted in a significant increase in patient medical resource utilisation. The findings of this study strongly indicate that the BB/LAN/ASA/S treatment plan is a highly cost-effective option for managing CCS compared with BB/ASA/S or ASA/S combinations. This option can potentially decrease costs by reducing the need for revascularisation.

Previous studies have emphasised the significance of statins and ASA in managing CCS.³¹,³³ Our research further highlights the importance of adding LANs into the combination of BBs, ASA and statins to manage CCS effectively. This approach helps reduce CV mortality, costs and disease burden, just as Castellano and team’s multicentre observational study shows that a polypill combining statin, antiplatelet and antihypertensive drugs significantly reduces mortality risk and other disabilities over a year.³⁴

In 2017, a group of researchers led by Liping conducted a meta-analysis on 106 002 patients with CAD. Their findings revealed that adhering to evidence-based MT combinations, including BB, angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, antiplatelets and statins, with good adherence (≥80%) can significantly reduce the risk of death (HR 0.56) and CV mortality (HR 0.66).³⁵ However, our study’s cohort phase discovered that the adherence rate was only about 60%, which is below the recommended threshold for good adherence. So, we believe that increasing patients’ adherence to the MT by accepted approaches, including better patient education, increased use of interventions to improve compliance and the use of performance measures focused on long-term optimal MT in ambulatory care settings,³⁶ can result in higher levels of cost-effectiveness.

Finally, although clinical guidelines recommend long-acting nitrates as a second-line treatment for CCS,¹³,¹⁵³⁷ our study found that adding this class of medicines to the combination of BB/ASA/statin is very cost-effective. Therefore, we suggest a further study to evaluate our findings with a prospective cohort study, which can contribute to developing and updating Iranian CCS clinical guidelines based on a large patient-level and real-world data study.

It is essential to be aware of certain limitations when interpreting the results of this study. First, despite providing valuable insights, this study is limited by its relatively short 6-year observational period. Chronic conditions like CAD often require long-term monitoring to fully understand their progression and the sustained impact of treatment interventions. Future longitudinal studies with extended follow-up periods are essential to capture long-term outcomes and cost-effectiveness more comprehensively. Such studies will enable a more thorough evaluation of the enduring benefits and potential drawbacks of medical therapies, ensuring that healthcare decisions are based on robust, long-term evidence. Second, our access to cause-of-death data was limited to questionnaires and data from the Iranian Civil Registry Organization as we did not have access to precise cause-of-death data based on the International Classification of Diseases. Third, we lacked data regarding patients' lifestyles, including smoking withdrawal, dietary habits and non-insured medications. Fourth, compliance with medication therapy was relatively low, which could be due to the nature of real-world data or the chronic nature of the disease. Lastly, socio-economic information and health literacy data for the patients included in the study were not accessible.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.