The results of this study indicate that BC patients in Vietnam incurred significant OOP costs and the cost of BC treatment was driven by use of therapies and presentation stage at diagnosis. It is important to note that the analysis was based on a sample in which the proportion of patients diagnosed at early stage (0/I/II) was much higher than the national average of 50.5%  though the age of respondents was consistent with other studies [21, 22].
The mean total OOP cost of D&T for BC in Vietnam, excluding the cost of targeted therapy, was 61.8 million VND ($2,667) (range: 9.6 to149.9 million VND ($414-6,468) - approximately five times higher than the mean cost of 11.7 million VND ($633) reported in the only other study about costs of BC in Vietnam . This difference in estimates may be due to several reasons. The former study used 2001–2006 patient data when medical equipment and medications were less advanced and their use was more restricted. Secondly, costs were calculated from the healthcare payer’s perspective using unit costs which, in Vietnam, are much lower than the real cost of the resources used due to underestimation of health workforce remuneration and capital depreciation [8, 23]. From the patient’s perspective, our study reported the OOP cost which reflected the cost borne by patients and reveal the financial burden they faced. For example, the unit cost of mastectomy regulated by the Ministry of Health in 2018 was 4.7 million VND ($204) . In our study, the cost borne by the patients were reported at 18.4 million VND ($794) which was almost four times higher.
Although the costs reported in this study were much higher than the previous study, it seems much lower than the neighbouring country China or high-income countries like the US (3 and 5–15 times lower, respectively) [24–26]. Comparison of treatment costs in Vietnam with other low- and middle-income countries is not feasible due to the absence of studies with comparable methods (i.e., time horizons of costs, source of data, costing perspective). Moreover, comparison should be used with great caution as between-country differences in treatment costs are likely to be influenced by the variation in treatment guidelines, availability of treatment therapy, stage at diagnosis, and access to healthcare services .
The wide range of initial treatment cost is similar with the previous study . Variation in costs may be influenced by types of healthcare services related to BC D&T and the utilisation of HI for each service. All respondents in the study had HI and the majority (82%) used their HI to pay 20% of healthcare services cost (coinsurance = 20%). However, possession of HI did not mean respondents could utilize their HI for every service related to BC D&T. There were two possible scenarios: 1) patients chose intentionally not to use HI and opted to access ‘services on demand’ at public hospitals or services at private hospitals (patients pay 100% OOP) which tended to be viewed as providing better quality healthcare; 2) patients could not use HI for services which were not covered. The lowest utilisation level of HI was for breast reconstruction surgery and mastectomy (80–88%). With ‘service on demand’, patients receive premium service with privileges (i.e., the right to choose a surgeon, time of operation, hospital room, and level of care) which might partly explain why patients chose this option instead of using standard service covered by the HI.
The analysis showed the overwhelming dominance of targeted therapy cost in relation to total cost of care for patient with HER2+ (patients eligible for targeted therapy). The mean cost of targeted therapy was 649.5 million VND ($28,025) - tenfold higher than the total cost of all other healthcare services related to BC D&T. The share of targeted therapy would be 86% if it was included in the total cost. A study in Portugal found that systemic therapy (targeted therapy + chemotherapy) accounted for 69.2% of the total treatment cost of HER2 + patients which, in turn, was four times higher than patients with other BC subtype . Although 100% of respondents used HI for targeted therapy, the cost borne by patients was still much higher compared to other service costs due to the high-cost nature of targeted therapy and low HI coverage (e.g., HI covered 60% of the cost of Trastuzumab meaning patients paid 40% of the total cost by OOP plus 20% coinsurance of the part covered by HI ). Targeted therapy for HER2 + patients does not tend to be prescribed in Vietnam when it is known that patients cannot afford the treatment (personal communication). The effect of cost on doctor’s prescribing behaviour and patient treatment decisions were documented elsewhere [28–30] and need further research in Vietnam. Polices involving higher HI coverage rate for targeted therapy and/or an OOP maximum (a cap on the amount of money that a patient pays for covered healthcare services plan/year) is likely to impact positively on access by patients to appropriate BC treatment.
When targeted therapy was excluded from total cost, the largest share of cost belonged to chemotherapy (33%) with a mean cost of 36.5 million VND ($1,575). Diagnosis accounted for the smallest share in total cost (2%) with a mean cost of 2.6 million VND ($112). This cost composition is consistent with the previously noted study in Vietnam . Cost of diagnosis contained the highest number of extreme outliers. The maximum value with and without outliers was 824 million VND ($35,554) and 17 million VND ($734), respectively. The outliers likely reflect patients having to go through multiple tests at different hospitals before reaching the definitive diagnosis of BC .
Multivariate analysis revealed that later stage at diagnosis and higher education level of respondents were associated with higher total OOP cost of D&T for BC. Age, household monthly income, and coverage rate of HI were not significantly associated with costs, similar to the previous study in Vietnam . The mean OOP costs of BC D&T at stage II and III were, respectively, 66% and 148% higher than stage I. The trend is similar though higher than the pooled result from a systematic review of global treatment costs of BC by stage in which the rate was 32% and 95% respectively . Higher costs borne by patients in stage II/III is understandable as their utilisation of mastectomy, chemotherapy, and radiotherapy, which accounted for nearly two-third of the total cost, was significantly higher than stage I patients. The higher costs of later cancer stage emphasise the importance of early detection through screening programs. Policies that help downstaging BC at diagnosis will lessen the costs of treatment borne by the patients and their financial toxicity as well as increase the access to care and outcomes of treatment.
This study provides updated and detailed OOP costs for BC D&T in Vietnam as well as associated factors, using patient-level data. This is only the second study on the subject about costs related to BC and the first study in the country that analysed data from the patient’s perspective. Apart from complementing the previous study which looked at costs from healthcare payer’s perspective, the study provides novel and valuable insights that will facilitate the evaluation of novel therapies in terms of cost-effectiveness in Vietnam including early detection. In turn, the results will help decision-making by policymakers regarding health system financing and service delivery. The sample consisted of patients who were treated in main public hospitals from all three regions of Vietnam (the North, the Central, and the South). It is important to note that the study has some limitations. All costs were self-reported by respondents and were subject to recall bias though by cross-checking information with service price lists in hospitals or regulated by government and market price of drugs, no unreasonable or inconsistent data were flagged. Due to risk of recall bias and difficulties in measurement, we could not collect the costs of follow-up treatment, direct non-medical costs (i.e., transportation, meal, accommodation), and indirect costs (i.e., lost income, premature death) were also not gathered. There were no patients diagnosed with stage IV in the sample and this fact affected cost comparisons. We applied a multiple imputation technique to deal with missing data. Although the method used for imputation was technically sound and several pre-cautionary steps were taken to obtain the best reliable imputed values, there remains a possibility of bias.