The cost analysis of breast cancer: A comparison between private and public hospitals

Backgrounds Breast cancer is the most prevalent cancer among women. Breast cancer imposes a considerable economic burden on the health system. This study aimed to compare the cost of breast cancer among patients who referred to private and public hospitals in Iran (2017). Methods This was a prevalence-based cost of illness study. A total of 179 patients were selected from private and public hospitals using the census method. The researcher-constructed checklist was used for data collection. Data were analyzed using SPSS software version 22. Results The estimated total mean (SD) direct cost of patients who referred to the private hospital and the public hospital was $10050 (19480) and $3960 (6780), respectively. Further, the total mean indirect cost of patients who referred to the private hospital was lower than those referring to the public hospital at $1870 (%15 of total costs) and $22350 (%85 of total costs), respectively. These differences were statistically signicant (P<0.05). Conclusion

each year, and 1063 cases result in death (12). In 2035 compared to 2012, the number of new cases will be nearly two times greater (13).
Breast cancer imposes a considerable economic burden on societies (14)(15)(16). For example, the total cost of breast cancer was more than three times the total cost of prostate cancer (17). A study by Figueiredo et al. indicated that between 2004 and 2014, public healthcare costs increased, and the correlation between breast cancer and public healthcare costs was positive, mainly in uenced by governmental strategies (18). Breast cancer imposes a signi cant nancial burden on healthcare systems of Iran (19,20). Policymakers and health planners are interested in understanding the economic burden of illnesses to assess the optimal allocation of health resources to various diseases and estimate the potential costs and bene ts of public health interventions (20).
Cost of Illness (COI) studies indicate the importance of a particular disease and provide a baseline for assessing new interventions (20) and nancial losses as a result of illness (21). The aim of the COIstudies is providing an estimate of how much society spends on a particular disease and identifying different cost components (22). The COI can be used as a criterion for decision making in allocating limited budgets and resources for governmental health policies in effective control of diseases (21). A comprehensive economic analysis demands consideration of both direct and indirect costs such as productivity losses as a result of individuals unable to work because of hospitalization or outpatient visits, and also premature death arising from the illness (21).
In future, the cost of cancer care will increase as new sophisticated, expensive treatment modalities are adopted to raise the standard of care {Xu, 2003 #2} (23). Breast cancer is on the rise in Iran, and since patients are mostly diagnosed at more advanced stages of the disease (24,25), mortality resulting from breast cancer is high (26). So, the presentation of accurate data about the economic burden of the disease will allow informed decision making by health care policymakers in Iran about the prevention and treatment of the disease. Therefore, the objective of this study was to compare the cost of breast cancer among patients who referred to private and public hospitals in Iran in 2017.

Database and study population
This was a prevalence-based cost of illness study, which was conducted from the societal perspective using bottom-up approach costing.
The statistical population in this study included all patients with breast cancer. One hundred seventy-nine patients with breast cancer who admitted to the private hospital (N=103) and the public hospital (N=76) in Rasht (a city in the north of Iran) between Aug 2016 and Aug 2017 included in this study.

Cost assessment
The cost of illness is divided into three general categories: direct costs, indirect costs, and intangible costs. In this study, we mainly focus on the rst two cost categories. The direct costs consist of medical costs and non-medical costs. The former includes medical care expenditures for diagnosis, treatment, and rehabilitation, etc., while the latter includes the consumption of non-healthcare resources like transportation, household expenditures, relocating. Indirect costs include lost productivity due to premature deaths and missed workdays and decreased productivity in the workplace due to morbidity.
Finally, intangible costs include the cost of pain and suffering in patients and their families and relatives. In this study, intangible costs were not calculated.
In this study, the economic burden of breast cancer was assessed by calculating direct medical costs, direct nonmedical costs, and indirect costs. Data related to the hospitalization part of direct medical costs were extracted from patients' records and data related to the outpatient part of direct medical costs, direct nonmedical costs and indirect costs were obtained via an interview with patients and their family members, respectively. The researcher-made checklist was used for data collection. The initial draft of the checklist extracted from two records: 1. "Cost-of-illness studies -a primer" (27) and 2. "Cost-of-illness studies: concepts, scopes, and methods" (28). Then, to complete the checklist, we interviewed 5 oncologists, 2 researchers who had conducted at least one cost of illness study, 2 professors in the eld of Health Economics and 8 breast cancer patients. The checklist consists of demographic variables (age, marital status, monthly income status, educational status, job status, supplemental insurance status, and the type of basic insurance), duration of the disease and treatment type and questions related to costs components incurred by patients during cancer diagnosis, and treatment procedures. In this study, direct medical costs were valued based on the medical tariffs of diagnostic and therapeutic services.
Indirect costs include the monetary value of resources loses due to morbidity and mortality. There are three approaches to estimate indirect costs: the human capital approach (HCA), the friction cost approach (FCA) and the willingness to pay approach (WTP). HCA measures the lost production, in terms of lost earnings, of a patient or caregiver. FCA measures only the production losses during the time it takes to replace a worker, and WTP measures the amount an individual would pay to reduce the probability of illness or mortality. HCA is the most common approach used to calculate the indirect costs of an illness. A criticism of this approach is that certain groups are assigned a higher value than others. A criticism of WTP is that this approach is often di cult to implement in COI studies. For speci c diseases, extensive surveys of people's preferences are needed, which the results rely heavily on the type of question and people's responses to very speci c hypothetical questions. For communicable diseases, surveys may ignore the cost of the disease because of externalities (cost of externalities incurred by disease). The WTP, therefore, is often not feasible for a cost-of-illness study. Proponents of the FCA criticize the HCA for overvaluing the indirect costs, claiming that the productivity losses are often eliminated after a new employee is trained and can replace the former employee. However, the FCA is rarely used because it requires extensive data to estimate losses in the friction period. On the other hand, the estimated cost is strongly in uenced by the labour supply situation (27,29).
In this study, indirect costs were calculated based on the HCA. These costs were estimated by summing two parts: 1) The costs of lost productivity due to patients and their families' missed workdays and 2) the cost of premature death due to breast cancer. First, in order to estimate the cost of missed workdays per patient, we calculated the average number of missed workdays by patients and their families because of breast cancer and then multiplied by the minimum daily wage rate (310000 (2017)), in this way we estimated the cost of missed workdays per patient. Also, by having the number and the mean age of premature death and retirement age (60 years old) in Iran, the total number of years lost due to premature death resulting from breast cancer was calculated and multiplied by the number of days of the year and the minimum daily wage rate, in this way the cost of premature death was calculated. Finally, the total cost of lost productivity calculated by summing these two parts.
The equations used for indirect costs calculation are as follows: 1. The cost of missed workdays = the mean (patients missed workdays + patient family's missed workdays) × minimum daily wage rate 2. C = the mean {(retirement age-age at premature death) × (the number of patients who died ÷sample size)} × (minimum daily wage rate × the number days of the year) To recall bias prevention, patients' treatment process were followed up every two months for one year. All costs in this study were expressed as US Dollars based on the Exchange rate of Central Bank of the Islam ic Republic of Iran (US$ 1 = 31389 Rials (2017)).

Data analysis
Data were analyzed using SPSS software version 22 and excel (2016). Descriptive statistics (mean (SD), frequency, and percent) were used to assess the demographic variables status. K-S test (Kolmogorov-Smirnov) was applied to assess the normality of data. Since the P-Value for all variables was less than 0.05 (P<0.05), non-parametric tests, including Mann-Withney and Kruskal-Wallis, were used to assess the association between demographic variables and costs. The Spearman correlation coe cient also was used to examine the correlation between age at diagnosis and costs. A multivariate regression model was used to control for confounding factors.

Results
A total of 179 patients with breast cancer were included in the analysis. The majority of patients were covered by the basic insurance (98.9%), and only 36.3% of patients were covered by supplemental insurance. Most of the patients (64.2% ) held a diploma degree and more than half of the patients were non-natives (54.2%). A statistically signi cant difference was found between supplemental insurance status and total medical direct cost (P<0.05) Table 1.
The mean(SD) of age at diagnosis, age and age at death was estimated at 45.41 (9.38), 47.98 (10.08) and 49.94 (11.80), respectively. The estimated mean(SD) number of hospital admission and the length of hospital stay of patients who referred to the private hospital was 1.35 (0.50) and 2.71 (2.49), respectively whereas those who referred to the public hospital was higher at 1.48 (085) and 8.63 (1049), respectively.
Additionally, 10.7% of patients who referred to the private hospital and 6.6% of those referring to the public hospital postponed their treatment process for more than two months due to nancial barriers. The direct medical costs of breast cancer patients who referred to the private hospital and the public hospital were $9880 (%82.90 of the total costs and 1.89 times GDP per capita) and $3620 (%13.74 of the total costs and 69.29 percent of GDP per capita), respectively. The hospitalization costs and outpatient costs of patients who referred to the private hospital were higher than those referring to the public hospital. The highest component of hospitalization costs of patients who referred to the private hospital was related to surgery cost at $980 (%53.73 of the total hospitalization cost), whereas that of patients who referred to the public hospital was related to hoteling costs at $380 (%30.26 of the total hospitalization cost).
Moreover, drug cost had the lowest rate in breast cancer patients who referred to the private hospital at $30 In contrast, the lowest cost among those referring to the public hospital was related to the diagnostic cost at $100. In summary, outpatient costs were the main component of the direct medical costs for breast cancer patients who referred to the private hospital and the public hospital.
Besides, the total mean nonmedical direct cost of patients who referred to the private hospital and the public hospital was $170 (%1.39 of the total costs) and $340 (%1.29 of the total costs), respectively. The highest component of the direct nonmedical cost of patients who referred to the private hospital and the public hospital was attributed to commuting and food costs at $150and $250, respectively Table 2.  Table 3. Hospitalization costs in the public hospital were 510, signi cantly lower as compared with the private hospital. Besides, outpatient costs and direct medical costs in the public hospital respectively were 3290 and 5270, lower compared to the private hospital, but this was not statistically signi cant.    (30). So it can be concluded that the age of breast cancer onset has decreased in Iran in recent years. The average mortality age of breast cancer is still lower than other cancers, and the economic burden of this disease will rise in the predictable future, according to one study in Japan (21).
In this study, the total mean cost of breast cancer among patients who referred to the public hospital was  (21), which the results of these studies are in line with our results.
The difference between direct and indirect costs in patients referred to private and public hospitals may be due to several reasons. Firstly, premature death was the major component of the total indirect cost of breast cancer patients who referred to the public hospital, whereas that did not occur among breast cancer patients in the private hospital. This may be because private hospitals offered better services which resulted in a higher survival rate and a lower mortality rate. Besides, given that the mean age of patients with breast cancer referring to the public hospital (49.776.66 (9.89)) was higher as compared with those referring to the private hospital (46.66 (10.06)), and this difference was statistically signi cant (p<0.05), the high mortality rate in the public hospital can be because most of the older patients referred to the public hospital. On the other hand, patients with advanced-stage cancer likely referred more to public hospitals for receiving services. Secondly, none of the patients who referred to the public hospital had supplementary insurance, while most of the patients who referred to the private hospital, in addition to basic health insurance, were covered by supplemental insurance. Supplemental insurance has increased patients access to more advanced and expensive treatment services and has made services more inelastic by reducing the patients' co-payment or have led to increased induced demand. In this study, the correlation between direct medical costs, outpatient costs, chemotherapy costs and age at diagnosis was statistically signi cant and negative at P<0.05.
Thirdly, tariffs in the private sector are 2-4 times higher than that of the public sector. Therefore, direct medical costs are higher in patients referring to the private sector.
Of note, in Iran, only people who have better socio-economic status and better income level are able to afford supplemental health insurance and refer to the private sector for receiving treatment, which in turn cause they receive more expensive and advanced services. Hence, it may cause the mortality rate among those to refer to the private hospital to be lower and incur lower indirect costs, but due to more treatment services utilization, they are more likely to incur greater direct medical costs than those referring to the public hospital.
Of the direct medical costs, outpatient costs were higher than hospitalization costs in both private and public hospitals. The outpatient cost of patients who referred to the private hospital was 3.4 times greater than those referring to the public hospital. The total cost of missed workdays for the patient and the patient's family, who referred to the private hospital was 1.7 times greater than those referring to the public hospital. Both in the private and the public hospital, the cost of missed workdays of patient's family members was greater than patients themselves. These costs (opportunity cost) are imposed on patients' families in real terms but are hidden from policymakers' view.
In our study, basic insurance played an important role in the reimbursement of direct medical costs and reducing the proportion of out-of-pocket expenses in direct medical costs. The majority of breast cancer costs in public hospitals was paid by basic insurance (%90.68), %6.39 of the costs were paid by the patient, and only a small proportion was paid from the targeted subsidy plan by the government (%2.92).
To the contrary, in the private hospitals, %35. 36  It is important to note that although most of the cancer patients in the private sector were covered by supplemental insurance, they paid higher co-payments. Since tariffs in the private sector are 2-4 times higher than that of the public sector, patients referring to private hospitals paid more out of pocket payments despite supplemental insurance. Therefore, these patients are likely to have better socioeconomic status and more ability to pay. On the other hand, despite higher costs, these patients may prefer to go to private hospitals because of the shorter waiting time and better service quality.
Since the present study was performed at cross-sectional and prevalence-based method, matching was not conducted between patients referring to the public and the private hospitals in terms of age, income level and disease stage and also the effect of confounding variables was not controlled. Since it is not possible to conclude with any certainty, it is necessary to investigate the cause of this difference in costs and mortality rate between patients referring to the public and the private hospitals in future studies using a perspective and controlled design. In Multivariate regression model after adjusting for confounding variables (e.g., age, education status, marital status, habitation status, type of basic insurance and supplemental insurance status), hospitalization costs in patients referring to private hospitals were signi cantly higher than those referring to the public hospital.

Limitations
This study had several limitations. First, since some patients refused to answer the questions asked of them, the selection bias (sampling bias and attrition) of respondents in reviewing the costs could not be avoided. Second, the indirect costs consisted of only the missed workdays and premature mortality, which would greatly undervalue the indirect economic burden of illness. The lack of data on permanent leaving the job by patients and caregivers during the recovery period could also underestimate the indirect cost estimates. Third, the cost of breastfeeding was not calculated due to the paucity of data. Fourth, intangible economic costs of breast cancer patients and their families, including the pain, sorrow, were not included because they are di cult to convert into a monetary value (36). Given this was a crosssectional and prevalence-based study, matching was not conducted between patients referring to the public and the private hospitals in terms of age, income level and disease stage and also the effect of confounding variables were not controlled. An additional limitation is that this study conducted in only two private and public hospitals that can limit the generalization of study ndings to all private and public sector.

Policy implications
Given that the cost of premature death in the private hospital was zero, it is not possible to conclude with certainty whether cancer patients who referred to the public hospital were at the nal stage of the disease or bene ted from better services or both? If the low mortality rate and low indirect costs in patients referred to the private hospital be attributed to the quantity and quality of services provided to cancer patients referring to the private sector and considering the high share of indirect costs of total costs in patients referred to the public hospital, it is necessary that health policymakers take the necessary measures to improve the quantity and quality of public sector services. Also, despite the insurance coverage, patients suffer a high amount of OOP payment, a substantial and wide-ranging effort is needed to support breast cancer patients. This suggests that insurance policies need to be revised to increase nancial support among cancer patients, especially for those who are currently uninsured. It is recommended that the results of this study to be used in future studies to evaluate the cost-effectiveness of screening interventions, early detection and preventive interventions, and health policymakers take an appropriate policy to reduce the economic burden of this disease. It is also suggested that future studies should examine whether the higher costs in private hospitals is due to disparities in tariffs of the private and public sector or due to greater quantity and quality services provided in private hospitals.

Conclusion
Breast cancer imposes a substantial economic burden on patients at private and at public hospitals, healthcare system and society. Indirect costs were considerably higher for breast cancer patients and their caregivers referring to the public hospital, especially in terms of premature mortality than those referring to the private hospital, which can show a signi cant proportion of the total costs. Because indirect costs do not impose on the health system and health insurance organizations, health policymakers do not pay enough attention to these costs. Therefore, these costs must be addressed at the macro level of economic policymaking. Support strategies also should be adopted for cancer patients and their family members at parliament and government level, and unemployment insurance, improved for cancer patients.
Abbreviations COI: Cost of Illness