2.1. Related Researches in the U.S.
To measure the economic impact of the disease, the common way of measurement includes the direct cost of medical cost, and the indirect cost by applying the human capital approach. The medical expenditure includes outpatient visits, inpatient visits, emergency visits, and drug fees
1 25% of men and 20% of women are suffered from hypertension in 2015 according to the WHO website.
2
(American Diabetes Association, 2008). As for the indirect cost, most studies include productivity loss from getting the disease (sick leave and mental problems), working hour loss due to the demand for caregiving, productivity loss due to mortality. The previous U.S. studies usually refer the disabled weight to the past survey for Activities of Daily Living (ADL) or other working ability surveys that assessing the impairment of the patient suffers from hypertension.
Then, they matched age and sex to project the whole patient population status of working ability.
In this study, several coefficients of the disabled weight of hypertension were adopted from those studies, while others were estimated for Taiwan cases.
2.2. Related Researches in Taiwan
Taiwan is known for its comprehensive and high-quality National Health Insurance Database. To estimate the health expenditure, there is no need to do sampling but use the whole patient population. Yet, there is no regular worker impairment survey every year. This makes it quite easy to estimate the direct cost but hard to estimate the indirect cost especially for the loss of productivity of labor and loss working hour due to caregiving demand. One study has been conducted to estimate the economic impact of enterovirus infection in Taiwan in 2015. The study streams the dataset of inpatient visits, outpatient visits, emergency visits, drug fees, and hospital stays to measurement the disease burden of enterovirus (Liu, et al., 2015). To estimate the economic impact, the direct cost of medical expenditures was calculated, and indirect cost includes travel costs, productivity loss of caregivers, and costs associated with premature mortality of patients are estimated by a human capital method.
Yet, there are several differences between this enterovirus study and the hypertension study.
First, the enterovirus is an acute disease. The syndrome would appear soon and is easy to be detected, which makes the cost is easy to be estimated. On the other hand, hypertension is a
3
chronic condition, which has a strong potential to cause other cardiovascular diseases but is hard to estimate the cost. Second, the target population of the enterovirus is children, there is no need to consider the production loss from the disease and children would need their parents to take them to see a doctor. However, in the hypertension case, the production loss of workers due to the there are a bunch of patients above age 15 must be estimated. Moreover, it is hard to estimate the loss of production due to caregiving since patients are capable to go to see a doctor by themselves.
2.3. Reduction of productivity for hypertension patients
To calculate the cost of sick leaves, Arno calculates days of sick leaves and multiplies with average earnings by specific age per day. The cost of caregivers taking leaves is calculated by days that patients are hospitalized or stay in ICU multiplied by average earnings for a worker in Taiwan. The reduction in productivity caused by hypertensive diseases is measured by the difference between earning of normal workers and hypertension workers multiplied with remaining life years (Arno, Levine, & Memmott, 1999).
To estimate the production loss of hypertension patients, most studies reviewing the survey on the elderly health status or ADL. Yet, according to the study from Hu, the loss of ADL caused by hypertension is not significant (Hu, Hu, Hsu, Hsieh, & Li, 2012). Therefore, in this study, the direct production loss from hypertension would not be calculated into the indirect cost but those derivative acute diseases.
According to the traditional cost-effectiveness analysis, one can measure the cost by adding up the cost of all the conditions that the disease may occur (here is hypertension) (Muennig, 2007). If there are more syndromes, then the highest cost of chronic disease should be seen as the dominating one. Types of indirect costs often include loss of productivity from morbidity and
4
premature mortality. Morbidity costs often include work loss among individuals, home productivity loss, and work loss for too sick to work. To estimate work loss among individuals due to too sick to work, the paper from American Diabetes Association matches the demographic data and the patient data and takes sick leave as the indicator of too sick to work (American Diabetes Association, 2008).
2.4. Labor force loss from mortality
The definition of the labor force refers to those who are currently working and the age above 15 from the Ministry of Labor in Taiwan. To estimate the loss of the labor force from mortality, one uses the growth model to capture the impact (Kalemli-Ozcan, Ryder, & Weil, 2000) (Finlay, 2007). Few studies discussed the loss from mortality via the cost estimation (Goetzel, et al., 2004).
2.5. Caregiver Cost Estimation
Caregiving includes formal and informal caregiving. In Taiwan, people often hire immigrants from South-East Asia for caregiving, and the reasonable wage is about 2000 to 2400 NTD per day (The Reporter, 2020). On the other hand, informal caregiving considers those family members to come to the hospital and take care of patients. To estimate the caregiver cost, the guideline proposed by Arno, Levine, and Memmott (Arno, Levine, & Memmott, 1999) asking the following questions: (1) What is the national prevalence of informal caregiving? (2) What is a reasonable market wage that would have to be paid to replace informal caregiving?
This study is not going to survey on Income and Program Participation, National Survey of Families and Households, National Health Interview Survey, or National Long-term Care Survey, but using the NHID to estimate the effect. From the estimation published by Labor
5
Ministry in Taiwan in 2015, there is one-fifth (2.31 million people) of the total workers (11.53 million people) in Taiwan are affected by informal caregiving (勞動部, 2010).
Critics for the indirect cost include the comment from Abegunde that the indirect cost may mislead the truth because with the high unemployment rate, for those patients who take sick leaves or cannot work, their family will replace their place. Therefore, the economic impact would be overestimated. He argues that the economic growth model would be a more accurate method to capture economic impact (Abegunde, Mathers, Adam, Ortegon, & Strong, 2007).
2.6. Hypertension issue in Economic Growth
The economic growth model proposed by Solow as AK type model with share belongs to (0,1), such as:
𝑌 = 𝐴𝐾𝛼𝐿1−𝛼 (Solow, 1956)
Y stands for the level of living, K stands for physical capital stock, L stands for the size of the labor force, A stands for the total factor productivity, 𝛼 stands for the share of physical capital stock.
After 1980, there are many researchers analyze how to improve the endogenous problem in the model. For the endogenous growth model, the macro-growth model needs to stand on the foundation of a micro-representative individual optimal decision problem. One of the factors they want to capture is a labor-augmented factor such as human capital. Moreover, human capital includes education and health status. Kalemli took care of the human capital investment with increasing life expectancy. (Kalemli-Ozcan, Ryder, & Weil, 2000) There are more factors rather than life expectancy, but few health topics have been discussed in the economic growth. Finlay’s
6
study proposed that education can affect health status. The model was adopted in a two-period utility maximization problem and considers the interaction with health and education (Finlay, 2007). Also, Chakraborty uses the economic growth model to examine the different health status groups that would reach a different equilibrium in the economy (Chakraborty, 2004). Cervellati constructs an overlapping generation model to capture the relationship between health and economic growth (Cervellati & Sunde, 2005). To combine these models above, it is reasonable to consider a stochastic process on losing their life, and the possibility for patients to occur hypertension follows Poisson Process. Also, to form the microeconomics foundation for Solow model, the model should consider (1) a two-period maximization problem (2) interaction between health and education (3) the medical expenditure (4) capital saving (5) fertility and mortality rate (6) the role of life expectancy and the disease weight.