Health Policy 119 (2015) 787–793
Contents lists available at ScienceDirect
Health Policy journal homepage: www.elsevier.com/locate/healthpol
The effectiveness of preventive care at reducing curative care risk for the Taiwanese elderly under National Health Insurance Chin-Shyan Chen a,1 , Yu-I Peng b,2 , Ping-Chang Lee b,3 , Tsai-Ching Liu b,∗ a b
Department of Economics, National Taipei University, 151, University Rd., San Shia, New Taipei City 237, Taiwan Department of Public Finance, National Taipei University, 151, University Rd., San Shia, New Taipei City 237, Taiwan
a r t i c l e
i n f o
Article history: Received 24 July 2014 Received in revised form 7 January 2015 Accepted 8 January 2015 Keywords: Preventive care Curative care Risk Elderly National Health Insurance Taiwan
a b s t r a c t Whether provision of free preventive care for the elderly under National Health Insurance has reduced the risk for curative care use raises much concern in Taiwan. This study analyzes the relationship by examining the impact of health examination utilization on the utilizations of outpatient care and inpatient care. Data come from the 2005 National Health Interview Survey and National Health Insurance Research Database. A two-stage method is used in the estimation. We found a negative relationship between the utilization of preventive care and hospitalization care in terms of length of stay and medical expenditures. On average, the elderly people who used preventive care tended to have 16 shorter hospitalization stays and NTD64,220 lower hospitalization expenditures than their counterparts. In order to improve the health of the elderly and reduce the escalation of medical expenditures due to aging, including preventive care in the health insurance is a very effective strategy. © 2015 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Population aging is observed in many countries around the world, with Taiwan among some of the fastest-aging societies as its elderly population and life expectancy have been increasing over time. The proportion of Taiwan’s population aged 65 and over was 6.53% in 1991, 8.81% in 2001, and hit 10.89% in 2011. Life expectancy at birth was
∗ Corresponding author. Tel.: +886 2 86741111x67398; fax: +886 2 86716108. E-mail addresses: [email protected] (C.-S. Chen), [email protected] (Y.-I. Peng), [email protected] (P.-C. Lee), [email protected] (T.-C. Liu). 1 Tel.: +886 2 86741111x67178; fax: +886 2 26739880. 2 Tel: +886 2 86741111x67391; fax: +886 2 86716108. 3 Tel: +886 2 86741111x67398; fax: +886 2 86716108. http://dx.doi.org/10.1016/j.healthpol.2015.01.004 0168-8510/© 2015 Elsevier Ireland Ltd. All rights reserved.
74.26 years in 1991, 76.75 years in 2001, and 79.15 years in 2011 [1]. With such a prolonged lifespan, chronic diseases have become much more prevalent among older people. In 2007 nearly 90% of elderly people aged 65 and above in Taiwan experienced at least one type of chronic disease. Chronic diseases, such as cancers, strokes, and heart diseases, are among the leading causes of death in the country [2]. Moreover, health expenditure by the elderly population is substantially higher than other age groups. In 2011 the average annual medical expenditure per person was NTD43,838 (US$1 = NTD30) for people aged 50–59, NTD73,984 for people aged 60–69, NTD104,799 for people aged 70–79, and NTD134,693 for people aged 80 and over [3]. The increasing instances of chronic diseases are largely attributable to a lack of a healthy lifestyle, appropriate health knowledge, and early detection of symptoms. As chronic diseases and their serious consequences are mostly preventable, health insurance could include preventive
788
C.-S. Chen et al. / Health Policy 119 (2015) 787–793
care services in the coverage, especially for the elderly people, as a measure to prevent chronic diseases turning into serious illnesses, resulting in escalations of medical expenditures and financial burdens on health insurance systems. The National Health Insurance (NHI) has been implemented in Taiwan since 1995. The NHI provides comprehensive coverage and medical care services for people in Taiwan, including low co-payments for outpatient care, inpatient care, and emergent care services. Taiwan’s NHI system launched the program of adult preventive care services in 1996, providing free annual services to people aged 65 and above. The preventive care services contain personal information inquiries (including personal and family medical history, medication history, health behaviors, depression detection, etc.), physical examination (including height, weight, blood pressure, BMI, waistline measurement, etc.), and laboratory tests (urine, blood, etc.). In 2001 the preventive care services were extended to contain health consultation services, in which people receive consultations on the examination results and health promotion recommendations after health examinations. The nation’s program of preventive care services is aimed at maintaining and promoting good health for the elderly and reducing medical burdens through early discovery and treatment of diseases. The proportion of elderly beneficiaries utilizing these services has been increasing over time since the program’s initiation. However, despite the free annual services, the utilization rate among people aged 65 and above remains low at around 35% [2]. The literature has found that the demand for preventive care is influenced by a variety of demographic and socio-economic characteristics, health status and behavior, health information, and medical insurance coverage [4–12]. Kenkel [4] indicates that the demand for preventive care could be higher for older people, because of higher risks of illness, but that the demand could also be lower as older individuals have less incentive to invest in their health, because the time period of their remaining life is shorter. Research on preventive care utilization among the elderly aged 65 and over is quite limited. Moreover, whether preventive care services are effective in maintaining and promoting individuals’ health can be evaluated through individuals’ subsequent health status and medical care utilization. The effect of preventive care service utilization on subsequent medical care utilization could be positive or negative, depending on whether the two services are complements or substitutes. Preventive care could reduce the number of illness incidences through early discovery and treatment, decreasing subsequent curative medical care utilization probability and expenditure. On the other hand, people who take up preventive care services are more aware and better informed of their medical conditions and thus have more frequent visits to hospitals. Research examining this relationship using the elderly population is more scant. Therefore, the purpose of this study is to examine those factors influencing the demand for preventive care services and the linkage of preventive care service utilization to subsequent medical care service utilization, including both inpatient and outpatient services, among the elderly people in Taiwan.
2. Method 2.1. Data This research’s data come from two databases. The first is the 2005 National Health Interview Survey (NHIS), which is a face-to-face interview survey conducted by the Bureau of Health Promotion, Department of Health, National Health Research Institutes, and the Food and Drug Administration, Department of Health in Taiwan. The NHIS views the 23 cities and counties of Taiwan as independent populations and uses a multi-stage stratified systematic sampling design to divide each city or county into one to four strata according to factors such as geographic location, population distribution, and urbanization. In addition, the samples are selected from each stratum by the probability proportional to size (PPS) method. This research takes the individual questionnaires for those people aged 65 and above. We exclusively focus on the elderly population aged 65 and above, because their health has a particularly higher risk and because elderly people in Taiwan have free access to preventive care services and a low financial barrier for seeking medical care under the NHI. The survey provides demographic and socioeconomic information, health status, and health behaviors. After excluding sets with missing data, a total of 2239 elderly people make up the study’s sample. This research links the NHIS and the National Health Insurance Research Database (NHIRD) to get the preventive care and medical utilization data of the elderly. The NHIRD encompasses all the medical utilization records of Taiwan’s National Health Insurance. Hence, we are able to get the most accurate information on medical utilization. 2.2. Analytical technique First, we present the data using descriptive statistics. Second, to identify the factors affecting the use of preventive care services by the elderly and the association between preventive care use and the use of curative care, including outpatient care and hospitalization care, we regress a two-stage model rather than perform a single equation so as to prevent the endogeneity problem. We present the structural model shown as follows: Y1∗ = ˇ1 X1 + ε1 ,
Y1 = 1
Y2∗ = Y1∗ + ˇ2 X2 + ε2 ,
Y1∗ > 0
Y2 = 1
(1)
Y2∗ > 0
(2)
[ε1 , ε2 ] ∼BVN (0, 0) 12 , 22 , . here Y1 and Y2 , respectively, denote the utilizations of preventive care and curative care, X1 and X2 are explanatory variables, ˇ and are assumed to be vectors of unknown coefficients, εs are unobserved disturbances that are assumed to be correlated across the equations, j is the standard deviation, and is the correlation of error terms (ε1 , ε2 ). The preventive care equation includes all variables believed to impact curative care use plus identifying those variables that affect the likelihood to obtain preventive care, but which do not have a direct impact on
C.-S. Chen et al. / Health Policy 119 (2015) 787–793
789
Table 1 Definitions of variables. Variables Dependent variables Preventive care Outpatient care Outpatient care use Number of outpatient Expense of outpatient Inpatient care Hospitalization LOS Expense of hospitalization Independent variables Gender Age 75 Education Elementary school (Ref.) High school College Living alone Household income ∼49,999 (Ref.) 50,000–99,999 100,000– Geographic location North (Ref.) Center South East Health status Good (Ref.) Moderate Bad ADLs Influenza vaccine
Definitions Receiving preventive care in 2005; yes = 1 others = 0 Receiving outpatient care in 2006; yes = 1, others = 0 The number of outpatient care visits in 2006 The total expense of outpatient care in 2006 Receiving inpatient care in 2006; yes = 1, others = 0 The total length of stay of hospitalization in 2006 The total expense of hospitalization in 2006 Gender of the elderly is male; yes = 1 other = 0 The elderly age 75 or above; yes = 1 other = 0 The elderly’s education is elementary school or below; yes = 1, others = 0 The elderly’s education is senior or junior school; yes = 1, others = 0 The elderly’s education is college or above; yes = 1, others = 0 The elderly is living alone; yes = 1, others = 0 Household monthly income is below 50,000; yes = 1, others = 0 Household monthly income is between 50,00 and 99,99; yes = 1, others = 0 Household monthly income is 100,000 or above; yes = 1, others = 0 The elderly lives in Ilan County, Keelung City, Taipei City, Taipei County, Taoyuan County, Hsinchu City, Hsinchu County, and Miaoli County; yes = 1, others = 0 The elderly lives in, Taichung City, Taichung County, Chunghua County, Nantou County, and Yunlin County; yes = 1, others = 0 The elderly lives in Chiayi city, Chiayi County, Tainan City, Tainan County, Kaohsiung City, Kaohsiung County, and Pingtung County; yes = 1, others = 0 The elderly lives in, Taitung County, Hualien County, and Penghu County; yes = 1, others = 0 The elderly’s health status is good, very good or extremely good; yes = 1, others = 0 The elderly’s health status is moderate; yes = 1, others = 0 The elderly’s health status is bad; yes = 1, others = 0 The elderly is having no difficulty in eating, bathing, dressing, toileting, going to bed, and walking in doors; yes = 1, others = 0 The elderly is taking influenza vaccine in the past one year; yes = 1, others = 0
Ref.: the reference group.
curative care use and which are found to be additional characteristics: activities of daily living (ADLs) and influenza vaccine. In other words, to reduce the potential endogeneity of preventive care use in the decision of medical care use, we adopt two instrumental variables (IV)—whether an individual has received an influenza vaccine and whether an individual has difficulty in ADLs. A valid IV should have a direct impact on the endogenous variable but have no direct impact on the outcome variable [13,14]. In our model, the IVs have to directly affect the receipt of preventive care but affect the use of curative care only through the use of preventive care. We assume that an individual who tends to take preventive actions such as receiving an influenza vaccine is more likely to receive preventive health examinations as well, while difficulty in ADLs might prevent an individual from taking effort to go through non-urgent health examinations. On the other hand, the receipt of an influenza vaccine and difficulty in ADLs do not directly reduce or increase the likelihood and expenses of curative care. To verify the validity of the two IVs we adopt, we ran the regression for the use of preventive care and found that two IVs had significant effects on the use preventive care. When we ran the regression for the use of curative care including the two IVs as additional explanatory variables, we found that both the
receipt of influenza vaccine and difficulty in ADL had no direct impact on the use of curative care. We are able to obtain consistent and asymptotically efficient parameters through the two-stage estimation. In the first stage of the empirical estimation, we get the fitted values of Y1 (Yˆ 1 ) by a logistic technique. In the second stage we estimate the structural equation of curative care after substituting Y1 by Yˆ 1 in Eq. (2). We next separate the utilization of curative care into two groups – outpatient care and hospitalization care – in this empirical study in order to differentiate between medical care use and the different severities of illnesses. We apply the two-part model to estimate the relationship between preventive care utilization and curative care utilization, including outpatient care and inpatient care. For outpatient care, the first part estimates the probability of receiving outpatient care and uses a logistic regression. The second part notes the frequency and expenditure of outpatient care visits, with the Poisson regression or negative binomial regression and OLS regression as the proper model. For inpatient care, the first part is the probability of receiving inpatient care and uses a logistic regression. The second part is the total length of stay and expenditure of hospitalization. Here, we respectively implement the Poisson regression or negative binomial regression and OLS regression for the
790
C.-S. Chen et al. / Health Policy 119 (2015) 787–793
estimations. Table 1 contains all variable names and definitions.
lower hospitalization expenses, compared to those who did not use preventive care (21.62% vs. 7.18%; 17.5 vs. 33.3; NTD106,952 vs. NTD171,173).
3. Results 3.2. Regression results 3.1. Descriptive statistics Table 2 provides descriptive statistics of preventive care with mean values and standard deviations for each variable. We find that those who tend to have a higher probability of preventive care use than their counterparts have the following characteristics: elderly people who are male, aged 75 or above, low educated, living alone, with a household income above NTD100,000, living in the eastern area, with a worse health status, with no disability, and have gotten an influenza vaccine. The characteristics of the sample for outpatient care utilization and hospitalization care use are presented in Tables A1 and A2 (see Supplementary material). Around 98% of the elderly who took health examinations have actually sought out sequential outpatient care, while only 37.6% of the elderly who did not have check-ups later utilized outpatient care. The number of outpatient care visits and medical expenses for the elderly with health examinations are much higher than that for those who did not have health examinations: 38 vs. 17.9 and NTD39,438 vs. NTD23,792, respectively. The elderly who used preventive care have a higher percentage of being hospitalized, but have a shorter length of stay and Table 2 Descriptive statistics of preventive care. Variables
Gender Male Female Age 75 Below 75 75 or above Education Elementary school High school College Living alone Yes No Household income Up to 49,999 50,000–99,999 100,000 onwards Geographic location North Center South East Health status Good Moderate Bad ADLs No difficulties Difficulties Influenza vaccine Yes No
Preventive care Mean (%)
S.D.
40.14 39.18
0.4904 0.4884
1116 1123
38.72 41.20
0.4873 0.4925
1387 852
40.33 36.91 37.58
0.4907 0.4833 0.4860
1773 317 149
48.03 38.71
0.5007 0.4872
229 2010
38.68 41.48 48.72
0.4872 0.4934 0.5020
1758 364 117
42.46 37.06 36.85 49.23
0.4945 0.4835 0.4827 0.5038
961 483 730 65
38.58 38.41 43.30
0.4871 0.4867 0.4959
801 893 545
40.67 33.54
0.4913 0.4729
1923 316
46.45 30.11
0.4989 0.4590
1309 930
N
Table 3 reports the result of factors affecting preventive care utilization. For the elderly who live alone, have no ADLs problems and have influenza vaccine, with a household income above NTD100,000 and with a worse health status, we find that they have a higher likelihood of using preventive care. Compared to those who live in the north, elderly who are located in the central or southern area are less likely to use preventive care. Table 4 shows the estimated effect of preventive care utilization for the elderly and the socio-demographic characteristics on outpatient care utilization in terms of the probability of seeking care, doctor visits, and medical expenses. Few factors are found to be significantly and consistently related to outpatient care utilization, with the most crucial factors being preventive care, household income, and regional inequality. This study finds that preventive care utilization has strongly positive significant (P < 0.001) coefficients on the likelihood of seeking outpatient care, suggesting that those elderly who use preventive care are more likely to seek outpatient care. Preventive care is also found to have a positive impact on both the number of outpatient visits and medical expenses. This indicates that once the elderly have had outpatient care, those who use outpatient care services tend to have a higher number of outpatient visits and bear greater medical expenses. This finding fails to reinforce our hypothesis that a negative association between preventive care utilization and curative care utilization. Table 4 also presents the result of hospitalization care regarding the probability of being admitted to a hospital, Table 3 Factors affecting preventive care utilization. Variables
Preventive care Coef.
Constant Gender Age 75 Education High school College Living alone Household income 50,000–99,999 100,000– Geographic location Center South East Health status Moderate Bad ADLs Influenza vaccine N * ** ***
P < 0.10. P < 0.05. P < 0.01.
S.E.
−1.13 0.09 0.05
0.17*** 0.09 0.09
−0.21 −0.18 0.43
0.13 0.19 0.14***
0.08 0.45
0.12 0.20**
−0.35 −0.30 0.21
0.12** 0.10** 0.26
0.01 0.18 0.32 0.73 2239
0.10 0.12 0.13** 0.09***
C.-S. Chen et al. / Health Policy 119 (2015) 787–793
791
Table 4 Effect of preventive care on outpatient care and hospitalization care. Number of outpatient
Outpatient care use
Constant Preventive care Gender Age 75 Education High school College Living alone Household income 50,000–99,999 100,000– Geographic location Center South East Health status Moderate Bad * ** ***
Log (expense of outpatient)
Hospitalization
LOS
Log (expense of hospitalization)
Coef.
S.E.
Coef.
S.E.
Coef.
S.E.
Coef.
S.E.
Coef.
S.E.
Coef.
S.E.
−0.44 2.59 0.06 0.10
0.23* 0.51*** 0.09 0.09
2.95 0.86 −0.01 −0.03
0.11*** 0.24*** 0.04 0.04
9.26 0.99 −0.01 −0.09
0.19*** 0.40** 0.07 0.07
−1.74 −0.22 −0.05 −0.02
0.32*** 0.72 0.13 0.13
4.03 −3.2 0.32 0.27
0.36*** 0.80*** 0.16** 0.15*
11.45 −2.01 0.55 0.19
0.42*** 0.92** 0.15*** 0.15
−0.14 0.04 −0.16
0.13 0.19 0.15
−0.20 0.38 0.08
0.20 0.24 0.22
0.24 0.28* 0.24
−0.53 0.26 0.24
0.26* 0.30 0.24
-0.52 0.38
0.22 0.34 0.21* 0.19 0.35
0.09 0.11 0.07
0.06 0.08 0.07
0.21 0.21 0.08
0.10** 0.14 0.11
0.05 0.50 0.39
0.12 0.23**
0.11 −0.10
0.06** 0.09
0.06 0.07
0.10 0.13
-0.25 0.51
0.19 0.26**
-0.34 0.94
−0.42 −0.27 −0.28
0.12*** 0.11** 0.27
0.03 0.11 −0.05
0.06 0.05** 0.12
−0.02 0.07 −0.14
0.10 0.08 0.17
−0.36 −0.16 0.46
0.18* 0.15 0.33
−0.19 −0.07 0.12
0.20 0.17 0.36
−0.39 −0.28 −0.45
−0.01 0.05
0.10 0.12
0.08 0.09**
−0.03 0.21
0.15 0.17
−0.27 0.09
0.17 0.19
0.17 0.12
0.08 0.53
0.01 0.05
0.05 0.06
0.06 0.20
0.22 0.27***
0.18 0.20
P < 0.10. P < 0.05. P < 0.01.
length of stay, and medical expense. We find that the relationship between preventive care and the probability of hospitalization care use is significant and negative, but it fails to reach the lowest significant level (P < 0.1). Although those who receive prevent care do not appear to reduce the likelihood of being hospitalized, they do seem to reduce hospitalization care use in terms of length of stay and total medical expenditures. The coefficient of preventive care is, as expected, found to be statistically significant negative, which reinforces our hypothesis that a negative association between preventive care utilization and curative care utilization. Household income is another crucial factor that consistently affects the frequency of hospitalization care use. 4. Discussion This study examines those factors influencing the demand for preventive care services and the linkage of preventive care service utilization to subsequent medical care service utilization, including both inpatient and outpatient services, among the elderly people in Taiwan. German et al. [15] show that individuals aged 65 and above who utilize preventive care services have a better health status and lower mortality rates than those who do not utilize such services. Nakanishi et al. [16] study the above relationship using city-level data in Japan and find that the utilization rate of preventive care is negatively associated with the utilization rate of inpatient care, as well as average inpatient and outpatient expenditures, but positively associated with the utilization of outpatient care. Tian et al. [12] find that people aged 50 and over in Taiwan who received preventive care services have lower probabilities of utilizing inpatient care services, suggesting a substitution relationship between the two services. Burton et al. [17] note that among people aged 65 and above in the U.S.,
compared to those who do not utilize preventive care services, those who do utilize the services have a lower number of outpatient visits. Unexpectedly, the elderly who use preventive care utilization does not show any obvious effect on reducing outpatient care use in this study. Conversely, it even significantly demands more outpatient care utilization in terms of the probability of seeking outpatient care, the number of outpatient care visits, and the total expenditures of outpatient care. We nevertheless have found negative relationships between the utilization of preventive care and hospitalization care in terms of length of stay and medical expenditures, though we did also find a positive link between the utilization of preventive care services and the probability that one will be hospitalized. It is possible that the elderly people who used preventive care would also be more likely to use curative care than their counterparts. Both the outpatient care utilization including the probability of seeking outpatient care, the number of outpatient care visits, the total expenditures of outpatient care and the hospitalization care utilization including the probability of seeking hospitalization care increased with those who had preventive care. However, on average, the elderly people who used preventive care tended to have 16 shorter hospitalization stays and NTD64,220 lower hospitalization expenditures than their counterparts (17.5 vs. 33.3; NTD106,952 vs. NTD171,173). This finding demonstrates that the elderly people who used preventive care service reduce later utilization of curative care, particularly for those who suffering from severe illness, leading to longer length of stay and higher medical expenditures through hospitalization care. A possible reason for an increase in outpatient care utilization and a reduction in hospitalization care utilization could be due to early detection of health problems and undergoing more-timely medical treatments with outpatient care
792
C.-S. Chen et al. / Health Policy 119 (2015) 787–793
and hospitalization care, which mitigates the dramatic increase in length of stay and medical expenditures later on. The National Health Insurance system that Taiwan implemented in 1995 now provides comprehensive care for its citizens who previously had no coverage at all and dramatically reduce their out-of-pocket medical expenditures. Furthermore, free preventive care is also provided for the elderly. However, no more than 35% of Taiwanese elderly actually utilize preventive care. One barrier is regional disparity. In the central and southern areas, the elderly are less likely to seek preventive care than those in the northern area, because more medical resources are centralized in the northern area where the capital city (Taipei City) is located. In non-northern areas, elderly people who consider traveling to be an inconvenience may not feel like using preventive services if they are not nearby. To further raise the accessibility of preventive care, time barriers need to be reduced to the home or community where elderly live as well as free shuttle services could be offered. Generally, the Bureau of NHI uses various incentives, including higher outpatient physician fees, no physician visit limits for payments, and increases in payments for circuit medical services, to encourage local independent practitioners to provide medical services in remote areas. Most of Taiwan’s elderly, afraid of facing the fact that they may have contacted or developed some type of serious illness, may refuse to get the check-ups they need. This fear of finding out the truth – a form of denial – prevents them from using preventive services and causes them to postpone needed medical attention as their health deteriorates. Taiwan’s health authorities should respond to these cultural and very common human responses to fear of bad news by educating the elderly and emphasizing the importance of preventive care. If the elderly fully understand the importance of preventive care for their health in reducing serious illness, they would more than likely prefer to receive preventive care to protect themselves from getting sick instead of seeking curative care after an illness has occurred. The BNHI needs to put forth more efforts at increasing the elderly’s awareness of the importance of preventive care and at disseminating information on the benefits of the NHI’s elderly preventive care package. Doing so can help lead the way to greater improvements in the health of Taiwan’s elderly population. Three major limitations of this study should be noted. First, the data we use are based on a cross-sectional design with a sample of only 2239 elderly people. Second, regarding the instrument variables we adopted, there might exist confounders that are associated with both the instrument and the outcome variables [18]. For example, health consciousness can be the instrument-outcome confounder. Those individuals who are more health-conscious (i.e., who care about their own health a lot) are more likely to take influenza vaccine, and are more likely to have healthy lifestyle and thus require less curative care. In that case, the estimates we obtain from the IV method could be biased. However, given the availability of our dataset, we are unable to address this by fully-accounting for every single potential instrument-outcome confounders. This limitation, along with the nature of our cross-sectional
dataset, makes the causal relationship difficulty to testify. Third, from the data we only know whether or not the respondent utilizes the standard package of preventive care services provided by the NHI. Future research with more detailed data on preventive care could explore the link between specific kinds of preventive care and subsequent utilization of medical care. 5. Conclusions Whether the provision of free preventive care for the elderly under the National Health Insurance system has reduced the risk for curative care use raises much concern in Taiwan, as it does in other countries. This study has analyzed the relationship between preventive care utilization and curative care utilization for the elderly by examining the impact of health examination utilization on the utilizations of outpatient care and inpatient care. Data come from the 2005 National Health Interview Survey and National Health Insurance Research Database. A two-stage model is used in the estimation. This study shows that a negative association between preventive care use and inpatient care use does exist in Taiwan’s health care system. The risk of having a serious illness, which leads to longer hospitalization stays and higher medical expenditures, drops if the illness can be detected through preventive care and treated as soon as possible. In order to both improve the health of the elderly and reduce the escalation of medical expenditures due to aging, including preventive care in the health insurance system is a very effective strategy. Conflict of interest statement None declared. Funding sources This work was supported by the National Science Council Taiwan through grant NSC 101-2410-H-305-054-MY3 and is based on data from the 2005 National Health Interview Survey (NHIS) and the National Health Insurance Research Database (NHIRD) in Taiwan. The NSC financial support and the NHIS and NHIRD data provision are gratefully acknowledged. Acknowledgements Constructive comments of editor and anonymous referees are gratefully acknowledged. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/ 10.1016/j.healthpol.2015.01.004. References [1] Council for Economic Planning and Development. Health and vital statistics (1): General health statistics. Taiwan: Department of Health, Executive Yuan Republic of China; 2012.
C.-S. Chen et al. / Health Policy 119 (2015) 787–793 [2] Bureau of Health Promotion. Elderly Health Promotion Plan 2009–2012. Taiwan: Bureau of Health Promotion, Department of Health, Executive Yuan Republic of China; 2009. [3] Department of Health. National Health Expenditure 2011. Taiwan: Department of Health, Executive Yuan Republic of China; 2011. [4] Kenkel D. The demand for preventive medical care. Applied Economics 1994;26:313–25. [5] Casey MM, Call KT, Klingner JM. Are rural residents less likely to obtain recommended preventive healthcare services? American Journal of Preventive Medicine 2001;21(3):182–8. [6] Yu SM, Bellamy HA, Schwalberg RH, Drum MA. Factors associated with use of preventive dental and health services among U.S. adolescents. Journal of Adolescent Health 2001;29:395–405. [7] Jenkins CNH, Le T, Mcphee SJ, Stewart S, Ha NT. Health care access and preventive care among Vietnamese immigrants: do traditional beliefs and practices pose barriers? Social Science & Medicine 1996;43(7):1049–56. [8] Ross JS, Bernheim SM, Bradley EH, Teng HM, Gallo WT. Use of preventive care by the working poor in the United States. Preventive Medicine 2006;44:254–9. [9] Powell-Griner E, Bolen J, Bland S. Health care coverage and use of preventive services among the near elderly in the United States. American Journal of Public Health 1999;89(6):882–6. [10] Wu S. Sickness and preventive medical behavior. Journal of Health Economics 2003;22:675–89.
793
[11] Chen JY, Diamant A, Pourat A, Kagawa-Singer M. Racial/ethnic disparities in the use of preventive services among the elderly. American Journal of Preventive Medicine 2005;29(5):388–95. [12] Tian WH, Chen CS, Liu TC. The demand for preventive care services and its relationship with inpatient services. Health Policy 2010;94:164–74. [13] Wooldridge JM. Introductory econometrics: a modern approach, 5th ed. Boston: Cengage Learning; 2012. [14] Newhouse JP, McClellan M. Econometrics in outcomes research: the use of instrumental variables. Annual Review of Public Health 1998;19:17–34. [15] German PS, Burton LC, Shapiro S, Steinwachs DM, Tsuji I, Paglia MJ, et al. Extended coverage for preventive services for the elderly: response and results in a demonstration population. American Journal of Public Health 1995;85(3):379–86. [16] Nakanishi N, Tatara K, Fujiwara H. Do preventive health services reduce eventual demand for medical care? Social Science & Medicine 1996;43(6):999–1005. [17] Burton LC, Steinwachs DM, German PS, Shapiro S, Brant LJ, Richards TM, et al. Preventive services for the elderly: would coverage affect utilization and costs under Medicare? American Journal of Public Health 1995;85(3):387–91. [18] Garabedian LF, Chu P, Toh S, Zaslavsky AM, Soumerai SB. Potential bias of instrumental variable analyses for observational comparative effectiveness research. Annals of Internal Medicine 2014;161:131–8.
Contents lists available at ScienceDirect
Health Policy journal homepage: www.elsevier.com/locate/healthpol
The effectiveness of preventive care at reducing curative care risk for the Taiwanese elderly under National Health Insurance Chin-Shyan Chen a,1 , Yu-I Peng b,2 , Ping-Chang Lee b,3 , Tsai-Ching Liu b,∗ a b
Department of Economics, National Taipei University, 151, University Rd., San Shia, New Taipei City 237, Taiwan Department of Public Finance, National Taipei University, 151, University Rd., San Shia, New Taipei City 237, Taiwan
a r t i c l e
i n f o
Article history: Received 24 July 2014 Received in revised form 7 January 2015 Accepted 8 January 2015 Keywords: Preventive care Curative care Risk Elderly National Health Insurance Taiwan
a b s t r a c t Whether provision of free preventive care for the elderly under National Health Insurance has reduced the risk for curative care use raises much concern in Taiwan. This study analyzes the relationship by examining the impact of health examination utilization on the utilizations of outpatient care and inpatient care. Data come from the 2005 National Health Interview Survey and National Health Insurance Research Database. A two-stage method is used in the estimation. We found a negative relationship between the utilization of preventive care and hospitalization care in terms of length of stay and medical expenditures. On average, the elderly people who used preventive care tended to have 16 shorter hospitalization stays and NTD64,220 lower hospitalization expenditures than their counterparts. In order to improve the health of the elderly and reduce the escalation of medical expenditures due to aging, including preventive care in the health insurance is a very effective strategy. © 2015 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Population aging is observed in many countries around the world, with Taiwan among some of the fastest-aging societies as its elderly population and life expectancy have been increasing over time. The proportion of Taiwan’s population aged 65 and over was 6.53% in 1991, 8.81% in 2001, and hit 10.89% in 2011. Life expectancy at birth was
∗ Corresponding author. Tel.: +886 2 86741111x67398; fax: +886 2 86716108. E-mail addresses: [email protected] (C.-S. Chen), [email protected] (Y.-I. Peng), [email protected] (P.-C. Lee), [email protected] (T.-C. Liu). 1 Tel.: +886 2 86741111x67178; fax: +886 2 26739880. 2 Tel: +886 2 86741111x67391; fax: +886 2 86716108. 3 Tel: +886 2 86741111x67398; fax: +886 2 86716108. http://dx.doi.org/10.1016/j.healthpol.2015.01.004 0168-8510/© 2015 Elsevier Ireland Ltd. All rights reserved.
74.26 years in 1991, 76.75 years in 2001, and 79.15 years in 2011 [1]. With such a prolonged lifespan, chronic diseases have become much more prevalent among older people. In 2007 nearly 90% of elderly people aged 65 and above in Taiwan experienced at least one type of chronic disease. Chronic diseases, such as cancers, strokes, and heart diseases, are among the leading causes of death in the country [2]. Moreover, health expenditure by the elderly population is substantially higher than other age groups. In 2011 the average annual medical expenditure per person was NTD43,838 (US$1 = NTD30) for people aged 50–59, NTD73,984 for people aged 60–69, NTD104,799 for people aged 70–79, and NTD134,693 for people aged 80 and over [3]. The increasing instances of chronic diseases are largely attributable to a lack of a healthy lifestyle, appropriate health knowledge, and early detection of symptoms. As chronic diseases and their serious consequences are mostly preventable, health insurance could include preventive
788
C.-S. Chen et al. / Health Policy 119 (2015) 787–793
care services in the coverage, especially for the elderly people, as a measure to prevent chronic diseases turning into serious illnesses, resulting in escalations of medical expenditures and financial burdens on health insurance systems. The National Health Insurance (NHI) has been implemented in Taiwan since 1995. The NHI provides comprehensive coverage and medical care services for people in Taiwan, including low co-payments for outpatient care, inpatient care, and emergent care services. Taiwan’s NHI system launched the program of adult preventive care services in 1996, providing free annual services to people aged 65 and above. The preventive care services contain personal information inquiries (including personal and family medical history, medication history, health behaviors, depression detection, etc.), physical examination (including height, weight, blood pressure, BMI, waistline measurement, etc.), and laboratory tests (urine, blood, etc.). In 2001 the preventive care services were extended to contain health consultation services, in which people receive consultations on the examination results and health promotion recommendations after health examinations. The nation’s program of preventive care services is aimed at maintaining and promoting good health for the elderly and reducing medical burdens through early discovery and treatment of diseases. The proportion of elderly beneficiaries utilizing these services has been increasing over time since the program’s initiation. However, despite the free annual services, the utilization rate among people aged 65 and above remains low at around 35% [2]. The literature has found that the demand for preventive care is influenced by a variety of demographic and socio-economic characteristics, health status and behavior, health information, and medical insurance coverage [4–12]. Kenkel [4] indicates that the demand for preventive care could be higher for older people, because of higher risks of illness, but that the demand could also be lower as older individuals have less incentive to invest in their health, because the time period of their remaining life is shorter. Research on preventive care utilization among the elderly aged 65 and over is quite limited. Moreover, whether preventive care services are effective in maintaining and promoting individuals’ health can be evaluated through individuals’ subsequent health status and medical care utilization. The effect of preventive care service utilization on subsequent medical care utilization could be positive or negative, depending on whether the two services are complements or substitutes. Preventive care could reduce the number of illness incidences through early discovery and treatment, decreasing subsequent curative medical care utilization probability and expenditure. On the other hand, people who take up preventive care services are more aware and better informed of their medical conditions and thus have more frequent visits to hospitals. Research examining this relationship using the elderly population is more scant. Therefore, the purpose of this study is to examine those factors influencing the demand for preventive care services and the linkage of preventive care service utilization to subsequent medical care service utilization, including both inpatient and outpatient services, among the elderly people in Taiwan.
2. Method 2.1. Data This research’s data come from two databases. The first is the 2005 National Health Interview Survey (NHIS), which is a face-to-face interview survey conducted by the Bureau of Health Promotion, Department of Health, National Health Research Institutes, and the Food and Drug Administration, Department of Health in Taiwan. The NHIS views the 23 cities and counties of Taiwan as independent populations and uses a multi-stage stratified systematic sampling design to divide each city or county into one to four strata according to factors such as geographic location, population distribution, and urbanization. In addition, the samples are selected from each stratum by the probability proportional to size (PPS) method. This research takes the individual questionnaires for those people aged 65 and above. We exclusively focus on the elderly population aged 65 and above, because their health has a particularly higher risk and because elderly people in Taiwan have free access to preventive care services and a low financial barrier for seeking medical care under the NHI. The survey provides demographic and socioeconomic information, health status, and health behaviors. After excluding sets with missing data, a total of 2239 elderly people make up the study’s sample. This research links the NHIS and the National Health Insurance Research Database (NHIRD) to get the preventive care and medical utilization data of the elderly. The NHIRD encompasses all the medical utilization records of Taiwan’s National Health Insurance. Hence, we are able to get the most accurate information on medical utilization. 2.2. Analytical technique First, we present the data using descriptive statistics. Second, to identify the factors affecting the use of preventive care services by the elderly and the association between preventive care use and the use of curative care, including outpatient care and hospitalization care, we regress a two-stage model rather than perform a single equation so as to prevent the endogeneity problem. We present the structural model shown as follows: Y1∗ = ˇ1 X1 + ε1 ,
Y1 = 1
Y2∗ = Y1∗ + ˇ2 X2 + ε2 ,
Y1∗ > 0
Y2 = 1
(1)
Y2∗ > 0
(2)
[ε1 , ε2 ] ∼BVN (0, 0) 12 , 22 , . here Y1 and Y2 , respectively, denote the utilizations of preventive care and curative care, X1 and X2 are explanatory variables, ˇ and are assumed to be vectors of unknown coefficients, εs are unobserved disturbances that are assumed to be correlated across the equations, j is the standard deviation, and is the correlation of error terms (ε1 , ε2 ). The preventive care equation includes all variables believed to impact curative care use plus identifying those variables that affect the likelihood to obtain preventive care, but which do not have a direct impact on
C.-S. Chen et al. / Health Policy 119 (2015) 787–793
789
Table 1 Definitions of variables. Variables Dependent variables Preventive care Outpatient care Outpatient care use Number of outpatient Expense of outpatient Inpatient care Hospitalization LOS Expense of hospitalization Independent variables Gender Age 75 Education Elementary school (Ref.) High school College Living alone Household income ∼49,999 (Ref.) 50,000–99,999 100,000– Geographic location North (Ref.) Center South East Health status Good (Ref.) Moderate Bad ADLs Influenza vaccine
Definitions Receiving preventive care in 2005; yes = 1 others = 0 Receiving outpatient care in 2006; yes = 1, others = 0 The number of outpatient care visits in 2006 The total expense of outpatient care in 2006 Receiving inpatient care in 2006; yes = 1, others = 0 The total length of stay of hospitalization in 2006 The total expense of hospitalization in 2006 Gender of the elderly is male; yes = 1 other = 0 The elderly age 75 or above; yes = 1 other = 0 The elderly’s education is elementary school or below; yes = 1, others = 0 The elderly’s education is senior or junior school; yes = 1, others = 0 The elderly’s education is college or above; yes = 1, others = 0 The elderly is living alone; yes = 1, others = 0 Household monthly income is below 50,000; yes = 1, others = 0 Household monthly income is between 50,00 and 99,99; yes = 1, others = 0 Household monthly income is 100,000 or above; yes = 1, others = 0 The elderly lives in Ilan County, Keelung City, Taipei City, Taipei County, Taoyuan County, Hsinchu City, Hsinchu County, and Miaoli County; yes = 1, others = 0 The elderly lives in, Taichung City, Taichung County, Chunghua County, Nantou County, and Yunlin County; yes = 1, others = 0 The elderly lives in Chiayi city, Chiayi County, Tainan City, Tainan County, Kaohsiung City, Kaohsiung County, and Pingtung County; yes = 1, others = 0 The elderly lives in, Taitung County, Hualien County, and Penghu County; yes = 1, others = 0 The elderly’s health status is good, very good or extremely good; yes = 1, others = 0 The elderly’s health status is moderate; yes = 1, others = 0 The elderly’s health status is bad; yes = 1, others = 0 The elderly is having no difficulty in eating, bathing, dressing, toileting, going to bed, and walking in doors; yes = 1, others = 0 The elderly is taking influenza vaccine in the past one year; yes = 1, others = 0
Ref.: the reference group.
curative care use and which are found to be additional characteristics: activities of daily living (ADLs) and influenza vaccine. In other words, to reduce the potential endogeneity of preventive care use in the decision of medical care use, we adopt two instrumental variables (IV)—whether an individual has received an influenza vaccine and whether an individual has difficulty in ADLs. A valid IV should have a direct impact on the endogenous variable but have no direct impact on the outcome variable [13,14]. In our model, the IVs have to directly affect the receipt of preventive care but affect the use of curative care only through the use of preventive care. We assume that an individual who tends to take preventive actions such as receiving an influenza vaccine is more likely to receive preventive health examinations as well, while difficulty in ADLs might prevent an individual from taking effort to go through non-urgent health examinations. On the other hand, the receipt of an influenza vaccine and difficulty in ADLs do not directly reduce or increase the likelihood and expenses of curative care. To verify the validity of the two IVs we adopt, we ran the regression for the use of preventive care and found that two IVs had significant effects on the use preventive care. When we ran the regression for the use of curative care including the two IVs as additional explanatory variables, we found that both the
receipt of influenza vaccine and difficulty in ADL had no direct impact on the use of curative care. We are able to obtain consistent and asymptotically efficient parameters through the two-stage estimation. In the first stage of the empirical estimation, we get the fitted values of Y1 (Yˆ 1 ) by a logistic technique. In the second stage we estimate the structural equation of curative care after substituting Y1 by Yˆ 1 in Eq. (2). We next separate the utilization of curative care into two groups – outpatient care and hospitalization care – in this empirical study in order to differentiate between medical care use and the different severities of illnesses. We apply the two-part model to estimate the relationship between preventive care utilization and curative care utilization, including outpatient care and inpatient care. For outpatient care, the first part estimates the probability of receiving outpatient care and uses a logistic regression. The second part notes the frequency and expenditure of outpatient care visits, with the Poisson regression or negative binomial regression and OLS regression as the proper model. For inpatient care, the first part is the probability of receiving inpatient care and uses a logistic regression. The second part is the total length of stay and expenditure of hospitalization. Here, we respectively implement the Poisson regression or negative binomial regression and OLS regression for the
790
C.-S. Chen et al. / Health Policy 119 (2015) 787–793
estimations. Table 1 contains all variable names and definitions.
lower hospitalization expenses, compared to those who did not use preventive care (21.62% vs. 7.18%; 17.5 vs. 33.3; NTD106,952 vs. NTD171,173).
3. Results 3.2. Regression results 3.1. Descriptive statistics Table 2 provides descriptive statistics of preventive care with mean values and standard deviations for each variable. We find that those who tend to have a higher probability of preventive care use than their counterparts have the following characteristics: elderly people who are male, aged 75 or above, low educated, living alone, with a household income above NTD100,000, living in the eastern area, with a worse health status, with no disability, and have gotten an influenza vaccine. The characteristics of the sample for outpatient care utilization and hospitalization care use are presented in Tables A1 and A2 (see Supplementary material). Around 98% of the elderly who took health examinations have actually sought out sequential outpatient care, while only 37.6% of the elderly who did not have check-ups later utilized outpatient care. The number of outpatient care visits and medical expenses for the elderly with health examinations are much higher than that for those who did not have health examinations: 38 vs. 17.9 and NTD39,438 vs. NTD23,792, respectively. The elderly who used preventive care have a higher percentage of being hospitalized, but have a shorter length of stay and Table 2 Descriptive statistics of preventive care. Variables
Gender Male Female Age 75 Below 75 75 or above Education Elementary school High school College Living alone Yes No Household income Up to 49,999 50,000–99,999 100,000 onwards Geographic location North Center South East Health status Good Moderate Bad ADLs No difficulties Difficulties Influenza vaccine Yes No
Preventive care Mean (%)
S.D.
40.14 39.18
0.4904 0.4884
1116 1123
38.72 41.20
0.4873 0.4925
1387 852
40.33 36.91 37.58
0.4907 0.4833 0.4860
1773 317 149
48.03 38.71
0.5007 0.4872
229 2010
38.68 41.48 48.72
0.4872 0.4934 0.5020
1758 364 117
42.46 37.06 36.85 49.23
0.4945 0.4835 0.4827 0.5038
961 483 730 65
38.58 38.41 43.30
0.4871 0.4867 0.4959
801 893 545
40.67 33.54
0.4913 0.4729
1923 316
46.45 30.11
0.4989 0.4590
1309 930
N
Table 3 reports the result of factors affecting preventive care utilization. For the elderly who live alone, have no ADLs problems and have influenza vaccine, with a household income above NTD100,000 and with a worse health status, we find that they have a higher likelihood of using preventive care. Compared to those who live in the north, elderly who are located in the central or southern area are less likely to use preventive care. Table 4 shows the estimated effect of preventive care utilization for the elderly and the socio-demographic characteristics on outpatient care utilization in terms of the probability of seeking care, doctor visits, and medical expenses. Few factors are found to be significantly and consistently related to outpatient care utilization, with the most crucial factors being preventive care, household income, and regional inequality. This study finds that preventive care utilization has strongly positive significant (P < 0.001) coefficients on the likelihood of seeking outpatient care, suggesting that those elderly who use preventive care are more likely to seek outpatient care. Preventive care is also found to have a positive impact on both the number of outpatient visits and medical expenses. This indicates that once the elderly have had outpatient care, those who use outpatient care services tend to have a higher number of outpatient visits and bear greater medical expenses. This finding fails to reinforce our hypothesis that a negative association between preventive care utilization and curative care utilization. Table 4 also presents the result of hospitalization care regarding the probability of being admitted to a hospital, Table 3 Factors affecting preventive care utilization. Variables
Preventive care Coef.
Constant Gender Age 75 Education High school College Living alone Household income 50,000–99,999 100,000– Geographic location Center South East Health status Moderate Bad ADLs Influenza vaccine N * ** ***
P < 0.10. P < 0.05. P < 0.01.
S.E.
−1.13 0.09 0.05
0.17*** 0.09 0.09
−0.21 −0.18 0.43
0.13 0.19 0.14***
0.08 0.45
0.12 0.20**
−0.35 −0.30 0.21
0.12** 0.10** 0.26
0.01 0.18 0.32 0.73 2239
0.10 0.12 0.13** 0.09***
C.-S. Chen et al. / Health Policy 119 (2015) 787–793
791
Table 4 Effect of preventive care on outpatient care and hospitalization care. Number of outpatient
Outpatient care use
Constant Preventive care Gender Age 75 Education High school College Living alone Household income 50,000–99,999 100,000– Geographic location Center South East Health status Moderate Bad * ** ***
Log (expense of outpatient)
Hospitalization
LOS
Log (expense of hospitalization)
Coef.
S.E.
Coef.
S.E.
Coef.
S.E.
Coef.
S.E.
Coef.
S.E.
Coef.
S.E.
−0.44 2.59 0.06 0.10
0.23* 0.51*** 0.09 0.09
2.95 0.86 −0.01 −0.03
0.11*** 0.24*** 0.04 0.04
9.26 0.99 −0.01 −0.09
0.19*** 0.40** 0.07 0.07
−1.74 −0.22 −0.05 −0.02
0.32*** 0.72 0.13 0.13
4.03 −3.2 0.32 0.27
0.36*** 0.80*** 0.16** 0.15*
11.45 −2.01 0.55 0.19
0.42*** 0.92** 0.15*** 0.15
−0.14 0.04 −0.16
0.13 0.19 0.15
−0.20 0.38 0.08
0.20 0.24 0.22
0.24 0.28* 0.24
−0.53 0.26 0.24
0.26* 0.30 0.24
-0.52 0.38
0.22 0.34 0.21* 0.19 0.35
0.09 0.11 0.07
0.06 0.08 0.07
0.21 0.21 0.08
0.10** 0.14 0.11
0.05 0.50 0.39
0.12 0.23**
0.11 −0.10
0.06** 0.09
0.06 0.07
0.10 0.13
-0.25 0.51
0.19 0.26**
-0.34 0.94
−0.42 −0.27 −0.28
0.12*** 0.11** 0.27
0.03 0.11 −0.05
0.06 0.05** 0.12
−0.02 0.07 −0.14
0.10 0.08 0.17
−0.36 −0.16 0.46
0.18* 0.15 0.33
−0.19 −0.07 0.12
0.20 0.17 0.36
−0.39 −0.28 −0.45
−0.01 0.05
0.10 0.12
0.08 0.09**
−0.03 0.21
0.15 0.17
−0.27 0.09
0.17 0.19
0.17 0.12
0.08 0.53
0.01 0.05
0.05 0.06
0.06 0.20
0.22 0.27***
0.18 0.20
P < 0.10. P < 0.05. P < 0.01.
length of stay, and medical expense. We find that the relationship between preventive care and the probability of hospitalization care use is significant and negative, but it fails to reach the lowest significant level (P < 0.1). Although those who receive prevent care do not appear to reduce the likelihood of being hospitalized, they do seem to reduce hospitalization care use in terms of length of stay and total medical expenditures. The coefficient of preventive care is, as expected, found to be statistically significant negative, which reinforces our hypothesis that a negative association between preventive care utilization and curative care utilization. Household income is another crucial factor that consistently affects the frequency of hospitalization care use. 4. Discussion This study examines those factors influencing the demand for preventive care services and the linkage of preventive care service utilization to subsequent medical care service utilization, including both inpatient and outpatient services, among the elderly people in Taiwan. German et al. [15] show that individuals aged 65 and above who utilize preventive care services have a better health status and lower mortality rates than those who do not utilize such services. Nakanishi et al. [16] study the above relationship using city-level data in Japan and find that the utilization rate of preventive care is negatively associated with the utilization rate of inpatient care, as well as average inpatient and outpatient expenditures, but positively associated with the utilization of outpatient care. Tian et al. [12] find that people aged 50 and over in Taiwan who received preventive care services have lower probabilities of utilizing inpatient care services, suggesting a substitution relationship between the two services. Burton et al. [17] note that among people aged 65 and above in the U.S.,
compared to those who do not utilize preventive care services, those who do utilize the services have a lower number of outpatient visits. Unexpectedly, the elderly who use preventive care utilization does not show any obvious effect on reducing outpatient care use in this study. Conversely, it even significantly demands more outpatient care utilization in terms of the probability of seeking outpatient care, the number of outpatient care visits, and the total expenditures of outpatient care. We nevertheless have found negative relationships between the utilization of preventive care and hospitalization care in terms of length of stay and medical expenditures, though we did also find a positive link between the utilization of preventive care services and the probability that one will be hospitalized. It is possible that the elderly people who used preventive care would also be more likely to use curative care than their counterparts. Both the outpatient care utilization including the probability of seeking outpatient care, the number of outpatient care visits, the total expenditures of outpatient care and the hospitalization care utilization including the probability of seeking hospitalization care increased with those who had preventive care. However, on average, the elderly people who used preventive care tended to have 16 shorter hospitalization stays and NTD64,220 lower hospitalization expenditures than their counterparts (17.5 vs. 33.3; NTD106,952 vs. NTD171,173). This finding demonstrates that the elderly people who used preventive care service reduce later utilization of curative care, particularly for those who suffering from severe illness, leading to longer length of stay and higher medical expenditures through hospitalization care. A possible reason for an increase in outpatient care utilization and a reduction in hospitalization care utilization could be due to early detection of health problems and undergoing more-timely medical treatments with outpatient care
792
C.-S. Chen et al. / Health Policy 119 (2015) 787–793
and hospitalization care, which mitigates the dramatic increase in length of stay and medical expenditures later on. The National Health Insurance system that Taiwan implemented in 1995 now provides comprehensive care for its citizens who previously had no coverage at all and dramatically reduce their out-of-pocket medical expenditures. Furthermore, free preventive care is also provided for the elderly. However, no more than 35% of Taiwanese elderly actually utilize preventive care. One barrier is regional disparity. In the central and southern areas, the elderly are less likely to seek preventive care than those in the northern area, because more medical resources are centralized in the northern area where the capital city (Taipei City) is located. In non-northern areas, elderly people who consider traveling to be an inconvenience may not feel like using preventive services if they are not nearby. To further raise the accessibility of preventive care, time barriers need to be reduced to the home or community where elderly live as well as free shuttle services could be offered. Generally, the Bureau of NHI uses various incentives, including higher outpatient physician fees, no physician visit limits for payments, and increases in payments for circuit medical services, to encourage local independent practitioners to provide medical services in remote areas. Most of Taiwan’s elderly, afraid of facing the fact that they may have contacted or developed some type of serious illness, may refuse to get the check-ups they need. This fear of finding out the truth – a form of denial – prevents them from using preventive services and causes them to postpone needed medical attention as their health deteriorates. Taiwan’s health authorities should respond to these cultural and very common human responses to fear of bad news by educating the elderly and emphasizing the importance of preventive care. If the elderly fully understand the importance of preventive care for their health in reducing serious illness, they would more than likely prefer to receive preventive care to protect themselves from getting sick instead of seeking curative care after an illness has occurred. The BNHI needs to put forth more efforts at increasing the elderly’s awareness of the importance of preventive care and at disseminating information on the benefits of the NHI’s elderly preventive care package. Doing so can help lead the way to greater improvements in the health of Taiwan’s elderly population. Three major limitations of this study should be noted. First, the data we use are based on a cross-sectional design with a sample of only 2239 elderly people. Second, regarding the instrument variables we adopted, there might exist confounders that are associated with both the instrument and the outcome variables [18]. For example, health consciousness can be the instrument-outcome confounder. Those individuals who are more health-conscious (i.e., who care about their own health a lot) are more likely to take influenza vaccine, and are more likely to have healthy lifestyle and thus require less curative care. In that case, the estimates we obtain from the IV method could be biased. However, given the availability of our dataset, we are unable to address this by fully-accounting for every single potential instrument-outcome confounders. This limitation, along with the nature of our cross-sectional
dataset, makes the causal relationship difficulty to testify. Third, from the data we only know whether or not the respondent utilizes the standard package of preventive care services provided by the NHI. Future research with more detailed data on preventive care could explore the link between specific kinds of preventive care and subsequent utilization of medical care. 5. Conclusions Whether the provision of free preventive care for the elderly under the National Health Insurance system has reduced the risk for curative care use raises much concern in Taiwan, as it does in other countries. This study has analyzed the relationship between preventive care utilization and curative care utilization for the elderly by examining the impact of health examination utilization on the utilizations of outpatient care and inpatient care. Data come from the 2005 National Health Interview Survey and National Health Insurance Research Database. A two-stage model is used in the estimation. This study shows that a negative association between preventive care use and inpatient care use does exist in Taiwan’s health care system. The risk of having a serious illness, which leads to longer hospitalization stays and higher medical expenditures, drops if the illness can be detected through preventive care and treated as soon as possible. In order to both improve the health of the elderly and reduce the escalation of medical expenditures due to aging, including preventive care in the health insurance system is a very effective strategy. Conflict of interest statement None declared. Funding sources This work was supported by the National Science Council Taiwan through grant NSC 101-2410-H-305-054-MY3 and is based on data from the 2005 National Health Interview Survey (NHIS) and the National Health Insurance Research Database (NHIRD) in Taiwan. The NSC financial support and the NHIS and NHIRD data provision are gratefully acknowledged. Acknowledgements Constructive comments of editor and anonymous referees are gratefully acknowledged. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/ 10.1016/j.healthpol.2015.01.004. References [1] Council for Economic Planning and Development. Health and vital statistics (1): General health statistics. Taiwan: Department of Health, Executive Yuan Republic of China; 2012.
C.-S. Chen et al. / Health Policy 119 (2015) 787–793 [2] Bureau of Health Promotion. Elderly Health Promotion Plan 2009–2012. Taiwan: Bureau of Health Promotion, Department of Health, Executive Yuan Republic of China; 2009. [3] Department of Health. National Health Expenditure 2011. Taiwan: Department of Health, Executive Yuan Republic of China; 2011. [4] Kenkel D. The demand for preventive medical care. Applied Economics 1994;26:313–25. [5] Casey MM, Call KT, Klingner JM. Are rural residents less likely to obtain recommended preventive healthcare services? American Journal of Preventive Medicine 2001;21(3):182–8. [6] Yu SM, Bellamy HA, Schwalberg RH, Drum MA. Factors associated with use of preventive dental and health services among U.S. adolescents. Journal of Adolescent Health 2001;29:395–405. [7] Jenkins CNH, Le T, Mcphee SJ, Stewart S, Ha NT. Health care access and preventive care among Vietnamese immigrants: do traditional beliefs and practices pose barriers? Social Science & Medicine 1996;43(7):1049–56. [8] Ross JS, Bernheim SM, Bradley EH, Teng HM, Gallo WT. Use of preventive care by the working poor in the United States. Preventive Medicine 2006;44:254–9. [9] Powell-Griner E, Bolen J, Bland S. Health care coverage and use of preventive services among the near elderly in the United States. American Journal of Public Health 1999;89(6):882–6. [10] Wu S. Sickness and preventive medical behavior. Journal of Health Economics 2003;22:675–89.
793
[11] Chen JY, Diamant A, Pourat A, Kagawa-Singer M. Racial/ethnic disparities in the use of preventive services among the elderly. American Journal of Preventive Medicine 2005;29(5):388–95. [12] Tian WH, Chen CS, Liu TC. The demand for preventive care services and its relationship with inpatient services. Health Policy 2010;94:164–74. [13] Wooldridge JM. Introductory econometrics: a modern approach, 5th ed. Boston: Cengage Learning; 2012. [14] Newhouse JP, McClellan M. Econometrics in outcomes research: the use of instrumental variables. Annual Review of Public Health 1998;19:17–34. [15] German PS, Burton LC, Shapiro S, Steinwachs DM, Tsuji I, Paglia MJ, et al. Extended coverage for preventive services for the elderly: response and results in a demonstration population. American Journal of Public Health 1995;85(3):379–86. [16] Nakanishi N, Tatara K, Fujiwara H. Do preventive health services reduce eventual demand for medical care? Social Science & Medicine 1996;43(6):999–1005. [17] Burton LC, Steinwachs DM, German PS, Shapiro S, Brant LJ, Richards TM, et al. Preventive services for the elderly: would coverage affect utilization and costs under Medicare? American Journal of Public Health 1995;85(3):387–91. [18] Garabedian LF, Chu P, Toh S, Zaslavsky AM, Soumerai SB. Potential bias of instrumental variable analyses for observational comparative effectiveness research. Annals of Internal Medicine 2014;161:131–8.
