Universal Health Coverage (UHC) can be defined as provision of quality health care for all people in a country that meets their basic health needs without them being exposed to financial hardship as a result of the cost of care. Despite international movements towards UHC, health resources are tightly limited in developing countries, including Thailand [1]. Resources must be carefully allocated and spent in a cost-effective manner, particularly in the areas of maternal and child health (MCH) where effective interventions are required [2, 3, 4]. The Millennium Development Goals called for a 75% reduction in maternal mortality between 1990 and 2015 (MDG 5). The World Health Organization has recommended that antenatal care services should be free of charge, planned and implemented within the community, and should yield evidence-based quality of care [3]. Thailand has become a good example for embedding UHC tracking into broader health system performance assessment [5].
Although Thailand had achieved the MDG 5, maternal mortality rate (MMR) is at 48/100,000 live birth, and infant mortality rate is at 14/1000 live births since 2009 [6], it is not clear how the universal coverage scheme (UCS) in Thailand—established in 2003—which allowed the country to achieve UHC— contributed to sustainability or any improvement of the outcomes. Under the UCS, resources allocated for MCH aimed to cover costs of disorder prevention and health promotion under benefit packages for family planning, antenatal care for pregnant women, and wellbaby clinics. This care could be provided at the primary care level, in health centers, or in public and private hospitals. In addition, the coverage included medical care to prevent and treat antenatal complications, and to ensure the safety of child delivery and newborn care in an inpatient setting. The global budget under the UCS was also allocated for inpatient obstetric services for high risk pregnancy and treatment of related complications. These services could be provided by hospitals at secondary- and tertiary-care levels. The reimbursements were based on the Thai diagnosis-related-group (Thai-DRG) system. In addition, the Ministry of Public Health needed to invest in MCH infrastructure, and be responsible for the cost of health care personnel. In case of morbidity and mortality as a result of adverse events in medical management, patients or their families could file requests to the National Health Security Office (NHSO) to consider payment for damage liabilities.
To our knowledge, the MCH costs under the UCS in Thailand have not yet been reviewed. Therefore, the present study aimed to explore how much the health system in Thailand spent on maternal and newborn health care under UHC. In other words, we explored the cost of care, including costs in preventing morbidity and mortality, costs in appraising and managing quality, and costs that were the result of failure to prevent morbidity and mortality, including liability claims. We also examined whether increasing costs of preventing the problems in MCH were related to decreasing costs of correcting the MCH problems (failure costs). Insights into the costs and cost structure might ensure financial sustainability and demonstrate whether moving towards UHC contributed to improvement in the MCH in Thailand.
After review and approval of the study protocol by the Institutional Review Board of the Faculty of Medicine, Chulalongkorn University (approval No. 332/2012, IRB No. 177/55) and permission from National Health Security Schemes Office, a cost-identification study was conducted by retrospective review of secondary data using the protocol as follows.
Maternal and newborn health care services provided by the hospitals and health centers under the Ministry of Public Health (MOPH) of Thailand in 20072011. The study setting included the providers in 75 provinces of Thailand, excluding Bangkok and Buengkan province. Bangkok was excluded because of its fragmented health services, whereas the public system of MOPH predominated in other provinces. Buengkan province was a new established province, separated from Nongkai province in the year 2011. Therefore, the Buengkan data was combined with Nongkai for comparability across the study period.
Applying a payer perspective, all payments, charges, and reimbursed expenses made by payers of the system were assumed to be costs. The costs of maternal and newborn health care were defined and categorized using a prevention–appraisal–failure cost of quality (PAF) model [7, 8]. According to the PAF model, prevention costs are incurred to provide services that lead to desirable outcomes, while minimizing undesirable ones. Appraisal costs are incurred in identifying, verifying, checking or evaluating services to assure quality of products or services. Prevention costs and appraisal costs produce positive effects on quality of care. By contrast, failure costs are costs incurred because of errors or defects that needed correction. Failure costs also include costs related to any claims consumers make for defective products or damages they may have from using them. The sum of all three categories of the costs become the costs of quality (CoQ). However, it is expected that effective spending on prevention and appraisal costs should reduce failure costs. The relevant costs are defined as follows.
Hospital payments were allocated from the global inpatient budget based on the DRG adjusted relative weight (adjusted RW), a unit of health care resource measurement of treatment of patient in each DRG. The payments were calculated from relative use of resources in managing each inpatient case in each DRG and were adjusted for low and high outliers according to an outlier trim point (OT) in the length of hospital stay (LOS), which was divided into three intervals: LOS greater than OT and lower than 2OT (OT < LOS ≤ 2OT), LOS greater than 2OT and lower than 3OT (2OT < LOS 3OT), and LOS greater than 3OT (LOS > 3OT). For each version of Thai-DRG system, outlier values were statistically calculated for medical cases and surgical cases [11].
In addition, the failure costs also include liabilities related to maternal and newborn care paid because of adverse events in medical management as indicated in the National Health Security Act. Details of the cost variables and their sources were summarized in
Cost variables and data analyses DRGs, diagnosis-related groups; NHSO, National Health Security Office; RW, relative weight is a unit of health care resource measurement of treatment of patient in each DRG = mean charge per DRG/aggregate mean charge; MOPH, Ministry of Public Health; adjusted RW is RW of DRG adjusted according to NHSO methodology [9, 11]. The payments were adjusted for low and high outliers according to an outlier trim point (OT) in the length of hospital stay (LOS), which was divided into three intervals: OT < LOS ≤ 2OT, 2OT < LOS ≤ 3OT, and LOS > 3OT. For each version of Thai-DRG system, outlier values were statistically calculated for medical cases and surgical cases.Variables Data Source Analytical method Data collection P1a: Cost of care based on benefit packages (e.g. antenatal care, family planning, post-natal care, vaccine) Value of budget allocation for providing services according to benefit package Budget allocation according to benefit package report Summation of prevention–promotion–Express Demand budget allocation per province NHSO budget allocation P1b: Reimbursement for prenatal complications Reimbursement according to DRG compared with adjusted RW In-patient database according to DRG Summation of reimbursement with adjusted RW standard price NHSO paid out each year Collect from NHSO reimbursement P2a: Investment in service infrastructure Investment budget MOPH’s health-resource database Standard price from Budget Bureau Bureau of Policy and strategy MOPH P2b: Health personnel Salaries of obstetric health personnel Specialist payment record Annual summation of medical staff payment Data from Bureau of Policy and Strategy, MOPH A1: Cost according to service quality indicators Payment made to service units based on quality indicators Pay for performance Expense according to quality indicators NHSO annual budget allocation EF1: Cost from comprehensive obstetric service with complications or newborn service with complications Reimbursement according to DRG compared with adjusted RW In-patient database according to DRG Summation of reimbursement with adjusted RW standard price NHSO paid out each year Collect from NHSO reimbursement EF2: Liabilities according to NHSO Act Payment for Liabilities Liabilities fund Summation of total Liabilities NHSO’s Law Office
The study spanned a 5-year period in which the general inflation ranged between –0.9%–3.8% [12]. All analyses are presented at their nominal rates for ease of understanding of descriptive information, and to avoid assumptions on any differential effects of general inflation and medical inflation, and used a province as a unit of analysis. The approximate rate of exchange rate at the end of 2007–2008 of 33.0 baht per US dollar was applied for international presentation. After data checking and verification, inpatient reimbursement data collected from patients’ case mix information records from 2007–2011 were entered into a MySQL database (Oracle, Cupertino, CA, USA) and imported as a large format text file. Then, data were processed with an SQL command set specifically written to process data for data cleaning, resulting in 2,146,938 records for newborn services, and 2,839,816 records for obstetric care, or 4,986,754 records in total. Combining provincial data, the data were analyzed using descriptive statistics, including percentage, mean and standard deviation. To examine whether increases in prevention and appraisal costs might lead to decreases in failure costs at a provincial level, univariate linear regression analyses were conducted to explore associations between year-on-year changes in the prevention (P) and appraisal (A) costs, and year-on-year changes in failure (F) costs. All of the analyses were conducted using STATA software (version 11.0; College Station, TX, USA). To protect patients’ rights to privacy and ensure confidentiality, the researchers were blinded to all patient identifiers.
As shown in Table 2, the per capita operating costs of providing maternal and child health services related to the benefits under the universal coverage scheme (UCS)—or P1 in
Amount of budget by type of Cost of Quality (US$ per Universal Coverage per person per year) CoQ, Cost of Quality; P+A+F, Prevention-Appraisal-Failure; The average reimbursement rate of the Universal Coverage System in 2009 was US$245.45 per adjusted relative weight of diagnosis-related groups; payments were adjusted for low and high outliers according to an outlier trim point (OT) in the length of hospital stay (LOS), which was divided into three intervals: OT < LOS ≤ 2OT, 2OT < LOS ≤ 3OT, and LOS > 3OT according to National Health Security Office methodology [9, 11]. The exchange rate at the end of year 2007"2011 was on average 33.0 Thai baht per US dollar.Type Categories 2007 Mean (min, max) n = 75 2008 Mean (min, max) n = 75 2009 Mean (min, max) n = 75 2010 Mean (min, max) n = 75 2011 Mean (min, max) n = 75 CoQ Total cost of quality (CoQ = P+A+F) 10.31 (6.45, 16.97) 12.47 (6.61, 18.64) 12.76 (8.15, 19.10) 12.98 (9.45, 18.76) 14.87 (6.64, 21.27) P Prevention cost (P=P1+P2) 7.77 (4.12, 14.48) 9.63 (4.39, 17.52) 9.81 (5.61, 15.78) 9.41 (6.36, 14.55) 11.69 (4.73,18.27) P1 Operating cost (P1=P1a+P1b) 4.16 (2.15, 5.32) 5.43 (3.15, 6.32) 5.59 (3.26, 6.38) 5.88 (3.75, 6.47) 8.04 (5.18, 8.98) P1a Benefit package: Prevention and prenatal care 1.13 (0.70, 2.36) 1.19 (0.94, 2.45) 1.25 (1.00, 2.61) 1.27 (1.09, 2.52) 1.32 (1.30, 2.98) P1b Reimbursement for prenatal complications 3.03 (1.70, 3.36) 4.34 (2.14, 5.17) 4.34 (2.13, 4.61) 3.91 (1.09, 4.52) 4.98 (1.94, 5.11) P2 Spending on service infrastructures (P2=P2a+P2b) 3.61 (1.44, 4.45) 4.20 (2.12, 5.06) 4.23 (1.24, 5.09) 3.52 (2.14, 4.25) 3.65 (1.09, 4.56) P2a Cost for specialist health personnel 1.42 (1.06, 4.32) 2.77 (1.15, 3.11) 3.11 (1.53, 3.21) 1.85 (1.14, 3.18) 1.26 (1.02, 2.97) P2b Investment for equipment 2.29 (1.03, 4.15) 1.53 (1.25, 3.24) 1.12 (1.03, 2.15) 2.67 (1.94, 3.97) 2.39 (1.94, 3.28) A Appraisal cost A1 Cost of pay for performance 0.63 (0.33, 1.03) 0.60 (0.24, 0.73) 0.60 (0.33, 1.24) 1.19 (0.55, 1.48) 0.63 (0.33, 0.82) F Failure Cost (F=EF1+EF2) 1.92 (1.06, 3.67) 2.24 (1.06, 4.18) 2.35 (1.30, 4.61) 2.48 (1.45, 4.64) 2.56 (1.58, 4.61) EF1 DRG-based payment for inpatient care 1.83 (1.00, 3.55) 2.18 (1.06, 4.15) 2.27 (1.30, 3.48) 2.39 (1.45, 3.55) 2.47 (1.58, 4.12) EF2 Liabilities 0.09 (0.00, 0.11) 0.06 (0.00, 0.03) 0.08 (0.00, 0.11) 0.07 (0.00, 0.09) 0.09 (0.00, 0.09)
The linear regression analyses of the year-on-year changes in the failure costs and the year-on-year changes in prevention and appraisal costs for providing maternal and child health services, as shown in
Pairwise univariate linear regression analyses of the year-on-year changes of the failure costs and the year-on-year changes in prevention costs and appraisal costs (n = 75 provinces in Thailand)Year-on-Year Regression coefficient ( 95% Confidence intervals Lower Upper 2008 vs 2007 –1.16 –1.44 –0.72 0.002 2009 vs 2008 –1.28 –1.84 –0.62 <0.001 2010 vs 2009 –1.12 –1.73 –0.74 <0.001 2011 vs 2010 –0.47 –1.62 +0.68 0.43
In light of the Universal Health Coverage in Thailand, we found that at least over the period from 2007 to 2011, the country increased spending on maternal and child health services, both in preventing the MCH problems and in correcting the problems. Although the nation-wide numbers showed increases in all three categories of the prevention–appraisal–failure (PAF) costs, the negative associations between the year-on-year changes indicated that for each province, those with increases in prevention and appraisal costs, tended to have decreases in the failure costs. This implied effective allocation of MCH resources.
We demonstrated Thailand has had continuous improvement in maternal and child health outcomes.
The overall failure costs in MCH also rose. Deep investigation of the reimbursement database showed that, during the 5-year period, there were increasing numbers of abortions and related complications, and ectopic and molar pregnancies, summed annual rates for O03–8 at 7.7%, 8.0%, 8.1%, 8.3%, and 8.4%, for the years 2007 to 2011 respectively. There were also increasing numbers of cesarean sections, from 21.1% to 23.7% of all deliveries. Although the increasing failure costs might indicate more problems in maternal and newborn care, the overall maternal mortality rates (MMR) during 2007 to 2011 decreased (62.51, 58.25, 46.69, 45.06, and 36.69 per 100,000 live births, respectively) [13].
The increase in the prevention and appraisal costs in each province were likely associated with reduction in the failure costs as indicated in
Not only are the 3 coefficients with statistical significance in
There were some other limitations to our findings. Although using claim data might help us explore the entire situation in terms of resources spent, the question needs further study to examine epidemiological details not available in the database to provide us with greater insight into what did happen more, and what less. More information on resource allocation within each province would, then, provide opportunities to determine how to allocate and use resources efficiently to provide effective maternal and child care. Finally, we used the perspective of the payer, namely the National Health Security Office of Thailand and the Thai Ministry of Public Health, in studying costs. It is possible that looking at costs from a provider or patient consumer perspective might yield different conclusions. For example, the liability claims under the NHSO Act was only a proxy of the cost of damages caused by morbidity and mortality. The amounts of compensation were fixed by regulations and judgement of committees who performed assessments. Moreover, we did not account for how much preventing maternal and child morbidity and mortality might help reduce indirect costs for patients’ families. Future studies in these areas are warranted to determine best practices at the provincial level, to learn from and share among others nationally and internationally, and to add sufficient evidence to convince policy makers to invest more in MCH.