Current Issue

Volume 64, Number 5, September 2012

 
Estimating Patient-Level Costs for Acute and Sub-Acute Inpatient Services in Thailand
Orathai Khiaocharoen, B.N., Ph.D.,Supasit Pannarunothai, M.D., Ph.D,Arthorn Riewpaiboon, B.Pharm, M.Sc., Ph.D.,Chairoj Zungsontiporn, M.D.
Research Development Division, Phitsanulok Provincial Health Office, Centre for Health Equity Monitoring, Faculty of Medicine, Naresuan University, Faculty of Pharmacy, Mahidol University, Central Office for Healthcare Information, Thailand.

Abstract

Objective: To determine patient-level cost of acute and sub-acute inpatient services and to examine factors predicting cost for patients requiring sub-acute care.
Methods: The study design was prevalence-based cost-of-illness approach. Subjects were 2,419 inpatients requiring sub-acute care in 2 regional hospitals. The costs of services were calculated employing standard costing methods. Inpatient costs were prospectively recorded from July 2008 to February 2009. Micro-costing approach was employed under a provider perspective. Paired-t test was used to evaluate the significant difference.
Results: The average cost during the sub-acute phase was much lower than that in the acute phase (14,877 vs. 23,089 Baht). Nursing service constituted the highest cost (42.6%) in sub-acute phase, whereas direct rehabilitation costs was only 2%. Burn with low physical function had the highest average costs (67,243 Baht). The average cost was significantly higher than the average charge (p<0.001). The physical function at admission, Rehabilitation Impairment Category (RIC) and length of stay (LOS) were the key determinants of sub-acute service costs (adjusted R2=0.688; p<0.001).
Conclusion: The cost in sub-acute phase was lower than the cost in acute phase. Physical function, RIC and LOS could be used to predict cost of sub-acute inpatient services.
Keywords: Patient-level costing, sub-acute inpatient, rehabilitation cost, Thailand
Page: 000 - 000

Go to top of page


INTRODUCTION

Estimating cost is important in the assessment for improvement of health system performance.1 Most published studies present the average costs per hospitalisation or cost per inpatient day. However, different patients have different health care needs depending upon their health conditions. Therefore, a patient-level costing system is needed for proper resource allocation.

Sub-acute care covers a variety of services provided to patients who recover from acute illness and need additional support to recuperate before discharge from the hospital such as rehabilitation, palliative, and geriatric care. Sub-acute care is increasingly in demand due to increases in the number of elderly,2 chronic diseases,3 accidents and gains in survival of patients with catastrophic injury.4 This has become an important concern in most health care systems.5

Currently, sub-acute inpatient service in Thailand is included in acute care services. Payment for sub-acute inpatients is inadequately covered within the diagnosis related group (DRG) payment which has been designed for acute inpatient care. In addition, information on patient-level cost of sub-acute service is lacking. This paper aimed to present the cost of sub-acute care compared with cost of acute care for inpatients who needed rehabilitation after acute care. Moreover, factors the predicting cost of sub-acute care were explored. The relationship between cost and charge was also examined.

Go to top of page


MATERIALS AND METHODS

This study was designed as a prevalence-based cost-of-illness from a provider perspective.6This study focused on inpatient rehabilitation services after acute episodes. Doctors’ decisions to refer patients to rehabilitation services for functional restoration were used as a criterion for triggering sub-acute care. The Naresuan University Ethics Committee for Research on Humans approved this study.

Data source and study population
This study was conducted in two regional hospitals in Thailand from July 2008 to March 2009. Each hospital had about 800 beds with a separate rehabilitation wards. Both have more than 1,500 outpatient visits per day and 60,000 inpatient admissions per year with occupancy rates about 90 percent.

Data were obtained from multiple sources: 1) administrative data both from computerized databases and paper documents covered patient demographic characteristics, medical treatments including operations, discharge diagnoses and costing data, 2) additional data using forms unique to this study included case summary, nursing activities, rehabilitation services, and functional status data.

For functional status data the Barthel Index (BI) and the Mini-Mental State Examination Thai 2002 version were used. The BI score ranged from 0 to 20 and total MMSE score ranged from 0 to 30. For each subject, BI and MMSE were performed at the beginning of the sub-acute phase, at discharge from the hospital only BI was performed.

The data of laboratory investigation, medical procedures, medical devices and drugs delivered to patients were prospectively collected during the study period. Total nursing activities of concerned wards and total rehabilitation services were collected from the recording books of each ward and each rehabilitation unit. Charge data of all patients in the hospital came from the hospital’s electronic data set.

The subjects were patients older than 17 years old diagnosed with diseases which needed intensive rehabilitation care and stayed in hospitals more than 2 days after acute care. Patients who were referred to other hospitals or died during the study period were excluded. There were 2,419 inpatients with diagnoses at admission as mentioned, 222 patients were excluded because they stayed in hospital less than three days after acute care, 7 died, 31 were referred to other hospitals, and 126 were excluded because their diagnoses were not confirmed at discharge. Finally 2,033 subjects remained. (Table 2)

For functional status data the Barthel Index (BI) and the Mini-Mental State Examination Thai 2002 version were used. The BI score ranged from 0 to 20 and total MMSE score ranged from 0 to 30. For each subject, BI and MMSE were performed at the beginning of the sub-acute phase, at discharge from the hospital only BI was performed.

Go to top of page

Cost analysis
The methodology used to estimate the cost of individual patients was the micro-costing or bottom-up method.

First, the standard top-down costing method using financial data, and total hospital costs were allocated down to the department level. Each hospital department was classified as supporting (non-patient care) or patient-care cost centres (medical departments and wards). The direct cost of each cost centre was the sum of capital, material and labour costs. The labour cost was the sum of salaries, wages, overtime payments and all fringe benefits (such as housing) of personnel in each cost centre. For persons who worked in more than one cost centre, a proportion of working time in each cost centre from self-reporting was used to apportion the labour cost. Material costs consisted of utilities and other materials consumed by each cost centre. Other materials were office materials, household materials, maintenance materials, medical materials (drugs excluded), and scientific materials. Capital costs were depreciation of office area and durable goods used in each cost centre. The direct costs of supporting cost centres were transferred to be indirect costs of patient-care cost centres using a simultaneous cost allocation method. Finally in each patient-care cost centre, there were direct and indirect costs of capital, material, and labour costs and the total costs were the sum of all these costs. The total costs of these cost centers were reallocated to service groups based on reimbursement schedules of the Civil Servant Medical Benefit Scheme.7

Second, unit cost was calculated. An average method8 was applied for the service groups with one service or homogeneous service, e.g. hotel cost per day. For service groups with heterogeneous services, the ratio of cost to charge (RCC), and relative value unit (RVU)9 were applied. The RCC for each group was directly computed from the total cost and total charge, and then the actual service charges were multiplied by the RCC to get the service costs. The costs of drug, laboratory, and other service groups were calculated via RCC. The costs per activity for nursing and rehabilitation were estimated by RVU. A service with RVU 2 would require twice as many resources as compared to a service with RVU 1.

The final step was a micro-costing approach. It is a bottom-up method to accumulate all costs for each patient. The cost for each service was estimated in the second step. Finally, the cost of services in acute and sub-acute phases of care for each patient was identified by adding up the cost of services provided in each phase.

Statistical analysis
Descriptive statistics including mean, and standard deviation were employed to explain unit costs. Stepwise multiple regression analysis was employed to analyse the relationship between the costs of sub-acute care (dependent variable) and several potential explanatory variables. Independent variables were the predictors of Sub-Acute and Non-Acute Patient Classification (SNAP) case mix such as RIC, functional score at admission to sub-acute phase measured by the Barthel Index, patient age, and length of stay in the sub-acute phase.10 An RIC is a collection of diagnoses related to conditions needing intensive rehabilitation care which are different from each other (Table 2). Altogether, 2,295 ICD-10 codes were defined for the 19 RICs. Paired-t test was employed to evaluate the difference between cost and charge.

Go to top of page


RESULTS

Hospital costs
The total cost of Ratchaburi Hospital was 1,300 million Baht in 2008. This consisted of 87% recurrent cost (37% labour and 50% non-labour) and 13% capital cost. Total cost of Udonthani Hospital was 1,660 million Baht, with 89% recurrent cost (35% labour and 54% non-labour) and 11% capital cost. Costs per bed day were 320 Baht for general wards in both hospitals, 480-590 Baht for private wards, 856-1,075 Baht for ICU. The cost of rehabilitation wards varied from 309 Baht per bed-day at Ratchaburi Hospital to 447 Baht per bed-day at Udonthani Hospital.

Profile of cost for acute and sub-acute care
Table 1 presents cost per admission by service group. The average cost of acute phase was 23,089 Baht, of which 24.4% was instruments, 23.4% operations, 21.2% nursing activities, and 13.5% drug costs. The average cost of sub-acute phase was 14,877 Baht, of which 42.6% was nursing activities, 19.8% room and board, and 15.4% drug costs. On the other hand, rehabilitation cost in the sub-acute phase was only 2%.

Patient characteristics,br> Table 2 shows that more than half of the patients were male (62%). The most common condition was stroke (383 cases; 18.8%), followed by orthopaedic conditions (370 cases; 18.2%) and cardiac disorders (236 cases; 11.6%). The average age was 53.5 years. The average length of stay was 3.2 days in the acute phase and 8.6 days in the sub-acute phase.

Costs for acute and sub-acute inpatient services Table 3 summarizes the average cost in each phase of care by Rehabilitation Impairment Category (RIC). In the acute phase, RIC 76 was the most costly (58,766 Baht), followed by RIC 63 (54,223 Baht), and RIC 77 (43,424 Baht). The highest cost at the sub-acute phase was RIC 78 (66,681 Baht), RIC 64 (38,343 Baht), followed by RIC 65 (20,698 Baht), and RIC 71 (19,332 Baht).   
Cost vs. charges
Table 3 shows that the overall difference between cost and charge by RIC was statistically significant (paired t-test p<0.001). The results also showed that average cost per admission was significantly higher than average charge.

Determinants of sub-acute care cost Table 4 presents the determinants of sub-acute care costs. Since the cost of the sub-acute phase was not normally distributed, a natural logarithmic transformation was undertaken. The fitted model achieved an adjusted R2 of 0.688, and the probability of F-test = 0.000. The significant predicting variables were the Barthel score at admission, RIC, and LOS in sub-acute phase.

Discussion and conclusion This study showed 3 remarkable findings. First, the micro-costing methodology in this study could estimate the cost of care of each individual patient and thus gave more accurate assessment of resource use. Cost components with a high impact on sub-acute care included nursing services, room and board, and drugs. It should be borne in mind that, if reimbursement rates are not correctly set on the basis of resource use, patients may be selected on the basis of financial considerations, not so much on the basis of diseases and servicing capacities.

Second, the costs of the sub-acute phase services were influenced by physical function which needed longer stays for restoring. Today, the main payment method for inpatient care in Thailand is based on diagnosis related group (DRG) including care for sub-acute and non-acute inpatients. The Thai DRG version 4 has only two out of 1,920 DRGs for rehabilitation services. This ignores the value of rehabilitation services in restoring cognitive and physical function of patients, especially those who have suffered stroke, spinal cord injury, traumatic brain injury or amputation. This result suggested that a sub-acute patient case mix system may be a better tool for payment in sub-acute care. In addition, functional status measurement is important in applying appropriate resource allocation according to the needs of patients in different disease groups.

Third, the results of this study indicate that most of the charges quoted were lower than costs. The costing approach in this study supported evidence-based cost accounting to identify the costs related to treatment strategies. The results could be useful for payers such as the National Health Security Office and the Comptroller General’s Department in improving payment so that it is more reasonable.

Despite the strengths of this prospective, patient-level cost study, there were several limitations. First, most of the patient treatments were focused on acute problems. This may not be a good sampling of the sub-acute service since they had too short LOS and lower cost for the sub-acute phase. This was different from a previous study in a rehabilitation hospital in Thailand which showed that the average LOS for rehabilitation was 27+18 (2-74) days.11

Second, no cost methodology is totally perfect. The cost methodology in this study was similar to that of the Australian National Sub-Acute and Non-Acute Patient Classification (AN-SNAP).12 Approaches for capturing cost in heterogeneous service cost centre cover the RCC, RUV, intensity weight, and activity based costing (ABC). For RCC in a cost centre with detailed bill data, it provides a reasonable compromise between accuracy and ease of implementation in estimating patient level costs.13 The RCC was also adopted in examining the average cost per DRG.14 Though RCC was simple to apply, the low resemblance of charge to the cost data was another drawback. The RVU provided a better cost driver approach as it took account of both time and resources consumed for calculating cost per activity.15 Even for the ABC which is considered to be the gold standard methodology, the application is lengthy, complicated, time consuming, and needs detailed information which might not be available or accurately registered.16,17

This study could conclude that cost analysis is important for setting the appropriate price of services and reimbursement rates. These findings generate cost data on individual patients which can help policy makers in considering a reimbursement system according to the phase of services provided: acute and sub-acute phases which consumed different medical resources. The micro-costing approach is very useful in terms of the accurate cost of each service. Although, this cost methodology was complex, standard costing method was necessary. The results can also provide illustrative ideas to both practitioners and managers of hospitals in

cost calculation methodology. In a case mix classification development, cost data at patient level is important in calculating the cost relative weight between classes.18,19 Physical function and RIC which affect sub-acute costs could help further development of this sub-acute case mix classification and reimbursement policy. Hospital information systems should be developed to routinely record activities and cost data for all patients. As the costing system becomes more detailed and accurate, hospitals will be able to make better pricing decisions and can become more efficient in cost and services management. In the Thai health care system, a suitable and practical costing method for similar types of hospitals, similar efficiency of hospitals, and similar contexts for standard costs should be implemented.

Go to top of page


ACKNOWLEDGMENTS

This study was financially supported by grants from the National Health Security Office (NHSO), the Health Intervention and Technology Assessment Program (HITAP), and the Health Promotion for People with Disabilities Programme. The authors are very grateful and wish to thank facilitators, nurses, computer programmers, research assistants and all patients that participated in this study.

Go to top of page


References

1.  Johns B, Baltussen R, Hutubessy R. Programme costs in the econimic evaluation of health interventions Cost Effectiveness and Resource Allocation. 2003;1:1-10.
2.  Jitapunkul S, Chayovan N. National Policies on Ageing in Thailand. 2001 [21/06/2007]; Available from: http://adrf.trf.or.th/Archive_data/LTCTHAILAND.doc.
3.  Bundhamcharoen K, Odton P, Phulkerd S, Tangcharoensathien V. Burden of disease in Thailand: changes in health gap between 1999 and 2004. BMC Public Health 2010;11(53):1-9.
4.  Tanaboriboon Y, Satiennam T. Traffic accidents in Thailand. IATSS research. 2005;29(1):88-100.
5.  Hyatt L. Subacute care: An inportant new trend. Long Term Care Management. 1993;42(42):21-2.
6.  Ekman M. Economic evidence in stroke: a review. Eur J Health Econ Suppl 1. 2004:S74-83.
7.  Cooper J, Suver J. Principles and methods of managerial cost-accounting systems. Am J Hosp Pharm. 1988;45:146-52.
8.  Chiang B. Estimating nursing costs-A methodological review. IJNS. 2009;46:716-22.
9.  Khiaocharoen O, Pannarunothai S, Zungsontiporn C, Riewpaiboon W. Casemix classification payment for sub-acute and non-acute inpatient care, Thailand. J Med Assoc Thai 2010;93(7):849-59.
10.  Kovindha A, Kuptniratsakul V, Dajpratham P, Massakulpan P, Piravej K, Archongka Y, et al. Thai Stroke Rehabilitation Registry (TSRR). J Thai Rehabili Med. 2007;17(1):31-6.
11.  Eagar K, Gordon R, Hodkinson A, Green J, Eagar L, Erven J, et al. The Australian National Sub-Acute and Non-Acute Patient Classification (AN-SNAP): report of the National Sub-Acute and Non-Acute Casemix Classification Study: Centre for Health Service Development, University of Wollongong; 1997. 39-48
12.  Drummond M, McGuire A. Economic Evaluation in Health Care Merging Theory with Practice. New York: Oxford University Press; 2001. 55-101
13.  Shwartz M, Young D, Slegrist R. The ratio of costs to charges: how good a basis for estimating costs? Inquiry. 1995-1996 Winter;32(4):476-81.
14.  Riewpaiboon A, Malarole S, Kongsawatt S. Effect of costing methods on unit cost of hospital medical services. TMIH. 2007;12(4):554-63.
15.  Lin BY, Chao T, Yao Y, Tu SM, Wu C, Chern J, et al. How can Activity-Based Costing methodology be performed as a powerful tool to calculate costs and secure appropriate patient care? J Med Syst. 2007;31:85-90.

16.  Tan SS, Rutten FFH, Ineveld BMv, Redekop WK, Roijen LH-v. Comparing methodologies for the cost estimation of hospital services. Eur J Health Econ. 2009;10:39-45.
17.  Jackson T. Using computerised patient-level costing data for setting DRG weights: the Victorian (Australia) cost weight studies. Health Policy. 2001;56:146-63.
18.  Street A, Dawson D. Costing hospital activity: the experience with healthcare resource groups in England. Eur J Health Econ. 2002;1:3-9.

Go to top of page