The Journal of Health, Population and Nutrition icddr,b ISSN: 1606-0997 EISSN: 2072-1315 Vol. 28, Num. 5, 2010, pp. 436-442

Journal of Health Population and Nutrition, Vol. 28, No. 5, September-October, 2010, pp. 436-442

Original Paper

Treatment costs of pneumonia, meningitis, sepsis, and other diseases among hospitalized children in Viet Nam

Dang Duc Anh¹, Arthorn Riewpaiboon², Le Huu Tho³, Soon Ae Kim⁴, Batmunkh Nyambat⁴, Paul Kilgore⁴

¹ National Institute of Hygiene and Epidemiology, Hanoi, Vietnam
² Division of Social and Administrative Pharmacy, Faculty of Pharmacy, Mahidol University, Bangkok, Thailand
³ Khanh Hoa Provincial Public Health Service, Nha Trang, Vietnam
⁴ International Vaccine Institute, Seoul, South Korea

Correspondence Address: Arthorn Riewpaiboon, Division of Social and Administrative Pharmacy, Department of Pharmacy Faculty of Pharmacy, Mahidol University, 447 Sri Ayutthaya Road, Ratchathevi, Bangkok 10400, Thailand, pyarp@mahidol.ac.th

Code Number: hn10057

Abstract

Keywords: Cost-of-illness; Healthcare costs; Meningitis; Pneumococcal diseases; Pneumonia; Sepsis; Viet Nam

Introduction

Globally, nearly two million children die each year due to acute respiratory tract infections (ARIs) ^[1] . Besides, an estimated 150 million episodes of child-hood pneumonia occur, of which 11- 20 million require hospitalization ^[2] . The incidence of pneu-monia among children aged less than five years is estimated to be 0. 29 episodes per child-year in developing and 0. 05 episodes per child-year in de-veloped countries ^[2] . Recent estimates from the United Nations Children′s Fund suggest that Strep-tococcus pneumoniae causes half of all deaths due to pneumonia in children, and 70% of these deaths occur in Africa and Asia ^[3] . A large proportion of these deaths could be averted through routine im-munization using available pneumococcal conju-gate vaccines ^[4],[5] . However, a substantial barrier to the use of pneumococcal conjugate vaccine is the non-availability of adequate information on costs associated with treatment of clinical condi-tions consistent with invasive pneumococcal dis-eases in children. In addition, an increasing num-ber of countries are now requesting or requiring cost-of-illness data on vaccine-preventable diseases before making long-term commitments for the introduction of new vaccines. At present, only a limited number of studies reported costs associated with syndromes of invasive bacterial infections in children from developing countries ^[6],[7],[8] . To ad-dress this gap in information, we determined costs for the treatment of invasive bacterial infection syndromes in children in Nha Trang, Viet Nam.

Materials And Methods

Study design and data collection

We conducted an incidence-based cost-of-illness analysis from the healthcare system perspective, a technique used for measuring the economic bur-den of patients from the onset to the end of illness ^[9],[10] . We selected the Khanh Hoa General Hos-pital (KHGH) for this study in Nha Trang city, the capital of Khanh Hoa province, Viet Nam. During 2005- 2006, a prospective surveillance for clinical syndromes associated with invasive pneumococ-cal diseases was conducted at the KHGH; its results were published elsewhere ^[11] . Before the enroll-ment of study patients, all clinicians and medical staff of the Department of Pediatrics and Infectious Diseases of KHGH were provided with standardized procedures and screening criteria for identifying children with pneumonia, meningitis, sepsis, and other syndromes consistent with invasive bacterial diseases. After the completion of informed consent procedures, appropriate clinical specimens (blood and/or cerebrospinal fluid) were obtained, and par-ents were interviewed to collect demographic in-formation and illness history. All specimens were tested using bacterial culture methods, and selected specimens underwent further testing by antigen de-tection and polymerase chain reaction. Definitions of diseases were based on the published guidelines ^[12],[13] , and patients′ demographics and clinical characteristics were drawn from the published re-search report ^[11] .

Costing methods

For our analysis, cost-of-illness was defined as the sum of direct medical, direct non-medical, and in-direct costs ^[9] . Direct medical cost was defined as costs directly associated with provision of health-care, including costs for detection, treatment, con-tinuing care, rehabilitation, and terminal care. For this study, we focused our analysis on direct medi-cal or treatment costs employing the micro-costing approach that estimates costs of treatment by sum-ming all medical services received by an individual patient ^[14] . Vietnamese children, aged less than six years, are now covered under a national health in-surance supported by the Vietnamese government. In Viet Nam, public hospitals receive government funding in the form of lump sum budget and insurance system reimbursements for delivery of medical services to patients. Hospitals receive reve-nue in the form of out-of-pocket payments directly from patients (approximately 60% of the total ex-penditure). For the health insurance, hospital costs are reimbursed by the Ministry of Health using a fee-for-service system. To collect data on individual patient′s cost, the KHGH uses a FoxPro^; (Microsoft, Redmond, Washington, DC) data-entry system that records costs of services provided (e.g. number of hospitalization days, laboratory tests, radiographic procedures, therapeutic procedures, and drugs) to an individual patient.

To compute costs of each medical service, we used the ratio of costs to charges (RCCs) method ^[15],[16] . In this method, the RCC takes into account all costs incurred by the hospital and the hospital′s total revenue (i.e. patient-charge in the form of out-of-pocket payment from uninsured patients and reimbursement from the Ministry of Health for insured patients). Then, the total annual hospi-tal costs were divided by total annual revenue (or total charges) to derive the overall RCC. We then multiplied charges for each service by the RCC to obtain costs of a service item for each patient. We summed the costs of all medical services and drugs of each patient to obtain total direct medical costs or treatment costs for each patient. Total hospital cost included labour cost of all hospital staff, mate-rial cost (e.g. use of electrical utilities), and capital cost. Capital costs associated with the use of depre-ciable assets, i.e. equipment, buildings, furnishings, and vehicles. The capital cost was calculated based on the straight-line equivalent annual economics-based approach ^[17] . This approach covers both de-preciation cost (the rate at which the capital asset is ′used up′) and opportunity cost (interest) of mak-ing the investment ^[18] . The cost is equally allo-cated to period of useful life. Based on the Ministry of Finance guidelines, we assumed that the useful life of vehicles, equipment, and buildings was 5, 10 and 25 years respectively ^[19] . We covered all items in use, although they were over official working years since they still had opportunity cost ^[18] . We applied a 3% discount rate to our cost analysis ^[18] . The costs were converted from Viet Nam dong to US dollar at the 2006 prices (VND 15, 976. 83=US$ 1) ^[20] .

Analysis of data

Demographic information was described by fre-quency and proportion. We calculated means with standard deviations (SDs) for continuous variables and used multiple regression analysis to explore the factors associated with variations in treatment costs. Covariates considered in this modelling included age (months), duration of stay (days), and diagnosis. In addition, since prices of drugs may vary substantially among hospitals, we performed a sensitivity analysis to explore the effect of variation in prices of antibiotics across a range of prices of antibiotics available from the KHGH databases.

Results

Characteristics of patients

Patients′ demographics are presented in [Table - 1]. In total, 980 children were enrolled in the study. Their average age was 12. 67 months (SD± 11. 38). The majority ( 56. 5%) were male. Sixty percent of the patients were currently breastfeeding, and 42% had received antibiotics during the 14-day period before hospital admission. Of the 980 children, 80. 4% were diagnosed with pneumonia, 7. 8% with meningitis, 3. 2% with very severe disease, and 8. 7% with other diseases. Of the 980 children with a spe-cific diagnosis, 43. 5% were classified as CXR-con-firmed pneumonia, followed by CXR-confirmed severe pneumonia ( 19. 7%) and suspected menin-gitis ( 6. 7%). All the enrolled patients were treated in the inpatient facilities of the KHGH, and the average duration of hospitalization was 6. 48 days (SD± 4. 14) [Table - 2].

Estimation of costs

To convert the payment to economic cost, the cost-to-charge ratio method was applied. The KHGH had total costs of VND 172, 482 million with rev-enue of VND 151, 219 million or a hospital cost-to-charge ratio of 1. 14. Of 48 antibiotics used for treat-ing the study patients, ceftriaxone, cefotaxime, and cefuroxime were the most expensive drugs [Table - 3]. These three antibiotics accounted for 38%, 29%, and 16% of all the drug costs in the study respec-tively. The average treatment or direct medical costs of pneumonia, meningitis, very severe diseases, and other diseases were US$ 31, US$ 57, US$ 73, and US$ 24 respectively [Table - 2]. The cost of drugs represented 40- 63% of the total treatment costs.

The highest costs (mean US$ 164) was incurred for definite (i.e. laboratory-confirmed) meningitis, fol-lowed by probable bacterial meningitis (US$ 121). Results of sensitivity analysis showed that treat-ment costs rose by 8% and decreased by 4% when the maximum and minimum prices of antibiotics were used for analysis instead of mean prices res-pectively. The costs increased by 13% when cal-culated using maximum compared to minimum prices of antibiotics.

To explore the co-factors affecting the treatment costs, we conducted a multiple regression analysis using age, days of hospitalization, and diagnosis as independent variables. All, except days of hos-pitalization, had significant effects on cost. The model had explanation power of 17. 2% (adjusted R ² = 0. 172). Estimated costs regarding age and di-agnosis were constructed. The costs of all disease-groups were inversely related to age-as the age of patients increased, costs decreased [Figure - 1]. When we compared diagnosis and age-groups, the average treatment cost of pneumonia ranged from US$ 18 to US$ 66. Similarly, the cost of treating meningitis ranged from US$ 25 to US$ 190.

Discussion

Characteristics of patients

This study used data on patients from a representa-tive group of hospitalized children in Viet Nam. As in previous studies ^[21],[22] , most ( 94. 6%) children hospitalized with acute respiratory disease were aged less than three years. Analysis of data from the prospective surveillance suggests that most ( 80. 4%) suspected invasive pneumococcal diseases in Viet-namese children manifests initially as pneumonia. While a few comparative studies across develop-ing countries have been performed, data from our surveillance study in Khanh Hoa province suggest that children suffered from various acute lower res-piratory tract diseases due to respiratory bacteria and viruses. While some children with bacterial infections may present with less severe disease due to prior outpatient treatment with antibiotics, such empiric treatment may put the child at risk for prolonged disease with antibiotic-resistant bacteria or respiratory viruses that do not respond to treat-ment with antibiotics ^[23] .

Costing of illness

Two previous studies in Pakistan estimated costs associated with clinical syndromes of invasive pneumococcal disease ^[24],[25] . For hospitalized care, the healthcare provider spent, on average, US$ 71per episode for pneumonia, US$ 235 for severe pneumonia, and US$ 2, 043 for meningitis ^[25] . Results of another study showed that the average treatment cost per episode was US$ 17. 77 for pneu-monia and US$ 125. 29 for severe pneumonia ^[24] . However, direct comparison of healthcare costs across countries is limited by the fact that these costs arise from systems with different approaches to healthcare delivery, treatment patterns, drug-pric-ing, costing of services, and different time periods for data analysis. Nevertheless, a comparison of trends in costs over time and the proportionate distribution of costs are reasonable. As in our analysis, data from Pakistan showed that the cost of treatment of meningitis was higher than that of pneumonia, and cost of treatment of severe pneu-monia was higher than that of non-severe pneu-monia. Our data in Viet Nam also showed that the cost of treatment of meningitis was approxi-mately twice compared to that of pneumonia. In contrast, this difference in the cost of treatment of meningitis compared to that of pneumonia was nearly 30-fold in Pakistan. In the subset of patients with laboratory-confirmed disease, the cost of treat-ment of definite meningitis (US$ 163. 95) was five times higher than that of CXR-confirmed pneumo-nia (US$ 32. 45). Another method of comparison is to compare with the country′s economy indicators. The average treatment cost (US$ 33. 56) was ap-proximately one-tenth of the Viet Nam per-capita gross national product (GNP). The per-capita 1998 GNP for Viet Nam was US$ 350 ^[26] .

Because of limitations in the availability of data, we could not determine an accurate count of material consumption. Instead, we assumed that quantities of materials used were equivalent to the amount of materials purchased. Regarding the cost-to-charge method applied in this study, the accuracy decreased if some patient-services were provided without charge as total hospital costs covered all costs of services produced. Thus, the cost of free services, e.g. health education, would be absorbed by services for which charges were levied. Accuracy was also related to accuracy of price-setting ^[27] . This refers to the relative prices of different servic-es (compared to each other in the same hospital) actually reflects the relative resource-use of those services, even if the level of those prices is much higher than the cost.

We considered resource-consumption patterns and the representativeness of the study hospital com-pared to other hospitals in Viet Nam. In the study hospital, costs of materials ( 54%), labour ( 28%), and capital ( 18%) were all similar to those of Thai-land ^{[27],[28],[29]} . For the cost-to-charge ratio, a 2003 study in the same hospital found that the ratios for the laboratory, imaging diagnosis, and the whole hospital were 1. 45, 0. 74, and 0. 98 respectively ^[30] . The difference might be due to different methods used for determining both costs and hospital rev-enues.

Factors affecting treatment costs

In general, consumption of drugs was positively associated with both body-weight and age. Due to the relatively low explanatory power of the fitted cost model (adjusted R ² = 0. 172), other variables not included in our modelling procedures may explain variations in cost. In addition, the relationship of the independent variables may be confounded by age. For example, younger children may be more vulnerable to pneumococcal diseases than older children. In addition, drug-prescribing behaviours may affect both use of certain drugs and costs of drugs provided ^[31] . We were unable to include factors relating to the attending physician in the model tested.

Generalizability

We believe that data used in our analysis are both reliable and internally valid. The KHGH has a well-organized computerized database, from which all costing data were drawn. Our findings regarding age- and disease-specific treatment costs can be used together with age-specific incidence rates for such economic evaluations as cost-effectiveness, cost-utility, and cost-benefit analyses ^[32] . Age-spe-cific data are also useful for assessing the most cost-effective ages for immunizations.

Conclusions

Treatment cost of pneumococcal diseases was con-siderable in Viet Nam. Costs of suspected/probable and confirmed pneumococcal disease were sta-tistically different as were costs of pneumococcal pneumonia and meningitis, which can differ by as much as seven times. Age of patient and type of disease affect costs of treating pneumococcal dis-eases in Viet Nam. These data will be important in the overall evidence to guide the development of vaccine policy in Viet Nam.

Ackowledgements

The study was supported by the PneumoADIP and by the Governments of Kuwait, Sweden, and South Korea.

The authors thank Dr. Le Thi Phuong, Mr. Nguyen Thanh Hien, Mr. Nguyen Vu Cuong, and Mr. Van Hoi for superb work in data collection and assis-tance for data management and Kathy Murray for editorial assistance.

References

Copyright 2010 - Journal of Health Population and Nutrition

The following images related to this document are available:

Photo images