Abstract
Objective: This study aimed to estimate the cost of patients admitted to the Intensive Care Unit (ICU) of the Teaching University Hospital of Thessaly (TUHT) in 2006 and to demonstrate discrepancies between actual hospitalisation cost and social funds’ reimbursement.
Methods: Cost analysis was performed using a macro-costing approach, which focused on the estimation of nominal and actual cost per ICU patient. Data were derived from the annual records of resources consumed in each hospital unit and from hospital balance sheets. Sensitivity analysis was also performed by inflating nominal costs to present values.
Results: There were 312 patients admitted to the ICU. Mean actual cost per ICU patient was estimated at €16,516, whereas actual reimbursement from social funds was only €1,671. This means that reimbursement accounted for just 10% of the actual hospitalisation cost. Once nominal costs were inflated to present values, the reimbursement accounted for 25% of the actual hospitalisation cost. The major cost drivers of ICU hospitalisation were personnel costs followed by infrastructure, hotel services and pharmaceutical expenditure. These results may be limited by a lack of consideration for clinical outcomes along with a high level of aggregation in cost data.
Conclusion: Reimbursement should be re-adjusted in order to balance public hospital deficits and make public-private mix viable. This way, intensive care capacity would increase and allow a more equitable distribution of healthcare resources.
Introduction
Utilisation of intensive care has generated a great deal of concern in several countries due to its continuously increasing demand and resource consumptionCitation[1–3]. Fast technological innovations and improved medical practices, in combination with the highly specialised personnel and the intense working conditions in intensive care units (ICUs), are in large part responsible for both increased patients’ survival and rising costsCitation[4–6]. Critical care is estimated as one of the most expensive components of inpatient care, since it consumes a disproportionate amount of healthcare resources. Although ICUs represent a small fraction of hospital beds, they account for approximately over 25% of hospital budgets and up to 2% of the gross domestic product (GDP)Citation[2],Citation[3],Citation[7],Citation[8]. An ICU hospitalisation is three to ten times the cost of a hospitalisation on a regular inpatient unitCitation[9].
The high cost of the ICU has been well documented across different countries and much of the attention has focused on the economic evaluation of critical careCitation[5],Citation[6]. In an era of increased financial scrutiny and competing demands for limited healthcare resources, technology assessment constitutes an especially useful tool for health professionals and decision makers to make more cost-conscious health-related decisions. In Greece, despite the fact that the number of beds in ICUs is insufficient to cover population needs, studies regarding the economics and economic evaluation of critical care are very few. Most of them focus on the unequal geographical distribution of ICUs in the country and the respective resource consumptionCitation[10–13].
Three-quarters of the total ICU beds in the country are found in the greater metropolitan cities of Greece, Athens and Thessaloniki. On the mainland, the region of Thessaly appears to have the lowest distribution of ICU beds, since it only absorbs 3.9% of total ICU bedsCitation[14]. Efforts have been made to increase bed capacity and meet patients’ needs in this region, such as the construction of the Teaching University Hospital and the operation of the respective ICU in 2000. However, only a part of its capacity has been used (8/24 beds) due to a lack of specialised personnel, equipment, etc. Consequently, approximately 40–50% of all patient flows are referred to the respective ICUs in Thessaloniki, demonstrating that the need to expand ICU facilities at the regional level is imperative. ICU beds in Greece constitute approximately 1.9% of total hospital beds, while the respective percentage in the international literature ranges from 12 to 20%Citation[15].
The lack of separate and analytical ICU accounts in the National Health Service (NHS) hospital budgetsCitation[16],Citation[17] is a severe obstacle in the estimation of ICU hospitalisation cost. Consequently, economic evaluation studies are limited and therefore, cannot support a policy makers’ ability to make cost-effective decisions that could serve to expand ICU beds in the country.
The aim of this study was to estimate the cost of patients admitted to the ICU of the Teaching University Hospital of Thessaly (TUHT) in 2006 and evaluate the difference between real and nominal cost. Identification of the ICU patients’ cost could facilitate public–private contracts to expand access to critical care and allow a more equitable distribution of ICU resources among regional authorities.
Methods
This study was conducted at the TUHT ICU. It is the only tertiary university ICU in Thessaly, accounting for 27.6% of the total ICU beds in the region.
Cost estimation of healthcare activity can be carried out using either micro-costing (bottom-up) or macro-costing (top-down) methods. In the former method, costs are derived for each element of intervention: staff time, supplies and medications, diagnostic and laboratory examinations and so on. In the latter, there is no detail available on the cost of every component of the inpatient stay. The degree of aggregation used in this method is high. The type of cost information available usually determines the method that will be used in the analysisCitation[18],Citation[19]. In this case, estimation of cost was performed using a top-down macro-costing approachCitation[9],Citation[20–22] since data were retrospectively collected and patients’ records did not provide analytical resources consumption.
Healthcare utilisation was not reported on a per patient basis but per hospital unit and data were derived from the 2006 budgetary control statements and balance sheets provided by the hospital's finance department. These were provided on a total cost basis. The National Health System (NHS) perspective was adopted. As such, the analysis included the annual cost of resource consumption (supplies, medication, laboratory and medical tests), infrastructure and overhead costs (electricity, watering, heating, building maintenance and repairs), and finally personnel cost (medical, nursing, administrative, paramedical and auxiliary staff). Depreciation of capital assets and machinery was excluded from the analysis.
Estimation of personnel cost was based on the wages of the medical and nursing staff that was fully employed in the ICU. The cost of the paramedical, administrative and other staff was calculated according to the number of patients admitted to the hospital. Laboratory and diagnostic imaging costs were based on the aggregate annual records of the ICU and not on a per-patient basis. Only drug utilisation was calculated per patient using data from the hospital pharmacy. The cost of oxygen and other consumables was obtained form the hospital supplies department on an annual basis. NHS prices have been used in the analysis while drug costs were taken from the wholesale NHS hospital price index. Infrastructure and general overhead costs were allocated with respect to the area occupied (square meters) by the ICU over the total area of the hospital. They were found in budgetary control statements provided by the hospital's finance department.
The analysis focused on the estimation of direct costs separated into nominal and actual cost per ICU patient. Nominal cost is defined as the prefixed price of the per diem payment according to which sickness funds reimburse NHS hospitals for each hospitalisation day of the insuredCitation[12]. This is set at €187.82 per ICU inpatient dayCitation[23]. Actual cost refers to the real patient's hospitalisation cost based on resource consumption during inpatient care.
A sensitivity analysis was performed in order to examine the reliability of the results. The per diem payment of an inpatient day has not been revised since 1991. As such, the difference in costs may be underestimated. By inflating reimbursement to present-day values, the ‘true’ difference between actual and nominal costs can be calculated. Inflation rates used were taken from annual data of the National Statistical Service of Greece (1991–2006)Citation[14].
Results
depicts comparative data among the four prefectures (Larissa, Volos, Karditsa and Trikala) of the Thessaly region. Unequal distribution in ICU beds’ availability and differences in patient flow to the TUHT, bed occupancy and length of stay are observed among the respective ICUs of the region. Specifically, the rate of patient flow in the prefectures’ hospitals are lower than that found in both hospitals of Larissa, where the TUHT is situated.
Table 1. Comparative data among hospitals in the Thessaly region, in 2006.
In 2006, 312 patients were admitted to the ICU of the TUHT, accounting for 2,745 inpatient days. The length of stay in the respective ICU was 8.9 days, almost 2.6 times the average length of stay in the hospital overall (3.4 days). From those patients, only 30% were residents of Larissa, while 70% came from the other prefectures.
The most common medical conditions mentioned were physiological abnormalities such as coma and tachycardia; chronic diagnoses such as renal failure and metastatic cancer; acute diagnoses such as acute renal failure, arrhythmia, gastrointestinal bleeding; as well as cardiopulmonary resuscitation or mechanical ventilation. According to the data collected, ICU patients were classified by medical condition. These included scheduled surgery patients (25%), trauma patients (20%), pneumonia (20%), stroke patients (15%), brain injury (10%) and COPD patients (10%).
Pharmaceutical cost included the cost of any drug used a during patient's overall stay in the ICU. In 2006, approximately 600 pharmaceutical products were used, including antibiotics, antithrombotics and drugs for anaemia or blood substitutes. The cost of antibiotic treatment represented an important part of hospital expenditure. The total annual cost of antibiotics was about €700,000, accounting for 64% of the total ICU pharmaceutical cost. Antibiotic therapy for pneumonia and for surgical patients was the most significant cost contributor. Laboratory and diagnostic costs included tests such as blood gas analysis, magnetic resonance imaging (MRI), computed tomography (CT) scans and more.
According to data shown in , scheduled surgery and trauma patients have the highest share of pharmaceutical expenses in the ICU followed by patients with respiratory diseases and stroke. The mean pharmaceutical cost per patient was similar among the various medical conditions.
Table 2. Pharmaceutical costs* per medical condition in 2006 euros.
depicts the breakdown of total ICU hospitalisation cost. The major cost components were personnel cost (29%) followed by infrastructure and hotel cost (23.8%) and pharmaceutical expenditure (21.2%). In 2006, the total annual cost for all patients’ hospitalisation in the TUHT ICU was estimated at €5,152,962. Overall, mean cost was €16,515.9 per patient during inpatient stay in the ICU. Further investigation of the cost data showed that there were differences in the mean cost per patient among the various staff categories which are related to the staff levels of each category. More specifically, mean cost per patient was estimated at €4,785 out of which nursing staff costs accounted for €2,604.8 and medical staff costs for €1,604.
Table 3. Breakdown of total annual hospitalisation cost* in the ICU in 2006.
presents the mean actual hospitalisation cost per ICU patient, which was estimated at €16,516 and the reimbursement of the social funds (nominal cost) which is only €1,671. In addition, the actual cost per hospitalisation day in the ICU is estimated at €1,856 while the social funds’ per diem payment in the ICU is fixed at €187.8. This result explains hospital deficits to a great extent, and also why social insurance funds struggle to repay medical expenses. Given that medical costs increase over time while the per diem payment has not changed since 1991, the per diem payment and the nominal cost per patient was inflated to present-day values so that the comparison could be realistic. Once nominal costs were inflated to present values, the reimbursement accounted for 25% of the actual hospitalisation cost, showing that the discrepancies between actual and nominal costs still remain, although reduced by 15%.
Table 4. Mean hospitalisation cost* per ICU patient in the TUHL, in 2006 euros.
Discussion
According to the results of a survey conducted by the Hellenic Intensive Care Society (HICS), in 2006, in Greece more than 4,500 patients die every year because of the high number of inoperative ICU beds (>150 beds) in the public hospitals. More precisely, in 2006, 23% of the total ICU beds in Greece were not used and this number increased to 25% in 200824.
The present cost analysis was performed in order to determine resource consumption and to estimate costs of ICU patients admitted to the TUHT in 2006. According to findings of the present study, there is an unequal distribution of ICU beds among the four prefectures of the region with lower availability being found in the Volos and Trikala prefectures. In addition, ICU availability in Thessaly is lower than the national average, despite the higher availability of ICU beds in Larissa and Karditsa, prefectures of Thessaly.
The major cost drivers of an ICU hospitalisation were personnel costs, followed by infrastructure and hotel services costs, and pharmaceutical costs. It is worth noting that the mean pharmaceutical cost per patient among the various medical conditions was similar. However, there were differences evident in the mean cost per patient among the various personnel categories. The mean nursing cost per patient was higher than the mean medical personnel cost. Despite the well-known differences between medical and nursing personnel salaries, the time needed for nursing care per patient per shift (three shifts per day) is greater than the respective time of physicians follow-upCitation[25]. Nursing care is more labour intensive.
This TUHT ICU was selected because it admits patients from the urban and rural areas of the neighbouring regions as well as from the other prefectures of Thessaly. The fact that after more than 6 years of the ICU's operation, only one-third of beds are in use, is of great importance. This issue has to be addressed especially in an era when the government, in order to meet the population needs for critical care, has legislated to reimburse NHS patients’ hospitalisation in private hospital ICUs.
The government's decision to create a public–private mix for the provision of intensive care was faced with scepticism from private ICU providers because of the vast variations between nominal and actual costs of delivering intensive care in the Greek NHS. If private hospitals accepted the proposed contracting of delivering critical care to NHS patients, they would face severe constraints in their budgets due to the existing underpayment of the social fundsCitation[12],Citation[16],Citation[17],Citation[26]. Once costs are adjusted, public–private contracts could be facilitated and expansion of critical care could be achieved. Consequently, ICU bed capacity could be enhanced and a more equitable distribution of resources would be available throughout Greece.
The mean actual daily cost per ICU patient was estimated at €1,856, while the per diem payment is €187.8. The discrepancies between costs of patients in the ICU are evident. More specifically, reimbursement paid by social insurance funds through the per diem payment reflects only 10% of the actual ICU patient hospitalisation cost.
When comparing these results with those reported in the international literature, it was found that, despite the different methodologies used, there are numerous similarities between the various studies. Firstly, the lack of and the unequal distribution of ICU beds in the country has been already reported in a number of studiesCitation[10–12],Citation[16],Citation[17],Citation[26]. Secondly, personnel costs were reported as the main ICU cost drivers in various internationalCitation[4],Citation[8],Citation[21],Citation[22] and national studiesCitation[10–12],Citation[26]. Thirdly, although the cost estimates in this study are based on aggregated costs and this method may be viewed as less accurateCitation[27], they are also comparable to those of other studies. The studies of Singer et al.Citation[4] and Edbrooke et al.Citation[28] quoted the average cost per patient day in an ICU as £1,148 and £1,152 respectively, while Noseworthy et al.Citation[29] estimated daily ICU costs ranging from $2,000 to $3,000 in many US hospitals. Additionally, Zhang et al.Citation[27] report that the mean aggregate costs in an ICU per patient in USA ranges from $1,876 to $2,917. However, they state that their ICU cost estimates are underestimated.
Some methodological limitations should be mentioned. Data used reflect the ICU costs of a single teaching university hospital in a specific region. Therefore, the cost estimations may not be easily generalised to the entire country. However, this study has shed light on the existing situation with regard to the provision of critical care and the related costs, as well as the difficulties of implementing the public–private mix decision. Additionally, the collection of cost data was recorded on an aggregated basis instead of a per-case or per procedure basis, due to the lack of analytical patients’ recordsCitation[30]. This led to the use of a macro-costing approach, which has been widely used by others to approximate the costs of ICUs due to the relative simplicity of the implied data collection. However, it is the bottom-up micro-costing methodology which is considered the gold standard methodology for costing hospital servicesCitation[28],Citation[31–33]. Another important limitation of the economic analysis is the lack of consideration of clinical outcomes and cost-effectiveness criteria of intensive care provision due to the lack of a standardised patient classification system in Greece. However, a broad classification of the ICU patients was performed for the purpose of the cost analysis.
It is believed that comparisons between ICUs internationally could contribute substantially to the evaluation of medical outcomes and to the adoption of evidence-based utilisation patterns through the exchange of clinical and economic information. Moreover, as hospital budgets become more constrained in Greece, fewer internal funds will be available to meet ICU supply. This study provides a detailed breakdown of where costs are incurred in the hospitalisation of patients admitted to the ICU. This could be useful for future research, where costs for ICU cases will be used as an input into economic models. As such, economic evaluations could be incorporated into decision making, which would allow health policy makers to make more cost-conscious decisions. The identification of ICU patients’ costs will increase awareness of real hospitalisation costs and discrepancies with reimbursement could be balanced resulting in substantial reductions in public hospital deficits. Moreover, the integration of an information system could allow the identification of performance indicators and their measurement and consequently, hospital outcomes and related costs could be easily isolated and analysed. Thus, the introduction of a prospective hospital payment system based on patients’ case–mix classification could be partially employed in the NHS hospitals.
Conclusion
Through this study it becomes evident that identification of ICU hospitalisation cost and adjustment of the respective social funds’ per diem payment can facilitate public–private contracts in order to improve access to critical care. The conduction of economic analysis studies is necessary in Greece, regardless of the type of healthcare provision, population group and provider. Such studies may enable health policy makers to incorporate economic assessment in their decision making.
Transparency
Declaration of funding:
This study was not funded.
Declaration of financial/other relationships:
M.G., L.A. and D.T. have disclosed that they have no relevant financial relationships.
The JME peer re1viewers 1 and 2 have not received an honorarium for their review work on this manuscript. Both have disclosed that they have no relevant financial relationships.
Acknowledgment:
The authors wish to thank Christina Vandoros for reviewing this manuscript.
References
- Birnbaum M. Cost-containment in critical care. Critical Care, LM Civetta, et al. JB Lippincott, Philadelphia 1986; 1977–97
- Jacobs P, Noseworthy TW. National estimates of intensive care utilization and costs: Canada and the United States. Crit Care Med 1990; 18: 1282–1286
- Miranda RD, Spangeberg FJ. Aspects of critical care organization in Europe. Principles of Critical Care, JB Hall, GA Schmidt. McGraw Hill, New York 1992; 2280–1302
- Singer M, Myers S, Hall G, et al. The cost of intensive care – a comparison on one unit between 1988 and 1991. Intensive Care Med 1994; 20: 542–549
- Chalfin DB, Cohen IL, Lambrinos J. The economics and cost-effectiveness of critical care medicine. Intensive Care Med 1995; 21: 952–961
- Kirton O, Civetta J, Hudson-Civetta J. Cost effectiveness in the intensive care unit. Surg Clin North Am 1996; 76: 175–199
- Bams JL, Miranda DR. Outcome and costs of intensive care. Intensive Care Med 1985; 11: 234–241
- Halpern NA, Bettes L, Greenstei R. Federal and nationwide intensive care units and healthcare costs: 1986–1992. Crit Care Med 1994; 22: 2001–2006
- Metnitz PG, Lenz K. Patient data management systems in the intensive care – the situation in Europe. Intensive Care Med 1995; 21: 703–715
- Plati C, Lemonidou C, Priami M, et al. The intensive care units in greater Athens: needs and resources. Intensive Crit Care Nurs 1996; 12: 340–345
- Baltopoulos G, Chryssanthou V, Moraitidis D. Intensive care medicine in Greece. Intensive Care World 1991; 8: 140–143
- Leftakis A, Geitona M. Cost analysis and estimation of thoracic surgical patients with lung cancer in Greece: the case of Sotiria ICU. Intensive Crit Care Nurs 2002; 17: 322–330
- Geitona M. Mapping of hospital care in Greece: regional inequalities in health. Iatriki 2001; 79: 259–267
- National Statistical Service in Greece 2006. www.statistics.gr, 02/09/09.
- Groeger JS, Strosberg MA, Halpern NA, et al. Descriptive analysis of critical care units in the United States. Crit Care Med 1992; 20: 846–863
- Geitona M, Yfantopoulos J. Financing public hospitals in Greece, 1922–1992. Iatriki 1995; 67: 133–141
- Liaropoulos L, Tragakes E. Public/private financing in the Greek health care system: implications for equity. Health Policy 1998; 43: 153–169
- Morris S, Devlin N, Parkin D. Economic evaluation methods. Economic Analysis in Healthcare, S Morris, et al. John Willey, Chichester 2007; 312–314
- Smith MW, Barnett PG, Phibbs CS, et al. Micro-cost methods for determining VA healthcare costs. Health Economics Resource Center U.S. Dept. of Veterans Affairs, Menlo Park, CA 2005
- Ridley S, Biggam M, Stone P. Cost of intensive therapy. A description of methodology and initial results. Anaesthesia 1991; 46: 523–530
- Holt AW, Bersen AD, Fuller S, et al. Intensive care costing methodology: cost benefit analysis of mask continuous positive airway pressure for severe cardiogenic pulmonary oedema. Anaesth Intensive Care 1994; 22: 170–174
- Gyldmark M. A review of cost studies of intensive care units: problems with the cost concept. Crit Care Med 1995; 23: 964–972
- Government Gazette 99b/10-2-98.
- Hellenic Society of Intensive Care. Report of the Hellenic Society of Intensive Care, (in Greek), January 2009, http://ygeia.tanea.gr/default.asp?pid=8&ct=98&articleID=5506&la=1, 2/9/09.
- Christoforidou A, Andrioti D, Kyriopoulos J. Health workforce in Greece: Trends and Projections. Human resources supply and cost containment in the health system, D Heffley, et al. FICOSSER- Exandas, Athens 1998; 75–88
- Geitona M, Hatzikou M. Cost estimation of neonatal intensive care in Greece: the case of Athens maternity hospitals. J Med Econ 2007; 10: 273–283
- Zhang B, Wright A, Huskamp AH, et al. Health care costs in the last week of life. Associations with end-of-life conversations. Arch Intern Med 2009; 169: 480–488
- Edbrooke DL, Stevens VG, Hibbert CL, et al. A new method of accurately identifying costs of individual patients in intensive care: the initial results. Intensive Care Med 1997; 23: 645–650
- Noseworthy TW, Konopad E, Shustack A, et al. Cost accounting of adult intensive care: methods and human and capital inputs. Critical Care Med 1996; 24: 1168–1172
- Apostolopoulou E, Nikoloudi P, Georgoudi E, Veldekis D. Utilisation of intensive care unit: a multicentre study in Greece. ICUS Nurs Web J 2005; 23: 1–6
- Edbrooke DL, Wilson AJ, et al. The Sheffield costing system for intensive care. Crit Care 1995; 11: 106–110
- Fein AI. Organization and management of critical care units. Intensive Care Medicine, J Rippe, R Irwin, M Fink, et al. Little, Brown & Co, Boston 1996; 2629–2636
- Jacobs P, Rapoport J, Edbrooke D. Economies of scale in British intensive care units and combined intensive care/high dependency units. Intensive Care Med 2004; 30: 660–664