How much does it cost to survive in the intensive care unit?

Clare L Hibbert BA (Hons)

Senior Researcher

Medical Economics and Research Centre, Sheffield (MERCS)

Intensive Care Unit, Royal Hallamshire Hospital

Glossop Road, Sheffield S10 2JF

United Kingdom

Tel +44 (0) 114 2713510

e-mail: medeconctr@aol.com

Margaret Corcoran BA (Hons)

Researcher

Medical Economics and Research Centre, Sheffield (MERCS)

Intensive Care Unit, Royal Hallamshire Hospital

Glossop Road, Sheffield S10 2JF

United Kingdom

Tel +44 (0) 114 2713510

e-mail: medeconctr@aol.com

Joanne Dean BA (Hons)

Research Associate

Medical Economics and Research Centre, Sheffield (MERCS)

Intensive Care Unit, Royal Hallamshire Hospital

Glossop Road, Sheffield S10 2JF

United Kingdom

Tel +44 (0) 114 2713510

e-mail: medeconctr@aol.com

David Edbrooke FRCA

Consultant in Anaesthesia and Intensive Care

Medical Economics and Research Centre, Sheffield (MERCS)

Intensive Care Unit, Royal Hallamshire Hospital

Glossop Road, Sheffield S10 2JF

United Kingdom

Tel +44 (0) 114 2713510

e-mail: medeconctr@aol.com

Elizabeth Coates BA (Hons)

Research Associate

Medical Economics and Research Centre, Sheffield (MERCS)

Intensive Care Unit, Royal Hallamshire Hospital

Glossop Road, Sheffield S10 2JF

United Kingdom

Tel +44 (0) 114 2713510

e-mail: medeconctr@aol.com

Philip Jacobs D Phil

Professor of Health Economics

Institute of Health Economics

Edmonton

Canada

e-mail: Philip.Jacobs@Ualberta.CA

Abstract

Aim: To determine the cost of survivors and non-survivors in the Intensive Care Unit (ICU).

Design: Retrospective, observational analysis of individual intensive care patient costs.

Setting: Adult six bedded general ICU in a university hospital located in the United Kingdom.

Patients: Two hundred and seventy patients consecutively admitted over a one-year period (1^st April 1996 - 31^st March 1997) to the ICU at the Royal Hallamshire Hospital, Sheffield (UK).

Measurements: Individual total costs per patient were measured using an activity-based costing system. Length of stay and ICU mortality data were obtained for each patient.

Results: The mean (SD) total cost of survivors was £3,514.72 ± 6801.24 and the mean total cost of non-survivors was £3,573.41 ± 5,201.56 using a study by Glance et al, the cost implications of using a scoring system to withdraw therapy at forty-eight hours were estimated. When non-survivors having a length of stay < 48 hours and non-survivors having a length of stay > 48 hours were analysed separately, a significant difference (p<0.0001) was observed between mean total cost of survivors compared to non-survivors having a length of stay > 48 hours. The mean total costs of non-survivors were reduced from £3,573 to £1,189.

Introduction

Intensive care is a specialty which monitors and supports failing vital functions in acutely ill patients in order to perform medical or surgical therapies [1]. The survival rate of patients treated in an adult general Intensive Care Unit (ICU) varies between approximately seventy-five [2] and eighty-seven percent [3]. Studies of specific disease categories, such as sepsis, report survival rates as low as forty percent [4]. Although average ICU survival rates are relatively high in comparison to high-risk groups such as sepsis patients, they are not as representative as hospital survival rates. Hospital survival is based on the number of patients surviving both intensive care and hospital stay, i.e. those patients well enough to return home. A number of studies confirm that hospital survival rates are substantially lower than ICU rates. For example, a study of ICU patients having a length of stay greater than fourteen days, observed an ICU survival rate of seventy-five percent [5]. The survival rate at hospital discharge however, was sixty-three percent. Similarly, a study of sepsis and non-sepsis patients observed a survival rate of eighty percent in the non-sepsis group at ICU discharge [4]. The survival rate of these patients at hospital discharge was only sixty-one percent.

The cost of intensive care provision in the UK has been estimated at £675 million per year [6] and in the US, approximately $62 billion [7]. If we estimate ICU survival rate to be approximately seventy-five percent, then £483 million is spent per year on survivors and £162 million is spent on non-survivors in the UK. In the US, $46 billion is spent on survivors and $16 billion is spent on non-survivors. A number of studies have examined strategies to reduce costs, by changing the way care is provided to critically ill patients [8]. In particular, efforts have been made to assess patients having a prolonged length of stay [9] [10] or patients requiring prolonged mechanical ventilation [11]. The former of these studies assessed whether it is worthwhile to continue treatment for patients with a length of stay greater than fourteen days in the ICU. The study found that the total cost of intensive care in patients with a prolonged length of stay having care withdrawn, to be $156,465, as opposed to $1,917,382 for the treatment of patients with a prolonged length of stay whose care was not withdrawn. However, the study found that 44% of patients survived at twelve months and that mean quality of life scores at twelve months did not differ between patients having a prolonged length of stay and patients having a shorter length of stay. A different study found a survival rate of 30% in patients having a length of stay greater than fourteen days in a surgical ICU [10].

Although the cost implications of withdrawing treatment have been investigated in a number of studies, there still remains the problem of how decisions of withdrawal are made. The cost-effectiveness, using decision analysis techniques of withdrawing care from ICU patients predicted to have a high probability of death has been investigated [3]. Two alternate strategies were evaluated, the first involving patients receiving ICU care until they were discharged, died or had care withdrawn based on subjective clinical criteria. The second strategy was identical to the first, with an additional factor where the use of a prognostic scoring system (APACHE II) after forty-eight hours in the ICU for determining probability of death was determined. Patients having a predicted risk of mortality greater than ninety percent had care withdrawn. The study found that not using a prognostic scoring system as the basis for withdrawing care resulted in a slightly higher survival rate (87.2 percent vs. 86.85 percent). This study concluded that existing scoring systems are not ideal for accurately predicting death, however we were interested in knowing the cost implications of withdrawing therapy at a set time in our own patient population..

Methods

Total patient costs of ICU care were determined for each individual patient using an activity based costing system. Information relating to the care of individual patients was extracted from a patient data management system (PDMS) located in the ICU. The PDMS has a computer terminal at each bedside and a central file server on which all data are stored. Data from bedside monitoring equipment were acquired automatically and nursing and medical staff entered information relating to therapeutic interventions manually. The activity based costing system, pre-configured with over four hundred activities of care was used to assign costs against the care delivered to each patient. An activity of care is defined as any patient related task requiring the use of ICU resources [6]. Activities of care included all drugs, treatments, major monitoring procedures used on the ICU and background nursing care. Total patient related costs of care for each individual patient were derived by summating the cost of each activity of care delivered to that patient. Average daily costs were also determined for each patient. The length of stay and survival of each individual patient were also obtained from the PDMS.

Patients were categorised by their survival status into three groups. The first group included non-survivors with a length of stay less than forty-eight hours, the second group included non-survivors with a length of stay greater than forty-eight hours and the third group were survivors. For non-surviving patients with a length of stay greater than forty-eight hours, the cost of the first forty-eight hours and the additional costs of care incurred after the initial forty-eight hours were determined.

Results

The mean (SD) total costs of care for survivors and non-survivors were £3,514.72 ± 6,801 and £3,573 ± 5,202 respectively. Average daily costs were £613 ± 262 for survivors and £1,001 ± 655 for non-survivors. The sum of the total costs for all survivors was £695,915 and for all non-survivors was £257,285 equating to seventy-three and twenty-seven percent of total cost respectively. Of all patients studied, seventy-four percent survived on ICU discharge. Approximately sixteen percent (n=44) of ICU admissions stayed in the ICU for less than forty-eight hours and died. Ten percent (n=28) of patients studied stayed in the ICU for greater than forty-eight hours and died. The mean total cost per patient of non-survivors with a length of stay less than forty-eight hours was £1,271 ± 932 and the mean total cost of non-survivors with a length of stay greater than forty-eight hours was £7,192 ± 6,896. Mean total costs of the three groups are shown in figure one. Figure two shows the proportion of total ICU expenditure spent on survivors and non-survivors, segregated into patients with a length of stay less than or greater than forty-eight hours. The net total cost of non-survivors with a length of stay greater than forty-eight hours (for all patients added together) for the first forty-eight hours of stay in the ICU was £45,251, and the net total cost of the subsequent days of stay in the ICU was £156,122. Had it been possible to accurately predict the mortality of these patients, and as a result withdraw care within forty-eight hours of their stay in the ICU, the cost of subsequent days of stay in the ICU (£156,000) could have been saved.

Discussion

The results of this study show that non-survivors with a length of stay in the ICU greater than forty-eight hours are more expensive to treat than survivors. If it had been possible to accurately predict the survival of patients within the first forty-eight hours of ICU stay, and as a result withdraw treatment in patients expected to die, the estimated cost saving would be £156,000. This money could have been used to treat between forty-four and one hundred and thirty-one additional patients depending on whether or not they survived.

Other studies incorporating cost data into analysis of outcome measures report varying results. The cost per day of ICU patients that died in the ICU was determined to be £816 (95% confidence interval = £649 - £982). The average cost per day of survivors however was £550 (£498 - £601), approximately £300 less than non-survivors [12]. A study of the cost per day of medical and surgical ICU patients compared to non-ICU days found that the average cost per day in the ICU for medical survivors was $1,357 (95% confidence interval = $1,396 - $1,540) and surgical survivors was $1,501 ($1,484 - $1,639). Compared to the cost per day of non-survivors, medical patients were more expensive ($1,502 ($1,260 - $3,207)) and surgical patients were less expensive ($1,463 ($1,236 - $2,064). Some studies have attempted to look at survival rates and costs per survivor associated with particular diagnostic groups [13] [14] [15]. One of these studies [14] looked at the patient related costs of treating sepsis patients. Patients were analysed by group according to the day on which sepsis was diagnosed. Patients septic on admission to the ICU and patients who became septic on day two in the ICU had an ICU survival rate of fifty percent. Patients that became septic after their second day in the ICU had a survival rate of sixty percent and non-sepsis patients had a survival rate of twenty percent. The median (interquartile range) total cost per patient of survivors was $7,745 ($3,802 - $25,519) in group one, $13,074 ($5,793 - $13,105) in group two and $23,699 (11,717 - 28,805) in group three. The median (interquartile range) total cost per patient of survivors in the non-sepsis patients was $1,583 (918 - 2,566). The median (interquartile range) total cost of non-survivors in comparison to survivors was lower in group one ($1,865 (963 - 3,329)), higher in group two ($17,857 (8,214 - 22,235)) and lower in group three ($15,639 (13,031 - 18,773)). The median (interquartile range) total cost of non-survivors in the non-sepsis group was $2,029 (1,274 - 3,063), higher than survivors.

Further examples of reducing costs through changes in care delivery include a study of the efficacy of an outcomes managed approach, using critical pathways for weaning patients from prolonged mechanical ventilation. The study observed reductions in duration of mechanical ventilation, hospital length of stay and cost per case, however these differences were not significant [11]. A similar study however found that the use of clinical pathways for patients having cerebral revascularisation or femoral revascularisation did not impair quality of care, did not lead to higher morbidity or mortality yet produced significant cost savings to the hospital (fifty-nine percent reduction). Further issues concerning the delivery of care that may be investigated for their ability to influence cost include admission, discharge and triage criteria [16] however, there is little rigorous scientific data to validate criteria and guidelines in these areas. This may be due to the inherent ethical problems associated with selective provision of critical care.

When considering the cost implications of intensive care delivery, not only should the survival rate be taken into account but also quality of life. Factors which improve ICU survival can be misleading, not only in the sense that patients may only survive a short length of time following ICU discharge, but also in that the quality of life following treatment in the ICU may be poor. In a study of the outcome of 1308 intensive care patients where patients were followed up for eight years following ICU discharge, ICU survival was observed to be eighty-two percent and hospital mortality was twenty-nine percent [17]. Further analysis of this data found that one-year after ICU admission, the survival rate had reduced to fifty-eight percent. Patients were asked to place themselves into one of the following categories by postal questionnaire: normal, limited activity, still ill (bed-ridden or convalescent), or still in hospital. Of the survivors, only forty four percent were recorded to have normal activity and twenty-six limited activity one year after admission to the ICU. It was noticed that increasing age and medical patients were associated with lower levels of activity and increased mortality.

Patients’ increasing expectations of health care and an ageing population [18] are conducive to continual increases in expenditure on intensive care. Technological advances further exacerbate the problem of escalating costs [19]. Limited resources necessitate the development of appropriate triage policies to direct resources towards patients more likely to benefit. Unless very reliable scoring systems are developed, practice is likely to continue in its current state.

References

1. Ferdinande, P., Recommendations on minimal requirements for Intensive Care Departments. Intensive Care Med, 1997. 23: p. 226-232.

2. Shiell, A., et al., An evaluation of the costs and outcome of adult intensive care in two units in the UK. Clin Intensive Care, 1990: p. 256-261.

3. Glance, L., T. Osler, and T. Shinozaki, Intensive Care Unit prognostic scoring systems to predict death: A cost-effectiveness analysis. Critical Care Medicine, 1998. 26(11): p. 1842-1849.

4. Edbrooke, D., et al., The patient related costs of care for sepsis patients in a United Kingdom adult general intensive care unit. Critical Care Medicine, 1999. 27(9): p. 1760-1767.

5. Fakhry, S., K. Kecher, and R. Rutledge, "Survival, Quality of life and charges in critically ill surgical patients requiring prolonged ICU stays". "The Journal of Trauma, Injury, Infection and Critical Care", 1996. 41: p. 999 - 1007.

6. Edbrooke, D., et al., A new method of accurately identifying costs of individual patients in intensive care: the initial results. Intensive Care Medicine, 1997. 23: p. 645 - 650.

7. Sonnefield, S., et al., Projections of national health expenditures through the year 2000. Health Care Finance Review, 1991. 13: p. 1-27.

8. Rapoport, J., D. Teres, and S. Lemeshow, Can futility be defined numerically. Critical Care Medicine, 1998. 26(11): p. 1781-1782.

9. Heyland, D.K., et al., Is it worthwhile to continue treating patients with a prolonged stay (>14 days) in the ICU? Chest, 1998. 114(1): p. 192-198.

10. Becker, G., G. Strauch, and H. Saranchak, Outcome and cost of prolonged stay in the surgical ITU. Arch Surg, 1984. 119: p. 1338-1342.

11. Burns, S., et al., Design, Testing and Results of an Outcomes-Managed Approach to Patients Requiring Prolonged Mechanical Ventilation. American Journal of Critical Care, 1998. 7(1): p. 45-57.

12. Ridley, S., M. Biggam, and P. Stone, A Cost-benefit analysis of intensive therapy. Anaesthesia, 1993. 48: p. 14-19.

13. Schapira, D., et al., "Intensive care, survival, and expense of treating critically ill cancer patients". JAMA, 1993. 269(6): p. 783-786.

14. Stevens, V., C. Hibbert, and D. Edbrooke, Evaluation of proposed casemix criteria as a basis for costing patients in the adult general intensive care unit. Anaesthesia, 1998. 53: p. 944-950.

15. Valta, P., et al., Acute respiratory distress syndrome: Frequency, clinical course, and costs of care. Critical Care Medicine, 1999. 27(11): p. 2367-2374.

16. Task Force of the American College of Critical Care Medicine, S.o.C.C.M., Society of Critical Care Medicine, Giudelines for intensive care unit admission, discharge, and triage. Critical Care Medicine, 1999. 27(3): p. 633-638.

17. Dragsted, L. and J. Qvist, Outcome from intensive care. I. A 5 year study of 1308 patients: methodology and patient population. Eur J Anaesthesiology, 1989. 6: p. 23-37.

18. Schneider, E. and J. Guralnik, The ageing of America: Impact on health care costs. Journal of the American Medical Association, 1990. 263: p. 2335 -40.

19. Sibbald, W., et al., "New technologies, critical care and economic realities". Critical Care Medicine, 1993. 21: p. 1777 - 1780.