+Model REDAR-446; No. of Pages 8
ARTICLE IN PRESS
Rev Esp Anestesiol Reanim. 2014;xxx(xx):xxx---xxx
Revista Española de Anestesiología y Reanimación www.elsevier.es/redar
ORIGINAL
Long-term admission to the intensive care unit: a cost-benefit analysis S. Rodríguez Villar a,∗ , R.M. Barrientos Yuste b a b
Intensive Care Medicine Department, Queen Elizabeth Hospital, London, United Kingdom Madrid Care Centre for the Disabled, Madrid, Spain
Received 21 January 2014; accepted 25 February 2014
KEYWORDS Long-stay patients; Intensive care; Cost; Mortality; Functional status post discharge
∗
Abstract Objective: To assess outcomes in long-term ICU patients, with follow-ups carried out at one year post discharge, in order to calculate the costs incurred by the hospital in relation to the benefits gained. Material: Of 3639 patients consecutively admitted over the course of three years to ICU, 235 (6.5%) were assessed for the purposes of the study, having spent a period exceeding 20 days in intensive care. Method: The survey tool used was the Spanish Minimum Data Set (MDS). The length of ICU stay and hospital stay following discharge from ICU were calculated, and one year post discharge the patient/next of kin was contacted in order to carry out a follow-up survey on survival and functional status (according to GOS-E scale). Results: The 235 study patients had a mean stay of 37 days, occupied 34% of ICU beds available and consumed 29% of the ICU’s economic resources ($14,400,175). Their stay on hospital wards was (mean) 33 days. Mortality in ICU and on hospital wards was 40% higher amongst older patients, and those with a higher APACHE II and Charlson index score. Mortality rates were three times higher among neurosurgical patients: mortality at follow-up was 25%, and only 21% recovered an acceptable functional status. Conclusions: Mortality rates in long-term ICU patients are high, both during their hospital stay and in the first year post discharge. Surviving patients do not exhibit a good level of recovery, and consume a large proportion of economic resources. © 2014 Sociedad Espa˜ nola de Anestesiología, Reanimación y Terapéutica del Dolor. Published by Elsevier España, S.L. All rights reserved.
Corresponding author. E-mail address: [email protected] (S. Rodríguez Villar).
0034-9356/$ – see front matter © 2014 Sociedad Espa˜ nola de Anestesiología, Reanimación y Terapéutica del Dolor. Published by Elsevier España, S.L. All rights reserved.
http://dx.doi.org/10.1016/j.redar.2014.02.008
Please cite this article in press as: Rodríguez Villar S, Barrientos Yuste RM. Long-term admission to the intensive care unit: a cost-benefit analysis. Rev Esp Anestesiol Reanim. 2014. http://dx.doi.org/10.1016/j.redar.2014.02.008
+Model REDAR-446; No. of Pages 8
ARTICLE IN PRESS
2
S. Rodríguez Villar, R.M. Barrientos Yuste
PALABRAS CLAVE pacientes de larga duración; Cuidados intensivos; Costes; Mortalidad; Estado funcional después del alta
Ingresos de larga duración en la unidad de cuidados intensivos: análisis de costo-beneficio Resumen Objetivo: Evaluar los resultados de los pacientes de larga duración en la UCI por medio de un seguimiento al a˜ no del alta a fin de calcular los gastos soportados por el hospital en relación con los beneficios obtenidos. Materiales: de los 3.639 pacientes consecutivos ingresados en la UCI durante tres a˜ nos, se evaluó a 235 (un 6,5%) para el presente estudio, todos ellos con una estancia superior a 20 días. Métodos: el instrumento de evaluación fue la base de datos espa˜ nola CMBD (conjunto mínimo de base de datos). Se calcularon la duración de la estancia en la UCI y en la planta después del alta de la UCI y, un a˜ no después del alta hospitalaria, se contactó con el paciente o su pariente más próximo para realizar una encuesta de seguimiento sobre su estado funcional (según la escala GOSE). Resultados: los 235 pacientes estudiados estuvieron ingresados un promedio de 37 días, ocuparon un 34% de las camas disponibles en la UCI y emplearon un 29% de los recursos económicos de dicha unidad (14.400.175$). Su estancia media en planta fue de 33 días. La mortalidad en la UCI y en planta fue un 40% más alta en los pacientes de mayor edad, puntuación del APACHE II e índice de Charlson. Las tasas de mortalidad se triplicaron en los pacientes neuroquirúrgicos. En el seguimiento después de un a˜ no, la tasa de mortalidad fue del 25%, y únicamente el 21% recuperó un estado funcional aceptable. Conclusiones: las tasas de mortalidad en pacientes de larga duración en la UCI son altas, tanto durante la estancia hospitalaria como durante el a˜ no posterior al alta. Los pacientes de larga duración no presentan una recuperación correcta y consumen una gran proporción de recursos económicos. © 2014 Sociedad Espa˜ nola de Anestesiología, Reanimación y Terapéutica del Dolor. Publicado por Elsevier España, S.L. Todos los derechos reservados.
Introduction
Patients and methods
The quality of resources currently available to intensive care services permits the treatment and survival of patients suffering from increasingly more serious and complex conditions. However, the main condition together with the failure of other associated organs often prolong the stay of patients in our units, and just as morbidity and mortality levels rise, so do the care and treatment costs involved. Financed by state taxes, the Spanish Healthcare System is free, and practitioners may use all available resources in their particular healthcare setting, according to their professional criteria. Healthcare expenditure in Spain accounts for 8.4% of Gross National Product (GNP) (54.4% of which is hospitalbased expenditure). The average expenditure in relation to GNP in the EU is 9.6% (OECD Health Data, 2008). Currently, life expectancy in Spain at birth is 81 years. According to some sources, the resources used in the critical care setting consume 15% of the hospital budget.1 In the USA, critical care accounts for 0.7% of GNP.2 The aim of this study is to investigate mortality in patients who remain in the Intensive Care Unit for over 20 days. The study focuses on survivors, following up their progress from the ICU to hospital wards and up to discharge. To determine survival rates and functional status, patients were followed-up one year after discharge from the ICU. The hospital administration provided the researchers with details of the cost of caring for these patients, both in the ICU and hospital wards.
Design An observational, longitudinal, descriptive study using data from patients admitted over a 3-year period to a multipurpose ICU at a tertiary-level hospital. The study was approved by the Independent Ethics Committee of the Hospital Virgen de la Salud in Toledo.
Materials The study was carried out in the Toledo Hospital Complex in Spain, which comprises two large hospitals and several specialist outpatient clinics. The complex has 777 beds with 26 ft multi-purpose ICU beds (3 beds exclusively for scheduled surgery patients). Data from patients admitted in 2010, 2011 and 2012 for a period exceeding 20 days were analyzed. We considered 20 days to be a prolonged period; however, others have applied different criteria in defining what may be considered a longterm patient, and have suggested hospital stays of anything between 7 and 30 days.3---5 Virgen de la Salud Hospital is a regional centre of excellence for neurosurgery. It also boasts a coronary unit and cardiac surgery resuscitation unit that are independent of the main ICU where the study was carried out. Over the 3 years studied, 3639 patients were admitted, of which 235 (6.5%) were hospitalized for more than 20 days (Fig. 1).
Please cite this article in press as: Rodríguez Villar S, Barrientos Yuste RM. Long-term admission to the intensive care unit: a cost-benefit analysis. Rev Esp Anestesiol Reanim. 2014. http://dx.doi.org/10.1016/j.redar.2014.02.008
+Model
ARTICLE IN PRESS
REDAR-446; No. of Pages 8
Long-term admission to the intensive care unit: a cost-benefit analysis
3
3.639 patients are admitted to ICU: .Over a period of 36 months (January 2008-December 2010). . Multiple-trauma, neurocritical, medical and surgical patients are included.
54 patients (23%) die on ICU and without having been discharged from hospital.
>20 days stay in ICU 235 patients (6,5%)
181 patients (77%) do not die in ICU.
161 patients were discharge to the hospital wards: -133 (80%) survived and are discharged at home. -28 (17,5%) died on hospital wards
20 patient sare repatriated to their original hospital
43 MEDICAL PATIENTS (21% DIED): GOSE (1) 9, GOSE (2-5) 11, GOSE (6-8) 21 patients. 153 patients survived: 27 SURGICAL PATIENTS (15% DIED) : GOSE (1) 4, GOSE (2-5) 11, GOSE (6-8) 12 patients.
-It was not possible to contact 27 patients by telephone. Their status is unknown -126 patients make up the final patient sample
56 NEUROSURGICAL PATIENTS (32% DIED): GOSE (1) 18, GOSE (2-5) 30, GOSE (6-8) 10 patients.
Figure 1
Patient flow-chart.
Method The basic survey tool used was a minimum data set (MDS) designed by the intensive Care Unit. Data such as the age, length of stay, origin and destination of the patient were collected in addition to the diagnosis and procedures evaluated according to the International Statistical Classification of Diseases and Related Health Problems (ICD-9), APACHE II, and NEMS score. SOFA is not routinely carried out for all patients, so it was not included in this analysis. Mean length of stay and destination following ICU discharge onto hospital wards were obtained from the hospital MDS. Further information was obtained from clinical reports of each of the 235 selected patients, both on discharge from ICU and from hospital wards, and based on these reports the patient was assigned a score on the Charlson comorbidity index6 (Annex 1). Where no comorbidity existed, the patient was assigned a score of zero. The Hospital’s administration department provided the researchers with information related to the cost of each patient’s care and treatment. In addition to calculating the mean cost of ICU stays, related costs were also categorized according to whether the patient was medical, surgical or neurosurgical. The same criteria were used during follow-up on costs incurred after discharge to hospital wards. The difference in cost between surgical patients and medical patients is due to the cost of the surgical procedures involved and the higher percentage of additional tests performed outside the ICU (generally radiological).
One year after discharge, the patients/next of kin were contacted on the telephone number given on their patient records. The researchers verified whether the patient had survived and assessed their functional status according to the Extended Glasgow Outcome Scale, with a score of 1---87 (Annex 2).
Statistics The statistical analysis was performed using SPSS 12.0 to calculate the mean, standard variation and percentages of the variables under study. For dichotomous variables, the chi-square test was used. The level of statistical significance was p < 0.05.
Results Patients admitted to the ICU for a period of less than 20 days For the 3400 patients hospitalized for less than 20 days over the 3-year study period the mean length of stay was 3.55 ± 4.18 days, with a mortality rate in the ICU of 13.4% and APACHE II score of 14.2 ± 9.0. On hospital wards postICU discharge, mortality was 6.1%, and mean length of stay was 11.6 ± 16.0 days. Of the foregoing patients, 30.1% were admitted from A&E, 15.5% from hospital wards, 34.8% from scheduled
Please cite this article in press as: Rodríguez Villar S, Barrientos Yuste RM. Long-term admission to the intensive care unit: a cost-benefit analysis. Rev Esp Anestesiol Reanim. 2014. http://dx.doi.org/10.1016/j.redar.2014.02.008
+Model REDAR-446; No. of Pages 8
ARTICLE IN PRESS
4
S. Rodríguez Villar, R.M. Barrientos Yuste
surgery, 10.3% from emergency surgery and 9.3% were transferred from other hospitals. By specialty: neurosurgical patients accounted for 28.9%, general surgery 18.4%, other surgery 17.5%, internal medicine patients 9.5%, cardiology 14.1% and other medical patients 10.8%.
Patients admitted to the ICU for over 20 days Of the 3635 patients, 235 were admitted to the ICU for a period that exceeded 20 days, with a mean length of stay of 37.2 ± 27.4 days, giving a total of 8742 days of stay over the three years studied, the equivalent of 7.9 beds/year. In other words, 6.5% of the total number of patients studied occupied 34.3% of the 23 ICU beds available. The 235 patients were categorized by specialty: 95 neurosurgical patients (41.1%), 59 other surgery (25.5%), and 81 medical patients (33.3%). The 95 neurosurgical patients had a mean length of stay of 31.8 ± 27.1 days, the 59 other surgical patients’ stay was 40.7 ± 23.2 days, and the 81 medical patients had a mean stay of 36.8 ± 18.7 days; 54 patients (22.9%) died in the ICU. Table 1 lists the most common procedures and resources used in the treatment of these patients. Table 2 shows the survival rate, age, APACHE II, NEMS and Charlson scores, both in the ICU and on hospital wards, of patients spending more than 20 days in ICU. Age and APACHE II were considered statistically significant (p < 0.5) when comparing survivors with deceased; however, the NEMS score was not found to be significant. According to some studies, SOFA score is a very significant factor; unfortunately, since it was not available for all the patients it could not be included in this study.8 As discussed above, patients were categorized into three pathological groups (neuro-surgical, other surgical and medical), among which advanced age, high APACHE II and Charlson scores continued to be significant factors for
Table 1 Some of the resources used (procedures and techniques) were. Long term invasive mechanical ventilation Tracheostomy Extracorporeal kidney support Extracorporeal liver support Poli-transfusions Cerebral arteriography-embolization Intracraneal pressure monitoring S.Ganz catheter CT-scan (Computed tomography) MRI-scan (Magnetic Resonance Imaging) Some of the most common surgical procedures were: 58 32 6 7 1 1 4
Craneotomy Laparotomy Thoracotomy Vascular surgery Nefrectomy Permanent pacemaker Orthopaedic surgery
99, 3% 57,1% 16,7% 3,7% 47,8% 7,6% 31,2% 10,8% 78,5% 10,4%
mortality. The neurosurgical patients were younger, with a high percentage of head injury cases. They had lower APACHE II and Charlson index scores and less associated comorbidity. As a result, fewer suffered from multiple organ failure, and the ICU mortality rate was therefore lower: 10.5%. In the other two groups (other surgical and medical), where multiple organ failure is more frequent and Charlson index scores higher, the ICU mortality rate was three times that of neurosurgical patients (31%). Patients were also older and have higher APACHE II scores.
Follow-up of patients transferred to hospital wards Of the 181 patients discharged from the ICU, 20 were transferred directly to another hospital (see Table 2). The 166 patients transferred to wards in Virgen de la Salud Hospital included 72 neurosurgical patients, 34 surgical and 54 medical patients. Mean hospital ward stay was 33.6 ± 37.6 days, which, added to the length of stay in the ICU, gives a mean hospital stay of 70.4 days. The hospital stay of neurosurgical patients was double that of the other categories, which is significant since they present more functional limitations. Of the 235 long-term patients, 82 (34.9%) died in hospital; 54 in the ICU, and 28 on hospital wards. Thirteen of the 28 deceased patients died on the wards in the first 1.8 ± 2.1 days. Death after a hospital stay of 34.4 days might be considered post-ICU limitation of therapeutic effort (LTE).
Use of resources and costs incurred in the treatment of ‘long-term’ patients The bed-day cost in this ICU is estimated to be D 1250/day (approximately $1637). The 8742 bed-days of the 235 longterm patients, therefore, represents an overall cost of D 10,927,500 (approximately $14,264,250) (Table 3). Considering these figures and the fact that 54 patients died in the ICU, the total cost per surviving patient per ICU stay is D 60,372, or approximately $78,807. The cost of running the Intensive Care Unit in 2010, 2011 and 2012 is estimated by hospital management to be D 37,695,557, (approximately $49,004,224). Long-term patients, 6.5% of the total, accounted for 29.5% of the total ICU budget. The bed-day cost of the three patient categories were: D 1430/day (approximately $1873/day) for neurosurgical patients, D 1260/day (approximately $1650) for other surgical patients, and D 1036/day (approximately $1357) for medical patients. The highest costs are those of staffing, while for surgical specialties costs increase due to certain surgical procedures and additional studies (see Table 3 for a break-down of costs per number of patients and length of stay in days). The hospital finance department also supplied us with main ward bed-day costs for the different specialties: neurosurgical D 707/day (approximately $926), other surgical D 586/day (approximately $767) and medical D 386/day (approximately $505).
Please cite this article in press as: Rodríguez Villar S, Barrientos Yuste RM. Long-term admission to the intensive care unit: a cost-benefit analysis. Rev Esp Anestesiol Reanim. 2014. http://dx.doi.org/10.1016/j.redar.2014.02.008
+Model
ARTICLE IN PRESS
REDAR-446; No. of Pages 8
Long-term admission to the intensive care unit: a cost-benefit analysis Table 2
5
235 patients spending 20+ days in ICUa,b
Patients
ICU-patients
Median ICU stay
age
NEMS*
APACHEII
Charlson
Ward patients
Median Ward stay
age
Charlson
Survived Died
181 54 (23%)
34.4 39.7
56.7 69.5
44.5 46.3
17.4 22.5
1.0 1.2
133 28 (12%)
35.5 29.9
55.2 63.3
1.3 1.1
Patientsc
ICU-patients
age
NEMS*
Apache II
Charlson
Ward patients
Median Ward stay
age
Charlson
Neurosurgical patients (n = 95) Survived 85 31.2 Died 10 (10%) 33.7
52.1 59.2
43.1 42.5
16.6 17.3
0.75 0.80
60 12 (13%)
43.2 50.4
52.4 60.0
0.75 0.80
Patientsd
age
NEMS*
Apache II
Charlson
Ward patients
Median Ward stay
age
Charlson
Patients from other surgeries (n= 59) Survived 39 40.4 Died 20 (34%) 41.9
59.8 71.1
45.1 47.6
15.6 23.2
1.0 1.4
27 7 (12%)
32.1 12.8
57.2 69.8
1.3 1.4
Patientse
age
NEMS*
Apache II
Charlson
Ward patients
Median Ward stay
age
Charlson
45.8 47.6
21.3 24.1
1.3 2.0
46 9 (11%)
26.1 19.2
56.6 63.1
1.6 1.8
ICU-patients
ICU-patients
Median ICU stay
Median ICU stay
Median ICU stay
Patients from medical specialities (n = 81) Survived 57 33.8 58.1 Died 24 (30%) 40.5 72.5 a
20 patients were excluded because they were discharged from ICU directly to another hospital. 28 died on hospital wards; 7 in the first 48 hours and a total of 13 (46.4%) in the first 5 days. The global hospital mortality of these patients (UCI and ward) was 35%. c 13 ward patients were excluded because they were discharged from ICU directly to another hospital. d 5 ward patients were excluded because they were discharged from ICU directly to another hospital. e 2 ward patients were excluded because they were discharged from ICU directly to another hospital. b
The cost of the 161 long-term ICU patients after transfer to hospital wards was D 3,388,823 (approximately $4,439,358), which, when added to the D 10,972,500 (approximately $14,373,975) cost of treatment in the ICU, gives in a total cost of D 14,316,323 (approximately $18,754,383). Only 133 patients survived and were
Table 3
discharged from hospital. The cost per surviving patient discharged from hospital was D 107,641 (approximately $141,000). The 20 patients who were transferred to other hospitals from this ICU were not included in the hospital-ward data.
Cost of stay for 2,235 patients admitted to ICU for over 20 days.a,b,c ICU admission costs
Patients
Speciality
Average stay
Daily cost
Total cost
N= 95 N= 59 N= 81
Neurosurgical surgeries Medicas
35.4 ± 35.0 días 40.7 ± 23.2 días 36.8 ± 18.7 días
1.430 D 1.260 D 1.036 D
4.809.090 D 3.025.638 D 3.088.108 D
Hospital ward costd,e,f Patients
Speciality
Average stay
Daily cost
Total cost
N = 72 N = 34 N= 55
Neurosurgical Surgeries Medicas
44.0 ± 52.2 días 30.5 ± 29.1 días 25.5 ± 19.0 días
707 D 586 D 386 D
2.239.776 D 607.682 D 541.365 D
a b c d e f
These 235 patients had an average stay of 37.21 days and total a cost of 10.927.500 D . During their admission to ICU, 54 died and 20 were transferred to another hospital. 161 patients were discharged to hospital wards. Total cost (ICU and ward admission) for the 235 patients was: 14.316.323 D . We do not have the cost of those patients transferred to other hospitals. All figures quoted were obtained from the Toledo Hospital Complex Management.
Please cite this article in press as: Rodríguez Villar S, Barrientos Yuste RM. Long-term admission to the intensive care unit: a cost-benefit analysis. Rev Esp Anestesiol Reanim. 2014. http://dx.doi.org/10.1016/j.redar.2014.02.008
+Model
ARTICLE IN PRESS
REDAR-446; No. of Pages 8
6
S. Rodríguez Villar, R.M. Barrientos Yuste Table 4
Functional status of long-term ICU patients by specialty, according to the GOS-E scale.
Total No follow-up 1-death 2 3 4 5 6 7 8 Total
Medical 27 31 6 18 17 11 7 17 19 153/126
17,6% 24,6% 4,7% 14,3% 13,5% 8,7% 5,5% 13,5% 15,0%
Surgical 5 9 1 2 5 3 2 10 9 46/41
11,6% 20,9% 2,3% 4,6% 1,6% 7,0% 4,6% 23,5% 20,9%
Neurosurgical 5 4 0 5 3 3 4 2 6 32/27
18,5% 14.8% 0 18,5% 11,1% 11,1% 14,8% 7,4% 22,2%
17 18 5 11 9 5 1 5 4 73/56
30,3% 32,1% 8,9% 19,6% 16,1% 8.9% 1,8% 8,9% 7,1%
27 patients were not contactable for follow-up; 17.6% were involved in traffic accidents, whose home address had a greater geographical spread. Percentages were calculated on the basis of patients contacted n = 208. Of these 208, 54 died in ICU (25.9%), 28 on wards (13.5%) and 31 post-discharge (14.9%). In total 54.3% died. According to the GOS-E scale, 52 patients had a significant functional limitation (25.0%) and 43 had an acceptable functional recovery, (20.7%).
Post-hospital discharge follow-up The 20 patients transferred to other hospitals were included in this follow-up (133 patients discharged from Virgin de la Salud plus 20 transferred to other hospitals), giving a total of 153 patients. At least one year after hospital discharge the researchers attempted to contact the patient/next of kin on the telephone number that appeared on the patient records. It was not possible to contact 27 patients/next of kin. When contact was made, the patient’s functional status was assessed using the GOS-E scale. Reviewing the clinical reports of the 27 patients with whom contact could not be made, it was found that their profile was very similar to that of the patients who were contacted for the follow-up survey. However, these patients were not included in the results (see Table 4). Of the 126 patients/next of kin contacted, 56 patients were neurosurgical (18 had died at 1 year of follow-up), 27 were surgical (4 had died) and 43 were medical (9 of whom had not survived). The mean age of the 31 deceased patients (24.6%) was 67.6 ± 13.8 years. Fifty-two patients (25.0%1 ) survived with serious lasting effects on functionality. Their mean age was 55.1 ± 16.4 years. Forty-three patients (20.7%1 ) recovered normal functional status. Their mean age was 49.6 ± 20.1 years.
Discussion A recently published multi-centre study on ICU bed-day costs in four European countries9 has shown that these costs range between D 1168 (approximately $1530) and D 2025 (approximately $2652), the highest cost being that of staffing, which is consistent with our data.
1
Out of 208 patients (the 27 patients who were not contacted were not included).
In our experience, Diagnosis Related Groups (DRG),10 even those currently termed as refined, do not accurately reflect the ICU cost involved in the treatment of patients, particularly long-term patients for whom greater resources are needed. DRGs are calculated at the time of hospital discharge. In his method for estimating Medicare system costs, Fetter (1980) grouped homogeneous patients who consume similar hospital resources, but did not take into account complex patients, particularly those hospitalized for more than 60 days, which is the group included in this study. Moreover, since DGRs are carried out at the time of hospital discharge, they can only relate to the patients who died in the ICU. In a paper published 16 years ago, Heyland DK et al.11 used a methodology very similar to ours to calculate the cost of long ICU stays (>14 days) with follow-up at one year, and reported similar results to ours. However, their conclusions were surprising in that they considered long term and short term stays to be similarly ‘‘efficient’’ in terms of the amount of resources used to achieve a good outcome. In terms of efficiency, we cannot ignore the fact that longterm patients cost 8.8 times more than short-term patients ($37,800 vs. $4300), with a hospital mortality rate 4.1 times higher (44.2% vs. 10.8%). Our sample is larger, 235 vs. 61 patients, we also included multiple-trauma and neurosurgical patients, and also we used the GOS-E scale to objectively measure functional status. Although patients with a long stay in the ICU have previously been studied by other authors,12---17 this investigation has introduced various additional factors such as cost by different patient categories, separating medical and surgical patients, and in this latter category, separating neurosurgical patients from those undergoing other types of surgery. The aim of this study was to analyze the evolution of long-term ICU patients, to determine the rate of mortality in the ICU, to estimate the cost of ICU and hospital ward stays, and by means of a follow-up one year after hospital discharge, to evaluate survival rates and functional status. The most important findings are that the 235 patients spending more than 20 days in the ICU represent only 6.4%
Please cite this article in press as: Rodríguez Villar S, Barrientos Yuste RM. Long-term admission to the intensive care unit: a cost-benefit analysis. Rev Esp Anestesiol Reanim. 2014. http://dx.doi.org/10.1016/j.redar.2014.02.008
+Model REDAR-446; No. of Pages 8
ARTICLE IN PRESS
Long-term admission to the intensive care unit: a cost-benefit analysis of the 3400 patients admitted, yet with a mean length of stay of 37.2 days, they occupied 34.3% of the total beds available in the unit and absorbed 29.5% of critical care expenditure. Added to the cost of care and treatment on hospital wards and the ICU, the 235 long-term patients had an overall hospitalization cost of D 14,361,323 (approximately $18,813,333) bearing in mind that at one year after hospital discharge, only 20.7% had recovered a relatively normal functional status. Furthermore, 55% of those patients died after discharge, an unacceptably high figure in our opinion. We consider the percentage of patients recovering normal functional status to be very low, especially in the case of neurosurgical patients. The Charlson index was used in this study to assess comorbidity prior to ICU admission: the higher the score, the more severe the comorbidity burden; however a good correlation was found between this score and the APACHE II, which is consistent with the findings of other authors.18 As mentioned above, similar results were obtained from two particular patient groups: on the one hand, neurosurgical patients (56% with head injury) and on the other, medical and complex surgical patients (see Table 4). This first group (neurosurgical) had a low mortality rate within the ICU, patients were younger, and had lower APACHE II and Charlson scores. However, mortality rates rose during their stay on hospital wards, and their overall hospital stay was longer. Moreover, at follow-up the results were even poorer, with a high rate of mortality (32.1%) and a high percentage experiencing functional limitation (53.3%). This group requires more extensive post-ICU follow-up treatment, which calls for additional evaluation of the consumption of resources that such long-term treatment involves. Various studies have shown that the recovery of these patients, and in particular those with head injury, should be viewed in terms of several years.19,20 In the second group (medical and complex surgery patients), mortality within the ICU was three times that of neurosurgical patients. Age, APACHE II and Charlson scores were also significantly higher; such findings are similar to those of other studies where patients admitted for medicalsurgical MODS (multi organ dysfunction syndrome) had a high one-year mortality rate. The factors influencing hospital mortality were age and poor overall functional status, but neither of which were modifiable factors.21 Similar results are found in another study where older age, longer treatment in the ICU and higher simplified acute physiology score (SAPS) at admission were associated with shorter 12-month survival.22 Patients who died after transfer to hospital wards were older and had a higher Charlson index. With these characteristics, this patient group has a poor prognosis; however, DNAR (do not attempt resuscitation) or WLPT (withdrawal of life prolonging treatment) criteria differ between ICUs. A search of the literature has shown the WLPT rate varies according to country and type of hospital, for example, in some US and Canadian ICUs figures range from 40 to 70%23---27 ; a French multi-centre study estimated that in 53% of cases some kind of WLPT was pursued,28 and in Spain this was estimated at 34---55%.29,30 In our study, the fact that out of 28 patients who died following transfer to hospital wards, 13 died (43.0%) within 1.8 ± 2.1 days shows that WLPT was
7
considered inappropriate and therefore not applied during the patient’s stay in the ICU.
Conclusions • The hospital mortality rate of long-term ICU patients, those spending more than 20 days, is nine times higher (35%) than those with shorter ICU stays. • Of these patients, a relatively small percentage of those admitted to the ICU (6.5%), consume one third of the unit’s entire resources: four times more than those with a short ICU stay. • One year after hospital discharge, only 21% of these patients had regained an acceptable level of functionality, and 25% survived with seriously impaired functionality.
Funding Financial support, including any institutional departmental funds, was not sought for this study.
Conflict of interests All faculty and staff in a position to control or affect the content of this paper declare that they have no competing financial interests.
Acknowledgements The authors of this study would like to thank all those who have worked towards the compilation of the database which has been central to this research. This paper is dedicated to Dr Rafael Barrientos Vega who have taught us an important part of the medical curriculum.
Annex 1. Charlson Comorbidity Index Condition
Weight
Myocardial infarction Congestive heart failure Peripheral vascular disease Cerebrovascular accident Dementia Chronic pulmonary disease Connective tissue disease Gastrointestinal ulcer disease Mild liver disease
ONE POINT
Hemiplegia Moderate to severe renal disease Diabetes with end-organ damage Any tumour
TWO POINTS
Leukaemia Lymphoma Moderate or severe liver disease
THREE POINTS
Autoimmune deficiency syndrome Metastatic solid tumour
SIX POINTS
Please cite this article in press as: Rodríguez Villar S, Barrientos Yuste RM. Long-term admission to the intensive care unit: a cost-benefit analysis. Rev Esp Anestesiol Reanim. 2014. http://dx.doi.org/10.1016/j.redar.2014.02.008
+Model REDAR-446; No. of Pages 8
ARTICLE IN PRESS
8
S. Rodríguez Villar, R.M. Barrientos Yuste
Annex 2. The Extended GOS (GOS-E) provides detailed categorization into eight categories by subdividing the categories of severe disability, moderate disability and good recovery into a lower and upper category 1 2 3 4 5 6 7 8
Death (D) Vegetative state (VS) Lower severe disability (SD −) Upper severe disability (SD +) Lower moderate disability (MD −) Upper moderate disability (MD +) Lower good recovery (GR −) Upper good recovery (GR +)
14.
15.
16.
17.
18.
References
19.
1. Ruiz Moreno J. Fundamentos de gestión sanitaria. Medicina critica práctica. Edika Med; 2006. 2. Halpern NA, Pastores SM. Critical care medicine in de USA 2000---2005: on analysis of numbers, occupancy rate, payer mix and costs. Crit Care Med. 2010;38:65---71. 3. Friedrich J, Wilson G, Chant C. Long-term outcomes and clinical predictors of hospital mortality in very long stay intensive care unit patients: a cohort study. Crit Care. 2006;10:R59. 4. Santana Cabrera L, Sánchez Palacios M, Hernández Medina E, Eugenio Robaina P, Villanueva-Hernández Á. Características y pronóstico de los paciente mayores con estancias prolongadas en una UCI. Med Intens. 2008:157---62. 5. Sricker KH, Cavegen R, Takala J, Rothen HU. Does ICU length of stay influence quality of life? Acta Anaesthesiol Scand. 2005;49:975---83. 6. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies. J Chronic Dis. 1987;40:373---83. 7. Fayol P, Carriere H, Habonimana D, Preux PM, Dumond JJ. French version of structured interviews for the Glasgow Outcome Scale: guidelines and first studies of validation. Ann Readapt Med Phys. 2004;47:142---56. 8. Cabre L, Mancebo J, Solsona JF, Saura P, Gich I, Blanch L, et al. Multicentre study of de multiple organ dysfunction syndrome in intensive care units: the usefulness of SOFA in decision making. Intens Care Med. 2005;31:927---33. 9. Tan SS, Bakker J, Hoogendoorn ME, Kapila A, Martin J, Pezzi A, et al. Direct cost analysis of intensive care unit stay in four European countries: applying a standardized costing methodology. Value Health. 2012;15:81---6. 10. Barrientos Vega R. Nuestra experiencia con los Grupos Relacionados por el Diagnostico en Cuidados Intensivos. Med Intens. 2003;27:391---8. 11. Heyland DK, Konopad E, Noseworthy TW. Is it ‘‘Wortthwhile’’ to continue treating patients with a prolonged stay (
ARTICLE IN PRESS
Rev Esp Anestesiol Reanim. 2014;xxx(xx):xxx---xxx
Revista Española de Anestesiología y Reanimación www.elsevier.es/redar
ORIGINAL
Long-term admission to the intensive care unit: a cost-benefit analysis S. Rodríguez Villar a,∗ , R.M. Barrientos Yuste b a b
Intensive Care Medicine Department, Queen Elizabeth Hospital, London, United Kingdom Madrid Care Centre for the Disabled, Madrid, Spain
Received 21 January 2014; accepted 25 February 2014
KEYWORDS Long-stay patients; Intensive care; Cost; Mortality; Functional status post discharge
∗
Abstract Objective: To assess outcomes in long-term ICU patients, with follow-ups carried out at one year post discharge, in order to calculate the costs incurred by the hospital in relation to the benefits gained. Material: Of 3639 patients consecutively admitted over the course of three years to ICU, 235 (6.5%) were assessed for the purposes of the study, having spent a period exceeding 20 days in intensive care. Method: The survey tool used was the Spanish Minimum Data Set (MDS). The length of ICU stay and hospital stay following discharge from ICU were calculated, and one year post discharge the patient/next of kin was contacted in order to carry out a follow-up survey on survival and functional status (according to GOS-E scale). Results: The 235 study patients had a mean stay of 37 days, occupied 34% of ICU beds available and consumed 29% of the ICU’s economic resources ($14,400,175). Their stay on hospital wards was (mean) 33 days. Mortality in ICU and on hospital wards was 40% higher amongst older patients, and those with a higher APACHE II and Charlson index score. Mortality rates were three times higher among neurosurgical patients: mortality at follow-up was 25%, and only 21% recovered an acceptable functional status. Conclusions: Mortality rates in long-term ICU patients are high, both during their hospital stay and in the first year post discharge. Surviving patients do not exhibit a good level of recovery, and consume a large proportion of economic resources. © 2014 Sociedad Espa˜ nola de Anestesiología, Reanimación y Terapéutica del Dolor. Published by Elsevier España, S.L. All rights reserved.
Corresponding author. E-mail address: [email protected] (S. Rodríguez Villar).
0034-9356/$ – see front matter © 2014 Sociedad Espa˜ nola de Anestesiología, Reanimación y Terapéutica del Dolor. Published by Elsevier España, S.L. All rights reserved.
http://dx.doi.org/10.1016/j.redar.2014.02.008
Please cite this article in press as: Rodríguez Villar S, Barrientos Yuste RM. Long-term admission to the intensive care unit: a cost-benefit analysis. Rev Esp Anestesiol Reanim. 2014. http://dx.doi.org/10.1016/j.redar.2014.02.008
+Model REDAR-446; No. of Pages 8
ARTICLE IN PRESS
2
S. Rodríguez Villar, R.M. Barrientos Yuste
PALABRAS CLAVE pacientes de larga duración; Cuidados intensivos; Costes; Mortalidad; Estado funcional después del alta
Ingresos de larga duración en la unidad de cuidados intensivos: análisis de costo-beneficio Resumen Objetivo: Evaluar los resultados de los pacientes de larga duración en la UCI por medio de un seguimiento al a˜ no del alta a fin de calcular los gastos soportados por el hospital en relación con los beneficios obtenidos. Materiales: de los 3.639 pacientes consecutivos ingresados en la UCI durante tres a˜ nos, se evaluó a 235 (un 6,5%) para el presente estudio, todos ellos con una estancia superior a 20 días. Métodos: el instrumento de evaluación fue la base de datos espa˜ nola CMBD (conjunto mínimo de base de datos). Se calcularon la duración de la estancia en la UCI y en la planta después del alta de la UCI y, un a˜ no después del alta hospitalaria, se contactó con el paciente o su pariente más próximo para realizar una encuesta de seguimiento sobre su estado funcional (según la escala GOSE). Resultados: los 235 pacientes estudiados estuvieron ingresados un promedio de 37 días, ocuparon un 34% de las camas disponibles en la UCI y emplearon un 29% de los recursos económicos de dicha unidad (14.400.175$). Su estancia media en planta fue de 33 días. La mortalidad en la UCI y en planta fue un 40% más alta en los pacientes de mayor edad, puntuación del APACHE II e índice de Charlson. Las tasas de mortalidad se triplicaron en los pacientes neuroquirúrgicos. En el seguimiento después de un a˜ no, la tasa de mortalidad fue del 25%, y únicamente el 21% recuperó un estado funcional aceptable. Conclusiones: las tasas de mortalidad en pacientes de larga duración en la UCI son altas, tanto durante la estancia hospitalaria como durante el a˜ no posterior al alta. Los pacientes de larga duración no presentan una recuperación correcta y consumen una gran proporción de recursos económicos. © 2014 Sociedad Espa˜ nola de Anestesiología, Reanimación y Terapéutica del Dolor. Publicado por Elsevier España, S.L. Todos los derechos reservados.
Introduction
Patients and methods
The quality of resources currently available to intensive care services permits the treatment and survival of patients suffering from increasingly more serious and complex conditions. However, the main condition together with the failure of other associated organs often prolong the stay of patients in our units, and just as morbidity and mortality levels rise, so do the care and treatment costs involved. Financed by state taxes, the Spanish Healthcare System is free, and practitioners may use all available resources in their particular healthcare setting, according to their professional criteria. Healthcare expenditure in Spain accounts for 8.4% of Gross National Product (GNP) (54.4% of which is hospitalbased expenditure). The average expenditure in relation to GNP in the EU is 9.6% (OECD Health Data, 2008). Currently, life expectancy in Spain at birth is 81 years. According to some sources, the resources used in the critical care setting consume 15% of the hospital budget.1 In the USA, critical care accounts for 0.7% of GNP.2 The aim of this study is to investigate mortality in patients who remain in the Intensive Care Unit for over 20 days. The study focuses on survivors, following up their progress from the ICU to hospital wards and up to discharge. To determine survival rates and functional status, patients were followed-up one year after discharge from the ICU. The hospital administration provided the researchers with details of the cost of caring for these patients, both in the ICU and hospital wards.
Design An observational, longitudinal, descriptive study using data from patients admitted over a 3-year period to a multipurpose ICU at a tertiary-level hospital. The study was approved by the Independent Ethics Committee of the Hospital Virgen de la Salud in Toledo.
Materials The study was carried out in the Toledo Hospital Complex in Spain, which comprises two large hospitals and several specialist outpatient clinics. The complex has 777 beds with 26 ft multi-purpose ICU beds (3 beds exclusively for scheduled surgery patients). Data from patients admitted in 2010, 2011 and 2012 for a period exceeding 20 days were analyzed. We considered 20 days to be a prolonged period; however, others have applied different criteria in defining what may be considered a longterm patient, and have suggested hospital stays of anything between 7 and 30 days.3---5 Virgen de la Salud Hospital is a regional centre of excellence for neurosurgery. It also boasts a coronary unit and cardiac surgery resuscitation unit that are independent of the main ICU where the study was carried out. Over the 3 years studied, 3639 patients were admitted, of which 235 (6.5%) were hospitalized for more than 20 days (Fig. 1).
Please cite this article in press as: Rodríguez Villar S, Barrientos Yuste RM. Long-term admission to the intensive care unit: a cost-benefit analysis. Rev Esp Anestesiol Reanim. 2014. http://dx.doi.org/10.1016/j.redar.2014.02.008
+Model
ARTICLE IN PRESS
REDAR-446; No. of Pages 8
Long-term admission to the intensive care unit: a cost-benefit analysis
3
3.639 patients are admitted to ICU: .Over a period of 36 months (January 2008-December 2010). . Multiple-trauma, neurocritical, medical and surgical patients are included.
54 patients (23%) die on ICU and without having been discharged from hospital.
>20 days stay in ICU 235 patients (6,5%)
181 patients (77%) do not die in ICU.
161 patients were discharge to the hospital wards: -133 (80%) survived and are discharged at home. -28 (17,5%) died on hospital wards
20 patient sare repatriated to their original hospital
43 MEDICAL PATIENTS (21% DIED): GOSE (1) 9, GOSE (2-5) 11, GOSE (6-8) 21 patients. 153 patients survived: 27 SURGICAL PATIENTS (15% DIED) : GOSE (1) 4, GOSE (2-5) 11, GOSE (6-8) 12 patients.
-It was not possible to contact 27 patients by telephone. Their status is unknown -126 patients make up the final patient sample
56 NEUROSURGICAL PATIENTS (32% DIED): GOSE (1) 18, GOSE (2-5) 30, GOSE (6-8) 10 patients.
Figure 1
Patient flow-chart.
Method The basic survey tool used was a minimum data set (MDS) designed by the intensive Care Unit. Data such as the age, length of stay, origin and destination of the patient were collected in addition to the diagnosis and procedures evaluated according to the International Statistical Classification of Diseases and Related Health Problems (ICD-9), APACHE II, and NEMS score. SOFA is not routinely carried out for all patients, so it was not included in this analysis. Mean length of stay and destination following ICU discharge onto hospital wards were obtained from the hospital MDS. Further information was obtained from clinical reports of each of the 235 selected patients, both on discharge from ICU and from hospital wards, and based on these reports the patient was assigned a score on the Charlson comorbidity index6 (Annex 1). Where no comorbidity existed, the patient was assigned a score of zero. The Hospital’s administration department provided the researchers with information related to the cost of each patient’s care and treatment. In addition to calculating the mean cost of ICU stays, related costs were also categorized according to whether the patient was medical, surgical or neurosurgical. The same criteria were used during follow-up on costs incurred after discharge to hospital wards. The difference in cost between surgical patients and medical patients is due to the cost of the surgical procedures involved and the higher percentage of additional tests performed outside the ICU (generally radiological).
One year after discharge, the patients/next of kin were contacted on the telephone number given on their patient records. The researchers verified whether the patient had survived and assessed their functional status according to the Extended Glasgow Outcome Scale, with a score of 1---87 (Annex 2).
Statistics The statistical analysis was performed using SPSS 12.0 to calculate the mean, standard variation and percentages of the variables under study. For dichotomous variables, the chi-square test was used. The level of statistical significance was p < 0.05.
Results Patients admitted to the ICU for a period of less than 20 days For the 3400 patients hospitalized for less than 20 days over the 3-year study period the mean length of stay was 3.55 ± 4.18 days, with a mortality rate in the ICU of 13.4% and APACHE II score of 14.2 ± 9.0. On hospital wards postICU discharge, mortality was 6.1%, and mean length of stay was 11.6 ± 16.0 days. Of the foregoing patients, 30.1% were admitted from A&E, 15.5% from hospital wards, 34.8% from scheduled
Please cite this article in press as: Rodríguez Villar S, Barrientos Yuste RM. Long-term admission to the intensive care unit: a cost-benefit analysis. Rev Esp Anestesiol Reanim. 2014. http://dx.doi.org/10.1016/j.redar.2014.02.008
+Model REDAR-446; No. of Pages 8
ARTICLE IN PRESS
4
S. Rodríguez Villar, R.M. Barrientos Yuste
surgery, 10.3% from emergency surgery and 9.3% were transferred from other hospitals. By specialty: neurosurgical patients accounted for 28.9%, general surgery 18.4%, other surgery 17.5%, internal medicine patients 9.5%, cardiology 14.1% and other medical patients 10.8%.
Patients admitted to the ICU for over 20 days Of the 3635 patients, 235 were admitted to the ICU for a period that exceeded 20 days, with a mean length of stay of 37.2 ± 27.4 days, giving a total of 8742 days of stay over the three years studied, the equivalent of 7.9 beds/year. In other words, 6.5% of the total number of patients studied occupied 34.3% of the 23 ICU beds available. The 235 patients were categorized by specialty: 95 neurosurgical patients (41.1%), 59 other surgery (25.5%), and 81 medical patients (33.3%). The 95 neurosurgical patients had a mean length of stay of 31.8 ± 27.1 days, the 59 other surgical patients’ stay was 40.7 ± 23.2 days, and the 81 medical patients had a mean stay of 36.8 ± 18.7 days; 54 patients (22.9%) died in the ICU. Table 1 lists the most common procedures and resources used in the treatment of these patients. Table 2 shows the survival rate, age, APACHE II, NEMS and Charlson scores, both in the ICU and on hospital wards, of patients spending more than 20 days in ICU. Age and APACHE II were considered statistically significant (p < 0.5) when comparing survivors with deceased; however, the NEMS score was not found to be significant. According to some studies, SOFA score is a very significant factor; unfortunately, since it was not available for all the patients it could not be included in this study.8 As discussed above, patients were categorized into three pathological groups (neuro-surgical, other surgical and medical), among which advanced age, high APACHE II and Charlson scores continued to be significant factors for
Table 1 Some of the resources used (procedures and techniques) were. Long term invasive mechanical ventilation Tracheostomy Extracorporeal kidney support Extracorporeal liver support Poli-transfusions Cerebral arteriography-embolization Intracraneal pressure monitoring S.Ganz catheter CT-scan (Computed tomography) MRI-scan (Magnetic Resonance Imaging) Some of the most common surgical procedures were: 58 32 6 7 1 1 4
Craneotomy Laparotomy Thoracotomy Vascular surgery Nefrectomy Permanent pacemaker Orthopaedic surgery
99, 3% 57,1% 16,7% 3,7% 47,8% 7,6% 31,2% 10,8% 78,5% 10,4%
mortality. The neurosurgical patients were younger, with a high percentage of head injury cases. They had lower APACHE II and Charlson index scores and less associated comorbidity. As a result, fewer suffered from multiple organ failure, and the ICU mortality rate was therefore lower: 10.5%. In the other two groups (other surgical and medical), where multiple organ failure is more frequent and Charlson index scores higher, the ICU mortality rate was three times that of neurosurgical patients (31%). Patients were also older and have higher APACHE II scores.
Follow-up of patients transferred to hospital wards Of the 181 patients discharged from the ICU, 20 were transferred directly to another hospital (see Table 2). The 166 patients transferred to wards in Virgen de la Salud Hospital included 72 neurosurgical patients, 34 surgical and 54 medical patients. Mean hospital ward stay was 33.6 ± 37.6 days, which, added to the length of stay in the ICU, gives a mean hospital stay of 70.4 days. The hospital stay of neurosurgical patients was double that of the other categories, which is significant since they present more functional limitations. Of the 235 long-term patients, 82 (34.9%) died in hospital; 54 in the ICU, and 28 on hospital wards. Thirteen of the 28 deceased patients died on the wards in the first 1.8 ± 2.1 days. Death after a hospital stay of 34.4 days might be considered post-ICU limitation of therapeutic effort (LTE).
Use of resources and costs incurred in the treatment of ‘long-term’ patients The bed-day cost in this ICU is estimated to be D 1250/day (approximately $1637). The 8742 bed-days of the 235 longterm patients, therefore, represents an overall cost of D 10,927,500 (approximately $14,264,250) (Table 3). Considering these figures and the fact that 54 patients died in the ICU, the total cost per surviving patient per ICU stay is D 60,372, or approximately $78,807. The cost of running the Intensive Care Unit in 2010, 2011 and 2012 is estimated by hospital management to be D 37,695,557, (approximately $49,004,224). Long-term patients, 6.5% of the total, accounted for 29.5% of the total ICU budget. The bed-day cost of the three patient categories were: D 1430/day (approximately $1873/day) for neurosurgical patients, D 1260/day (approximately $1650) for other surgical patients, and D 1036/day (approximately $1357) for medical patients. The highest costs are those of staffing, while for surgical specialties costs increase due to certain surgical procedures and additional studies (see Table 3 for a break-down of costs per number of patients and length of stay in days). The hospital finance department also supplied us with main ward bed-day costs for the different specialties: neurosurgical D 707/day (approximately $926), other surgical D 586/day (approximately $767) and medical D 386/day (approximately $505).
Please cite this article in press as: Rodríguez Villar S, Barrientos Yuste RM. Long-term admission to the intensive care unit: a cost-benefit analysis. Rev Esp Anestesiol Reanim. 2014. http://dx.doi.org/10.1016/j.redar.2014.02.008
+Model
ARTICLE IN PRESS
REDAR-446; No. of Pages 8
Long-term admission to the intensive care unit: a cost-benefit analysis Table 2
5
235 patients spending 20+ days in ICUa,b
Patients
ICU-patients
Median ICU stay
age
NEMS*
APACHEII
Charlson
Ward patients
Median Ward stay
age
Charlson
Survived Died
181 54 (23%)
34.4 39.7
56.7 69.5
44.5 46.3
17.4 22.5
1.0 1.2
133 28 (12%)
35.5 29.9
55.2 63.3
1.3 1.1
Patientsc
ICU-patients
age
NEMS*
Apache II
Charlson
Ward patients
Median Ward stay
age
Charlson
Neurosurgical patients (n = 95) Survived 85 31.2 Died 10 (10%) 33.7
52.1 59.2
43.1 42.5
16.6 17.3
0.75 0.80
60 12 (13%)
43.2 50.4
52.4 60.0
0.75 0.80
Patientsd
age
NEMS*
Apache II
Charlson
Ward patients
Median Ward stay
age
Charlson
Patients from other surgeries (n= 59) Survived 39 40.4 Died 20 (34%) 41.9
59.8 71.1
45.1 47.6
15.6 23.2
1.0 1.4
27 7 (12%)
32.1 12.8
57.2 69.8
1.3 1.4
Patientse
age
NEMS*
Apache II
Charlson
Ward patients
Median Ward stay
age
Charlson
45.8 47.6
21.3 24.1
1.3 2.0
46 9 (11%)
26.1 19.2
56.6 63.1
1.6 1.8
ICU-patients
ICU-patients
Median ICU stay
Median ICU stay
Median ICU stay
Patients from medical specialities (n = 81) Survived 57 33.8 58.1 Died 24 (30%) 40.5 72.5 a
20 patients were excluded because they were discharged from ICU directly to another hospital. 28 died on hospital wards; 7 in the first 48 hours and a total of 13 (46.4%) in the first 5 days. The global hospital mortality of these patients (UCI and ward) was 35%. c 13 ward patients were excluded because they were discharged from ICU directly to another hospital. d 5 ward patients were excluded because they were discharged from ICU directly to another hospital. e 2 ward patients were excluded because they were discharged from ICU directly to another hospital. b
The cost of the 161 long-term ICU patients after transfer to hospital wards was D 3,388,823 (approximately $4,439,358), which, when added to the D 10,972,500 (approximately $14,373,975) cost of treatment in the ICU, gives in a total cost of D 14,316,323 (approximately $18,754,383). Only 133 patients survived and were
Table 3
discharged from hospital. The cost per surviving patient discharged from hospital was D 107,641 (approximately $141,000). The 20 patients who were transferred to other hospitals from this ICU were not included in the hospital-ward data.
Cost of stay for 2,235 patients admitted to ICU for over 20 days.a,b,c ICU admission costs
Patients
Speciality
Average stay
Daily cost
Total cost
N= 95 N= 59 N= 81
Neurosurgical surgeries Medicas
35.4 ± 35.0 días 40.7 ± 23.2 días 36.8 ± 18.7 días
1.430 D 1.260 D 1.036 D
4.809.090 D 3.025.638 D 3.088.108 D
Hospital ward costd,e,f Patients
Speciality
Average stay
Daily cost
Total cost
N = 72 N = 34 N= 55
Neurosurgical Surgeries Medicas
44.0 ± 52.2 días 30.5 ± 29.1 días 25.5 ± 19.0 días
707 D 586 D 386 D
2.239.776 D 607.682 D 541.365 D
a b c d e f
These 235 patients had an average stay of 37.21 days and total a cost of 10.927.500 D . During their admission to ICU, 54 died and 20 were transferred to another hospital. 161 patients were discharged to hospital wards. Total cost (ICU and ward admission) for the 235 patients was: 14.316.323 D . We do not have the cost of those patients transferred to other hospitals. All figures quoted were obtained from the Toledo Hospital Complex Management.
Please cite this article in press as: Rodríguez Villar S, Barrientos Yuste RM. Long-term admission to the intensive care unit: a cost-benefit analysis. Rev Esp Anestesiol Reanim. 2014. http://dx.doi.org/10.1016/j.redar.2014.02.008
+Model
ARTICLE IN PRESS
REDAR-446; No. of Pages 8
6
S. Rodríguez Villar, R.M. Barrientos Yuste Table 4
Functional status of long-term ICU patients by specialty, according to the GOS-E scale.
Total No follow-up 1-death 2 3 4 5 6 7 8 Total
Medical 27 31 6 18 17 11 7 17 19 153/126
17,6% 24,6% 4,7% 14,3% 13,5% 8,7% 5,5% 13,5% 15,0%
Surgical 5 9 1 2 5 3 2 10 9 46/41
11,6% 20,9% 2,3% 4,6% 1,6% 7,0% 4,6% 23,5% 20,9%
Neurosurgical 5 4 0 5 3 3 4 2 6 32/27
18,5% 14.8% 0 18,5% 11,1% 11,1% 14,8% 7,4% 22,2%
17 18 5 11 9 5 1 5 4 73/56
30,3% 32,1% 8,9% 19,6% 16,1% 8.9% 1,8% 8,9% 7,1%
27 patients were not contactable for follow-up; 17.6% were involved in traffic accidents, whose home address had a greater geographical spread. Percentages were calculated on the basis of patients contacted n = 208. Of these 208, 54 died in ICU (25.9%), 28 on wards (13.5%) and 31 post-discharge (14.9%). In total 54.3% died. According to the GOS-E scale, 52 patients had a significant functional limitation (25.0%) and 43 had an acceptable functional recovery, (20.7%).
Post-hospital discharge follow-up The 20 patients transferred to other hospitals were included in this follow-up (133 patients discharged from Virgin de la Salud plus 20 transferred to other hospitals), giving a total of 153 patients. At least one year after hospital discharge the researchers attempted to contact the patient/next of kin on the telephone number that appeared on the patient records. It was not possible to contact 27 patients/next of kin. When contact was made, the patient’s functional status was assessed using the GOS-E scale. Reviewing the clinical reports of the 27 patients with whom contact could not be made, it was found that their profile was very similar to that of the patients who were contacted for the follow-up survey. However, these patients were not included in the results (see Table 4). Of the 126 patients/next of kin contacted, 56 patients were neurosurgical (18 had died at 1 year of follow-up), 27 were surgical (4 had died) and 43 were medical (9 of whom had not survived). The mean age of the 31 deceased patients (24.6%) was 67.6 ± 13.8 years. Fifty-two patients (25.0%1 ) survived with serious lasting effects on functionality. Their mean age was 55.1 ± 16.4 years. Forty-three patients (20.7%1 ) recovered normal functional status. Their mean age was 49.6 ± 20.1 years.
Discussion A recently published multi-centre study on ICU bed-day costs in four European countries9 has shown that these costs range between D 1168 (approximately $1530) and D 2025 (approximately $2652), the highest cost being that of staffing, which is consistent with our data.
1
Out of 208 patients (the 27 patients who were not contacted were not included).
In our experience, Diagnosis Related Groups (DRG),10 even those currently termed as refined, do not accurately reflect the ICU cost involved in the treatment of patients, particularly long-term patients for whom greater resources are needed. DRGs are calculated at the time of hospital discharge. In his method for estimating Medicare system costs, Fetter (1980) grouped homogeneous patients who consume similar hospital resources, but did not take into account complex patients, particularly those hospitalized for more than 60 days, which is the group included in this study. Moreover, since DGRs are carried out at the time of hospital discharge, they can only relate to the patients who died in the ICU. In a paper published 16 years ago, Heyland DK et al.11 used a methodology very similar to ours to calculate the cost of long ICU stays (>14 days) with follow-up at one year, and reported similar results to ours. However, their conclusions were surprising in that they considered long term and short term stays to be similarly ‘‘efficient’’ in terms of the amount of resources used to achieve a good outcome. In terms of efficiency, we cannot ignore the fact that longterm patients cost 8.8 times more than short-term patients ($37,800 vs. $4300), with a hospital mortality rate 4.1 times higher (44.2% vs. 10.8%). Our sample is larger, 235 vs. 61 patients, we also included multiple-trauma and neurosurgical patients, and also we used the GOS-E scale to objectively measure functional status. Although patients with a long stay in the ICU have previously been studied by other authors,12---17 this investigation has introduced various additional factors such as cost by different patient categories, separating medical and surgical patients, and in this latter category, separating neurosurgical patients from those undergoing other types of surgery. The aim of this study was to analyze the evolution of long-term ICU patients, to determine the rate of mortality in the ICU, to estimate the cost of ICU and hospital ward stays, and by means of a follow-up one year after hospital discharge, to evaluate survival rates and functional status. The most important findings are that the 235 patients spending more than 20 days in the ICU represent only 6.4%
Please cite this article in press as: Rodríguez Villar S, Barrientos Yuste RM. Long-term admission to the intensive care unit: a cost-benefit analysis. Rev Esp Anestesiol Reanim. 2014. http://dx.doi.org/10.1016/j.redar.2014.02.008
+Model REDAR-446; No. of Pages 8
ARTICLE IN PRESS
Long-term admission to the intensive care unit: a cost-benefit analysis of the 3400 patients admitted, yet with a mean length of stay of 37.2 days, they occupied 34.3% of the total beds available in the unit and absorbed 29.5% of critical care expenditure. Added to the cost of care and treatment on hospital wards and the ICU, the 235 long-term patients had an overall hospitalization cost of D 14,361,323 (approximately $18,813,333) bearing in mind that at one year after hospital discharge, only 20.7% had recovered a relatively normal functional status. Furthermore, 55% of those patients died after discharge, an unacceptably high figure in our opinion. We consider the percentage of patients recovering normal functional status to be very low, especially in the case of neurosurgical patients. The Charlson index was used in this study to assess comorbidity prior to ICU admission: the higher the score, the more severe the comorbidity burden; however a good correlation was found between this score and the APACHE II, which is consistent with the findings of other authors.18 As mentioned above, similar results were obtained from two particular patient groups: on the one hand, neurosurgical patients (56% with head injury) and on the other, medical and complex surgical patients (see Table 4). This first group (neurosurgical) had a low mortality rate within the ICU, patients were younger, and had lower APACHE II and Charlson scores. However, mortality rates rose during their stay on hospital wards, and their overall hospital stay was longer. Moreover, at follow-up the results were even poorer, with a high rate of mortality (32.1%) and a high percentage experiencing functional limitation (53.3%). This group requires more extensive post-ICU follow-up treatment, which calls for additional evaluation of the consumption of resources that such long-term treatment involves. Various studies have shown that the recovery of these patients, and in particular those with head injury, should be viewed in terms of several years.19,20 In the second group (medical and complex surgery patients), mortality within the ICU was three times that of neurosurgical patients. Age, APACHE II and Charlson scores were also significantly higher; such findings are similar to those of other studies where patients admitted for medicalsurgical MODS (multi organ dysfunction syndrome) had a high one-year mortality rate. The factors influencing hospital mortality were age and poor overall functional status, but neither of which were modifiable factors.21 Similar results are found in another study where older age, longer treatment in the ICU and higher simplified acute physiology score (SAPS) at admission were associated with shorter 12-month survival.22 Patients who died after transfer to hospital wards were older and had a higher Charlson index. With these characteristics, this patient group has a poor prognosis; however, DNAR (do not attempt resuscitation) or WLPT (withdrawal of life prolonging treatment) criteria differ between ICUs. A search of the literature has shown the WLPT rate varies according to country and type of hospital, for example, in some US and Canadian ICUs figures range from 40 to 70%23---27 ; a French multi-centre study estimated that in 53% of cases some kind of WLPT was pursued,28 and in Spain this was estimated at 34---55%.29,30 In our study, the fact that out of 28 patients who died following transfer to hospital wards, 13 died (43.0%) within 1.8 ± 2.1 days shows that WLPT was
7
considered inappropriate and therefore not applied during the patient’s stay in the ICU.
Conclusions • The hospital mortality rate of long-term ICU patients, those spending more than 20 days, is nine times higher (35%) than those with shorter ICU stays. • Of these patients, a relatively small percentage of those admitted to the ICU (6.5%), consume one third of the unit’s entire resources: four times more than those with a short ICU stay. • One year after hospital discharge, only 21% of these patients had regained an acceptable level of functionality, and 25% survived with seriously impaired functionality.
Funding Financial support, including any institutional departmental funds, was not sought for this study.
Conflict of interests All faculty and staff in a position to control or affect the content of this paper declare that they have no competing financial interests.
Acknowledgements The authors of this study would like to thank all those who have worked towards the compilation of the database which has been central to this research. This paper is dedicated to Dr Rafael Barrientos Vega who have taught us an important part of the medical curriculum.
Annex 1. Charlson Comorbidity Index Condition
Weight
Myocardial infarction Congestive heart failure Peripheral vascular disease Cerebrovascular accident Dementia Chronic pulmonary disease Connective tissue disease Gastrointestinal ulcer disease Mild liver disease
ONE POINT
Hemiplegia Moderate to severe renal disease Diabetes with end-organ damage Any tumour
TWO POINTS
Leukaemia Lymphoma Moderate or severe liver disease
THREE POINTS
Autoimmune deficiency syndrome Metastatic solid tumour
SIX POINTS
Please cite this article in press as: Rodríguez Villar S, Barrientos Yuste RM. Long-term admission to the intensive care unit: a cost-benefit analysis. Rev Esp Anestesiol Reanim. 2014. http://dx.doi.org/10.1016/j.redar.2014.02.008
+Model REDAR-446; No. of Pages 8
ARTICLE IN PRESS
8
S. Rodríguez Villar, R.M. Barrientos Yuste
Annex 2. The Extended GOS (GOS-E) provides detailed categorization into eight categories by subdividing the categories of severe disability, moderate disability and good recovery into a lower and upper category 1 2 3 4 5 6 7 8
Death (D) Vegetative state (VS) Lower severe disability (SD −) Upper severe disability (SD +) Lower moderate disability (MD −) Upper moderate disability (MD +) Lower good recovery (GR −) Upper good recovery (GR +)
14.
15.
16.
17.
18.
References
19.
1. Ruiz Moreno J. Fundamentos de gestión sanitaria. Medicina critica práctica. Edika Med; 2006. 2. Halpern NA, Pastores SM. Critical care medicine in de USA 2000---2005: on analysis of numbers, occupancy rate, payer mix and costs. Crit Care Med. 2010;38:65---71. 3. Friedrich J, Wilson G, Chant C. Long-term outcomes and clinical predictors of hospital mortality in very long stay intensive care unit patients: a cohort study. Crit Care. 2006;10:R59. 4. Santana Cabrera L, Sánchez Palacios M, Hernández Medina E, Eugenio Robaina P, Villanueva-Hernández Á. Características y pronóstico de los paciente mayores con estancias prolongadas en una UCI. Med Intens. 2008:157---62. 5. Sricker KH, Cavegen R, Takala J, Rothen HU. Does ICU length of stay influence quality of life? Acta Anaesthesiol Scand. 2005;49:975---83. 6. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies. J Chronic Dis. 1987;40:373---83. 7. Fayol P, Carriere H, Habonimana D, Preux PM, Dumond JJ. French version of structured interviews for the Glasgow Outcome Scale: guidelines and first studies of validation. Ann Readapt Med Phys. 2004;47:142---56. 8. Cabre L, Mancebo J, Solsona JF, Saura P, Gich I, Blanch L, et al. Multicentre study of de multiple organ dysfunction syndrome in intensive care units: the usefulness of SOFA in decision making. Intens Care Med. 2005;31:927---33. 9. Tan SS, Bakker J, Hoogendoorn ME, Kapila A, Martin J, Pezzi A, et al. Direct cost analysis of intensive care unit stay in four European countries: applying a standardized costing methodology. Value Health. 2012;15:81---6. 10. Barrientos Vega R. Nuestra experiencia con los Grupos Relacionados por el Diagnostico en Cuidados Intensivos. Med Intens. 2003;27:391---8. 11. Heyland DK, Konopad E, Noseworthy TW. Is it ‘‘Wortthwhile’’ to continue treating patients with a prolonged stay (
