Conserving resources after carotid endarterectomy: Selective use of the care unit

Conserving resources after carotid endarterectomy: Selective use of the care unit intensive Monica S. O'Brien, MS, RN, and John J. Ricotta, MD, Buffalo, N.Y. A retrospective review was undertaken of a random sample (N = 73) comprising 50% of carotid endarterectomies performed during 1986 to evaluate the necessity of routine postoperative intensive care unit (ICU) admission after carotid endarterectomy. Severity of illness was determined with use of the Acute Physiology Score of the APACHE II system. The Therapeutic Index Scoring System was used to quantify postoperative services used. Postoperative morbidity was analyzed. Financial impact was extrapolated with use of 1990 billing data. Length of ICU stay was 24.5 hours. Only 13 of 73 patients (18%) required ICU services. In 10 (77%) of these patients therapy was initiated in the recovery room and discontinued in six patients within 3 hours oflcu admission. Only two patients required ICU services for 16 hours after surgery. The mean Acute Physiology Score was low (4.96) and could not identify patients who required unique ICU services. Neurologic deficits were seen in five patients (6.9%). In three cases deficits were recognized in the recovery room; deficits developed in two patients after discharge from the ICU. Observation in the recovery room with transfer of stable patients would have eliminated ICU admission in 60 patients (82%). In 1990 the incremental ICU charge was $720/patient day. This represents 12.5% of the hospital charges for carotid endarterectomy. The ICU is an expensive and highly used hospital resource. Only a few patients need unique ICU services after carotid endarterectomy, and this is usually apparent within 2 hours of surgery. Prolonged recovery room observation or use of intermediate care units can avoid ICU admission for most patients undergoing carotid endarterectomy thereby conserving this precious hospital resource. (J VAsc SURG 1991;14:796-802.) The introduction of prospective payment during the early 1980s challenged hospitals to explore various avenues for containing costly hospital resources. Intensive care Unit (ICU) utilization practices have come under close scrutiny. More specifically, the necessity for routine postoperative ICU admission after various surgical procedures has become a topic of debate. Patients undergoing carotid endarterectomy (CEA) have been identified as a low-risk subgroup who require monitoring services rather than the unique care (active therapy) provided in an ICU. 1-~ This study examined the practice of routine postoperative ICU admission after CEA, possible mechanisms to identify patients requiring From the State University of New York at Buffalo Department of Surgery Division of Vascular Surgery and Millard Fillmore Hospitals Buffalo. Presented at the Fifth Annual Meeting of the Eastern Vascular Society, Pittsburgh Pa., May 2-5, 1991. Reprint requests: John J. Ricotta, MD, Department of Surgery, Millard Fillmore Hospital, 3 Gates Circle, Buffalo, NY 14209. 24/6/33418 796 ICU admission, and the financial impact of a selective ICU admission policy. METHODS A retrospective medical record review of a random sample comprising 50% (N = 73) of all CEAs performed at a university medical center during 1986 was completed. Severity ofiuness was determined by use of the Acute Physiology Score (APS) of the Acute Physiology and Health Evaluation 6,7 tool (APACHE 1I). The Therapeutic Index Scoring System 7,8 (TISS) was used to describe and quantify postoperative services used. Preoperative risk factors and postoperative morbidity were examined. The frequency of myocardial infarction, stroke, hypertension, or other acute medical conditions that might require ICU services was recorded. Financial impact was extrapolated with use of 1990 billing information. Data were analyzed with the personal computer version of the Statistical Package for Social Sciences (SPSS/PC; Chicago, I11.). The APS reflects the level of physiologic derange-

Volume 14 Number 6 December 1991 Conserving ICU resources after carotid endarterectomy 797 ment within seven body systems by assigning a score between 0 and 4 to 12 indicators. The composite score (APS) summarizes the patient's physiologic disturbance. 6 Content validity of the original APACHE tool has been established by its developers. The APS was found to be excellent in predicting subsequent need for active therapy in patients admitted for monitoring purposes. 5 It was found to correlate significantly with TISS (r = 0.59 p = 0.01-1 )7 In this investigation the APS was calculated with use of the most deranged values during the first 24 hours after admission to the ICU. The TISS first developed by Cullen and later revised by Knaus et al. 8 assigns a score between 1 and 4 to 73 potential therapeutic and diagnostic tasks provided to patients in the ICU. The cumulative score represents the total TISS or therapeutic effort. The indicators are further categorized as follows to differentiate the type of service rendered: Active therapy--care that is unique to the ICU. Monitoring services-observational tasks that require the level of staffing or sophisticated facilities available in an ][CU. Standard floor care- Treatments frequently provided to patients in the IGU, but which are routinely provided on general units. In this investigation TISS calculations were derived from the medical record at 8-hour intervals for the period in which subjects remained in the ICU up to 24 hours after operation. For patients admitted directly to general surgical units, TISS was calculated every 8 hours for a 24-hour period. RESULTS Thirty-seven (51%) of the 73 subjects studied were men. The mean age was 65 years, with a range of 43 to 83 years. Hypertension was the most common risk factor, found in 74% of subjects. Other risk factors included smoking (52%), coronary artery disease (43%), previous vascular surgery (35%), peripheral vascular disease (26%), and diabetes mellitus (25%) These preoperative characteristics were examined and appeared to be representative of patients undergoing CEA at the study institution. Postoperative morbidity and mortality rates were reviewed. Postoperative neurologic deficits developed in five patients (6.9%) in the sample, a~td two of these patients subsequently died, for a mortality rate of 2.8%. Examination of this subgroup revealed that the sample was not representative of the whole for this variable. The true incidence of postoperative deficit for the entire CEA population at the study site during 1986 was 4%, with an overall mortality rate of 1.4%. Major neurologic deficits developed in four patients in the sample, and a minor deficit occurred in one patient. Major deficits were noted immediately in two patients as they awoke from anesthesia, whereas deficits developed in the remaining two patients 60 hours and 11 days after discharge from the ICU. The patient who suffered minor stroke awoke from anesthesia with the deficit. Consequently, 60% (3/5) of all postoperative strokes were identified early in the recovery room, and 40% (2/5) occurred after discharge from the ICU. Three of the five patients in whom neurologic deficits developed used active therapy. One patient had an early major neurologic deficit and required ventilatory support. The second patient in whom an early minor deficit developed underwent active diuresis with mannitol. In the third patient, an asymptomatic thrombosis was detected in the recovery room, and the patient was returned to surgery for thrombectomy and revision. A stroke subsequently developed in this patient 3 days after discharge from the ICU. Other than reoperation this patient used no specific ICU resources. Postoperative hypertension was defined as a blood pressure greater than 180/110 mm Hg sustained for more than 1 hour during the initial 24-hour postoperative period, or any blood pressure requiring new drug treatment in addition to the patient's preoperative regime. Hypertension developed in 43 subjects (58.9%). Seventy-seven percent (33) of patients with postoperative hypertension were identified in the recovery room within the first postoperative hour. Identification improved to 90% after 3 hours of observation. Seventeen (40%) of the patients with hypertension required intravenous medication to control blood pressure. Nine patients (21%) required intravenous bolus medication, which is considered standard floor care, whereas eight patients (19%) required active treatment with continuous vasoactive drug infusion. This therapy was used in each of the patients to maintain blood pressure within a preestablished range. Control was achieved with low-dose infusion, and frequent titration was not required. Infusion was initiated in the recovery room in six of the eight patients (75%). Two patients were identified as hypertensive in the recovery room, but infusion therapy was not required until after admission to the ICU. Five of the eight patients who received active treatment in the ICU for hypertension required therapy for 2 to 3 hours. Only one patient required infusion for longer than 16 hours.

798 O'Brien and Ricot~a ~ournal of VASCULAR SURGERY Table I. Frequencies of monitoring services and active therapy used by patients after CEA (N = 73) Service No. Percentage + + ECG monitoring 71 970 + +Hourly vital signs 68 93% + -Systemic arterial monitoring 65 89% + + Hourly neurologic checks 55 75% + * +Vasoactive drug infusion 9 12.3% + * +Active diuresis 2 2.7% + - +Intermittent mandatory ventilation 1 1.4% + + + Emergency operative procedure 1 1.4% + +Monitoring services. + + + Active ICU therapy. Hypotension, defined as any decrease in blood pressure requiring intervention, occurred in four (5.5%) patients. Seventy-five percent (3) became hypotensive in the recovery room and were treated at that time. One patient was resuscitated with intravenous fluid, the other two received vasoactive drug infusion. This treatment was discontinued in one patient while in the recovery room. Therefore he was not considered to have used active ICU therapy. The other patient required drug infusion while in the ICU for 9 hours. The final patient became hypotensive 14 hours after operation and received treatment with intravenous fluid bolus. Arrhythmia developed in 40 (52%) patients after operation. Only six patients received treatment ([5] intravenous potassium supplements; [ 1] intermittent intravenous medication). These treatments are considered standard floor care therapies. Two patients (2.7%) experienced myocardial infarction. One patient exhibited signs within 2 hours of surgery, the other within 6 hours. Both patients were hypertensive within the first postoperative hour. One patient required active therapy with vasoactive medication. After operation patients who underwent CEA used only 19 of the potential 73 TISS indicators, most of which were standard floor care treatments. Monitoring services used by the group included systemic arterial pressure and continuous electrocardiogram (ECG) monitoring, hourly neurologic observations, and vital signs measurement. Only 18% (13/73) of the sample required care unique to the ICU. The active ICU services used were vasoactive drug infusion nine (12.3%), active diuresis two (2.7%), intermittent mandatory mechanical ventilation one (1.4%) and emergency reoperation one (1.4%) (Table I). No single patient required more than one active therapy. Active treatment was initiated in the recovery room in 77% (10/13) of patients requiring therapy and discontinued in 46% (6) within 2 to 3 hours after ICU admission. Only 5% (3/63) of patients initially admitted to the ICU for monitoring required subsequent active therapy. Acute Physiology Scores for the sample ranged from 0 to 11, with a mean of 4.96 and a mode of 4.0. The mean APS among patients requiring active therapy was 5.4 _+ 2.5. Patients in the ICU who did not use unique ICU services had a mean APS of 4.8 + 2.2. (20 = 0.4284). Therapeutic index scoring system scores ranged from 7 to 21, with a mean of 12.69 and a mode of 12. Weak but significant correlation was observed between APS and TISS scores (r = 0.20051; p = 0.0445). The mean TISS score for patients receiving active therapy was 16.6 _+ 2.2 and 11.8 + 2.6 (d0 < 0.001) for those who used only monitoring services and standard floor care. By use of a classification system developed by Knaus et al.,2 subjects were categorized into four classes based on TISS composite. Scores of 1 to 12 points were assigned to class I; 13 to 20 points to class II; 21 to 30 points were assigned to class III; and more than 30 points to class IV. Results are sunamarized in Table II. Ninety-nine percent of the subjects were categorized as class I or II admissions. Most (56%) were class I admissions who used standard floor care and monitoring services, but not active therapy. One class III patient received 21 TISS points. No class IV patients were in the sample. During periods of ICU bed shortage, four patients undergoing CEA were admitted directly to general surgical units from the recovery room. These patients had a lower severity of illness ( APS = 3.25) than did patients in the ICU ( APS = 5.03) and used fewer TISS points (~TISS = 7.75) than those admitted to the ICU (~TISS = 12.99). Therapeutic index scoring system scores progressively declined in both groups as time elapsed. The mean length of ICU stay was 24.5 hours. At 8 hours after operation, the mean TISS score for patients in the ICU was 12.99, decreasing to 8.1 after 16 hours. The mean TISS score for patients admitted to general units was 7.75, decreasing to 6.5 at 16 hours (Table III). According to 1990 billing data, the average hospital charge for patients undergoing CEA was $5728.62. The charge for an ICU bed per patient day was $1000 as compared to $280 for a general surgical bed. The difference ($720) represents 12.5% of total CEA hospital charges.

Volume 14 Number 6 December 1991 Conserving ICU resources after carotid endarterectomy 799 DISCUSSION This study demonstrates that most patients after operation for CEA have minimal physiologic derangement, and rarely use unique ICU services. The mean APS (4.96',1 of patients undergoing CEA in the sample is consistent with scores reported for low-risk monitor admissions. Knaus et al. 9 suggested that ICU admission may not be warranted in patients with APS less than 5 who do not receive active therapy. By use of those standards, 48.3% of the sample would not have required ICU admission. The mean APS (5.4) of subjects who used active therapy was only slightly' above that standard and considerably lower than APS ( x 11.34) reported for patients likely to require active therapy. 1 In our study even patients receiving active treatment were relatively stable. Although 18% of the sample required :active therapy, only four of the 33 active treatment indicators were used. No single patient used more than one active treatment. Vasoactiv drug infusion, the most common active service used, could have been delivered in less intensive settings such as a step down unit. Composite TISS scores further support the contention that patients undergoing CEA can be cared for in less intensive environments. The mean TISS score for the sample was 12.69 and decreased to 8.1 after 16 hours. Cullen et al. n suggested that TISS scores of < 10 may constitute inappropriate ICU admission other than to rule out myocardial infarction. Monitoring interventions and standard floor care treatments rather than active therapy contributed to the preponderance of TISS points in these patients and could have been provided outside the ICU. The results of this study indicate that patients who require active IC, U treatments can be identified early in the postoperative period. Ten of the 13 patients requiting active therapy were identified in the recovery room. Active therapy was discontinued in most patients requiring treatment, within 2 to 3 hours after ICU admission and was reflected in the progressive decline in TISS scores. Early discharge from the ICU or prolonged recovery room stay is a potential alternative to the standard 24-hour ICU stay. Comparison of resource use between patients admitted directly to surgical units and those admitted to the ICU revealed that patients in the ICU continued to use more resources at 8 to 16 hours after operation than general unit patients used during the initial 8 hours after operation. One explanation is that patients may receive treatments based on ICU admission protocol. Hourly vital signs, neurologic observations, and use of ECG and arterial pressure Table II. Categorization of patients based on total TISS (N = 73) Composite Active Category TISS score No. Percentage treatment I 1-12 41 56.2 0 II 13-20 31 42.5 12 III 21-30 1 1.4 1 IV > 30 0 0 0 Table III. Measures of central tendency of APS and TISS scores in general surgical unit and ICU postoperative CEA admissions (N = 73) Mean TISS Mean TISS Category No. Mean APS at 8 hours at 16 hours Floor 4 3.25 7.75 6.5 ICU 69 5.03 12.99 8.1 monitoring are examples of services that may be routine. The weak correlation between APS and TISS supports this hypothesis. Many services were provided to patients whose severity of illness was relatively low. The need for close monitoring to prevent postoperative morbidity cannot be refuted. However, consideration must be given as to whether invasive equipment is needed to achieve adequate surveillance and to the length of time such monitoring is required. Efficient resource use demands matching patients' real needs to the services provided. Patients are often admitted to the ICU after CEA to monitor their neurologic status, cardiac rhythm, and blood pressure. None of the neurologic deficits that occurred in our series were detected in the ICU; three occurred in the recovery room and two after discharge from the ICU. Ninety percent of patients with hypertension and 75 % of patients with hypotension patients were identified within 3 hours of surgery. Infusion lasted < 3 hours in most (56%) of patients requiring this therapy. Myocardial infarction occurred in 2.4% of patients in this stud),, consistent with previous reports. 12 Although arrhythmias occurred in 50% of patients, only 8% of the entire group required intervention. Certainly, close observation with capability for telemetry and intravenous drug infusion can be provided in a step down unit rather than an ICU. Discussions of"charges" and "cost" are confusing in the current climate where reimbursement is fixed by diagnosis, and hospital charges may have tittle meaning. However imprecise, these charges are the

800 O'Brien and Ricotta Journal of VASCULAR SURGERY current mechanism that hospitals use to quantify their resource use for a particular diagnosis. The effect of fixed and variable costs further complicates this problem, especially in areas where resources are plentiful and may be "underutilized." It is incontrovertible that care delivered in an ICU requires more resources than similar care given elsewhere. Optimal resource use, matching the care delivered to the patient's needs, is to the ultimate benefit of the hospital, society, and above all, the patient. One of the most expensive components of ICU care is the personnel involved in care. Development and use of step down units, with lower nurse/patient ratios allows personnel savings (or reallocation) while not compromising patient care. The most pressing argument for efficient use of ICUs is medical not economic. As inpatient acuity increases and hospitals are forced to reduce costs because of limited reimbursement, critical care beds are fast becoming a treasured resource. In many parts of the country these units function at or near capacity most of the time, and it is not uncommon to cancel elective surgery because of a lack of critical care beds. Physicians can best serve all their patients by using these beds appropriately. This study demonstrates the need for critical appraisal of our current practices and physician-directed efforts to improve use without compromising patient care. Special thanks to Barbara Peterangelo for her help with the manuscript and to Richard Green, MD, and James DeWeese, MD, for permission to study their patients. REFERENCES 1. Draper EA. Benefits and cost of intensive care. Image: J Nurs Scholar 1983; 15:90-4. 2. Knaus WA, Wagner DP, Draper EA, Lawrence DE, Zimmerman JE. The range of intensive care services today. JAMA 1981;246:2711-6. 3. Knans WA, Draper EA, Wagner DP. The use of intensive care: new research initiatives and their implications for national health policy. Milbank Memorial Fund, Quarterly/Health and Society 1983;61:561-79. 4. Nelson JB. The role of an intensive care unit in a community hospital: a ten-year review with observations on utilization past, present, and future. Arch Surg 1985;120:1233-6. 5. Wagner DP, Knans WA, Draper EA, Zimmerman JE. Identification of low-risk monitor patients within a medicalsurgical intensive care unit. Med Care 1983;21:425-34. 6. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med 1985;13:818-29. 7. Knaus WA, Zimmerman JE, Wagner DP, Draper EA, Lawrence BS. APACHE - acute physiology and chronic health evaluation: a physiologically based classification system. Crit Care Med 1981;4:591-7. 8. Knaus WA, Draper EA, Lawrence DE, Wagner DP, Zimmerman JE. Neurosurgical admissions to the intensive care unit: intensive monitoring versus intensive therapy. Neurosurgery 1982;8:438-41. 9. Knaus WA, Draper EA, Wagner DP. Toward quality review in intensive care: the APACHE system. Q Rep Bull 1983;9: 196-204. 10. Wagner P, Knaus WA, Draper EA. Identification of low risk monitor admissions to medical-surgical ICUs. Chest 1987; 92:423-8. 11. Cullen DJ, Civetta JM, Birggs BA, Ferrara LC. Therapeutic intervention scoring system: a method for quantitative comparison of patient care. Crit Care Med 1974;2:57-60. 12. Kirshner D, O'Brien M, Ricotta JJ. Risk factors in a community experience with carotid endarterectomy. J VASC SURG 1989;10:178-86. 13. Godin MS, Bell WH, Schwedler M, Kerstein MD. Cost effectiveness of regional anesthesia in carotid endarterectomy. Am Surg 1989;55:656-9. Submitted May 13, 1991; accepted Aug. 29, 1991. DISCUSSION Dr. Dhiraj Shah (Albany, N.Y.). Cost containment in health care is a concern for everyone, and the authors are showing us a way to cut down the cost of CEA. The main point is that if all patients do not routinely go to the ICU after operation for CEA, then it saves approximately 12.5% of the total cost in the hospital. To identify those patients who need ICU care one may use certain physiologic scores. Our approach, however, to carotid surgery is somewhat different, keeping in mind both cost and quality of care. We do CEA under regional block anesthesia, and the patients do not go to ICU. Over the last 10 years we have done 732 CEAs, of which 389 were done under general anesthesia in prior years, and all went to the ICU after operation. More recently we have done 336 CEAs with the patient under regional anesthesia, and only three patients needed to go to the ICU, although seven of the patients had to be converted to general anesthesia. One patient was admitted to the ICU 2 days after stroke and aspiration. The overall stroke rate in our patients is 1.2%, and the mortality rate is 1%. Furthermore, since we started using regional anesthesia, the length of the stay in the hospital for carotid artery surgery is 2 days to 4 days shorter than the DRG allowance. Thus CEA under regional anesthesia provides for a significant cost savings in our hospital. We have also