From the Departments of Health Services Research and Orthopedic Surgery, Cedars-Sinai Medical Center, Cedars-Sinai Research Institute, and UCLA School ...
26
Articles
Total Knee Replacement A Guideline to Reduce Postoperative Length of Stay SCOTT R. WEINGARTEN, MD, MPH; LAURA CONNER, RN; MARY RIEDINGER, RN, MSN; ANTHONY ALTER, MD; WILLIAM BRIEN, MD; and A. GRAY ELLRODT, MD, Los Angeles, California
In a retrospective study in an academic, acute-care community hospital, we studied the possible safety and effectiveness of a practice guideline recommending early discharge from the hospital for patients having uncomplicated total knee replacement. Of 206 patients receiving knee replacements, 162 (79%) were classified by the guideline as being at low risk for complications between the 4th and 7th postoperative days. Use of the guideline could have reduced the postoperative length of stay from 7.3 ± 2.6 days to 4 days for the 1 2 patients (54%) who became low risk on the 4th postoperative day. Explicit and implicit review of the quality of care determined that 157 patients (96.9%; 95% confidence interval, 92.9%, 99.0%) could have been safely transferred from the acute-care hospital to an appropriate setting when they became classified at low risk between the 4th and 7th postoperative days. Clinical practice guidelines can possibly be used to reduce the postoperative length of acute-care hospital stay for patients having knee replacements. This guideline requires further study in a controlled clinical trial before it can be recommended for use. (Weingarten SR, Conner L, Riedinger M, Alter A, Brien W, Ellrodt AG: Total knee replacement-A guideline to reduce postoperative length of stay. West J Med 1995; 163:26-30) bout 125,000 knee replacement operations are done annually in the United States at a cost of more than $3 billion per year.',2 Even though the frequency of knee replacement exceeds that of hip replacement,' little is known about the cost-effectiveness of the care of these patients after the operation. Previous studies of common medical conditions have shown relatively safe ways to shorten lengths of hospital stay and reduce hospital costs.36 Patients undergoing knee replacements may achieve similar benefits. Practice guidelines and clinical pathways to enhance safe and effective care for hospitalized patients with common conditions are being developed.7'-" Valid and reliable clinical data are essential in this effort, but are often lacking.7 The American Academy of Orthopaedic Surgeons' discharge guidelines for patients receiving total knee arthroplasty do not have recommendations about the medically or surgically appropriate length of hospital stay.'2 We retrospectively studied a clinically derived practice guideline that prescribed the appropriate length of stay in an acute-care hospital for patients undergoing knee replacement operations. In a hypothetical experiment, we determined the possible benefits (reduced lengths of stay) and risks (premature hospital discharges) of this guideline if it were to be used to identiA
fy patients for early discharge from the acute-care hospital to a more appropriate level of care. We hypothesized that acute hospital care could be provided at lower cost without compromising the quality of care.
Patients and Methods The study was performed at Cedars-Sinai Medical Center, a large teaching community hospital that serves West Los Angeles, California. Most patients admitted to this hospital are cared for by physicians in private practice. The study was approved by the Institutional Review Board at the study hospital. Patients Patients consecutively admitted to the hospital were potentially eligible for inclusion in the study if they received a total knee replacement between January 1991 and April 1993. Patients were selected from the hospital database by International Classification of Diseases (9th revised), Clinical Modification (ICD-9-CM) codes.13
Patients who underwent revision of a previous knee replacement or who had bilateral knee replacement were excluded from the study. Practice Guideline The guideline was developed through the consensus of two board-certified orthopedic surgeons and a board-
From the Departments of Health Services Research and Orthopedic Surgery, Cedars-Sinai Medical Center, Cedars-Sinai Research Institute, and UCLA School of Medicine, Los Angeles, California. Reprint requests to Scott Weingarten, MD, Cedars-Sinai Medical Center, Becker 146, 8700 Beverly Blvd, Los Angeles, CA 90048.
WJM, July 1995-Vol 163, No. I
.1995-Vol163,
1
TABLE 1.-Risk Classification Criteria for Low-Risk Patients Having Total Knee Replacement* Criteria
Class ...........
Hypovolemia (at time of classification) Dehydration
Ila .
llb.
Cardiopulmonary complications (at any time before
classification) Acute myocardial infarction Pulmonary embolism Ventricular arrhythmias Supraventricular arrhythmias Pacemaker malfunction Respiratory acidosis True syncopal episode Cardiopulmonary complications (at any time before classification) Hypotension Tachypnea
Hypoxemia
Acute-onset congestive heart failure or pulmonary
edema III
Surgical complications (at any time before the time of classification)
Joint infection latrogenic fracture Wound dehiscence Wound infection Wound necrosis Seroma Dislocation of the implant Return to the operating room for orthopedic reasons IV ......... Infectious complications (at time of classification) Fever Nosocomial infections being treated with intravenous antibiotics Va ......... Other unstable conditions (at time of classification) Acute changes in mental state Severe anemia Hyponatremia Hypokalemia or hyperkalemia Severe hyperglycemia Drug toxicity
Vb .
Postoperative ileus Patient requires total parenteral nutrition Observed inability to take medications or fluids orally Substantial gastrointestinal hemorrhage New-onset stroke Serious medication reaction Other unstable comorbid conditions requiring continued acute care Other unstable comorbid conditions (at any time before classification) Deep venous thrombosis Return to operating room for nonorthopedic reasons
VI
Functional status (at time of classification) Patients not yet able to transfer out of bed
'if a patient satisfies anv of these crteria, the patient mav not be at loin risk and suitable for transfer or discharge from the acute-care hospital.
certified general internist (A.A., W.B., S.R.W.). We had previously studied a similar guideline for patients having total hip replacement.'4 Patients who did not have obvious reasons for continued hospital stay (as explicitly defined by the guideline) on the fourth through the seventh postoperative days were classified as being at "low risk" (Table 1). These patients were considered
Total Knee
Replacement-Weinqarten et al
27
as possibly no longer requiring treatment in an acutecare hospital. Complications-Explicit Review We used a two-stage process to identify and classify patient complications. In the first stage, the data were abstracted from the medical record by a research nurse (L.C.). Complications of total knee replacement as defined for purposes of the study included death, the requirement of treatment in the intensive care unit, a systolic blood pressure of less than 100 mm of mercury, gastrointestinal distress and an inability to take food or medications orally, dehydration requiring intravenous therapy, acute mental state changes, hypoxemia while the patient was breathing room air (Po2 < 60 torr or pulse oximetry < 90%), hypercapnia (Pco2 > 50 torr), congestive heart failure, acute myocardial infarction, serious change in comorbid diseases, ventricular fibrillation, sustained ventricular tachycardia, asystole, unstable atrial fibrillation, unstable atrial flutter, supraventricular tachycardia, wound dehiscence, wound seroma, fever (temperature > 37.8°C [100.0°F]), postoperative pneumonia, anemia requiring transfusion, dislocation of the implant, peroneal nerve palsy, sciatic nerve palsy, seizure, postoperative hematoma, hemarthrosis, and deep venous thrombosis. Patients who were determined to be at low risk on the fourth through the seventh postoperative days and who then suffered any of the explicitly defined complications after that day were considered to be possibly misclassified by the guideline. The medical records of these patients were then submitted for implicit review. Complications-Implicit Review In the second stage of the review, all complications occurring in low-risk patients after they became low risk were implicitly and independently reviewed by two physicians (a board-certified orthopedist and a boardcertified general internist [A.A., S.R.W.]). These physicians rated whether transfer to a skilled nursing facility or discharge of the patient to home would have worsened the quality of care.15 This method of judging quality of care has been used in previous investigations.15 Functional Status Each patient's ability to transfer and walk was determined by examining the physical therapists' progress notes. If the physical therapy record was missing documentation, the last recorded value was used. Resource Use-Lengths of Stay The expected benefit afforded by the guideline was the projected reduction in the number of acute-care hospital bed-days. Lengths of stay were determined as per the midnight census. For each low-risk patient, the hypothetical number of acute-care hospital bed-days saved was defined as the number of actual acute-care hospital bed-days less the number of days in the acutecare hospital recommended by the practice guideline.
28
28-
Total Knee
WJM, July 1995-Vol 163, No. I WIM-Jul
195-o
Replacement-Weingarten et al
-16,N.1TtlKeRelemnWic
Patients who were deemed not suitable for transfer or discharge by the guideline were not included in the analysis.
neta
TABLE 2.-Patient Characteristics (n = 206) Patients
Statistical Analysis
Characteristic
No.
%
The primary end points of the study were the projected effect of the guideline on quality of care (misclassification rate) and the number of hospital bed-days that could have been saved. The misclassification rate was defined as the number of low-risk patients who suffered complications after they became low risk according to the guideline. Means are reported with standard deviations. The 95% confidence intervals (CIs) were calculated using the software package True Epi-Stat."1 End points were tested using a type I error of .05.
Sex, female .....1...
1 37
68
196 6
97 3
30 21 21 21 18 3
15 10 10 10 9 1
Data Quality
The inter-rater agreement on the guideline classification of whether a patient had an obvious reason for a continued hospital stay on the fourth postoperative day was 95% (K 0.89); fifth postoperative day, 95% (K 0.85); sixth postoperative day, 95% (K 0.64); and seventh postoperative day, 89% (K 0.46). Inter-rater agreement on the implicit review of complications between the orthopedic surgeon and internist was 93% (K 0.71). Results Demographics A total of 206 patients were included in the study. Medical records were available for abstraction on 202 (98%) eligible patients. The mean patient age was 68.9 ± 10.9 years (mean ± standard deviation). Patient characteristics are described in Table 2. The mean hospital length of stay was 8.0 ± 3.8 days, and the mean postoperative length of stay was 7.8 ± 3.6 days. Patients were able to transfer out of bed and begin physical therapy in an average of 1.6 ± 0.7 and 1.3 ± 0.6 days after their surgical procedure, respectively. The discharge dispositions of the patients were as follows: 125 patients (61%) were transferred to a hospital-based skilled nursing facility, 64 (31%) were discharged to home, 3 (1%) were transferred to a freestanding skilled nursing facility, 1 (0.5%) died during the hospital stay, and 13 (6%) had a disposition other than these categories. Guideline Classification A total of 162 patients (78.6%) were at low risk during the fourth through the seventh postoperative days (Table 3). For example, for the patients who became low risk on the fourth hospital day, the postoperative length of stay was 7.3 ± 2.6 days (Table 3). None of these patients were discharged from the hospital before the fourth postoperative day. Had the guideline been applied to patients who became low risk at any time during the specified period, costs associated with 538 bed-days (223 bed-days per year) might have been reduced. All patients were transferring out of bed, and many patients were walking substantial distances when they
Surgery type Unilateral knee arthroplasty ............... Unicompartmental arthroplasty ........ ... Comorbid diseases Coronary artery disease .................. Chronic obstructive pulmonary disease Diabetes mellitus ................ Rheumatoid arthritis ..................... Congestive heart failure .................. Cerebrovascular accident ........3........
classified as low risk according to the guideline (Table 4). Explicit and Implicit Judgment of Quality of Care A total of 29 patients (18%; 95% CI, 12%, 25%) suffered minor complications after they became low risk (Table 5). Implicit review determined that for 157 of these patients (97%; 95% CI, 93%, 99%), the quality of care would not have been compromised had they been transferred (or sometimes discharged home) to a more appropriate location when they became low risk. Patients whose care might have been compromised had they been transferred are listed in Table 6. Discussion Most patients recovering after total knee replacement had an uncomplicated clinical course and were possible candidates for early transfer or discharge from the acutecare hospital to a more appropriate level of care. Correspondingly, these patients suffered few clinically important complications after they were classified as low risk. Patients' actual length of stay exceeded the recommendation in all cases, which signifies an opportunity to safely shorten lengths of stay for these patients. There was also great variation among physicians in length of stay for low-risk patients, which indicates a lack of consensus about the most appropriate length of stay.'7 Clinical data such as those collected for this study may reduce this uncertainty. Reducing the length of stay for patients having a total knee replacement could subwere
TABLE 3.-Patients With No Obvious Reason for Continued Acute Care in Hospital Postoperotive Day
Patients at "Low Risk" No. (%)
Length of Stay, days*
4
112
(54)
7.3±2.6
5
32
(16)
7.4 ± 1.4
6
13
(6)
11.0±4.7
7
5
(2)
9.0 ±1.2
*Postoperative days. The numbers represent the mean standard deviation. ±
WJM, July 1995-Vol 163, No. 1
Total Knee Replacement-Weingarten et al 29
WIMI
TABLE 6.-Patients Who May Have Been Adversely Affected Had They Been Transferred or Discharged When They Became Low Risk*
TABLE 4.-Distance Walked on the 4th Through 6th Postoperative Days by Low-Risk Patients Having a Knee Replacement (n = 157)
Feet Walked
No.
Unknown. .................
1-10. 11-25 ..... 26-50. 51-150 ..... >1 50 .8
1 ......
5th,
6th,
5 2 1 7 7 8 2
2 1 0 4 2 3 1
No.
No.
12 6 14 3 19 40
stantially reduce the cost of caring for these patients both at our hospital and on a national level. Our study is limited in several ways. First, the data were collected retrospectively and only during the acutecare hospital stay (patient data were not collected after discharge from the acute-care hospital), and we cannot be certain that some patients would not have derived at least some benefit from continued stay in the acute-care hospital. A prospective study done with the data collected after hospital discharge would be necessary to exclude this possibility, however. Second, although we were able to conclude that low-risk patients may be discharged from an acute-care hospital earlier than is current practice, we are unable to determine whether these patients would be better suited going to a skilled nursing facility or to their home. Finally, this study was done at only one hospital. The possible advantage of using an objectively derived guideline to prescribe the appropriate length of stay for patients with common conditions is the ability to predict the safety and efficacy of earlier hospital discharge. Our analysis allows discharge decisions to be based on clinical data rather than on economically derived data or on "customary and usual practice." These data also may be used to develop practice guidelines and clinical pathways and to determine the usual progress of patients after a knee replacement operation. Furthermore, early hospital discharge from an acute-care TABLE 5.-Complications Occurring After the Patient Was Designated 'Low Risk' by Practice Guideline*Results of Explicit Review Patients, N'o*
Compilcation Fever (temperature >37.8-C [1 00lF]) ....... ......... 20 Anemia requiring transfusion .......... ............. 9 Systolic
blood pressure '100
mm
of mercury
4 4
Hemarthrosis
Deep venous thrombosis .............. Congestive heart failure ................ Hematoma
............. ............
3 2
...................................
Substantial change in comorbid disease
......
Total
'Several patients hao more tnan 1 complication. tcomoiicat on rate, 1 7.9%5 f95Qi. Cl; 12.3-%, 24.7'10).
.......
Age. yr
Sex
Case Description, Treotment, ond Outcome
67
F
2 ....
90
F
3 ........
73
M
Deep venous thrombosis developed on 8th postoperative day; anticoagulation therapy given; recovery uneventful Received 2 units of packed erythrocytes on 5th postoperative day; mild congestive heart failure developed; oral furosemide given; recovery uneventful Deep venous thrombosis developed on 6th postoperative day; anticoagulation therapy given; recovery
Patient
Postoperative Doys
4th,
1 29t
I
........
uneventful
4 ....
81
F
5.
72
F
Fever and mild congestive heart failure developed after 6th postoperative day; 40 mg of furosemide IV and packed erythrocytes given; recovery uneventful Fever and changes on ECG developed after 5th postoperative day; serial cardiac enzymes showed no evidence of acute myocardial infarction; 2 conflicting ventilation-perfusion scans showed no definite pulmonary embolism; treated with packed erythrocytes; recovery otherwise uneventful
ECc = e'ectrocardiogram. 'IV in travenous 'According to implic.t revie.v
hospital, when appropriate, may benefit patients by reducing the risk of nosocomial infections and other iatrogenic complications.'8 Acknowledgment Ms Vanessa Walker and the Medical Records Department of the Cedars-Sinai Medical Center assisted with this project. REFERENCES 1. Data from the National Hospital Discharge Survey: Detailed diagnoses and procedures, 1990. Vital Health Stat [13] 1992, No. 113 2. Harris WH, Sledge CB: Total hip and total knee replacement. N Engl J Med 1990; 323:801-807 3. Weingarten S, Ermann B, Riedinger M, Shah PK, Ellrodt AG: Selecting the best triage rule for patients hospitalized with chest pain. Am J Med 1989; 87:494-500 4. Weingarten S, Ermann B, Bolus R, et al: Early 'step-down' transfer of low-risk patients with chest pain-A controlled interventional trial. Ann Intem Med 1990; 113:283-289 5. Weingarten S, Agocs L, Tankel N, Sheng A, Ellrodt AG: Reducing lengths of stay for patients hospitalized with chest pain using medical practice guidelines and opinion leaders. Am J Cardiol 1993; 71:259-262 6. Weingarten SR, Riedinger MS, Shinbane J, et al: Triage practice guideline for patients hospitalized with congestive heart failure: Improving the effectiveness of the coronary care unit. Am J Med 1993; 94:483-490 7. Field MJ, Lohr KN (Eds): Guidelines for Clinical Practice. Washington, DC, Institute of Medicine, National Academy Press, 1992 8. Audet AM, Greenfield S, Field M: Medical practice guidelines: Current activities and future directions. Ann Intem Med 1990; 113:709-714 9. Woolf S: Practice guidelines: A new reality in medicine-I. Recent developments. Arch Intem Med 1990; 150:1811-1818 10. Directory of Practice Parameters: Titles, Sources, and Updates-Quality Assessment, Quality Assurance and Quality Improvement Programs. Chicago, American Medical Assoc, 1993 III, 11. Coffey RJ, Richards JS, Remmert CS, LeRoy SS, Schoville RR, Baldwin PJ: An introduction to critical paths. Qual Manage Healthcare 1992; 1:45-54
30
WJM, July 1995-Vol 163, No. 195Vl13
o
Total Knee Replacement-Weingarten TtlKe elcmetWIgre ettaal
I
15. Donabedian A: Explorations in Quality Assessment and Monitoring, Vol 2. Ann Arbor, Mich, Health Administration Press, 1982, pp 19-21 16. Gustafson T: True Epi-Stat, 4th edition. Richardson, Texas, 1992 17. Eddy DM: Variations in physician practice: The role of uncertainty. Health Aff (Millwood) 1984; 3:74-89 18. Myers JD: Preventing iatrogenic complications. N Engl J Med 1981; 304:664-665
12. Clinical Policies. Park Ridge, Ill, American Academy of Orthopaedic Surgeons, 1992 13. Commission on Professional and Hospital Activities: Intemational Classification of Diseases, 9th revised-Clinical Modification. Ann Arbor, Mich, National Center for Health Statistics and WHO, 1978 14. Weingarten SR, Riedinger M, Conner L, et al: Hip replacement and hip hemiarthroplasty surgery-Opportunities to shorten lengths of hospital stay. Am J Med, in press
*
*
*
Sleep's Siren Calls Three hours of sleep is not enough to forget the world: a 4 AM darkness pants damply against the window, with deep, dull clangs and high frequency hisses the radiator launches while beeper's unbloodied blade repeatedly pierces splayed-out senses. I'd been dreaming too sweetly for this offensivethe mind's first stirring is to question everything. The inky apparitions piled in my clothes surrounded by name tag, 3 x5 cards and clipboard menace my mind into chanting: "Swarming vat of mental vapors let rise barely contained visions of marked surprise to put the steam again in sleepy eyes." Nearly awake, I still question everything. Like the wayward chutist wrapped in a web of branches I wonder if I can make my arms and legs move. As a creepy panic scrambles over exposed nerves I count how many more call nights I have to endure and wonder how other doctors shake this stupordo they also question everything? Talks with patients, serious and humorous, and conversations with 5 AM-punchy colleagues drag back with me to my suite "to get my one last hour of sleep" crowded with dreams who visit, nurture and leave me marveling that I questioned anything. TIM VAN ERT, MD0
Selah, Washington