Venous thrombosis incidence in burn patients: preliminary results of a prospective study.

There are few prospective data on the incidence of deep venous thrombosis (DVT) in burn patients. In an on-going prospective study, hospitalized burn patients 18 years or older with an expected hospital length of stay more than 72 hours were imaged with baseline venous duplex ultrasound of all extremities within the first 48 hours after admission and weekly until discharge. Patient demographics and clinical risk factors for DVT were assessed. At the time of submission, 40 patients met screening criteria, and 30 were enrolled. Ultrasound diagnosed seven patients with 11 acute DVT for an incidence of 23%. One pulmonary embolism was documented. DVT patients had a mean age of 49 +/- 23 years with an average TBSA burn of 15 +/- 4% compared with those without thrombosis with a mean age of 44 +/- 17 years (P = NS) and TBSA burn of 18 +/- 25% (P = NS). There were no statistically significant differences for DVT patients in terms of age, number of central line days, hospital length of stay, or TBSA burned. Given the preliminary findings of this small study, we believe that all hospitalized burn patients are at risk for DVT. On-going investigation will be helpful in defining level of risk and improved prevention strategies for thromboembolic complications in burn patients.

Computed tomographic venography is specific but not sensitive for diagnosis of acute lower-extremity deep venous thrombosis in patients with suspected pulmonary embolus.

PURPOSE: Duplex ultrasound scanning (US) is the accepted standard means of diagnosis for lower-extremity suprageniculate deep venous thrombosis (LE-DVT). Computed tomographic venography (CTV) has been proposed as an alternative modality for diagnosis of LE-DVT in patients with suspected pulmonary embolism (PE). This study compared CTV with US as a means of diagnosing acute LE-DVT. METHODS: A retrospective review of US and CTV scans from 136 patients with suspected PE who underwent both studies to exclude acute LE-DVT at a single institution was performed. Studies were reviewed and coded in a blinded manner. US was considered to be the reference test. Direct costs of each study were determined by using commercial software. RESULTS: The sensitivity and specificity rates of CTV were 71% and 93%, respectively. The positive predictive value, negative predictive value, and accuracy rates of CTV were 53%, 97%, and 90%, respectively. DVT localization was the same in eight of 10 cases in which the results of both US and CTV were positive. CTV costs and charges per study were greater than those of US by $46.88 and $602.00, respectively. CONCLUSION: CTV is specific, but has a lower sensitivity rate and positive predictive value for the diagnosis of acute LE-DVT compared with US. Additionally, CTV is more costly than US scanning. Because of the lower sensitivity rate and positive predictive value and the increased cost of CTV, US remains the screening study of choice in cases of suspected acute LE-DVT.

Burnout among American surgeons.

BACKGROUND: The long-term consequences of stress on the surgeon are unknown. One manifestation of stress is burnout. The purpose of this study was to measure the prevalence of burnout in actively practicing American surgeons. METHODS: The Maslach Burnout Inventory and a questionnaire of our own design were sent to 1706 graduates of various University of Michigan surgical residencies (1222) and members of the Midwest Surgical Association (484). The response rate was 44%. Responses from 582 actively practicing surgeons were the sample used for analysis. RESULTS: Thirty-two percent of actively practicing surgeons showed "high" levels of emotional exhaustion, 13% showed "high" levels of depersonalization, and 4% showed evidence for low personal accomplishment. Younger surgeons were more susceptible to burnout (r = -0.28, P <.01). Burnout was not related to caseload, practice setting, or percent of patients insured by a health maintenance organization. Important etiologic factors were a sense that work was "overwhelming" (r = 0.61, P <.01), a perceived imbalance between career, family, and personal growth (r = -0.56), P <.01), perceptions that career was unrewarding (r = -0.42, P <.01), and lack of autonomy or decision involvement (r = -0.39, P <.01). A strong association was noted between burnout elements and a desire to retire early (r = 0.50, P <.01). CONCLUSIONS: Burnout is an important problem for actively practicing American surgeons. These data could be used to modify existing surgical training curricula or as an aid to surgical leadership when negotiating about the surgical work environment.

A clinical comparison of pneumatic compression devices: the basis for selection.

PURPOSE: The five pneumatic compression devices (PCDs) that are marketed provide mechanical protection from deep venous thrombosis (DVT). They differ with respect to patterns of compression and the length of the sleeve. Evidence linking differences to clinical outcomes is lacking. Our purpose was twofold: to evaluate each of the marketed PCDs with respect to effectiveness, compliance, and patient and nursing satisfaction and to determine whether there is a clinical basis for the selection of one device over another. METHODS: Each of the marketed devices was used exclusively for a 4-week period. Patients participated in an evaluation including venous duplex ultrasound scan, DVT risk assessment, and device evaluation. Vascular laboratory records were used to document DVT. Compliance was measured by meters installed on all pumps. A ranking matrix was stratified by compression pattern: rapid graduated sequential compression, graduated compression, and intermittent compression, and each device was rated by patients and nurses. RESULTS: The PCDs were used in 1350 cases with a DVT rate of 3.5% ranging from 2% to 9.8% depending on the method of compression. Patients with DVT were older (58 vs 54 years), had better compliance (67% vs 50%), and had more compression days (11 vs 7.2). When thigh-length sleeves were used, a greater proportion of DVT occurred in the proximal segments (71%) as compared with the number of proximal DVT when the calf-length devices were used (52%; P =.21). Devices W, X, and Y had comparable rates of DVT, which were lower than those for V and Z. Compression device W, [correction] with calf and thigh sleeves, achieved the best overall ranking largely because of high scores for patient and nurse satisfaction. CONCLUSION: Our data appear at odds with commonly held beliefs. We were unable to show a difference in DVT incidence based on the length of the device or the method of compression. Randomized studies are needed to confirm our findings and evaluate hypotheses derived from this study.

Renal artery aneurysms: a 35-year clinical experience with 252 aneurysms in 168 patients.

OBJECTIVE: To define the relevance of treating renal artery aneurysms (RAAs) surgically. SUMMARY BACKGROUND DATA: Most prior definitions of the clinical, pathologic, and management features of RAAs have evolved from anecdotal reports. Controversy surrounding this clinical entity continues. METHODS: A retrospective review was undertaken of 168 patients (107 women, 61 men) with 252 RAAs encountered over 35 years at the University of Michigan Hospital. Aneurysms were solitary in 115 patients and multiple in 53 patients. Bilateral RAAs occurred in 32 patients. Associated diseases included hypertension (73%), renal artery fibrodysplasia (34%), systemic atherosclerosis (25%), and extrarenal aneurysms (6.5%). Most RAAs were saccular (79%) and noncalcified (63%). The main renal artery bifurcation was the most common site of aneurysms (60%). RAAs were often asymptomatic (55%), with a diagnosis made most often during arteriographic study for suspected renovascular hypertension (42%). RESULTS: Surgery was performed in 121 patients (average RAA size 1.5 cm), including 14 patients undergoing unilateral repair with contralateral RAA observation. The remaining 47 patients (average RAA size 1.3 cm) were not treated surgically. Operations included aneurysmectomy and angioplastic renal artery closure or segmental renal artery reimplantation, aneurysmectomy and renal artery bypass, and planned nephrectomy for unreconstructable renal arteries or advanced parenchymal disease. Eight patients underwent unplanned nephrectomy, being considered a technical failure of surgical therapy. Dialysis-dependent renal failure occurred in one patient. There were no perioperative deaths. Late follow-up (average 91 months) was available in 145 patients (86%). All but two arterial reconstructions remained clinically patent. Secondary renal artery procedures included percutaneous angioplasty, branch embolization, graft thrombectomy, and repeat bypass for late aneurysmal change of a vein conduit. Among 40 patients with clearly documented preoperative and postoperative blood pressure measurements, 60% had a significant decline in blood pressure after surgery while taking fewer antihypertensive medications. Late RAA rupture did not occur in the nonoperative patients, but no lessening of this group's hypertension was noted. CONCLUSION: Surgical therapy of RAAs in properly selected patients provides excellent long-term clinical outcomes and is often associated with decreased blood pressure.

Thromboprophylaxis in an academic medical center.

INTRODUCTION: Questions regarding which patients require prophylaxis for thromboembolism, what methods should be used and the appropriate duration of treatment remain unanswered. METHODS AND MATERIALS: A retrospective review from a single academic medical center was undertaken to evaluate prophylactic strategies. Multiple sources of data were used to identify patients who were prophylaxed and those who developed deep vein thrombosis or pulmonary embolism. These data were analyzed to determine factors associated with successful prophylaxis including age, type of prophylaxis and admitting services. RESULTS: A total of 22,030 patients were admitted of whom 7520 (36%) received prophylaxis and there were 523 thromboembolic events. Pneumatic compression devices and antiembolic stockings had the lowest incidence of failure, 2.2% and 3.2% respectively. There were significant differences in the rates of prophylaxis used by the five admitting services, being highest in surgery (40.8%) and lowest in gynecology (11.4%). However, these groups had the lowest incidence of venous thromboembolism (4.3%, 2.3%). Both of these groups used pneumatic compression as the method of choice (64.3% and 65.2%). CONCLUSIONS: Thromboembolism prophylaxis reduces the incidence of DVT and PE, however, our study demonstrates the variable effectiveness of each method in different types of patients. Our data suggest that patient risk of DVT should be individually assessed and an appropriate method of prophylaxis should be applied when warranted.

Efficacy and durability of autogenous saphenous vein conduits for lower extremity arterial reconstructions in preadolescent children.

OBJECTIVE: Limb length discrepancies (LLDs) in growing children may accompany extremity arterial occlusions. Revascularization with vein grafts has been questioned because of degenerative graft changes observed at other sites. This study was undertaken to define vein graft durability and efficacy in lower extremity revascularizations in preadolescent children. STUDY DESIGN: Fourteen children (10 boys, 4 girls) with a mean age of 7.3 years (range, 2-11 years) who underwent 16 lower extremity revascularizations with greater saphenous vein grafts were subjected to follow-up with graft ultrasonography, ankle/brachial indices (ABIs) with and without exercise, and limb length determinations. A mean of 5.7 years elapsed between the onset of ischemia and operation. Arterial occlusions resulted from cardiac catheterizations (11), arteritis (1), dialysis cannulation (1), and penetrating trauma (1). Indications for operation included LLD (6), claudication (4), both LLD and claudication (3), markedly diminished ABIs with a potential for LLD (2), and a traumatic transection with hemorrhage (1). The reconstructions with 15 reversed and one in situ vein grafts included iliofemoral (11), femorofemoral (1), aortofemoral (1), femoropopliteal (1), popliteal-popliteal (1), and popliteal-posterior tibial (1) arterial bypass grafts. RESULTS: Among patent grafts available for follow-up, 36% (5 of 14) remained unchanged, 50% (7 of 14) developed nonaneurysmal dilatation, and 14% (2 of 14) exhibited nonprogressive aneurysmal expansion. One graft became occluded, and one graft was lost to follow-up. Collectively, the grafts manifest an 11.2% expansion at an average of 10.7 years postoperatively. ABIs increased from 0.75 preoperatively to 0.97, at an average of 11.0 years postoperatively. LLDs were reduced from 1.66 to 1.24 cm, at an average of 11.4 years postoperatively. CONCLUSION: Vein graft reconstructions of lower extremity arteries in preadolescent children are durable. They provide an efficacious means of restoring normal blood flow, and in 70% of children their preexisting LLDs were reduced.

Diagnosis and endovascular treatment of iliocaval compression syndrome.

PURPOSE: The purpose of this study was to evaluate the clinical presentation, diagnosis, and endovascular treatment of iliocaval compression syndrome (ICS). PATIENTS AND METHODS: During a 3-year period, 18 patients (17 women, 1 man; mean age, 42 years) presented with clinical and imaging findings consistent with ICS. All patients were evaluated with venography and Doppler ultrasound (DUS), 13 of 18 with intravascular pressure measurements, 12 of 18 with intravascular ultrasound, 9 of 18 with air plethysmography (APG), and 4 of 18 with magnetic resonance venography. Seventeen patients were treated with endovascular stenting, one was treated with angioplasty alone, and six received adjunct thrombolysis. RESULTS: Despite the presence of stenosis or occlusion in all cases, APG indicated no iliac vein obstruction (outflow fraction > or = 40%) in nine patients. DUS revealed acute (6) or chronic (7) unilateral iliofemoral deep venous thrombosis in 13 of 18 patients, whereas the results of five of 18 DUS studies were normal. Recanalization and stent placement (n = 17) or angioplasty (n = 1) was achieved in all patients. The average pressure gradient was 5.6 mm Hg preprocedure and 0.6 mm Hg postprocedure. The primary patency rate demonstrated with DUS (n = 17) and venography (n = 7) at 6 months was 89%. The primary patency rate at 12 months was 79%. CONCLUSIONS: ICS often presents as sudden unilateral left lower extremity pain and swelling in young to middle-aged female patients after pregnancy, surgery, or a period of inactivity. Venography, intravascular ultrasound, and magnetic resonance venography demonstrate high sensitivity, whereas APG-outflow fraction demonstrates low sensitivity in the diagnosis of ICS. Endovascular stenting and angioplasty provide safe and effective early and intermediate-term treatment of symptomatic ICS.

How DRGs hurt academic health systems.

BACKGROUND: Academic health centers continue their mission of clinical care, education, and research. This mission predisposes them to accept patients regardless of their individual clinical variation and financial risk. The purpose of this study is to assess the variation in costs and the attendant financial risk associated with these patients. In addition, we propose a new reimbursement methodology for academic health center high-end DRGs that better aligns financial risks. STUDY DESIGN: We reviewed clinical and financial data from the University of Michigan data warehouse for FY1999 (n = 39,804). The diagnosis-related groups were classified by volume (group 1, low volume to group 4, high volume). The coefficient of variation for total cost per admission was then calculated for each DRG classification. A regression analysis was also performed to assess how costs in the first 3 days estimated total costs. A hybrid methodology to estimate costs was then determined and its accuracy benchmarked against actual Medicare and Blue Cross reimbursements. RESULTS: Low-volume DRGs (< 75 annual admissions) had the highest coefficient of variation relative to each of the three other DRG classifications (moderate to high volume, groups 2, 3, and 4). The regression analysis accurately estimated costs (within 25% of actual costs) in 64.7% of patients with a length of stay > or = 4 days (n = 16,287). This regression fared well compared with actual FY 1999 DRG-based Medicare and Blue Cross reimbursements (n = 9,085 with length of stay > or = 4 days), which accurately reimbursed the University of Michigan Health System in only 43.9% of cases. CONCLUSIONS: Academic health centers receive a disproportionate number of admissions to low-volume, high-variation DRGs. This clinical variation translates into financial risk. Traditional risk management strategies are difficult to use in health care settings. The application of our proposed reimbursement methodology better distributes risk between payers and providers, and reduces adverse selection and incentive problems ("moral hazard").

Outcomes after surgery for thoracic outlet syndrome.

PURPOSE: This study determined whether there is an association between psychological and socioeconomic characteristics and the long-term outcome of operative treatment for patients with sensory neurogenic thoracic outlet syndrome (N-TOS). METHODS: Clinical records, preoperative psychological testing results, and long-term follow-up questionnaire data were reviewed for consecutive patients who underwent surgery for N-TOS from 1990 to 1999. Multivariate logistic regression models were developed as a means of identifying independent risk factors for postoperative disability. RESULTS: Operative decompression of the brachial plexus via a supraclavicular approach was performed for upper extremity pain and paresthesia with no mortality and minimal morbidity in 170 patients. After an average follow-up period of 47 months, 65% of patients reported improved symptoms, and 64% of patients were satisfied with their operative outcome. However, 35% of patients remained on medication, and 18% of patients were disabled. Preoperative factors associated with persistent disability include major depression (odds ratio [OR], 15.7; P =.02), not being married (OR, 7.9; P =.04), and having less than a high school education (OR, 8.1; P =.09). CONCLUSION: Operative decompression was beneficial for most patients. Psychological and social factors, including depression, marital status, and education, are associated with self-reported disability. The impact of the preoperative treatment of depression on the outcome of TOS decompression should be studied prospectively.

P-selectin antagonism causes dose-dependent venous thrombosis inhibition.

Inhibition of P-selectin by antibody or selectin antagonist decreases inflammation and thrombosis. This study evaluates the dose-response relationship using a selectin receptor antagonist. Eight male baboons (Papio anubis) underwent inferior vena caval thrombosis using a 6 h balloon occlusion model. Three animals received 500 microg/kg P-selectin antagonist (rPSGL-Ig) and five 1 mg/kg rPSGL-Ig with or without a non-anticoagulant dose of Dalteparin. These animals were compared to our published results in this model with 4 saline controls and 8 animals that received 4 mg/kg rPSGL-Ig. A statistically significant dose-response relationship existed between rPSGL-Ig dose and thrombosis (p < 0.01), and between rPSGL-Ig dose and spontaneous recanalization (p<0.05). Inflammatory assessment revealed decreased gadolinium enhancement in all rPSGL-Ig groups compared to previously reported control, despite no significant differences in inflammatory cell extravasation. No dose of rPSGL-Ig caused anticoagulation. Selectin antagonism results in a dose-dependent decrease in thrombosis and increase in spontaneous recanalization.

Recurrent thromboembolism in patients with vena cava filters.

BACKGROUND: Patients with venous thromboembolic disease are treated with anticoagulation or vena cava filter placement to prevent pulmonary embolism. A recent report suggested that filter placement may increase the risk of recurrent deep venous thrombosis (DVT) and prompted a review of our experience. METHODS: Prospectively collected data on 2109 consecutive patients receiving filters were evaluated for recurrent thromboembolism, vena cava occlusion, or venous stasis ulceration. Outcomes were stratified and analyzed according to the use of anticoagulants at the time of insertion and at follow-up. Incidence rates were also compared with reports in the literature. RESULTS: Of 1191 patients with DVT at filter placement, complete follow-up data at a mean of 9 years were available for 465. Recurrent DVT was found in 12% of the 241 patients who were given anticoagulants and 15% of the 224 who were not (P >.05). We also failed to find a significant association between the use of anticoagulation and the incidence of pulmonary embolism (2%), stasis ulceration (2%), and vena cava occlusion (0.0). CONCLUSIONS: Recurrent DVT in patients with existing thromboembolic disease is not an unexpected event, which, in our experience, is not associated with anticoagulant or filter use. Anticoagulation should be used when possible to treat existing DVT to reduce thrombus progression and potentially to reduce subsequent complications but does not seem to reduce the rate of recurrent DVT. Rates of recurrent thromboembolism were consistently less than the 20% to 50% reported in the literature.

Impact of chronic obstructive pulmonary disease on elective and emergency abdominal aortic aneurysm repair.

OBJECTIVE: Chronic obstructive pulmonary disease (COPD) is associated with abdominal aortic aneurysm (AAA) expansion and is considered by some to be a relative contraindication to conventional aortic surgery. This study was undertaken to determine if COPD increases operative death, morbidity, intensive care unit (ICU) length of stay (LOS), and hospital LOS, after AAA repair. METHODS: Data from national administrative records supplemented with laboratory data previously obtained for a system-wide study were analyzed in a retrospective review of 1053 consecutive patients (264 with and 789 without COPD) undergoing operation for intact or ruptured AAAs in Veterans Administration Hospitals from 1997 to 1998. Bivariate comparisons and multivariate regression were used to evaluate the impact of COPD on the number of days of ventilation, ICU LOS, total hospital LOS, and death, while controlling for other known risk factors, including acute myocardial infarction, renal failure, and age. RESULTS: The mortality rate in elective aneurysm patients did not differ (P =.99) between patients with (3.7%) or without COPD (3.7%). However, elective AAA repair was associated with longer hospital LOS (14.4 vs 12.3 days, P =.01), longer ICU LOS (6.5 vs 5.4 days, P =.01), and a higher incidence of requiring 96 hours or more ventilation (6.9% vs 3.6%, P =.02) in patients with COPD. Ruptured AAA affected 4.9% of patients and was strongly associated with COPD (P =.02); however, COPD did not result in a statistically significant increase in death (P =.25). CONCLUSIONS: Although COPD does not appear to increase operative death, it is associated with an increased risk of rupture. Elective repair of AAA should not be deferred in patients with COPD despite their higher LOSs and need for postoperative ventilation.

The percutaneous greenfield filter: outcomes and practice patterns.

OBJECTIVE: The percutaneous steel Greenfield filter (PSGF) is similar in appearance to the titanium Greenfield filter (TGF) but differs in the length and orientation of the attachment hooks and in the over-the-wire delivery system. Because these differences improve ease of insertion and attachment, they may affect patient outcomes and physician practices. The purpose of this study was to evaluate the performance of the PSGF relative to the TGF and to determine whether there had been a change in physician practices. METHODS: The Michigan Filter Registry contains data for a prospective cohort of 2188 patients with Greenfield filters. Procedural and long-term outcomes for patients with a PSGF were abstracted. These events were compared with rates for Registry patients who had a TGF. Trends for indication for placement, delivery route, and filter location were also compared with published series. RESULTS: Since 1995, 600 PSGFs have been placed in 599 patients. A 1-year mortality rate of 42% left 349 patients available for annual follow-up, and studies were completed for 231 (66%). Periprocedural events occurred in 2.5% of cases with associated morbidity in 1.5%. The rate of new pulmonary embolism was 2.6%, and vena caval patency was 98.3%. The combined rate of new venous thromboembolic events was 12.5%. Left-sided femoral vein placements increased to 20%, and the major indication for filter placement has become prophylaxis (46%). CONCLUSIONS: The PSGF is similar to the TGF with respect to patient outcomes, and it provides decreased rates of asymmetry along with excellent fixation. The flexible carrier system has allowed more frequent access through the left femoral vein. The ease of use and favorable patient outcomes have resulted in more frequent placement for prophylactic indications.

Filter complications and their management.

Complications from the use of vena caval filters are rare but can occur at placement, during other procedures, or late in follow-up. Placement problems such as bleeding from the insertion site or embolism from thrombus manipulation are largely avoidable. Technical difficulties with introduction, positioning, or misplacement have been minimized by use of flexible carriers and guidewires. However, later insertion of guidewires without fluoroscopy for line changes can lead to problems of entrapment. Filter capture of emboli can result in further thrombus propagation at that site or in filter obstruction. Management of these events requires adequate visualization to determine whether to use lytic therapy or a second filter. Late complications include filter strut fracture in 0.05% of cases and filter penetration, of even rarer functional significance. Pain that can be attributed to the filter has been found only in a single case of pelvic misplacement. The limited consequences of mechanical changes in the filter argue strongly against unnecessary interventions, which frequently result in an increase in morbidity.

Prophylactic vena caval filters in trauma: the rest of the story.

OBJECTIVE: The purpose of this study was to describe outcomes for patients with trauma who had vena caval filters placed in the absence of venous thromboembolic disease (group P) and compare them with outcomes for patients with trauma who had filters placed after either deep venous thrombosis or pulmonary embolism (group T). DESIGN: The study is a case series of consecutive patients who received vena caval filters after traumatic injury. Data were collected prospectively at the time of filter placement from reports of diagnostic studies obtained for clinical indications and during the annual follow-up examinations. Event rate findings are based on objective tests. Data were obtained from the Michigan Vena Cava Filter Registry. RESULTS: Filters were placed in 385 patients with trauma; 249 of these filters were prophylactic (group P). Event rates were similar in the two groups. New pulmonary embolism was diagnosed in 1.5% of the patients in group P and 2% of the patients in group T. Caval occlusion rates were 3.5% for group P and 2.3% for group T. In all, 15.6% of the patients in group P had deep venous thrombosis or pulmonary embolism after placement. The frequencies of lower extremity swelling and use of support hose were higher in group T than in group P (43% vs 25% and 25% vs 3.5%, respectively; P <.005). Outcomes were comparable in the two groups with respect to mechanical stability of the filter. CONCLUSIONS: The prophylactic indication for vena caval filter placement in patients with trauma is associated with a low incidence of adverse outcomes while providing protection from fatal pulmonary embolism. The current challenge is to limit the number of unnecessary placements through improved methods of risk stratification.

Length of stay has minimal impact on the cost of hospital admission.

BACKGROUND: Hospital cost containment, cost reduction, and alternative care delivery systems continue to preoccupy health care providers, payers, employers, and policy makers throughout the United States. The universal metric for gauging the success of these efforts is hospital length of stay (LOS). Reducing the LOS purportedly yields large cost savings. The purpose of this study is to assess precisely how much hospitals save by shortening LOS. STUDY DESIGN: We reviewed the cost-accounting records of all surviving patients (n = 12,365) discharged from our academic medical center during fiscal year 1998 with LOS of 4 days or more. Actual costs were identified through the University of Michigan cost-accounting system. Individual patient costs were broken out on a daily basis and then decomposed further into variable direct, fixed direct, and indirect categories. The population was analyzed by determining the incremental resource cost of the last full day of stay versus the total cost for the entire stay. The data were also stratified by LOS and by surgical costs. An analysis of all trauma patients was then performed on all patients discharged from the hospital's adult level I trauma center (n = 665). Costs were determined on specific days, including admission day, each ICU day, day of discharge from the ICU, and each of the last 2 days before the discharge day. RESULTS: The incremental costs incurred by patients on their last full day of hospital stay were $420 per day on average, or just 2.4% of the $17,734 mean total cost of stay for all 12,365 patients. Mean end-of-stay costs represented only a slightly higher percentage of total costs when LOS was short (e.g., 6.8% for patients with LOS of 4 days). Even when the data were stratified to focus on patients without major operations, the $432 average last-day variable direct cost was only 3.4% of the $12,631 average total cost of care. A focus on the trauma center helps to explain this phenomenon. For our trauma center, variable direct costs accounted for 42% of the mean total cost per patient of $22,067. The remaining 58% was hospital overhead (fixed and indirect costs). The median variable direct cost on the first day of admission is $1,246, and the median variable direct cost on discharge is $304. Approximately 40% of the variable costs are incurred during the first 3 days of admission. CONCLUSIONS: For most patients, the costs directly attributable to the last day of a hospital stay are an economically insignificant component of total costs. Reducing LOS by as much as 1 full day reduces the total cost of care on average by 3% or less. Going forward, physicians and administrators must deemphasize LOS and focus instead on process changes that better use capacity and alter care delivery during the early stages of admission, when resource consumption is most intense.

Academic health systems management: the rationale behind capitated contracts.

OBJECTIVE: To determine why hospitals enter into "capitated" contracts, which often generate accounting losses. The authors' hypothesis is that hospitals coordinate contracts to keep beds full and that in principal, capitated contracts reflect sound capacity management. SUMMARY BACKGROUND DATA: In high-overhead industries, different consumers pay different prices for similar services (e.g., full-fare vs. advanced-purchase plane tickets, full tuition vs. financial aid). Some consumers gain access by paying less than total cost. Hospitals, like other high-overhead business enterprises, must optimize the use of their capacity, amortizing overhead over as many patients as possible. This necessity for enhanced throughput forces hospitals and health systems to discount empty beds, sometimes to the point where they incur accounting losses serving some payors. METHODS: The authors analyzed the cost accounting system at their university teaching hospital to compare hospital and intensive care unit (ICU) lengths of stay (LOS), variable direct costs (VDC), overhead of capitated patients, and reimbursement versus other payors for all hospital discharges (n = 29,036) in fiscal year 1998. The data were analyzed by diagnosis-related groups (DRGs), length of stay (LOS), insurance carrier, proximity to hospital, and discharge disposition. Patients were then distinguished across payor categories based on their resource utilization, proximity to the hospital, DRG, LOS, and discharge status. RESULTS: The mean cost for capitated patients was $4,887, less than half of the mean cost of $10,394 for the entire hospitalized population. The mean capitated reimbursement was $928/day, exceeding the mean daily VDC of $616 but not the total cost of $1,445/day. Moreover, the mean total cost per patient day of treating a capitated patient was $400 less than the mean total cost per day for noncapitated patients. The hospital's capitated health maintenance organization (HMO) patients made up 16. 0% of the total admissions but only 9.4% of the total patient days. Both the mean LOS of 3.4 days and the mean ICU LOS of 0.3 days were significantly different from the overall values of 5.8 days and 1 day, respectively, for the noncapitated population. For patients classified with a DRG with complication who traveled from more than 60 miles away, the mean LOS was 10.7 days and the mean total cost was $21,658. This is in contrast to all patients who traveled greater than 60 miles, who had an LOS of 7.2 days and a mean total cost of $12,569. CONCLUSION: The capitated payor directed the bulk of its subscribers to one hospital (other payors transferred their sicker patients). This was reflected in the capitated group's lower costs and LOS. This stable stream of relatively low-acuity patients enhanced capacity utilization. For capitated patients, the hospital still benefits by recovering the incremental cost (VDC) of treating these patients, and only a portion of the assigned overhead. Thus, in the short run, capitated patients provide a positive economic benefit. Other payors' higher-acuity patients arrive more randomly, place greater strains on capacity, and generate higher overhead costs. This results in differential reimbursement to cover this incremental overhead. Having a portfolio of contracts allows the hospital to optimize capacity both in terms of patient flows and acuity. One risk of operating near capacity is that capitated patients could displace other higher-paying patients.