Society of University Surgeons. Presidential address: discovery.
Explore the source record for details and available documents.
General Surgery↗
Biomedical subjects
D L Dunn
Publications and source records attributed to D L Dunn.
Design of a potent novel endotoxin antagonist.
BACKGROUND: Bactericidal permeability increasing protein (BPI) binds to and neutralizes lipopolysaccharide (LPS, endotoxin). Small synthetic peptides based on the amino acid sequence of the LPS binding domain of BPI neutralize LPS, albeit inefficiently. Although the LPS binding domain of native BPI possesses a beta-turn secondary structure, this structure is not present in small derivative peptides. The purpose of this study was to determine whether the addition of a beta-turn to a BPI-derived peptide is associated with more potent endotoxin antagonism. METHODS: We generated a hybrid peptide (BU3) on the basis of (1) a portion of the LPS binding domain from BPI and (2) amino acids known to initiate a beta-turn. BU3 folds with a beta-turn, and we tested its effects on LPS neutralization and LPS-induced tumor necrosis factor-alpha secretion, comparing it with BPI-derived peptide BG22 that lacks a beta-turn and to an irrelevant peptide (BG16). RESULTS: Compared with BG22, BU3 demonstrated enhanced LPS neutralization and inhibition of LPS-induced tumor necrosis factor-alpha secretion in vitro and a similar diminution of endotoxemia and tumor necrosis factor-alpha secretion in a murine model of endotoxemia. CONCLUSIONS: These data demonstrate the potential for enhancing the biologic activity of a BPI-derived peptide endotoxin antagonist via manipulation of its conformational structure.
Amino Acid Sequence↗
Duodenal complications in bladder-drained pancreas transplantation.
BACKGROUND: The most common type of pancreas transplantation is whole pancreaticoduodenal (with bladder drainage) from a cadaver donor. Complications can arise not only from the pancreas itself but also from the simultaneously transplanted duodenum. The purpose of this study was to analyze the incidence, diagnosis, and treatment of duodenal complications and their impact on patient and pancreas graft survival rates. METHODS: Our retrospective study is based on 425 pancreaticoduodenal transplantations performed between July 1, 1986, and June 30, 1994. Complications pertaining to the duodenal segment were labeled early if they occurred within the first postoperative month and late otherwise. Mean follow-up was 55 months (range, 13 to 108 months). RESULTS: We noted 85 (20%) duodenal complications: duodenal leaks (n = 42), hematuria (n = 26), recurrent urinary tract infections (n = 9), duodenal ulceration or necrosis (n = 6), and bladder stones (n = 2). Of these complications, 40 (48%) required surgical intervention. In all, duodenal complications resulted in 14 (16%) enteric conversions and eight (9%) pancreas graft losses (six because of duodenal leak and 2 because of hematuria). The mortality rate from duodenal complications was 0%. CONCLUSIONS: Duodenal complications were common, but they were not associated with a high rate of pancreas graft loss (only 9%). With early diagnosis and treatment, morbidity can be reduced and death avoided in pancreas transplant recipients.
Duodenal Diseases↗
The surgical risk of pancreas transplantation in the cyclosporine era: an overview.
BACKGROUND: Pancreas transplants are still associated with the highest surgical complication rate of all routinely performed solid organ transplants. To date, the impact of serious surgical complications in the cyclosporine era on perioperative patient morbidity, graft and patient survival, and hospital costs has not been analyzed in detail. STUDY DESIGN: We retrospectively studied surgical complications after 445 consecutive pancreas transplants (45% simultaneous pancreas-kidney [SPK], 24% pancreas after kidney [PAK], and 31% pancreas transplant alone [PTA]). Of these, 80% were primary transplants, 20% were retransplants. Cadaver donors were used in 92%, living related donors in 8%. To develop guidelines for their prevention and management, we studied the impact of significant surgical complications (intra-abdominal infections, vascular graft thrombosis, and anastomotic leak) requiring relaparotomy on graft and patient survival. RESULTS: Relaparotomy was required after 32% of all pancreas transplants (SPK: 36%, PAK: 25%, PTA: 16% [p = 0.04]). Perioperative mortality was 9%. Graft and patient survival rates were significantly lower for recipients with (versus without) relaparotomy. The most common procedures were drainage of intra-abdominal abscess with graft necrosectomy (50% of all relaparotomies) and transplant pancreatectomy (34%). The most common causes of relaparotomy were intra-abdominal infection, vascular graft thrombosis, and anastomotic leak. Intra-abdominal infection occurred in 20% (SPK: 18%, PAK: 24%, PTA: 20% [p = NS]). The rate was significantly higher for living related donor (42%) versus cadaver donor (18%) recipients and for those with enteric-drained (39%) versus bladder-drained (18%) transplants. Graft and patient survival rates were significantly lower for recipients with (versus without) intra-abdominal infection. Outcome was better after bacterial (versus fungal) infections. For SPK recipients, those not on dialysis before the transplant had significantly higher graft survival than those on dialysis. Vascular graft thrombosis occurred in 12% of all recipients. The rate was significantly higher for PAK (21%) than for PTA (10%) and SPK (9%) recipients. It was significantly lower for recipients of grafts with donor iliac Y-graft reconstruction (versus all other types of arterial reconstruction) and with right-sided (versus left-sided) graft placement. Of note, patient survival was not different for recipients with versus without vascular graft thrombosis. The incidence of anastomotic or duodenal stump leaks was 10%; of these recipients, 70% required relaparotomy. Patient and graft survival rates were no different for recipients with versus without leaks. CONCLUSIONS: Serious surgical complications occurred in 35% of pancreas recipients and had a significant impact on patient and graft survival. Based on multivariate risk factor analyses, we recommend the following: donors over 45 years and those dying of cerebrocardiovascular disease should not be used; recipients over 45 years and those with a history of cardiac disease should be considered for a kidney transplant alone (KTA); surgical technique for graft procurement, preparation, and implantation should be meticulous; right-sided implantation and arterial Y-graft reconstruction should be performed when possible, since they had the highest success rates; when complications require relaparotomy, the focus must switch from graft salvage to life preservation; and the threshold for pancreatectomy should be low. Diagnosis should be timely, and treatment and relaparotomy expeditious. These cornerstones of success should help decrease the risk of surgical complications and mortality after pancreas transplants.
Abdominal Abscess↗
Risk factors for prolonged hospitalization after kidney transplants.
A major variable in the cost of kidney transplants is the length of initial hospitalization. Using multivariate analysis, we studied risk factors for hospital stay > 10 d post-transplant. Between 1 January 1985 and 31 August 1995 a total of 1588 patients underwent first or second kidney transplants at the University of Minnesota. Antibody was used for 1 wk in cadaver donor recipients and for 2 wk in pediatric recipients (resulting in a long stay for all pediatric recipients). Adult living related donor recipients were immunosuppressed with triple therapy. Donor risk factors studied were age (< 15, 15-50, > 50 yr) and,- for cadaver recipients, preservation time (< 12, 12-18, 18-24, 24-30, > 30 h) and cause of death (trauma, cerebrovascular accident, or cardiac). Recipient risk factors studied were age (< 18, 18-55, > 55 yr); sex; transplant number; antigen mismatch; peak PRA; PRA at transplant (< 11, 11-50, > 50); diabetic status; pretransplant dialysis (vs. pre-emptive transplant); pretransplant cardiac, peripheral vascular, or respiratory disease; and delayed graft function (DGF) (dialysis in the first week vs. no dialysis). Risk factors were analyzed separately for living donor and cadaver donor recipients. For cadaver donor recipients, DGF was the major risk factor (p < 0.0001); others were age 55 yr (p = 0.03) and diabetes (p = 0.02). For living donor recipients, DGF was also a risk factor (p = 0.003); others were diabetes (p = 0.01), retransplant (p = 0.006), PRA at transplant > 50 (p < 0.0001), age > 55 yr (p = 0.02), pretransplant respiratory disease (p = 0.005), and pretransplant dialysis (p = 0.005). Because DGF was the major risk factor for a prolonged stay, we then studied risk factors for DGF using multivariate analysis. For cadaver donor recipients, risk factors were recipient weight > 90 kg (p = 0.004), preservation time 24 h (p = 0.03), PRA at transplant > 50 (p = 0.03), and donor age < 15 or > 50 yr (p = 0.002). For living donor recipients, risk factors were recipient age < 18 yr (p = 0.01), donor age > 50 yr (p = 0.03), female sex (p = 0.05), pretransplant respiratory disease (p = 0.1), pretransplant peripheral vascular disease (p = 0.05), and recipient weight > 90 kg (p = 0.1). From our data, a profile emerged of recipients likely to have a longer hospital stay. Important variables, either simultaneous with or related to DGF, include donor and recipient age, diabetes, pretransplant recipient weight, PRA at transplant, preservation time, and pretransplant respiratory or peripheral vascular disease.
Adolescent↗
Pregnancy after donor nephrectomy.
Potential female donors frequently ask whether unilateral nephrectomy will impair future childbearing capabilities. To address this question, we surveyed 220 women who underwent donor nephrectomy between 1985 and 1992. Of the 144 women who responded, 33 became pregnant after donation for a total of 45 pregnancies. Seventy-five percent of the pregnancies were carried to term without difficulty. Complications incurred during gestation included miscarriage (13.3%), preeclampsia (4.4%), gestational hypertension (4.4%), proteinuria (4.4%), and tubal pregnancy (2.2%). Four of the 45 pregnancies (excluding miscarriages) required preterm hospitalization, resulting in an overall morbidity of 8.8%. There were no pregnancy-related deaths, and no fetal abnormalities were reported. Problems with persistent hypertension, proteinuria, or changes in renal function were not noted. None of the above complications exceeded what has been noted for the general population. Infertility was a problem in 8.3% (3/36) of our respondents, compared with a worldwide incidence of 16.7%. Based on these results, we conclude that donor nephrectomy is not detrimental to the prenatal course or outcome of future pregnancies.
Female↗
Renal transplantation exposes patients with previous Kaposi's sarcoma to a high risk of recurrence.
It is currently estimated that about 0.5% of patients will develop Kaposi's sarcoma (KS) after kidney transplantation. Tapering of immunosuppression often leads to KS remission, but also results in graft loss in more than 50% of cases. Whether retransplantation is safe in these patients is unknown. We here report on eight patients who developed KS recurrence after kidney transplantation-(A) Patients with previously treated KS: There were 4 patients who had clinical remission of KS (including three posttransplantation) for periods ranging from 5 months up to 19 years before transplantation. All 4 developed KS recurrence within months after transplantation. In 3 patients, KS regressed only when all immunosuppression was discontinued, at the price of allograft removal. Partial remission occurred in the fourth patient following reduction of immunosuppression and gancyclovir administration; (B) Patients with recurrent KS during a single transplant: 4 patients developed KS after transplantation that regressed following reduction of immunosuppressive therapy. Increased immunosuppression, in the form of steroid pulses in 3 patients was associated with recurrence of KS. Subsequent reduction of immunosuppression caused regression of KS in all 4 patients, but 2 recipients lost their allografts. These data emphasize the high risk of recurrence of KS after renal transplantation. If physicians decide to transplant patients with a history of KS, they should inform the future recipient of the possibility of KS recurrence.
Adult↗
Delayed graft function in the absence of rejection has no long-term impact. A study of cadaver kidney recipients with good graft function at 1 year after transplantation.
We previously reported that delayed graft function (DGF) in the absence of biopsy-proven acute rejection (Rej) had no effect on outcome of primary cadaver kidney transplantation (TX). By contrast, DGF in combination with Rej strongly predicted poor long-term graft survival. We asked whether this poor long-term outcome was due to early graft loss associated with DGF, or to an ongoing process leading to late graft loss. To answer this question, we studied a subset of 298 cadaver kidney transplant recipients who had not suffered early graft loss and had a serum creatinine level < or = 2.0 mg/dl at 1 year after TX. The incidence of DGF (defined by dialysis during the first week after TX) in this subset was 19%. DGF was associated with cold ischemia time >24 hr (P = 0.0003) and Rej (P = 0.06). For grafts with versus without DGF, the incidence of late acute Rej (>1 year after TX) was similar. Actuarial graft survival was similar for Rej-free recipients with versus without DGF (P = 0.9) and was worse for those with Rej and no DGF (P < 0.02). Importantly, however, in our recipients who all had a serum creatinine level < or = 2.0 mg/dl at 1 year after TX, the worst long-term outcome was noted in the subgroup with both DGF and Rej (P < 0.0001). By multivariate analysis, DGF was also only a risk factor in combination with Rej (P = 0.002, relative risk = 3.7), while a 0-antigen HLA mismatch had no impact. Patient survival decreased for recipients with both DGF and Rej by univariate (P = 0.009) and multivariate (P = 0.02, relative risk = 2.9) analyses. We conclude that DGF without Rej has no impact on long-term survival. However, our data for recipients with both DGF and Rej suggest that a chronic ongoing process leads to late graft failure. Further research is necessary to identify the exact pathophysiology of this process, which appears to be, at least in part, HLA antigen independent.
Acute Disease↗
Employment patterns after successful kidney transplantation.
We studied 822 kidney transplant recipients followed 1-9 years to determine the dynamics of their entering and leaving the work force. Multivariate analysis revealed that not being diabetic and that being employed pretransplant were associated with a higher rate of posttransplant employment. Some recipients in all pretransplant employment categories, including those receiving disability benefits pretransplant, returned to full-time work posttransplant. The most rapid return to work was in those who had been working full-time or attending school pretransplant. After returning to work, a higher percentage of diabetic recipients stopped working; of those who stopped working, 50% received disability benefits. In contrast, nondiabetic recipients who stopped working full-time were more likely to be retired or working part-time; only 22% received disability benefits.
Adult↗
Bacterial translocation in a large-animal model of small-bowel transplantation. Portal vs systemic venous drainage and the effect of tacrolimus immunosuppression.
OBJECTIVE: To study whether bacterial translocation is more prevalent after small-bowel transplantation with systemic venous drainage (SVD) vs portal venous drainage (PVD) and whether it is influenced by immunosuppression. DESIGN: We performed 15 small-bowel transplantations in pigs. Group 1 (n = 5) had SVD and no immunosuppression; group 2 (n = 6), PVD and no immunosuppression; and group 3 (n = 4), PVD and immunosuppression with tacrolimus and methylprednisolone sodium succinate. Portal and systemic blood, portal and mesenteric lymph nodes, and liver were cultured in donors and recipients on postoperative day 0 (POD 0) and in recipients on postoperative day 3 (POD 3). Jejunal and ileal contents were also sampled at these times. SUBJECTS: Outbred male Yorkshire-Landrace pigs. MAIN OUTCOME MEASURES: (1) Blood and tissue bacterial cultures, (2) blood endotoxin levels, and (3) histopathologic examination. RESULTS: Cultures were positive for bacteria in 32% (16/50) of samples on POD 0 and 88% (22/25) on POD 3 in group 1, in 18% (11/60) of samples on POD 0 and 97% (29/30) on POD 3 in group 2, and in 8% (3/40) of samples on POD 0 and 95% (19/20) on POD 3 in group 3. Systemic blood cultures were positive for bacteria on POD 3 in 60% (3/5) of pigs in group 1, 83% (5/6) in group 2, and 100% (4/4) in group 3. Significantly more bacteria were present in the ileum than in the jejunum on POD 0 in group 2; this difference approached significance in groups 1 and 3. Bacterial numbers were identical in the ileum and jejunum by POD 3 in all groups. Circulating endotoxin levels were significantly elevated on POD 3 vs POD 0 only in group 1. Endotoxin levels were not significantly different between the SVD group (group 1) and either PVD group (groups 2 and 3). CONCLUSIONS: Bacterial translocation is prevalent after small-bowel transplantation in pigs whether PVD or SVD is used. Immunosuppression with tacrolimus does not prevent bacterial translocation but may reduce systemic endotoxemia.
Animals↗
Pancreas graft loss caused by intra-abdominal infection. A risk factor for a subsequent pancreas retransplantation.
OBJECTIVE: To investigate whether previous pancreas graft loss caused by intra-abdominal infection is a risk factor for a subsequent pancreas retransplantation. DESIGN: Retrospective case-series analysis. SETTING: Large university hospital. PATIENTS: Of 97 pancreatic retransplantations (July 1, 1985 to June 30, 1994), 13 (13%) were performed after previous pancreas grafts had been lost because of intra-abdominal infection. MAIN OUTCOME MEASURES: Cause of retransplant graft loss; patient survival. RESULTS: Of 13 cases of pancreatic retransplantations performed after previous grafts had been lost because of intra-abdominal infection, 12 (92%) were again complicated by intra-abdominal infection. Of these 12 retransplantation infections, 10 were caused by the same microbial species that had caused failure of the previous graft 1 to 5 years earlier (8 required graft pancreatectomy [mortality rate, 25%], and 2 were successfully treated) and 2 were caused by different microbial species (both required graft pancreatectomy [mortality rate, 0%]). In 1 of the 13 retransplantations, prophylactic antimicrobial treatment was directed at the microbial species that had caused failure of the previous graft; no recurrent intra-abdominal infection developed. CONCLUSIONS: Intra-abdominal infection after a previous pancreas transplantation is a risk factor for recurrence of infection by the same microbial species after a subsequent retransplantation. For the selected patients who are considered for retransplantation in spite of their previous graft loss caused by intra-abdominal infection, periretransplantation antimicrobial prophylaxis should include a prolonged course of an agent directed against the previously identified microbial species.
Abdomen↗
Macrophages expressing a fusion protein derived from bactericidal/permeability-increasing protein and IgG are resistant to endotoxin.
OBJECTIVES: To generate a recombinant fusion protein (FP) based on the endotoxin-binding domain of bactericidal/permeability-increasing protein (BPI) and the constant domain of IgG and to test its ability to inhibit lipopolysaccharide (LPS)-induced macrophage tumor necrosis factor alpha (TNF-alpha) secretion. DESIGN: A murine macrophage cell line, RAW 264.7, was transfected with a BPI-IgG FP before incubation with LPS. The amount of LPS-induced TNF-alpha protein secreted was measured and compared with that secreted by cells transfected with a control construct. SETTING: Basic science research laboratory. MAIN OUTCOME MEASURES: Secreted TNF-alpha protein concentration. RESULTS: After transfection, RAW 264.7-cell FP expression was detected in cell lysates and supernatants. At each LPS dose tested, cells transfected with the FP gene secreted less TNF-alpha than did cells transfected with a control construct. CONCLUSIONS: The FP possesses substantial antiendotoxin activity, as delineated by inhibition of LPS-induced TNF-alpha secretion by murine macrophages transfected with the fusion gene construct. In the future, such FP may be used as a clinical reagent to reduce the morbidity and mortality associated with serious gram-negative bacterial infections in surgical patients.
Animals↗
Differential sensitivity to Escherichia coli infection in mice lacking tumor necrosis factor p55 or interleukin-1 p80 receptors.
OBJECTIVE: To determine the effect of targeted disruption of the cellular receptors of either tumor necrosis factor alpha (TNF-alpha) or interleukin-1 beta (IL-1 beta) during experimental gram-negative bacterial infection and endotoxemia. DESIGN: Transgenic (knockout [KO]) mice deficient in either the p55 TNF receptor (TNF RI) or the p80 IL-1 receptor (IL-1 RI) were challenged with intravenous lipopolysaccharide (endotoxin) or intraperitoneal live Escherichia coli 0111:B4. Mortality was assessed daily for 7 days. Serum endotoxin levels and quantitative blood cultures were monitored at multiple times during infection. SETTING: Surgical infectious disease research laboratory. MAIN OUTCOME MEASURES: Mortality, results of quantitative blood cultures, and serum endotoxin levels. RESULTS: Both TNF and IL-1 RI KO mice were resistant to endotoxin challenge (0% mortality for both groups) compared with control mice (100% mortality [P < .01]). In contrast, only the IL-1 RI KO mice were resistant to infection caused by viable gram-negative bacteria (43% mortality) compared with control mice (100% mortality [P < .01]). Infection led to 100% mortality in TNF RI KO mice. The IL-1 RI KO mice exhibited less bacteremia and diminished endotoxemia compared with control and TNF RI KO mice 18 and 24 hours after infection. CONCLUSION: The absence of either the TNF or the IL-1 RI receptor prevents cellular activation by each respective cytokine. Absence confers protection against intravenous endotoxin, which stimulates massive rapid release of cytokines into the systemic circulation. However, bacterial infection within the peritoneal cavity is known to cause more delayed cytokine release, and cytokines may act at the site of infection to enhance host defenses. We believe that IL-1 signaling may be more critical in provoking lethal systemic toxic effects than TNF signaling. However, TNF signaling may be an important component of host defense enhancement at the local site of infection.
Animals↗
A novel endotoxin antagonist attenuates tumor necrosis factor-alpha secretion.
Twenty-seven amino acid peptides with sequences corresponding to a proposed endotoxin binding region of bactericidal permeability increasing protein (BPI):1) inhibit lipopolysaccharide induced macrophage tumor necrosis factor-alpha (TNF-alpha) secretion, 2) have bactericidal activity against gram-negative bacteria, and 3) protect mice from a lethal lipopolysaccharide (LPS) challenge. Unfortunately, peptides have a short halflife in vivo. Therefore, we have chemically conjugated the BPI based peptide, BG38, to a larger carrier protein, keyhole limpet hemocyanin (KLH), and characterized its ability: 1) to inhibit LPS induced macrophage TNF-alpha secretion and 2) to decrease plasma endotoxin and TNF-alpha levels following an i.v. injection of E. coli 0111:B4 LPS. BG38-KLH inhibited cultured macrophage TNF-alpha secretion in response to LPS derived from four pathogenic strains of gram-negative bacteria in a dose dependent manner (>90% inhibition at 50 microgram/ml, P < 0.05 Student's t test). BG38-KLH also decreased serum endotoxin (>90%, P < 0.05 Student's t test) and peak TNF-alpha levels (>30% inhibition, P < 0.05 Student's t test) following E. coli LPS challenge in a murine gram-negative bacterial sepsis model. Novel endotoxin antagonists based upon a small domain of BPI represent promising reagents for the treatment of serious gram-negative bacterial infections.
Animals↗
Mycoplasma hominis infections occurring in cardiovascular surgical patients.
BACKGROUND: Postoperative Mycoplasma hominis sternal would or mediastinal infections are uncommon and difficult to diagnose. Atypical growth characteristics in routine bacterial culture, and the inability to demonstrate the organism on Gram stain, lead to delayed diagnosis of M hominis infections and increased morbidity. METHODS: Postoperative purulent would drainage or acute mediastinitis caused by M hominis developed in 3 cardiovascular surgery patients. These patients were considered along with 9 patients previously reported in the literature. RESULTS: Operative findings included moderately thick, gray purulent fluid with the degree of tissue necrosis related to duration of infection. Intraoperative Gram stain of wound or mediastinal drainage demonstrated no microorganisms, and initial bacterial cultures did not reveal microbial growth. After an average of 4.5 days of culture, minute translucent colonies of M hominis were identified. The institution of appropriate antimycoplasma therapy (doxycycline and clindamycin) was associated with clinical or microbiological cure in all patients. Sternal wound complications developed in 3 patients, and a chronic infection developed in 1 patient. CONCLUSIONS: Empiric therapy for M hominis infection should be considered in patients with mediastinitis or a sternal wound infection in which organisms are not observed on Gram stain and are not readily cultured.
Acute Disease↗
Diagnosis and treatment of bacteremia and intravascular catheter infections.
The past decade has witnessed an explosive rise in the rate of bacteremia and intravascular catheter infection. Although gram-negative organisms continue to account for up to one third of these infections, gram-positive organisms have become increasingly prevalent pathogens. Virulent antibiotic-resistant bacterial strains have emerged and present a formidable treatment challenge. Simultaneously, management of catheter infection has evolved. Although patients who develop fungemia, gram-negative bacteremia, or sepsis syndrome are best treated by catheter removal in addition to antimicrobial therapy, an increasing body of evidence suggests that many gram-positive bacterial catheter infections can be treated by use of antimicrobial agents without catheter removal. Advances in catheter design and immunotherapy for sepsis syndrome also hold promise. Despite these innovations, determining the initial need for catheter placement, adherence to meticulous sterile surgical technique during insertion, and subsequent fastidious catheter maintenance remain the mainstays of preventing these potentially disastrous infections.
Anti-Bacterial Agents↗
The measurement of physician work and alternative uses of the resource based relative value scale.
On January 1, 1992, the Medicare program implemented a new payment system for physician services based on the Resource Based Relative Value Scale (RBRVS). The RBRVS has been widely accepted as a rational and systematic approach to measuring the resource costs associated with physician services. In addition to deriving physician payment rates, the RBRVS provides a useful metric that allows the measurement and comparison of provider utilization rates and productivity across physicians performing a varied mix of services. In this study we describe the measurement of physician work, discuss alternative ways in which work values can be used to monitor physician service utilization (e.g., profiling physician practice patterns), measure physician productivity, and determine physician compensation.
Efficiency↗