Free
Education  |   February 2007
Influence of Patient and Hospital Characteristics on Anesthesia Time in Medicare Patients Undergoing General and Orthopedic Surgery
Author Affiliations & Notes
  • Jeffrey H. Silber, M.D., Ph.D.
    *
  • Paul R. Rosenbaum, Ph.D.
  • Xuemei Zhang, M.S.
  • Orit Even-Shoshan, M.S.
    §
  • * Professor, The Center for Outcomes Research, The Children’s Hospital of Philadelphia; Department of Anesthesiology and Critical Care and Department of Pediatrics, The University of Pennsylvania School of Medicine; and The Leonard Davis Institute of Health Original Investigations, Philadelphia, Pennsylvania. † Professor, Department of Statistics, The Wharton School, The University of Pennsylvania. ‡ Senior Statistical Programmer, The Children’s Hospital of Philadelphia. § Associate Center Director, The Center for Outcomes Research, The Children’s Hospital of Philadelphia, and The Leonard Davis Institute of Health Original Investigations.
Article Information
Education / Practice Management
Education   |   February 2007
Influence of Patient and Hospital Characteristics on Anesthesia Time in Medicare Patients Undergoing General and Orthopedic Surgery
Anesthesiology 2 2007, Vol.106, 356-364. doi:
Anesthesiology 2 2007, Vol.106, 356-364. doi:
PROCEDURE time is a common variable reported in the anesthesia and surgical literature.1–10 However, because the study of procedure time generally requires chart abstraction, the variable is usually reported from single institutions where such information is more easily abstracted. Utilizing an algorithm to estimate anesthesia procedure time from the Medicare anesthesia bill, we report the influence of specific principal procedure, patient characteristics, and hospital characteristics on procedure time. Use of anesthesia bills from Medicare allows for large samples of cases to be studied, among many different hospital settings; in so doing, one can better estimate patient and hospital effects that may be lost, or poorly estimated, in a smaller single-institution study.
Materials and Methods
Patients and Databases
Medicare data for patients 65 yr and older comprise the most representative sample of healthcare data in the United States because Medicare is an entitlement. The only significant group of elderly citizens not represented in the Medicare claims data is those who opt out of the Medicare fee-for-service program and join a Medicare-approved prepaid health maintenance organization. As part of the Surgical Outcomes Study,11–13 we obtained the Medicare Inpatient (Part A), Outpatient Standard Analytic Files, and Physician Part B files for all admissions in general and orthopedic surgical Diagnosis Related Groups in Pennsylvania during 1995 and 1996 for patients between the ages of 65 and 85 yr. These files represent the fee-for-service Medicare population, comprising approximately 90% of all beneficiaries for 1995-1996.14,15 We created a longitudinal record by including all inpatient and outpatient claims and physicians’ claims during that time interval for each patient. Data also included the Medicare Vital Status File, American Hospital Association Annual Surveys for 1996, and the Pennsylvania Health Care Cost Containment Council Hospital Discharge Database for similar years, which included the MedisGroups (MediQual, Inc., Marlborough, MA) severity score to supplement the Medicare record.16 Medicare electronically stores all bills submitted by hospitals and physicians and all other caregivers, and stores all payments made. These billing data include limited diagnosis codes collected using International Classification of Diseases (ICD9-CM) coding and other procedure specific codes.
When anesthesiologists and anesthetists bill Medicare for services performed, these bills include time units. Before January 1, 1994,1-2anesthesia time was reported in units of 15-min intervals. Thereafter, actual anesthesia time in minutes is reported on the claims in Medicare Physician Part B, the file that includes physicians’ claims. By noting the time units in the physicians’ Medicare claims and by abstracting the medical records that generated them, one can develop a model to estimate the number of chart minutes for each procedure based on the anesthesia bill.
The claims data set consisted of 102,781 Medicare beneficiaries collected as part of the Surgical Outcomes Study. Patients were aged 65-85 yr and had undergone general surgical or orthopedic procedures. If more than one surgical admission was recorded for a patient, we chose the first surgical admission of that patient, so that each patient was represented only once in the data set. We then identified the 20 most common procedures in general surgery and the 20 most common orthopedic procedures. We excluded patients that did not have Part B anesthesia bills or who had anesthesia bills less than or equal to 10 min or anesthesia bills greater than 600 min. We also excluded patients from small hospitals, which in our data set had fewer than 50 patients per hospital. The final data set contained 77,638 patients.
Estimating Anesthesia Time from the Medicare Bill
We used the Anesthesia Time Units1as a measure of anesthesia procedure time for each patient in the sample. Data were derived from Medicare Part B claims obtained from the Centers for Medicare and Medicaid Services. We defined “anesthesia chart time” as time of recovery room entrance minus time of induction and expressed this in minutes.17 In previous work, we utilized chart abstraction on 1,931 patients to test the validity of the information from the Medicare bill and the abstracted anesthesia chart time. This work has been published elsewhere in Anesthesiology,17 and our algorithm for relating the billed time denoted in the Part B claim requires the selection of the anesthesia provider who billed the most minutes on the day that the principal procedure was performed. Using such a method, we achieved a Kendal’s τ of 0.85, or produced a probability of concordance of 0.93 between the length of the longest anesthesia bill from Medicare and the anesthesia time recorded in the chart.17 
Statistical Methods
We report the anesthesia time in minutes for each of the 20 general surgical and orthopedic procedures, based on the anesthesia bill (claim) using the methods described above. We also performed an accompanying analysis that included only patients with one surgical procedure noted on the day of the principal procedure. An indicator variable describing whether there were single or multiple procedures on the day of the principal procedure was also included in all regression modeling. When performing multiple regression, we used Huber’s robust m-estimation as implemented in SAS Version 9 (SAS Institute, Inc., Cary, NC) using Huber weights.18–20 For these regressions, we report as R2the square of the Spearman rank correlation between the observed and fitted dependent variables. We defined income levels based on patient ZIP code. Utilizing the median household income at the ZIP code level in the 2000 Census and the patients’ ZIP codes from our Medicare claims file, we appended to each patient’s record in our analysis file an income category based on the following definitions that divided our sample into three equal parts: Lower-Income (< $31,388), Middle-Income ($31,388–$41,858), and Higher Income (>$41,858). MedisGroups Severity Score on admission was used as a physiologic severity adjustment. The score was abstracted on admission for all patients in the data set.16 We used MedisGroups score 0 as the reference category and compared this to scores 1, 2, and grouped 3 + 4 (due to low numbers in the two highest severity scores).
Results
Length of Anesthesia Time in General Surgical and Orthopedic Procedures
Table 1displays the distribution of anesthesia times in the 20 general surgical procedures, and table 2displays the distribution of times for orthopedic procedures. For each procedure, we report the 25th, 50th, and 75th percentiles for all cases, and we also report these same percentiles for cases in which no other procedure was billed except the principal procedure for the date the principal procedure was performed. These are unadjusted times, so that variations across hospitals and patient characteristics may account for some of the dispersion in anesthesia minutes. As can be observed, the median (50th percentile) time varied greatly among procedures. Furthermore, multiple procedures had some influence on the median lengths noted for each procedure.
Table 1. Estimated Anesthesia Time for the 20 Most Frequent Surgical Procedures in the Surgical Outcomes Study 
Image not available
Table 1. Estimated Anesthesia Time for the 20 Most Frequent Surgical Procedures in the Surgical Outcomes Study 
×
Table 2. Estimated Anesthesia Time for the 20 Most Frequent Orthopedic Procedures in the Surgical Outcomes Study 
Image not available
Table 2. Estimated Anesthesia Time for the 20 Most Frequent Orthopedic Procedures in the Surgical Outcomes Study 
×
Estimating the Influence of Patient Factors: Models to Adjust for Procedure and Hospital
We next developed a model to explore how patient characteristics influenced anesthesia time, adjusting for procedure and hospital. In the model, we combined general surgery and orthopedic surgery populations and adjusted for a hospital effect using an indicator for 182 of the 183 hospitals in the study with at least 50 patients. The median hospital in the data set served as the reference group. Table 3displays the model coefficients on patient characteristics, adjusting for procedure and hospital. For a given procedure, there were some patient covariates that were associated with longer anesthesia procedure times and others that were associated with shorter times. For example, covariates associated with longer anesthesia time included: multiple procedures on the same day as the principal procedure, +18.3 min (P  < 0.0001); transfer in from another acute care facility, +6.7 min (P  = 0.00002); black race, +5.5 min (P  < 0.0001); paraplegia, +4.5 min (P  = 0.0006); male sex +4.4 min (P  < 0.0001); cancer, +2.7 min (P  < 0.0001); diabetes, +2.4 min (P  < 0.0001); and hypertension, +2.0 min (P  < 0.0001). Factors associated with shorter anesthesia time included age older than 80 yr versus  ages 65–70 yr, -4.1 min (P  < 0.0001); and history of stroke, -3.5 min (P  < 0.0001).
Table 3. The Contribution of Patient Characteristics to Anesthesia Time 
Image not available
Table 3. The Contribution of Patient Characteristics to Anesthesia Time 
×
We also studied the individual hospital effect after adjusting for procedure and patient characteristics. Figure 1displays the distribution of hospital coefficients in the regression used to produce table 3. There were 182 hospital coefficients in this regression, with the reference hospital coded as a 0. The first box plot uses all 183 hospitals and provides a view of the variability in hospital procedure times. The next two plots are subsets of the first plot, describing hospitals with and without residency programs. There is considerable variation among hospital procedure times, adjusting for patient characteristics. Some of the variation seems to be explained by the presence of residency programs; however, within the groups of hospitals with and without residents, there continued to be considerable variation.
Fig. 1. Differences in hospital-adjusted procedure time. Box plots of the 183 hospitals (with at least 50 patients). As shown from the median values, the typical procedure performed in hospitals with residency programs was approximately 6 min slower than those performed in hospitals without residency programs. 
Fig. 1. Differences in hospital-adjusted procedure time. Box plots of the 183 hospitals (with at least 50 patients). As shown from the median values, the typical procedure performed in hospitals with residency programs was approximately 6 min slower than those performed in hospitals without residency programs. 
Fig. 1. Differences in hospital-adjusted procedure time. Box plots of the 183 hospitals (with at least 50 patients). As shown from the median values, the typical procedure performed in hospitals with residency programs was approximately 6 min slower than those performed in hospitals without residency programs. 
×
Taking a Closer Look at Race and Income
We next explored the interaction between race and income. We created three groups of income: Higher, Middle, and Lower Income (defined in Materials and Methods) and interacted these groups with race. We found an interesting pattern (table 4). After adjusting for patient characteristics and hospital identifiers, the lowest adjusted anesthesia time was found in the higher-income white group. Compared with this group, lower-income black patients had, on average, 8.5 min longer anesthesia time (P  < 0.0001). To place this in context, history of paraplegia contributed 4.5 additional minutes to anesthesia time (P  < 0.0001). Middle-income black patients displayed 4.8-min longer anesthesia time (P  = 0.011) than higher-income white patients, whereas higher-income black patients showed no difference in anesthesia time compared with the higher-income white group (-0.28 min, P  = 0.880). Anesthesia time for the lower-income white group was 1.25 min longer than that for the higher-income white group (P  < 0.016), but middle-income white patients showed no significant difference from higher-income white patients (0.67 min, P  = 0.146).
Table 4. The Influence of Patient Race and Income on Anesthesia Time 
Image not available
Table 4. The Influence of Patient Race and Income on Anesthesia Time 
×
To better appreciate the impact of the selection of hospitals by race and income groups, we presented the race and income procedure times in two comparative analyses in table 5. Each cell in the table provides two results, both adjusted for patient characteristics and procedures as in table 3. The top value provides the difference in minutes between the row group and the column group, and the associated P  value. The bottom result displays the same result also adjusted for the 183 hospitals in the data set, providing the within-hospital differences. For example, low-income black patients had an extra 29 min of procedure time compared with low-income white patients, and this was significant at the P  < 0.0001 level. However, after adjusting for the hospital, the difference in time decreased to 7 min (P  < 0.0001). It is interesting to note that before adjustment by hospital, there seems to be a large difference in the procedure times between black and white patients for each income level, but this disappears after adjusting by hospital. This suggests that black patients are going to hospitals that are generally associated with longer procedure times than white patients, for all income levels. Overall, even after adjusting for the hospital, there was a significant effect of race and income on operative time.
Table 5. The Influence of Patient Race and Income on Anesthesia Procedure Time 
Image not available
Table 5. The Influence of Patient Race and Income on Anesthesia Procedure Time 
×
Exploring Difference in Procedure Time between Black and White Patients within Hospitals
Although we have seen large differences in procedure time among hospitals, we have not yet closely examined differences in procedure time in individual hospitals by race. Table 6examines the 15 hospitals with the largest population of black surgical patients in Pennsylvania. Not surprisingly, most of these hospitals had residency programs. For each hospital, we provide the numbers of black and white patients, the type of hospital (Very Major Teaching; Major Teaching; Minor Teaching; and Very Minor or Non-Teaching), the hospital’s excess procedure time for its white patients compared with the “other nonresident hospital” white patients, the hospital’s excess procedure time for its black patients compared with “other nonresident hospital” white patients, and the adjusted difference in procedure time (black vs.  white) with the associated P  value. Within-hospital differences were modeled using the same model as in table 3but included hospital by race interaction terms. The overall Wald test of race-by-hospital interaction21 was significant at the 0.029 level (suggesting 15 hospitals plus two hospital groups differed in black versus  white excess procedure times). Adjusted procedure times in hospitals 4, 5, 6, 10, 11, other resident hospitals, and other nonresident hospitals were significantly longer for their black patients than their white patients. The pattern is complex, however. Hospital 3 is much slower than most others, but the difference between black and white procedure times is negligible and much smaller than at most hospitals. Hospital 5 is faster than most other hospitals, even for black patients, but there is 15-min difference between procedure times for black versus  white patients.
Table 6. Exploring Differences in Operative Procedure Time between Black and White Patients within Hospitals 
Image not available
Table 6. Exploring Differences in Operative Procedure Time between Black and White Patients within Hospitals 
×
There was considerable variability in hospital procedure length within this group of hospitals. For example, when comparing hospitals 5 and 11 in table 6, we see that hospital 11’s adjusted procedure time was more than 1.5 hours longer than that of hospital 5. For white patients, hospital 11 procedures were 90 min longer than the “other nonresident” hospital reference white group; for black patients, procedures took 106 min longer than that for white patients in the “other nonresident” group, compared with hospital 5, where procedures for white patients were 21 min faster and black patients were 7 min faster than the white “other nonresident” group. Other large differences in procedure length among individual hospitals were common.
We also fit a separate regression (results not shown) in which the 15 hospitals in table 6were combined into a single group and found that procedure times were 24 min slower for white patients in this group (P  < 0.0001) than for white patients at both “other residency hospitals” and “other nonresidency hospitals,” whereas for black patients, the difference was 41 min longer than the white patients at the “other residency hospitals” and “other nonresidency hospitals,” which corresponds to an additional 17 min for black patients over the difference for white patients (P  < 0.0001) (that is, 24 + 17 = 41 min).
The Influence of Race and Obesity on Procedure Time
One possible explanation to account for slower procedure times in black patients as compared to white patients could relate to obesity. If there was differential obesity between groups, and if obesity were responsible for longer surgery, then this may have lead to a spurious relationship between race and procedure length. Although Medicare claims did not record height and weight, we did have such information on a subset of the data set consisting of 1931 charts that were abstracted for the Surgical Outcomes Study. This was the data set used in the companion paper in this issue17 and in previous publications.12,13 Following convention, a body mass index (BMI) in the recommended range 18.5 to <25 is considered normal (used as the reference group), 25 to <30 is overweight, 30 to <35 is obese, >35 is severely obese, and <18.5 is underweight. In a model that included patient characteristics and procedures and hospital identifiers for the 19 largest hospitals in the data set, the black versus  white procedure time difference was 43.8 min (P  < 0.0001). When BMI information was added to the model, the black versus  white procedure time difference remained almost unchanged at 44.8 min (P  < 0.0001). BMI did have an effect on procedure time (consistent with others22–24): patients with BMI <18.5 had procedures 12.4 min (P  = 0.02) faster than the reference group; those with BMI 25-30 had 11.9 min slower procedure time (P  = 0.0017) than the reference group. Patients with BMI 30-35 had procedures 17.2 min slower than the reference group (P  = 0.003), and those with BMI >35 had a 14.4-min longer procedure time than the reference group (P  = 0.066). Despite the association of BMI with procedure time, we found that including BMI information did not influence the longer procedure time associated with black patients compared with white patients in this Medicare population. Of note, 10.6% of black patients had BMI values ≥30 versus  11.4% in white patients (P  = 0.20).
The Influence of Other Hospital Characteristics on Procedure Length
After adjustment for patient characteristics, principal procedure, multiple procedures, race, and income, we next explored whether other hospital factors were associated with procedure length as measured through anesthesia time units. We found that the presence of a trauma center predicted an additional 9.6 min of procedure length (P  < 0.0001); however, hospitals with 200 or more beds (compared with smaller hospitals) had only slightly shorter procedure length (-2.0 min, P  < 0.0001). Location of a hospital in an urban versus  rural setting (metropolitan area population ≥100,000 vs.  <100,000) had a slightly longer length (+1.8 min, P  = 0.0006).
Discussion
After adjusting for procedure, we found that specific patient and hospital characteristics predicted anesthesia time. Whereas anesthesia and surgical times are often discussed and analyzed in the anesthesia and surgical literature, few large-scale studies have been reported, in part because costly chart review would be necessary at most institutions. To our knowledge, estimates of anesthesia time by procedure using Medicare claims have not been reported on a large scale among multiple institutions before this study. Results from the Veteran’s Administration studies have explored variations in procedure time among hospitals in the Veteran’s Administration system,2 but studies utilizing Medicare claims obtained through Medicare, allowing for modeling using a vast number of hospitals and patients, has not been previously reported.
There are a number of important findings that should be noted from this analysis. First, we have identified specific medical conditions that may influence procedure length as defined using Anesthesia Time Units from Medicare. Although not surprising, such differences can be considerable, as we have reported the independent effect of these comorbidities, each adding some time to the procedure. One mechanism by which time is influenced by the presence of comorbidities may be in the set-up time, in that patients with comorbidities may require more monitoring before incision. However, the most important factor associated with procedure length based in Anesthesia Time Units is the incision to closure time, as we demonstrated in the companion article, in which the Spearman correlation between the time from induction to entrance to the recovery room time versus  the time from incision to closure was 90%.17 
We observed important variation in adjusted procedure length among hospitals. This was especially striking when we observed variation in procedure length among some individual large teaching hospitals. For example, when comparing hospitals 5 and 11 in table 6, we found that hospital 11’s adjusted procedure time was more than 1.5 hours longer than that of hospital 5. Other large differences in procedure length between individual hospitals were not unusual. The question then becomes what explains these differences. As we have described in figure 1, the mere presence of a residency program cannot explain the magnitude of the differences in procedure length—indeed, there seems to be great variability inside hospitals with residency programs. Clearly, there is considerable heterogeneity in what we mean by “residency program,” and we have learned from this analysis that defining a hospital by whether it has residents does little to help to explain our findings.
We also found striking differences in the procedure length of black patients versus  white patients, and income level also played an important role. For example, we observed that black patients with lower incomes had adjusted procedure times that were 29 min longer than those in similar lower-income white patients. However, this difference decreased to 12 min when comparing high-income black patients with high-income white patients. More important, the 29-min difference between lower-income black patients and white patients decreased to 7 min when adjusting for the individual hospital, and the 12-min difference between higher-income black patients and white patients vanished completely after accounting for the hospital in which each patient underwent surgery. This suggests that black patients seem to be going to hospitals with very different procedure lengths than white patients. When examining the disparity in procedure time between black and white patients, differences in procedure length between where the patients had surgery explains far more of the disparity than differences between how black patients and white patients are actually treated inside of hospitals. Efforts to understand why hospitals where black patients undergo surgery (often the large urban teaching hospitals listed in table 6) require so much longer procedure time for both black and white patients may aid in helping to reducing this apparent disparity. Such reductions in procedure length would benefit both black and white patients but have a disproportionately larger benefit for black patients, simply because a greater percentage of black patients are admitted to these slower hospitals.
Although most of the differences between procedure lengths for black and white patients can be explained by where patients are undergoing surgery, we also found significant differences between white and black patients inside some hospitals, as shown in table 6. In Pennsylvania, of the 15 hospitals with the largest black populations in the state, 5 had significant differences in adjusted procedure time ranging from 9 to 16 min. There were also considerable (and statistically significant differences) between black and white patients in some hospitals with smaller black populations not displayed in table 6.
Although our analysis may shed light on individual procedure lengths, it does not directly allow us to analyze hospital efficiency or clinical productivity.1,25–28 Our data only involve Medicare patients, so we do not know how many and when other non-Medicare patients are being treated.
In summary, we used the Medicare anesthesia bill to analyze information on anesthesia time in a large sample of surgical patients among many hospitals. We have attempted to control for patient characteristics and procedures and to explore how different patient and hospital characteristics influence procedure time. In the process, we found that hospital and patient characteristics do influence anesthesia procedure time. Important variables include patient characteristics such as multiple procedures, transfer status, paraplegia, male sex, hypertension, diabetes, and, perhaps surprisingly for this insured Medicare population, race and income. The most important factor explaining why race and income influence procedure time was the hospital in which these patients underwent surgery, not differential treatment inside hospitals. Lower-income black patients tended to go to hospitals that had longer procedure times for both black and white patients. Analysis within hospitals displayed far less differences in procedure times for black and white patients, but we did observe some significant differences between black and white patients inside a number of hospitals. Reducing the procedure length in large urban teaching hospitals may benefit both black and white patients at these hospitals and would have a disproportionately beneficial effect on black patients simply because a greater percentage of the black patient population is admitted to these hospitals than the white patient population.
We thank William J. Greeley, M.D., M.B.A. (Anesthesiologist-in-Chief, Department of Anesthesiology and Critical Care Medicine, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania), and Lee Fleisher, M.D. (Chair, Department of Anesthesiology and Critical Care Medicine, The University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania), for their aid in conducting this research. We thank the anonymous reviewers for their insightful comments and suggestions regarding this manuscript. We also thank Traci Frank (Administrative Coordinator) and Andrea Millman, B.A. (Research Assistant), from the Center for Outcomes Research, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, for aid in preparing this manuscript.
References
Abouleish AE, Prough DS, Zornow MH, Hughes J, Whitten CW, Conlay LA, Abate JJ, Horn TE: The impact of longer-than-average anesthesia times on the billing of academic anesthesiology departments. Anesth Analg 2001; 93:1537–43Abouleish, AE Prough, DS Zornow, MH Hughes, J Whitten, CW Conlay, LA Abate, JJ Horn, TE
Khuri SF, Najjar SF, Daley J, Krasnicka B, Hossain M, Henderson WG, Aust JB, Bass B, Bishop MJ, Demakis J, DePalma R, Fabri PJ, Fink A, Gibbs J, Grover F, Hammermeister K, McDonald G, Neumayer L, Roswell RH, Spencer J, Turnage RH: Comparison of surgical outcomes between teaching and nonteaching hospitals in the Department of Veterans Affairs. Ann Surg 2001; 234:370–82Khuri, SF Najjar, SF Daley, J Krasnicka, B Hossain, M Henderson, WG Aust, JB Bass, B Bishop, MJ Demakis, J DePalma, R Fabri, PJ Fink, A Gibbs, J Grover, F Hammermeister, K McDonald, G Neumayer, L Roswell, RH Spencer, J Turnage, RH
Schuster M, Standl T, Reissmann H, Kuntz L, Am Esch JS: Reduction of anesthesia process times after the introduction of an internal transfer pricing system for anesthesia services. Anesth Analg 2005; 101:187–94Schuster, M Standl, T Reissmann, H Kuntz, L Am Esch, JS
Sandberg WS, Daily B, Egan M, Stahl JE, Goldman JM, Wiklund RA, Rattner D: Deliberate perioperative systems design improves operating room throughput. Anesthesiology 2005; 103:406–18Sandberg, WS Daily, B Egan, M Stahl, JE Goldman, JM Wiklund, RA Rattner, D
Torkki PM, Marjamaa RA, Torkki MI, Kallio PE, Kirvela OA: Use of anesthesia induction rooms can increase the number of urgent orthopedic cases completed within 7 hours. Anesthesiology 2005; 103:401–5Torkki, PM Marjamaa, RA Torkki, MI Kallio, PE Kirvela, OA
Brunicardi FC, Andersen DK, Billiar TR, Dunn DL, Hunter JG, Pollock RE: Schwartz’s Principles of Surgery, 8th edition. New York, McGraw-Hill, 1994.Brunicardi, FC Andersen, DK Billiar, TR Dunn, DL Hunter, JG Pollock, RE New York McGraw-Hill
Haley RW, Culver DH, Morgan WM, White JW, Emori TG, Hooton TM: Identifying patients at high risk of surgical wound infection. A simple multivariate index of patient susceptibility and wound contamination. Am J Epidemiol 1985; 121:206–15Haley, RW Culver, DH Morgan, WM White, JW Emori, TG Hooton, TM
Culver DH, Horan TC, Gaynes RP, Martone WJ, Jarvis WR, Emori TG, Banerjee SN, Edwards JR, Tolson JS, Henderson TS: Surgical wound infection rates by wound class, operative procedure, and patient risk index. National Nosocomial Infections Surveillance System. Am J Med 1991;91:152S–7SCulver, DH Horan, TC Gaynes, RP Martone, WJ Jarvis, WR Emori, TG Banerjee, SN Edwards, JR Tolson, JS Henderson, TS
Scott CF Jr:Length of operation and morbidity: Is there a relationship? Plast Reconstr Surg 1982;69:1017–21.Scott, CF
Grossmann EM, Longo WE, Virgo KS, Johnson FE, Oprian CA, Henderson W, Daley J, Khuri SF: Morbidity and mortality of gastrectomy for cancer in Department of Veterans Affairs Medical Centers. Surgery 2002; 131:484–90Grossmann, EM Longo, WE Virgo, KS Johnson, FE Oprian, CA Henderson, W Daley, J Khuri, SF
Silber JH, Rosenbaum PR, Trudeau ME, Even-Shoshan O, Chen W, Zhang X, Mosher RE: Multivariate matching and bias reduction in the surgical outcomes study. Med Care 2001; 39:1048–64Silber, JH Rosenbaum, PR Trudeau, ME Even-Shoshan, O Chen, W Zhang, X Mosher, RE
Silber JH, Rosenbaum PR, Trudeau ME, Chen W, Zhang X, Rapaport Kelz R, Mosher RE, Even-Shoshan O: Changes in prognosis after the first postoperative complication. Med Care 2005; 43:122–31Silber, JH Rosenbaum, PR Trudeau, ME Chen, W Zhang, X Rapaport Kelz, R Mosher, RE Even-Shoshan, O
Silber JH, Rosenbaum PR, Trudeau ME, Chen W, Zhang X, Lorch SA, Kelz RR, Mosher RE, Even-Shoshan O: Preoperative antibiotics and mortality in the elderly. Ann Surg 2005; 242:107–14Silber, JH Rosenbaum, PR Trudeau, ME Chen, W Zhang, X Lorch, SA Kelz, RR Mosher, RE Even-Shoshan, O
US Department of Health and Human Services: Statistical supplement 1997. Health Care Finance Rev 1997:31.
US Department of Health and Human Services: Statistical supplement 1998. Health Care Finance Rev 1998:177.
Steen PM, Brewster AC, Bradbury RC, Estabrook E, Young JA: Predicted probabilities of hospital death as a measure of admission severity of illness. Inquiry 1993; 30:128–41Steen, PM Brewster, AC Bradbury, RC Estabrook, E Young, JA
Silber JH, Rosenbaum PR, Rapaport Kelz R, Zhang X, Even-Shoshan O: Estimating anesthesia and surgical time from Medicare anesthesia claims. Anesthesiology 2007; 106:346–55Silber, JH Rosenbaum, PR Rapaport Kelz, R Zhang, X Even-Shoshan, O
Huber PJ: The basic types of estimates, Robust Statistics, New York, John Wiley & Sons, 1981, pp 43–55.Huber, PJ New York John Wiley & Sons
Hampel FR, Ronchett EM, Rousseeuw PJ, Stahel WA: Linear models: Robust estimation, Robust Statistics: The Approach Based on Influence Functions, New York, John Wiley & Sons, 1986, pp 315–28.Hampel, FR Ronchett, EM Rousseeuw, PJ Stahel, WA New York John Wiley & Sons
Huber PJ: Robust regression: Asymptotics, conjectures and the Monte Carlo. Ann Stat 1973; 1:799–821Huber, PJ
Hampel FR, Ronchett EM, Rousseeuw PJ, Stahel WA: Linear models: Robust testing, Robust Statistics. The Approach Based on Influence Functions. New York, John Wiley & Sons, 1986, pp 364–6.Hampel, FR Ronchett, EM Rousseeuw, PJ Stahel, WA New York John Wiley & Sons
Lee CT, Dunn RL, Chen BT, Joshi DP, Sheffield J, Montie JE: Impact of body mass index on radical cystectomy. J Urol 2004; 172:1281–5Lee, CT Dunn, RL Chen, BT Joshi, DP Sheffield, J Montie, JE
Chang SS, Jacobs B, Wells N, Smith JA Jr, Cookson MS: Increased body mass index predicts increased blood loss during radical cystectomy. J Urol 2004; 171:1077–9Chang, SS Jacobs, B Wells, N Smith, JA Cookson, MS
Dindo D, Muller MK, Weber M, Clavien P-A: Obesity in general elective surgery. Lancet 2003;361:2032–5.Dindo, D Muller, MK Weber, M Clavien, P-A
Abouleish AE, Prough DS, Whitten CW, Zornow MH, Lockhart A, Conlay LA, Abate JJ: Comparing clinical productivity of anesthesiology groups. Anesthesiology 2002; 97:608–15Abouleish, AE Prough, DS Whitten, CW Zornow, MH Lockhart, A Conlay, LA Abate, JJ
Abouleish AE, Prough DS, Barker SJ, Whitten CW, Uchida T, Apfelbaum JL: Organizational factors affect comparisons of the clinical productivity of academic anesthesiology departments. Anesth Analg 2003; 96:802–12Abouleish, AE Prough, DS Barker, SJ Whitten, CW Uchida, T Apfelbaum, JL
Abouleish AE, Prough DS, Whitten CW, Zornow MH: The effects of surgical case duration and type of surgery on hourly clinical productivity of anesthesiologists. Anesth Analg 2003; 97:833–8Abouleish, AE Prough, DS Whitten, CW Zornow, MH
Abouleish AE, Dexter F, Whitten CW, Zavaleta JR, Prough DS: Quantifying net staffing costs due to longer-than-average surgical case durations. Anesthesiology 2004; 100:403–12Abouleish, AE Dexter, F Whitten, CW Zavaleta, JR Prough, DS
Fig. 1. Differences in hospital-adjusted procedure time. Box plots of the 183 hospitals (with at least 50 patients). As shown from the median values, the typical procedure performed in hospitals with residency programs was approximately 6 min slower than those performed in hospitals without residency programs. 
Fig. 1. Differences in hospital-adjusted procedure time. Box plots of the 183 hospitals (with at least 50 patients). As shown from the median values, the typical procedure performed in hospitals with residency programs was approximately 6 min slower than those performed in hospitals without residency programs. 
Fig. 1. Differences in hospital-adjusted procedure time. Box plots of the 183 hospitals (with at least 50 patients). As shown from the median values, the typical procedure performed in hospitals with residency programs was approximately 6 min slower than those performed in hospitals without residency programs. 
×
Table 1. Estimated Anesthesia Time for the 20 Most Frequent Surgical Procedures in the Surgical Outcomes Study 
Image not available
Table 1. Estimated Anesthesia Time for the 20 Most Frequent Surgical Procedures in the Surgical Outcomes Study 
×
Table 2. Estimated Anesthesia Time for the 20 Most Frequent Orthopedic Procedures in the Surgical Outcomes Study 
Image not available
Table 2. Estimated Anesthesia Time for the 20 Most Frequent Orthopedic Procedures in the Surgical Outcomes Study 
×
Table 3. The Contribution of Patient Characteristics to Anesthesia Time 
Image not available
Table 3. The Contribution of Patient Characteristics to Anesthesia Time 
×
Table 4. The Influence of Patient Race and Income on Anesthesia Time 
Image not available
Table 4. The Influence of Patient Race and Income on Anesthesia Time 
×
Table 5. The Influence of Patient Race and Income on Anesthesia Procedure Time 
Image not available
Table 5. The Influence of Patient Race and Income on Anesthesia Procedure Time 
×
Table 6. Exploring Differences in Operative Procedure Time between Black and White Patients within Hospitals 
Image not available
Table 6. Exploring Differences in Operative Procedure Time between Black and White Patients within Hospitals 
×