The Financial Impact of Teaching SurgicalResidents in the Operating Room

Matthew Bridges, MD, Daniel L. Diamond, MD, Knoxville, Tennessee

BACKGROUND: There have been no published dataregarding the cost of training surgical residentsin the operating room.

METHODS: At the University of Tennessee MedicalCenter-Knoxville, in addition to resident-per-formed teaching cases, some cases are per-formed without the assistance of residents bythe same faculty.

RESULTS: Sixty-two case categories involving14,452 cases were compared for operative timesalone. In 46 case categories (10,787 procedures),resident operative times were longer than facultyalone. In 16 case categories, resident operatingtimes were shorter than faculty times. The netincremental operative time cost was 2,050 hoursbetween July 1993 and March 1997. Assuming 4years of operative training for 11 graduatingchief residents, the cost per graduating residentwas $47,970.

CONCLUSION: Extrapolated to a national annualcost for the 1,014 general surgery residents whocompleted training in the 1997 academic year,the annual cost of training residents in the oper-ating room is $53 million. This high monetarycost suggests the need for digital skills, selec-tion criteria, the development of training curricu-lum and resource facilities, the pre-operatingroom need for suturing and stapling techniques,and perhaps the acquisition of virtual surgerytraining modules. Am J Surg. 1999;177:28–32.© 1999 by Excerpta Medica, Inc.

Never sine the beginning of an organized residencysystem in the United States has the future ofgraduate medical education been more unclear.

With the merger of hospitals and hospital systems, not tomention the large shifts in patient base that have affectedindividual hospitals because of HMO and insurance con-tracting, the stability of patient base for education hasbecome a major issue for graduate medical education. Theemphasis on faculty economic productivity to support med-ical school budgets and to provide self-support for faculty

whose teaching and research salaries have disappearedthreaten the stability of teaching faculty. The decliningreimbursement per operative case for attending surgeonsand faculty has further assaulted the time commitmentsrequired to educate surgical and other trainees.

Perhaps most significant of all has been the gradual shiftin Medicare reimbursement towards support for primarycare fields and away from surgical specialty training. Inaddition, the change from indemnity, cost-based reim-bursement which included hospital expenses for graduatemedical education toward a system of managed care, wherepricing has not included the cost of graduate medicaleducation, has left the United States Government as theprimary monetary support for graduate medical training.1

The Medicare program alone is estimated to have paid$6.8 billion in 1997 for direct medical expense andindirect medical expense. There has been and is no clearconsensus and no national vision as to who or whatentities should pay for graduate medical education in thefuture.

Unfortunately, there is little knowledge of the true costof graduate medical education and even less objectivedata reflective of the true cost of surgical education. Atcountless meetings in which residency training has beendiscussed, one may hear such comments as “Surgicalresidents are the best deal a hospital ever had,” “Surgicalresidents get paid about thirty cents an hour,” and“Hospitals are making out like bandits on indirect anddirect medical education reimbursement from Medi-care.” In spite of this plethora of self-righteous opinions,there have been no data indicating the true monetarycost of training a surgical resident. One of the criticalpieces of missing data is the cost to a hospital fortraining residents in the operating room. This study wasundertaken to attempt to quantitate the hospital costsincurred as a result of the training of residents in theoperating room.

METHODSSince our residency is not large enough to allow resident

participation in all operative procedures, surgical facultyperform some cases without resident participation. Sincethe faculty surgeon is constant, the variable of residentparticipation was thought to be an adequate means ofseparating cases into those definitely not done by a residentand those probably done by a resident. Both resident andfaculty evaluations have suggested that over 95% of theprocedures that involve a resident are performed by theresident with the faculty as teaching assistant.

Sixty-two procedures were selected from the Univer-sity of Tennessee hospital operating room records fortheir relative frequency and the likelihood that therewould be significant numbers of cases performed by

From the Department of Surgery, University of Tennessee Med-ical Center-Knoxville, Knoxville, Tennessee.

Requests for reprints should be addressed to Daniel L. Dia-mond, MD, Department of General and Peripheral Vascular Sur-gery, 382 W. Chestnut Street, Suite 106, Washington, Pennsyl-vania 15301.

Manuscript submitted August 17, 1998 and accepted in revisedform September 28, 1998.

28 © 1999 by Excerpta Medica, Inc. 0002-9610/99/$–see front matterAll rights reserved. PII S0002-9610(98)00289-X

residents. Most of the cases were types that a generalsurgery resident would be expected to complete withinan approved residency (Table I, Table II). They weredrawn from the fields of general, pediatric, vascular,plastic, urologic, and trauma surgery. Cases that wereperformed in low numbers or those in which a residentwas always present, eg, Whipple procedure, were ex-cluded from the analysis. The cases of surgeons who do

not participate with the surgical residency as teachingfaculty were also excluded from the study.

A total procedure time was determined from operativerecords used for hospital billing. The beginning of theprocedure time was the incision time and the end of theprocedure time was the “left room time,” both recordedby the circulating nurse. A total procedure time for eachtype of procedure, as well as a mean procedure time and

TABLE IForty-Six Categories Where Resident Time Was Longer

Procedures

Without Residents With Residents Difference

#Cases

Avg.Min.

#Cases

Avg.Min.

Avg.Min.

TotalMin.

Appendectomy 35 57 125 78 21 2,625Appendectomy, laparoscopic 26 72 154 94 22 3,388AAA 7 105 254 178 73 18,542AV fistula/graft 41 88 290 105 17 4,930AV revision 29 65 126 95 30 3,780Amputation, toe/transmet 10 20 59 28 8 472Anoplasty/endorectal advance 30 28 12 54 26 312Bowel resection, low anterior 4 113 48 210 97 4,656Breast mass biopsy or removal 638 43 648 54 11 7,976Breast reconstruction, TRAM 14 292 12 453 161 1,932Biopsy, node 77 43 86 56 13 1,118Cystectomy, radical w/ileocon 20 216 10 263 47 470CEA 28 126 609 128 2 1,218Cholecystectomy, laparoscopic 294 80 604 89 9 5,436Colostomy 11 85 65 93 8 520Colostomy closure 15 110 59 118 8 472Cysto w/RPG/cystogram 361 24 49 34 10 490Embolectomy UE 11 74 54 86 12 648Embolectomy LE 8 52 118 117 65 7,670Excision lesion/lipoma 284 46 249 76 30 7,470Excision cyst 63 35 56 45 10 560Exploratory laparotomy 60 111 723 121 10 7,230Feeding jejunostomy 7 47 44 60 13 572Femoral–popliteal bypass graft 10 148 277 176 28 7,758Femoral–tibial bypass graft 8 194 111 206 12 1,3321 & D abscess/wound/mass 84 26 178 61 35 6,230IVC filter 40 23 163 44 21 3,423Hernia, inguinal, unilateral 240 70 420 81 11 4,620Hernia, umbilical 75 45 119 90 45 5,359Hernia, ventral/incisional 78 87 230 96 9 2,070Hydrocelectomy, unilateral Ing. 29 25 9 39 14 126Laparotomy/LOA 11 75 32 120 45 1,440Mastectomy, bilateral 7 167 42 194 27 1,134Mastectomy, modified radical 60 119 252 141 22 5,544Mastectomy, simple/partial 56 69 101 94 25 2,525Nephrectomy 90 91 38 105 14 532Parathyroidectomy, total/subtotal 22 87 90 120 33 2,970PEG tube insertion 19 37 40 63 26 1,040Prostatectomy, radical 42 100 19 188 88 1,672Raz/Stamey urethropexy 84 52 28 78 26 728Rectal biopsy 3 12 7 37 25 175STSG 230 62 88 78 16 1,408Thoracotomy 83 115 151 152 37 5,587Thyroidectomy 47 99 175 133 34 5,950Tracheostomy 71 35 212 50 15 3,180VVSL, unilateral 12 71 77 86 15 1,159Total 3,474 7,313 148,812

AAA, abdominal aortic aneurysm; AV, arteriovenous; CEA, carotid endarterectomy; IVC, inferior vena cava; LE, lower extremity; LOA, lysis of adhesions; PEG,percutaneous endoscopic gastrostomy; RPG, retrograde pyelogram; STSG, split thickness skin graft; TRAM, transfer rectus abdominis muscle; VVSL, varicose veinstripping and ligation.

TEACHING SURGICAL RESIDENT IN OPERATING ROOM/BRIDGES AND DIAMOND

THE AMERICAN JOURNAL OF SURGERY® VOLUME 177 JANUARY 1999 29

a mean difference in the length of the procedure with orwithout a resident were determined using regression andanalysis of means. The mean time lost or gained wascalculated in each case category. The total time lost wasgenerated by subtracting the total net time difference incategories when resident cases were shorter from thetotal net time difference in categories where residentcases were longer.

During the course of the study (July 1993 through March1997), there were 11 surgical trainees who completedtraining. This number was divided into the total “lost” timeto yield the lost time per graduated resident. Thus, the losttime is an averaged figure that includes time lost by bothjunior and senior residents over a 4-year period.

The cost per hour of operating time was determined bydividing the total minutes of all operations in 1 year intothe non-supply cost associated with running the operatingroom for that year. The cost per minute was then multi-plied by 60 to generate an hourly cost. The non-supplycosts include administrative salaries, hourly wages, over-time, on-call pay, longevity pay, and the cost of all benefits(including retirement) of all operating room personnel and20 CRNAs (Table III). For purposes of this calculation,the cost of surgeon and anesthesiologist time was not

considered, since this is not a hospital expense in mostsystems. The 1997 fiscal year (July 1 to June 30) wasselected as the index year for purposes of cost calculationformula.

RESULTSThere were 14,452 total cases for analysis, 9,733 with a

resident and 4,719 with no resident (Table IV). Therewere 46 operative case categories (10,787 cases) in whichprocedures performed with a resident required longer op-erative times (103 minutes average time) when comparedwith cases done without a resident (61 minutes averagetime). This resulted in a total of 148,812 more minutes(2,480 hours), of operating time than would have beenexpected if the case times had been the same as thosewhich were performed without residents in these 46 cate-gories (Table V). In the 16 categories (3,665 cases) inwhich operating time of cases performed with residents wasactually shorter than cases performed without a resident,the total time difference was 25,788 minutes (430 hours).

Analysis of all of the 62 categories yields a net incremen-tal operative time “cost” of 123,024 minutes (2,050 hours).Given a cost figure of $4.29 per minute (excludes supplies,

TABLE IISixteen Case Categories Where Resident Time Was Shorter

Procedures

Without Residents With Residents Difference

#Cases

Avg.Min.

#Cases

Avg.Min.

Avg.Min.

TotalMin

AV access declot 31 78 185 74 24 2740ASD 32 137 18 131 26 2108AVR 50 228 11 209 219 2209Amputation, LE 58 67 361 63 24 21,444Axillary dissection 32 89 103 76 213 21,339Bowel resection, small 22 125 133 121 24 2532Cholecystectomy, lap w/chol 180 93 501 82 211 25,511Colectomy/hemicolectomy 27 190 223 158 232 27,136Fistulotomy/fistulectomy 44 28 31 25 23 293Hemorrhoidectomy 149 68 100 40 228 22,800Kidney transplant, cad/LRD 16 190 108 189 21 2108Laparoscopy, diagnostic 32 105 130 74 231 24,030Muscle biopsy 17 31 15 25 26 290Port placement 537 47 478 45 22 2956Thoracotomy, lobectomy 7 152 12 11 241 2492VSD 11 188 11 168 220 2220Total 1,245 2,420 25,788

ASD, atrial septic defect; AV, arteriovenous; AVR, aortic valve replacement; LE, lower extremity; LRD, live related donor; VSD, ventricular septic defect.

TABLE IIIUniversity of Tennessee Medical Center-Knoxville

Operating Room 1997 Fiscal Year (July 1 through June 30)

Salaries, wages, benefits (OR personnel) $6,384,075Salaries, wages, benefits (20 CRNA) $2,196,814Total $8,580,889Total number of surgical cases (includes

nonteaching and surgeons courtesy) 18,330Total operating minutes 2,001,180Cost per operating minute $ 4.29Cost per operating hour without anesthesiologist $ 257.27

TABLE IVUniversity of Tenessee Medical Center-Knoxville Surgical

Teaching Case Data (July 1993–March 1997)

Total cases for analysis 14,452Cases with resident 9,733Cases with no resident 4,719Forty-six categories where resident time was longer

7,313 cases with a resident3,474 cases with no resident

Sixteen categories where resident time was shorter2,420 cases with a resident1,245 cases with no resident

TEACHING SURGICAL RESIDENT IN OPERATING ROOM/BRIDGES AND DIAMOND

30 THE AMERICAN JOURNAL OF SURGERY® VOLUME 177 JANUARY 1999

indirect costs, anesthesiologist fees or costs, and surgeonsfees), the total cost for this “lost time” was $527,772.96.Within the 62 case categories that were analyzed for thisreport, there were 9,733 operative cases performed withresidents or 885 per graduating chief resident (Table VI).Given a total lost or extra minute figure of 123,024 minutesgenerated by 9,733 cases performed with residents, the4-year training figure is 11,184 lost minutes of operativetime, or $47,979 of cost per graduating resident.

COMMENTSOne can certainly take exception to the methodologies

used in this analysis. Some of the cases listed with theresident may actually have been done by faculty, but ourinternal analysis seems to indicate that cases done with aresident listed by operating room personnel as assistantserves as a reasonable proxy for “case done by resident withfaculty in attendance.” If anything, the potential to havesome resident cases actually done by faculty would mini-mize the effect of lost time rather than to magnify it, sincethere are more cases that were prolonged than shortened byresident participation.

The operative times may not be totally accurate, but theyserve as the basis for hospital billing. One may reasonablyassume that there are the same inaccuracies whether aresident is scrubbed or not scrubbed. Some case categorieshave a high percentage of resident participation whereasothers have a low resident participation, thus creating anonrandom distribution. There is also the possibly of fac-ulty or resident selection bias because of resident partici-pation in more difficult or index type cases. While all ofthese are valid potential criticisms, the large number ofcases should tend to nullify the potentially confoundingeffects of some of these concerns.

It would have been easier and neater to use 5 years ofdata, but our hospital operating room information systemhad slightly less than 4 years of data that could substantiateresident participation and length of procedure. Unfortu-nately, the hospital information system does not use thesame 277 case categories that exist in the Surgical Opera-tive Log of the Residency Review Committee for Surgery,

thus partially explaining the disparity between the 885cases per graduating resident generated by our data and theaverage 938 reported to the RRC for the 11 graduatingresidents.

The reasons for some case categories in which residentsseemingly operated faster than their teaching faculty are ofobvious interest. Some, such as laparoscopic cholecystec-tomy with cholangiogram, colectomy, or kidney transplant,are cases in which the individual resident has as high avolume of experience as the operating surgeon or is moreaccustomed to doing that case with the teaching faculty asan assistant than the teaching faculty is without a resident.In addition, cases that are performed by teaching facultywith no resident participation are often performed withsignificantly less assistance, as it is rare for faculty surgeonsto assist each other and there are no paid surgical assistantsat University of Tennessee Medical Center at Knoxville.Whatever assistance is provided, if any, is usually an addi-tional scrub technician.

The cost figures are thought to represent easily measur-able costs in most institutions. There are undoubtedly localissues in each hospital that would make an analysis of thecosts somewhat different. By defining the costs that weused in the analysis, other institutions should be able toarrive at a comparison.

Most discussions of indirect medical expense paymentsto hospitals list such things as additional diagnostic testsand lower hospital staff productivity as one of the ra-tionales for indirect medical expense payment fromMedicare. To our knowledge, there has never been afinancial analysis of incremental cost associated withtraining a resident in the operating room. Given the factthat there are approximately 21,000 residents enrolledin core surgical programs in the United States,2 it wouldseem prudent to study this issue in more detail sincethere is a direct financial cost over and above that paidfrom the patients and income of teaching surgeons.When and if there is a national consensus regardinggraduate medical education funding, the cost of opera-tive training should be factored in.

If one attempts to extrapolate from this study to a na-tional annual cost figure for general surgery alone, the costis somewhere in the $53 million range. This assumes aper-case loss of $54.23 for the 11 residents who finished theprogram involved in this study, 961 average cases reportedin the Surgical Operative Log for 1,014 general surgeryresidents who completed training in the 1997 academicyear.3

If the dictates of optimal patient care and the stressesimposed on faculty surgeons did not already demand it, thedirect economic cost of training residents on patientswould suggest that we need to look for ways to make this amore productive and efficient process. Perhaps there is aneed for an articulated technical training curriculum thatincludes the setting and achievement of goals and stan-dards. There is a need for development of technical re-source facilities where residents can practice and can com-pete in technical exercises. There is a need for animal orcadaver laboratories in which residents are taught suturingand stapling techniques. There is a need for the promise ofvirtual surgery to be realized for the benefit of training

TABLE VUniversity of Tennessee Medical Center-Knoxville

Net Operating Time Gain or Loss Analysis

Net time loss (46 categories) 148,812 min (2,480 hr)Net time gained (16 categories) 25,788 min (430 hr)“Net of nets” time ‘lost’ to

resident training 123,024 minNet cost @ $4.29/minutes

(11 graduating residentsover 4 years) $527,772.96

TABLE VIUniversity of Tennessee Medical Center-Knoxville Four-

Year Operating Room Training Cost per GraduatingResident

Cases with residents 9,733Cases per graduating resident 885Net time lost per case 12.64 minNet lost time per resident 11,816 minCost per graduating resident $47.797

TEACHING SURGICAL RESIDENT IN OPERATING ROOM/BRIDGES AND DIAMOND

THE AMERICAN JOURNAL OF SURGERY® VOLUME 177 JANUARY 1999 31

surgical residents. While these alternatives may be expen-sive, they are one-time costs that do not recur each time aresident steps in the operating room. And finally, perhapsit is not too early for us to recognize that surgery is andalways has been a technical exercise at which some peopleare more adept than others and to reflect this recognitionin our selection processes.

Although the figures generated by this analysis are likelyto be different in each institution, it is hoped that all willreflect a net time cost to training residents in the operatingroom. If that is not the case, perhaps those of us who are

responsible for that training are in need of some remedialtechnical training.

REFERENCES1. Fishman LE. Medicare payments with an education label: fun-damentals and the future. Association of American MedicalColleges; 1996.2. Kwakwa F, Jonasson O. The longitudinal study of surgical resi-dents, 1993 to 1994. J Am Coll Surg. 1996;183:425–433.3. Residency Review Committee for Surgery. Program NationalData. Reporting period 7/1/96–6/3/97. Accreditation Council forGraduate Medical Education, Chicago, IL; 1997.

TEACHING SURGICAL RESIDENT IN OPERATING ROOM/BRIDGES AND DIAMOND

32 THE AMERICAN JOURNAL OF SURGERY® VOLUME 177 JANUARY 1999