Original communications

Comparison of data extractionfrom standardized versus traditionalnarrative operative reports fordatabase-related research andquality controlA. Harvey, MD, MSc,a H. Zhang, MEng, MSc,b J. Nixon, MD,a and C. J. Brown, MD, MSc,c Calgaryand Vancouver, Canada

Background. The purpose of this study was to compare the completeness and reproducibility of dataextracted from a standardized operative report (SOR) with the non–standardized operative report(NSOR).Methods. Between July and December 2003, operative data were collected from all laparoscopiccholecystectomy procedures performed at the Peter Lougheed Centre Hospital. A standardized format fordictating laparoscopic cholecystectomy operative reports was introduced on October 1, 2003.Non–standardized operative reports dictated in the first 3 months of the study period were comparedwith SORs dictated in the final 3 months. Two physicians independently extracted data from eachoperative report into a surgical database.Results. During the study period, 221 cholecystectomy reports were analyzed (119 SOR and 102NSOR). Completeness of data extraction for identifying variables (eg, patient name, age, and date ofprocedure) was similar in the 2 types of reports. However, most other operative and perioperative detailswere more completely reported in the SOR (95% to 100%) when compared to the NSOR (14% to100% complete). Furthermore, interobserver agreement between 2 independent data extractors was betterfor the SOR than the NSOR (0.9972 vs 0.9809, P � .0001).Conclusions. Standardized operative reports result in more complete and reliably interpretable operativedata compared with NSORs. (Surgery 2007;141:708-14.)

From the Division of General Surgery, Peter Lougheed Center, Calgarya; Centre for Health Evaluation and

Outcomes Sciencesb and Division of General Surgery,c St. Paul’s Hospital, Vancouver; Canada

Traditional, dictated operative reports are theofficial medical documentation of an operation. Thecontent of these documents is not usually standard-ized or regulated. During their residency training,surgeons receive little or no formal teaching1 onhow to properly document an operative procedure.Operative reports contain information critical for

Accepted for publication January 6, 2007.

Reprint requests: Carl J. Brown, MD, Rm c310, St. Paul’s Hos-pital, 1081 Burrard St, Vancouver, Canada V6Z 1Y6. E-mail:cbrown@providencehealth.bc.ca.

0039-6060/$ - see front matter

© 2007 Mosby, Inc. All rights reserved.

doi:10.1016/j.surg.2007.01.022

708 SURGERY

effective patient care and are often scrutinized inmedicolegal proceedings. Moreover, operative re-ports are often examined for research purposesand for quality assurance and improvement initia-tives.

Quality assurance and improvement has becomea priority in health care systems.2-4 Quality assur-ance typically involves assessment of the structure(ie, human and structural resources), process (ie,interventions and policies), and outcomes (eg,morbidity, cost, etc) of care. Surgical proceduresare an important component of the process ofhealth care. However, operative reports are notori-ously inadequate in the documentation of surgical

procedures.5,6

Surgery Harvey et al 709Volume 141, Number 6

There is limited research related to standard-ized recordkeeping in the context of surgicalprocedures. Recently, several investigations haveimplemented medical report templates and dem-onstrated improved efficiency,7,8 physician satisfac-tion,9 and reduced administrative costs.10

The purpose of this study was to compare thecompleteness and reproducibility of operative datacollected in a standardized operative report (SOR)for laparoscopic cholecystectomy to data extractedfrom traditional, non–standardized operative re-ports (NSOR).

PATIENTS AND METHODSDevelopment of the SOR. In July 2003, 9 gen-

eral surgeons at the Peter Lougheed CenterHospital developed an SOR for laparoscopic cho-lecystectomy procedures. The report included 2parts: a narrative report outlining the steps of theprocedure and a standardized summary at the endof the report.

The elements included in the standardized sum-mary were generated through a review of the liter-ature on the technique and complications oflaparoscopic cholecystectomy. A Medline searchwas performed using the keywords “laparoscopic,”“cholecystectomy,” “operative technique,” “opera-tion,” “complications,” “morbidity,” and “mortal-ity.” The search was limited to human, Englishlanguage, and publication dates of 1990–2003. Alarge series of patients documenting complications,morbidity, and mortality was selected from thesearch results and reviewed.11-18 In addition, the au-thors surveyed operative and procedural databasesused at the University of Calgary, including theGeneral Surgery Residents Operative database, theDivision of Transplant Surgery database, and the Pe-ter Lougheed Center Endoscopy database. Fromthis process, a list of operative details was identifiedthat all participating surgeons agreed should bedocumented in laparoscopic cholecystectomy oper-ative reports. This final list of operative details wasformulated into a summary paragraph at the end ofeach operative report (Fig).

On October 1, 2003, the SOR for laparoscopiccholecystectomy was implemented at the PeterLougheed Center. There were 9 participating staffsurgeons. All cholecystectomies that were initiatedas laparoscopic procedures were included in thestudy regardless of conversion to an open opera-tion, performance of a cholangiogram, commonbile duct exploration, or additional proceduresperformed during the same operation. All chole-cystectomies that were initiated as open procedures

were excluded from the study.

Study design. This study was approved by theUniversity of Calgary Research Ethics Board. Theintervention (SOR) group included all SORs com-pleted in the first 3 months after the change inoperative dictation. The control (NSOR) groupincluded all operative reports for laparoscopic cho-lecystectomies from the 3 months prior to the ini-tiation of the SOR.

Two surgeons independently extracted datafrom each operative report into the LaparoscopicCholecystectomy Operative database using Mi-crosoft Excel software (Microsoft, Redmond,Wash). Data completeness was determined for allelements of each report as a dichotomous variable:present according to both data extractors or not.Surgeon compliance was evaluated by comparingthe number of standardized reports completed tothe number of eligible procedures performed dur-ing the study period.

Statistical analysis. Completeness of data extrac-tion was compared between the 2 report types us-ing the Fisher exact test. Based on an expecteddifference of 10% in the completeness of data anda power of 0.8, a sample size of 100 reports in eachgroup was required. Compliance was analyzed us-ing descriptive statistics. Simple Kappa statisticswere calculated to evaluate component-wise agree-ment. For the comparison of overall agreementbefore and after standardization, a logistic randomeffect model was used with each operative reportconsidered independent and the agreement ofeach element considered a repeated measure (ex-cluding identifying data). The statistical analysiswas carried out using SAS®, version 9.1 software,(SAS Institute Inc, Cary, NC). A P value of .05 wasconsidered statistically significant for all analyses.

RESULTSDuring the entire study period, 102 NSORs were

compared with 119 SORs. Mean data extractiontime was 156 seconds versus 124 seconds. The over-all compliance with the SOR was 63% (119/189).Compliance improved from 53% in the first monthof the study period to 67% over the final 2 months.

Operative data extracted from the SOR and theNSOR are shown in Table I. Identifying data (sur-geon, anesthetist, patient, and date of procedure)were similarly reported in the SOR and NSOR.Most perioperative and operative data were morecompletely reported in the standardized report(Table I). Two important exceptions were the useof a gallbladder retrieval bag (59/102 vs 75/119,P � .45) and the size of the operative trocars (57/

102 vs 69/119, P � .75). These data were extracted

710 Harvey et al SurgeryJune 2007

1. Operative Urgency: Emergent (Surgical emergency – procedure necessary within 6 hours of assessment) Urgent (Surgery during admission after emergency visit but not “emergent”) Elective

2. Inpatient or Outpatient i.Date of admission (if inpatient)

3. Indication: Biliary colic Cholecystitis Gall stone pancreatitis Cholangitis Acalculous cholecystitis Other

4. DVT Prophylaxis Pneumatic compression stockings Heparin Stockings and Heparin Neither Other

5. Preoperative Antibiotics Specify Antibiotic

6. Abdominal Entry: Hassan Veres Needle Visible Port Insertion Device Other

7. Gall Bladder Appearance Normal Inflamed Scarred (previous inflammation) Gangrenous Other

8. Conversion to Open Yes or No

i. Reason: Surgical complication (e.g. Bleeding,CBD injury) Technical problems (e.g. Inability to complete) Anatomic abnormality CBD exploration Malignancy suspected Other

9. Cystic Artery Ligation : # Clips

10. Cystic Duct Ligation: # Clips Stapler Laparoscopic Loop Suture Application

11. Bile Spillage: Yes or No

12. Gall Stone Spillage: Yes or No i. Recovered: None,

Some (<50%) Most (>50%) All

13. Intraoperative Cholangiogram: Yes or No

i. Indication: Routine Liver enzyme abnormality U/S findings (eg. CBD diameter) Intraoperative findings (anatomy,etc) Pancreatitis Cholangitis Other

ii.Findings: Normal Filling defect CBD injury Abnormal anatomy (specify)

14. Fascial Closure: Umbilical only Sub-xiphoid only BothNeither * CBD – common bile duct

Fig. Operative details documented in standardized operative summary.

Surgery Harvey et al 711Volume 141, Number 6

from the reports but were not included in thestandardized summary.

Interobserver agreement between the 2 sur-geons who performed the data extraction was highfor each element of both the standardized andtraditional operative reports (Table II). Similarly,analyzing overall agreement with the logistic ran-dom effect model demonstrated excellent interob-server agreement for both the non–standardized(0.9809, 95% confidence interval [CI]0.9734–0.9862) and the standardized (0.9972, 95%CI 0.9907–0.9991 operative reports). However,even at this high level of agreement with both

Table I. Completeness of data from standardizedoperative reports (SOR) vs non–standardizedoperative reports (NSOR)

Operative detailSOR

(n � 119)NSOR

(n � 102)P

value

I. Identifying dataPatient name 119 (100%) 102 (100%) 1.00Patient age 112 (94%) 91 (89%) .18Patient sex 119 (100%) 102 (100%) 1.00Surgeon 119 (100%) 102 (100%) 1.00Anesthetist 81 (68%) 63 (61%) .33Operative date 118 (99%) 101 (99%) .91

II. Perioperative dataIndication for

surgery119 (100%) 102 (100%) 1.00

Operative urgency 118 (99%) 65 (64%) �.001Inpatient status 116 (97%) 50 (49%) �.001Antibiotics 113 (95%) 15 (15%) �.001DVT prophylaxis 116 (97%) 14 (14%) �.001

III. Qualitativeoperative details

Abdominal entrytechnique

117 (98%) 102 (100%) .29

Gallbladderappearance

119 (100%) 36 (35%) �.001

Conversion toopen

119 (100%) 102 (100%) 1.00

Cholangiogram 119 (100%) 102 (100%) 1.00Use of retrieval

bag75 (63%) 59 (58%) .43

Bile spillage 119 (100%) 60 (59%) �.001Gallstone spillage 119 (100%) 57 (54%) �.001Fascial closure 119 (100%) 100 (98%) .212

IV. Quantitativeoperative details

No. clips on cysticduct

115 (97%) 75 (74%) �.001

No. clips on cysticartery

118 (99%) 72 (71%) �.001

Trocar size 69 (58%) 57 (56%) .75

DVT, Deep vein thrombosis.

strategies, the interobserver agreement was better

with the SOR when compared with the NSOR (P �.0001).

DISCUSSIONAccurate and complete operative data is a criti-

cal element in both clinical research and qualityassurance in surgery. However, traditional methodsof documentation have been criticized for inade-quately reporting key aspects of the surgical proce-dure. Warsi et al6 analyzed 3 major oncologydatabases and found that operative details weremissing in up to 89% of entries. Furthermore, datalabeled as part of “process of care,” such as surgicaldata, was an independent predictor of missing data(odds ratio, 11.3; P � .0001). Similarly, Edhemovicet al5 used a rectal cancer template to extract datafrom narrative operative reports and found that amean of 54.1% of important perioperative datawere not reported. In fact, the most importantaspects of the laparotomy and tumor resectionwere clearly reported in only 33.5% to 47.5% of thedictated reports.

To ameliorate this documentation deficit, oper-ative report templates using standardized dataforms and computerized databases have been pro-posed.5,19-21 Christiaens et al20 implemented a stan-dard data collection form for breast cancer surgeryat 7 centers that were participating in the EuropeanOrganization for Research and Treatment of Can-cer (EORTC) breast cancer trials. Despite previousparticipation in several multicenter randomized tri-als, the new operative data collection form revealedthat surgeons at each center varied substantively inthe length of incision, width of tumor excision,timing and duration of procedure, and antibioticprophylaxis. Clearly, this unrecognized variabilityin surgical technique may confound important out-comes and compromise external validity in studiesof adjuvant therapy.19

Similarly, Scherer et al21 compared informationgathered from standard research data collectionforms and narrative operative reports from 123 pa-tients enrolled in the Ischemic Optic NeuropathyDecompression Trial (IONDT). Identifying data suchas age and name were equally well reported andextracted from both documents. The same was truefor most categorical data (eg, type of anesthesia, in-cision). However, agreement between the 2 reportswas less consistent for open-ended variables. Periop-erative medications (34% agreement) and time ofsurgery (23% agreement) were more consistently ex-tracted from the standard form than from the oper-ative report. In contrast, intraoperative complicationswere more completely documented on narrative op-

erative reports. This study suggests that a standard-

712 Harvey et al SurgeryJune 2007

ized data format is superior in recording most of thepertinent operative data. However, these data suggestthat traditional narrative operative reports have someadvantages, particularly in documenting unusual orcomplicated cases.

While standardized operative data sheets anddatabases have been used successfully in trials, theiracceptance in broad clinical use is uncertain. Arecent study analyzing the implementation of com-puter-based medical documentation revealed thatthese systems were partly or wholly rejected by prac-ticing physicians in all 3 centers studied.22 Keyreasons for physician resistance to these new tech-nologies were disruption of the physicians’ currentpractice patterns and increased physician workloadrelated to data entry. In a related editorial, Zint-ner23 commented that implementation of suchdocumentation strategies should be responsive to

Table II. Interobserver agreement between 2 inde

Component

SO

Kappa ASE

Identifying dataDate of procedure 1.0000 0.0000Surgeon 1.0000 0.0000Assistant 1.0000 0.0000Anesthetist 1.0000 0.0000Patient age 1.0000 0.0000Sex 1.0000 0.0000

Perioperative dataUrgency of procedure 0.9627 0.0255Inpatient status 1.0000 0.0000Indication for surgery 1.0000 0.0000Antibiotics 1.0000 0.0000DVT prophylaxis 1.0000 0.0000

Qualitative operative dataAbdominal entry technique 1.0000 0.0000Conversion to open 1.0000 0.0000Cystic duct ligation technique 1.0000 0.0000Cholangiogram 1.0000 0.0000Indication for cholangiogram 1.0000 0.0000Findings of cholangiogram 0.9505 0.0471Fascial closure method 1.0000 0.0000Gallbladder appearance 1.0000 0.0000Bile spillage 1.0000 0.0000Gallstone spillage 1.0000 0.0000

Quantitative operative dataNo. clips on cystic duct 1.0000 0.0000No. clips on cystic artery 0.9836 0.0162Gallstone retrieval 1.0000 0.0000

Control dataTrocar sizes 0.9667 0.0230Retrieval bag use 0.5952 0.0492

ASE, asymptotic standard error; CI, confidence interval; DVT, deep veoperative reports.

physicians in order to be successful.

We developed a standardized operative sum-mary to augment current narrative reports usingthe model of the synoptic report adopted by pa-thologists.24-27 This model maintains the flexibilityof the narrative report while facilitating more com-plete and consistent operative data reporting in theform of summative information at the end of eachreport. The goal was to produce more completeand consistent reporting of important operativefeatures. This should be contrasted with the goalsof time and effort reduction central to the produc-tion of more common, inflexible operative tem-plates. We chose this model after preliminarymeetings with participating surgeons establishedthat current dictation systems were more accept-able than either standardized data sheets or com-puterized data entry systems.

In this study, data collected from the SOR were

ent data extractors

NSOR

95% CI Kappa ASE 95% CI

000, 1.0000 1.0000 0.0000 1.0000, 1.0000000, 1.0000 1.0000 0.0000 1.0000, 1.0000000, 1.0000 1.0000 0.0000 1.0000, 1.0000000, 1.0000 1.0000 0.0000 1.0000, 1.0000000, 1.0000 1.0000 0.0000 1.0000, 1.0000000, 1.0000 1.0000 0.0000 1.0000, 1.0000

127, 1.0000 0.8814 0.0404 0.8023, 0.9606000, 1.0000 0.8128 0.0504 0.7140, 0.9116000, 1.0000 0.9862 0.0138 0.9592, 1.0000000, 1.0000 1.0000 0.0000 1.0000, 1.0000000, 1.0000 0.9600 0.0395 0.8825, 1.0000

000, 1.0000 1.0000 0.0000 1.0000, 1.0000000, 1.0000 1.0000 0.0000 1.0000, 1.0000000, 1.0000 1.0000 0.0000 1.0000, 1.0000000, 1.0000 1.0000 0.0000 1.0000, 1.0000000, 1.0000 0.9674 0.0316 0.9055, 1.0000583, 1.0000 1.0000 0.0000 (1.0000, 1.0000)000, 1.0000 1.0000 0.0000 (1.0000, 1.0000)000, 1.0000 0.9491 0.0290 0.8923, 1.0000000, 1.0000 0.9556 0.0252 0.9061, 1.0000000, 1.0000 0.9671 0.0231 0.9219, 1.0000

000, 1.0000 0.9664 0.0235 0.9203, 1.0000518, 1.0000 0.9672 0.0229 0.9224, 1.0000000, 1.0000 1.0000 0.0000 1.0000, 1.0000

217, 1.0000 1.0000 0.0000 1.0000, 1.0000989, 0.6916 0.6537 0.0537 0.5486, 0.7589

mbosis; NSOR, non–standardized operative report; SOR, Standardized

pend

R

1.01.01.01.01.01.0

0.91.01.01.01.0

1.01.01.01.01.00.81.01.01.01.0

1.00.91.0

0.90.4

in thro

superior to data from the NSOR in completeness.

Surgery Harvey et al 713Volume 141, Number 6

Identifying data variables (eg, patient name, gen-der, surgeon, etc.) were similarly reported in bothreports. At our institution, identifying data are al-ready presented in a standardized format at thebeginning of the narrative operative report. Thisstudy confirms previous findings that many impor-tant operative details are consistently omitted fromtraditional narrative reports. These omissions in-clude perioperative factors, as well as quantitativeand qualitative operative details.

After the initial 3-month introductory phase ofthe SOR, participating surgeons identified 2 addi-tional elements to be added to the template: theuse of a gallbladder retrieval bag and the size andposition of trocars used for the procedure. Whilethese variables were not on the original template,we extracted these data from reports in the SORand NSOR groups. These operative details weresimilarly reported in the SOR and NSOR groups.This finding suggests that the observed advantagesin data completeness in the SOR were a directresult of the standardized template and not due toimproved dictating practices related to SOR imple-mentation. From this study, it is apparent that thedevelopment of these documents is an ongoingprocess. The flexibility provided by this model al-lows additional features to be identified by sur-geons and added to the summative paragraph asneeded. For example, operative features related toachieving the “critical view of safety” related to thetriangle of Calot have been included in subsequentiterations of this standardized report.

Compliance with the standardized dictation was63% over the entire study period. Compliance im-proved with increasing familiarity with the study,and rose from 53% in the first month to 67% in thefollowing two months. In addition, compliance wasobserved to be better for surgeons doing a highervolume of laparoscopic cholecystectomy proce-dures. Compliance was 75% for the 4 surgeonscompleting 20 or more procedures during the 3months and just 30% for the 5 surgeons complet-ing less than 20. Interestingly, the majority of thesereports were completed by residents who work withall of the surgeons at this site. As such, the differ-ences in compliance are difficult to explain. It ispossible that bias may be introduced by this vari-able compliance. For example, procedures thatwere particularly challenging or performed late atnight may be less likely to be dictated in the stan-dard format. Further study of these factors is nec-essary. Identification and modification of thesefactors may be used to improve implementation ofthis operative documentation strategy. Moreover,

there may be an opportunity to improve compli-

ance by encouraging lower volume surgeons toadopt the new documentation paradigm.

Interobserver reliability was found to be superiorwith the SOR than the NSOR (0.9972 vs 0.9808,P � .0001). While this observed difference in reli-ability is small, it is important to consider that thisoverall indicator of agreement is influenced by anumber of clinically meaningful differences in in-dividual operative features including procedure ur-gency, inpatient status, gallbladder appearance,and bile spillage. In addition, because both dataextractors were surgeons, their ability to interpretthe narrative reports is clearly superior to dataextractors who would be employed for data collec-tion for research or quality assurance purposes.The advantage of the SOR with regard to interob-server agreement is likely to be amplified in circum-stances where non-surgeons are extracting thesedata.

The laparoscopic cholecystectomy was chosenfor study because of its high volume and relativeconsistency in surgical approach among surgeons.The application of this system to more complex orvariable surgical procedures will be more challeng-ing. However, the flexibility of the combined nar-rative and standardized report is ideally suited toadoption for more complex surgical proceduresthan the more rigidly structured template methods.

In summary, we have developed a method ofstandardizing the reporting of an operative proce-dure that maintains the flexibility of the narrativereport, enhances the completeness and extractabil-ity of key elements of the procedure, and minimallyimpacts current surgeon practices. This method isan alternative to computer-based operative reporttemplates and standardized data collection formsthat may be more acceptable to surgeons in practice.

The authors would like to thank Drs F. Sutherland,J. Heine, D. Johnson, W. Rosen, B. Rothwell, G. Hollar, P.Mitchell, M. Nutley, M. Raval, D. Fenech, M. Dunham,and S. Mackenzie for their participation in this study.

REFERENCES1. Stanley-Brown EG. How to dictate operative notes. Resid

Staff Physician 1983;29:109-10.2. Khuri SF. The NSQIP: a new frontier in surgery. Surgery

2005;138:837-43.3. Khuri SF, et al. The Department of Veterans Affairs’ NSQIP:

the first national, validated, outcome-based, risk-adjusted,and peer-controlled program for the measurement and en-hancement of the quality of surgical care. National VASurgical Quality Improvement Program. Ann Surg 1998;228:491-507.

4. Birkmeyer JD, et al. Surgeon volume and operative mortality

in the United States. N Engl J Med 2003;349:2117-27.

714 Harvey et al SurgeryJune 2007

5. Edhemovic I, Temple WJ, de Gara CJ, Stuart GC. Thecomputer synoptic operative report—a leap forward in thescience of surgery. Ann Surg Oncol 2004;11:941-7.

6. Warsi AA, White S, McCulloch P. Completeness of dataentry in three cancer surgery databases. Eur J Surg Oncol2002;28:850-6.

7. van Walraven C, Laupacis A, Seth R, Wells G. Dictated versusdatabase-generated discharge summaries: a randomizedclinical trial. CMAJ 1999;160:319-26.

8. Grogan EL, et al. Improving documentation of patient acu-ity level using a progress note template. J Am Coll Surg2004;199:468-75.

9. Marill KA, et al. Prospective, randomized trial of template-assisted versus undirected written recording of physicianrecords in the emergency department. Ann Emerg Med1999;33:500-9.

10. DeOrio JK. Surgical templates for orthopedic operative re-ports. Orthopedics 2002;25:639-42.

11. Gadacz TR. U.S. experience with laparoscopic cholecystec-tomy. Am J Surg 1993;165:450-4.

12. Hannan EL, Imperato PJ, Nenner RP, Starr H. Laparo-scopic and open cholecystectomy in New York State: mor-tality, complications, and choice of procedure. Surgery1999;125:223-31.

13. Kumar V, Pande GK. Complications of cholecystectomy in theera of laparoscopic surgery. Trop Gastroenterol 2001;22:72-9.

14. Lau H, Brooks DC. Contemporary outcomes of ambulatorylaparoscopic cholecystectomy in a major teaching hospital.World J Surg 2002;26:1117-21.

15. Matthews BD, Williams GB. Laparoscopic cholecystectomyin an academic hospital: evaluation of changes in perioper-ative outcomes. JSLS 1999;3:9-17.

16. Shea JA, et al. Mortality and complications associated withlaparoscopic cholecystectomy. A meta-analysis. Ann Surg

1996;224:609-20.

17. Shea JA, et al. Indications for and outcomes of cholecystec-tomy: a comparison of the pre and postlaparoscopic eras.Ann Surg 1998;227:343-50.

18. Zacks SL, Sandler RS, Rutledge R, Brown RS Jr. A popula-tion-based cohort study comparing laparoscopic cholecys-tectomy and open cholecystectomy. Am J Gastroenterol2002;97:334-40.

19. Christiaens MR, van der SE, Vantongelen K. More detaileddocumentation of operative procedures in breast conserv-ing treatment: what good will it do us? Eur J Surg Oncol1996;22:326-30.

20. Christiaens MR, et al. Comparison of the surgical proce-dures for breast conserving treatment of early breast cancerin seven EORTC centres. Eur J Cancer 1996;32A:1866-75.

21. Scherer R, et al. Comparison of information obtained byoperative note abstraction with that recorded on a standard-ized data collection form. Surgery 2003;133:324-30.

22. Lapointe L, Rivard S. Getting physicians to accept newinformation technology: insights from case studies. CMAJ2006;174:1573-8.

23. Zitner D. Physicians will happily adopt information technol-ogy. CMAJ 2006;174:1583-4.

24. Srigley JR. Key issues in handling and reporting radicalprostatectomy specimens. Arch Pathol Lab Med 2006;130:303-17.

25. Zhang P, Brooks JS. Modern pathological evaluation ofsoft tissue sarcoma specimens and its potential role in softtissue sarcoma research. Curr Treat Options Oncol 2004;5:441-50.

26. Thompson JF, Scolyer RA. Cooperation between surgicaloncologists and pathologists: a key element of multidisci-plinary care for patients with cancer. Pathology 2004;36:496-503.

27. Hammond EH, Flinner RL. Clinically relevant breast cancerreporting: using process measures to improve anatomic pa-

thology reporting. Arch Pathol Lab Med 1997;121:1171-5.