The American College of Surgeons/Association of Program Directors in Surgery National Skills Curriculum: adoption rate, challenges and strategies for effective implementation into surgical residency programs.

BACKGROUND: The American College of Surgeons/Association of Program Directors in Surgery (ACS/APDS) National Skills Curriculum is a 3-phase program targeting technical and nontechnical skills development. Few data exist regarding the adoption of this curriculum by surgical residencies. This study attempted to determine the rate of uptake and identify implementation enablers/barriers. METHODS: A web-based survey was developed by an international expert panel of surgical educators (5 surgeons and 1 psychologist). After piloting, the survey was sent to all general surgery program directors via email link. Descriptive statistics were used to determine the residency program characteristics and perceptions of the curriculum. Implementation rates for each phase and module were calculated. Adoption barriers were identified quantitatively and qualitatively using free text responses. Standardized qualitative methodology of emergent theme analysis was used to identify strategies for success and details of support required for implementation. RESULTS: Of the 238 program directors approached, 117 (49%) responded to the survey. Twenty-one percent (25/117) were unaware of the ACS/APDS curriculum. Implementation rates for were 36% for phase I, 19% for phase II, and 16% for phase III. The most common modules adopted were the suturing, knot-tying, and chest tube modules of phase I. Over 50% of respondents identified lack of faculty protected time, limited personnel, significant costs, and resident work-hour restrictions as major obstacles to implementation. Strategies for effective uptake included faculty incentives, adequate funding, administrative support, and dedicated time and resources. CONCLUSION: Despite the availability of a comprehensive curriculum, its diffusion into general surgery residency programs remains low. Obstacles related to successful implementation include personnel, learner, and administrative issues. Addressing these issues may improve the adoption rate of the curriculum.

Characterization of resident surgeon participation during carotid endarterectomy and impact on perioperative outcomes.

INTRODUCTION: The impact of resident surgeon participation during vascular procedures on postoperative outcomes is incompletely understood. We characterized resident physician participation during carotid endarterectomy (CEA) procedures within the 2005-2009 American College of Surgeons National Surgical Quality Improvement Participant Use Datafile and evaluated associations with procedural characteristics and perioperative adverse events. METHODS: CEAs were identified using primary current procedural terminology codes; those performed simultaneously with other major procedures or unknown resident participation status were excluded. Group-wise comparisons based on resident participation status were performed using χ(2) or Fisher's exact test for categorical variables and t tests or nonparametric methods for continuous variables. Associations with perioperative adverse events (major = stroke, death, myocardial infarction, or cardiac arrest; minor = peripheral nerve injury, bleeding requiring transfusion, surgical site infection, or wound disruption) were assessed using multivariable logistic regression models adjusting for other known risk factors. RESULTS: A total of 25,280 CEA procedures were analyzed, of which residents participated in 13,705 (54.2%), while residents were absent in 11,575 (45.8%). Among CEAs with resident physician participation, resident level was categorized as junior (postgraduate year [PGY] 1-2) in 21.9%, senior (PGY 3-5) in 52.7%, and fellow (PGY ≥6) in 25.3%. Major adverse event rates with and without resident participation were 1.9% versus 2.1%, and minor adverse event rates with and without resident participation were 0.9% versus 1.0%, respectively. In multivariable models, resident physician participation was not associated with perioperative risk for major adverse events (odds ratio [OR], 0.90; 95% confidence interval [CI], 0.75-1.08) or minor adverse events (OR, 0.93; 95% CI, 0.72-1.21). CONCLUSIONS: Resident surgeon participation during CEA is not associated with risk of adverse perioperative events.

Positive predictive value of the AHRQ Patient Safety Indicator "postoperative wound dehiscence".

BACKGROUND: The Agency for Healthcare Research and Quality patient safety indicator (PSI) 14, or "postoperative wound dehiscence," is 1 of 4 PSIs recently adopted by the Centers for Medicare & Medicaid Services to compare quality and safety across hospitals. We determined how well it identifies true cases of postoperative wound dehiscence by examining its positive predictive value (PPV). STUDY DESIGN: A retrospective cross-sectional study of hospitalization records that met PSI 14 criteria was conducted within the Veterans Health Administration hospitals from fiscal years 2003 to 2007. Trained abstractors used standardized abstraction instruments to review electronic medical records. We determined the PPV of the indicator and performed descriptive analyses of cases. RESULTS: Of the 112 reviewed cases, 97 were true events of postoperative wound dehiscence, yielding a PPV of 87% (95% CI 79% to 92%). Sixty-one percent (n = 59) of true positive cases had at least 1 risk factor, such as low albumin level, COPD, or superficial wound infection. False positives were due to coding errors, such as cases in which the patient's abdomen was intentionally left open during the index procedure. CONCLUSIONS: PSI 14 has relatively good predictive ability to identify true cases of postoperative wound dehiscence. It has the highest PPV among all PSIs evaluated within the Veterans Health Administration system. Inaccurate coding was the reason for false positives. Providing additional training to medical coders could potentially improve the PPV of this indicator. At present, this PSI is a promising measure for both quality improvement and performance measurement; however, its use in pay-for-performance efforts seems premature.

Design and development of a surgical skills simulation curriculum.

BACKGROUND: A number of significant changes have forced surgical educators to re-evaluate the adequacy of traditional forms of surgical skills training. MATERIALS: A review of the literature reveals that surgical simulation has emerged as a useful adjunct to help educators adjust to the demands of an ever-changing surgical practice environment. As such, integration of simulation technology into a busy surgical training program has now become a priority for training programs worldwide. RESULTS: Successful integration requires a disciplined and dedicated approach to the appropriate use of all forms of available simulation in a well-designed curriculum. CONCLUSION: This manuscript provides a discussion of how this can be achieved using a sequential, modular, criterion-based framework, providing details of the rationale behind such an approach and current examples of how it can be integrated.

Proficiency-based virtual reality training significantly reduces the error rate for residents during their first 10 laparoscopic cholecystectomies.

BACKGROUND: Virtual reality (VR) training has been shown previously to improve intraoperative performance during part of a laparoscopic cholecystectomy. The aim of this study was to assess the effect of proficiency-based VR training on the outcome of the first 10 entire cholecystectomies performed by novices. METHODS: Thirteen laparoscopically inexperienced residents were randomized to either (1) VR training until a predefined expert level of performance was reached, or (2) the control group. Videotapes of each resident's first 10 procedures were reviewed independently in a blinded fashion and scored for predefined errors. RESULTS: The VR-trained group consistently made significantly fewer errors (P = .0037). On the other hand, residents in the control group made, on average, 3 times as many errors and used 58% longer surgical time. CONCLUSIONS: The results of this study show that training on the VR simulator to a level of proficiency significantly improves intraoperative performance during a resident's first 10 laparoscopic cholecystectomies.

A comparison between fluoroscopically guided radiofrequency energy delivery and conventional technique in an animal model of fundoplication failure.

BACKGROUND: The use of radiofrequency energy (RFe) treatment at the gastroesophageal junction (GEJ) has been considered an alternative to surgery after fundoplication disruption. It is unknown whether the recommended delivery technique for primary gastroesophageal reflux disease applies to an anatomically altered GEJ following fundoplication. The aim of this study was to determine whether modifications to the standard technique using fluoroscopic guidance more accurately localizes ablation zones compared with standard technique alone. METHODS: Ten pigs were randomized to either conventional or fluoroscopically guided RFe ablation. All pigs had a laparoscopic Nissen fundoplication that was subsequently disrupted by severing all but the most cranial fundoplication stitch. Conventional RFe delivery included usage of markers located on the Stretta catheter. After labeling the z-line via submucosal contrast injection, fluoroscopic guidance involved using fluoroscopic markers to guide RFe ablation. Ablations were acutely marked, measured, and agreed upon by a panel of three researchers analyzing harvested tissue. Distances from the target zone for each ablation line (e.g., 1 cm was the target zone for line 1) were calculated and analyzed using Mann-Whitney and Fischer's tests. RESULTS: Fluoroscopic guidance was significantly more accurate than the conventional technique (0.2 +/- 0.2 cm vs. 1.8 +/- 0.8 cm, p < 0.0001). Analyzing the individual distances for each of the six ablation lines revealed that all within Group B were closer than Group A (p < 0.01 for all except lines 1 and 2). Overall, the total ablation treatment length for conventionally treated animals was 4.48 +/- 0.7 cm and for those who underwent fluoroscopic guidance it was 2.92 +/- 0.5 cm (p < 0.001). CONCLUSION: In a porcine model of fundoplication disruption, fluoroscopic guidance improved RFe accuracy.

Delayed gastric emptying in patients undergoing antireflux surgery: analysis of a treatment algorithm.

PURPOSE: Controversy remains about the treatment of patients with mild delayed gastric emptying (90 min < emptying half-time [T(1/2)] < 180 min) who undergo antireflux surgery. This retrospective, nonrandomized study reviewed the records of patients treated from January 1996 through October 2003, during which time we applied two treatment algorithms for patients with mild delayed gastric emptying. The goal of this study was to determine whether the most recent treatment algorithm was effective in reducing the need for a concomitant gastric drainage procedure, pyloroplasty. MATERIALS AND METHODS: Eighteen patients with mild delayed gastric emptying underwent antireflux surgery plus pyloroplasty (group A) before 2001, and 13 patients with mild delayed gastric emptying underwent antireflux surgery plus gastric decompression with percutaneous endoscopic gastrostomy placement (group B) starting in 2001. We reviewed indications for the procedure, complications, and outcomes. Primary outcome measures for this study were recurrence of gastroparesis symptoms and need for pyloroplasty. RESULTS: The average T(1/2) was similar for both groups A and B: 129 min and 123 min, respectively. Eleven of 13 patients (85%) in group B experienced resolution of gastroparesis symptoms, improved gastric emptying times, or both; only 1 patient (8%) underwent subsequent pyloroplasty for treatment failure. Only one serious percutaneous endoscopic gastrostomy-related event occurred (tube migration), and no patients died. Significantly fewer patients in group B required total pyloroplasty (8% vs. 56% in group A; P < 0.008), and significantly fewer required pyloroplasty for symptomatic control (15% vs. 56% in group A; P < 0.03). CONCLUSION: A treatment algorithm incorporating percutaneous endoscopic gastrostomy tube placement at the time of antireflux surgery for gastric decompression successfully managed antireflux surgery patients with mild delayed gastric emptying. This approach allows for a more selective use of pyloroplasty.

Groin hernia: anatomical and surgical history.

The history of surgical repair of groin hernia is a lengthy record of assorted techniques in search of a cure for an ailment that comes in many sizes and shapes and that has plagued humanity for thousands of years. Although improvements are still being sought and found, for several decades surgeons have had the means to relieve most hernia sufferers. A remaining issue is whether the wide array of surgical procedures can or should be whittled down to a few "standard" operations that are safe, effective, and cost-efficient. The history of the anatomy of groin hernia shows how much there was to learn and how much remains to be learned. It also shows how important it is for the surgeon to know and understand both the anatomy of the area and the formation of groin hernia.

Perceptual, visuospatial, and psychomotor abilities correlate with duration of training required on a virtual-reality flexible endoscopy simulator.

BACKGROUND: Trainees acquire endoscopic skills at different rates. Fundamental abilities testing could predict the amount of training required to reach a performance goal on a virtual-reality simulator. METHODS: Eleven medical students were tested for fundamental abilities. Baseline endoscopic proficiency was evaluated with the GI Mentor II VR simulator (Simbionix, USA, Cleveland, OH). Subjects trained on the simulator with a defined performance goal. Subjects who achieved the goal were then reassessed. RESULTS: All subjects completed at least 10 trials or reached the performance goal. The <10 trial group (n=6) tested better for all fundamental abilities and baseline endoscopic performance than the >10 trial group (n=5). The number of trials required to reach the performance goal correlated significantly with both perceptual (r=.92, P=0.001) and visuospatial ability (r=.76, P=.03). Multiple regression showed strong correlation of all three abilities with duration of training (r=.95, P=.015). CONCLUSIONS: Most of the variability in acquisition of endoscopic skills can be accounted for by differences in fundamental abilities of trainees. Testing of fundamental abilities could help identify trainees who will require additional training to achieve desired performance objectives.

Real-time objective assessment of knot quality with a portable tensiometer is superior to execution time for assessment of laparoscopic knot-tying performance.

OBJECTIVE: Laparoscopic intracorporeal knot tying is a difficult skill to acquire. Currently, time to complete a knot is the most commonly used metric to assess the acquisition of this skill; however, without a measure of knot quality, time is a poor indicator of skills mastery. Others have shown that knot quality can be accurately assessed with a tensiometer, but obtaining this type of assessment has typically been cumbersome. We investigated a new method of real-time assessment of knot quality that allows for more practical use of knot quality as a performance metric. METHODS: Eleven experienced endoscopic surgeons tied 100 intracorporeal knots in a standard box trainer. Each of the knots was immediately tested using the InSpec 2200 benchtop tensiometer (INSTRON, Canton MA) where a knot quality score (KQS) is generated based on the load handling properties of the knotted suture. The execution time was also recorded for each knot. RESULTS: The assessment of all knots ended with one of two end points: knots that slipped (n=48) or knots that held until the suture broke (n=52). Knots that slip are generally of poorer quality than those that held. Execution time did not correlate with knot-quality score (r=0.009, P=.9), and the mean execution time did not differ significantly between slipped and held knots (65 vs 68 seconds, P=.8). No completion time criteria were able to accurately predict slipped versus held knots. The mean KQS difference between held and slipped knots was highly significant (24 vs 12, P<.0001). A knot with a KQS exceeding 20 was nearly 10 times more likely to hold than slip. CONCLUSION: Time to complete a knot is a poor metric for the objective assessment of intracorporeal knot-tying performance in the absence of a measure of knot quality. Real-time evaluation of the knot quality can accurately distinguish well-tied knots from poorly tied knots. This mode of assessment should be incorporated into training curriculum for surgical knot tying.

Relationship between tissue ingrowth and mesh contraction.

Contraction is a well-documented phenomenon occurring within two months of mesh implantation. Its etiology is unknown, but it is suggested to occur as a result of inadequate tissue ingrowth into the mesh and has been associated with hernia recurrence. In continuation of our previous studies, we compared tissue ingrowth characteristics of large patches of polyester (PE) and heavyweight polypropylene (PP) and their effect on mesh contraction. The materials used were eight PE and eight PP meshes measuring 10 x 10 cm2. After random assignment to the implantation sites, the meshes were fixed to the abdominal wall fascia of swine using interrupted polypropylene sutures. A necropsy was performed three months after surgery for evaluation of mesh contraction/shrinkage. Using a tensiometer, tissue ingrowth was assessed by measuring the force necessary to detach the mesh from the fascia. Histologic analysis included inflammatory and fibroblastic reactions, scored on a 0-4 point scale. One swine developed a severe wound infection that involved two PP meshes and was therefore excluded from the study. The mean area covered by the PE meshes (87 +/- 7 cm2) was significantly larger than the area covered by the PP meshes (67 +/- 14 cm2) (p = 0.006). Tissue ingrowth force of the PE meshes (194 +/- 37 N) had a trend toward being higher than that of the PP meshes (159 +/- 43 N), although it did not reach statistical significance. There was no difference in histologic inflammatory and fibroblastic reactions between mesh types. There was a significant correlation between tissue ingrowth force and mesh size (p = 0.03, 95% CI: 0.05-0.84). Our results confirm those from previous studies in that mesh materials undergo significant contraction after suture fixation to the fascia. PE resulted in less contraction than polypropylene. A strong integration of the mesh into the tissue helps prevent this phenomenon, which is evidenced by a significant correlation between tissue ingrowth force and mesh size.

When fundoplication fails: redo?

OBJECTIVE: The largest series in the literature dealing with redo fundoplication was presented and published in 1999 and included 100 patients. Herein we update this initial series of 100, with 207 additional patients who have undergone redo fundoplication (n = 307). SUMMARY BACKGROUND DATA: Increasing numbers of patients are failing esophagogastric fundoplication and requiring redo procedures. Data regarding the nature of these failures have been scant. METHODS: Data on all patients undergoing foregut surgery are collected prospectively. Between 1991 and 2004, 307 patients underwent redo fundoplication for the management of anatomic complications or recurrent GERD. Statistical analysis was performed with multiple chi2 and Mann-Whitney U analyses, as well as ANOVA. RESULTS: Between 1991 and 2004, 1892 patients underwent primary fundoplication for GERD (1734) or paraesophageal hernia (158). Of these, 54 required redo fundoplication (2.8%). The majority of failures (73%) were managed within 2 years of the initial operation (P = 0.0001). The mechanism of failure was transdiaphragmatic wrap herniation in 33 of 54 (61%). In the 231 patients who underwent fundoplication elsewhere, 109 had transdiaphragmatic herniation (47%, P = NS). In this group of 285 patients, 22 (8%) required another redo (P = NS). The majority of the procedures were initiated laparoscopically (240/307, 78%), with 20 converted (8%). Overall mortality was 0.3%. CONCLUSIONS: Failure of fundoplication is unusual in experienced hands. Most are managed within 2 years of the initial operation. Wrap herniation has now become the most common mechanism of failure requiring redo. Redo fundoplication was successful in 93% of patients, and most could be safely handled laparoscopically.

Video-assisted surgery represents more than a loss of three-dimensional vision.

BACKGROUND: Loss of depth cues is a major challenge facing surgeons performing video-assisted surgery (VAS). Whether the degradation of image quality from a video-displayed image plays a direct role in performance of VAS has not been studied. METHODS: Twenty-four volunteer novice subjects were randomized to binocular direct-vision (BDV), monocular direct-vision (MDV), or video-imaging (VI) conditions. Each subject completed ten trials of a simple cutting task in a box trainer using standard laparoscopic instruments. RESULTS: VI subjects made significantly fewer correct incisions than both of the other groups for all trials. Differences between the BDV and MDV groups did not reach statistical significance. Improvement in performance was more rapid in the BDV group than in either the MDV or VI groups. CONCLUSIONS: The degradation of image quality with VI has a detrimental influence on VAS performance above and beyond the loss of binocular vision.

The history of anatomy and surgery of the preperitoneal space.

Preperitoneal (properitoneal) space is the space between the peritoneum and transversalis fascia. Bogros (1786-1825) described a triangular space in the iliac region between the iliac fascia, transversalis fascia, and parietal peritoneum. In the modern concept, this space lies between the peritoneum and posterior lamina of the transversalis fascia. In 1858, Retzius described the homonymous space, situated anterior and lateral to the urinary bladder (prevesical space). In 1975, Fowler reported that the preperitoneal fascia of the groin is distinct from the transversalis fascia. Preperitoneal herniorrhaphy may be subdivided into 2 approaches: transperitoneal and inguinal. We present herein the evolution of approaches to the preperitoneal space from use of the transperitoneal (or posterior) to use of the anterior preperitoneal and posterior preperitoneal approaches. As anatomic knowledge has increased, the evolution of laparoscopic surgery has paralleled that of open procedures.

Evaluation of a tensiometer to provide objective feedback in knot-tying performance.

Training and assessment methods for knot tying by medical students or residents have traditionally been subjective. Objective methods for evaluating creation of a tied knot should include assessing the strength and quality of the knotted suture. The purpose of this study was to evaluate the use of a tensiometer as a feedback device for improving knot-tying performance. Twelve medical students with no knot-tying experience were selected to perform three-throw instrument ties with 00 silk suture. Students were randomly assigned to perform between 10 and 20 baseline knots and then received one of four feedback training conditions followed by 10 completion knots. Subjects were timed, and all knots were pulled in the tensiometer to assess for strength and slippage. Differences between baseline and completed knots for each subject were analyzed with an unpaired t test. Subjects receiving both subjective and tensiometer feedback demonstrated the greatest improvements in knot quality score (KQS) and slip percentage (Subject 1: 0.15 +/- 0.9 vs 0.21 +/- 0.05, P < 0.04, 75% vs 60%, P = NS; Subject 2: 0.22 +/- 0.10 vs 0.29 +/- 0.05, P < 0.02, 33% vs 0%, P < 0.05; Subject 3: 0.10 +/- 0.07 vs 0.25 +/- 0.07, P < 0.0001, 60% vs 10%, P < 0.01). Objective assessment of knot-tying performance is possible using the tensiometer device. Introduction of the tensiometer during the learning phase produced improved KQS and slip percentage in most students regardless of the number of baseline knots tied. Greatest improvements in performance were seen when the tensiometer was used in combination with subjective instruction.

Laparoscopic approach reduces likelihood of perioperative complications in patients undergoing adrenalectomy.

Laparoscopy has been reported advantageous over the conventional open technique for adrenalectomy. However, most comparative series include the relatively more challenging cases in the open group. The aim of this study is to assess the actual role of laparoscopy in reducing perioperative complications compared to open surgery in patients undergoing adrenalectomy. Between January 1992 and December 2002, we performed 148 adrenalectomies in 138 patients. Depending on the approach, patients were divided into laparoscopic (LA) or open adrenalectomy (OA) groups. Demographics, tumor characteristics, operative data, and outcomes were analyzed. Linear and logistic regressions identified factors influencing perioperative outcomes. Multivariate-adjusted logistic regression assessed independent relationship between factors and perioperative outcomes. A total of 78 cases were performed laparoscopically and 70 open. Patients were matched for age and sex. Tumor size was smaller (3 +/- 2 vs 5 +/- 3 cm), operative time was shorter (133 +/- 65 vs 165 +/- 100 min), estimated blood loss was less (114 +/- 152 vs 350 +/- 417 cc), length of stay was shorter (3 +/- 2 vs 7 +/- 3 days), and overall complication rate was lower (7% vs 20%) in the LA compared to the OA group. The incidence of cancer in tumors > or = 6 cm (31%) was higher than in those < 6 cm (4%). All patients with cancer underwent OA. LA was the only factor independently associated with a decreased likelihood of intraoperative bleeding and postoperative pulmonary complications. Large and malignant adrenal tumors are more frequently removed through an open approach. However, this fact has no influence on the advantages of the LA over the OA. Laparoscopy reduces perioperative adrenalectomy perioperative complication rates. It has a positive impact on intraoperative bleeding and postoperative pulmonary complications.

The effects of gastric surgery on systemic ghrelin levels in the morbidly obese.

HYPOTHESIS: Circulating ghrelin, produced primarily in the stomach, is a powerful orexigen. Ghrelin levels are elevated in states of hunger, but rapidly decline postprandially. Early alterations in ghrelin levels in morbidly obese patients undergoing weight reduction surgery may be attributed to gastric partitioning. DESIGN AND PATIENTS: Thirty-four patients underwent Roux-en-Y gastric bypass with a completely divided gastroplasty to create a 15-mL vertically oriented gastric pouch. Eight other patients underwent other gastric procedures that did not involve complete division of the stomach, including 4 vertical banded gastroplasties and 4 antireflux surgical procedures. Six additional patients undergoing antireflux surgery served as lean control subjects. Plasma samples were obtained before surgery and immediately after surgery. In a substudy, plasma was collected after Roux-en-Y limb formation and after dividing the stomach to identify any changes in plasma ghrelin levels. SETTING: Tertiary university medical center. MAIN OUTCOME MEASURES: Ghrelin levels at different stages of surgical intervention. RESULTS: Mean +/- SEM preoperative and postoperative ghrelin levels in the gastric bypass group were 355 +/- 20 and 246 +/- 13 pg/mL, respectively (P<.001). In the vertical banded gastroplasty group and in all patients undergoing antireflux surgery, ghrelin levels were not significantly changed. CONCLUSIONS: Compared with morbidly obese humans, lean controls had significantly higher plasma ghrelin levels at baseline. A divided gastroplasty creating a small proximal gastric pouch results in significant early declines in circulating ghrelin levels that are not observed with other gastric procedures. This may explain, in part, the loss of hunger sensation and rapid weight loss observed following gastric bypass surgery.

Objective psychomotor skills assessment of experienced and novice flexible endoscopists with a virtual reality simulator.

The objective of this study was to determine whether the GI Mentor II virtual reality simulator can distinguish the psychomotor skills of intermediately experienced endoscopists from those of novices, and do so with a high level of consistency and reliability. A total of five intermediate and nine novice endoscopists were evaluated using the EndoBubble abstract psychomotor task. Each subject performed three repetitions of the task. Performance and error data were recorded for each trial. The intermediate group performed better than the novice group in each trial. The differences were significant in trial 1 for balloons popped (P=.001), completion time (P=.04), and errors (P=.03). Trial 2 showed significance only for balloons popped (P=.002). Trial 3 showed significance for balloons popped (P=.004) and errors (P=.008). The novice group showed significant improvement between trials 1 and 3 (P<0.05). No improvement was noted in the intermediate group. Measures of consistency and reliability were greater than 0.8 in both groups with the exception of novice completion time where test-retest reliability was 0.74. The GI Mentor II simulator can distinguish between novice and intermediate endoscopists. The simulator assesses skills with levels of consistency and reliability required for high-stakes assessment.