"Bending the cost curve" in gastroenterology.

INTRODUCTION: Increasing attention is being focused on reigning in escalating costs of healthcare, i.e. trying to 'bend the cost curve'. In gastroenterology (GI), inpatient hospital care represents a major component of overall costs. This study aimed to characterize the trend in cost of care for GI-related hospitalizations in recent years and to identify the most costly diagnostic groups. METHODS: All hospital inpatients admitted between January 2008 and December 2009 with a primary diagnosis of one of the six most common GI-related Diagnosis Related Groups (DRGs) in this hospital system were identified; all DRGs contained at least 40 patients during the study period. Patient Level Costing (PLC) was used to express the total cost of hospital care for each patient; PLC comprised a weighted daily bed cost plus cost of all medical services provided (e.g., radiology, pathology tests) calculated according to an activity-based costing approach; cost of medications were excluded. All costs were discounted to 2009 values. Mean length of stay (LOS) was also calculated for each DRG. RESULTS: Over 2 years, 470 patients were admitted with one of the six most common GI DRGs. Mean cost of care increased from 2008 to 2009 for all six DRGs with the steepest increases seen in 'GI hemorrhage (non-complex)' (31 % increase) and 'Cirrhosis/Alcoholic hepatitis (non-complex)' (45 % increase). No differences in readmission rates were observed over time. There was a strong correlation between year-to-year change in costs and change in mean LOS, r = 0.93. CONCLUSION: The cost of GI-related inpatient care appears to be increasing in recent years with the steepest increases observed in non-complex GI hemorrhage and non-complex Cirrhosis/Alcoholic hepatitis. Efforts to control the increasing costs should focus on these diagnostic categories.

Does the cost of care differ for patients with fee-for-service vs. capitation of payment? A case-control study in gastroenterology.

INTRODUCTION: There is growing evidence to demonstrate overuse of medical resources in fee for service (FFS) payment models (in which physicians are reimbursed according to volume of care provided) compared to capitation payment models (in which physicians receive a fixed salary regardless of level of care provided). In this medical centre, patients with and without insurance are admitted through the same access point (emergency room) and cared for by the same physicians. Therefore, apart from insurance status, all other variables influencing delivery of care are similar for both patient groups. However, physician reimbursement differs for both groups: FFS for patients with private insurance (i.e. the admitting physician's reimbursement escalates progressively with each day that the patient spends in hospital) and base salary irrespective of care provided for patients with universal insurance (capitation payment model). All admitting physicians are aware of the patient's insurance status and the duration of hospitalization is at the discretion of the admitting physician. This study aimed to compare cost of care of patients with and without insurance admitted to a teaching hospital with a primary gastroenterology or hepatology (GIH) diagnosis. METHODS: All hospital inpatients admitted between January 2008 and December 2009 with a primary GI-related diagnosis related group (DRG) were identified. Patients were classified as uninsured (state-funded) or privately insured. Only DRGs with at least five patients in both the insured and uninsured patient groups were analyzed to ensure a precise estimate of inpatient costs. Patient level costing (PLC) was used to express the total cost of hospital care for each patient; PLC comprised a weighted daily bed cost plus cost of all medical services provided (e.g. radiology, pathology tests) calculated according to an activity-based costing approach, cost of medications were excluded. An overall mean cost of care per patient was calculated for both groups. All costs were discounted to 2009 values. RESULTS: In total, 630 patients were admitted with one of 11 GIH DRGs, 181 (29 %) with private insurance. Pooled mean cost of care was higher for uninsured (6,781 euros/patient) compared to insured patients (6,128 euros/patient). Apart from patients with 'non-cirrhotic non-alcoholic liver disease (non-complex)' in whom mean cost was higher for insured patients, there were no significant differences in mean cost of care nor mean patient age for insured and uninsured groups for any other diagnoses. CONCLUSION: Inpatient hospital costs were equivalent for patients with and without private health insurance when care was provided in a single hospital. Provision of care for all patients in a common hospital setting regardless of health insurance status may reduce disparities in healthcare utilization.

Randomized controlled trial of feedback on cost of hospital care among gastroenterology inpatients.

INTRODUCTION: Spending on hospital inpatients comprises a major proportion of healthcare costs. This study assessed the impact of systematic feedback to gastroenterologists on the cost of care provided to inpatients on a gastrointestinal/hepatology (GIH) hospital service. METHODS: Patients with a GIH diagnosis were randomly assigned to be cared for by one of two hospital services. Over 3 months, teams were randomized to receive feedback (GIH A) or no feedback (GIH B, control group); feedback consisted of an email sent twice weekly to all physicians on the GIH A service detailing the length of stay (LOS) and real-time cost of care accrued by each inpatient. RESULTS: Over 3 months, care was provided to 56 (GIH A) and 47 (GIH B) inpatients with a GIH illness. Patient complexity level was similar for both services as demonstrated by mean relative value: 1.11 (GIH A) vs. 1.27 (GIH B), p=0.2. Weighted LOS and weighted cost of care values were calculated to adjust for the respective RV of each patient. Mean weighted LOS (10.8 [GIH A] vs. 13.8 days/pt [GIH B], p=0.02) and mean weighted cost of care (9,904 [GIH A] vs. 12,654 euros/pt [GIH B], p=0.02) were significantly lower in the feedback group. Subsequent hospital readmission rates did not differ among both groups. CONCLUSION: Systematic feedback on cost of care was associated with lower healthcare costs without compromising quality. Incorporating a running total of patient costs into computer software used to order patient tests may represent one approach to controlling healthcare expenses.

Specialty-specific admission: a cost-effective intervention?

INTRODUCTION: Cost effectiveness of healthcare has become an important component in its delivery. Current practices need to be assessed and measured for variations that may lead to financial savings. Speciality specific admission is known not only to lead improved clinical outcomes but also to lead important cost reductions. METHODS: All patients admitted to an Irish teaching hospital via the emergency department over a 2-year period with a gastroenterology (GI) related illness were included in this analysis.GI illness was classified using the Disease related grouping (DRG) system. Mean length of stay (LOS) and patient level costing (PLC) were calculated. Differences between DRGs with respect to speciality (i.e. specialist vs. non-specialist) were calculated for the five commonest DRGs. RESULTS: Significant variations in LOS and PLC were demonstrated in the DRGs. Mean LOS varied with increasing complexity, from 3.2 days for non-complex GI haemorrhage to 14.4 days for complex alcohol related cirrhosis as expected. A substantial difference in LOS within DRG groups was demonstrated by large standard deviations in the mean (up to 8.1 days in some groups) and was independent of complexity of cases. PLC also varied widely in both complex and non-complex cases with standard deviations of up to 17,342 noted. Specialty-specific admission was associated with shorter LOS for most GI admissions. CONCLUSION: Significant disparity exists for both LOS and PLC for most GI diagnoses. Specialty-specific admissions are associated with reduced LOS. Specialty-specific admission would appear to be cost-effective which may also lead to improved clinical outcomes.

Quality of colonoscopy performance among gastroenterology and surgical trainees: a need for common training standards for all trainees?

BACKGROUND AND STUDY AIM: Cecal intubation and polyp detection rates are objective measures of colonoscopy performance. Minimum cecal intubation rates greater than 90% have been endorsed by the American Society for Gastrointestinal Endoscopy (ASGE) and the Joint Advisory Group (JAG) UK. Performance data for medical and surgical trainee endoscopists are limited, and we used endoscopy quality parameters to compare these two groups. METHODS: Retrospective review of all single-endoscopist colonoscopies done by gastroenterology and surgical trainees ("registrars," equivalent to fellows, postgraduate year 5) with more than two years' endoscopy experience, in 2006 and 2007 at a single academic medical center. Completion rates and polyp detection rates for endoscopists performing more than 50 colonoscopies during the study period were audited. Colonoscopy withdrawal time was prospectively observed in a representative subset of 140 patients. RESULTS: Among 3079 audited single-endoscopist colonoscopies, seven gastroenterology trainees performed 1998 procedures and six surgery trainees performed 1081. The crude completion rate was 82%, 84% for gastroenterology trainees and 78% for surgery trainees (P < 0.0001). Adjusted for poor bowel preparation quality and obstructing lesions, the completion rate was 89%; 93% for gastroenterology trainees, and 84% for surgical trainees (P < 0.0001). The polyp detection rate was 19% overall, with 21% and 14% for gastroenterology and surgical trainees, respectively (P < 0.0001). The adenoma detection rate in patients over 50 was 12%; gastroenterology trainees 14% and surgical trainees 9% (P = 0.0065). In the prospectively audited procedures, median withdrawal time was greater in the gastroenterology trainee group and polyp detection rates correlated closely with withdrawal time (r = 0.99). CONCLUSION: The observed disparity in endoscopic performance between surgical and gastroenterology trainees suggests the need for a combined or unitary approach to endoscopy training for specialist medical and surgical trainees.

Randomised clinical trial: a 'nudge' strategy to modify endoscopic sedation practice.

BACKGROUND: In behavioural economics, a 'nudge' describes configuration of a choice to encourage a certain action without taking away freedom of choice. AIM: To determine the impact of a 'nudge' strategy - prefilling either 3mL or 5mL syringes with midazolam - on endoscopic sedation practice. METHODS: Consecutive patients undergoing sedation for EGD or colonoscopy were enrolled. On alternate weeks, midazolam was prefilled in either 3mL or 5mL syringes. Preprocedure sedation was administered by the endoscopist to achieve moderate conscious sedation; dosages were at the discretion of the endoscopist. Meperidine was not prefilled. RESULTS: Overall, 120 patients received sedation for EGD [59 (5mL), 61 (3mL)] and 86 patients were sedated for colonoscopy [38 (5mL), 48 (3mL)]. For EGDs, average midazolam dose was significantly higher in the 5-mL group (5.2mg) vs. 3-mL group (3.3mg), (P<0.0001); for colonoscopies, average midazolam dose was also significantly higher in the 5-mL group (5.1mg) vs. 3-mL group (3.3mg), (P<0.0001). There was no significant difference in mean meperidine dose (42.1mg vs. 42.8mg, P=0.9) administered to both colonoscopy groups. No adverse sedation-related events occurred; no patient required reversal of sedation. CONCLUSIONS: These findings demonstrate that 'nudge' strategies may hold promise in modifying endoscopic sedation practice. Further research is required to explore the utility of 'nudges' in impacting other aspects of endoscopic practice.

Utilization of resource leveling to optimize ERCP efficiency.

BACKGROUND: Optimizing endoscopy efficiency is becoming increasingly important. This study profiled ERCP availability and assessed resource leveling as a strategy to enhance efficiency. DESIGN: All ERCPs performed at an academic teaching hospital between January 2007 and December 2008 were reviewed. Procedure timeliness (time between admission and ERCP) and demand were analyzed to assess resource utilization. RESULTS: Data were recorded for 393 ERCPs. Profiling identified an unequal distribution of waiting times from admission to procedure due to restricted ERCP availability. Use of resource leveling methodology demonstrated that a small increase in procedure availability (one additional half day per week) would significantly reduce the hospital stay of ERCP patients. CONCLUSIONS: Resource leveling can be applied to balance procedure provision with demand to cope with fluctuations in demand. The impact of resource leveling can be truly measured only by implementing these changes and prospectively studying the effect.

What do gastroenterology trainees want: recognition, remuneration or recreation?

BACKGROUND: Occupational psychologists have identified three factors important in motivating physicians: financial reward, academic recognition, time off. AIM: To assess motivators among gastroenterology (GI) trainees. METHODS: A questionnaire was distributed to GI trainees to assess their motivators: (1) work fewer hours for less lucrative rate, (2) reduction in salary/increase in hours for academic protected time, and (3) work longer hours for higher total salary, but less lucrative hourly rate. RESULTS: Overall, 61 trainees responded; 52% of trainees would work shorter hours for less lucrative rate; 60% would accept a disproportionate reduction in salary/increase in hours for academic protected time; 54% would work longer hours for more money but less lucrative rate. Most trainees (93%) accepted at least one scenario. CONCLUSIONS: Most GI trainees are willing to modify their job description to align with their personal values. Tailoring job descriptions according to these values can yield economic benefits to GI Divisions.

Prospective, controlled assessment of the impact of formal evidence-based medicine teaching workshop on ability to appraise the medical literature.

BACKGROUND: The ability to critically appraise the calibre of studies in medical literature is increasingly important for medical professionals. AIM: This prospective controlled study evaluated the impact of a 6-h Evidence Based Medicine (EBM) Workshop on the critical appraisal skills of medical trainees. METHODS: Individuals attended three 2-h workshops over a 3-week period, incorporating didactic lectures in statistics, clinical trial design, appraising research papers and practical examples. Appraisal skills were assessed pre- and post-training based on grading the quality of randomised control studies (level 1 evidence), cohort studies (level 2 evidence) and case-control studies (level 3 evidence) [From Oxford Centre for Evidence Based Medicine Levels of Evidence (2001), http://www.cebm.net/critical_appraisal.asp ]. RESULTS: Overall grading improved from 39% (pre-course) to 74% (post-course), P = 0.002, with grading of levels 1, 2 and 3 studies improving from 42 to 75%, 53 to 61% and 21 to 84%, respectively. CONCLUSIONS: We conclude that a 6-h formal EBM workshop is effective in enhancing the critical appraisal skills of medical trainees.

Potential impact of enhanced practice efficiency on endoscopy waiting times.

BACKGROUND: With the growing demand on endoscopy services, optimising practice efficiency has assumed increasing importance. Prior research has identified practice changes, which increase the efficiency in endoscopy. In this study, the potential impact of these practice changes on the current and projected future endoscopy waiting times at our institution was assessed. METHODS: The annual volume of endoscopic procedures performed at a major teaching hospital and the annual procedure demand from 2000 to 2007 were reviewed. Procedure demand and waiting times were projected until 2012. The impact of three practice changes, which have been shown to increase efficiency was assessed: 1. routinely obtaining i.v. access and consent in patients prior to endoscopy (approach 1); 2. routinely obtaining i.v. access and consent, and sedating the patient prior to endoscopy (approach 2); 3. utilizing a two-room per endoscopist model (approach 3). RESULTS: There has been a significant increase in annual procedure volume (36%) and annual procedure demand (69%) from 2000 to 2007. Annual waiting times for routine procedures have lengthened, from 6 weeks (2000) to 22 weeks (2007). Assuming continued linear growth in demand up to 2012, the projected waiting times will continue to rise reaching 40 weeks in 2012. Routinely obtaining i.v. access/consent prior to procedure (approach 1) would shorten the average routine waiting times so that 8 weeks (recommended HSE maximum) would not be exceeded until early 2006; routinely obtaining i.v. access/consent and sedating patient prior to procedure (approach 2) would shorten the average routine waiting time so that 8 weeks would not be exceeded until 2008; utilising two rooms per endoscopist (approach 3) would shorten the average routine waiting time so that 8 weeks would not be exceeded until early 2012. CONCLUSIONS: Maintaining timely access to endoscopic services is becoming more challenging in the face of growing demand. Modifications in routine clinical practice can significantly impact procedure waiting times. In an era where economic aspects of medical care are becoming increasingly important and where there is growing focus on waiting times as a measure of clinical performance, these findings underscore the importance of providing clinical care in the most efficient manner possible.

Assessment of colorectal cancer knowledge and patient attitudes towards screening: is Ireland ready to embrace colon cancer screening?

INTRODUCTION: The level of awareness among the Irish public regarding colorectal cancer (CRC) remains uncertain. This study aimed to characterise CRC knowledge levels among a cohort of Irish patients. METHODS: A survey evaluating CRC knowledge levels was distributed among outpatients at a gastroenterology clinic in a Dublin teaching hospital. RESULTS: In total, 472 surveys were distributed of which 465 (98.5%) were returned. Twenty-nine percent of respondents correctly judged CRC to be the commonest cause of cancer death among the options provided while 26% correctly judged the lifetime risk of CRC; 59% underestimated and 15% overestimated the risk. Most patients (91%) were willing to pay 300 euros for a prompt colonoscopy if recommended by their physician while 7% opted to wait 6 months for a free colonoscopy. CONCLUSIONS: There is a willingness to embrace CRC screening and to shoulder some of the financial burden that this entails.

Assessment of adherence to published surveillance guidelines--opportunity to enhance efficiency of endoscopic practice.

Strict adherence to recommended surveillance intervals is important in ensuring timely access for patients awaiting endoscopy. This study aimed to characterise adherence rates to surveillance endoscopy guidelines. All surveillance procedures scheduled between January and December 2006 were reviewed. Surveillance procedures were classified as: a) Barrett's oesophagus, b) chronic IBD, c) prior adenomatous colorectal polyps and, d) prior surgical resection of colorectal cancer. 441 endoscopies were scheduled for surveillance of which 195 (44.2%) were scheduled at an inappropriate interval; all were scheduled prematurely. Of these, 50 of 133 (37.6%) Barrett's patients, 92 of 213 (43.2%) patients with prior colonic polyps, 36 of 48 (75.0%) patients with prior colonic malignancy and 17 of 47 (36.2%) patients for IBD surveillance were scheduled prematurely. Almost half of all surveillance procedures were scheduled inappropriately early. This 'over-surveillance' represents an unnecessary additional burden on the current endoscopic workload.

Evaluation of formal feedback on endoscopic competence among trainees: the EFFECT trial.

INTRODUCTION: The medical literature describes disparity in colonoscopy performance. This randomised, controlled study aimed to characterise the impact of feedback on colonoscopy performance among gastroenterology (GI) trainees. METHODS: Gastroenterology trainees of similar experience levels who independently performed 581 colonoscopies over the study period were randomised to receive feedback/no feedback on their colonoscopy performance. RESULTS: Baseline colonoscopy performance was similar in both groups. Following feedback, caecal intubation improved by 10.5% (from 72.9 to 83.4%, p = 0.04) in the feedback group and declined by 6.1% (from 78 to 71.9%, p = 0.2) in the control group; polyp detection improved by 5.1% (from 12.9 to 18.0%, p = 0.2) in the feedback group and by 2.9% (from 16.7 to 19.6%, p = 0.5) in the control group. CONCLUSIONS: Systematic feedback appears to enhance colonoscopy performance among GI trainees.

Impact of endoscopist withdrawal speed on polyp yield: implications for optimal colonoscopy withdrawal time.

BACKGROUND: In 2002, a U.S. Multi-Society Task Force on Colorectal Cancer recommended that the withdrawal phase for colonoscopy should average at least 6-10 min. This was based on 10 consecutive colonoscopies by two endoscopists with different adenoma miss rates. AIMS: To characterize the relationship between endoscopist withdrawal time and polyp detection at colonoscopy, and to determine the withdrawal time that corresponds to the median polyp detection rate. DESIGN: Procedural data from out-patient colonoscopies performed at the Mayo Clinic, Rochester during 2003 were reviewed. Endoscopists were characterized by their mean withdrawal time for a negative procedure and individual polyp detection rate. RESULTS: A total of 10 955 colonoscopies performed by 43 endoscopists were analysed. Median withdrawal time was 6.3 min (range: 4.2-11.9); polyp detection rate was 44.0% (all polyps), 29.8% (< or = 5 mm), 5.9% (6-9 mm), 6.7% (10-19 mm), 2.1% (> or = 20 mm). Longer withdrawal time was associated with higher polyp detection rate (r = 0.76; P < 0.0001); this relationship weakened for larger polyps (r = 0.19 for polyps 6-9 mm, r = 0.28 for polyps 10-19 mm, r = 0.02 for polyps > or = 20 mm). Overall median polyp detection rate corresponded to a withdrawal time of 6.7 min. CONCLUSION: Our findings support a colonoscopy withdrawal time of at least 7 min, which correlates with higher colon polyp detection rates.

Prospective assessment of the impact of feedback on colonoscopy performance.

BACKGROUND: Colonoscopy is an operator-dependent procedure. The medical literature describes disparity in colonoscopy performance with respect to polyp detection, caecal intubation rates and procedural times. AIM: To assess prospectively the impact of feedback among a large cohort of colonoscopists on three performance parameters: caecal intubation rate, insertion time and withdrawal time. METHOD: In a prospective clinical study, procedural data from all out-patient colonoscopies performed by attending gastroenterologists at our institution were recorded routinely in a computerized database. Enhanced serial feedback was provided on a quarterly basis for three procedure parameters: intubation to caecum, insertion time and withdrawal time. Feedback (absolute value, % rank and group distribution) was sent by email every 3 months beginning with January 2005 feedback for all of 2004, and subsequently quarterly in April 2005 (for January-March 2005), July 2005 (for April-June 2005) and October 2005 (for July-September 2005). RESULTS: Feedback was provided to 58 endoscopists with a median experience level of 8 years. There was a relative decline of 19% in incomplete procedures, with median caecal non-intubation rates decreasing from 4.7% to 3.8% following the introduction of feedback while median insertion times declined from 10.6 to 9.5 mins, P = 0.02. Median withdrawal times did not change significantly, 9.1-8.9 mins, P = 0.6. CONCLUSIONS: Feedback by email appears to improve colonoscopy performance, enhancing completion rates and shortening insertion times without compromising withdrawal times.

Variation in practice of ileal intubation among diverse endoscopy settings: results from a national endoscopic database.

BACKGROUND: Terminal ileum intubation rates at colonoscopy are variable. One of the major indications for terminal ileum intubation is to identify Crohn's disease. Signs and symptoms which raise a suspicion of Crohn's include abdominal pain/bloating, anaemia and diarrhoea. AIM: To determine the proportion of terminal ileal intubation in patients undergoing evaluation of abdominal pain/bloating, anaemia or diarrhoea with normal endoscopic findings at colonoscopy. METHODS: The Clinical Outcomes Research Initiative national endoscopic database was analysed to determine the proportion of terminal ileum intubation in patients undergoing evaluation of either abdominal pain/bloating, anaemia or diarrhoea with normal endoscopic findings at colonoscopy and to characterize this population of patients. Patients with known or suspected inflammatory bowel disease were excluded from the analysis. RESULTS: Between January 2000 and December 2003, 21 638 patients underwent complete colonoscopy for evaluation of either abdominal pain/bloating, anaemia or diarrhoea with normal colon findings. Overall, 3858 patients (18%) underwent terminal ileum evaluation. Intubation rates differed according to procedure indication: abdominal pain (13%), anaemia (13%), diarrhoea (28%). Terminal ileum assessment declined with advancing patient age and was least frequent in Black patients (12% vs. 18% in non-Blacks, P < 0.0001). Ileal intubation rates also varied among endoscopy site types: community (17%), academic (21%), Veterans Affairs Medical Centres (17%), P < 0.0001. Multiple logistic regression identified patients with the indication of diarrhoea (OR: 2.58) as more likely to undergo terminal ileum intubation when compared with those with abdominal pain/bloating. Patients in Veterans Affairs (OR: 1.26) and academic (OR: 1.29) sites were more likely to undergo terminal ileum intubation compared with community sites. CONCLUSION: Less than one-fifth of patients with either abdominal pain/bloating, anaemia or diarrhoea underwent ileal intubation in the setting of a normal colonoscopy. Significant practice variation was observed in rates of terminal ileum evaluation. Further study is required to determine whether terminal ileum examination impacts patient management or outcome.

PSC-IBD: a unique form of inflammatory bowel disease associated with primary sclerosing cholangitis.

BACKGROUND: Inflammatory bowel disease associated with primary sclerosing cholangitis (PSC-IBD) may have a high prevalence of rectal sparing, backwash ileitis, and colorectal neoplasia. AIMS: To describe the clinical features and outcomes of PSC-IBD and compare these to a group of chronic ulcerative colitis (CUC) patients. METHODS: The medical records of all patients with PSC-IBD evaluated at the Mayo Clinic Rochester between 1987 and 1992 were abstracted for information on endoscopic and histological features, colorectal neoplasia, surgery, and other clinical outcomes. Patients referred for colorectal neoplasia and those who did not undergo colonoscopy with biopsies were excluded. A control group of CUC patients matched for sex, duration of IBD at first clinic visit, and calendar year of first clinic visit was identified, and similar information was abstracted. RESULTS: Seventy one PSC-IBD patients and 142 CUC patients without PSC were identified. Rectal sparing and backwash ileitis were more common in the PSC-IBD group (52% and 51%, respectively) than in controls (6% and 7%, respectively). Overall, colorectal neoplasia developed in 18 cases and 15 controls, including 11 cancers (seven cases and four controls). An increased risk of colorectal neoplasia or death was not detected in a matched analysis. Although the cumulative incidence of colorectal neoplasia was higher in cases (33%) than in controls (13%) at five years, this was of borderline statistical significance (p=0.054, unmatched log rank test). Overall survival from first clinic visit was significantly worse among cases (79% v 97%) at five years (p<0.001, unmatched log rank test). CONCLUSION: PSC-IBD is frequently characterised by rectal sparing and backwash ileitis. Colorectal neoplasia develops in a substantial fraction and overall survival is worse. PSC-IBD may represent a distinct IBD phenotype.

Cost-effectiveness analysis of alternative strategies for palliation of distal biliary obstruction after a failed cannulation attempt.

OBJECTIVES: Occasionally alternative techniques such as precut sphincterotomy or percutaneous transhepatic cholangiography (PTC) are required to achieve access to the common bile duct. Tradeoffs exist, however, with respect to their complications and costs. Some experts believe that precut sphincterotomy should not be performed at all. We aimed to compare the cost-effectivenesses of metallic biliary stent placement after an initial failed cannulation attempt at ERCP utilizing precut sphincterotomy and placement utilizing PTC for palliation of jaundice. A cost-effectiveness analysis was performed, as viewed from the societal perspective. METHODS: A decision analysis model was designed comparing precut sphincterotomy and PTC approaches for placement of a metallic biliary stent for palliation of jaundice in a patient with inoperable malignant distal biliary obstruction in whom an initial attempt at ERCP cannulation had failed. Baseline probabilities, obtained from the published literature, were varied through plausible ranges using sensitivity analysis. Charges were based on Medicare professional plus facility fees or diagnosis-related group rates for out- and inpatients, respectively. The outcome measured was cost per year of life. RESULTS: Sensitivity analysis showed that precut sphincterotomy with subsequent PTC, if necessary, was the most cost-effective strategy provided the precut complication rate was <51% ($9,033/yr), versus $14,741/yr for PTC. CONCLUSIONS: Precut sphincterotomy followed by PTC (if necessary) is the most cost-effective strategy for palliative biliary stenting in the setting of malignant distal biliary obstruction after a failed ERCP attempt. The endoscopic approach is best practiced by experienced endoscopists who minimize precut complication rates.

An assessment of the learning curve for precut biliary sphincterotomy.

OBJECTIVE: Precut sphincterotomy is considered unsafe when used by inexperienced endoscopists. We sought to determine whether procedural experience with precut sphincterotomy predicted either successful cannulation or development of complications in these patients. METHODS: We describe the experience of 253 consecutive patients who underwent precut biliary sphincterotomy done by one endoscopist between September, 1993 and April, 2001. Data were prospectively collected on procedure indication and outcome. All patients were contacted by phone 30 days after the procedure to determine outcome. We also described precut utilization over time. RESULTS: All 253 precut procedures were divided chronologically into five groups of 50, with 53 in the final group. The rates of successful cannulation after precutting were 88%, 94%, 90%, 88%, and 98%, respectively (p = 0.05 for groups 1-4 vs group 5). Overall complication rates were similar in all groups (12%, 18%, 20%, 12%, and 14%), with no difference in complication severity. Female gender predicted successful cannulation (OR = 2.9 [p = 0.02]), whereas an indication of sphincter of Oddi dysfunction predicted development of complications (OR = 1.7 [p = 0.03]). The total number of ERCP procedures performed increased over time, whereas the proportion of precut sphincterotomies performed decreased. CONCLUSIONS: Although the success rate for precut sphincterotomy may increase with procedural experience, the complication rate does not seem to decrease. Precut sphincterotomy continues to carry an increased complication rate over standard sphincterotomy even when performed by an experienced biliary endoscopist. The need for precut sphincterotomy appears to decrease with increasing ERCP experience.