The exact definition of what constitutes quality of

Transcription

1 Clinical Gastroenterology and Hepatology 2016;14: Poor Documentation of Inflammatory Bowel Disease Quality Measures in Academic, Community, and Private Practice Joseph D. Feuerstein,* Natalia E. Castillo,* Sana S. Siddique, Jeffrey J. Lewandowski, Kathy Geissler, k Manuel Martinez-Vazquez, Chandrashekhar Thukral, #, ** Daniel A. Leffler,* and Adam S. Cheifetz* *Department of Medicine and Division of Gastroenterology, Department of Emergency Medicine, Beth Israel Deaconess Medical Center, Department of Medicine, Mt Auburn Hospital, Harvard Medical School, Boston, Massachusetts; k Rockford Gastroenterology Associates, Rockford, Illinois; Gastroenterology Service, Dr. José Eleuterio González University Hospital, Monterrey, Nuevo León, Mexico; # University of Illinois at Chicago College of Medicine, Rockford, Illinois; **Rockford and Rockford Gastroenterology Associates, Rockford, Illinois BACKGROUND & AIMS: METHODS: RESULTS: CONCLUSIONS: Quality measures are used to standardize health care and monitor quality of care. In 2011, the American Gastroenterological Association established quality measures for inflammatory bowel disease (IBD), but there has been limited documentation of compliance from different practice settings. We reviewed charts from 367 consecutive patients with IBD seen at academic practices, 217 patients seen at community practices, and 199 patients seen at private practices for compliance with 8 outpatient measures. Records were assessed for IBD history, medications, comorbidities, and hospitalizations. We also determined the number of patient visits to gastroenterologists in the past year, whether patients had a primary care physician at the same institution, and whether they were seen by a specialist in IBD or in conjunction with a trainee, and reviewed physician demographics. A univariate and multivariate statistical analysis was performed to determine which factors were associated with compliance of all core measures. Screening for tobacco abuse was the most frequently assessed core measure (89.6% of patients; n [ 701 of 783), followed by location of IBD (80.3%; n [ 629 of 783), and assessment for corticosteroid-sparing therapy (70.8%; n [ 275 of 388). The least-frequently evaluated measures were pneumococcal immunization (16.7% of patients; n [ 131 of 783), bone loss (25%; n [ 126 of 505), and influenza immunization (28.7%; n [ 225 of 783). Only 5.8% of patients (46 of 783) had all applicable core measures documented (24 in academic practice, none in clinical practice, and 22 in private practice). In the multivariate model, year of graduation from fellowship (odds ratio [OR], 2.184; 95% confidence interval [CI], ; P <.001), year of graduation from medical school (OR, 0.500; 95% CI, ; P <.001), and total number of comorbidities (OR, 1.089; 95% CI, ; P [.016) were associated with compliance with all core measures. We found poor documentation of IBD quality measures in academic, clinical, and private gastroenterology practices. Interventions are necessary to improve reporting of quality measures. Keywords: AGA; Crohn s Disease; Ulcerative Colitis; Guidelines; Outcome. The exact definition of what constitutes quality of care has varied over time. 1 The Institute of Medicine defines quality as, the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge. 2 Quality care is a multifaceted goal aimed at improving patient outcomes, creating a healthier population, and reducing health care costs. 3 To this end, government organizations and medical societies have developed quality measures and practice guidelines aimed at improving the overall quality of care. In support of these measures, Mehta et al 4 found that when practice guidelines were followed, quality of care was improved across different institutions and patients. Based on these potential benefits, several commercial and Abbreviations used in this paper: AGA, American Gastroenterological Association; AP, academic practice; BIDMC, Beth Israel Deaconess Medical Center; CI, confidence interval; CP, community practice; EHR, electronic health record; IBD, inflammatory bowel disease; OR, odds ratio; PP, private practice; TNF, tumor necrosis factor. Most current article 2016 by the AGA Institute /$

2 422 Feuerstein et al Clinical Gastroenterology and Hepatology Vol. 14, No. 3 public insurance companies have provided incentives for compliance with quality measures and established penalties for those failing to adequately document compliance with quality measures. 5,6 Although governmental agencies and physicians alike have advocated for the widespread use of electronic health record (EHR) systems to improve reporting of these indices, a number of studies have not found an improvement in quality of care with the use of EHRs alone. 7,8 Inflammatory bowel disease (IBD) is a chronic medical condition associated with significant morbidity. 9 Although the disease often can be treated successfully with pharmacologic therapies, these medications carry significant, albeit rare, risks. Nonetheless, given the efficacy of these agents, they are used frequently in the management of IBD. 10 However, a recent analysis of US health care utilization in IBD from 2010 to 2012 indicated that corticosteroids still are used frequently with low rates of steroid-sparing therapies. 11 In 2011, the American Gastroenterological Association (AGA) established 10 quality metrics for the care of patients with IBD. 12 The measures focus on means to improve the care provided to patients with IBD while also reducing the risk of treatment-related complications. Despite the publication of these measures in 2011, the current adoption of these measures is inadequate. A recent abstract by Melmed et al 13 using a large administrative database showed that overall compliance with the AGA quality measures is suboptimal. The adequacy of gastroenterologists documentation of the AGA s outpatient IBD core measures in consecutive patients in different outpatient practice settings. We sought to appraise the current state of compliance with the AGA core measures in various types of practices including a large tertiary care academic faculty practice, a community hospital based practice, and a large private practice group. Methods The 10 AGA IBD core measures were reviewed for measures relevant to the outpatient setting. 12 Two of the indices, inpatient testing of Clostridium difficile infection and inpatient prevention of venous thromboembolism, therefore were excluded (Table 1). The study was conducted at 3 health care facilities. The first center was a tertiary care academic medical center (academic practice [AP]), Beth Israel Deaconess Medical Center (BIDMC) in Boston, Massachusetts. BIDMC has a specialized center for inflammatory bowel disease consisting of 5 physicians in addition to 24 other gastroenterologists. The second institution, Mount Auburn Hospital, was a community-based, Harvard Medical School affiliated program (community practice [CP]) consisting of a gastroenterology practice of 5 physicians. The third center was a large private gastroenterology practice (private practice [PP]) of 15 physicians Table 1. Inflammatory Bowel Disease Quality Measures Measure Location of care 1: IBD: type, anatomic location, and activity all Outpatient assessed 2: IBD preventive care: corticosteroid-sparing Outpatient therapy 3: IBD preventive care: corticosteroid-related Outpatient iatrogenic injury bone loss assessment 4: IBD preventive care: influenza immunization Outpatient 5: IBD preventive care: pneumococcal Outpatient immunization 6: Testing for latent tuberculosis before starting Outpatient anti-tnf therapy 7: Assessment of hepatitis B virus before Outpatient starting anti-tnf therapy 8: Testing for C difficile a Inpatient 9: Prophylaxis for venous thromboembolism a Inpatient 10: IBD preventive care: tobacco screening Outpatient and cessation intervention a Excluded measures from our study. and 1 nurse practitioner in Rockford, Illinois. We reviewed consecutive medical charts of patients with IBD who were seen at each center. At BIDMC, we reviewed consecutive charts during the month of April 2013, at Mt. Auburn Hospital we reviewed consecutive charts from January 2013 to December 2013, and at Rockford Gastroenterology Associates, we reviewed charts from January 2014 to February The time point of chart review was chosen by each center with a goal to obtain 200 consecutive patients and at least a 1-month time interval. The patient sample size was one of convenience because there were no studies available to project if any differences would be found between an AP, CP, and PP and which center would be better. In addition, none of the 3 centers had any checklist or review of the AGA quality measures aside from publications sent out to the AGA membership by the AGA. Each center used an EHR system. Each chart was reviewed manually by one or more of the authors. All gastroenterology notes were reviewed for documented compliance with the outpatient AGA IBD quality measures. As per the AGA recommendations, each chart was reviewed for 12 months before the patient was included to assess for documentation of compliance with the quality measures. The only exception was measure 6 (screening for latent tuberculosis before anti tumor necrosis factor [TNF]), which was reviewed for the 6 months preceding study inclusion based on AGA recommendations. In addition, when available in the EHR, medications, immunizations, laboratory data, and tobacco screening were reviewed. When reviewing immunizations and tuberculosis skin testing, compliance was considered only if there was documentation of administration. Lastly, if a patient already had an anti-tnf started at one of the study centers, the charts were reviewed to determine if an assessment for latent tuberculosis or hepatitis B was performed before initiation of

3 March 2016 Documented Reports With IBD Quality Measures 423 the anti-tnf even if this was before the study period dates. These cases were included because of the limited number of patients seen with planned initiation of anti- TNF during the study inclusion dates. In addition, only gastroenterology notes were reviewed with regard to quality measures. Primary care notes were not reviewed to determine compliance with the measures because the focus was on reporting by gastroenterologists. However, if a gastroenterology note commented that a test/procedure was performed with the primary care physician or in an outside institution, this was considered adequate documentation of the quality measure. Patient demographic information and physician training background also were obtained. Type of IBD, year of IBD diagnosis, total number of current medications documented in the last clinical note during the study period, comorbidities as listed in the patient s past medical history, hospitalizations in the past year, and gastroenterology visits in the past year were obtained. When applicable, it was noted if the patient was seen by a physician in the BIDMC Center for Inflammatory Bowel Disease, or by a gastroenterology trainee (BIDMC only), and whether the patient had a primary care physician at the same institution (BIDMC and Mt. Auburn). The physician s year of medical school graduation, fellowship graduation, and board certification was assessed as well. Patients were categorized into 4 groups based on which AGA quality measures were applicable. Group 1 included patients never exposed to steroids or anti- TNF medications (measures 1, 4, 5, and 10) (Table 1). Group 2 included patients who were previously, but not currently, on corticosteroids, and had never been treated with anti-tnf. These patients were not exposed to steroids in the past year and had been on a stable dose of a non steroid-sparing agent or off medication without flares for at least 1 year or longer. Therefore, this group was not assessed within this study for steroid-sparing medications given the remote use of steroids (measures 1, 3, 4, 5, and 10). Group 3 included patients currently receiving steroids, but not considering therapy with anti-tnf (measures 1, 2, 3, 4, 5, and 10). Group 4 included patients who previously were exposed/ currently on steroids, and anti-tnf was being discussed. This group also included patients who already had been started on an anti-tnf at 1 of the 3 centers (measures 1, 2, 3, 4, 5, 6, 7, and 10). The primary end point of the study was adherence to all applicable core measures. We assessed compliance with all applicable core measures in the 3 centers combined and within each center individually. Given the importance of proving quality care, the optimal compliance rate was considered 100%, which is higher than the current standards set forth by the AGA. Nonetheless, for this study of quality measures, the goal was to provide all patients with optimal quality of care in all situations. Data were analyzed further to determine patient variables and physician variables predictive of compliance with all relevant measures. Statistical analysis Statistical analysis was performed using Stata (version 12; StataCorp LP, TX). Categoric variables were expressed as frequencies and proportions, and continuous variables were expressed as medians and ranges. Differences between continuous variables were tested using Kruskall Wallis and analysis of variance after assessing each continuous variable for normality with the Shapiro Wilk normality test. Categoric variables were evaluated using the chi-square or the Fisher exact test as appropriate. Statistical significance was set at a P value less than.05. Multiple pairwise comparisons at an adjusted P value less than.016 were performed between centers. Predictors of compliance with all relevant core measurements were assessed using a univariate logistic regression reporting the odds ratio and 95% confidence intervals (CIs). Explanatory variables in the final multivariate model were evaluated by a forward selection model at a significant P level of.10. The final multivariate model included year of graduation from fellowship, year of graduation from medical school, total comorbidities, year of IBD diagnosis, and medications at last gastroenterology visit. Variables that were potentially redundant in nature in characterizing physician demographics were excluded in the multivariate model. All authors had access to the data and approved the final manuscript. The study was approved by the Institutional Review Boards at BIDMC, Mt. Auburn, and Northwestern University. Results Demographics Of the 783 patients from the 3 participant centers, 367 were seen in AP, 217 at CP, and 199 at PP. Fifty-eight percent (n ¼ 457 of 783) of the cohort were female. No significant differences in sex were seen between institutions (P ¼.544). The median age was 44 years (range, y; P ¼.506). Of the 783 subjects, 340 patients (43.4%) had ulcerative colitis (P ¼.323), 55.2% (432 of 783) had Crohn s disease (P ¼.383), and 1.4% (n ¼ 11 of 783) had indeterminate colitis. Tables 2 and 3 show further demographic details. When the patients were categorized into groups, 34.3% (n ¼ 269 of 783) were in group 1 (never exposed to steroids or use of anti-tnf), 16.1% (n ¼ 126 of 783) were in group 2 (prior steroid exposure but not currently or in the past year), 21.5% (n ¼ 168 of 783) were in group 3 (prior/current steroid exposure but no anti-tnf), and 28.1% (n ¼ 220 of 783) were in group 4 (prior/current steroid exposure on/planning use of anti-tnf) (Supplementary Table 1). The relative number of patients in each category was different at every center in the study.

4 424 Feuerstein et al Clinical Gastroenterology and Hepatology Vol. 14, No. 3 Table 2. Demographics of Patients at Each Center Academic practice (n ¼ 367) Community practice (n ¼ 217) Private practice (n¼ 199) Total (n ¼ 783) P value Males, n (%) a 145 (39.5) 91 (41.9) 88 (44.2) 324 (41.4).544 Median age, y (range) 43 (17 88) 45 (21 91) 45 (18 92) 44 (17 92).506 Median year of IBD diagnosis (range) 2002 ( ) 2007 ( ) 2004 ( ) 2004 ( ) <.001 b Ulcerative colitis, n (%) 159 (43.3) 102 (47.0) 79 (39.7) 340 (43.4).323 Crohn s disease, n (%) 199 (54.2) 115 (53.0) 118 (59.3) 432 (55.2).383 Indeterminate colitis, n (%) 9 (2.5) - 2 (1.0) 11 (1.4).344 Seen by gastroenterology trainee, n (%) 158 (43.0) (20.2) - Seen by IBD specialist, n (%) 234 (63.8) (29.8) - Primary care physician at same medical center 173 (47.1) 90 (41.5) (33.6).184 Patients hospitalized in the past year, n (%) 66 (17.9) 39 (17.9) 45 (22.6) 150 (19.2).357 Number of hospitalizations in the past year, 0(0 15) 0 (0 4) 0 (0 7) 0 (0 15).469 Number of gastroenterology visits in past year, 3(0 13) 1 (1 7) 2 (0 6) 2 (0 13) <.001 b Number of medications at last gastroenterology 7(0 35) 3 (0 19) 7 (0 23) 6 (0 35) <.001 c visit, Number of comorbidities, 4 (0 36) 3 (0 14) 3 (1 13) 3 (0 36) <.001 c a Sex is missing for 2 participants at the community center. b Pairwise comparisons were significant for AP vs PP and AP vs CP with a P value of <.016. c Pairwise comparisons were significant for AP vs CP and CP vs PP with a P value of <.016. Table 4 shows the compliance rates with individual core measures in aggregate and per institution. Overall, vaccination for influenza (measure 4) (28.7%; n ¼ 225 of 783), vaccination for pneumonia (measure 5) (16.7%; n ¼ 131 of 783), and assessing for bone loss secondary to steroid use (measure 3) (25.0%; n ¼ 126 of 505) were the least adhered to measures. Only assessing for tobacco abuse (measure 10) was documented consistently (89.6%; n ¼ 701 of 783). Table 4 details compliance with the core measures for individual institutions. The measures that were significantly different between the centers were as follows: documenting location of IBD (measure 1) (AP, 70.0%; CP, 81.1%; and PP, 98.5%; P.001), corticosteroid-sparing therapy (measure 2) (AP, 82.2%; CP, 8.3%; and PP, 26.1%; P.001), assessing for bone loss secondary to steroid use (measure 3) (AP, 30.0%; CP, 8.3%; and PP, 26.1%; P.001), influenza immunization (measure 4) (AP, 33.8%; CP, 4.6; and PP, 45.7%), pneumococcal immunization (measure 5) (AP, 20.7%; CP, 0.5%; and PP, 27.1%; P.001), testing for latent tuberculosis before anti-tnf (measure 6) (AP, 67.2%; CP, 45.8%; and PP, 35.0%; P.001), assessing for latent hepatitis B before anti-tnf (measure 7) (AP, 74.1%; CP, ¼2.5%; and PP, 56.3%; P ¼.023), and assessing for tobacco abuse (measure 10) (AP, 95.9%; CP, 76.5%; and PP, 91.9%; P.001). Pairwise comparisons are specified in Table 4. Primary Outcome: Overall Core Measures Compliance When analyzing all the centers, only 5.8% (n ¼ 46 of 783) of patients had all applicable core measures documented appropriately. Only AP (n ¼ 24) and PP (n ¼ 22) Table 3. Physician Demographics at Each Institution Academic practice (n ¼ 367) Community practice (n ¼ 217) Private practice (n ¼ 187) Total a (n ¼ 771) P value Year of graduation from medical school, 1990 ( ) 1990 ( ) 1997 ( ) 1990 ( ) <.001 b Year of graduation from fellowship, median 1996 ( ) 2001 ( ) 2004 ( ) 2001 ( ) <.001 b (range) Board certified (%) 322 (87.7) 217 (100.0) 187 (100.0) 726 (94.2) <.001 c Year of board certification, 1983 ( ) 2002 ( ) 2007 ( ) 2004 ( ) <.001 b US Medical School graduates (%) 230 (62.7) 149 (68.7) 187 (100.0) 566 (73.4) <.001 d a Excluding physician (5) n ¼ 12; total n ¼ 771. b Pairwise comparisons were significant for AP vs CP, AP vs PP, and CP vs PP, with a P value of less than.016. c Pairwise comparisons were significant for AP vs CP and AP vs PP, with a P value of less than.016. d Pairwise comparison were significant for AP vs PP and CP vs PP, with a P value of less than.016.

5 March 2016 Documented Reports With IBD Quality Measures 425 Table 4. Compliance of Individual IBD Core Measures by Institution Measures Academic practice (n ¼ 367) Community practice (n ¼ 217) Private practice (n ¼ 199) (%) Total (n ¼ 783) P value Measure 1: type, anatomic location, and activity 257/367 (70.0) 176/217 (81.1) 196/199 (98.5) 629/783 (80.3) <.001 a all assessed, n (%) Measure 2: corticosteroid-sparing therapy, n (%) 180/219 (82.2) 11/61 (18.0) 84/108 (77.7) 275/388 (70.8) <.001 b Measure 3: corticosteroid-related iatrogenic 88/293 (30.0) 8/97 (8.3) 30/115 (26.1) 126/505 (25.0) <.001 b injury bone loss assessment, n (%) Measure 4: influenza immunization, n (%) 124/367 (33.8) 10/217 (4.6) 91/199 (45.7) 225/783 (28.7) <.001 a Measure 5: pneumococcal immunization, n (%) 76/367 (20.7) 1/217 (0.5) 54/199 (27.1) 131/783 (16.7) <.001 b Measure 6: testing for latent tuberculosis before 78/116 (67.2) 11/24 (42.3) 28/80 (35.0) 117/220 (52.7) <.001 c starting anti-tnf therapy, n (%) Measure 7: assessment of hepatitis B virus 86/116 (74.1) 15/24 (62.5) 45/80 (56.3) 146/220 (66.4).023 c before starting anti-tnf therapy, n (%) Measure 10: tobacco screening and cessation intervention, n (%) 352/367 (95.9) 166/217 (76.5) 183/199 (91.9) 701/783 (89.6) <.001 b NOTE. Denominator in the table is the number of subjects. a Pairwise comparisons were significant for all comparisons AP vs CP, AP vs PP, and CP vs PP. b Pairwise comparisons were significant for AP vs CP and CP vs PP. c Pairwise comparisons were only significant for AP vs PP. reported compliance with all core measures in a subset of patients (Supplementary Tables 2 and 3). When assessing patient demographics, univariate predictors of compliance with applicable core measures were as follows: age (odds ratio [OR], 1.018; 95% CI, ; P ¼.042), more recent year of IBD diagnosis (OR, 0.967; 95% CI, , P ¼.002), greater number of medications at last gastroenterology visit (OR, 1.075; 95% CI, ; P ¼.002), and the greater the mean number of medical comorbidities (OR, 1.076; 95% CI, ; P ¼.003). Additional predictors included being seen by a gastroenterology trainee (OR, 2.235; 95% CI, ; P ¼.015). Trending toward significance was being seen by a specialist in IBD (OR, 1.586; 95% CI, ; P ¼.127). Physician demographics including year of graduation from medical school (OR, 1.092; 95% CI, ; P.001), year of graduation from fellowship (OR, 1.086; 95% CI, ; P.001), and year of board certification (OR, 1.133; 95% CI, ; P.001) were all statistically significant (Table 5). No significant differences were seen across practices (AP and PP) between those who complied with all core measures vs those who did not report all core measures (P ¼.076). CP was excluded from analysis given that none of the subjects were evaluated for all relevant core measures. Similarly, graduating from a US medical school was excluded from the analysis given that all 46 patients with all core measures documented were seen by physicians graduating from US medical schools. Multivariate analysis showed that year of graduation from fellowship (OR, 2.184; 95% CI, ; P.001), year of graduation from medical school (OR, 0.500; 95% CI, ; P.001), and total number of comorbidities (OR, 1.089; 95% CI, ; P ¼.016) were significant. Trending toward significance were year of IBD diagnosis (OR, 0.974; 95% CI, ; P ¼.051), and number of medications at the last gastroenterology visit (OR, 0.072; 95% CI, ; P ¼.072) (Table 6). Post Hoc Analysis of Results by Practice In review of documented compliance at each practice, physicians at each site were queried as to methods used that provided greater adherence/lack of adherence to the quality measures. None of the centers provided vaccinations in their clinics, providing some explanation for the low vaccination rates. Both the AP and PP used an EHR system with alerts regarding tobacco abuse necessitating clicking a checkbox to remove the alert. In contrast, the CP did not have such an alert system and instead relied on the physician transcribing this into the record system from a nursing intake form. The higher rates of documented compliance with measure 1 at PP and CP seemed more related to a standardized practice in the group documentation compared with the AP, which did not report any standardized documentation. In contrast, the CPs lower rates of compliance with measures 2 and 3 were believed to be secondary to inadequate documentation as opposed to a lack of assessment. Finally, the higher rates of pre anti-tnf reporting of tuberculosis testing and hepatitis B testing at the AP was secondary to a common practice within the group to document in the clinical notes any records obtained from external sources. Discussion This was a multicenter study to assess documented compliance with IBD core measures. We show that the current documented reporting of the IBD quality measures is inadequate to prove adequate adherence with the AGA quality measures. This was present regardless of

6 426 Feuerstein et al Clinical Gastroenterology and Hepatology Vol. 14, No. 3 Table 5. Univariate Predictors of Compliance With All Relevant Core Measures Patient with all core measures evaluated (n ¼ 46) Patients without all measures evaluated (n ¼ 737) P value Odds ratio (95% CI) Sex, male/female, n (%) 20 (43.5)/26 (56.5) 304 (41.4)/431 (58.6) a ( ) Median age, y (range) 51.5 (21 89) 44 (17 92) ( ) Median year of IBD diagnosis, median 1998 ( ) 2004 ( ) ( ) Type of IBD UC/Crohn s disease, a n (%) 21 (45.7)/25 (54.3) 319 (44.0)/407 (56.0) ( ) Seen by gastroenterology trainee, n (%) 16/46 (34.8) 142/737 (19.3) ( ) Seen by specialist in IBD, n (%) 23/46 (50.0) 285/737 (38.7) ( ) Primary care physician at same medical center, n (%) 18/46 (60.9) 245/737 (33.2) ( ) Patients hospitalized in the past year, n (%) 11/46 (23.9) 141/737 (19.1) ( ) Number of hospitalizations in the past year, 0(0 4) 0 (0 15) ( ) Number of gastroenterology visits in past year, 2(1 7) 2 (0 13) ( ) Number of medications at last gastroenterology visit, 8(3 24) 6 (0 35) ( ) Number of comorbidities, 5 (1 18) 3 (0 36) ( ) Year of graduation from medical school, 1997 ( ) 1990 ( ) < ( ) Year of graduation from fellowship, 2004 ( ) 2001 ( ) < ( ) Year of board certification, 2004 ( ) 2002 ( ) < ( ) Type of practice, b academic practice/private practice 24 (6.5)/22 (11.1) 343 (46.5)/177 (24.0) ( ) NOTE. The variable of US medical school was excluded from the univariate analysis and multivariate analysis given that all 46 subjects (with all core measurements) were from a US medical school, thus failing to predict this variable in any of the models. UC, ulcerative colitis. a Two of 737 patients refused to report their sex. b For statistical analysis, the 11 patients with inflammatory bowel disease unclassified were excluded. the type of clinical practice setting academic medical center, community hospital, or private practice. Predictors of compliance with the measures in our study included year of graduation from fellowship and total number of comorbidities at the last gastroenterology visit. Because the IBD quality measures only recently were developed (2011), more recent graduates may be more up to date with the latest recommendations. This finding is consistent with Choudhry et al, 14 who reported that more experienced physicians and longer intervals since graduating medical school did not necessarily improve the quality of care. In our study, we set the ideal benchmark for documented compliance with quality measures at 100%. In theory, if quality measures are meant to improve the overall quality of care, then all patients should receive all Table 6. Multivariate Analysis of Variable Predicting Compliance With All Applicable Core Measures P value Odds ratio (95% CI) Year of graduation from < ( ) fellowship Year of graduation from < ( ) medical school Total number of comorbidities ( ) Year of IBD diagnosis ( ) Number of medications at last gastroenterology visit ( ) measures. This goal is higher than that which has been set by the Health Care Incentives Improvement Institute Bridges to Excellence IBD care recognition program. 15 In this program, each measure is assigned a number of points totaling 100. A passing score requires compliance with 60 points. However, in the age of accountable care organizations, care is now being assessed by the quality provided to patients. 16,17 Ideally, by providing patients with 100% quality care, patient health should be improved and overall costs to the medical system subsequently should be reduced. 16,17 Therefore, we set our benchmark at 100% compliance. Unfortunately, none of the centers achieved this benchmark for any measures. The overall suboptimal documentation noted in our study, however, may be more indicative of inadequate dissemination and publicizing of the core measures and/ or poor acceptance by practitioners of their importance in clinical practice. Quality measures are meant to improve the quality of care provided, but, importantly, this quality should result in improved outcomes (eg, mortality, re-admissions, patient satisfaction). 18 However, this connection to improved outcomes in quality measures is inconsistent and often lacking. 18 Although the AGA IBD quality measure document provides significant background supporting the basis for each measure, additional studies are necessary to prove that compliance with these measures truly improves patient outcomes. 12 Although our study did not assess the reasons for poor compliance, some argue that compliance with current

7 March 2016 Documented Reports With IBD Quality Measures 427 quality measures have not shown improvements in patient care, and that guidelines used to formulate recommendations for the quality measures often are based on expert opinion and not clinical studies This may explain why even when IBD specialists and patients primary care physicians were at the same institution, compliance with the measures still was not improved. In contrast, the one measure that has been shown to improve outcomes, assessing for tobacco cessation, which also is a component of the meaningful use measures, was the most frequently documented measure. 23,24 However, with the threat of reductions in reimbursements and public reporting of quality measures, compliance with the current quality measures is critical. When analyzing the specific documentation of the quality measures in aggregate, compliance with the majority of the measures was superior to that noted by Melmed et al, 13 who used the OptumInsight administrative claims database. The underlying improvement in this reporting is unclear. There may be a component of improved knowledge of the AGA quality measures given the advertising campaign surrounding these measures. In addition, however, it is likely that the improved compliance in some areas, such as screening for tuberculosis (21.7% in the OptumInsight database vs 52.7% in our study) and assessing for hepatitis B virus (20.3% in the OptumInsight database vs 66.4% in our study), before starting anti-tnf therapy may have more to do with the increased accuracy associated with reviewing all the gastroenterology clinical documentation. This potential difference highlights a key issue in the ability to prove adequate compliance with quality measures. Although compliance with these measures is based on documentation of adequate laboratory testing, documentation of these laboratory results may not be in the result section of an EHR, but instead, may be documented in the middle of a clinician s note. Ideally, all quality measures need to be documented in a method that is easily exportable from an EHR to public reporting systems to enable proof of compliance. Unfortunately, without the ability to readily export evidence of compliance, physicians inadvertently may be penalized for providing suboptimal care. The strengths of our study include its multicenter design. We have included all types of clinical practices including academic, community, and private, as well as specialists in IBD and general gastroenterologists. In addition, the cross-sectional design and inclusion of consecutive patients with IBD limited the potential bias of selectively choosing patients to include in this study. The difference in time periods at the centers could lead to bias if there were increased publications about the core measures. However, none of the centers reported any center-specific checklist or discussion of the quality measures before the inclusion periods. The only increased awareness would relate to the AGA publicizing the measures. This may have resulted in some improved compliance at the CP and PP, but our results indicated that the compliance still was inadequate overall. Conclusions Current documented compliance with the AGA quality measures is inadequate across all practice types academic, community hospital, and private practice. Further studies are necessary to improve the reporting of these measures. Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at and at References 1. Blumenthal D. Quality of care what is it? N Engl J Med 1996; 335: Chassin MR, Galvin RW. The urgent need to improve health care quality. Institute of Medicine National Roundtable on Health Care Quality. JAMA 1998;280: Siegel CA, Allen JI, Melmed GY. Translating improved quality of care into an improved quality of life for patients with inflammatory bowel disease. Clin Gastroenterol Hepatol 2013; 11: Mehta RH, Montoye CK, Gallogly M, et al. Improving quality of care for acute myocardial infarction: the Guidelines Applied in Practice (Gap) initiative. JAMA 2002;287: Center for Medicare and Medicaid Services. Medicare and Medicaid Ehr Incentive Program Basics Available: IncentivePrograms/Basics.html. Accessed: November 26, A conceptual use to meaningful use Available: IncentivePrograms/downloads/MU_Stage1_ReqOverview.pdf. Accessed: October 15, Bates DW, Gawande AA. Improving safety with information technology. N Engl J Med 2003;348: Romano MJ, Stafford RS. Electronic health records and clinical decision support systems: impact on national ambulatory care quality. Arch Intern Med 2011;171: Park KT, Bass D. Inflammatory bowel disease-attributable costs and cost-effective strategies in the United States: a review. Inflamm Bowel Dis 2011;17: Feuerstein JD, Cheifetz AS. Miscellaneous adverse events with biologic agents (excludes infection and malignancy). Gastroenterol Clin North Am 2014;43: van Deen WK, van Oijen MGH, Myers KD, et al. A nationwide analysis of U.S. health care utilization in inflammatory bowel diseases. Inflamm Bowel Dis 2014;20: American Gastroenterological Association. Adult inflammatory bowel disease physician performance measures set Available: Measures.pdf. Accessed: January 1, Melmed G, Ozbay AB, Skup M, et al. Analysis of an administrative claims database suggests poor quality of care for

8 428 Feuerstein et al Clinical Gastroenterology and Hepatology Vol. 14, No. 3 inflammatory bowel disease (abstr). Am J Gastroenterol 2014; 109:S Choudhry NK, Fletcher RH, Soumerai SB. Systematic review: the relationship between clinical experience and quality of health care. Ann Intern Med 2005;142: Health Care Incentives Improvement Institute. IBD care recognition program. Available: programs/ibd_care_recognition. Accessed: January 1, Berwick DM. Launching accountable care organizations the proposed rule for the Medicare Shared Savings Program. N Engl J Med 2011;364:e Fisher ES, Shortell SM. Accountable care organizations: accountable for what, to whom, and how. JAMA 2010; 304: Chatterjee P, Joynt KE. Do cardiology quality measures actually improve patient outcomes? J Am Heart Assoc 2014; 3:e Feuerstein JD, Akbari M, Gifford AE, et al. Systematic review: the quality of the scientific evidence and conflicts of interest in international inflammatory bowel disease practice guidelines. Aliment Pharmacol Ther 2013;37: Feuerstein JD, Gifford AE, Akbari M, et al. Systematic analysis underlying the quality of the scientific evidence and conflicts of interest in gastroenterology practice guidelines. Am J Gastroenterol 2013;108: Jha AK, Joynt KE, Orav EJ, et al. The long-term effect of premier pay for performance on patient outcomes. N Engl J Med 2012; 366: Snyder C, Anderson G. Do quality improvement organizations improve the quality of hospital care for Medicare beneficiaries? JAMA 2005;293: US Preventive Services Task Force. Counseling and interventions to prevent tobacco use and tobacco-caused disease in adults and pregnant women: US Preventive Services Task Force Reaffirmation Recommendation Statement. Ann Intern Med 2009;150: Blumenthal D, Tavenner M. The meaningful use regulation for electronic health records. N Engl J Med 2010;363: Reprint requests Address requests for reprints to: Joseph D. Feuerstein, MD, Department of Medicine and Division of Gastroenterology, Beth Israel Deaconess Medical Center, Harvard Medical School, 110 Francis Street, 8E Gastroenterology, Boston, Massachusetts jfeuerst@bidmc.harvard.edu; fax: (617) Conflicts of interest These authors disclose the following: Daniel Leffler has performed consulting or received grants from Prometheus Diagnostics, Alba Pharmaceuticals, Alvine Therapeutics, and Shire Therapeutics; and Adam Cheifetz has performed consulting or received grants from AbbVie, Janssen Pharmaceuticals, Pfizer, Takeda, and Prometheus. The remaining authors disclose no conflicts.

9 March 2016 Documented Reports With IBD Quality Measures 428.e1 Supplementary Table 1. Grouping of Patients Based on Applicable Core Measures Groups Academic practice (n ¼ 367) Community practice (n ¼ 217) Private practice (n ¼ 199) Total (n ¼ 783) P value Group 1, n (%) 68 (18.5) 120 (55.3) 81 (40.7) 269 (34.3) <.001 a Group 2, n (%) 80 (21.8) 36 (16.6) 10 (5.0) 126 (16.1) <.001 b Group 3, n (%) 103 (28.1) 37 (17.1) 28 (14.1) 168 (21.5) <.001 c Group 4, n (%) 116 (31.6) 24 (11.1) 80 (40.2) 220 (28.1) <.001 d a Pairwise comparisons were significant for all comparisons: AP vs CP, AP vs PP, and CP vs PP, with a P value of less than.016. b Pairwise comparisons were significant for AP vs PP and CP vs PP, with a P value of less than.016. c Pairwise comparisons were significant for AP vs CP and AP vs PP, with a P value of less than.016. d Pairwise comparisons were significant for AP vs CP and CP vs PP, with a P value of less than.016. Supplementary Table 2. Patients With All Applicable Core Measures Reported Patients with all core measures (n ¼ 46) Patients without all measures evaluated (n ¼ 738) BIDMC, n (%) 24 (52.2) 343 (46.8) Mount Auburn, n (%) (29.4) Rockland, n (%) 22 (47.8) 178 (24.0)

10 428.e2 Feuerstein et al Clinical Gastroenterology and Hepatology Vol. 14, No. 3 Supplementary Table 3. Demographics of 46 Patients With All Applicable Core Measures Achieved Sex, n 56.5 (26/46) female 43.5 (20/46) male Type of IBD, n Ulcerative colitis, Crohn s disease, 45.6 (21/46) 54.4 (25/46) Yes No IBD attending, n a 96 (23/24) 4 (1/24) Seen by PCP a 75 (18/24) 25 (6/24) Seen by gastroenterology fellow a 33.3 (8/24) 66.6 (16/24) PCP, primary care physician. a IBD attending, seen PCP and seen by gastroenterology fellow data only from BIDMC.