The Rome IV Diagnostic Questionnaires were developed to screen for functional gastrointestinal disorders, serve as inclusion criteria in clinical trials, and support epidemiologic surveys. Separate questionnaires were developed for adults, children and adolescents, and infants and toddlers. For the adult questionnaire, we first surveyed 1162 adults without gastrointestinal disorders, and recommended the 90th percentile symptom frequency as the threshold for defining what is abnormal. Diagnostic questions were formulated and verified with clinical experts using a recursive process. The diagnostic sensitivity of the questionnaire was tested in 843 patients from 9 gastroenterology clinics, with a focus on clinical diagnoses of irritable bowel syndrome (IBS), functional constipation (FC), and functional dyspepsia (FD). Sensitivity was 62.7% for IBS, 54.7% for FD, and 32.2% for FC. Specificity, assessed in a population sample of 5931 adults, was 97.1% for IBS, 93.3% for FD, and 93.6% for FC. Excess overlap among IBS, FC, and FD was a major contributor to reduced diagnostic sensitivity, and when overlap of IBS with FC was permitted, sensitivity for FC diagnosis increased to 73.2%. All questions were understandable to at least 90% of individuals, and Rome IV diagnoses were reproducible in three-fourths of patients after 1 month. Validation of the pediatric questionnaires is ongoing. The Rome IV Diagnostic Questionnaires were developed to screen for functional gastrointestinal disorders, serve as inclusion criteria in clinical trials, and support epidemiologic surveys. Separate questionnaires were developed for adults, children and adolescents, and infants and toddlers. For the adult questionnaire, we first surveyed 1162 adults without gastrointestinal disorders, and recommended the 90th percentile symptom frequency as the threshold for defining what is abnormal. Diagnostic questions were formulated and verified with clinical experts using a recursive process. The diagnostic sensitivity of the questionnaire was tested in 843 patients from 9 gastroenterology clinics, with a focus on clinical diagnoses of irritable bowel syndrome (IBS), functional constipation (FC), and functional dyspepsia (FD). Sensitivity was 62.7% for IBS, 54.7% for FD, and 32.2% for FC. Specificity, assessed in a population sample of 5931 adults, was 97.1% for IBS, 93.3% for FD, and 93.6% for FC. Excess overlap among IBS, FC, and FD was a major contributor to reduced diagnostic sensitivity, and when overlap of IBS with FC was permitted, sensitivity for FC diagnosis increased to 73.2%. All questions were understandable to at least 90% of individuals, and Rome IV diagnoses were reproducible in three-fourths of patients after 1 month. Validation of the pediatric questionnaires is ongoing. The Rome working teams, which are composed of clinical investigators and clinicians who are experts in the functional gastrointestinal disorders (FGIDs) that affect specific regions of the gut, devise diagnostic criteria for these disorders that are intended for use by other clinicians and researchers (not patients). Their goal is to make the criteria as sensitive and specific as possible, and to accomplish this they sometimes develop complex diagnostic algorithms and combine requirements for laboratory evaluations with symptom criteria. However, there is also a need for a questionnaire that translates the diagnostic criteria into questions that are understandable to most patients, in order to enable standardized diagnostic assessment of individuals. Consequently, the Rome Foundation appointed a Questionnaire Development Committee (QDC) of individuals with expertise in test development to develop a patient questionnaire that incorporates the Rome diagnostic criteria. These questions can be used as inclusion criteria in clinical trials, as case definitions in epidemiological surveys, or for clinic screening. The mandate of the QDC was to develop the Rome IV Diagnostic Questionnaire for Adults based on the new Rome IV criteria and to assess its performance with respect to understandability by patients, test−retest reliability, concordance with independent diagnoses by experienced clinicians, and ability to discriminate patients with the 3 most common FGIDs, which are irritable bowel syndrome (IBS), functional constipation (FC), and functional dyspepsia (FD), from nonpatient controls recruited from the population. Two pediatric questionnaires were also developed by a QDC subcommittee, but testing of these is still underway and will not be described here. The process of developing and validating the Rome IV Diagnostic Questionnaire for Adults consisted of 5 different project steps, and these will be described in sequence. The symptoms that the Rome working teams identified as diagnostic of FGIDs are not pathognomonic for GI disease; they are symptoms such as abdominal pain, nausea, vomiting, and heartburn that also occur occasionally in healthy nonpatients. These symptoms are only considered clinically significant and indicative of an FGID if they occur at an abnormally high frequency. This observation has 2 important consequences for the design of symptom-based diagnostic questionnaires: (1) the response scales used for questions about the frequency of symptom occurrence must have small enough steps to capture clinically significant differences between individuals, and (2) the thresholds used to define what is an abnormal frequency of occurrence will vary for different symptoms because the normal frequency of occurrence of these symptoms differs. To address this, the QDC first developed and validated new response scales for the Rome IV diagnostic questionnaire, with more response steps than the Rome III.1Whitehead W.E. Palsson O.S. Cascade E. Validation of Response Scales for ROME Diagnostic Questionnaire.Gastroenterology. 2013; 144 (S-916)Google Scholar The committee then conducted a survey of a nationally representative sample of US adults in order to provide the Rome IV working teams with the data needed to set thresholds for identifying meaningful, clinically significant deviations from the normal frequency of occurrence of GI symptoms. A sample of 1665 US adults stratified by sex (50% males), age, race, and ethnicity was recruited by a market research firm, CINT USA, Inc. (Los Angeles, CA) to complete an Internet survey. After inconsistent responders (identified by 3 repeated survey questions) were eliminated, response sets from 1277 individuals were retained for analysis. For the purpose of setting frequency thresholds for defining what should be considered abnormal, we identified the 90th percentile for all questions (males and females combined) and reported these to the committees as recommended thresholds. We reasoned that using these thresholds for diagnoses would result in no more than 10% of healthy subjects being misclassified as patients, and when these 90th percentile thresholds were combined for multiple symptoms used to diagnose a disorder, the specificity could be expected to be in excess of 90%. In the calculation of these cutoffs, we excluded subjects who reported having a prior medical diagnosis of upper GI diagnoses from analysis of upper GI symptom thresholds, and conversely excluded those reporting lower GI diagnoses from analysis of lower GI symptoms. Figure 1 illustrates how these data were analyzed: The cardinal symptom defining IBS is the frequency of abdominal pain, and the figure shows a histogram of the frequency of all responses to the question on abdominal pain. If the threshold for clinical significance is set at once a week as the Rome IV Bowel Committee recommended, the proportion of these population controls who might be misclassified as IBS is only 6.7%, indicating that the specificity of this symptom criterion for IBS diagnosis is 93.3%. Figure 1 also shows how the specificity of the diagnosis would be impacted if an alternative threshold for clinical significance were selected, and it shows that abdominal pain or discomfort is reported significantly more frequently by females than by males. Because males have a lower prevalence of abdominal pain and discomfort than females, the 90th percentile for males could be set at 2 to 3 days per month rather than once a week, and the 90th percentile for females occurs once a week, as is the case for the combined sample. For simplicity, the QDC recommended that the frequency thresholds for IBS diagnosis should be based on the 90th percentile for women and men combined. A summary report2Whitehead W.E. Palsson O.S. Report on the Rome III Normative Gastrointestinal Symptom Survey. University of North Carolina, Chapel Hill, NC2013Google Scholar on the distribution of symptom occurrence rates for all the Rome III symptoms was distributed to the Rome IV working teams together with recommendations for selecting frequency thresholds for diagnosis. This report is available as a supplement to this article. Most of the working teams adopted these suggestions. The 6 working teams that were tasked with updating the Rome diagnostic criteria for the FGIDs in each region of the GI tract were appointed in 2013 and were requested to complete draft documents, including revised diagnostic criteria by May 2014. These committees worked by e-mail and conference calls, and met together for the first time at a satellite meeting held in conjunction with Digestive Disease Week 2014. Previous to this, the QDC completed its survey of the base rates of symptom occurrence in the population and also revised and validated new, more sensitive response scales1Whitehead W.E. Palsson O.S. Cascade E. Validation of Response Scales for ROME Diagnostic Questionnaire.Gastroenterology. 2013; 144 (S-916)Google Scholar for patients to use when reporting their symptoms. The QDC developed a draft of the Rome IV Diagnostic Questionnaire (R4DQ) and distributed the questions appropriate to each region of the GI tract to the chair and co-chair of the working teams immediately before 3-day meetings of all the committees in December 2014. The chair and co-chair of the QDC met with each committee to discuss how well the draft diagnostic questions embodied the revised diagnostic criteria. The QDC’s interactions with the working teams were iterative: When the working teams revised their diagnostic criteria based on feedback from the Rome Foundation’s Editorial Committee, or the critiques submitted by outside reviewers, the QDC revised the draft diagnostic questions. After each revision to either the diagnostic criteria or the diagnostic questions, the QDC asked the committees to again review the diagnostic questions for consistency with the diagnostic criteria. When the QDC was confident that the diagnostic questionnaire was in a near-final form, it was submitted to the professional translation company, Transperfect Inc. (New York, NY), for an assessment of the translatability into the following major languages: Arabic, French, German, Hindi, Italian, Japanese, Korean, Russian, and simplified Chinese. This translatability assessment resulted in no changes to the English version of the questionnaire, but did provide guidance that will be useful to the translators when the Rome IV questionnaire is translated into these and other languages. In order to assess the understandability of the R4DQ, the QDC recruited a new US general population sample of 589 adult subjects stratified by sex, age, race, ethnicity, and years of education to complete an Internet survey. Each subject was asked to evaluate one-third (approximately 28) of the total questions. Subjects first answered each symptom question and were then asked (1) whether the question was difficult to understand (yes or no) and, if the question was difficult, (2) what about the question made it difficult to understand and (3) to give suggestions for alternative more understandable wording. On average, 1.5 questions of 28 were rated as difficult to understand, and neither older age nor fewer years of education were associated with the number of questions rated difficult to understand. Seven questions were rated difficult by 10% or more subjects, and these were revised to enhance their simplicity and clarity based on the suggestions of the subjects. The QDC concluded that all questions in the final form of the R4DQ are understandable to >90% of US adults, and that understandability is not influenced by age or education. After completing the understandability and translatability analyses and making minor changes to the questionnaire, the chair and co-chair of the QDC committee again compared the questions with the Rome IV diagnostic criteria and developed scoring criteria for assigning provisional diagnoses based on questionnaire responses. This review identified a need for the working teams to make minor changes to the wording of the diagnostic criteria to ensure agreement with the questionnaire, and these suggestions were communicated to the working teams and incorporated into the Rome IV diagnostic criteria. The preferred method for assessing the accuracy of a new diagnostic test is to compare classifications based on the new test to classifications based on an objectively measured biological marker that is known to be related to the pathophysiology of the disease. However, this is not possible for the FGIDs because there is no consensus on specific pathophysiological mechanisms for them, and no biomarkers exist that can identify the FGIDs with acceptable precision.3Spiller R.C. Potential biomarkers.Gastroenterol Clin North Am. 2011; 40: 121-139Abstract Full Text Full Text PDF PubMed Scopus (22) Google Scholar Another barrier to validation of symptom criteria for the FGIDs is that similar symptoms (eg, abdominal pain, nausea, vomiting, constipation, and diarrhea) are present in other diseases; examples are inflammatory bowel disease, gastroesophageal reflux disease, celiac disease, and GI cancers. Many view the exclusion of other disease explanations for symptoms as necessary for the diagnosis of the FGIDs. In the absence of an objective reference standard for diagnosis of FGIDs, 2 models of validation have been used4Whitehead W.E. Drossman D.A. Validation of symptom-based diagnostic criteria for irritable bowel syndrome: a critical review.Am J Gastroenterol. 2010; 105 (quiz 813, 821): 814-820Crossref PubMed Scopus (86) Google Scholar: In the first model, the reference standard is a negative endoscopy, alternative imaging study, or laboratory test to exclude other diseases that could cause the symptoms, and the measure of diagnostic accuracy is whether the symptom criteria correctly identify patients who do not have an alternative basis for their symptoms.5Ford A.C. Talley N.J. Veldhuyzen van Zanten S.J. et al.Will the history and physical examination help establish that irritable bowel syndrome is causing this patient's lower gastrointestinal tract symptoms?.JAMA. 2008; 300: 1793-1805Crossref PubMed Scopus (90) Google Scholar, 6Sood R. Gracie D.J. Law G.R. et al.Systematic review with meta-analysis: the accuracy of diagnosing irritable bowel syndrome with symptoms, biomarkers and/or psychological markers.Aliment Pharmacol Ther. 2015; 42: 491-503Crossref PubMed Scopus (58) Google Scholar The second model uses positive diagnoses made by experienced clinicians as the reference standard, and the measure of diagnostic accuracy is whether the symptom criteria are concordant with the clinical diagnosis.7Whitehead W.E. Palsson O.S. Feld A.D. et al.Utility of red flag symptom exclusions in the diagnosis of irritable bowel syndrome.Aliment Pharmacol Ther. 2006; 24: 137-146Crossref PubMed Scopus (87) Google Scholar, 8Whitehead W.E. Development and validation of the Rome III diagnostic questionnaire.in: Drossman D.A. Corazziari E. Delvaux M. Rome III: The Functional Gastrointestinal Disorders. Degnon Associates, McLean, VA2006: 835-853Google Scholar A third model is a hybrid of the first 2; for example, Vanner et al9Vanner S.J. Depew W.T. Paterson W.G. et al.Predictive value of the Rome criteria for diagnosing the irritable bowel syndrome.AmJ Gastroenterol. 1999; 94: 2912-2917Crossref PubMed Scopus (236) Google Scholar first excluded from consideration any patient with alarm signs or symptoms suggestive of possible structural disease (eg, blood in stools or family history of GI cancer) and then examined the agreement of the Rome II symptom criteria with clinical diagnosis; he found that the positive predictive value of the Rome II criteria was 100%. The QDC utilized a hybrid model for validation of the criteria for IBS and FD: To be included as a reference case of IBS, patients must have had a negative endoscopy within the last 5 years as well as a positive clinical diagnosis; and to be included as a reference case of FD, patients must have had a negative upper endoscopy within the last 5 years plus a positive clinical diagnosis. However, red flag signs or alarm symptoms, such as blood in stools or a family history of colon cancer, did not result in exclusion of the patient. The primary goal of the validation study was to test the sensitivity of the R4DQ for identifying patients who were diagnosed IBS, FC, or FD by clinicians. These reference clinical diagnoses were made by experienced clinicians before the patients completing the R4DQ, thus guaranteeing that the reference diagnoses were independent. In addition, a diagnosis of IBS required that the patient have a colonoscopy within the last 5 years, and a diagnosis of FD required that the patient have an upper endoscopy within the same time frame. Additional goals of this study were to assess the test−retest reliability of the R4DQ by having approximately 30 patients at each site complete the R4DQ a second time after a 30-day interval, and to examine the overlap between FGIDs diagnosed by the R4DQ. The study was conducted at 9 clinical sites managed by academic gastroenterologists who are familiar with the FGIDs and with previous versions of the Rome criteria. Each site was asked to recruit 100−150 patients, including 25% with a clinical diagnosis of IBS, 25% with a clinical diagnosis of FC, 25% with a clinical diagnosis of FD, and 25% with other FGID diagnoses. The purpose of concentrating recruitment on IBS, FC, and FD was to insure that sufficient numbers of patients could be recruited to have adequate statistical power to test sensitivity. These are the most common FGID diagnoses made in clinical practice. To participate, subjects were required to have been diagnosed with an FGID in a medical clinic, have personal access to an Internet-connected computer or tablet and be able to use it to answer questionnaires online; be able read and write English fluently; be at least 18 years old; have had GI symptoms for at least 6 months; and for the subsample of patients with IBS or FD clinical diagnosis, have had an endoscopy (colonoscopy for IBS, upper GI endoscopy for FD) with negative findings within the past 5 years. Individuals who had been diagnosed with the following organic health problems likely to affect GI symptoms were excluded from participation in the study: inflammatory bowel disease (Crohn’s disease or ulcerative colitis), cancer anywhere in the GI tract, current infection of the GI tract, celiac disease, diabetes mellitus, and/or an eating disorder. Individuals who had undergone bariatric surgery or resection of any part of their bowels except appendix or gallbladder operations were also excluded from participation. The research coordinator at each site reviewed the medical records for the past 3−6 months to retrospectively identify all patients assigned a clinical diagnosis of any FGID. At sites where this retrospective medical record review did not yield enough patients in all categories, prospective enrollment of new patients was permitted. Enrollment and data collection occurred between April 1 and September 15, 2015. Research coordinators at the clinical sites sent e-mails or letters to eligible patients explaining the purpose of the study. Subjects were given the web address where they could complete the questionnaire and, to ensure that the database contained no information that could be used to identify them, they were assigned a randomly selected ID and a unique password to enter when they took the questionnaire. When potential study participants accessed the study website, they first reviewed a study description and recorded their consent before completing the online study questionnaire. The questionnaire included the Rome IV Diagnostic Questionnaire (from 26 to 86 questions depending on skip patterns); the Rome III Diagnostic Questionnaire modules for IBS, FC, and FD (up to 27 questions); 6 demographic questions; and 6 questions about the frequency and types of medications used for GI symptoms. They were also asked about excluded diagnoses to confirm eligibility. Completion of the online questionnaires required 15 to 20 minutes, and patients received $25 for participation. The research coordinator at each site had access on the study website to a password-protected study-management interface to view a list of the IDs of all patients from their site who had completed the questionnaire. The coordinators were encouraged to contact all subjects who had not completed the questionnaires within 2 weeks to remind them to do so. Research coordinators entered the primary FGID clinical diagnosis (reference diagnosis) into the website interface for all subjects who completed the questionnaire. The coordinators also reviewed the medical records of questionnaire completers and abstracted data on all other GI diagnoses and relevant medical tests, such as endoscopy, transit study, gastric emptying study, esophageal pH study, esophageal manometry, barium enema or magnetic resonance imaging of the colon, and fecal stool test for ova/parasites. The research coordinators randomly selected a subset of questionnaire completers with clinical diagnoses of IBS, FD, and functional constipation and invited them to complete the questionnaire a second time 30 days after the first completion, for an additional payment of $25. Reminders to complete the repeat test were sent by e-mail as needed. For the primary analysis of diagnostic test sensitivity, the data from all clinical sites were pooled. Sensitivity and 95% confidence intervals for sensitivity were computed for each diagnosis (IBS, FC, and FD) separately. Sensitivity was defined as the proportion of all patients with a primary clinical diagnosis of the index disorder who fulfilled Rome IV criteria for the same diagnosis based on questionnaire responses. Patients receiving a secondary clinical diagnosis of the index disorder were excluded from the analysis due to a concern that clinicians may vary in the amount of time and effort devoted to secondary diagnoses in a busy clinic and may make secondary diagnoses of IBS or FD without endoscopy, thus causing secondary diagnoses to be less reliable. Specificity, which is defined as the proportion of all patients without the index clinical diagnosis who also do not fulfill Rome IV questionnaire criteria for the index diagnosis, was not a goal of the analysis of data collected from the clinical sites for the following reasons: (1) The ability to discriminate patients with the index FGID from patients with other FGIDs is less relevant to the performance of the diagnostic questionnaire than is the ability to discriminate true cases from general population controls because it is known that there is a high degree of overlap between FGIDs in specialty medical clinics. (2) The design of the study was not appropriate for estimating the ability to discriminate the index disorder from other FGIDs even in the medical clinic because the patient sample was not representative of the clinic population; instead, patients were selectively recruited based on having established clinical diagnoses of specific FGIDs. Consequently, we estimated specificity of the Rome IV criteria from a large, representative population sample (see Three Country Population Survey), which was recruited specifically for this purpose. Secondary sensitivity analyses were performed to assess the influence of symptom frequency thresholds and other specifics of the diagnostic criteria on test accuracy. Additionally, the degree of overlap between FGID diagnoses based on Rome IV criteria was calculated. In this analysis, all FGID diagnoses (primary and secondary) recorded in the medical record were included. Test−retest reliability was assessed for each of the 3 key FGID diagnoses in the study by computing percent agreement and κ statistics between 2 administrations of the questionnaire to the same subjects approximately 1 month apart, for each of the 3 primary diagnoses separately. A total of 881 patients enrolled and completed the study questionnaire across the 9 sites, but 38 were disqualified because they were found not to meet study criteria after enrollment, leaving 843 for analysis. Females comprised 76.3% of the 843 evaluable patients, and age ranged from 18 to 81 years (mean 43.6 years). Race/ethnicity distribution was 85.3% white, 4.3% black, 4.5% Asian, and 6.0% other or not disclosed. Among US patients, 8.2% reported Hispanic ethnicity. Table 1 shows the frequency of all FGID clinical diagnoses (not Rome IV diagnoses based on the questionnaire) found in the medical records of the 843 patients in the analysis sample.Table 1Frequency of All Clinical Functional Gastrointestinal Disorder Diagnoses in the Sample of 843 Gastrointestinal Patients, Including Multiple Diagnoses for the Same PatientFGID diagnosisNo. of cases% of Total sampleIrritable bowel syndrome44252.4Functional constipation23527.9Functional dyspepsia15318.1Functional bloating9611.4Functional diarrhea799.4Functional abdominal pain718.4Functional fecal incontinence556.5Chronic idiopathic nausea344Dysphagia131.5Globus111.3Functional heartburn111.3Functional vomiting91.1Unspecified belching70.8Cyclic vomiting70.8Functional chest pain50.6Proctalgia fugax30.4Rumination syndrome20.2Sphincter of Oddi dysfunction20.2Chronic proctalgia20.2Aerophagia10.1Levator ani10.1 Open table in a new tab Columns 1−3 in Table 2 show the sensitivity statistics for the 3 most prevalent FGIDS: IBS, FC, and FD. (The data on specificity in columns 4−5 come from the Three Country Population Survey and will be discussed later.)Table 2Sensitivity and Specificity of the 3 Most Common Functional Gastrointestinal DisordersDiagnosisClinical validation samplePopulation control sampleSensitivity, %95% CISpecificity, %95% CIIBS62.757.8−67.697.196.6−97.6FD54.746.3−63.193.392.5−94.0FC excluding OIC and IBS33.927.0−40.894.593.9−95.2FC including OIC and IBS70.563.8−77.293.192.5−93.9CI, confidence interval. Open table in a new tab CI, confidence interval. The sensitivity of the Rome IV criteria for IBS is comparable with levels previously reported for Rome III,4Whitehead W.E. Drossman D.A. Validation of symptom-based diagnostic criteria for irritable bowel syndrome: a critical review.Am J Gastroenterol. 2010; 105 (quiz 813, 821): 814-820Crossref PubMed Scopus (86) Google Scholar but is suboptimal. The QDC examined the impact on diagnostic sensitivity of each of the 3 changes made to the Rome III criteria:1.The threshold frequency of abdominal pain was made more stringent, changing from 2−3 days per month in Rome III to at least once a week in Rome IV. If the frequency of abdominal pain was relaxed to 2−3 days per month and all other Rome IV criteria remained the same, this would increase the sensitivity of Rome IV to 69.9%.2.“Discomfort” was dropped from the key qualifying symptom, “abdominal discomfort or pain”; Rome IV requires “abdominal pain.” However, this more restrictive phrasing did not reduce diagnostic sensitivity: The proportion of patients with a primary clinical diagnosis of IBS who reported “abdominal discomfort or pain” at least weekly on the Rome III pain question was similar to the proportion who reported “abdominal pain” at least weekly on the equivalent Rome IV question (74.9% vs 78.4% respectively).3.Rome III required that pain or discomfort improve after defecation and that changes in stool frequency and/or consistency occur “when the pain started.” By contrast, Rome IV requires only that pain and defecation are “associated” in time or that changes in stool consistency and frequency are “associated” with abdominal pain. However, this change did not have a significant impact on the sensitivity of the Rome criteria. To understand the causes of misclassifications in IBS, we examined which Rome IV diagnoses were assigned by the questionnaire to the 140 patients who received a clinical diagnosis of IBS but did not meet Rome IV criteria for IBS. The most common Rome IV diagnoses were FC (35 of 140), FD (34 of 140), functional diarrhea (22 of 140), and levator ani syndrome (20 of 140). We also tabulated the clinical diagnoses assigned to 192 patients who met the Rome IV criteria for IBS but did not receive a clinical diagnosis of IBS by their physician. The most common clinical diagnoses (either primary or secondary diagnosis) were FC (95 of 192), FD (58 of 192), functional bloating (27 of 192), functional abdominal pain (25 of 192), and functional diarrhea (17 of 192). These data show that the core symptoms that combine to diagnose IBS are the ones accounting for most misclassifications, and they suggest that clinicians likely base their clinical diagnoses on the predominant or most bothersome symptom when patients present with the constellation of constipation, diarrhea, abdominal pain, and bloating. A significant number of misclassifications also result from the overlap of FD with IBS. The Rome IV criteria for unspecified FD had a sensitivity of 54.7% (
