Computer-generated vs. physician-documented history of present illness (HPI): results of a blinded comparison

Are medical history data fit for risk stratification of patients with chest pain in emergency care? Comparing data collected from patients using computerized history taking with data documented by physicians in the electronic health record in the CLEOS-CPDS prospective cohort study

OBJECTIVE

In acute chest pain management, risk stratification tools, including medical history, are recommended. We compared the fraction of patients with sufficient clinical data obtained using computerized history taking software (CHT) versus physician-acquired medical history to calculate established risk scores and assessed the patient-by-patient agreement between these 2 ways of obtaining medical history information.

MATERIALS AND METHODS

This was a prospective cohort study of clinically stable patients aged ≥ 18 years presenting to the emergency department (ED) at Danderyd University Hospital (Stockholm, Sweden) in 2017-2019 with acute chest pain and non-diagnostic ECG and serum markers. Medical histories were self-reported using CHT on a tablet. Observations on discrete variables in the risk scores were extracted from electronic health records (EHR) and the CHT database. The patient-by-patient agreement was described by Cohen's kappa statistics.

RESULTS

Of the total 1000 patients included (mean age 55.3 ± 17.4 years; 54% women), HEART score, EDACS, and T-MACS could be calculated in 75%, 74%, and 83% by CHT and in 31%, 7%, and 25% by EHR, respectively. The agreement between CHT and EHR was slight to moderate (kappa 0.19-0.70) for chest pain characteristics and moderate to almost perfect (kappa 0.55-0.91) for risk factors.

CONCLUSIONS

CHT can acquire and document data for chest pain risk stratification in most ED patients using established risk scores, achieving this goal for a substantially larger number of patients, as compared to EHR data. The agreement between CHT and physician-acquired history taking is high for traditional risk factors and lower for chest pain characteristics.

CLINICAL TRIAL REGISTRATION

ClinicalTrials.gov NCT03439449.

Effects of Combinational Use of Additional Differential Diagnostic Generators on the Diagnostic Accuracy of the Differential Diagnosis List Developed by an Artificial Intelligence-Driven Automated History-Taking System: Pilot Cross-Sectional Study

BACKGROUND

Low diagnostic accuracy is a major concern in automated medical history-taking systems with differential diagnosis (DDx) generators. Extending the concept of collective intelligence to the field of DDx generators such that the accuracy of judgment becomes higher when accepting an integrated diagnosis list from multiple people than when accepting a diagnosis list from a single person may be a possible solution.

OBJECTIVE

The purpose of this study is to assess whether the combined use of several DDx generators improves the diagnostic accuracy of DDx lists.

METHODS

We used medical history data and the top 10 DDx lists (index DDx lists) generated by an artificial intelligence (AI)-driven automated medical history-taking system from 103 patients with confirmed diagnoses. Two research physicians independently created the other top 10 DDx lists (second and third DDx lists) per case by imputing key information into the other 2 DDx generators based on the medical history generated by the automated medical history-taking system without reading the index lists generated by the automated medical history-taking system. We used the McNemar test to assess the improvement in diagnostic accuracy from the index DDx lists to the three types of combined DDx lists: (1) simply combining DDx lists from the index, second, and third lists; (2) creating a new top 10 DDx list using a 1/n weighting rule; and (3) creating new lists with only shared diagnoses among DDx lists from the index, second, and third lists. We treated the data generated by 2 research physicians from the same patient as independent cases. Therefore, the number of cases included in analyses in the case using 2 additional lists was 206 (103 cases × 2 physicians' input).

RESULTS

The diagnostic accuracy of the index lists was 46% (47/103). Diagnostic accuracy was improved by simply combining the other 2 DDx lists (133/206, 65%, P<.001), whereas the other 2 combined DDx lists did not improve the diagnostic accuracy of the DDx lists (106/206, 52%, P=.05 in the collective list with the 1/n weighting rule and 29/206, 14%, P<.001 in the only shared diagnoses among the 3 DDx lists).

CONCLUSIONS

Simply adding each of the top 10 DDx lists from additional DDx generators increased the diagnostic accuracy of the DDx list by approximately 20%, suggesting that the combinational use of DDx generators early in the diagnostic process is beneficial.

Evaluating an app for digital medical history taking in urgent care practices: study protocol of the cluster-randomized interventional trial 'DASI'

BACKGROUND

In out-of-hours urgent care practices in Germany, physicians of different specialties care for a large number of patients, most of all unknown to them, resulting in a high workload and challenging diagnostic decision-making. As there is no common patient file, physicians have no information about patients' previous conditions or received treatments. In this setting, a digital tool for medical history taking could improve the quality of medical care. This study aims to implement and evaluate a software application (app) that takes a structured symptom-oriented medical history from patients in urgent care settings.

METHODS

We conduct a time-cluster-randomized trial in two out-of-hours urgent care practices in Germany for 12 consecutive months. Each week during the study defines a cluster. We will compare participants with (intervention group) and without app use (control group) prior to consultation and provision of the self-reported information for the physician. We expect the app to improve diagnostic accuracy (primary outcome), reduce physicians' perceived diagnostic uncertainty, and increase patients' satisfaction and the satisfaction with communication of both physician and patient (secondary outcomes).

DISCUSSION

While similar tools have only been subject to small-scale pilot studies surveying feasibility and usability, the present study uses a rigorous study design to measure outcomes that are directly associated with the quality of delivered care.

TRIAL REGISTRATION

The study was registered at the German Clinical Trials Register (No. DRKS00026659 registered Nov 03 2021. World Health Organization Trial Registration Data Set, https://trialsearch.who.int/Trial2.aspx? TrialID = DRKS00026659.

Comparison of clinical note quality between an automated digital intake tool and the standard note in the emergency department

BACKGROUND

Medical encounters require an efficient and focused history of present illness (HPI) to create differential diagnoses and guide diagnostic testing and treatment. Our aim was to compare the HPI of notes created by an automated digital intake tool versus standard medical notes created by clinicians.

METHODS

Prospective trial in a quaternary academic Emergency Department (ED). Notes were compared using the 5-point Physician Documentation Quality Instrument (PDQI-9) scale and the Centers for Medicare & Medicaid Services (CMS) level of complexity index. Reviewers were board certified emergency medicine physicians blinded to note origin. Reviewers received training and calibration prior to note assessments. A difference of 1 point was considered clinically significant. Analysis included McNemar's (binary), Wilcoxon-rank (Likert), and agreement with Cohen's Kappa.

RESULTS

A total of 148 ED medical encounters were charted by both digital note and standard clinical note. The ability to capture patient information was assessed through comparison of note content across paired charts (digital-standard note on the same patient), as well as scores given by the reviewers. Reviewer agreement was kappa 0.56 (CI 0.49-0.64), indicating moderate level of agreement between reviewers scoring the same patient chart. Considering all 18 questions across PDQI-9 and CMS scales, the average agreement between standard clinical note and digital note was 54.3% (IQR 44.4-66.7%). There was a moderate level of agreement between content of standard and digital notes (kappa 0.54, 95%CI 0.49-0.60). The quality of the digital note was within the 1 point clinically significant difference for all of the attributes, except for conciseness. Digital notes had a higher frequency of CMS severity elements identified.

CONCLUSION

Digitally generated clinical notes had moderate agreement compared to standard clinical notes and within the one point clinically significant difference except for the conciseness attribute. Digital notes more reliably documented billing components of severity. The use of automated notes should be further explored to evaluate its utility in facilitating documentation of patient encounters.

Need assessment for history-taking instruction program using chatbot for nursing students: A qualitative study using focus group interviews

Purpose

A comprehensive health history contributes to identifying the most appropriate interventions and care priorities. However, history-taking is challenging to learn and develop for most nursing students. Chatbot was suggested by students to be used in history-taking training. Still, there is a lack of clarity regarding the needs of nursing students in these programs. This study aimed to explore nursing students' needs and essential components of chatbot-based history-taking instruction program.

Methods

This was a qualitative study. Four focus groups, with a total of 22 nursing students, were recruited. Colaizzi's phenomenological methodology was used to analyze the qualitative data generated from the focus group discussions.

Results

Three main themes and 12 subthemes emerged. The main themes included limitations of clinical practice for history-taking, perceptions of chatbot used in history-taking instruction programs, and the need for history-taking instruction programs using chatbot. Students had limitations in clinical practice for history-taking. When developing chatbot-based history-taking instruction programs, the development should reflect students' needs, including feedback from the chatbot system, diverse clinical situations, chances to practice nontechnical skills, a form of chatbot (i.e., humanoid robots or cyborgs), the role of teachers (i.e., sharing experience and providing advice) and training before the clinical practice.

Conclusion

Nursing students had limitations in clinical practice for history-taking and high expectations for chatbot-based history-taking instruction programs.

Digital Tools Designed to Obtain the History of Present Illness From Patients: Scoping Review

Background

Many medical conditions, perhaps 80% of them, can be diagnosed by taking a thorough history of present illness (HPI). However, in the clinical setting, situational factors such as interruptions and time pressure may cause interactions with patients to be brief and fragmented. One solution for improving clinicians’ ability to collect a thorough HPI and maximize efficiency and quality of care could be to use a digital tool to obtain the HPI before face-to-face evaluation by a clinician.

Objective

Our objective was to identify and characterize digital tools that have been designed to obtain the HPI directly from patients or caregivers and present this information to clinicians before a face-to-face encounter. We also sought to describe outcomes reported in testing of these tools, especially those related to usability, efficiency, and quality of care.

Methods

We conducted a scoping review using predefined search terms in the following databases: MEDLINE, CINAHL, PsycINFO, Web of Science, Embase, IEEE Xplore Digital Library, ACM Digital Library, and ProQuest Dissertations & Theses Global. Two reviewers screened titles and abstracts for relevance, performed full-text reviews of articles meeting the inclusion criteria, and used a pile-sorting procedure to identify distinguishing characteristics of the tools. Information describing the tools was primarily obtained from identified peer-reviewed sources; in addition, supplementary information was obtained from tool websites and through direct communications with tool creators.

Results

We identified 18 tools meeting the inclusion criteria. Of these 18 tools, 14 (78%) used primarily closed-ended and multiple-choice questions, 1 (6%) used free-text input, and 3 (17%) used conversational (chatbot) style. More than half (10/18, 56%) of the tools were tailored to specific patient subpopulations; the remaining (8/18, 44%) tools did not specify a target subpopulation. Of the 18 tools, 7 (39%) included multilingual support, and 12 (67%) had the capability to transfer data directly into the electronic health record. Studies of the tools reported on various outcome measures related to usability, efficiency, and quality of care.

Conclusions

The HPI tools we identified (N=18) varied greatly in their purpose and functionality. There was no consensus on how patient-generated information should be collected or presented to clinicians. Existing tools have undergone inconsistent levels of testing, with a wide variety of different outcome measures used in evaluation, including some related to usability, efficiency, and quality of care. There is substantial interest in using digital tools to obtain the HPI from patients, but the outcomes measured have been inconsistent. Future research should focus on whether using HPI tools can lead to improved patient experience and health outcomes, although surrogate end points could instead be used so long as patient safety is monitored.

User experience of self-reported computerized medical history taking for acute chest pain: The Clinical Expert Operating System Chest Pain Danderyd Study

BACKGROUND AND OBJECTIVE

Chest pain is one of the most common complaints in emergency departments (EDs). Self-reported computerized history taking (CHT) programmes can be used for interpretation of the clinical significance of medical information coming directly from patients. The adoption of CHT in clinical practice depends on reactions and attitudes to the technology from patients and their belief that the technology will have benefits for their medical care. The study objective was to explore the user experience of the self-reported CHT programme Clinical Expert Operating System (CLEOS) in the setting of patients visiting an ED for acute chest pain.

METHODS

This qualitative interview study is part of the ongoing CLEOS-Chest Pain Danderyd Study. A subset (n = 84) of the larger sample who had taken part in self-reported history taking during waiting times at the ED were contacted by telephone and n = 54 (64%) accepted participation. An interview guide with open-ended questions was used and the text was analysed using directed content analysis.

RESULTS

The patients' experiences of the CLEOS programme were overall positive although some perceived it as extensive. The programme was well accepted and despite the busy environment, patients were highly motivated and deemed it helpful to make a diagnosis. Six categories of user experience emerged: The clinical context, The individual context, Time aspect, Acceptability of the programme, Usability of the programme and Perceptions of usefulness in a clinical setting.

CONCLUSIONS

The programme was well accepted by most patients in the stressful environment at ED although some found it difficult to answer all the questions. Adjustments to the extent of an interview to better suit the context of the clinical use should be a future development of the programme. The findings suggest that CHT programmes can be integrated as a standard process for collecting self-reported medical history data in the ED setting.

Obtaining patients' medical history using a digital device prior to consultation in primary care: study protocol for a usability and validity study

Background

With the help of digital tools patients’ medical histories can be collected quickly and transferred into their electronic medical records. This information can facilitate treatment planning, reduce documentation work, and improve care. However, it is still unclear whether the information collected from patients in this way is reliable. In this study, we assess the accuracy of the information collected by patients using an app for medical history taking by comparing it with the information collected in a face-to-face medical interview. We also study the app’s usability from the patients’ point of view and analysing usage data.

Methods

We developed a software application (app) for symptom-oriented medical history taking specialized for general practice. Medical history taking will take place involving patients with acute somatic or psychological complaints (1) using the app and (2) verbally with trained study staff. To assess the perceived usability, patients will complete a questionnaire for the System Usability Scale. We will collect sociodemographic data, information about media use and health literacy, and app usage data.

Discussion

Digital tools offer the opportunity to improve patient care. However, it is not self-evident that the medical history taken by digital tools corresponds to the medical history that would be taken in an interview. If simply due to a design flaw patients answer questions about signs and symptoms that indicate possible serious underlying conditions ‘wrong’, this could have severe consequences. By additionally assessing the app’s usability as perceived by a diverse group of patients, potential weaknesses in content, design and navigation can be identified and subsequently improved. This is essential in order to ensure that the app meets the need of different groups of patients.

Trial registration German Clinical Trials Register DRKS00026659, registered Nov 03 2021. World Health Organization Trial Registration Data Set, https://trialsearch.who.int/Trial2.aspx? TrialID = DRKS00026659.

Incidence of diagnostic errors in unplanned hospitalized patients using an automated medical history-taking system with differential diagnosis generator: retrospective observational study (Preprint)

Background

Automated medical history–taking systems that generate differential diagnosis lists have been suggested to contribute to improved diagnostic accuracy. However, the effect of these systems on diagnostic errors in clinical practice remains unknown.

Objective

This study aimed to assess the incidence of diagnostic errors in an outpatient department, where an artificial intelligence (AI)–driven automated medical history–taking system that generates differential diagnosis lists was implemented in clinical practice.

Methods

We conducted a retrospective observational study using data from a community hospital in Japan. We included patients aged 20 years and older who used an AI-driven, automated medical history–taking system that generates differential diagnosis lists in the outpatient department of internal medicine for whom the index visit was between July 1, 2019, and June 30, 2020, followed by unplanned hospitalization within 14 days. The primary endpoint was the incidence of diagnostic errors, which were detected using the Revised Safer Dx Instrument by at least two independent reviewers. To evaluate the effect of differential diagnosis lists from the AI system on the incidence of diagnostic errors, we compared the incidence of these errors between a group where the AI system generated the final diagnosis in the differential diagnosis list and a group where the AI system did not generate the final diagnosis in the list; the Fisher exact test was used for comparison between these groups. For cases with confirmed diagnostic errors, further review was conducted to identify the contributing factors of these errors via discussion among three reviewers, using the Safer Dx Process Breakdown Supplement as a reference.

Results

A total of 146 patients were analyzed. A final diagnosis was confirmed for 138 patients and was observed in the differential diagnosis list from the AI system for 69 patients. Diagnostic errors occurred in 16 out of 146 patients (11.0%, 95% CI 6.4%-17.2%). Although statistically insignificant, the incidence of diagnostic errors was lower in cases where the final diagnosis was included in the differential diagnosis list from the AI system than in cases where the final diagnosis was not included in the list (7.2% vs 15.9%, P=.18).

Conclusions

The incidence of diagnostic errors among patients in the outpatient department of internal medicine who used an automated medical history–taking system that generates differential diagnosis lists seemed to be lower than the previously reported incidence of diagnostic errors. This result suggests that the implementation of an automated medical history–taking system that generates differential diagnosis lists could be beneficial for diagnostic safety in the outpatient department of internal medicine.

From statistics to clinics: the visual feedback of PROMIS® CATs

Background

To reduce the burden of completing Patient-Reported Outcome Measures (PROMs), PROMIS® Computerized Adaptive Tests (CATs) are being implemented in pediatric clinical practice. We aimed to develop recommendations for visual feedback options for PROMIS CATs on individual item and domain score level as an evidence-based feedback recommendation for PROMIS CATs is lacking.

Methods

Focus groups were held with clinicians who use the KLIK PROM portal. Literature-based feedback options were provided to initiate group discussion. Data was analyzed using thematic coding method. Additionally, a questionnaire was sent out to assess patients’ (12-18y) and parents’ (child 0-18y) preference for individual item feedback. Data was analyzed using descriptive statistics.

Results

Six focus groups were held (N = 28 clinicians). Regarding individual item feedback, showing the complete item bank, with only responses to administered items in traffic light colors was preferred. For domain scores, line graphs were preferred, including numerical (T-)scores, reference and cut-off lines, and traffic light colors. Separate graphs per domain, ranked in order of importance and harmonization of directionality (‘higher = better’) were considered important. Questionnaire results (N = 31 patients/N = 131 parents) showed that viewing their own item responses was preferred above receiving no item feedback by 58.1% of the patients and 77.1% of the parents.

Conclusions

Based on the outcomes and after discussion with the Dutch-Flemish PROMIS National Center, recommendations for PROMIS CAT feedback options were developed. PROMIS CATs can now be used in clinical practice to help clinicians monitor patient outcomes, while reducing the burden of completing PROMs for patients significantly.

Impact of an Online Gastrointestinal Symptom History Taker on Physician Documentation and Charting Time: Pragmatic Controlled Trial

Background

A potential benefit of electronic health records (EHRs) is that they could potentially save clinician time and improve documentation by auto-generating the history of present illness (HPI) in partnership with patients prior to the clinic visit. We developed an online patient portal called AEGIS (Automated Evaluation of Gastrointestinal [GI] Symptoms) that systematically collects patient GI symptom information and then transforms the data into a narrative HPI that is available for physicians to review in the EHR prior to seeing the patient.

Objective

This study aimed to compare whether use of an online GI symptom history taker called AEGIS improves physician-centric outcomes vs usual care.

Methods

We conducted a pragmatic controlled trial among adults aged ≥18 years scheduled for a new patient visit at 4 GI clinics at an academic medical center. Patients who completed AEGIS were matched with controls in the intervention period who did not complete AEGIS as well as controls who underwent usual care in the pre-intervention period. Of note, the pre-intervention control group was formed as it was not subject to contamination bias, unlike for post-intervention controls. We then compared the following outcomes among groups: (1) documentation of alarm symptoms, (2) documentation of family history of GI malignancy, (3) number of follow-up visits in a 6-month period, (4) number of tests ordered in a 6-month period, and (5) charting time (difference between appointment time and time the encounter was closed). Multivariable regression models were used to adjust for potential confounding.

Results

Of the 774 patients who were invited to complete AEGIS, 116 (15.0%) finished it prior to their visit. The 116 AEGIS patients were then matched with 343 and 102 controls in the pre- and post-intervention periods, respectively. There were no statistically significant differences among the groups for documentation of alarm symptoms and GI cancer family history, number of follow-up visits and ordered tests, or charting time (all P>.05).

Conclusions

Use of a validated online HPI-generation portal did not improve physician documentation or reduce workload. Given universal adoption of EHRs, further research examining how to optimally leverage patient portals for improving outcomes are needed.

Use of Self-Reported Computerized Medical History Taking for Acute Chest Pain in the Emergency Department - the Clinical Expert Operating System Chest Pain Danderyd Study (CLEOS-CPDS): Prospective Cohort Study

Background

Chest pain is one of the most common chief complaints in emergency departments (EDs). Collecting an adequate medical history is challenging but essential in order to use recommended risk scores such as the HEART score (based on history, electrocardiogram, age, risk factors, and troponin). Self-reported computerized history taking (CHT) is a novel method to collect structured medical history data directly from the patient through a digital device. CHT is rarely used in clinical practice, and there is a lack of evidence for utility in an acute setting.

Objective

This substudy of the Clinical Expert Operating System Chest Pain Danderyd Study (CLEOS-CPDS) aimed to evaluate whether patients with acute chest pain can interact effectively with CHT in the ED.

Methods

Prospective cohort study on self-reported medical histories collected from acute chest pain patients using a CHT program on a tablet. Clinically stable patients aged 18 years and older with a chief complaint of chest pain, fluency in Swedish, and a nondiagnostic electrocardiogram or serum markers for acute coronary syndrome were eligible for inclusion. Patients unable to carry out an interview with CHT (eg, inadequate eyesight, confusion or agitation) were excluded. Effectiveness was assessed as the proportion of patients completing the interview and the time required in order to collect a medical history sufficient for cardiovascular risk stratification according to HEART score.

Results

During 2017-2018, 500 participants were consecutively enrolled. The age and sex distribution (mean 54.3, SD 17.0 years; 213/500, 42.6% women) was similar to that of the general chest pain population (mean 57.5, SD 19.2 years; 49.6% women). Common reasons for noninclusion were language issues (182/1000, 18.2%), fatigue (158/1000, 15.8%), and inability to use a tablet (152/1000, 15.2%). Sufficient data to calculate HEART score were collected in 70.4% (352/500) of the patients. Key modules for chief complaint, cardiovascular history, and respiratory history were completed by 408 (81.6%), 339 (67.8%), and 291 (58.2%) of the 500 participants, respectively, while 148 (29.6%) completed the entire interview (in all 14 modules). Factors associated with completeness were age 18-69 years (all key modules: Ps<.001), male sex (cardiovascular: P=.04), active workers (all key modules: Ps<.005), not arriving by ambulance (chief complaint: P=.03; cardiovascular: P=.045), and ongoing chest pain (complete interview: P=.002). The median time to collect HEART score data was 23 (IQR 18-31) minutes and to complete an interview was 64 (IQR 53-77) minutes. The main reasons for discontinuing the interview prior to completion (n=352) were discharge from the ED (101, 28.7%) and tiredness (95, 27.0%).

Conclusions

A majority of patients with acute chest pain can interact effectively with CHT on a tablet in the ED to provide sufficient data for risk stratification with a well-established risk score. The utility was somewhat lower in patients 70 years and older, in patients arriving by ambulance, and in patients without ongoing chest pain. Further studies are warranted to assess whether CHT can contribute to improved management and prognosis in this large patient group.

Trial Registration

ClinicalTrials.gov NCT03439449; https://clinicaltrials.gov/ct2/show/NCT03439449

International Registered Report Identifier (IRRID)

RR2-10.1136/bmjopen-2019-031871

Efficacy of Artificial-Intelligence-Driven Differential-Diagnosis List on the Diagnostic Accuracy of Physicians: An Open-Label Randomized Controlled Study

Background: The efficacy of artificial intelligence (AI)-driven automated medical-history-taking systems with AI-driven differential-diagnosis lists on physicians’ diagnostic accuracy was shown. However, considering the negative effects of AI-driven differential-diagnosis lists such as omission (physicians reject a correct diagnosis suggested by AI) and commission (physicians accept an incorrect diagnosis suggested by AI) errors, the efficacy of AI-driven automated medical-history-taking systems without AI-driven differential-diagnosis lists on physicians’ diagnostic accuracy should be evaluated. Objective: The present study was conducted to evaluate the efficacy of AI-driven automated medical-history-taking systems with or without AI-driven differential-diagnosis lists on physicians’ diagnostic accuracy. Methods: This randomized controlled study was conducted in January 2021 and included 22 physicians working at a university hospital. Participants were required to read 16 clinical vignettes in which the AI-driven medical history of real patients generated up to three differential diagnoses per case. Participants were divided into two groups: with and without an AI-driven differential-diagnosis list. Results: There was no significant difference in diagnostic accuracy between the two groups (57.4% vs. 56.3%, respectively; p = 0.91). Vignettes that included a correct diagnosis in the AI-generated list showed the greatest positive effect on physicians’ diagnostic accuracy (adjusted odds ratio 7.68; 95% CI 4.68–12.58; p < 0.001). In the group with AI-driven differential-diagnosis lists, 15.9% of diagnoses were omission errors and 14.8% were commission errors. Conclusions: Physicians’ diagnostic accuracy using AI-driven automated medical history did not differ between the groups with and without AI-driven differential-diagnosis lists.

Partners in Narrative: Empowering Patient-Physician Partnerships in the Electronic Health Record

Amidst the chorus of valid laments about the electronic health record (EHR) are voices calling our attention to its potential to enhance transmission of information, patient communication, and decision-making. Herein, we propose ideas which, in addition, may enhance the potential of physicians and patients to become better at storytelling through the EHR. Clinicians can partner with patients to create meaningful, personalized narratives which restore inclusivity and patient agency to the EHR.

Impact of a Commercial Artificial Intelligence-Driven Patient Self-Assessment Solution on Waiting Times at General Internal Medicine Outpatient Departments: Retrospective Study

Background

Patient waiting time at outpatient departments is directly related to patient satisfaction and quality of care, particularly in patients visiting the general internal medicine outpatient departments for the first time. Moreover, reducing wait time from arrival in the clinic to the initiation of an examination is key to reducing patients’ anxiety. The use of automated medical history–taking systems in general internal medicine outpatient departments is a promising strategy to reduce waiting times. Recently, Ubie Inc in Japan developed AI Monshin, an artificial intelligence–based, automated medical history–taking system for general internal medicine outpatient departments.

Objective

We hypothesized that replacing the use of handwritten self-administered questionnaires with the use of AI Monshin would reduce waiting times in general internal medicine outpatient departments. Therefore, we conducted this study to examine whether the use of AI Monshin reduced patient waiting times.

Methods

We retrospectively analyzed the waiting times of patients visiting the general internal medicine outpatient department at a Japanese community hospital without an appointment from April 2017 to April 2020. AI Monshin was implemented in April 2019. We compared the median waiting time before and after implementation by conducting an interrupted time-series analysis of the median waiting time per month. We also conducted supplementary analyses to explain the main results.

Results

We analyzed 21,615 visits. The median waiting time after AI Monshin implementation (74.4 minutes, IQR 57.1) was not significantly different from that before AI Monshin implementation (74.3 minutes, IQR 63.7) (P=.12). In the interrupted time-series analysis, the underlying linear time trend (–0.4 minutes per month; P=.06; 95% CI –0.9 to 0.02), level change (40.6 minutes; P=.09; 95% CI –5.8 to 87.0), and slope change (–1.1 minutes per month; P=.16; 95% CI –2.7 to 0.4) were not statistically significant. In a supplemental analysis of data from 9054 of 21,615 visits (41.9%), the median examination time after AI Monshin implementation (6.0 minutes, IQR 5.2) was slightly but significantly longer than that before AI Monshin implementation (5.7 minutes, IQR 5.0) (P=.003).

Conclusions

The implementation of an artificial intelligence–based, automated medical history–taking system did not reduce waiting time for patients visiting the general internal medicine outpatient department without an appointment, and there was a slight increase in the examination time after implementation; however, the system may have enhanced the quality of care by supporting the optimization of staff assignments.

Prevalence of Gastroesophageal Reflux Disease and Proton Pump Inhibitor-Refractory Symptoms

BACKGROUND & AIMS

There are few data on the prevalence of gastroesophageal reflux disease (GERD) in the United States. We performed a population-based study to determine the prevalence of GERD symptoms and persistent GERD symptoms despite use of proton pump inhibitors (PPIs).

METHODS

We conducted the National Gastrointestinal Survey in 2015 using MyGiHealth, an app that guides participants through National Institutes of Health gastrointestinal Patient-Reported Outcomes Measurement Information System surveys. Primary outcomes were prevalence of GERD symptoms in the past and persistence of GERD symptoms (heartburn or regurgitation 2 or more days in past week) among participants taking PPIs. Population weights were applied to the data and multivariable regression was used to adjust for confounding.

RESULTS

Among 71,812 participants, 32,878 (44.1%) reported having had GERD symptoms in the past and 23,039 (30.9%) reported having GERD symptoms in the past week. We also found that 35.1% of those who had experienced GERD symptoms were currently on therapy (55.2% on PPIs, 24.3% on histamine-2 receptor blockers, and 24.4% on antacids). Among 3229 participants taking daily PPIs, 54.1% had persistent GERD symptoms. Younger individuals, women, Latino individuals, and participants with irritable bowel syndrome or Crohn's disease were more likely to have continued symptoms, even when taking PPIs.

CONCLUSIONS

Using a population-based survey, we found GERD symptoms to be common: 2 of 5 participants have had GERD symptoms in the past and 1 of 3 had symptoms in the past week. We also found that half of PPI users have persistent symptoms. Given the significant effect of GERD on quality of life, further research and development of new therapies are needed for patients with PPI-refractory GERD symptoms.

A prospective cohort study of self-reported computerised medical history taking for acute chest pain: protocol of the CLEOS-Chest Pain Danderyd Study (CLEOS-CPDS)

INTRODUCTION

Management of acute chest pain focuses on diagnosis or safe rule-out of an acute coronary syndrome (ACS). We aim to determine the additional value of self-reported computerised history taking (CHT).

METHODS AND ANALYSIS

Prospective cohort study design with self-reported, medical histories collected by a CHT programme (Clinical Expert Operating System, CLEOS) using a tablet. Women and men presenting with acute chest pain to the emergency department at Danderyd University Hospital (Stockholm, Sweden) are eligible. CHT will be compared with standard history taking for completeness of data required to calculate ACS risk scores such as History, ECG, Age, Risk factors and Troponin (HEART), Global Registry of Acute Coronary Events (GRACE), and Thrombolysis in Myocardial Infarction (TIMI). Clinical outcomes will be extracted from hospital electronic health records and national registries. The CLEOS-Chest Pain Danderyd Study project includes (1) a feasibility study of CHT, (2) a validation study of CHT as compared with standard history taking, (3) a paired diagnostic accuracy study using data from CHT and established risk scores, (4) a clinical utility study to evaluate the impact of CHT on the management of chest pain and the use of resources, and (5) data mining, aiming to generate an improved risk score for ACS. Primary outcomes will be analysed after 1000 patients, but to allow for subgroup analysis, the study intends to recruit 2000 or more patients. This ongoing project may lead to new and more effective ways for collecting thorough, accurate medical histories with important implications for clinical practice.

ETHICS AND DISSEMINATION

This study has been reviewed and approved by the Stockholm Regional Ethical Committee (now Swedish Ethical Review Authority). Results will be published, regardless of the outcome, in peer-reviewed international scientific journals.

TRIAL REGISTRATION NUMBER

This study is registered at https://www.clinicaltrials.gov (unique identifier: NCT03439449).

Burden of Gastrointestinal Symptoms in the United States: Results of a Nationally Representative Survey of Over 71,000 Americans

OBJECTIVES

Digestive diseases account for >100 million ambulatory care visits annually in the U.S. Yet, comparatively less is known about the true burden of gastrointestinal (GI) symptoms in the general U.S.

POPULATION

The aim of this study was to use data from the "National GI Survey"-a population-based audit of GI symptoms in >71,000 participants-to determine the prevalence and predictors of GI symptoms in community-dwelling Americans.

METHODS

We conducted the National GI Survey using a mobile app called MyGiHealth, which employs a computer algorithm that systematically collects participants' GI symptoms. We recruited a nationally representative sample of Americans to complete the survey, which guided respondents through National Institutes of Health (NIH) GI Patient Reported Outcome Measurement Information System (PROMIS®) scales along with questions about relevant comorbidities and demographics. We measured the prevalence of GI symptoms in the past week and employed logistic regression to adjust for confounding.

RESULTS

Overall, 71,812 individuals completed the survey, of which 61% reported having had ≥1 GI symptom in the past week. The most commonly reported symptoms were heartburn/reflux (30.9%), abdominal pain (24.8%), bloating (20.6%), diarrhea (20.2%), and constipation (19.7%). Less common symptoms were nausea/vomiting (9.5%), dysphagia (5.8%), and bowel incontinence (4.8%). Females, non-Hispanic whites, and individuals who were younger, highly educated, and had medical comorbidities were more likely to have symptoms (all adjusted p < 0.05).

CONCLUSIONS

In this large population-based study that combined digital health technology with NIH PROMIS questionnaires, we found that GI symptoms are highly prevalent, as nearly two thirds of surveyed Americans are burdened by these symptoms.

Unique Clinical Language Patterns Among Expert Vestibular Providers Can Predict Vestibular Diagnoses

OBJECTIVE

To identify novel language usage by expert providers predictive of specific vestibular conditions.

STUDY DESIGN

Retrospective chart review and natural language processing. Level IV.

SETTING

Tertiary referral center.

PATIENTS

Patients seen for vestibular complaint.

INTERVENTION(S)

Natural language processing and machine learning analyses of semantic and syntactic patterns in clinical documentation from vestibular patients.

MAIN OUTCOME MEASURE

Accuracy of Naïve Bayes predictive models correlating language usage with clinical diagnoses.

RESULTS

Natural language analyses on 866 physician-generated histories from vestibular patients found 3,286 unique examples of language usage of which 614 were used 10 or greater times. The top 15 semantic types represented only 11% of all Unified Medical Language System semantic types but covered 86% of language used in vestibular patient histories. Naïve Bayes machine learning algorithms on a subset of 255 notes representing benign paroxysmal positional vertigo, vestibular migraine, anxiety-related dizziness and central dizziness generated strong predictive models showing an average sensitivity rate of 93.4% and a specificity rate of 98.2%. A binary model for assessing whether a subject had a specific diagnosis or not had an average AUC for the receiver operating characteristic curves of .995 across all conditions.

CONCLUSIONS

These results indicate that expert providers utilize unique language patterns in vestibular notes that are highly conserved. These patterns have strong predictive power toward specific vestibular diagnoses. Such language elements can provide a simple vocabulary to aid nonexpert providers in formulating a differential diagnosis. They can also be incorporated into clinical decision support systems to facilitate accurate vestibular diagnosis in ambulatory settings.

Prevalence of and Factors Associated With Fecal Incontinence: Results From a Population-Based Survey

BACKGROUND & AIMS

Fecal incontinence (FI) is characterized by uncontrolled passage of solid or liquid stool. We aimed to determine the prevalence and severity of FI in a large sample of US residents.

METHODS

We recruited a representative sample of patients in October 2015 to complete the National Gastrointestinal (GI) Survey; a mobile app called MyGiHealth was used to systematically collect data on GI symptoms. FI was defined as accidental leakage of solid or liquid stool. Severity of FI was determined by responses to the National Institutes of Health FI Patient Reported Outcomes Measurement Information System questionnaire. Multivariable regression models were used to identify factors associated with FI prevalence and severity.

RESULTS

Among 71,812 individuals who completed the National GI Survey, 14.4% reported FI in the past; of these, 33.3% had FI within the past 7 days. Older age, male sex, and Hispanic ethnicity increased the likelihood of having FI within the past week. Individuals with Crohn's disease, ulcerative colitis, celiac disease, irritable bowel syndrome, or diabetes were more likely to report FI. Non-Hispanic black and Hispanic individuals and individuals with Crohn's disease, celiac disease, diabetes, human immunodeficiency virus/acquired immunodeficiency syndrome, or chronic idiopathic constipation had more severe symptoms of FI than individuals without these features.

CONCLUSIONS

In a large population-based survey, 1 in 7 people reported previous FI. FI is age-related and more prevalent among individuals with inflammatory bowel disease, celiac disease, irritable bowel syndrome, or diabetes than people without these disorders. Proactive screening for FI among these groups is warranted.

Lower and Upper Gastrointestinal Symptoms Differ Between Individuals With Irritable Bowel Syndrome With Constipation or Chronic Idiopathic Constipation

BACKGROUND/AIMS

We evaluated the distribution of lower and upper gastrointestinal (GI) symptoms among individuals with irritable bowel syndrome with constipation (IBS-C) and chronic idiopathic constipation (CIC) in a nationwide survey.

METHODS

Individuals (≥ 18 years of age) were identified from a nationwide sample of ＞ 70 000 United States adults. Participants completed the National Institutes of Health GI Patient Reported Outcomes Measurement Information System (NIH GI-PROMIS) questionnaire. Symptom frequency and intensity in the prior 7 days were assessed using validated PROMIS scores. Odds ratios (OR) with 95% confidence intervals (CI) were calculated to compare symptom prevalence in IBS-C vs CIC, and one-way ANOVA was used to assess differences in PROMIS scores. Regression analysis was performed to adjust for demographic variables.

RESULTS

Nine hundred and seventy adults met eligibility criteria (275 with IBS-C, 734 with CIC). Demographics were similar among groups except for education, marital and employment status, and income. Adjusting for demographic differences, GI-PROMIS scores of global GI symptoms were higher in IBS-C (251.1; 95% CI, 230.0-273.1) compared to CIC (177.8; 95% CI 167.2-188.4) (P ＜ 0.001). Abdominal pain was more prevalent (OR, 4.3; 95% CI, 2.9-6.6) and more severe (P = 0.007) in IBS-C. Constipation was more severe in IBS-C (P = 0.011). Incontinence was more common (OR, 2.9; 95% CI, 1.3-6.3) but just as severe (P = 0.389) in IBS-C versus CIC. Regarding upper GI symptoms, the prevalence of dysphagia, heartburn, and nausea were similar. However, IBS-C individuals had more severe heartburn (P = 0.001).

CONCLUSION

GI symptoms are generally more severe in IBS-C compared to CIC, however abdominal pain, bloating, and upper GI symptoms still commonly occur in CIC.

Developing a Mobile Health Application to Assist With Clinic Flow, Documentation, Billing, and Research in a Specialty Clinic

In specialty clinics, a staff physician is often required to direct patient flow through the clinic and performs all documentation for coding/billing. In response to the workload created by increased patient volume, many specialty clinics have implemented protocols for both disease treatment and coordination of clinic flow. In this article, we review the literature on using mobile technology to assist with patient care, clinic flow, disease treatment, and documentation/billing. We also describe the development and implementation of a mobile application in our pulmonary clinic designed to automate patient flow, assist the physician in documentation/billing, and gather research data including review of initial user data and lessons learned.

MERS-CoV pathogenesis and antiviral efficacy of licensed drugs in human monocyte-derived antigen-presenting cells

Middle East respiratory syndrome coronavirus (MERS-CoV) presents an emerging threat to public health worldwide by causing severe respiratory disease in humans with high virulence and case fatality rate (about 35%) since 2012. Little is known about the pathogenesis and innate antiviral response in primary human monocyte-derived macrophages (MDMs) and dendritic cells (MDDCs) upon MERS-CoV infection. In this study, we assessed MERS-CoV replication as well as induction of inflammatory cytokines and chemokines in MDMs and immature and mature MDDCs. Immature MDDCs and MDMs were permissive for MERS-CoV infection, while mature MDDCs were not, with stimulation of proinflammatory cytokine and chemokine upregulation in MDMs, but not in MDDCs. To further evaluate the antiviral activity of well-defined drugs in primary antigen presenting cells (APCs), three compounds (chloroquine, chlorpromazine and toremifine), each with broad-spectrum antiviral activity in immortalized cell lines, were evaluated in MDMs and MDDCs to determine their antiviral effect on MERS-CoV infection. While chloroquine was not active in these primary cells, chlorpromazine showed strong anti-MERS-CoV activity, but it was associated with high cytotoxicity narrowing the potential window for drug utilization. Unlike in established cells, toremifene had marginal activity when tested in antigen presenting cells, with high apparent cytotoxicity, also limiting its potential as a therapeutic option. These results demonstrate the value of testing drugs in primary cells, in addition to established cell lines, before initiating preclinical or clinical studies for MERS treatment and the importance of carefully assessing cytotoxicity in drug screen assays. Furthermore, these studies also highlight the role of APCs in stimulating a robust protective immune response to MERS-CoV infection.

Compliance Is Contagious: Using Informatics Methods to Measure the Spread of a Documentation Standard From a Preoperative Clinic

PURPOSE

Preoperative documentation is essential to coordinated care and has the potential for standardization, which may facilitate downstream clinical management.

DESIGN

An observational pre/post standardization design was used.

METHODS

We analyzed the implementation of a preoperative documentation standardization intervention in Vanderbilt's Preoperative Evaluation Clinic (VPEC) and its impact outside VPEC. A phased intervention consisted of clinician education with monthly feedback, followed by the development of a compliance dashboard and inclusion in Ongoing Professional Performance Evaluation system by VPEC. A follow-up survey was administered to measure the impact on clinical management.

FINDINGS

Adherence to standardization was improved with the addition of electronic feedback. Implementation of this system in the preoperative clinic had significant impact outside VPEC. Trainee status was a significant predictor of adoption of the standardized format.

CONCLUSIONS

Adoption of a preoperative documentation standard in a clinic had a positive impact on standardization practices in a perioperative system.

Safety Evaluation of an Automated Remote Monitoring System for Heart Failure in an Urban, Indigent Population

Heart Failure (HF) is the most expensive preventable condition, regardless of patient ethnicity, race, socioeconomic status, sex, and insurance status. Remote telemonitoring with timely outpatient care can significantly reduce avoidable HF hospitalizations. Human outreach, the traditional method used for remote monitoring, is effective but costly. Automated systems can potentially provide positive clinical, fiscal, and satisfaction outcomes in chronic disease monitoring. The authors implemented a telephonic HF automated remote monitoring system that utilizes deterministic decision tree logic to identify patients who are at risk of clinical decompensation. This safety study evaluated the degree of clinical concordance between the automated system and traditional human monitoring. This study focused on a broad underserved population and demonstrated a safe, reliable, and inexpensive method of monitoring patients with HF.

Impact of National Institutes of Health Gastrointestinal PROMIS Measures in Clinical Practice: Results of a Multicenter Controlled Trial

OBJECTIVES

The National Institutes of Health (NIH) created the Patient Reported Outcomes Measurement Information System (PROMIS) to allow efficient, online measurement of patient-reported outcomes (PROs), but it remains untested whether PROMIS improves outcomes. Here, we aimed to compare the impact of gastrointestinal (GI) PROMIS measures vs. usual care on patient outcomes.

METHODS

We performed a pragmatic clinical trial with an off-on study design alternating weekly between intervention (GI PROMIS) and control arms at one Veterans Affairs and three university-affiliated specialty clinics. Adults with GI symptoms were eligible. Intervention patients completed GI PROMIS symptom questionnaires on an e-portal 1 week before their visit; PROs were available for review by patients and their providers before and during the clinic visit. Usual care patients were managed according to customary practices. Our primary outcome was patient satisfaction as determined by the Consumer Assessment of Healthcare Providers and Systems questionnaire. Secondary outcomes included provider interpersonal skills (Doctors' Interpersonal Skills Questionnaire (DISQ)) and shared decision-making (9-item Shared Decision Making Questionnaire (SDM-Q-9)).

RESULTS

There were 217 and 154 patients in the GI PROMIS and control arms, respectively. Patient satisfaction was similar between groups (P>0.05). Intervention patients had similar assessments of their providers' interpersonal skills (DISQ 89.4±11.7 vs. 89.8±16.0, P=0.79) and shared decision-making (SDM-Q-9 79.3±12.4 vs. 79.0±22.0, P=0.85) vs.

CONCLUSIONS

This is the first controlled trial examining the impact of NIH PROMIS in clinical practice. One-time use of GI PROMIS did not improve patient satisfaction or assessment of provider interpersonal skills and shared decision-making. Future studies examining how to optimize PROs in clinical practice are encouraged before widespread adoption.

2015 American Journal of Gastroenterology Lecture: How Digital Health Will Transform Gastroenterology

Our patients spend most of their lives far away from an examination room. If we are truly going to capture our patients' attention and engage them in their care, then we must reach beyond the four walls of the clinic, hospital, or endoscopy suite. This is the vision of the digital health movement-an effort to monitor patients remotely and dynamically with mobile health ("mHealth") smartphone applications, electronic health record portals, social media, and wearable biosensors to improve health care outside of the clinical trenches. This article explores how advances in digital health may improve health-care delivery, focusing on gastroenterology and hepatology. It describes how technology can monitor patients remotely, improve face-to-face care, drive clinical decisions, and offer value to health-care organizations, their patients, and their staff. The article also describes pitfalls and shortcomings of digital technologies and concludes by describing a new model for how digital health can be deployed at scale to improve coordination and outcomes of care.

Computer versus physician identification of gastrointestinal alarm features

OBJECTIVE

It is important for clinicians to inquire about "alarm features" as it may identify those at risk for organic disease and who require additional diagnostic workup. We developed a computer algorithm called Automated Evaluation of Gastrointestinal Symptoms (AEGIS) that systematically collects patient gastrointestinal (GI) symptoms and alarm features, and then "translates" the information into a history of present illness (HPI). Our study's objective was to compare the number of alarms documented by physicians during usual care vs. that collected by AEGIS.

METHODS

We performed a cross-sectional study with a paired sample design among patients visiting adult GI clinics. Participants first received usual care by their physicians and then completed AEGIS. Each individual thus contributed both a physician-documented and computer-generated HPI. Blinded physician reviewers enumerated the positive alarm features (hematochezia, melena, hematemesis, unintentional weight loss, decreased appetite, and fevers) mentioned in each HPI. We compared the number of documented alarms within patient using the Wilcoxon signed-rank test.

RESULTS

Seventy-five patients had both physician and AEGIS HPIs. AEGIS identified more patients with positive alarm features compared to physicians (53% vs. 27%; p<.001). AEGIS also documented more positive alarms (median 1, interquartile range [IQR] 0-2) vs. physicians (median 0, IQR 0-1; p<.001). Moreover, clinicians documented only 30% of the positive alarms self-reported by patients through AEGIS.

CONCLUSIONS

Physicians documented less than one-third of red flags reported by patients through a computer algorithm. These data indicate that physicians may under report alarm features and that computerized "checklists" could complement standard HPIs to bolster clinical care.