Diagnostic errors reported in primary healthcare and emergency departments: A retrospective and descriptive cohort study of 4830 reported cases of preventable harm in Sweden

Effect of contextual factors on the prevalence of diagnostic errors among patients managed by physicians of the same specialty: a single-centre retrospective observational study

BACKGROUND

There has been growing recognition that contextual factors influence the physician's cognitive processes. However, given that cognitive processes may depend on the physicians' specialties, the effects of contextual factors on diagnostic errors reported in previous studies could be confounded by difference in physicians.

OBJECTIVE

This study aimed to clarify whether contextual factors such as location and consultation type affect diagnostic accuracy.

METHODS

We reviewed the medical records of 1992 consecutive outpatients consulted by physicians from the Department of Diagnostic and Generalist Medicine in a university hospital between 1 January and 31 December 2019. Diagnostic processes were assessed using the Revised Safer Dx Instrument. Patients were categorised into three groups according to contextual factors (location and consultation type): (1) referred patients with scheduled visit to the outpatient department; (2) patients with urgent visit to the outpatient department; and (3) patients with emergency visit to the emergency room. The effect of the contextual factors on the prevalence of diagnostic errors was investigated using logistic regression analysis.

RESULTS

Diagnostic errors were observed in 12 of 534 referred patients with scheduled visit to the outpatient department (2.2%), 3 of 599 patients with urgent visit to the outpatient department (0.5%) and 13 of 859 patients with emergency visit to the emergency room (1.5%). Multivariable logistic regression analysis showed a significantly higher prevalence of diagnostic errors in referred patients with scheduled visit to the outpatient department than in patients with urgent visit to the outpatient department (OR 4.08, p=0.03), but no difference between patients with emergency and urgent visit to the emergency room and outpatient department, respectively.

CONCLUSION

Contextual factors such as consultation type may affect diagnostic errors; however, since the differences in the prevalence of diagnostic errors were small, the effect of contextual factors on diagnostic accuracy may be small in physicians working in different care settings.

Exploring diagnostic events and first referrals in cancer patient pathways in primary care. A questionnaire survey

BACKGROUND

Cancer diagnostic pathways in general practice are often nonlinear, and several events can delay timely diagnosis.

OBJECTIVES

To explore cancer diagnostic processes in general practice, examining how patients' symptom presentations, sex, and age are associated with the occurrence of predefined potentially delaying events and the first referrals.

METHOD

General practices in 3 Danish Regions were invited to participate in a questionnaire survey, addressing patient's symptom presentation, diagnostic process events, and first referral. The general practitioners (GPs) received a list of their incident cancer patients from the preceding 2 years.

RESULTS

In total 187 general practices participated, including 5,908 patients with the cancer diagnostic pathways initiated in general practice. Presenting with nonspecific symptoms was associated with potentially delaying events, even when the patient also had specific symptoms. Almost half of the patients were referred to a cancer patient pathway (CPP) first, men more often than women, and 10% were referred for acute hospitalization. In 23% of the diagnostic processes, GPs initially treated or referred patients on suspicion of another disease rather than cancer and waited due to normal examinations in 1 out of 20 patients. Excluding sex-specific cancers, these 2 events were more prevalent in women. Men less often complied to the follow-up agreement. Younger patients were less often first referred to a CPP and together with older patients more often first acutely hospitalized.

CONCLUSION

In cancer diagnostic processes in general practice, first referrals and the occurrence of potentially delaying events are associated with the patient's age, sex, and specificity of symptoms.

Prevalence of atypical presentations among outpatients and associations with diagnostic error

OBJECTIVES

This study aimed to assess the prevalence of atypical presentations and their association with diagnostic errors in various diseases.

METHODS

This retrospective observational study was conducted using cohort data between January 1 and December 31, 2019. Consecutive outpatients consulted by physicians from the Department of Diagnostic and Generalist Medicine at a university hospital in Japan were included. Patients for whom the final diagnosis was not confirmed were excluded. Primary outcomes were the prevalence of atypical presentations, and the prevalence of diagnostic errors in groups with typical and atypical presentations. Diagnostic errors and atypical presentations were assessed using the Revised Safer Dx Instrument. We performed primary analyses using a criterion; the average score of less than five to item 12 of two independent reviewers was an atypical presentation (liberal criterion). We also performed additional analyses using another criterion; the average score of three or less to item 12 was an atypical presentation (conservative criterion).

RESULTS

A total of 930 patients were included out of a total of 2022 eligible. The prevalence of atypical presentation was 21.7 and 6.7 % when using liberal and conservative criteria for atypical presentation, respectively. Diagnostic errors (2.8 %) were most commonly observed in the cases with slight to moderate atypical presentation. Atypical presentation was associated with diagnostic errors with the liberal criterion for atypical presentation; however, this diminished with the conservative criterion.

CONCLUSIONS

An atypical presentation was observed in up to 20 % of outpatients with a confirmed diagnosis, and slight to moderate atypical presentation may be the highest risk population for diagnostic errors.

Exploring why European primary care physicians sometimes do not think of, or act on, a possible cancer diagnosis. A qualitative study

BACKGROUND

While primary care physicians (PCPs) play a key role in cancer detection, they can find cancer diagnosis challenging, and some patients have considerable delays between presentation and onward referral.

AIM

To explore European PCPs' experiences and views on cases where they considered that they had been slow to think of, or act on, a possible cancer diagnosis.

DESIGN & SETTING

A multicentre European qualitative study, based on an online survey with open-ended questions, asking PCPs for their narratives about cases when they had missed a diagnosis of cancer.

METHOD

Using maximum variation sampling, PCPs in 23 European countries were asked to describe what happened in a case where they were slow to think of a cancer diagnosis, and for their views on why it happened. Thematic analysis was used to analyse the data.

RESULTS

A total of 158 PCPs completed the questionnaire. The main themes were as follows: patients' descriptions did not suggest cancer; distracting factors reduced PCPs' cancer suspicions; patients' hesitancy delayed the diagnosis; system factors not facilitating timely diagnosis; PCPs felt that they had acted wrongly; and problems with communicating adequately.

CONCLUSION

The study identified six overarching themes that need to be addressed. Doing so should reduce morbidity and mortality in the small proportion of patients who have a significant, avoidable delay in their cancer diagnosis. The 'Swiss cheese' model of accident causation showed how the themes related to each other.

The ethical void: a critical analysis of commissioned expert reports on Swedish healthcare governance

PURPOSE

The purpose of this paper is to contribute knowledge on ethical issues and reasoning in expert reports concerning healthcare governance, commissioned by the Swedish healthcare system.

DESIGN/METHODOLOGY/APPROACH

An in-depth analysis of ethical issues and reasoning in 36 commissioned expert reports was performed. Twenty-seven interviews with commissioners and producers of the reports were also carried out and analysed.

FINDINGS

Some ethical issues were identified in the reports. But ethical reasoning was rarely evident. The meaning of ethical concepts could be devalued and changed over time and thereby deviate from statutory ethical goals and values. Several ethical issues of great concern for the Swedish public healthcare were also absent.

PRACTICAL IMPLICATIONS

The commissioner of expert reports needs to ensure that comprehensive ethical considerations and ethical analysis are integrated in the expert reports.

ORIGINALITY/VALUE

Based on an extensive data material this paper reveals an ethical void in expert reports on healthcare governance. By avoiding ethical issues there is a risk that the expert reports could bring about reforms and control models that have ethically undesirable consequences for people and society.

Patients' and clinicians' views on the appropriate use of safety-netting advice in consultations-an interview study from Sweden

BACKGROUND

A promising approach to manage clinical uncertainty and thereby reduce the risk of preventable diagnostic harm is to use safety-netting advice (ie, communicating structured information to patients about when and where to reconsult healthcare).

AIM

To explore clinicians' and patients' views on when and how safety-netting can be successfully applied in primary-care and emergency-care settings.

DESIGN AND SETTING

An exploratory qualitative research design; we performed focus groups and interviews in a Swedish setting.

PARTICIPANTS

Nine physicians working in primary or emergency care and eight patients or caregivers participated. The participants were an ethnically homogeneous group, originating from Western European or Australian backgrounds.

METHOD

Data were analysed inductively, using the framework method. The results are reported according to the Standards for Reporting Qualitative Research guidelines for reporting qualitative research.

RESULTS

In order to manage diagnostic uncertainty using safety-netting, clinicians and patients emphasised the need to understand the preconditions for the consultation (ie, the healthcare setting, the patient's capacity and existing power imbalance). Furthermore, participants raised the importance of establishing a mutual understanding regarding the patient's perspective and the severity of the situation before engaging in safety-netting advice.

CONCLUSION

The establishment of a shared mental model between clinician and patient of the preconditions for the clinical encounter is a vital factor affecting how safety-netting advice is communicated and received and its ability to support patients in problem detection and planning after the visit. We suggest that successful safety-netting can be viewed as a team activity, where the clinician and patient collaborate in monitoring how the patient's condition progresses after the care visit. Furthermore, our findings suggest that to be successfully implemented, safety-netting advice needs to be tailored to the clinical context in general and to the patient-clinician encounter in particular.

Deep learning assisted diagnosis system: improving the diagnostic accuracy of distal radius fractures

Objectives

To explore an intelligent detection technology based on deep learning algorithms to assist the clinical diagnosis of distal radius fractures (DRFs), and further compare it with human performance to verify the feasibility of this method.

Methods

A total of 3,240 patients (fracture: n = 1,620, normal: n = 1,620) were included in this study, with a total of 3,276 wrist joint anteroposterior (AP) X-ray films (1,639 fractured, 1,637 normal) and 3,260 wrist joint lateral X-ray films (1,623 fractured, 1,637 normal). We divided the patients into training set, validation set and test set in a ratio of 7:1.5:1.5. The deep learning models were developed using the data from the training and validation sets, and then their effectiveness were evaluated using the data from the test set. Evaluate the diagnostic performance of deep learning models using receiver operating characteristic (ROC) curves and area under the curve (AUC), accuracy, sensitivity, and specificity, and compare them with medical professionals.

Results

The deep learning ensemble model had excellent accuracy (97.03%), sensitivity (95.70%), and specificity (98.37%) in detecting DRFs. Among them, the accuracy of the AP view was 97.75%, the sensitivity 97.13%, and the specificity 98.37%; the accuracy of the lateral view was 96.32%, the sensitivity 94.26%, and the specificity 98.37%. When the wrist joint is counted, the accuracy was 97.55%, the sensitivity 98.36%, and the specificity 96.73%. In terms of these variables, the performance of the ensemble model is superior to that of both the orthopedic attending physician group and the radiology attending physician group.

Conclusion

This deep learning ensemble model has excellent performance in detecting DRFs on plain X-ray films. Using this artificial intelligence model as a second expert to assist clinical diagnosis is expected to improve the accuracy of diagnosing DRFs and enhance clinical work efficiency.

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.

Patients' Perceptions of Safety in Primary Healthcare Settings: A Cross-Sectional Study in the Qassim Region of Saudi Arabia

This study assessed patients' perceptions of safety and experiences in primary healthcare in the Qassim region of Saudi Arabia. Between July and September 2022, 730 patients from primary healthcare centers were surveyed using a multi-staged cluster random sampling approach. The Patient-Reported Experiences and Outcomes of Safety in Primary Care (PREOS-PC) questionnaire was used to measure patients' perceived safety and experience in primary healthcare settings within the past year. Descriptive analyses were performed to report patients' perceived safety experiences. The statistical analysis examined individual items and scales. A considerable proportion of patients reported encountering safety problems, ranging from 11% (vaccine-related) to 27% (diagnosis-related). Diagnostic errors were the most common perceived safety problem (26.7%), followed by communication issues (24.1%) and medication errors (16.3%). Between 26% and 40% experienced harm, including financial problems (40%), increased care needs (32.4%), physical health issues (32%), limitations in activities (30.6%), increased healthcare needs (30.2%), and mental health concerns (26.8%). Patient-reported safety experiences reported in our study offer valuable insights into primary care safety in Saudi Arabia. Collecting routine patient feedback is crucial for addressing identified safety problems and implementing standardized procedures.

Artificial Intelligence for Neuroimaging and Musculoskeletal Radiology: Overview of Current Commercial Algorithms

There is a rapidly increasing number of artificial intelligence (AI) products cleared by the Food and Drug Administration (FDA) for quantification, identification, and even diagnosis in clinical radiology. This review article aims to summarize the landscape of current commercial software products in neuroimaging and musculoskeletal radiology. We will discuss key applications, provide an overview of current FDA cleared products, and summarize relevant peer reviewed publications of these products when available.

Deep Learning Assistance Closes the Accuracy Gap in Fracture Detection Across Clinician Types

BACKGROUND

Missed fractures are the most common diagnostic errors in musculoskeletal imaging and can result in treatment delays and preventable morbidity. Deep learning, a subfield of artificial intelligence, can be used to accurately detect fractures by training algorithms to emulate the judgments of expert clinicians. Deep learning systems that detect fractures are often limited to specific anatomic regions and require regulatory approval to be used in practice. Once these hurdles are overcome, deep learning systems have the potential to improve clinician diagnostic accuracy and patient care.

QUESTIONS/PURPOSES

This study aimed to evaluate whether a Food and Drug Administration-cleared deep learning system that identifies fractures in adult musculoskeletal radiographs would improve diagnostic accuracy for fracture detection across different types of clinicians. Specifically, this study asked: (1) What are the trends in musculoskeletal radiograph interpretation by different clinician types in the publicly available Medicare claims data? (2) Does the deep learning system improve clinician accuracy in diagnosing fractures on radiographs and, if so, is there a greater benefit for clinicians with limited training in musculoskeletal imaging?

METHODS

We used the publicly available Medicare Part B Physician/Supplier Procedure Summary data provided by the Centers for Medicare & Medicaid Services to determine the trends in musculoskeletal radiograph interpretation by clinician type. In addition, we conducted a multiple-reader, multiple-case study to assess whether clinician accuracy in diagnosing fractures on radiographs was superior when aided by the deep learning system compared with when unaided. Twenty-four clinicians (radiologists, orthopaedic surgeons, physician assistants, primary care physicians, and emergency medicine physicians) with a median (range) of 16 years (2 to 37) of experience postresidency each assessed 175 unique musculoskeletal radiographic cases under aided and unaided conditions (4200 total case-physician pairs per condition). These cases were comprised of radiographs from 12 different anatomic regions (ankle, clavicle, elbow, femur, forearm, hip, humerus, knee, pelvis, shoulder, tibia and fibula, and wrist) and were randomly selected from 12 hospitals and healthcare centers. The gold standard for fracture diagnosis was the majority opinion of three US board-certified orthopaedic surgeons or radiologists who independently interpreted the case. The clinicians' diagnostic accuracy was determined by the area under the curve (AUC) of the receiver operating characteristic (ROC) curve, sensitivity, and specificity. Secondary analyses evaluated the fracture miss rate (1-sensitivity) by clinicians with and without extensive training in musculoskeletal imaging.

RESULTS

Medicare claims data revealed that physician assistants showed the greatest increase in interpretation of musculoskeletal radiographs within the analyzed time period (2012 to 2018), although clinicians with extensive training in imaging (radiologists and orthopaedic surgeons) still interpreted the majority of the musculoskeletal radiographs. Clinicians aided by the deep learning system had higher accuracy diagnosing fractures in radiographs compared with when unaided (unaided AUC: 0.90 [95% CI 0.89 to 0.92]; aided AUC: 0.94 [95% CI 0.93 to 0.95]; difference in least square mean per the Dorfman, Berbaum, Metz model AUC: 0.04 [95% CI 0.01 to 0.07]; p < 0.01). Clinician sensitivity increased when aided compared with when unaided (aided: 90% [95% CI 88% to 92%]; unaided: 82% [95% CI 79% to 84%]), and specificity increased when aided compared with when unaided (aided: 92% [95% CI 91% to 93%]; unaided: 89% [95% CI 88% to 90%]). Clinicians with limited training in musculoskeletal imaging missed a higher percentage of fractures when unaided compared with radiologists (miss rate for clinicians with limited imaging training: 20% [95% CI 17% to 24%]; miss rate for radiologists: 14% [95% CI 9% to 19%]). However, when assisted by the deep learning system, clinicians with limited training in musculoskeletal imaging reduced their fracture miss rate, resulting in a similar miss rate to radiologists (miss rate for clinicians with limited imaging training: 9% [95% CI 7% to 12%]; miss rate for radiologists: 10% [95% CI 6% to 15%]).

CONCLUSION

Clinicians were more accurate at diagnosing fractures when aided by the deep learning system, particularly those clinicians with limited training in musculoskeletal image interpretation. Reducing the number of missed fractures may allow for improved patient care and increased patient mobility.

LEVEL OF EVIDENCE

Level III, diagnostic study.

Artificial intelligence in emergency radiology: A review of applications and possibilities

Artificial intelligence (AI) applications in radiology have been rising exponentially in the last decade. Although AI has found usage in various areas of healthcare, its utilization in the emergency department (ED) as a tool for emergency radiologists shows great promise towards easing some of the challenges faced daily. There have been numerous reported studies examining the application of AI-based algorithms in identifying common ED conditions to ensure more rapid reporting and in turn quicker patient care. In addition to interpretive applications, AI assists with many of the non-interpretive tasks that are encountered every day by emergency radiologists. These include, but are not limited to, protocolling, image quality control and workflow prioritization. AI continues to face challenges such as physician uptake or costs, but is a long-term investment that shows great potential to relieve many difficulties faced by emergency radiologists and ultimately improve patient outcomes. This review sums up the current advances of AI in emergency radiology, including current diagnostic applications (interpretive) and applications that stretch beyond imaging (non-interpretive), analyzes current drawbacks of AI in emergency radiology and discusses future challenges.

Identifying and prioritizing educational content from a malpractice claims database for clinical reasoning education in the vocational training of general practitioners

Diagnostic reasoning is an important topic in General Practitioners' (GPs) vocational training. Interestingly, research has paid little attention to the content of the cases used in clinical reasoning education. Malpractice claims of diagnostic errors represent cases that impact patients and that reflect potential knowledge gaps and contextual factors. With this study, we aimed to identify and prioritize educational content from a malpractice claims database in order to improve clinical reasoning education in GP training. With input from various experts in clinical reasoning and diagnostic error, we defined five priority criteria that reflect educational relevance. Fifty unique medical conditions from a malpractice claims database were scored on those priority criteria by stakeholders in clinical reasoning education in 2021. Subsequently, we calculated the mean total priority score for each condition. Mean total priority score (min 5-max 25) for all fifty diagnoses was 17,11 with a range from 13,89 to 19,61. We identified and described the fifteen highest scoring diseases (with priority scores ranging from 18,17 to 19,61). The prioritized conditions involved complex common (e.g., cardiovascular diseases, renal insufficiency and cancer), complex rare (e.g., endocarditis, ectopic pregnancy, testicular torsion) and more straightforward common conditions (e.g., tendon rupture/injury, eye infection). The claim cases often demonstrated atypical presentations or complex contextual factors. Including those malpractice cases in GP vocational training could enrich the illness scripts of diseases that are at high risk of errors, which may reduce diagnostic error and related patient harm.

Automated fracture screening using an object detection algorithm on whole-body trauma computed tomography

The emergency department is an environment with a potential risk for diagnostic errors during trauma care, particularly for fractures. Convolutional neural network (CNN) deep learning methods are now widely used in medicine because they improve diagnostic accuracy, decrease misinterpretation, and improve efficiency. In this study, we investigated whether automatic localization and classification using CNN could be applied to pelvic, rib, and spine fractures. We also examined whether this fracture detection algorithm could help physicians in fracture diagnosis. A total of 7664 whole-body CT axial slices (chest, abdomen, pelvis) from 200 patients were used. Sensitivity, precision, and F1-score were calculated to evaluate the performance of the CNN model. For the grouped mean values for pelvic, spine, or rib fractures, the sensitivity was 0.786, precision was 0.648, and F1-score was 0.711. Moreover, with CNN model assistance, surgeons showed improved sensitivity for detecting fractures and the time of reading and interpreting CT scans was reduced, especially for less experienced orthopedic surgeons. Application of the CNN model may lead to reductions in missed fractures from whole-body CT images and to faster workflows and improved patient care through efficient diagnosis in polytrauma patients.

Role of primary care physician factors on diagnostic testing and referral decisions for symptoms of possible cancer: a systematic review

BACKGROUND

Missed opportunities for diagnosing cancer cause patients harm and have been attributed to suboptimal use of tests and referral pathways in primary care. Primary care physician (PCP) factors have been suggested to affect decisions to investigate cancer, but their influence is poorly understood.

OBJECTIVE

To synthesise evidence evaluating the influence of PCP factors on decisions to investigate symptoms of possible cancer.

METHODS

We searched MEDLINE, Embase, Scopus, CINAHL and PsycINFO between January 1990 and March 2021 for relevant citations. Studies examining the effect or perceptions and experiences of PCP factors on use of tests and referrals for symptomatic patients with any cancer were included. PCP factors comprised personal characteristics and attributes of physicians in clinical practice.

DATA EXTRACTION AND SYNTHESIS

Critical appraisal and data extraction were undertaken independently by two authors. Due to study heterogeneity, data could not be statistically pooled. We, therefore, performed a narrative synthesis.

RESULTS

29 studies were included. Most studies were conducted in European countries. A total of 11 PCP factors were identified comprising modifiable and non-modifiable factors. Clinical judgement of symptoms as suspicious or 'alarm' prompted more investigations than non-alarm symptoms. 'Gut feeling' predicted a subsequent cancer diagnosis and was perceived to facilitate decisions to investigate non-specific symptoms as PCP experience increased. Female PCPs investigated cancer more than male PCPs. The effect of PCP age and years of experience on testing and referral decisions was inconclusive.

CONCLUSIONS

PCP interpretation of symptoms as higher risk facilitated testing and referral decisions for possible cancer. However, in the absence of 'alarm' symptoms or 'gut feeling', PCPs may not investigate cancer. PCPs require strategies for identifying patients with non-alarm and non-specific symptoms who need testing or referral.

PROSPERO REGISTRATION NUMBER

CRD420191560515.

Improving Radiographic Fracture Recognition Performance and Efficiency Using Artificial Intelligence

Background Missed fractures are a common cause of diagnostic discrepancy between initial radiographic interpretation and the final read by board-certified radiologists. Purpose To assess the effect of assistance by artificial intelligence (AI) on diagnostic performances of physicians for fractures on radiographs. Materials and Methods This retrospective diagnostic study used the multi-reader, multi-case methodology based on an external multicenter data set of 480 examinations with at least 60 examinations per body region (foot and ankle, knee and leg, hip and pelvis, hand and wrist, elbow and arm, shoulder and clavicle, rib cage, and thoracolumbar spine) between July 2020 and January 2021. Fracture prevalence was set at 50%. The ground truth was determined by two musculoskeletal radiologists, with discrepancies solved by a third. Twenty-four readers (radiologists, orthopedists, emergency physicians, physician assistants, rheumatologists, family physicians) were presented the whole validation data set (n = 480), with and without AI assistance, with a 1-month minimum washout period. The primary analysis had to demonstrate superiority of sensitivity per patient and the noninferiority of specificity per patient at -3% margin with AI aid. Stand-alone AI performance was also assessed using receiver operating characteristic curves. Results A total of 480 patients were included (mean age, 59 years ± 16 [standard deviation]; 327 women). The sensitivity per patient was 10.4% higher (95% CI: 6.9, 13.9; P < .001 for superiority) with AI aid (4331 of 5760 readings, 75.2%) than without AI (3732 of 5760 readings, 64.8%). The specificity per patient with AI aid (5504 of 5760 readings, 95.6%) was noninferior to that without AI aid (5217 of 5760 readings, 90.6%), with a difference of +5.0% (95% CI: +2.0, +8.0; P = .001 for noninferiority). AI shortened the average reading time by 6.3 seconds per examination (95% CI: -12.5, -0.1; P = .046). The sensitivity by patient gain was significant in all regions (+8.0% to +16.2%; P < .05) but shoulder and clavicle and spine (+4.2% and +2.6%; P = .12 and .52). Conclusion AI assistance improved the sensitivity and may even improve the specificity of fracture detection by radiologists and nonradiologists, without lengthening reading time. © RSNA, 2021 Online supplemental material is available for this article. See also the editorial by Link and Pedoia in this issue.

Diagnostic errors: Under-appreciated, under-reported and under-researched

Diagnostic errors, were given relatively little attention, compared with the effort invested in treatment errors. However, erroneous diagnoses continue to be quite prevalent (10%-15% in every setting investigated), and are often associated with substantial patient harm including increased mortality and frequent permanent disability. Physicians may not be aware of the fact that despite the wide availability of sophisticated diagnostic imaging and new and sensitive tests, diagnosis remains far from infallible, because of a complex interplay of physician, patient and illness factors. Research devoted to misdiagnosis remains difficult to perform and insufficient in scope, but the search for the optimal means to improve diagnostic accuracy continues. Newly achieved insights regarding diagnostic errors are presented, and essential system and individual approaches to improve diagnostic accuracy are proposed.

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.

Symptoms and signs of colorectal cancer, with differences between proximal and distal colon cancer: a prospective cohort study of diagnostic accuracy in primary care

Background

In an abdominal symptom study in primary care in six European countries, 511 cases of cancer were recorded prospectively among 61,802 patients 16 years and older in Norway, Denmark, Sweden, Netherlands, Belgium and Scotland. Colorectal cancer is one of the main types of cancer associated with abdominal symptoms; hence, an in-depth subgroup analysis of the 94 colorectal cancers was carried out in order to study variation in symptom presentation among cancers in different anatomical locations.

Method

Initial data capture was by completion of standardised forms containing closed questions about symptoms recorded during the consultation. Follow-up data were provided by the GP after diagnosis, based on medical record data made after the consultation. GPs also provided free text comments about the diagnostic procedure for individual patients. Fisher’s exact test was used to analyse differences between groups.

Results

Almost all symptoms recorded could indicate colorectal cancer. ‘Rectal bleeding’ had a specificity of 99.4% and a PPV of 4.0%. Faecal occult blood in stool (FOBT) or anaemia may indicate gastrointestinal bleeding: when these symptoms and signs were combined, sensitivity reached 57.5%, with 69.2% for cancer in the distal colon. For proximal colon cancers, none of 18 patients had ‘Rectal bleeding’ at the initial consultation, but three of the 18 did so at a later consultation. ‘Abdominal pain, lower part’, ‘Constipation’ and ‘Distended abdomen, bloating’ were less specific and also less sensitive than ‘Rectal bleeding’, and with PPV between 0.7% and 1.9%.

Conclusions

Apart from rectal bleeding, single symptoms did not reach the PPV 3% NICE threshold. However, supplementary information such as a positive FOBT or persistent symptoms may revise the PPV upwards. If a colorectal cancer is suspected by the GP despite few symptoms, the total clinical picture may still reach the NICE PPV threshold of 3% and justify a specific referral.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12875-021-01452-6.

The Odyssey from Symptom to Diagnosis of Pulmonary Hypertension from the Patients and Spouses Perspective

INTRODUCTION/OBJECTIVE

Diagnostic delays in pulmonary arterial hypertension (PAH) and chronic thromboembolic pulmonary hypertension (CTEPH) are related to increased morbidity and mortality. The risk of a delayed, or even a missed, diagnosis is high as the conditions are rare. The aim was to describe patients' and spouses' experiences of the journey from the first symptom to an established diagnosis.

METHODS

A secondary analysis of 31 transcripts, based on 2 primary datasets containing interviews with 17 patients and 14 spouses, was carried out and analyzed according to qualitative content analysis.

RESULTS

One overarching category was revealed from the content analysis; "The journey from doubt and hope to receive the diagnosis." Five subcategories were identified as: overall experiences; ignoring symptoms; seeking primary care/hospital specialty care; blame and stigma; and finding a pulmonary hypertension specialist clinic. The main finding was that both patients and spouses experienced that waiting for a diagnosis and the deteriorating state of health led to anxiety and frustration. The knowledge about rare diseases among health professionals needs to be improved to enable a timelier diagnosis and initiation of treatment.

CONCLUSION

Patients' and spouses' lives were negatively affected by having to search for a correct diagnosis. In order for health care to identify rare diseases earlier, a well-functioning and responsive health care system, in primary care as well as in specialist care, is needed. Symptoms like breathlessness and fatigue are often unspecific but should not be ignored. Keeping the patient and spouse in the loop, and providing information that the search for an answer might take time is essential for health care providers to create trust.

Validation and initial results of surveys exploring perspectives on risks and solutions for diagnostic and medication errors in primary care in Sweden

OBJECTIVE

To (1) validate and (2) display initial results of surveys to health care professionals and patients on the importance and mitigation of specified risks for diagnostic and medication errors.

DESIGN

For validation, psychometric properties were analysed by assessment of construct validity and internal consistency by factor analysis. Non-parametric analyses were used concerning areas of risk, and top ranking of solutions were reported descriptively.

SETTING

Primary health care in Sweden.

PARTICIPANTS

Health care professionals (HCPs); including physicians, nurses and practice managers, as well as patients who had experienced diagnostic or medication errors.

MAIN OUTCOME MEASURES

Psychometric properties of the surveys. Median ratings for risks and top rankings of solutions for professionals and patients.

RESULTS

There were 939 respondents to the HCP survey. Construct validity resulted in a model with four dimensions: Patient-provider level; Support systems for every day clinical work; Shared information and cooperation between different caregivers; Risks in the environment. Internal consistency was acceptable with Cronbach's α values above 0.7. Confirmatory factor analysis generally showed an acceptable fit. Initial results from the professionals showed the importance of continuity of care, a nationwide on-line medical platform and cooperation in transfer of care. The patient survey could not be validated because of low response rate.

CONCLUSION

The HCP survey showed some contradicting results regarding model fit and may be tentatively acceptable but validity needs further study. HCP survey answers indicated that relational continuity of care and a nationwide on-line medical platform are highly valued. Current awareness Health care professionals and patients are rather untapped sources of knowledge regarding patient safety in primary health care Main statements Validation is performed on a new survey capturing rating of risks and solutions. The validation of the health care professional survey is tentatively acceptable. Survey answers indicate that health care professionals' and patients' perspectives are complementary.

Patient and provider perspectives on reducing risk of harm in primary health care: a qualitative questionnaire study in Sweden

Objective: To explore how patients, that had experienced harm in primary care, and how primary providers and practice managers understood reasons for harm and possibilities to reduce risk of harm.Design: Inductive qualitative analysis of structured questionnaires with free text answers.Setting: Primary health care in Sweden.Patients/subjects: Patients (n = 22) who had experienced preventable harm in primary health care, and primary care providers and practice managers, including 15 physicians, 20 nurses and 24 practice managers.Main outcome measures: Categories and overarching themes from the qualitative analysis.Results: The three categories identified as important for safety were continuity of care, communication and competence. With flaws in these, risks were thought to be greater and if these were strengthened the risks could be reduced. The overarching theme for the patient was the experience of being neglected, like not having been properly examined. The overarching theme for primary care providers and practice managers was lack of continuity of care.Conclusion: Primary care providers, practice managers and patients understood the risks and how to reduce the risks of patient safety problems as related to three main categories: continuity of care, communication and competence. Future work towards a safer primary health care could therefore benefit from focusing on these areas.Key pointsCurrent awareness:  • Patients and primary care providers are rather untapped sources of knowledge regarding patient safety in primary health care.Main statements:  • Patients understood the risk of harm as stemming from that they were not properly examined.  • Primary care providers understood the risk of harm to a great extent as stemming from poor continuity of care.  • Patients, primary care providers and practice managers believed continuity, communication and competence play an important role in reducing risks.

Patient-related factors associated with an increased risk of being a reported case of preventable harm in first-line health care: a case-control study

Background

Patient safety issues in primary health care and in emergency departments have not been as thoroughly explored as patient safety issues in the hospital setting. Knowledge is particularly sparse regarding which patients have a higher risk of harm in these settings. The objective was to evaluate which patient-related factors were associated with risk of harm in patients with reports of safety incidents.

Methods

A case–control study performed in primary health care and emergency departments in Sweden. In total, 4536 patients (cases) and 44,949 controls were included in this study. Cases included patients with reported preventable harm in primary health care and emergency departments from January 1st, 2011 until December 31st, 2016.

Results

Psychiatric disease, including all psychiatric diagnoses regardless of severity, nearly doubled the risk of being a reported case of preventable harm (odds ratio, 1.96; p < 0.001). Adjusted for income and education there was still an increased risk (odds ratio, 1.69; p < 0.001). The preventable harm in this group was to 46% diagnostic errors of somatic disease.

Conclusion

Patients with psychiatric illness are at higher risk of preventable harm in primary care and the emergency department. Therefore, this group needs extra attention to prevent harm.