Overcoming Diagnostic Errors in Medical Practice

Classification of major depression disorder via using minimum spanning tree of individual high-order morphological brain network

BACKGROUND

Major depressive disorder (MDD) is an overbroad and heterogeneous diagnosis with no reliable or quantifiable markers. We aim to combine machine-learning techniques with the individual minimum spanning tree of the morphological brain network (MST-MBN) to determine whether the network properties can provide neuroimaging biomarkers to identify patients with MDD.

METHOD

Eight morphometric features of each region of interest (ROI) were extracted from 3D T1 structural images of 106 patients with MDD and 97 healthy controls. Six feature distances of the eight morphometric features were calculated to generate a feature distance matrix, which was defined as low-order MBN. Further linear correlations of feature distances between ROIs were calculated on the basis of low-order MBN to generate individual high-order MBN. The Kruskal's algorithm was used to generate the MST to obtain the core framework of individual low-order and high-order MBN. The regional and global properties of the individual MSTs were defined as the feature. The support vector machine and back-propagation neural network was used to diagnose MDD and assess its severity, respectively.

RESULT

The low-order and high-order MST-MBN constructed by cityblock distance had the excellent classification performance. The high-order MST-MBN significantly improved almost 20 % diagnostic accuracy compared with the low-order MST-MBN, and had a maximum R² value of 0.939 between the predictive and true Hamilton Depression Scale score. The different group-level connectivity strength mainly involves the central executive network and default mode network (no statistical significance after FDR correction).

CONCLUSION

We proposed an innovative individual high-order MST-MBN to capture the cortical high-order morphological correlation and make an excellent performance for individualized diagnosis and assessment of MDD.

The alterations of brain functional connectivity networks in major depressive disorder detected by machine learning through multisite rs-fMRI data

BACKGROUND

The current diagnosis of major depressive disorder (MDD) is mainly based on the patient's self-report and clinical symptoms. Machine learning methods are used to identify MDD using resting-state functional magnetic resonance imaging (rs-fMRI) data. However, due to large site differences in multisite rs-fMRI data and the difficulty of sample collection, most of the current machine learning studies use small sample sizes of rs-fMRI datasets to detect the alterations of functional connectivity (FC) or network attribute (NA), which may affect the reliability of the experimental results.

METHODS

Multisite rs-fMRI data were used to increase the size of the sample, and then we extracted the functional connectivity (FC) and network attribute (NA) features from 1611 rs-fMRI data (832 patients with MDD (MDDs) and 779 healthy controls (HCs)). ComBat algorithm was used to harmonize the data variances caused by the multisite effect, and multivariate linear regression was used to remove age and sex covariates. Two-sample t-test and wrapper-based feature selection methods (support vector machine recursive feature elimination with cross-validation (SVM-RFECV) and LightGBM's "feature_importances_" function) were used to select important features. The Shapley additive explanations (SHAP) method was used to assign the contribution of features to the best classification effect model.

RESULTS

The best result was obtained from the LinearSVM model trained with the 136 important features selected by SVMRFE-CV. In the nested five-fold cross-validation (consisting of an outer and an inner loop of five-fold cross-validation) of 1611 data, the model achieved the accuracy, sensitivity, and specificity of 68.90 %, 71.75 %, and 65.84 %, respectively. The 136 important features were tested in a small dataset and obtained excellent classification results after balancing the ratio between patients with depression and HCs.

CONCLUSIONS

The combined use of FC and NA features is effective for classifying MDDs and HCs. The important FC and NA features extracted from the large sample dataset have some generalization performance and may be used as a reference for the altered brain functional connectivity networks in MDD.

Time to Diagnosis of Autism Spectrum Disorders in Children with Coexisting Developmental Behavioral Disorders

OBJECTIVE

Our study evaluates whether having an alternate developmental behavioral disorder (DBDs) diagnosis before diagnosis of autism spectrum disorders (ASD) is associated with delays in diagnosis in a nationally representative sample.

METHODS

Data were obtained from the 2011 National Survey of Pathways to Diagnosis and Services, a survey of children aged 6 to 17 years with ASD, developmental delay, or intellectual disability. A total of 1049 children met inclusion criteria for this study. Of these, 799 children were identified as "late" diagnosis if >12 months elapsed between the age parents reported concerns to a provider and age of ASD diagnosis and 250 as "timely" diagnosis if the gap was ≤12 months. Univariate and multivariate logistic regressions were used to look for association between having an alternate DBDs diagnosed before ASD and "timely" versus "late" ASD diagnosis.

RESULTS

The mean time elapsed between the age parents reported concerns to a provider and age of ASD diagnosis was 51 months for children with an alternate DBDs diagnosis before receiving ASD diagnosis and 29 months for those diagnosed with alternate DBDs concurrently with ASD. Having alternate DBDs diagnosis before diagnosis with ASD was associated with "late" ASD diagnosis as follows: developmental delay (adjusted odds ratio [aOR,] 3.46; 95% confidence interval [CI], 1.86-6.42; p < 0.001), intellectual disability (aOR, 9.75; 95% CI, 3.0-31.60; p = 0.04), attention-deficit disorder (aOR, 11.07; 95% CI, 3.43-35.71; p < 0.001), depression (aOR, 8.05; 95% CI, 1.07-60.03; p = 0.0495), and behavioral conduct disorder (aOR, 9.9; 95% CI, 3.55-27.62; p < 0.001).

CONCLUSION

These findings highlight the importance of research to improve the early diagnosis of ASD even in the presence of coexisting developmental behavioral disorders.

A pause in pediatrics: implementation of a pediatric diagnostic time-out

BACKGROUND

Diagnostic errors are frequently the product of cognitive biases that arise when heuristic-based approaches fail. The efficiency-thoroughness tradeoff (ETTO) principle states sacrificing thoroughness for efficiency is normal and occurs frequently in medicine. The goal of a diagnostic timeout was to provide an actionable template for when providers transition to an analytical mindset and to help incorporate the ETTO principle during the diagnostic process.

METHODS

A diagnostic time-out was adapted for use in pediatric hospital medicine (PHM). In this prospective study, a group of eight PHM providers piloted the time-out in the hospitalized setting. Data was collected over 12 months and descriptive statistics were used for analysis.

RESULTS

Cases were most frequently chosen for time-out use due to clinician intuition. In more than half the cases the time-out didn't confirm the initial diagnosis and alternate diagnoses for the wrong diagnosis were pursued. There was only one case of the time-out being burdensome from a time perspective. Learners participated in all cases. As a result of the diagnostic time-out, new actions were taken in all cases.

CONCLUSIONS

Implementation of a diagnostic time out provides an actionable template for providers to actively change their mindset to an analytical thinking process to counteract cognitive biases and potentially reduce diagnostic errors in the pediatric inpatient setting.

Long COVID-19 in Children: From the Pathogenesis to the Biologically Plausible Roots of the Syndrome

Long Coronavirus disease-19 (COVID-19) refers to the persistence of symptoms related to the infection with severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2). This condition is described as persistent and can manifest in various combinations of signs and symptoms, such as fatigue, headache, dyspnea, depression, cognitive impairment, and altered perception of smells and tastes. Long COVID-19 may be due to long-term damage to different organs-such as lung, brain, kidney, and heart-caused by persisting viral-induced inflammation, immune dysregulation, autoimmunity, diffuse endothelial damage, and micro thrombosis. In this review, we discuss the potential and biologically plausible role of some vitamins, essential elements, and functional foods based on the hypothesis that an individual's dietary status may play an important adjunctive role in protective immunity against COVID-19 and possibly against its long-term consequences.

Decision Making: Healthy Heuristics and Betraying Biases

Critical care settings are unpredictable, dynamic environments where clinicians face high decision density in suboptimal conditions (stress, time constraints, competing priorities). Experts have described two systems of human decision making: one fast and intuitive; the other slow and methodical. Heuristics, or mental shortcuts, a key feature of intuitive reasoning, are often accurate, applied instinctively, and essential for efficient diagnostic decision making. Heuristics are also prone to failures, or cognitive biases, which can lead to diagnostic errors. A variety of strategies have been proposed to mitigate biases; however, current understanding of such interventions to optimize diagnostic safety is still incomplete.

Diagnostic Errors in Initial Misdiagnosis of Foreign Body Aspiration in Children: A Retrospective Observational Study in a Tertiary Care Hospital in China

Background: Foreign body aspiration (FBA) in children is a common emergency that can easily be missed, leading to delays in treatment. Few large cohort studies have focused on errors in diagnostic assessment. The main purpose of this study was to analyze factors contributing to the initial misdiagnosis of FBA in children. Methods: We retrospectively reviewed the charts of 226 children diagnosed with FBA at the Second Affiliated Hospital and Yuying Children's Hospital of Wenzhou Medical University from January 2018 to November 2020. Cases were divided into two groups according to whether or not patients were initially misdiagnosed. The clinical characteristics of the two groups were then compared. The Diagnosis Error Evaluation and Research (DEER) taxonomy tool was applied to cases with initial misdiagnosis. Results: Of the 226 included children with a final diagnosis of FBA, 153 (67.7%) were boys. Ninety percent of patients were under 3 years old. More than half (61.9%) of the children were referred from primary institutions, and 38.1% visited tertiary hospitals directly. A total of 80 (35.4%) patients were initially misdiagnosed. More than half of misdiagnosed children received an alternative diagnosis of bronchiolitis (51.3%), the most common alternative diagnosis. Test failures (i.e., errors in test ordering, test performance, and clinician processing) were primarily responsible for the majority of initial diagnostic errors (76.3%), followed by failure or delay in eliciting critical case history information (20.0%). Characteristics significantly associated with initial misdiagnosis were: presentation over 24 h (OR 9.2, 95% CI 4.8-17.5), being referred from primary institutions (OR 8.8, 4.1-19.0), no witnessed aspiration crisis (OR 7.8, 3.0-20.3), (4) atypical signs or symptoms (OR 3.2, 1.8-5.7), foreign body not visible on CT (OR 36.2, 2.1-636.8), foreign body located in secondary bronchi (OR 4.8, 1.3-17.2), organic foreign body (OR 6.2, 1.4-27.2), and history of recurrent respiratory infections (OR 2.7, 1.4-5.3). Children with misdiagnosis tended to have a longer time from symptom onset to the definitive diagnosis of FBA (P < 0.001). Conclusions: More than one-third of children with FBA were missed at first presentation. Errors in diagnostic testing and history taking were the main reasons leading to initial misdiagnosis.

Development and internal validation of prognostic models for recovery in patients with non-specific neck pain presenting in primary care

OBJECTIVES

Development and internal validation of prognostic models for post-treatment and 1-year recovery in patients with neck pain in primary care.

DESIGN

Prospective cohort study.

SETTING

Primary care manual therapy practices.

PARTICIPANTS

Patients with non-specific neck pain of any duration (n=1193).

INTERVENTION

Usual care manual therapy.

OUTCOME MEASURES

Recovery defined in terms of pain intensity, disability, and global perceived improvement directly post-treatment and at 1-year follow-up.

RESULTS

All post-treatment models exhibited acceptable discriminative performance after derivation (AUC≥0.7). The developed post-treatment disability model exhibited the best overall performance (R²=0.24; IQR, 0.22-0.26), discrimination (AUC=0.75; 95% CI, 0.63-0.84), and calibration (slope 0.92; IQR, 0.91-0.93). After internal validation and penalization, this model retained acceptable discriminative performance (AUC=0.74). The five other models, including those predicting 1-year recovery, did not reach acceptable discriminative performance after internal validation. Baseline pain duration, disability, and pain intensity were consistent predictors across models.

CONCLUSION

A post-treatment prognostic model for disability was successfully developed and internally validated. This model has potential to inform primary care clinicians about a patient's individual prognosis after treatment, but external validation is required before clinical use can be recommended.

Higher psychological distress in patients seeking care for a knee disorder is associated with diagnostic discordance between health care providers: a secondary analysis of a diagnostic concordance study

Background

Knee disorders are highly prevalent and may be a disabling condition. An accurate diagnosis is necessary to guide toward a rapid and efficient management of knee disorders. However, the ability to make a valid diagnosis is often complex for clinicians and evidence is mainly focused on clinician cognitive biases or errors produced during clinical reasoning. The aim of this secondary exploratory analysis is to identify patient-specific characteristics associated with diagnostic discordance between health care providers in making a diagnosis for a new knee disorder.

Methods

We performed a secondary analysis of a diagnostic study comparing the diagnostic ability of a physiotherapist to medical musculoskeletal specialists. Patients’ socio-demographic, psychosocial and clinical characteristics were compared between the concordant and discordant diagnostic groups. Psychosocial symptoms were evaluated using the validated Kessler 6 (K6) questionnaire. We performed multivariable logistic regressions using the Bayesian Information Criterion to identify the most probable model including patients’ characteristics associated with diagnostic discordance. Overall probability of identified variables to explain diagnostic discordance and associated odd ratios (OR) with 95% credibility intervals (95% CrI) were calculated.

Results

Overall, 279 participants were evaluated by a physiotherapist and medical musculoskeletal specialists. The mean age of the participants was 49.1 ± 15.8 years and 57.7% were female. The most common disorder was osteoarthritis (n = 117, 18.8% of cases were discordant). The most probable model explaining diagnostic discordance (11.13%) included having depressive symptoms, which was associated with an increased probability of diagnostic discordance (OR: 3.9; 95% CrI: 1.9 – 8.0) and having a higher number of comorbidities, which was associated with a decreased probability of diagnostic discordance (OR: 0.6; 95% CrI: 0.5 – 0.9). The depression item of the K6 questionnaire had a 99.4% chance to be included in a model explaining diagnostic discordance. Other variables taken separately had less than 50% chance to be included in a model explaining diagnostic discordance and cannot be considered significant.

Conclusion

Our results suggest that depressive symptoms may increase the risk of knee diagnostic discordance. Clinicians may be more likely to make diagnostic errors and should be more cautious when evaluating patients with knee disorders suffering from psychological distress.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12891-021-04534-9.

What Contributes to Diagnostic Error or Delay? A Qualitative Exploration Across Diverse Acute Care Settings in the United States

OBJECTIVES

Diagnostic error and delay is a prevalent and impactful problem. This study was part of a mixed-methods approach to understand the organizational, clinician, and patient factors contributing to diagnostic error and delay among acutely ill patients within a health system, as well as recommendations for the development of tailored, targeted, feasible, and effective interventions.

METHODS

We did a multisite qualitative study using focus group methodology to explore the perspectives of key clinician stakeholders. We used a conceptual framework that characterized diagnostic error and delay as occurring within 1 of 3 stages of the patient's diagnostic journey-critical information gathering, synthesis of key information, and decision making and communication. We developed our moderator guide based on the sociotechnical frameworks previously described by Holden and Singh for understanding noncognitive factors that lead to diagnostic error and delay. Deidentified focus group transcripts were coded in triplicate and to consensus over a series of meetings. A final coded data set was then uploaded into NVivo software. The data were then analyzed to generate overarching themes and categories.

RESULTS

We recruited a total of 64 participants across 4 sites from emergency departments, hospital floor, and intensive care unit settings into 11 focus groups. Clinicians perceive that diverse organizational, communication and coordination, individual clinician, and patient factors interact to impede the process of making timely and accurate diagnoses.

CONCLUSIONS

This study highlights the complex sociotechnical system within which individual clinicians operate and the contributions of systems, processes, and institutional factors to diagnostic error and delay.

Review of the Basics of Cognitive Error in Emergency Medicine: Still No Easy Answers

Emergency physicians (EP) make clinical decisions multiple times daily. In some instances, medical errors occur due to flaws in the complex process of clinical reasoning and decision-making. Cognitive error can be difficult to identify and is equally difficult to prevent. To reduce the risk of patient harm resulting from errors in critical thinking, it has been proposed that we train physicians to understand and maintain awareness of their thought process, to identify error-prone clinical situations, to recognize predictable vulnerabilities in thinking, and to employ strategies to avert cognitive errors. The first step to this approach is to gain an understanding of how physicians make decisions and what conditions may predispose to faulty decision-making. We review the dual-process theory, which offers a framework to understand both intuitive and analytical reasoning, and to identify the necessary conditions to support optimal cognitive processing. We also discuss systematic deviations from normative reasoning known as cognitive biases, which were first described in cognitive psychology and have been identified as a contributing factor to errors in medicine. Training physicians in common biases and strategies to mitigate their effect is known as debiasing. A variety of debiasing techniques have been proposed for use by clinicians. We sought to review the current evidence supporting the effectiveness of these strategies in the clinical setting. This discussion of improving clinical reasoning is relevant to medical educators as well as practicing EPs engaged in continuing medical education.

Overmedicalization in Children with Medical Complexity

Children with medical complexity (CMC) are a unique pediatric patient population with increased exposure and interactions with the health care system and reliance on family caregivers. These attributes place CMC at high risk of overmedicalization (OM). This article reviews the risk factors for OM in CMC and presents an algorithm that primary providers can use to recognize and address this issue. Involvement of a broad multidisciplinary team, including child advocacy when needed, is recommended. The article also focuses on challenges and additional considerations that arise when medical child abuse as the cause of OM is suspected in this population. [Pediatr Ann. 2020;49(11):e478-e485.].

The competence of village clinicians in the diagnosis and management of childhood epilepsy in Southwestern China and its determinants: A cross-sectional study

BACKGROUND

Due to lack of neurologists in low- and middle-income countries, communities of patients living with epilepsy are calling for task-shifting of diagnosis and management from physicians to paramedical providers in the primary health care systems to narrow the huge treatment gap. Evidence to guide this work has been limited. This study assesses the competence of village clinicians (VC)- mostly paramedical providers- in the diagnosis and management of a presumptive case of childhood epilepsy and its determinants.

METHODS

A cross-sectional study was conducted in rural areas of a province in Southwestern China from July 2017 to January 2018. We randomly selected 370 VCs who practiced Western medicine and assumed the main responsibility of providing medical services in his/her clinic. A standardized clinical vignette based on national clinical practice guidelines was used to evaluate clinicians' competence in three domains: number and proportion of recommended (and essential) checklist (questions, examinations, and tests) completed, correctness of diagnosis, and correctness of case management.

FINDINGS

Though VCs completed 14•3% (IQR 9•5%-19•1%) of the recommended checklist, 63•2% (234/370, 95%CI 58•2%-68•0%) provided a correct diagnosis. Only 1•6% of VCs (6/370, 95%CI 0•7%-3•5%) gave correct management with both correct medication and referral, however 90•3% (334/370, 95%CI 86•8%-92•9%) provided partially correct management by referring patients to upper-level health facilities (89•5%, 331/370, 95%CI 85•9%-92•2%) or prescribing anti-epileptic drugs (AEDs) correctly (0•8%, 3/370, 95%CI 0•3%-2•4%). Around 1/4 VCs referred patients to Township Health Centers which usually were not staffed with pediatric neurologists. Fewer provided helpful medical advice to patients for daily management. The heuristic process was found to be negatively associated with the proportion of the recommended checklist that VCs completed, which is positively associated with correctness of diagnosis.

INTERPRETATION

Most VCs could diagnose and refer childhood epilepsy patients correctly; however, they lacked competence when it came to assuming the responsibility of primary care providers, referring efficiently, refilling AEDs, as well as supervising and instructing daily management of patients.

FUNDING

HY received the funding for this study from the "Health and Hope Fund" of the Business Development Center of the RCSC (Beijing) and UCB (Belgium). UCB provided support in the form of a salary for author DET.

Undiagnosed and Rare Diseases in Perinatal Medicine: Lessons in Context and Cognitive Diagnostic Error

Critically ill neonates experience high rates of morbidity and mortality. Major diagnostic errors are identified in up to 20% of autopsied neonatal intensive care unit deaths. Neonates with undiagnosed or rare congenital disorders may mimic critically ill neonates with more common acquired conditions. The context of the diagnostic evaluation can introduce unique biases that increase the likelihood of diagnostic error. Herein is presented a framework for understanding diagnostic errors in perinatal medicine, and individual, team, and systems-based solutions for improving diagnosis learned through the implementation and administration of an undiagnosed and rare disease program.

Diagnostic Errors in Initial Misdiagnosis of Optic Nerve Sheath Meningiomas

Importance

Diagnostic errors can lead to the initial misdiagnosis of optic nerve sheath meningiomas (ONSM), which can lead to vision loss.

Objective

To identify factors contributing to the initial misdiagnosis of ONSM.

Design, Setting, and Participants

We retrospectively reviewed 35 of 39 patients with unilateral ONSM (89.7%) who were seen in the tertiary neuro-ophthalmology practice at Emory University School of Medicine between January 2002 and March 2017. The Diagnosis Error Evaluation and Research taxonomy tool was applied to cases with missed/delayed diagnoses.

Exposures

Evaluation in a neuro-ophthalmology clinic.

Main Outcomes and Measures

Identifying the cause of diagnostic errors for patients who initially received a misdiagnosis who were found to have ONSM.

Results

Of 35 patients with unilateral ONSM (30 women [85.7%]; mean [SD] age, 45.26 [15.73] years), 25 (71%) had a diagnosis delayed for a mean (SD) of 62.60 (89.26) months. The most common diagnostic error (19 of 25 [76%]) was clinician assessment failure (errors in hypothesis generation and weighing), followed by errors in diagnostic testing (15 of 25 [60%]). The most common initial misdiagnosis was optic neuritis (12 of 25 [48%]), followed by the failure to recognize optic neuropathy in patients with ocular disorders. Five patients who received a misdiagnosis (20%) underwent unnecessary lumbar puncture, 12 patients (48%) unnecessary laboratory tests, and 6 patients (24%) unnecessary steroid treatment. Among the 16 patients who initially received a misdiagnosis that was later correctly diagnosed at our institution, 11 (68.8%) had prior magnetic resonance imaging (MRI) results that were read as healthy; 5 (45.5%) showed ONSM but were misread by a non-neuroradiologist and 6 (54.5%) were performed incorrectly (no orbital sequence or contrast). Sixteen of the 25 patients (64%) had a poor visual outcome.

Conclusions and Relevance

Biased preestablished diagnoses, inaccurate funduscopic examinations, a failure to order the correct test (MRI brain/orbits with contrast), and a failure to correctly interpret MRI results were the most common sources of diagnostic errors and delayed diagnosis with worse visual outcomes and increased cost (more visits and tests). Easier access to neuro-ophthalmologists, improved diagnostic strategies, and education regarding neuroimaging should help prevent diagnostic errors.

Pediatric faculty knowledge and comfort discussing diagnostic errors: a pilot survey to understand barriers to an educational program

Background

Improving Diagnosis in Healthcare calls for improved training in diagnostic reasoning and establishing non-judgmental forums to learn from diagnostic errors arising from heuristic-driven reasoning. Little is known about pediatric providers’ familiarity with heuristics or the culture surrounding forums where diagnostic errors are discussed. This study aimed to describe pediatric providers’ familiarity with common heuristics and perceptions surrounding public discussions of diagnostic errors.

Methods

We surveyed pediatric providers at a university-affiliated children’s hospital. The survey asked participants to identify common heuristics used during clinical reasoning (five definitions; four exemplar clinical vignettes). Participants answered questions regarding comfort publicly discussing their own diagnostic errors and barriers to sharing them.

Results

Seventy (30.6% response rate) faculty completed the survey. The mean number of correctly selected heuristics was 1.60/5 [standard deviation (SD)=1.13] and 1.01/4 (SD=1.06) for the definitions and vignettes, respectively. A low but significant correlation existed between correctly identifying a definition and selecting the correct heuristic in vignettes (Spearman’s ρ=0.27, p=0.02). Clinicians were significantly less likely to be “pretty” or “very” comfortable discussing diagnostic errors in public vs. private conversations (28.3% vs. 74.3%, p<0.01). The most frequently cited barriers to discussing errors were loss of reputation (62.9%) and fear of knowledge-base (58.6%) or decision-making (57.1%) being judged.

Conclusions

Pediatric providers demonstrated limited familiarity with common heuristics leading to diagnostic error. Greater years in practice is associated with more comfort discussing diagnostic errors, but negative peer and personal perceptions of diagnostic performance are common barriers to discussing errors publicly.

Hidden diagnosis behind viral infection: the danger of anchoring bias

Anchoring bias is one of the most common diagnostic biases that may lead to closed-minded thinking and could result in unnecessary tests, inappropriate patient management and even misdiagnosis. A 4-year-old boy was brought to the emergency department because of shaking chills. On the basis of bilateral swollen preauricular areas, high level of serum amylase and the prevalence of mumps, he initially received a diagnosis of mumps in spite of the shaking chills. However, blood culture turned out to be positive for two different kinds of bacteria. The patient finally received a diagnosis of polymicrobial bacteraemia resulting from suppurative appendicitis. We must consider and rule out bacteraemia in the differential diagnosis for patients who present with shaking chills, even in the presence of symptoms or information consistent with a more common viral infection such as mumps. In addition, intra-abdominal infection should be ruled out in the presence of polymicrobial enterobacteriaceae bacteraemia.

Multivariate Classification of Major Depressive Disorder Using the Effective Connectivity and Functional Connectivity

Major depressive disorder (MDD) is a mental disorder characterized by at least 2 weeks of low mood, which is present across most situations. Diagnosis of MDD using rest-state functional magnetic resonance imaging (fMRI) data faces many challenges due to the high dimensionality, small samples, noisy and individual variability. To our best knowledge, no studies aim at classification with effective connectivity and functional connectivity measures between MDD patients and healthy controls. In this study, we performed a data-driving classification analysis using the whole brain connectivity measures which included the functional connectivity from two brain templates and effective connectivity measures created by the default mode network (DMN), dorsal attention network (DAN), frontal-parietal network (FPN), and silence network (SN). Effective connectivity measures were extracted using spectral Dynamic Causal Modeling (spDCM) and transformed into a vectorial feature space. Linear Support Vector Machine (linear SVM), non-linear SVM, k-Nearest Neighbor (KNN), and Logistic Regression (LR) were used as the classifiers to identify the differences between MDD patients and healthy controls. Our results showed that the highest accuracy achieved 91.67% (p < 0.0001) when using 19 effective connections and 89.36% when using 6,650 functional connections. The functional connections with high discriminative power were mainly located within or across the whole brain resting-state networks while the discriminative effective connections located in several specific regions, such as posterior cingulate cortex (PCC), ventromedial prefrontal cortex (vmPFC), dorsal cingulate cortex (dACC), and inferior parietal lobes (IPL). To further compare the discriminative power of functional connections and effective connections, a classification analysis only using the functional connections from those four networks was conducted and the highest accuracy achieved 78.33% (p < 0.0001). Our study demonstrated that the effective connectivity measures might play a more important role than functional connectivity in exploring the alterations between patients and health controls and afford a better mechanistic interpretability. Moreover, our results showed a diagnostic potential of the effective connectivity for the diagnosis of MDD patients with high accuracies allowing for earlier prevention or intervention.