Development of a Social Network for People Without a Diagnosis (RarePairs): Evaluation Study.

BACKGROUND: Diagnostic delay in rare disease (RD) is common, occasionally lasting up to more than 20 years. In attempting to reduce it, diagnostic support tools have been studied extensively. However, social platforms have not yet been used for systematic diagnostic support. This paper illustrates the development and prototypic application of a social network using scientifically developed questions to match individuals without a diagnosis. OBJECTIVE: The study aimed to outline, create, and evaluate a prototype tool (a social network platform named RarePairs), helping patients with undiagnosed RDs to find individuals with similar symptoms. The prototype includes a matching algorithm, bringing together individuals with similar disease burden in the lead-up to diagnosis. METHODS: We divided our project into 4 phases. In phase 1, we used known data and findings in the literature to understand and specify the context of use. In phase 2, we specified the user requirements. In phase 3, we designed a prototype based on the results of phases 1 and 2, as well as incorporating a state-of-the-art questionnaire with 53 items for recognizing an RD. Lastly, we evaluated this prototype with a data set of 973 questionnaires from individuals suffering from different RDs using 24 distance calculating methods. RESULTS: Based on a step-by-step construction process, the digital patient platform prototype, RarePairs, was developed. In order to match individuals with similar experiences, it uses answer patterns generated by a specifically designed questionnaire (Q53). A total of 973 questionnaires answered by patients with RDs were used to construct and test an artificial intelligence (AI) algorithm like the k-nearest neighbor search. With this, we found matches for every single one of the 973 records. The cross-validation of those matches showed that the algorithm outperforms random matching significantly. Statistically, for every data set the algorithm found at least one other record (match) with the same diagnosis. CONCLUSIONS: Diagnostic delay is torturous for patients without a diagnosis. Shortening the delay is important for both doctors and patients. Diagnostic support using AI can be promoted differently. The prototype of the social media platform RarePairs might be a low-threshold patient platform, and proved suitable to match and connect different individuals with comparable symptoms. This exchange promoted through RarePairs might be used to speed up the diagnostic process. Further studies include its evaluation in a prospective setting and implementation of RarePairs as a mobile phone app.

Common pre-diagnostic features in individuals with different rare diseases represent a key for diagnostic support with computerized pattern recognition?

BACKGROUND: Rare diseases (RD) result in a wide variety of clinical presentations, and this creates a significant diagnostic challenge for health care professionals. We hypothesized that there exist a set of consistent and shared phenomena among all individuals affected by (different) RD during the time before diagnosis is established. OBJECTIVE: We aimed to identify commonalities between different RD and developed a machine learning diagnostic support tool for RD. METHODS: 20 interviews with affected individuals with different RD, focusing on the time period before their diagnosis, were performed and qualitatively analyzed. Out of these pre-diagnostic experiences, we distilled key phenomena and created a questionnaire which was then distributed among individuals with the established diagnosis of i.) RD, ii.) other common non-rare diseases (NRO) iii.) common chronic diseases (CD), iv.), or psychosomatic/somatoform disorders (PSY). Finally, four combined single machine learning methods and a fusion algorithm were used to distinguish the different answer patterns of the questionnaires. RESULTS: The questionnaire contained 53 questions. A total sum of 1763 questionnaires (758 RD, 149 CD, 48 PSY, 200 NRO, 34 healthy individuals and 574 not evaluable questionnaires) were collected. Based on 3 independent data sets the 10-fold stratified cross-validation method for the answer-pattern recognition resulted in sensitivity values of 88.9% to detect the answer pattern of a RD, 86.6% for NRO, 87.7% for CD and 84.2% for PSY. CONCLUSION: Despite the great diversity in presentation and pathogenesis of each RD, patients with RD share surprisingly similar pre-diagnosis experiences. Our questionnaire and data-mining based approach successfully detected unique patterns in groups of individuals affected by a broad range of different rare diseases. Therefore, these results indicate distinct patterns that may be used for diagnostic support in RD.

Künstliche Intelligenz zur diagnostischen Unterstützung ausgewählter seltener lysosomaler Speichererkrankungen: Ergebnisse einer Pilotstudie.

BACKGROUND: Diagnosing a rare metabolic disease challenges physicians and affected individuals and their families. To support the diagnostic pathway, a diagnostic tool was developed using the experiences of the affected individuals gained in interviews. METHODS: 17 interviews with parents or individuals with a selected rare metabolic disease (Mucopolysaccharidosis (MPS), M. Fabry and M. Gaucher) were performed and qualitatively analysed using the standardized methods of Colaizzi. The results are reflected in diagnostic questionnaires. The questionnaires were distributed and answered by parents or individuals with an established diagnosis of MPS, M. Fabry or M. Gaucher and a control group. Four combined data mining classifiers were trained to detect suspicious answer patterns in the questionnaires. RESULTS: 56 questionnaires were used for training and cross-validation tests of the binary data mining system resulting in a sensitivity value of 91% for the diagnosis 'MPS'. Another 20 questionnaires which have not been used for the training process could be evaluated as a preliminary prospective test. Out of these 20 questionnaires the test delivered 18 correct diagnoses (90%). DISCUSSION AND CONCLUSIONS: Questionnaires for diagnostic support based on interviews with parents and affected individuals were developed and answer patterns were analysed with an ensemble of classifiers. Although preliminary, the results illustrate the potential of answer pattern recognition using data mining techniques. This approach might prove useful for diagnostic support in selected metabolic diseases.

Patient's Experience in Pediatric Primary Immunodeficiency Disorders: Computerized Classification of Questionnaires.

INTRODUCTION: Primary immunodeficiency disorders (PIDs) are a heterogeneous group of more than 200 rare diseases. Timely diagnosis is of uttermost importance. Therefore, we aimed to develop a diagnostic questionnaire with computerized pattern-recognition in order to support physicians to identify suspicious patient histories. MATERIALS AND METHODS: Standardized interviews were conducted with guardians of children with PID. The questionnaire based on parental observations was developed using Colaizzis' framework for content analysis. Answers from 64 PID patients and 62 controls were analyzed by data mining methods in order to make a diagnostic prediction. Performance was evaluated by k-fold stratified cross-validation. RESULTS: The diagnostic support tool achieved a diagnostic sensitivity of up to 98%. The analysis of 12 interviews revealed 26 main phenomena observed by parents in the pre-diagnostic period. The questions were systematically phrased and selected resulting in a 36-item questionnaire. This was answered by 126 patients with or without PID to evaluate prediction. Item analysis revealed significant questions. DISCUSSION: Our approach proved suitable for recognizing patterns and thus differentiates between observations of PID patients and control groups. These findings provide the basis for developing a tool supporting physicians to consider a PID with a questionnaire. These data support the notion that patient's experience is a cornerstone in the diagnostic process.

Diagnostic needs for rare diseases and shared prediagnostic phenomena: Results of a German-wide expert Delphi survey.

BACKGROUND: Worldwide approximately 7,000 rare diseases have been identified. Accordingly, 4 million individuals live with a rare disease in Germany. The mean time to diagnosis is about 6 years and patients receive several incorrect diagnoses during this time. A multiplicity of factors renders diagnosing a rare disease extremely difficult. Detection of shared phenomena among individuals with different rare diseases could assist the diagnostic process. In order to explore the demand for diagnostic support and to obtain the commonalities among patients, a nationwide Delphi survey of centers for rare diseases and patient groups was conducted. METHODS: A two-step Delphi survey was conducted using web-based technologies in all centers for rare diseases in Germany. Moreover, the leading patient support group, the German foundation for rare diseases (ACHSE), was contacted to involve patients as experts in their disease. In the survey the experts were invited to name rare diseases with special need for diagnostic improvement. Secondly, communal experiences of affected individuals were collected. RESULTS: 166 of 474 contacted experts (35%) participated in the first round of the Delphi process and 95 of 166 (57%) participated in the second round. Metabolic (n = 74) and autoimmune diseases (n = 39) were ranked the highest for need for diagnostic support. For three diseases (i.e. scleroderma, Pompe's disease, and pulmonary arterial hypertension), a crucial need for diagnostic support was explicitly stated. A typical experience of individuals with a rare disease was stigmatization of having psychological or psychosomatic problems. In addition, most experts endured an 'odyssey' of seeing many different medical specialists before a correct diagnosis (n = 38) was confirmed. CONCLUSION: There is need for improving the diagnostic process in individuals with rare diseases. Shared experiences in individuals with a rare disease were observed, which could possibly be utilized for diagnostic support in the future.

Diagnostic support for selected neuromuscular diseases using answer-pattern recognition and data mining techniques: a proof of concept multicenter prospective trial.

BACKGROUND: Diagnosis of neuromuscular diseases in primary care is often challenging. Rare diseases such as Pompe disease are easily overlooked by the general practitioner. We therefore aimed to develop a diagnostic support tool using patient-oriented questions and combined data mining algorithms recognizing answer patterns in individuals with selected neuromuscular diseases. A multicenter prospective study for the proof of concept was conducted thereafter. METHODS: First, 16 interviews with patients were conducted focusing on their pre-diagnostic observations and experiences. From these interviews, we developed a questionnaire with 46 items. Then, patients with diagnosed neuromuscular diseases as well as patients without such a disease answered the questionnaire to establish a database for data mining. For proof of concept, initially only six diagnoses were chosen (myotonic dystrophy and myotonia (MdMy), Pompe disease (MP), amyotrophic lateral sclerosis (ALS), polyneuropathy (PNP), spinal muscular atrophy (SMA), other neuromuscular diseases, and no neuromuscular disease (NND). A prospective study was performed to validate the automated malleable system, which included six different classification methods combined in a fusion algorithm proposing a final diagnosis. Finally, new diagnoses were incorporated into the system. RESULTS: In total, questionnaires from 210 individuals were used to train the system. 89.5 % correct diagnoses were achieved during cross-validation. The sensitivity of the system was 93-97 % for individuals with MP, with MdMy and without neuromuscular diseases, but only 69 % in SMA and 81 % in ALS patients. In the prospective trial, 57/64 (89 %) diagnoses were predicted correctly by the computerized system. All questions, or rather all answers, increased the diagnostic accuracy of the system, with the best results reached by the fusion of different classifier methods. Receiver operating curve (ROC) and p-value analyses confirmed the results. CONCLUSION: A questionnaire-based diagnostic support tool using data mining methods exhibited good results in predicting selected neuromuscular diseases. Due to the variety of neuromuscular diseases, additional studies are required to measure beneficial effects in the clinical setting.

Diagnostic Support for Selected Paediatric Pulmonary Diseases Using Answer-Pattern Recognition in Questionnaires Based on Combined Data Mining Applications--A Monocentric Observational Pilot Study.

BACKGROUND: Clinical symptoms in children with pulmonary diseases are frequently non-specific. Rare diseases such as primary ciliary dyskinesia (PCD), cystic fibrosis (CF) or protracted bacterial bronchitis (PBB) can be easily missed at the general practitioner (GP). OBJECTIVE: To develop and test a questionnaire-based and data mining-supported tool providing diagnostic support for selected pulmonary diseases. METHODS: First, interviews with parents of affected children were conducted and analysed. These parental observations during the pre-diagnostic time formed the basis for a new questionnaire addressing the parents' view on the disease. Secondly, parents with a sick child (e.g. PCD, PBB) answered the questionnaire and a data base was set up. Finally, a computer program consisting of eight different classifiers (support vector machine (SVM), artificial neural network (ANN), fuzzy rule-based, random forest, logistic regression, linear discriminant analysis, naive Bayes and nearest neighbour) and an ensemble classifier was developed and trained to categorise any given new questionnaire and suggest a diagnosis. For estimating the diagnostic accuracy, we applied ten-fold stratified cross validation. RESULTS: All questionnaires of patients suffering from CF, asthma (AS), PCD, acute bronchitis (AB) and the healthy control group were correctly diagnosed by the fusion algorithm. For the pneumonia (PM) group 19/21 (90.5%) and for the PBB group 17/18 (94.4%) correct diagnoses could be reached. The program detected the correct diagnoses with an overall sensitivity of 98.8%. Receiver operating characteristics (ROC) analyses confirmed the accuracy of this diagnostic tool. Case studies highlighted the applicability of the tool in the daily work of a GP. CONCLUSION: For children with symptoms of pulmonary diseases a questionnaire-based diagnostic support tool using data mining techniques exhibited good results in arriving at diagnostic suggestions. In the hands of a doctor, this tool could be of value in arousing awareness for rare pulmonary diseases such as PCD or CF.

Supporting diagnostic decisions using hybrid and complementary data mining applications: a pilot study in the pediatric emergency department.

INTRODUCTION: This article demonstrates the capacity of a combination of different data mining (DM) methods to support diagnosis in pediatric emergency patients. By using a novel combination of these DM procedures, a computer-based diagnosis was created. METHODS: A support vector machine (SVM), artificial neural networks (ANNs), fuzzy logics, and a voting algorithm were simultaneously used to allocate a patient to one of 18 diagnoses (e.g., pneumonia, appendicitis). Anonymized data sets of patients who presented in the emergency department (ED) of a pediatric care clinic were chosen. For each patient, 26 identical clinical and laboratory parameters were used (e.g., blood count, C-reactive protein) to finally develop the program. RESULTS: The combination of four DM operations arrived at a correct diagnosis in 98% of the cases, retrospectively. A subgroup analysis showed that the highest diagnostic accuracy was for appendicitis (97% correct diagnoses) and idiopathic thrombocytopenic purpura or erythroblastopenia (100% correct diagnoses). During the prospective testing, 81% of the patients were correctly diagnosed by the system. DISCUSSION: The combination of these DM methods was suitable for proposing a diagnosis using both laboratory and clinical parameters. We conclude that an optimized combination of different but complementary DM methods might serve to assist medical decisions in the ED.