Automated extraction of potential migraine biomarkers using a semantic graph.

PROBLEM: Biomedical literature and databases contain important clues for the identification of potential disease biomarkers. However, searching these enormous knowledge reservoirs and integrating findings across heterogeneous sources is costly and difficult. Here we demonstrate how semantically integrated knowledge, extracted from biomedical literature and structured databases, can be used to automatically identify potential migraine biomarkers. METHOD: We used a knowledge graph containing more than 3.5 million biomedical concepts and 68.4 million relationships. Biochemical compound concepts were filtered and ranked by their potential as biomarkers based on their connections to a subgraph of migraine-related concepts. The ranked results were evaluated against the results of a systematic literature review that was performed manually by migraine researchers. Weight points were assigned to these reference compounds to indicate their relative importance. RESULTS: Ranked results automatically generated by the knowledge graph were highly consistent with results from the manual literature review. Out of 222 reference compounds, 163 (73%) ranked in the top 2000, with 547 out of the 644 (85%) weight points assigned to the reference compounds. For reference compounds that were not in the top of the list, an extensive error analysis has been performed. When evaluating the overall performance, we obtained a ROC-AUC of 0.974. DISCUSSION: Semantic knowledge graphs composed of information integrated from multiple and varying sources can assist researchers in identifying potential disease biomarkers.

Benefits and problems of electronic information exchange as perceived by health care professionals: an interview study.

BACKGROUND: Various countries are currently implementing a national electronic patient record (n-EPR). Despite the assumed positive effects of n-EPRs, their overall adoption remains low and meets resistance from health care providers. This study aims to increase our understanding of health care providers' attitude towards the n-EPR, by investigating their perceptions of the benefits and problems of electronic information exchange in health care and the n-EPR in particular. METHODS: The study was conducted in three Dutch health care settings: acute care, diabetes care, and ambulatory mental health care. Two health care organisations were included per setting. Between January and June 2010, interviews were conducted with 17 stakeholders working in these organisations. Relevant themes were deduced by means of thematic qualitative analysis. RESULTS: Health care providers perceived electronic information exchange to promote the efficiency and quality of care. The problems they perceived in electronic information exchange mainly concerned the confidentiality and safety of information exchange and the reliability and quality of patient data. Many problems perceived by health care providers did not specifically apply to the n-EPR, but to electronic information exchange in general. CONCLUSIONS: The implementation of the Dutch n-EPR has mainly followed a top-down approach, thereby neglecting the fact that the perceptions and preferences of its users (health care providers) need to be addressed in order to achieve successful implementation. The results of this study provide valuable suggestions about how to promote health care providers' willingness to adopt electronic information exchange, which can be useful for other countries currently implementing an n-EPR. Apart from providing information about the benefits and usefulness of electronic information exchange, efforts should be focused on minimising the problems as perceived by health care providers. The safety and confidentiality of electronic information exchange can be improved by developing tools to evaluate the legitimacy of access to electronic records, by increasing health care providers' awareness of the need to be careful when using patient data, and by measures to limit access to sensitive patient data. Improving health care providers' recording behaviour is important to improve the reliability and quality of electronically exchanged patient data.

Towards patient-related information needs.

The quality of health care depends, among other factors, on the quality of a physician's domain knowledge. Since it is impossible to keep up with all new findings and developments, physicians usually have gaps in their domain knowledge. To handle exceptional cases, access to the full range of medical literature is required. The specific literature needed for appropriate treatment of the patient is described by a physician's information need. Physicians are often unaware of their information needs. To support them, this paper presents a first step towards automatically formulating patient-related information needs. We start investigating how we can model a physician's information needs in general. Then we propose an approach to instantiate the model into a representation of a physician's information needs using the patient data as stored in a medical record. Our experiments show that this approach is feasible. Since the number of formulated patient-related information needs is rather high, it has to be reduced. To reduce the number of formulated information needs we propose the use of additional knowledge. Four types of knowledge are discussed, viz. (a) knowledge about temporal aspects, (b) domain knowledge, (c) knowledge about a physician's specialism, and (d) a user model. Future research has to clarify which type of knowledge (or combination thereof) is most appropriate for our purpose. It is expected that the resultant set of information needs will have a manageable size and contributes to the quality of health care.

Avoiding literature overload in the medical domain.

The retrieval of patient-related literature is hampered by the large size of medical literature. Various computer systems have been developed to assist physicians during information retrieval. However, in general, physicians lack the time and skills required to employ such systems effectively. Our goal is to investigate to what extent a physician can be provided with patient-related literature without spending extra time and without acquiring additional skills. In previous research we developed a method to formulate a physician's patient-related information needs automatically, without requiring any interaction between the physician and the system. The formulated information needs can be used as a starting point for literature retrieval. As a result we found that the number of information needs formulated per physician was quite high and had to be reduced to avoid a literature overload. In this paper we present four types of knowledge that may be used to accomplish a reduction in the number of information needs. The usefulness of each of these knowledge types depends heavily on the specific cause underlying the multitude of information needs. To determine the nature of the cause, we performed an experimental analysis. The results of the analysis led us to conclude that the knowledge types can be ordered according to their appropriateness as follows: (1) knowledge concerning temporal aspects, (2) knowledge concerning a physician's specialism, (3) domain knowledge, and (4) a user model. Further research has to be performed, in particular on precisely assessing the performance of each type of knowledge within our domain.

From patient data to information needs.

The goal of this paper is to contribute to the improvement of the quality of care. For physicians, it is a problem that they are often not aware of gaps in their knowledge and the corresponding information needs. Our research aim is to resolve this problem by formulating information needs automatically. Based on these information needs, patient-specific literature can be retrieved. As a first step, we investigate how to model a physician's information needs. Thereafter, we design and analyse an approach to instantiate the model with patient data, resulting in information-need templates that are able to represent patient-specific information needs. Our experiments show that a physician's information needs can be modelled adequately and can be substantiated into patient-specific information needs. Since the number of formulated information needs is rather high, future research will focus on methods that restrict the set of automatically formulated information needs to a more specialized set.

Towards patient-related information needs.

The quality of health care depends, among others, on the quality of a physician's domain knowledge. Since it is impossible to keep up with all new findings and developments, physicians usually have gaps in their domain knowledge. To handle exceptional cases, access to the full range of medical literature is required. The specific literature needed for appropriate treatment of the patient is described by a physician's information need. Physicians are often unaware of their information needs. To support them, this paperThis research is part of the MIA project (Medical Information Agent), which is funded by NWO (grant number 634.000.021). aims at presenting a first step towards automatically formulating patient-related information needs. We start investigating how we can model a physician's information needs in general. Then we propose an approach to instantiate the model into a representation of a physician's information needs using the patient data as stored in a medical record. Our experiments show that this approach is feasible. Since the number of formulated patient-related information needs is rather high, we propose the use of filters. Future research will focus on the combination of personalization and filtering. It is expected that the resultant set of information needs will have a manageable size and contributes to the quality of health care.