Integrating unified medical language system and association mining techniques into relevance feedback for biomedical literature search.

BACKGROUND: Finding highly relevant articles from biomedical databases is challenging not only because it is often difficult to accurately express a user's underlying intention through keywords but also because a keyword-based query normally returns a long list of hits with many citations being unwanted by the user. This paper proposes a novel biomedical literature search system, called BiomedSearch, which supports complex queries and relevance feedback. METHODS: The system employed association mining techniques to build a k-profile representing a user's relevance feedback. More specifically, we developed a weighted interest measure and an association mining algorithm to find the strength of association between a query and each concept in the article(s) selected by the user as feedback. The top concepts were utilized to form a k-profile used for the next-round search. BiomedSearch relies on Unified Medical Language System (UMLS) knowledge sources to map text files to standard biomedical concepts. It was designed to support queries with any levels of complexity. RESULTS: A prototype of BiomedSearch software was made and it was preliminarily evaluated using the Genomics data from TREC (Text Retrieval Conference) 2006 Genomics Track. Initial experiment results indicated that BiomedSearch increased the mean average precision (MAP) for a set of queries. CONCLUSIONS: With UMLS and association mining techniques, BiomedSearch can effectively utilize users' relevance feedback to improve the performance of biomedical literature search.

A functional temporal association mining approach for screening potential drug-drug interactions from electronic patient databases.

AIMS: Drug-drug interactions (DDIs) can result in serious consequences, including death. Existing methods for identifying potential DDIs in post-marketing surveillance primarily rely on spontaneous reports. These methods suffer from severe underreporting, incompleteness, and various bias. The aim of this study was to more effectively screen potential DDIs using patient electronic data and temporal association mining techniques. METHODS: We focus on discovery of potential DDIs by analyzing the temporal relationships between the concurrent use of two drugs of interest and the occurrences of various symptoms. We introduced innovative functional temporal association rules where the degree of temporal association between two events within a patient case was defined by a function. RESULTS: Preliminary test results on two drug pairs (i.e., and ) were classified into 260 clinically meaningful categories. These categories were evaluated by physicians and the results exhibited that all the potential DDIs were confined to top 20 of the 260 outcomes. CONCLUSIONS: Our methodology can be used to dramatically reduce a long list of association rules to a manageable list for further analysis and investigation by drug safety professionals.

Patient care complexity as perceived by primary care physicians.

INTRODUCTION: Currently there are various definitions of patient care complexity with little consensus. The numbers of patients with complex care needs are increasing. To improve interventions for "complex patients" and appropriately reimburse healthcare providers it is important to determine the characteristics or contextual factors contributing to complexity. METHOD: Action research methods were used to enhance an explicit understanding of complexity. Several conferences were organized and primary care physicians, nurses, social science faculty, and patients shared their perspectives on patient care complexity. A subset of attendees created a complex patient screening tool, which was piloted by 12 primary care physicians with 267 patients to identify which factors contribute to complexity. RESULTS: Complex patients were found to differ significantly from noncomplex patients based on factors associated with complexity. Based on latent class analysis, 58% of complex patients were characterized by multiple diagnoses, mental health issues, and a lack of effective participation in their care plans, while 42% of patients were considered complex because of multiple diagnoses only. In contrast, 90% of the noncomplex patients had no discernable pattern of health issues, while 10% of noncomplex patients had mental health and insurance issues that were easily managed. These results identify several factors that distinguish patients with complex care needs from those without complex care needs. The results also illustrate the heterogeneity within classes of patients identified as having complex care needs or non-complex needs. DISCUSSION: By identifying factors contributing to complexity, this research has important implications for enhancing the management of patients with complex care needs.

A temporal interestingness measure for drug interaction signal detection in post-marketing surveillance.

Drug-drug interactions (DDIs) can result in serious consequences, including death. Existing methods for identifying potential DDIs in post-marketing surveillance primarily rely on the FDA's (Food and Drug Administration) spontaneous reporting system. However, this system suffers from severe underreporting, which makes it difficult to timely collect enough valid cases for statistical analysis. In this paper, we study how to signal potential DDIs using patient electronic health data. Specifically, we focus on discovery of potential DDIs by analyzing the temporal relationships between the concurrent use of two drugs of interest and the occurrences of various symptoms using novel temporal association mining techniques we developed. A new interestingness measure called functional temporal interest was proposed to assess the degrees of temporal association between two drugs of interest and each symptom. The measure was employed to screen potential DDIs from 21,405 electronic patient cases retrieved from the Veterans Affairs Medical Center in Detroit, Michigan. The preliminary results indicate the usefulness of our method in finding potential DDIs for further analysis (e.g., epidemiology study) and investigation (e.g., case review) by drug safety professionals.

A potential causal association mining algorithm for screening adverse drug reactions in postmarketing surveillance.

Early detection of unknown adverse drug reactions (ADRs) in postmarketing surveillance saves lives and prevents harmful consequences. We propose a novel data mining approach to signaling potential ADRs from electronic health databases. More specifically, we introduce potential causal association rules (PCARs) to represent the potential causal relationship between a drug and ICD-9 (CDC. (2010). International Classification of Diseases, Ninth Revision (ICD-9). [Online]. Available: http://www.cdc.gov/nchs/icd/icd9.html) coded signs or symptoms representing potential ADRs. Due to the infrequent nature of ADRs, the existing frequency-based data mining methods cannot effectively discover PCARs. We introduce a new interestingness measure, potential causal leverage, to quantify the degree of association of a PCAR. This measure is based on the computational, experience-based fuzzy recognition-primed decision (RPD) model that we developed previously (Y. Ji, R. M. Massanari, J. Ager, J. Yen, R. E. Miller, and H. Ying, "A fuzzy logic-based computational recognition-primed decision model," Inf. Sci., vol. 177, pp. 4338-4353, 2007) on the basis of the well-known, psychology-originated qualitative RPD model (G. A. Klein, "A recognition-primed decision making model of rapid decision making," in Decision Making in Action: Models and Methods, 1993, pp. 138-147). The potential causal leverage assesses the strength of the association of a drug-symptom pair given a collection of patient cases. To test our data mining approach, we retrieved electronic medical data for 16,206 patients treated by one or more than eight drugs of our interest at the Veterans Affairs Medical Center in Detroit between 2007 and 2009. We selected enalapril as the target drug for this ADR signal generation study. We used our algorithm to preliminarily evaluate the associations between enalapril and all the ICD-9 codes associated with it. The experimental results indicate that our approach has a potential to better signal potential ADRs than risk ratio and leverage, two traditional frequency-based measures. Among the top 50 signal pairs (i.e., enalapril versus symptoms) ranked by the potential causal-leverage measure, the physicians on the project determined that eight of them probably represent true causal associations.

A distributed, collaborative intelligent agent system approach for proactive postmarketing drug safety surveillance.

Discovering unknown adverse drug reactions (ADRs) in postmarketing surveillance as early as possible is of great importance. The current approach to postmarketing surveillance primarily relies on spontaneous reporting. It is a passive surveillance system and limited by gross underreporting (<10% reporting rate), latency, and inconsistent reporting. We propose a novel team-based intelligent agent software system approach for proactively monitoring and detecting potential ADRs of interest using electronic patient records. We designed such a system and named it ADRMonitor. The intelligent agents, operating on computers located in different places, are capable of continuously and autonomously collaborating with each other and assisting the human users (e.g., the food and drug administration (FDA), drug safety professionals, and physicians). The agents should enhance current systems and accelerate early ADR identification. To evaluate the performance of the ADRMonitor with respect to the current spontaneous reporting approach, we conducted simulation experiments on identification of ADR signal pairs (i.e., potential links between drugs and apparent adverse reactions) under various conditions. The experiments involved over 275,000 simulated patients created on the basis of more than 1000 real patients treated by the drug cisapride that was on the market for seven years until its withdrawal by the FDA in 2000 due to serious ADRs. Healthcare professionals utilizing the spontaneous reporting approach and the ADRMonitor were separately simulated by decision-making models derived from a general cognitive decision model called fuzzy recognition-primed decision (RPD) model that we recently developed. The quantitative simulation results show that 1) the number of true ADR signal pairs detected by the ADRMonitor is 6.6 times higher than that by the spontaneous reporting strategy; 2) the ADR detection rate of the ADRMonitor agents with even moderate decision-making skills is five times higher than that of spontaneous reporting; and 3) as the number of patient cases increases, ADRs could be detected significantly earlier by the ADRMonitor.

Hospitalized patients with asthma who leave against medical advice: characteristics, reasons, and outcomes.

BACKGROUND: A discharge against medical advice (AMA) after an asthma hospitalization is a frustrating problem for health care providers, yet little is known about this occurrence. OBJECTIVE: To determine the baseline characteristics, reasons for leaving, and clinical outcomes of patients with asthma who leave AMA. METHODS: A retrospective study from 1999 to 2004 of all asthma discharges from 3 large hospitals in Detroit compared those who left AMA with those who left with medical approval. RESULTS: There were 180 patients who left AMA and 3457 patients who had a standard discharge. Patients with asthma who left AMA were more likely to be younger, male, have Medicaid or lack insurance, require intensive care unit admission, and have a lower socioeconomic status than patients with asthma discharged with approval (P < .05 for all comparisons). There was no difference in race, day of the week admitted, or month admitted. Among records that documented a reason for leaving AMA, the most common was dissatisfaction with care, although a variety of motives were found. Finally, patients who left AMA were more likely to have an asthma relapse within 30 days. This included both emergency department revisits (21.7% vs 5.4%; P < .001) and readmission to the hospital (8.5% vs 3.2%; P < .001). CONCLUSION: Patients with asthma who leave AMA have demographic and hospital admission characteristics that differ from those who leave with approval. The reasons why patients with asthma leave AMA are varied. Within 30 days, patients with asthma who leave AMA have much higher readmission and emergency department return rates. CLINICAL IMPLICATIONS: Patients with asthma who leave AMA are at increased risk of relapse.

An evaluation of the cost effectiveness of drotrecogin alfa (activated) relative to the number of organ system failures.

BACKGROUND: While drotrecogin alfa (activated) was shown to decrease absolute 28-day mortality by 6.1% in patients with severe sepsis in the Recombinant Human Protein C Worldwide Evaluation in Severe Sepsis (PROWESS) study, no mortality benefit was observed in the subset of patients with only one organ system failure. Consequently, some institutions restrict drotrecogin alfa (activated) use to patients with severe sepsis with >/=2 organ system failures. OBJECTIVE: To measure the cost effectiveness of drotrecogin alfa (activated) for treatment of severe sepsis in relation to the number of organ system failures and determine the economic impact of restricting drotrecogin alfa (activated) use based on the number of organ system failures. PERSPECTIVE: Policy perspective specific to our 340-bed, level I trauma centre. METHODS: A Monte Carlo simulation analysis was conducted to evaluate a hypothetical cohort of 10 000 patients with severe sepsis in four scenarios restricting treatment with drotrecogin alfa (activated) to patients with >/=1, >/=2, >/=3 or >/=4 organ system failures. The primary outcomes of 28-day all-cause mortality and serious bleeding were obtained from the PROWESS study. Costs (year 2002 values) were obtained from institutional financial records and literature estimates. The incremental cost per life saved at 28 days with drotrecogin alfa (activated) plus best standard care versus best standard care alone (placebo) was calculated. The incidence of severe sepsis and number of drotrecogin alfa (activated) candidates were estimated through chart review, and projected annual institutional expenditures were derived according to these data. RESULTS: With increasing number of organ system failures, the proportion of lives saved with drotrecogin alfa (activated) increased, and consequently the ICER decreased. Restriction of drotrecogin alfa (activated) to patients with >/=4 organ system failures was the most cost-effective scenario (0.11 lives saved; 56727 US dollars per life saved). For the nine patients that would be treated annually by our institution under this policy, one life would be saved at a total additional cost of 56160 US dollars per year. Use of the drug in patients with >/=1 or >/=2 organ system failures would save the greatest number of lives per year (4-5); however, restricting drotrecogin alfa (activated) to patients with >/=2 organ system failures would be the cheaper alternative (total additional cost 356022 US dollars vs 462204 US dollars . CONCLUSION: While restriction of drotrecogin alfa (activated) use to patients with sepsis with >/=4 organ system failures is the most cost-effective alternative, restriction to those with >/=2 organ system failures is the preferred alternative for our institution according to the number of lives saved and available financial resources.