Tool-supported Interactive Correction and Semantic Annotation of Narrative Clinical Reports.

OBJECTIVES: Our main objective is to design a method of, and supporting software for, interactive correction and semantic annotation of narrative clinical reports, which would allow for their easier and less erroneous processing outside their original context: first, by physicians unfamiliar with the original language (and possibly also the source specialty), and second, by tools requiring structured information, such as decision-support systems. Our additional goal is to gain insights into the process of narrative report creation, including the errors and ambiguities arising therein, and also into the process of report annotation by clinical terms. Finally, we also aim to provide a dataset of ground-truth transformations (specific for Czech as the source language), set up by expert physicians, which can be reused in the future for subsequent analytical studies and for training automated transformation procedures. METHODS: A three-phase preprocessing method has been developed to support secondary use of narrative clinical reports in electronic health record. Narrative clinical reports are narrative texts of healthcare documentation often stored in electronic health records. In the first phase a narrative clinical report is tokenized. In the second phase the tokenized clinical report is normalized. The normalized clinical report is easily readable for health professionals with the knowledge of the language used in the narrative clinical report. In the third phase the normalized clinical report is enriched with extracted structured information. The final result of the third phase is a semi-structured normalized clinical report where the extracted clinical terms are matched to codebook terms. Software tools for interactive correction, expansion and semantic annotation of narrative clinical reports has been developed and the three-phase preprocessing method validated in the cardiology area. RESULTS: The three-phase preprocessing method was validated on 49 anonymous Czech narrative clinical reports in the field of cardiology. Descriptive statistics from the database of accomplished transformations has been calculated. Two cardiologists participated in the annotation phase. The first cardiologist annotated 1500 clinical terms found in 49 narrative clinical reports to codebook terms using the classification systems ICD 10, SNOMED CT, LOINC and LEKY. The second cardiologist validated annotations of the first cardiologist. The correct clinical terms and the codebook terms have been stored in a database. CONCLUSIONS: We extracted structured information from Czech narrative clinical reports by the proposed three-phase preprocessing method and linked it to electronic health records. The software tool, although generic, is tailored for Czech as the specific language of electronic health record pool under study. This will provide a potential etalon for porting this approach to dozens of other less-spoken languages. Structured information can support medical decision making, quality assurance tasks and further medical research.

Formalization of clinical practice guidelines.

Clinical practice guidelines are textual recommendations based on the consensus of medical experts with the aim to solve diagnostic and therapeutic problems. For more advanced use in real medical applications it is necessary to find out mathematical models of physicians' decision-making processes. The acquisition of a formal model from text-based guidelines is a crucial point for development of decision support systems. We introduce a system for formalization and presentation of medical knowledge contained in clinical practice guidelines where knowledge formalization is based on the GLIF model.

Medical knowledge representation system.

The aim of this article is to present a design of a Medical Knowledge Representation System (MEKRES). The system automatically offers relevant formalized knowledge by extended GLIF (Guidelines Interchange Format) models to participants (patient, physician, operator, ..) on the basis of acquired data. This selection algorithm is based on key attributes and cooperation with knowledge ontologies.

SAH gene variants revisited in the European Project On Genes in Hypertension.

OBJECTIVE: Previous studies found significant association of hypertension and hypertension-related phenotypes with genetic variation in SAH (Spontaneously hypertensive rat-clone A-Hypertension-associated). We sought independent confirmation of these findings in the European Project On Genes in Hypertension. METHODS AND RESULTS: We randomly recruited 2603 relatives from 560 families and 31 unrelated subjects from six European populations (mean age 38.8 +/- 15.7 years; 52.1% women). We measured systolic/diastolic blood pressure (mean, 122.4/76.6 mmHg), body mass index (24.9 kg/m2), triceps skinfold (1.7 cm), waist-to-hip ratio (0.83 units), serum total and high-density lipoprotein (HDL) cholesterol (5.14 and 1.33 mmol/l), serum triglycerides (1.95 mmol/l) and blood glucose (4.90 mmol/l). We genotyped the G-1606A and -962del/ins polymorphisms. In all subjects, the allele frequencies were 11.8 and 29.5% for -1606A and -962del, respectively. Lewontin's D' was 0.97 (P < 0.0001). Haplotype frequencies were 58.8% for -1606G plus -962ins, 29.5% for -1606G plus -962del, and 11.7% for -1606A plus -962ins. Both before and after adjustment for covariates, none of the phenotype-genotype associations approached statistical significance. Our study had 80% power to detect on two-sided tests (P = 0.05), effect sizes of 1.8/1.3 mmHg for systolic/diastolic blood pressure, 0.52 kg/m2 for body mass index, 0.01 units for the waist-to-hip ratio, 0.96 mm for the triceps skinfold, 0.13 and 0.05 mmol/l for total and HDL cholesterol, 0.18 mmol/l for serum triglycerides, and 0.11 mmol/l for blood glucose. The family-based analyses did not reveal population stratification (P > or = 0.67). CONCLUSION: The evidence supporting an association of hypertension or hypertension-related phenotypes with the SAH gene remains equivocal in human studies.

Sympathetic activity, assessed by power spectral analysis of heart rate variability, in white-coat, masked and sustained hypertension versus true normotension.

OBJECTIVE: To assess, in a population-based approach, sympathetic nervous system activity by the use of power spectral analysis of heart rate variability, in normotension, white-coat hypertension, masked hypertension and sustained hypertension. METHODS: The electrocardiographic RR interval was registered in the supine and standing positions and the low-frequency and high-frequency components of its variability were quantified. Cut-off values of 140/90 mmHg for conventional blood pressure and 135/85 mmHg for daytime ambulatory blood pressure were used to define the four blood pressure groups. RESULTS: After exclusion of patients with diabetes, myocardial infarction or treated hypertension, 1485 subjects with complete data remained for the analysis in the supine position. Age averaged 39 +/- 14 years; 54% were women. Conventional and ambulatory blood pressure averaged, respectively, 122 +/- 16/79 +/- 11 mmHg and 124 +/- 12/77 +/- 8 mmHg. After adjusting for demographic, anthropometric and lifestyle characteristics, the low-frequency to high-frequency ratio (geometric mean) averaged 0.81 in normotension and was significantly higher in white-coat hypertension (1.11; P < 0.001), based on a higher low-frequency component and a lower high-frequency component (P < 0.01). This ratio was not significantly different between normotension, masked hypertension (0.97) and sustained hypertension (0.93). The adjusted standing-to-supine ratio of the high-frequency component (geometric mean) was significantly higher in sustained hypertension (0.50) than in normotension (0.39; P < 0.01), but not in white-coat (0.40) and masked hypertension (0.45). CONCLUSION: The findings at rest are compatible with increased sympathetic activity and decreased parasympathetic modulation in white-coat hypertension, with normal autonomic cardiac regulation in masked and sustained hypertension. In addition, sustained hypertension is characterized by a blunted decrease of the high-frequency component on standing.

Sodium excretion as a modulator of genetic associations with cardiovascular phenotypes in the European Project on Genes in Hypertension.

Hypertension is a chronic age-related disorder, affecting nearly 20% of all adult Europeans. This disease entails debilitating cardiovascular complications and is the leading cause for drug prescriptions in Europeans older than 50 years. Intensive research over the past two decades has so far failed to identify common genetic polymorphisms with a major impact on blood pressure or associated cardiovascular phenotypes, suggesting that multiple genes each with a minor impact, along with gene-gene and gene-environment interactions, play a role. The European Project on Genes in Hypertension (EPOGH) is a large-scale, family-based study in which participants from seven different populations were phenotyped and genotyped according to standardized procedures. This review article summarizes the initial 5-year findings and puts these observations into perspective against other published studies. The EPOGH demonstrated that phenotype-genotype relations strongly depend on host factors such as gender and lifestyle, in particular salt intake as reflected by the 24-h urinary excretion of sodium. The EPOGH therefore highlights the concept that phenotype-genotype relations can only be studied within a defined ecogenetic context.

Medical guidelines presentation and comparing with Electronic Health Record.

Electronic Health Record (EHR) systems are now being developed in many places. More advanced systems provide also reminder facilities, usually based on if-then rules. In this paper we propose a method how to build the reminder facility directly upon the guideline interchange format (GLIF) model of medical guidelines. The method compares data items on the input of EHR system with medical guidelines GLIF model and is able to reveal if the input data item, that represents patient diagnosis or proposed patient treatment, contradicts with medical guidelines or not. The reminder facility can be part of EHR system itself or it can be realized by a stand-alone reminder system (SRS). The possible architecture of stand-alone reminder system is described in this paper and the advantages of stand-alone solution are discussed. The part of the EHR system could be also a browser that would present graphical GLIF model in easy to understand manner on the user screen. This browser can be data driven and focus attention of user to the relevant part of medical guidelines GLIF model.

Analysis of guideline compliance--a data mining approach.

While guideline-based decision support is safety-critical and typically requires human interaction, offline analysis of guideline compliance can be performed to large extent automatically. We examine the possibility of automatic detection of potential non-compliance followed up with (statistical) association mining. Only frequent associations of non-compliance patterns with various patient data are submitted to medical expert for interpretation. The initial experiment was carried out in the domain of hypertension management.

Genetic variation in CYP11B2 and AT1R influences heart rate variability conditional on sodium excretion.

Sympathetic tone increases with stimulation of the renin-angiotensin system and is under the influence of salt intake. In the European Project On Genes in Hypertension (EPOGH), we investigated whether polymorphisms in the genes encoding aldosterone synthase (CYP11B2 C-344T) and the type-1 angiotensin II receptor (AT1R A1166C) affect the autonomic modulation of heart rate at varying levels of salt intake. We measured the low frequency (LF) and high frequency (HF) components of heart rate variability and their ratio (LF:HF) in the supine and standing positions in 1797 participants (401 families and 320 unrelated subjects) randomly selected from 6 European populations, whose average urinary sodium excretion ranged from 163 to 245 mmol/d. In multivariate analyses with sodium excretion analyzed as a continuous variable, we explored the phenotype-genotype associations using generalized estimating equations and quantitative transmission disequilibrium tests. Across populations, there was no heterogeneity in the phenotype-genotype relations. The genotypic effects differed according to sodium excretion. In subjects with sodium excretion <190 mmol/d (median), supine heart rate, LF, and LF:HF increased and HF decreased with the number of CYP11B2 -344T alleles, and the orthostatic changes in LF, HF, and LF:HF were blunted in carriers of the AT1R 1166C allele. In subjects with sodium excretion >190 mmol/d, these associations with the CYP11B2 and AT1R polymorphisms were nonsignificant or in the opposite direction, respectively. Thus, CYP11B2 C-344T and AT1R A1166C polymorphisms affect the autonomic modulation of heart rate, but these genetic effects depend on sodium excretion.

Association of -344/T/C aldosterone synthase polymorphism (CYP11B2) with left ventricular structure and humoral parameters in young normotensive men.

BACKGROUND: Aldosterone plays an important role in development of left ventricular (LV) hypertrophy and myocardial fibrosis. We assessed the influence of the T-344C polymorphism of aldosterone synthase - the rate-limiting enzyme in aldosterone biosynthesis - on the structure of the left ventricle in young normotensive men. DESIGN AND METHODS: The population included 113 normotensive mid-European Caucasian men aged 18-40 years (mean 27 +/- 5 years). The genotype was assessed using polymerase chain reaction with subsequent cleavage with restriction enzyme HAEIII (restriction fragment length polymorphism method) and visualization with ethidium bromide. Plasma renin activity (PRA) and plasma aldosterone were measured. All subjects were examined by echocardiography and LV mass was assessed by using M-mode based ASE formula. RESULTS: The distribution of the genotypes was TT 23%:TC 55%:CC 22%. There were no differences in blood pressure among the groups. Men with the TT genotype had significantly higher levels of PRA (2.7 +/- 1.7 vs 1.8 +/- 1.0 vs 1.8 +/- 1.1 ng/ml/h, p < 0.01) and slightly higher plasma levels of aldosterone (113 +/- 64 vs 93 +/- 43 vs 87 +/- 39 pg/ml, p = 0,12). In the whole population, LV mass index (LVMI) did not differ significantly among the genotypes (92 +/- 16 vs 86 +/- 18 vs 84 +/- 16 g/m, p=0.20). In the population divided according to PRA, subjects with high renin had significantly higher LVMI in presence of the TT genotype (95 +/- 17 vs 84 +/- 16 vs 81 +/- 15 g/m, p < 0.05). CONCLUSIONS: In agreement with previous studies, we found that the TT genotype of T-344C polymorphism of aldosterone synthase gene was associated with significantly higher levels of PRA in normotensive men. In subjects with high PRA, the TT genotype was associated with higher values of the LVMI.

Presentation of medical guidelines on a computer.

Free-text forms of medical guidelines that are used in medical care are often cumbersome and difficult to memorize. Therefore a system that is able to present guidelines in a user-friendly manner has been designed. Guidelines are at first formalized by means of the popular GLIF3 model. Subsequently, the GLIF3 model is coded in XML. The system uses patient data and an XML-coded GLIF3 model graph evaluating the conditions of decision steps. If some conditions cannot be evaluated, (due to the unavailability of required data items), the system stops and highlights a branch from the root of the decision tree to the final step. Then, the user can input missing data into the system so that the system can continue to provide visualization. Thus, the presentation of guidelines is data-driven, making its use easier. The visualization system and the Electronic Health Record can be integrated in a system that can, during examination of a patient, suggest subsequent medical actions according to medical guidelines. Such a system has been designed and is now under development.

Quality control of the blood pressure phenotype in the European Project on Genes in Hypertension.

OBJECTIVES: In the European Project on Genes in Hypertension (EPOGH) standardized epidemiological methods were used to determine complex phenotypes consisting of blood pressure (BP) in combination with other traits. In this report, we present the quality control of one of the BP phenotypes. METHODS: In seven European countries eight different research groups recruited random samples of nuclear families. Trained observers measured the BP five times consecutively with the participants in the seated position at each of two separate home visits, 1 to 3 weeks apart, according to the guidelines of the British Hypertension Society. Quality assurance and quality control of this BP phenotype were implemented according to detailed instructions defined in the protocol of the EPOGH study. RESULTS: On 31 August 2001, BP measurements of 2476 subjects were available for analysis. Fewer BP readings than the five planned per visit occurred in one of the eight centres, but only in 0.4% of the home visits. Across centres the relative frequency of identical consecutive readings for systolic or diastolic blood pressure varied from 0 to 6%. The occurrence of odd readings ranged from 0 to 0.1%. Of the 49,488 systolic and diastolic BP readings, 24.0% ended on a zero (expected 20%). In most EPOGH centres there was a progressive decline in the BP from the first to the second home visit. Overall, these decreases averaged 2.36 mmHg [95% confidence interval (CI): 1.98-2.74, P < 0.001] for systolic BP and 1.74 mmHg (95% CI: 1.46-2.02, P < 0.001) for diastolic BP. CONCLUSIONS: Quality assurance and control should be planned at the design stage of a project involving BP measurement and implemented from its very beginnings until the end. The procedures of quality assurance set up in the EPOGH study for the BP measurements resulted in a well-defined BP phenotype, which was consistent across centres.

Association between tyrosine hydroxylase polymorphisms and left ventricular structure in young normotensive men.

Tyrosine hydroxylase (TH) is a rate-limiting enzyme for catecholamine biosynthesis. Increased sympathetic activity is associated with an increased left ventricular (LV) mass. However, the influence of TH gene polymorphisms on LV structure and function has yet to be investigated. Here, we analyse the association of Val-81-Met and tetranucleotide TCAT repeat TH polymorphisms with LV structure and function (assessed by echocardiography) in 108 normotensive men aged < or = 35 years (mean age: 25+/-4 years) with body mass index (BMI) < or = 30 kg/m2 (mean BMI: 23+/-3 kg/m2). The distribution of genotypes was VV homozygotes (n=42), VM heterozygotes (n=49) and MM homozygotes (n=17). The Val-81-Met polymorphism showed significant linkage disequilibrium with the TCAT polymorphism (P<0.0001). No differences were seen between the subgroups with respect to age, BMI and blood pressure. Compared with the VV and VM genotypes, subjects with the MM genotype showed significantly (all P<0.05) increased LV cavity diameter (VV: 52.8+/-3.9 mm, VM: 52.9+/-3.6 mm, MM: 56.1+/-3.2 mm), global LV mass (VV: 159+/-31 g, VM: 165+/-36 g, MM: 187+/-30 g) and LV mass index (VV: 81+/-14 g/m2, VM: 84+/-17 g/m2, MM: 93+/-12 g/m2). No differences were seen between the subgroups in parameters of LV function. In addition, plasma epinephrine and norepinephrine levels were comparable in the three subgroups. The results suggest an important association between the MM genotype of Val-81-Met TH gene polymorphism and increased LV cavity dimension and mass in a young normotensive male population, indicating an important role for genetic determination of the sympathetic system in LV growth.

Development of electronic form of the 1999 WHO/ISH hypertension guidelines.

An electronic form of 1999 ISH/WHO Guidelines for Management of Hypertension was developed. It concentrates on two main problems--assessment of a patient's cardiovascular risk and selecting drug treatment. Moreover, it can remind the risk of unfilled items both for a cardiovascular risk calculation and drug contraindications. It can be stored for comparison with other records from follow-up both in an electronic and paper form.

Enhanced care of hypertensive patients using the Internet.

BACKGROUND: Medical guidelines provide recommendations that ought to help physicians in medical decision-making under different circumstances. To develop electronic medical guidelines and to make them available for physicians using the Internet can further enhance the quality and efficiency of health care, especially with the simultaneous use of electronic health records. OBJECTIVES: This study aims to outline the needs for web-based systems that support the use of medical guidelines in practice; and to focus on the development of web-based electronic medical guidelines for treatment of arterial hypertension. METHODS: The importance of electronic health record in cardiology for data acquisition, data storage and data mining is considered. The anonymized database of approximately 1800 hypertensive patients has been created to compare medical practice with guidelines and discover features of diseases that can help with their management. Using this database we evaluated several web-based electronic guidelines systems. RESULTS: The 1999 WHO/ISH Guidelines for the Management of Hypertension were formalized and interpreted using the Guide-X methodology, using the Apollo system and web-based electronic guidelines. The web-based electronic guidelines were tested on the smaller anonymous recent data set of 840 hypertensive patients. CONCLUSIONS: An easy transfer of knowledge from medical guidelines to structured electronic guidelines opens new possibilities for easy reusability of medical knowledge by general practitioners and clinicians.