Effects of personal identifier resynthesis on clinical text de-identification

Objective

De-identified medical records are critical to biomedical research. Text de-identification software exists, including “resynthesis” components that replace real identifiers with synthetic identifiers. The goal of this research is to evaluate the effectiveness and examine possible bias introduced by resynthesis on de-identification software.

Design

We evaluated the open-source MITRE Identification Scrubber Toolkit, which includes a resynthesis capability, with clinical text from Vanderbilt University Medical Center patient records. We investigated four record classes from over 500 patients'' files, including laboratory reports, medication orders, discharge summaries and clinical notes. We trained and tested the de-identification tool on real and resynthesized records.

Measurements

We measured performance in terms of precision, recall, F-measure and accuracy for the detection of protected health identifiers as designated by the HIPAA Safe Harbor Rule.

Results

The de-identification tool was trained and tested on a collection of real and resynthesized Vanderbilt records. Results for training and testing on the real records were 0.990 accuracy and 0.960 F-measure. The results improved when trained and tested on resynthesized records with 0.998 accuracy and 0.980 F-measure but deteriorated moderately when trained on real records and tested on resynthesized records with 0.989 accuracy 0.862 F-measure. Moreover, the results declined significantly when trained on resynthesized records and tested on real records with 0.942 accuracy and 0.728 F-measure.

Conclusion

The de-identification tool achieves high accuracy when training and test sets are homogeneous (ie, both real or resynthesized records). The resynthesis component regularizes the data to make them less “realistic,” resulting in loss of performance particularly when training on resynthesized data and testing on real data.     

Cost-effectiveness Analysis for Technology Acquisition

Background

In a developing country with limited resources, it is important to utilize the total cost visibility approach over the entire life-cycle of the technology and then analyse alternative options for acquiring technology.

Methods

The present study analysed cost-effectiveness of an “In-house” magnetic resonance imaging (MRI) scan facility of a large service hospital against outsourcing possibilities. Cost per unit scan was calculated by operating costing method and break-even volume was calculated. Then life-cycle cost analysis was performed to enable total cost visibility of the MRI scan in both “In-house” and “outsourcing of facility” configuration. Finally, cost-effectiveness analysis was performed to identify the more acceptable decision option.

Result

Total cost for performing unit MRI scan was found to be Rs 3,875 for scans without contrast and Rs 4,129 with contrast. On life-cycle cost analysis, net present value (NPV) of the “In-house” configuration was found to be Rs-(4,09,06,265) while that of “outsourcing of facility” configuration was Rs-(5,70,23,315). Subsequently, cost-effectiveness analysis across eight Figures of Merit showed the “In-house” facility to be the more acceptable option for the system.

Conclusion

Every decision for acquiring high-end technology must be subjected to life-cycle cost analysis.Key Words: Technology assessment, Cost benefit analysis, Cost-effectiveness analysis     

Study of Behaviour Problems in a Paediatric Outpatient Department

Background

Behaviour problems in children still needs precise definition, explicit criterion and assessment on multiple paradigms.

Methods

Fifty children of the age group 6-14 years, from paediatric outpatient department, selected after randomisation were assessed for behaviour problems with the child behaviour checklist. The data collected was analysed using appropriate statistical tests.

Results

40% children were above cutoff score. Mean child behaviour check list (CBCL) score was 40.6. Total of 72% children were from armed forces background of whom 9% were siblings of officers. 30.6% children from the armed forces background were above the cutoff score. There was no significant difference in the behaviour problems between different age groups and sex. There was no significant difference in behaviour problems between children of officers, other ranks or various income groups. Female children had behaviour problems like “too concerned with neatness or cleanliness”, “feels has to be perfect” and “argues a lot” where as male children had behaviour problems like “Does not feel guilty after misbehaving”, “argues a lot” and “restless”.

Conclusion

Behaviour problems in the subjects were externalizing ones. No specific trend was found in children of defence personnel vis-a-vis children of civilian population.Key Words: Child behaviour problems, Child behaviour checklist     

“Not all my friends need to know”: a qualitative study of teenage patients,privacy, and social media

Background

The literature describes teenagers as active users of social media, who seem to care about privacy, but who also reveal a considerable amount of personal information. There have been no studies of how they manage personal health information on social media.

Objective

To understand how chronically ill teenage patients manage their privacy on social media sites.

Design

A qualitative study based on a content analysis of semistructured interviews with 20 hospital patients (12–18 years).

Results

Most teenage patients do not disclose their personal health information on social media, even though the study found a pervasive use of Facebook. Facebook is a place to be a “regular”, rather than a sick teenager. It is a place where teenage patients stay up-to-date about their social life—it is not seen as a place to discuss their diagnosis and treatment. The majority of teenage patients don''t use social media to come into contact with others with similar conditions and they don''t use the internet to find health information about their diagnosis.

Conclusions

Social media play an important role in the social life of teenage patients. They enable young patients to be “regular” teenagers. Teenage patients'' online privacy behavior is an expression of their need for self-definition and self-protection.     

Development and Application of a Framework for Maintenance of Medical Terminological Systems

Objective

Terminological Systems (TSs) need to be maintained in order to sustain their utility. This paper describes a study aiming at the standardization of the maintenance processes of medical TSs by capturing the criteria for the management of the maintenance processes into a framework. Furthermore, this paper describes application of the framework, which sheds light on the current practice of TS maintenance.

Design

Observational study.

Measurements

By means of a literature study, criteria for the maintenance of TSs were obtained and categorized into a framework. The current practice of TS maintenance was explored by a survey among organizations that maintain a TS. Results were stratified by the size of the TS being maintained.

Results

From Sixty-three relevant articles, criteria for the maintenance processes of TSs were extracted and organized into four components. The primary component “Execution” concerns the core activities of the maintenance process. The other three components “Process management,” “Change specifications,” and “Editing tools” support the core activities of the component “Execution.”The survey had a response rate of 40% (37 of 93). The answers reflect the large variation in the number of criteria that are satisfied for the participating organizations. Overall, maintenance of larger TSs seems to satisfy more criteria.

Conclusions

The framework is an important step towards standardization of the maintenance of medical TSs and can be used to eliminate shortcomings in this process. Surveying the current practice showed that there is ample room to improve the maintenance processes of medical TSs, especially for the smaller TSs.     

A pilot study to evaluate the diagnostic accuracy of local terminology for malaria screening among children in rural Malawi

Background

Malaria is a serious health problem in Malawi. It is responsible for 43% of all out patient visits and 19% of all deaths occurring to children under five years of age. Rapid diagnosis and appropriate treatment can avert most malaria deaths. However this is not always possible in resource limited settings where functioning laboratories are almost nonexistent.

Methods

This paper assesses the accuracy of local terminology in detecting parasitemia in children using blood smears as the reference standard.

Results

The study observes that there are local terms that can be used as an inexpensive, readily available and easily implementable malaria screening test in Malawian children in rural areas. These terms include “malungo” (official name for malaria), “kutentha thupi” (hot body), “kutsegula m''mimba” (official term for diarrhoea) and “kukhosomola” (coughing). The local terms “malungo” and “kutentha thupi” yielded better results.

Conclusion

Although the local terminology produced results that are less than optimal, the study concludes that the knowledge of sensitivity and specificity of local terminology can be used by local healthcare practitioners to identify children who could benefit from malaria confirmation testing and presumptive treatment. The study, however, cautions that these terms should be used as an entry point to malaria case management as they do not distinguish the severity of the malaria infection and all of them produced a sensitivity of less than 50%.     

“Different trend” in multiple primary lung cancer and intrapulmonary metastasis

Background

The distinguishing of intrapulmonary metastatic tumors from multiple primary lung cancers is difficult but of great importance for the therapeutic management and prognosis of these patients.

Methods

We used genomic DNA analyzed by six microsatellites (D7S1824, D15S822, D2S1363, D10S1239, D6S1056, and D22S689) with PCR to identify discordant allelic variation from 12 patients. There are five patients with multiple primary lung cancers and seven patients who were diagnosed with intrapulmonary metastases from 850 patients with primary lung cancer in our hospital. The experiments were approved by the West China Hospital Ethics committee (No. 2013 (33)) and all patients agreed to participate in the study and signed an informed consent form.

Results

In the group of metachronous lung tumor, three of five patients have different histological types and one of five patients have the same histological type which showed “contradictory trend”. The other one showed “unique trend”. In the second group (intrapulmonary metastasis lung tumor), one patient showed “contradictory trend” and the others showed “unique trend”.

Conclusions

“Different trends” are useful in discrimination of intrapulmonary metastasis lung cancer and multiple primary lung cancer even diagnosed with the histopathological evaluation.     

A novel approach to improve undergraduate surgical teaching

Background:

Undergraduate surgery is at an important crossroads. Many departments report significant difficulties delivering effective teaching. Our student feedback indicated a dated surgical curriculum lacking structure, quality and uniformity. We report on a new ”blended” approach employing a combination of professional DVDs, case based discussions, online material and traditional bedside teaching designed to provide structure, standardization, and equality of learning .

Methods:

Year 4 students who had undertaken the new course and year 5 students who had participated in the traditional teaching programme were compared. Students completed a 20 item questionnaire about their experiences of the surgical teaching programme.

Results:

One hundred and seventy-one year 4 (70%) and148 year 5 students (66%) responded. Domains relating to “Overall Satisfaction with the course”, “Approval of innovative teaching methods and interactivity” and “Satisfaction with the clarity of course information” showed improvements when comparing the new and old programmes. However bedside teaching was not rated as highly in the new programme (p<0.05).

Conclusion:

This blended approach has resulted in improved student understanding and engagement. The apparent compromise of bedside teaching may be a reflection of higher expectations. We believe that a similar blended approach has the potential to re-invigorate surgical teaching elsewhere.     

Perceptions of normal and abnormal ejaculatory latency times: an observational study in Ghanaian males and females

Background

This cross-sectional study aimed at quantifying the perceptions of Ghanaian men and women on how long they thought sex should last, from intromission until ejaculation.

Method

A random sample of 568 heterosexual men and women within the Kumasi Metropolis was surveyed from December 2009 to February 2010. The question of primary interest in the present study includes perceived intravaginal ejaculatory latency (IELT), in minutes, for four different conditions: “adequate,” “desirable,” “too short,” and “too long” IELT.

Results

The interquartile range for the respondent’s judgment of an “adequate” length for IELT was from 7.0 to 20.3 min; “desirable” from 10.0 to 25.0 min; “too short” from 2.0 to 5.0 min; “too long” from 10.5 to 60.0 min. However, the “actual” IELT (i.e. what the respondents are capable of doing) as found in this study was from 6 to 15 min. Ghanaian perceptions about ejaculatory latencies were in part consistent with data from Germany and contrary to data from the USA on ejaculatory latency and were not affected by age or educational level.

Conclusion

These results suggest that the average Ghanaian believes that intercourse that lasts 7.0–25.0 min is normal. Dissemination of the present finding to the public may modify their expectations for IELT which will lead to a realistic replica of sexuality and hence help prevent sexual disappointments and dysfunctions. It will also be beneficial to couples who are being treated for sexual problems by normalizing their expectations.     

Feature engineering combined with machine learning and rule-based methods for structured information extraction from narrative clinical discharge summaries

Objective

A system that translates narrative text in the medical domain into structured representation is in great demand. The system performs three sub-tasks: concept extraction, assertion classification, and relation identification.

Design

The overall system consists of five steps: (1) pre-processing sentences, (2) marking noun phrases (NPs) and adjective phrases (APs), (3) extracting concepts that use a dosage-unit dictionary to dynamically switch two models based on Conditional Random Fields (CRF), (4) classifying assertions based on voting of five classifiers, and (5) identifying relations using normalized sentences with a set of effective discriminating features.

Measurements

Macro-averaged and micro-averaged precision, recall and F-measure were used to evaluate results.

Results

The performance is competitive with the state-of-the-art systems with micro-averaged F-measure of 0.8489 for concept extraction, 0.9392 for assertion classification and 0.7326 for relation identification.

Conclusions

The system exploits an array of common features and achieves state-of-the-art performance. Prudent feature engineering sets the foundation of our systems. In concept extraction, we demonstrated that switching models, one of which is especially designed for telegraphic sentences, improved extraction of the treatment concept significantly. In assertion classification, a set of features derived from a rule-based classifier were proven to be effective for the classes such as conditional and possible. These classes would suffer from data scarcity in conventional machine-learning methods. In relation identification, we use two-staged architecture, the second of which applies pairwise classifiers to possible candidate classes. This architecture significantly improves performance.     

Intestinal Microbiota Metabolism and Atherosclerosis

Objective:

This review aimed to summarize the relationship between intestinal microbiota metabolism and cardiovascular disease (CVD) and to propose a novel CVD therapeutic target.

Data Sources:

This study was based on data obtained from PubMed and EMBASE up to June 30, 2015. Articles were selected using the following search terms: “Intestinal microbiota”, “trimethylamine N-oxide (TMAO)”, “trimethylamine (TMA)”, “cardiovascular”, and “atherosclerosis”.

Study Selection:

Studies were eligible if they present information on intestinal microbiota metabolism and atherosclerosis. Studies on TMA-containing nutrients were also included.

Results:

A new CVD risk factor, TMAO, was recently identified. It has been observed that several TMA-containing compounds may be catabolized by specific intestinal microbiota, resulting in TMA release. TMA is subsequently converted to TMAO in the liver. Several preliminary studies have linked TMAO to CVD, particularly atherosclerosis; however, the details of this relationship remain unclear.

Conclusions:

Intestinal microbiota metabolism is associated with atherosclerosis and may represent a promising therapeutic target with respect to CVD management.     

Insulin Neuroprotection and the Mechanisms

Objective:

To analyze the mechanism of neuroprotection of insulin and which blood glucose range was benefit for insulin exerting neuroprotective action.

Data Sources:

The study is based on the data from PubMed.

Study Selection:

Articles were selected with the search terms “insulin”, “blood glucose”, “neuroprotection”, “brain”, “glycogen”, “cerebral ischemia”, “neuronal necrosis”, “glutamate”, “γ-aminobutyric acid”.

Results:

Insulin has neuroprotection. The mechanisms include the regulation of neurotransmitter, promoting glycogen synthesis, and inhibition of neuronal necrosis and apoptosis. Insulin could play its role in neuroprotection by avoiding hypoglycemia and hyperglycemia.

Conclusions:

Intermittent and long-term infusion insulin may be a benefit for patients with ischemic brain damage at blood glucose 6–9 mmol/L.     

Automatic extraction of medication information from medical discharge summaries

Objective

This article describes a system developed for the 2009 i2b2 Medication Extraction Challenge. The purpose of this challenge is to extract medication information from hospital discharge summaries.

Design

The system explored several linguistic natural language processing techniques (eg, term-based and token-based rule matching) to identify medication-related information in the narrative text. A number of lexical resources was constructed to profile lexical or morphological features for different categories of medication constituents.

Measurements

Performance was evaluated in terms of the micro-averaged F-measure at the horizontal system level.

Results

The automated system performed well, and achieved an F-micro of 80% for the term-level results and 81% for the token-level results, placing it sixth in exact matches and fourth in inexact matches in the i2b2 competition.

Conclusion

The overall results show that this relatively simple rule-based approach is capable of tackling multiple entity identification tasks such as medication extraction under situations in which few training documents are annotated for machine learning approaches, and the entity information can be characterized with a set of feature tokens.     

Estimating Consumer Familiarity with Health Terminology: A Context-based Approach

Objectives

Effective health communication is often hindered by a “vocabulary gap” between language familiar to consumers and jargon used in medical practice and research. To present health information to consumers in a comprehensible fashion, we need to develop a mechanism to quantify health terms as being more likely or less likely to be understood by typical members of the lay public. Prior research has used approaches including syllable count, easy word list, and frequency count, all of which have significant limitations.

Design

In this article, we present a new method that predicts consumer familiarity using contextual information. The method was applied to a large query log data set and validated using results from two previously conducted consumer surveys.

Measurements

We measured the correlation between the survey result and the context-based prediction, syllable count, frequency count, and log normalized frequency count.

Results

The correlation coefficient between the context-based prediction and the survey result was 0.773 (p < 0.001), which was higher than the correlation coefficients between the survey result and the syllable count, frequency count, and log normalized frequency count (p ≤ 0.012).

Conclusions

The context-based approach provides a good alternative to the existing term familiarity assessment methods.     

Malaria on the Move : Ecological Considerations for the Armed Forces

Background

Armed forces personnel deployed in the North Eastern states of India are vulnerable to falciparum malaria. This vulnerability increases during mobilization of troops.

Methods

Epidemiological case sheet was used for recording individual movement, clinical features and laboratory investigations of each case of malaria. Immunochromotography test (ICT) or Paracheck Pf was used as a rapid test for falciparum malaria at the regimental aid post (RAP). Subsequently, a case control approach was used to ascertain whether the cases of malaria differed significantly from healthy controls in observing antimalaria measures such as the use of mosquito nets, repellants and chemoprophylaxis.

Result

Nineteen out of 623 soldiers suffered from falciparum malaria during a short period of ten days during operational mobilization. Use of mosquito nets and repellants was significantly less among the cases as compared to healthy controls. There was no significant difference among the two groups regarding compliance with chemoprophylaxis.

Conclusion

A paradigm of “malaria on the move” or “operational malaria” has been proposed.Key Words: Malaria, Armed Forces, Mobilization     

Automated concept-level information extraction to reduce the need for custom software and rules development

Objective

Despite at least 40 years of promising empirical performance, very few clinical natural language processing (NLP) or information extraction systems currently contribute to medical science or care. The authors address this gap by reducing the need for custom software and rules development with a graphical user interface-driven, highly generalizable approach to concept-level retrieval.

Materials and methods

A ‘learn by example’ approach combines features derived from open-source NLP pipelines with open-source machine learning classifiers to automatically and iteratively evaluate top-performing configurations. The Fourth i2b2/VA Shared Task Challenge''s concept extraction task provided the data sets and metrics used to evaluate performance.

Results

Top F-measure scores for each of the tasks were medical problems (0.83), treatments (0.82), and tests (0.83). Recall lagged precision in all experiments. Precision was near or above 0.90 in all tasks.

Discussion

With no customization for the tasks and less than 5 min of end-user time to configure and launch each experiment, the average F-measure was 0.83, one point behind the mean F-measure of the 22 entrants in the competition. Strong precision scores indicate the potential of applying the approach for more specific clinical information extraction tasks. There was not one best configuration, supporting an iterative approach to model creation.

Conclusion

Acceptable levels of performance can be achieved using fully automated and generalizable approaches to concept-level information extraction. The described implementation and related documentation is available for download.     

MITRE system for clinical assertion status classification

Objective

To describe a system for determining the assertion status of medical problems mentioned in clinical reports, which was entered in the 2010 i2b2/VA community evaluation ‘Challenges in natural language processing for clinical data’ for the task of classifying assertions associated with problem concepts extracted from patient records.

Materials and methods

A combination of machine learning (conditional random field and maximum entropy) and rule-based (pattern matching) techniques was used to detect negation, speculation, and hypothetical and conditional information, as well as information associated with persons other than the patient.

Results

The best submission obtained an overall micro-averaged F-score of 0.9343.

Conclusions

Using semantic attributes of concepts and information about document structure as features for statistical classification of assertions is a good way to leverage rule-based and statistical techniques. In this task, the choice of features may be more important than the choice of classifier algorithm.     

External Validation of EPICON: A Grouping System for Estimating Morbidity Rates Using Electronic Medical Records

Objective

To externally validate EPICON, a computerized system for grouping diagnoses from EMRs in general practice into episodes of care. These episodes can be used for estimating morbidity rates.

Design

Comparative observational study.

Measurements

Morbidity rates from an independent dataset, based on episode-oriented EMRs, were used as the gold standard. The EMRs in this dataset contained diagnoses which were manually grouped by GPs. The authors ungrouped these diagnoses and regrouped them automatically into episodes using EPICON. The authors then used these episodes to estimate morbidity rates that were compared to the gold standard. The differences between the two sets of morbidity rates were calculated and the authors analyzed large as well as structural differences to establish possible causes.

Results

In general, the morbidity rates based on EPICON deviate only slightly from the gold standard. Out of 675 diagnoses, 36 (5%) were considered to be deviating diagnoses. The deviating diagnoses showed differences for two main reasons: “differences in rules between the two methods of episode construction” and “inadequate performance of EPICON.”

Conclusion

The EPICON system performs well for the large majority of the morbidity rates. We can therefore conclude that EPICON is useful for grouping episodes to estimate morbidity rates using EMRs from general practices. Morbidity rates of diseases with a broad range of symptoms should, however, be interpreted cautiously.