Surveillance of venous thromboembolisms (VTEs) is necessary for improving patient safety in acute... more Surveillance of venous thromboembolisms (VTEs) is necessary for improving patient safety in acute care hospitals, but current detection methods are inaccurate and inefficient. With the growing availability of clinical narratives in an electronic format, automated surveillance using natural language processing (NLP) techniques may represent a better method. We assessed the accuracy of using symbolic NLP for identifying the 2 clinical manifestations of VTE, deep vein thrombosis (DVT) and pulmonary embolism (PE), from narrative radiology reports. A random sample of 4000 narrative reports was selected among imaging studies that could diagnose DVT or PE, and that were performed between 2008 and 2012 in a university health network of 5 adult-care hospitals in Montreal (Canada). The reports were coded by clinical experts to identify positive and negative cases of DVT and PE, which served as the reference standard. Using data from the largest hospital (n=2788), 2 symbolic NLP classifiers were trained; one for DVT, the other for PE. The accuracy of these classifiers was tested on data from the other 4 hospitals (n=1212). On manual review, 663 DVT-positive and 272 PE-positive reports were identified. In the testing dataset, the DVT classifier achieved 94% sensitivity (95% CI, 88%-97%), 96% specificity (95% CI, 94%-97%), and 73% positive predictive value (95% CI, 65%-80%), whereas the PE classifier achieved 94% sensitivity (95% CI, 89%-97%), 96% specificity (95% CI, 95%-97%), and 80% positive predictive value (95% CI, 73%-85%). Symbolic NLP can accurately identify VTEs from narrative radiology reports. This method could facilitate VTE surveillance and the evaluation of preventive measures.This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivitives 3.0 License, where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially. http://creativecommons.org/licenses/by-nc-nd/3.0.
We examined the immune responses of patients with active pulmonary tuberculosis (TB) and their he... more We examined the immune responses of patients with active pulmonary tuberculosis (TB) and their healthy household contacts to short-term culture filtrate (ST-CF) of Mycobacterium tuberculosis or molecular mass fractions derived from it. Our goal was to identify fractions strongly recognized by donors and differences among the donor groups of possible relevance for vaccine development. The study population consisted of 65 human immunodeficiency virus-negative donors from the Hossana Regional Hospital, Hossana, Ethiopia. Peripheral blood leukocytes from the donors were stimulated with different antigens and immune responses were determined. Household contacts produced significantly higher levels of gamma interferon (IFN-gamma) than the TB patients in response to antigens present in ST-CF and the 10 narrow-molecular-mass fractions. A similar difference in leukocyte proliferative responses to the antigens between the two groups was also found. In general, while all fractions stimulated i...
Individually identifiable health information needs to be masked or deleted in the secondary use o... more Individually identifiable health information needs to be masked or deleted in the secondary use of data for research. We defined and identified rare visible disorders as part of de-identification of individually identifiable health information. The prevalence of these disorders was evaluated using an administrative database and conformed to our definition. Due emphasis should be given to rare visible disorders in the de-identification process of health information.
Off-label prescribing may lead to adverse drug events. Little is known about its prevalence and d... more Off-label prescribing may lead to adverse drug events. Little is known about its prevalence and determinants resulting from challenges in documenting treatment indication. We used the Medical Office of the XXI Century electronic health record network in Quebec, Canada, where documentation of treatment indication is mandatory. One hundred thirteen primary care physicians wrote 253 347 electronic prescriptions for 50 823 patients from January 2005 through December 2009. Each drug indication was classified as on-label or off-label according to the Health Canada drug database. We identified off-label uses lacking strong scientific evidence. Alternating logistic regression was used to estimate the association between off-label use and drug, patient, and physician characteristics. The prevalence of off-label use was 11.0%; of the off-label prescriptions, 79.0% lacked strong scientific evidence. Off-label use was highest for central nervous system drugs (26.3%), including anticonvulsants (66.6%), antipsychotics (43.8%), and antidepressants (33.4%). Drugs with 3 or 4 approved indications were associated with less off-label use compared with drugs with 1 or 2 approved indications (6.7% vs 15.7%; adjusted odds ratio [AOR], 0.44; 95% CI, 0.41-0.48). Drugs approved after 1995 were prescribed off-label less often than were drugs approved before 1981 (8.0% vs 17.0%; AOR, 0.46; 95% CI, 0.42-0.50). Patients with a Charlson Comorbidity Index of 1 or higher had lower off-label use than did patients with an index of 0 (9.6% vs 11.7%; AOR, 0.94; 95% CI, 0.91-0.97). Physicians with evidence-based orientation were less likely to prescribe off-label (AOR, 0.93; 95% CI, 0.88-0.99), a 7% reduction per 5 points in the evidence section of the Evidence-Practicality-Conformity Scale. Off-label prescribing is common and varies by drug, patient, and physician characteristics. Electronic prescribing should document treatment indication to monitor off-label use.
Journal of the American Medical Informatics Association : JAMIA, Jan 10, 2015
Computer-based decision support has been effective in providing alerts for preventive care. Our o... more Computer-based decision support has been effective in providing alerts for preventive care. Our objective was to determine whether a personalized asthma management computer-based decision support increases the quality of asthma management and reduces the rate of out-of-control episodes. A cluster-randomized trial was conducted in Quebec, Canada among 81 primary care physicians and 4447 of their asthmatic patients. Patients were followed from the first visit for 3-33 months. The physician control group used the Medical Office of the 21st century (MOXXI) system, an integrated electronic health record. A custom-developed asthma decision support system was integrated within MOXXI and was activated for physicians in the intervention group. At the first visit, 9.8% (intervention) to 12.9% (control) of patients had out-of-control asthma, which was defined as a patient having had an emergency room visit or hospitalization for respiratory-related problems and/or more than 250 doses of fast-a...
The incidence of gastric cancer has been increasing among elderly persons in Japan. This study ai... more The incidence of gastric cancer has been increasing among elderly persons in Japan. This study aimed to clarify risk factors for postoperative complications in oldest old patients with gastric cancer. One-hundred ninety patients more than 75 years old with gastric cancer underwent gastrectomy between 2000 and 2011. Patients were classified into two groups: group A included 29 patients who were 85 years or older (oldest old patients), and group B included 161 patients who were 75-84 years of age. Perioperative parameters associated with complications were compared in each group. The preoperative estimated glomerular filtration rate was significantly lower in group A (p = 0.03). The two groups significantly differed in performance status (p = 0.018). Patients in group A received a lesser extent of lymph node dissection and had fewer lymph nodes excised. As a result, the duration of the operation was significantly shorter in group A. There were no significant differences in the frequency or grade of total complications or mortality between the two groups. Operative hemorrhage (>300 ml) and Hiroshima POSSUM (predicted morbidity risk >40) were risk factors in both groups A and B; the risk factors of preoperative serum albumin level and prognostic nutritional index (PNI) were specific to group A. Adjustments to the extent of surgery among oldest old patients most likely reduces the incidence of postoperative complications in this group. Preoperative serum albumin level and PNI are significant predictors of postoperative complications in oldest old patients with gastric cancer.
Surveillance of venous thromboembolisms (VTEs) is necessary for improving patient safety in acute... more Surveillance of venous thromboembolisms (VTEs) is necessary for improving patient safety in acute care hospitals, but current detection methods are inaccurate and inefficient. With the growing availability of clinical narratives in an electronic format, automated surveillance using natural language processing (NLP) techniques may represent a better method. We assessed the accuracy of using symbolic NLP for identifying the 2 clinical manifestations of VTE, deep vein thrombosis (DVT) and pulmonary embolism (PE), from narrative radiology reports. A random sample of 4000 narrative reports was selected among imaging studies that could diagnose DVT or PE, and that were performed between 2008 and 2012 in a university health network of 5 adult-care hospitals in Montreal (Canada). The reports were coded by clinical experts to identify positive and negative cases of DVT and PE, which served as the reference standard. Using data from the largest hospital (n=2788), 2 symbolic NLP classifiers were trained; one for DVT, the other for PE. The accuracy of these classifiers was tested on data from the other 4 hospitals (n=1212). On manual review, 663 DVT-positive and 272 PE-positive reports were identified. In the testing dataset, the DVT classifier achieved 94% sensitivity (95% CI, 88%-97%), 96% specificity (95% CI, 94%-97%), and 73% positive predictive value (95% CI, 65%-80%), whereas the PE classifier achieved 94% sensitivity (95% CI, 89%-97%), 96% specificity (95% CI, 95%-97%), and 80% positive predictive value (95% CI, 73%-85%). Symbolic NLP can accurately identify VTEs from narrative radiology reports. This method could facilitate VTE surveillance and the evaluation of preventive measures.This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivitives 3.0 License, where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially. http://creativecommons.org/licenses/by-nc-nd/3.0.
We examined the immune responses of patients with active pulmonary tuberculosis (TB) and their he... more We examined the immune responses of patients with active pulmonary tuberculosis (TB) and their healthy household contacts to short-term culture filtrate (ST-CF) of Mycobacterium tuberculosis or molecular mass fractions derived from it. Our goal was to identify fractions strongly recognized by donors and differences among the donor groups of possible relevance for vaccine development. The study population consisted of 65 human immunodeficiency virus-negative donors from the Hossana Regional Hospital, Hossana, Ethiopia. Peripheral blood leukocytes from the donors were stimulated with different antigens and immune responses were determined. Household contacts produced significantly higher levels of gamma interferon (IFN-gamma) than the TB patients in response to antigens present in ST-CF and the 10 narrow-molecular-mass fractions. A similar difference in leukocyte proliferative responses to the antigens between the two groups was also found. In general, while all fractions stimulated i...
Individually identifiable health information needs to be masked or deleted in the secondary use o... more Individually identifiable health information needs to be masked or deleted in the secondary use of data for research. We defined and identified rare visible disorders as part of de-identification of individually identifiable health information. The prevalence of these disorders was evaluated using an administrative database and conformed to our definition. Due emphasis should be given to rare visible disorders in the de-identification process of health information.
Off-label prescribing may lead to adverse drug events. Little is known about its prevalence and d... more Off-label prescribing may lead to adverse drug events. Little is known about its prevalence and determinants resulting from challenges in documenting treatment indication. We used the Medical Office of the XXI Century electronic health record network in Quebec, Canada, where documentation of treatment indication is mandatory. One hundred thirteen primary care physicians wrote 253 347 electronic prescriptions for 50 823 patients from January 2005 through December 2009. Each drug indication was classified as on-label or off-label according to the Health Canada drug database. We identified off-label uses lacking strong scientific evidence. Alternating logistic regression was used to estimate the association between off-label use and drug, patient, and physician characteristics. The prevalence of off-label use was 11.0%; of the off-label prescriptions, 79.0% lacked strong scientific evidence. Off-label use was highest for central nervous system drugs (26.3%), including anticonvulsants (66.6%), antipsychotics (43.8%), and antidepressants (33.4%). Drugs with 3 or 4 approved indications were associated with less off-label use compared with drugs with 1 or 2 approved indications (6.7% vs 15.7%; adjusted odds ratio [AOR], 0.44; 95% CI, 0.41-0.48). Drugs approved after 1995 were prescribed off-label less often than were drugs approved before 1981 (8.0% vs 17.0%; AOR, 0.46; 95% CI, 0.42-0.50). Patients with a Charlson Comorbidity Index of 1 or higher had lower off-label use than did patients with an index of 0 (9.6% vs 11.7%; AOR, 0.94; 95% CI, 0.91-0.97). Physicians with evidence-based orientation were less likely to prescribe off-label (AOR, 0.93; 95% CI, 0.88-0.99), a 7% reduction per 5 points in the evidence section of the Evidence-Practicality-Conformity Scale. Off-label prescribing is common and varies by drug, patient, and physician characteristics. Electronic prescribing should document treatment indication to monitor off-label use.
Journal of the American Medical Informatics Association : JAMIA, Jan 10, 2015
Computer-based decision support has been effective in providing alerts for preventive care. Our o... more Computer-based decision support has been effective in providing alerts for preventive care. Our objective was to determine whether a personalized asthma management computer-based decision support increases the quality of asthma management and reduces the rate of out-of-control episodes. A cluster-randomized trial was conducted in Quebec, Canada among 81 primary care physicians and 4447 of their asthmatic patients. Patients were followed from the first visit for 3-33 months. The physician control group used the Medical Office of the 21st century (MOXXI) system, an integrated electronic health record. A custom-developed asthma decision support system was integrated within MOXXI and was activated for physicians in the intervention group. At the first visit, 9.8% (intervention) to 12.9% (control) of patients had out-of-control asthma, which was defined as a patient having had an emergency room visit or hospitalization for respiratory-related problems and/or more than 250 doses of fast-a...
The incidence of gastric cancer has been increasing among elderly persons in Japan. This study ai... more The incidence of gastric cancer has been increasing among elderly persons in Japan. This study aimed to clarify risk factors for postoperative complications in oldest old patients with gastric cancer. One-hundred ninety patients more than 75 years old with gastric cancer underwent gastrectomy between 2000 and 2011. Patients were classified into two groups: group A included 29 patients who were 85 years or older (oldest old patients), and group B included 161 patients who were 75-84 years of age. Perioperative parameters associated with complications were compared in each group. The preoperative estimated glomerular filtration rate was significantly lower in group A (p = 0.03). The two groups significantly differed in performance status (p = 0.018). Patients in group A received a lesser extent of lymph node dissection and had fewer lymph nodes excised. As a result, the duration of the operation was significantly shorter in group A. There were no significant differences in the frequency or grade of total complications or mortality between the two groups. Operative hemorrhage (>300 ml) and Hiroshima POSSUM (predicted morbidity risk >40) were risk factors in both groups A and B; the risk factors of preoperative serum albumin level and prognostic nutritional index (PNI) were specific to group A. Adjustments to the extent of surgery among oldest old patients most likely reduces the incidence of postoperative complications in this group. Preoperative serum albumin level and PNI are significant predictors of postoperative complications in oldest old patients with gastric cancer.
Uploads
Papers by Tewodros Eguale