Rheumatoid arthritis (RA) is a chronic, destructive, inflammatory arthritis that places significant burden on the individual and society. This thesis represents the most comprehensive effort to date to determine the accuracy of administrative data for detecting RA patients; and describes the development and validation of an administrative data algorithm to establish a province-wide RA database. Beginning with a systematic review to guide the conduct of this research, two independent, multicentre, retrospective chart abstraction studies were performed amongst two random samples of patients from rheumatology and primary care family physician practices, respectively. While a diagnosis by a rheumatologist remains the gold standard for establishing a RA diagnosis, the high prevalence of RA in rheumatology clinics can falsely elevate positive predictive values. It was therefore important we also perform a validation study in a primary care setting where prevalence of RA would more closely approximate that observed in the general population. The algorithm of [1 hospitalization RA code] OR [3 physician RA diagnosis codes (claims) with ≥1 by a specialist in a 2 year period)] demonstrated a high degree of accuracy in terms of minimizing both the number of false positives (moderately good PPV; 78%) and true negatives (high specificity: 100%). Moreover, this algorithm has excellent sensitivity at capturing contemporary RA patients under active rheumatology care (>96%). Application of this algorithm to Ontario health administrative data to establish the Ontario RA administrative Database (ORAD) identified 97,499 Ontarians with RA as of 2010, yielding a cumulative prevalence of (0.9%). Age/sex-standardized RA prevalence has doubled from 473 per 100,000 in 1996 to 784 per 100,000 in 2010, with approximately 50 new cases of RA emerging per 100,000 Ontarians each year. Our findings will inform future population-based research and will serve to improve arthritis surveillance activities across Canada and abroad.

Rheumatoid arthritis (RA) is a chronic, destructive, inflammatory arthritis that places significant burden on the individual and society. This thesis represents the most comprehensive effort to date to determine the accuracy of administrative data for detecting RA patients; and describes the development and validation of an administrative data algorithm to establish a province-wide RA database. Beginning with a systematic review to guide the conduct of this research, two independent, multicentre, retrospective chart abstraction studies were performed amongst two random samples of patients from rheumatology and primary care family physician practices, respectively. While a diagnosis by a rheumatologist remains the gold standard for establishing a RA diagnosis, the high prevalence of RA in rheumatology clinics can falsely elevate positive predictive values. It was therefore important we also perform a validation study in a primary care setting where prevalence of RA would more closely approximate that observed in the general population. The algorithm of [1 hospitalization RA code] OR [3 physician RA diagnosis codes (claims) with ≥1 by a specialist in a 2 year period)] demonstrated a high degree of accuracy in terms of minimizing both the number of false positives (moderately good PPV; 78%) and true negatives (high specificity: 100%). Moreover, this algorithm has excellent sensitivity at capturing contemporary RA patients under active rheumatology care (>96%). Application of this algorithm to Ontario health administrative data to establish the Ontario RA administrative Database (ORAD) identified 97,499 Ontarians with RA as of 2010, yielding a cumulative prevalence of (0.9%). Age/sex-standardized RA prevalence has doubled from 473 per 100,000 in 1996 to 784 per 100,000 in 2010, with approximately 50 new cases of RA emerging per 100,000 Ontarians each year. Our findings will inform future population-based research and will serve to improve arthritis surveillance activities across Canada and abroad.

Background: Primary subarachnoid hemorrhage (1°SAH) is an important disease that causes significant morbidity and mortality. The sparse Canadian epidemiologic literature on 1° SAH is outdated and relies on diagnostic coding for case ascertainment which misses true cases and incorrectly labels non-cases. Objectives: Primary objective was to identify all patients with 1° SAH presenting to the Ottawa Hospital (TOH) between July 1, 2002 and June 30, 2011 by deriving and validating a search algorithm using an enriched administrative database. Secondary objectives included: 1) determine incidence and case-fatality rates (CFR) of 1° SAH at TOH; and 3) derive and validate a method to identify 1° SAH using routinely collected administrative data. Methods: A cohort of 1° SAH patients were identified with a case-defining algorithm that was derived and validated using a combination of cerebrospinal fluid analysis results and text-search algorithms of both cranial imaging and post-mortem reports. The incidence of 1° SAH was calculated using the total number of hospital encounters over the same time period. CFR was calculated by linking to vital statistic data of hospitalized patients at discharge. An optimal1° SAH prediction model was derived and validated using binomial recursive partitioning built with independent variables obtained from routinely collected administrative data. Results: Using the case-defining algorithm, 831 patients were identified with a 1° SAH over the study period. Hospital incidence of 1° SAH was 17.2 events per 10,000 inpatient encounters (or 0.17% of encounters) with a case-fatality rate of 18.1%. A validated SAH prediction model based on administrative data using a recursive partitioning model had a sensitivity of 96.5% (95% CI 93.9-98.0), a specificity of 99.8% (95%CI 99.6-99.9), and a +LR of 483 (95% CI 254-879). This results in a post-test probability of disease of 45%. Conclusion: We identified almost all cases of 1° SAH at our hospital using an enriched administrative data. Accurately identifying such patients with routinely collected health administrative data is possible, providing important opportunities to examine and study this patient population. Further studies, involving multiple centres are needed to reproduce these results.

This study determined the sensitivity and specificity of questionnaires and administrative databases compared to a clinical reference standard for asthma. 208 schoolchildren from a population-based sample participated. They underwent a physician assessment, spirometry, methacholine challenge, exhaled nitric oxide and skin testing. Data was linked to the Ontario Asthma Surveillance Information System. “Questionnaire diagnosis” was an affirmative response to physician-diagnosed asthma. “Database diagnosis” was 2 outpatient visits or 1 hospitalization within 2 years. “Clinical diagnosis” required a physician assessment and objective findings of asthma. “Questionnaire diagnosis” of asthma was specific (92.1%) but not sensitive (75.3%) compared to the “clinical diagnosis”. “Database diagnosis” was sensitive (87.5%) but not specific (64.8%). Both sources had an excellent negative predictive value (97-98%) but poor positive predictive value (24-55%). Epidemiologic methods accurately identify those without asthma but are poor at identifying those with asthma, leading to an overestimation of asthma prevalence and dilution of risk estimates.

This study determined the sensitivity and specificity of questionnaires and administrative databases compared to a clinical reference standard for asthma. 208 schoolchildren from a population-based sample participated. They underwent a physician assessment, spirometry, methacholine challenge, exhaled nitric oxide and skin testing. Data was linked to the Ontario Asthma Surveillance Information System. “Questionnaire diagnosis” was an affirmative response to physician-diagnosed asthma. “Database diagnosis” was 2 outpatient visits or 1 hospitalization within 2 years. “Clinical diagnosis” required a physician assessment and objective findings of asthma. “Questionnaire diagnosis” of asthma was specific (92.1%) but not sensitive (75.3%) compared to the “clinical diagnosis”. “Database diagnosis” was sensitive (87.5%) but not specific (64.8%). Both sources had an excellent negative predictive value (97-98%) but poor positive predictive value (24-55%). Epidemiologic methods accurately identify those without asthma but are poor at identifying those with asthma, leading to an overestimation of asthma prevalence and dilution of risk estimates.

京都大学 / 0048 / 新制・課程博士 / 博士(社会健康医学) / 甲第22383号 / 社医博第105号 / 新制||社||医11(附属図書館) / 京都大学大学院医学研究科社会健康医学系専攻 / (主査)教授 古川 壽亮, 教授 川上 浩司, 教授 滝田 順子 / 学位規則第4条第1項該当 / Doctor of Public Health / Kyoto University / DFAM

Background: Pre-eclampsia (PET) is one of the major causes of maternal and neonatal morbidity and mortality1. Misclassification of PET can lead to biased or erroneous results in epidemiologic studies resulting in false conclusions. Objectives: The objectives of this thesis are to determine the validity of PET diagnosis in pregnant women in administrative database using the ICD-10-CA codes, to explore the nature of misclassification, and to estimate whether misclassification of PET diagnosis in administrative database may result in biased conclusions. Methods: Pregnant women who participated in the Ottawa and Kingston (OaK) Birth Cohort study and delivered in the Ottawa Hospital were included in the study. All cases with hypertensive disorder of pregnancy in the study population were adjudicated to confirm diagnosis of PET. This adjudicated dataset was used as a reference standard. The PET incidence in hospital discharge database was compared with PET incidence calculated from the reference standard database. Results: 2887 of the requested charts were available for review. The PET incidence was much lower in administrative database (1.47%) than in the OaK Birth Cohort Study (3.6%). The results of the study demonstrated that hospital discharge database via ICD-10-CA was not very sensitive to determine incidence of PET since sensitivity of ICD-10-CA diagnostic codes for PET was low (35.92% with 95% Confidence Intervals (CI): 26.7; 45.9) but specificity, PPV, and NPV were high. The majority of misclassified cases belonged to the category (according to the proposed classification) “PET pregnancies coded with incorrect ICD-10-CA code” (78.88%) followed by the category “Pregnancies affected by PET coded as normal” (14.08%). Conclusion: Using hospital discharge database and ICD-10-CA coding to determine incidence of PET in certain settings may yield low sensitivity. Researchers should validate the results when using the hospital discharge database for PET research to ensure that the findings based on analyses of such data demonstrate what they claimed to demonstrate.

Background: Pre-eclampsia (PET) is one of the major causes of maternal and neonatal morbidity and mortality1. Misclassification of PET can lead to biased or erroneous results in epidemiologic studies resulting in false conclusions. Objectives: The objectives of this thesis are to determine the validity of PET diagnosis in pregnant women in administrative database using the ICD-10-CA codes, to explore the nature of misclassification, and to estimate whether misclassification of PET diagnosis in administrative database may result in biased conclusions. Methods: Pregnant women who participated in the Ottawa and Kingston (OaK) Birth Cohort study and delivered in the Ottawa Hospital were included in the study. All cases with hypertensive disorder of pregnancy in the study population were adjudicated to confirm diagnosis of PET. This adjudicated dataset was used as a reference standard. The PET incidence in hospital discharge database was compared with PET incidence calculated from the reference standard database. Results: 2887 of the requested charts were available for review. The PET incidence was much lower in administrative database (1.47%) than in the OaK Birth Cohort Study (3.6%). The results of the study demonstrated that hospital discharge database via ICD-10-CA was not very sensitive to determine incidence of PET since sensitivity of ICD-10-CA diagnostic codes for PET was low (35.92% with 95% Confidence Intervals (CI): 26.7; 45.9) but specificity, PPV, and NPV were high. The majority of misclassified cases belonged to the category (according to the proposed classification) “PET pregnancies coded with incorrect ICD-10-CA code” (78.88%) followed by the category “Pregnancies affected by PET coded as normal” (14.08%). Conclusion: Using hospital discharge database and ICD-10-CA coding to determine incidence of PET in certain settings may yield low sensitivity. Researchers should validate the results when using the hospital discharge database for PET research to ensure that the findings based on analyses of such data demonstrate what they claimed to demonstrate.

Background: Pre-eclampsia (PET) is one of the major causes of maternal and neonatal morbidity and mortality1. Misclassification of PET can lead to biased or erroneous results in epidemiologic studies resulting in false conclusions. Objectives: The objectives of this thesis are to determine the validity of PET diagnosis in pregnant women in administrative database using the ICD-10-CA codes, to explore the nature of misclassification, and to estimate whether misclassification of PET diagnosis in administrative database may result in biased conclusions. Methods: Pregnant women who participated in the Ottawa and Kingston (OaK) Birth Cohort study and delivered in the Ottawa Hospital were included in the study. All cases with hypertensive disorder of pregnancy in the study population were adjudicated to confirm diagnosis of PET. This adjudicated dataset was used as a reference standard. The PET incidence in hospital discharge database was compared with PET incidence calculated from the reference standard database. Results: 2887 of the requested charts were available for review. The PET incidence was much lower in administrative database (1.47%) than in the OaK Birth Cohort Study (3.6%). The results of the study demonstrated that hospital discharge database via ICD-10-CA was not very sensitive to determine incidence of PET since sensitivity of ICD-10-CA diagnostic codes for PET was low (35.92% with 95% Confidence Intervals (CI): 26.7; 45.9) but specificity, PPV, and NPV were high. The majority of misclassified cases belonged to the category (according to the proposed classification) “PET pregnancies coded with incorrect ICD-10-CA code” (78.88%) followed by the category “Pregnancies affected by PET coded as normal” (14.08%). Conclusion: Using hospital discharge database and ICD-10-CA coding to determine incidence of PET in certain settings may yield low sensitivity. Researchers should validate the results when using the hospital discharge database for PET research to ensure that the findings based on analyses of such data demonstrate what they claimed to demonstrate.

L’endocardite infectieuse (EI) est une maladie rare, de diagnostic difficile et de pronostic réservé. Staphylococcus aureus (SA) et les streptocoques oraux en sont les principaux microorganismes responsables. L’évaluation du risque de survenue de l’EI et l’amélioration des connaissances justifiant la stratégie de prévention nécessitent la mise en place de grandes cohortes cliniques et l’utilisation de bases médico-administratives. Chez les 2 008 patients (pts) présentant une bactériémie à SA de la cohorte multicentrique nationale VIRSTA, nous avons développé et validé un score prédictif d’EI comportant les caractéristiques initiales des pts et celles initiales et évolutives de la bactériémie. Les pts dont le score était ≤ 2 avaient un très faible risque d’EI (1% ; valeur prédictive négative [IC95%] = 99% [98;99]) comparés à ceux dont le score était ≥ 3, à risque d’EI élevé (17%) pour lesquels une échocardiographie devrait être effectuée. Utilisant la base médico-administrative du SNIIRAM, nous avons évalué la relation entre la pratique de gestes buccodentaires invasifs (GBDI) et la survenue d’EI à streptocoques oraux à partir d’une cohorte de 138 876 porteurs de prothèses valvulaires cardiaques ainsi que d’un plan expérimental de type case-crossover incluant 648 EI à streptocoques oraux. L’incidence d’EI à streptocoques oraux [IC95%] était de 93,7 pour 100 000 PA [82,4;104,9] sans augmentation significative du risque dans les 3 mois suivant un GBDI (RR= 1,25 [0,82;1,82]). Dans l’analyse case-crossover, la fréquence d’exposition à un GBDI dans les 3 mois précédent l’EI était faible mais plus élevée que lors de périodes contrôles antérieures (5,1% vs 3,2% ; OR : 1,66 [1,05;2,63]). Les GBDI pourraient contribuer au développement des EI à streptocoques oraux dans la population de pts porteurs de prothèses valvulaires cardiaques.La qualité des données de VIRSTA associée à la puissance du SNIIRAM ont permis l’identification des pts à risque d’EI à SA et la clarification de la contribution des GBDI dans les EI à streptocoques oraux. / Infective endocarditis (IE) is a rare disease, difficult to diagnose, with high morbidity and mortality rates. Main involved microorganisms are Staphylococcus aureus and oral streptococci. Clinical research to improve IE risk assessment and IE prevention strategy requires the establishment of large clinical cohort studies and the use of medico-administrative databases. Using data from the multicenter French prospective VIRSTA cohort study on 2 008 adult patients (pts) with Staphylococcus aureus bloodstream infection (SAB), we have developed and validated an IE prediction score taking into account pts’ background and initial SAB characteristics. Pts with a score ≤ 2 had a very low risk of IE (1%, negative predictive value [95% CI] = 99% [98;99]) compared to those with a score ≥ 3, at higher risk of IE (17%) for whom an echocardiography is needed. Using the medico-administrative SNIIRAM database, we assessed the relation between invasive dental procedures (IDP) and oral streptococcal IE in a population-based cohort study of 138 876 pts with prosthetic heart valves and a case-crossover study including 648 pts with oral streptococcal IE. Incidence rate of oral streptococcal IE [95% CI] was 93.7 per 100 000 PA [82.4;104.9] without significant increase within the 3 months following IDP (RR = 1.25 [0.82;1.82]). In the case-crossover analysis, exposure to IDP was more frequent in the 3 months preceding IE than during previous control periods (5.1% vs. 3.2%, OR: 1.66 [1.05;2.63]). IDP may contribute to the development of oral streptococcal IE in pts with prosthetic heart valves.The quality of data from VIRSTA study combined with the power of SNIIRAM database made possible the identification of IE at-risk SAB pts and the evaluation of the IDP contribution in oral streptococcal IE.