Use of laboratory and administrative data to understand the potential impact of human parainfluenza virus 4 on cases of bronchiolitis, croup, and pneumonia in Alberta, Canada.

BACKGROUND: Human Parainfluenza Virus (hPIV) causes severe respiratory illness in infants and adults. Our study describes the association of hPIV1-4 with bronchiolitis, croup, and pneumonia using retrospective laboratory, administrative and public health data. Due to issues including the historic lack of hPIV4 in some commercial respiratory virus panels, the description of the impact of hPIV4 on croup, bronchiolitis, and pneumonia at population levels has often been limited. This study will use routine clinical laboratory data, and administrative data to provide a preliminary description of the impact of hPIV4 on these diseases in our population. METHODS: A three year cohort of patients positive for hPIV was linked with data from physician visits and hospital admissions to define cases and hospitalization status. International Classification of Disease (ICD-9) codes were used to determine if cases had croup, bronchiolitis, and pneumonia. We also looked at differences in hospitalization status, age and gender among hPIV1-4. All statistical analysis was done using SPSS (Version 19.0.0, IBM Corp© 2010) and Graphpad Prism V6 (GraphPad Software, Inc., 2012). RESULTS: Only hPIV1 and hPIV4 specimens had positivity rates greater than 5 % of all specimens sent for respiratory virus panel testing. hPIV1 exhibited a biennial pattern while the pattern for hPIV3 was less interpretable due to lower positivity rates. Circulation patterns for hPIV2 and hPIV4 were not assessed due to the low positivity rates of theses specimens. From 2010 to 2013, there were 2300 hPIV cases with hPIV3 (46 %) being the most common, followed by hPIV1 (27 %), hPIV4 (16 %) and hPIV2 (11 %). The median age was 2 years for all hPIV types. Males were slightly greater than females for hPIV1 and hPIV2, with an equal distribution for hPIV3 and slightly more females than males for hPIV4. hPIV1 and hPIV2 had the highest or proportion of croup while hPIV3 and hPIV4 had the highest proportion of pneumonia. Within hPIV4 cases, distributions of diseases were; pneumonia (21 %, 95 % CI 17.1-25.7), bronchiolitis (18 %, 95 % CI 14.3-22.5), croup (2 %, 95 % CI 0.8-3.9), mixed illness of any of pneumonia, bronchiolitis or croup (4 %, 95 % CI 2.5-7.0) or other respiratory diseases (54 %, 95 % CI 49.1-59.6). CONCLUSIONS: We used laboratory and administrative data to undertake a descriptive analysis of the association of hPIV1-4 with croup, bronchiolitis and pneumonia. hPIV4 appears to be more associated more with bronchiolitis and pneumonia and less with croup in our population.

Molecular profiling of Escherichia coli O157:H7 and non-O157 strains isolated from humans and cattle in Alberta, Canada.

Virulence markers in Shiga toxin-producing Escherichia coli (STEC) and their association with diseases remain largely unknown. This study determines the importance of 44 genetic markers for STEC (O157 and non-O157) from human clinical cases and their correlation to disease outcome. STEC isolated from a cattle surveillance program were also included. The virulence genes tested were present in almost all O157:H7 isolates but highly variable in non-O157 STEC isolates. Patient age was a significant determinant of clinical outcome.

Factors associated with poor mental health status among homeless women with and without dependent children.

The purpose of this study was to estimate the prevalence of mental health problems among a representative sample of homeless women with and without dependent children and determine if the effects of risk factors for mental health are modified by the presence of dependent children. Homeless women (n = 522) were recruited in 2004-2005 from shelters and meal programs in Toronto, Canada. Linear and logistic regression was performed to identify factors associated with mental health status. Poor mental health was associated with low perceived access to social support, physical/sexual assault in the past 12 months, presence of a chronic health condition, and presence of a drug use problem in the past month. Efforts to improve mental health in this population will need to address the associated problems of victimization, substance abuse, and lack of social supports.

Impact of routine real-time PCR testing of imported malaria over 4 years of implementation in a clinical laboratory.

In clinical laboratories, diagnosis of imported malaria is commonly performed by microscopy. However, the volume of specimens is generally low and maintaining proficiency in reading blood smears, particularly at the species level, is challenging in this setting. To address this problem, the Provincial Laboratory for Public Health (ProvLab) in Alberta, Canada, implemented real-time PCR for routine confirmation of all smear-positive samples in the province. Here we report our experience over a 4-year period (2008 to 2012) with this new diagnostic algorithm. While detection of Plasmodium falciparum by microscopy alone was accurate, real-time PCR served as an important adjunct to microscopy for the identification of non-falciparum species. In 18% of cases, the result was reported as non-falciparum or the species could not be identified by microscopy alone, and in all cases, the species was resolved by real-time PCR. In another 4% of cases, the species was misidentified by microscopy. To enhance surveillance for malaria, we integrated our demographic, clinical, and laboratory data into a new system developed by the Canadian Network for Public Health Intelligence, called the Malaria System for Online Surveillance (SOS). Using this application, we characterized our patient populations and travel history to identify risk factors associated with malaria infection abroad.

Return on investment of the lymphoma diagnostic pathway implementation in Alberta, Canada.

The Lymphoma Diagnostic Pathway (LDP) was developed based upon clinical best practice guidelines and implemented in large urban centers where lymphoma treatment is provided in Alberta, Canada. A return-on-investment analysis of the implementation of this care pathway was conducted to inform future sustainability and expansion. A cohort design with propensity score matching and difference-in-difference estimation methods were used comparing both cost and return (reduced health service utilization) between patients who were diagnosed within the LDP and those who were diagnosed outside the LDP. LDP resulted in $1800 avoided HSU costs per patient. The LDP has been found to be cost-saving with an ROI of 5.3 (ranging from 3.95-8.97) - for every $1 invested, LDP resulted in a $5.30 return for the health system due to capacity improvements in ED, inpatient, outpatient, and a reduction in GP service utilization. Further study of implementation including patient/provider satisfaction and uptake is recommended.

Modeling continuous diagnostic test data using approximate Dirichlet process distributions.

There is now a large literature on the analysis of diagnostic test data. In the absence of a gold standard test, latent class analysis is most often used to estimate the prevalence of the condition of interest and the properties of the diagnostic tests. When test results are measured on a continuous scale, both parametric and nonparametric models have been proposed. Parametric methods such as the commonly used bi-normal model may not fit the data well; nonparametric methods developed to date have been relatively complex to apply in practice, and their properties have not been carefully evaluated in the diagnostic testing context. In this paper, we propose a simple yet flexible Bayesian nonparametric model which approximates a Dirichlet process for continuous data. We compare results from the nonparametric model with those from the bi-normal model via simulations, investigating both how much is lost in using a nonparametric model when the bi-normal model is correct and how much can be gained in using a nonparametric model when normality does not hold. We also carefully investigate the trade-offs that occur between flexibility and identifiability of the model as different Dirichlet process prior distributions are used. Motivated by an application to tuberculosis clustering, we extend our nonparametric model to accommodate two additional dichotomous tests and proceed to analyze these data using both the continuous test alone as well as all three tests together.

Standardized observation of neighbourhood disorder: does it work in Canada?

BACKGROUND: There is a growing body of evidence that where you live is important to your health. Despite numerous previous studies investigating the relationship between neighbourhood deprivation (and structure) and residents' health, the precise nature of this relationship remains unclear. Relatively few investigations have relied on direct observation of neighbourhoods, while those that have were developed primarily in US settings. Evaluation of the transferability of such tools to other contexts is an important first step before applying such instruments to the investigation of health and well-being. This study evaluated the performance of a systematic social observational (SSO) tool (adapted from previous studies of American and British neighbourhoods) in a Canadian urban context. METHODS: This was a mixed-methods study. Quantitative SSO ratings and qualitative descriptions of 176 block faces were obtained in six Toronto neighbourhoods (4 low-income, and 2 middle/high-income) by trained raters. Exploratory factor analysis was conducted with the quantitative SSO ratings. Content analysis consisted of independent coding of qualitative data by three members of the research team to yield common themes and categories. RESULTS: Factor analysis identified three factors (physical decay/disorder, social accessibility, recreational opportunities), but only 'physical decay/disorder' reflected previous findings in the literature. Qualitative results (based on raters' fieldwork experiences) revealed the tool's shortcomings in capturing important features of the neighbourhoods under study, and informed interpretation of the quantitative findings. CONCLUSIONS: This study tested the performance of an SSO tool in a Canadian context, which is an important initial step before applying it to the study of health and disease. The tool demonstrated important shortcomings when applied to six diverse Toronto neighbourhoods. The study's analyses challenge previously held assumptions (e.g. social 'disorder') regarding neighbourhood social and built environments. For example, neighbourhood 'order' has traditionally been assumed to be synonymous with a certain degree of homogeneity, however the neighbourhoods under study were characterized by high degrees of heterogeneity and low levels of disorder. Heterogeneity was seen as an appealing feature of a block face. Employing qualitative techniques with SSO represents a unique contribution, enhancing both our understanding of the quantitative ratings obtained and of neighbourhood characteristics that are not currently captured by such instruments.

Effectiveness of Live Attenuated vs Inactivated Influenza Vaccines in Children During the 2012-2013 Through 2015-2016 Influenza Seasons in Alberta, Canada: A Canadian Immunization Research Network (CIRN) Study.

Importance: Recent observational studies report conflicting results regarding the effectiveness of live attenuated influenza vaccine (LAIV), particularly against influenza A(H1N1)pdm09. Objective: To compare the effectiveness of LAIV and inactivated influenza vaccine (IIV) against laboratory-confirmed influenza. Design, Setting, and Participants: A test-negative study to estimate influenza vaccine effectiveness (VE) using population-based, linked, individual-level laboratory, health administrative, and immunization data. Data were obtained from 10 169 children and adolescents aged 2 to 17 years (children) who were tested for influenza in inpatient or outpatient settings during periods when influenza was circulating based on a threshold level of 5% weekly test positivity for the province during the 4 influenza seasons spanning from November 11, 2012, to April 30, 2016, in Alberta, Canada. Logistic regression was used to estimate VE by vaccine type, influenza season, and influenza type and subtype. The relative effectiveness of each vaccine type was assessed by comparing the odds of laboratory-confirmed influenza infection for LAIV recipients with that for IIV recipients. Exposures: The primary exposure was receipt of LAIV or IIV before testing for influenza. Main Outcomes and Measures: The primary outcome was influenza case status as determined by reverse-transcriptase polymerase chain reaction testing. Results: A total of 10 779 respiratory specimens (from 10 169 children) collected and tested for influenza during the 4 influenza seasons were included, with 53.4% from males; the mean (SD) age was 7.0 (4.6) years. Across the 4 influenza seasons, 3161 children tested positive for influenza. Combining the 4 influenza seasons, the adjusted VE against influenza A(H1N1)pdm09 was 69% (95% CI, 56%-78%) for LAIV compared with 79% (95% CI, 70%-86%) for IIV. Vaccine effectiveness against influenza A(H3N2) was 36% (95% CI, 14%-53%) for LAIV and 43% (95% CI, 22%-59%) for IIV. Against influenza B, VE was 74% (95% CI, 62%-82%) for LAIV and 56% (95% CI, 41%-66%) for IIV. There were no significant differences in the odds of influenza infection for LAIV recipients compared with IIV recipients except for influenza B during the 2015-2016 season, when LAIV recipients had lower odds of infection than IIV recipients (odds ratio, 0.36; 95% CI, 0.17-0.76). Conclusions and Relevance: There was no evidence to support the lack of effectiveness of LAIV against influenza A(H1N1)pdm09. These results support administration of either vaccine type in this age group.

Contact investigation and genotyping to identify tuberculosis transmission to children.

BACKGROUND: Tuberculosis (TB) in young children is an indicator of ongoing community transmission. We examined contact investigations related to pediatric TB, yield for source case identifications and genotypes for relevant Mycobacterium tuberculosis isolates in a low-incidence setting. METHODS: We reviewed public health data for all patients with TB aged <18 years reported to Montreal authorities during 1996 to 2000. M. tuberculosis isolates from patients of all ages were subjected to IS6110-based genotyping, supplemented by spoligotyping, to compare isolates from children and adults during the same years. RESULTS: Sixty-six patients aged <18 years were diagnosed with active TB from 1996 to 2000. Mean age was 11.1 years (standard deviation 6.7 years). Twenty-five children (38%) were Canadian-born, all with at least one foreign-born parent. Nineteen children were diagnosed after contact investigations of known adult cases; 8 underwent no contact investigation. For the remaining 39 children, a total of 616 contacts were identified. The median number of contacts per child was 9 (interquartile range, 6-10). Four hundred eighty-one contacts (78%) underwent tuberculin testing; 188 (39%) were reactors and 186 (39%) began treatment of latent TB. Investigations uncovered 4 probable source cases, all involving parents or other relatives. M. tuberculosis genotyping for 38 children identified up to 14 additional possible source cases; in only one was a possible epidemiologic link evident from public health records. CONCLUSIONS: Among largely foreign-born children with active TB, contact investigations were extensive and often identified latent tuberculosis infection--but rarely source cases. However, genotyping suggested substantial, previously unrecognized transmission to children despite low overall incidence.

Public health surveillance response following the southern Alberta floods, 2013.

OBJECTIVE: In June of 2013, southern Alberta underwent flooding that affected approximately 100,000 people. We describe the process put in place for public health surveillance and assessment of the impacts on health. METHODS: Public health surveillance was implemented for the six-week period after the flood to detect anticipated health events, including injuries, mental health problems and infectious diseases. Data sources were emergency departments (EDs) for presenting complaints, public health data on the post-exposure administration of tetanus vaccine/immunoglobulin, administrative data on prescription drugs, and reportable diseases. RESULTS: An increase in injuries was detected through ED visits among Calgary residents (rate ratio [RR] 1.28, 95% confidence interval [CI]: 1.14-1.43) and was supported by a 75% increase in the average weekly administration of post-exposure prophylaxis against tetanus. Mental health impacts in High River residents were observed among females through a 1.64-fold (95% CI: 1.11-2.43) and 2.32-fold (95% CI: 1.45-3.70) increase in new prescriptions for anti-anxiety medication and sleep aids respectively. An increase in sexual assaults presenting to EDs (RR 3.18, 95% CI: 1.29-7.84) was observed among Calgary residents. No increases in infectious gastrointestinal disease or respiratory illness were identified. Timely identification and communication of surveillance alerts allowed for messaging around the use of personal protective equipment and precautions for personal safety. CONCLUSION: Existing data sources were used for surveillance following an emergency situation. The information produced, though limited, was sufficiently timely to inform public health decision-making.

Bayesian modelling of tuberculosis clustering from DNA fingerprint data.

A combination of continuous and categorical tests, none of which is a gold standard, is often available for classification of subject status in epidemiologic studies. For example, tuberculosis (TB) molecular epidemiology uses select mycobacterial DNA sequences to provide clues about which cases of active TB are likely clustered, implying recent transmission between these cases, versus reactivation of previously acquired infection. The proportion of recently transmitted cases is important to public health, as different control methods are implemented as transmission rates increase. Standard typing methods include IS6110 restriction fragment length polymorphism (IS6110 RFLP), but recently developed polymerase chain reaction based genotyping modalities, including mycobacterial interspersed repetitive unit-variable-number tandem repeat and spoligotyping provide quicker results. In addition, it has recently been suggested that results from IS6110 RFLP can be used to create a continuous measure of genetic relatedness, called the nearest genetic distance. Whichever method is used, estimation of cluster rates is rendered difficult by the lack of a gold standard method for classifying cases as clustered or not. Since many of these methods are relatively new, their properties have not been extensively investigated. Misclassification errors subsequently lead to sub-optimal estimation of risk factors for clustering. Here we show how Bayesian latent class models can be used in such situations, for example to simultaneously analyse Mycobacterium tuberculosis DNA data from all three of the above methods. Using the data collected at the Public Health Unit in Montreal, we estimate the proportion of clustered cases and the operating characteristics of each method using information from all three methods combined, including both continuous and dichotomous measures from IS6110 RFLP. A misclassification-adjusted regression model provides estimates of the effects of risk factors on the clustering probabilities. We also discuss how one must carefully interpret any inferences that arise from a combination of continuous and dichotomous tests.

Sensitivities and specificities of spoligotyping and mycobacterial interspersed repetitive unit-variable-number tandem repeat typing methods for studying molecular epidemiology of tuberculosis.

The development of PCR-based genotyping modalities (spoligotyping and mycobacterial interspersed repetitive unit-variable-number tandem repeat [MIRU-VNTR] typing) offers promise for real-time molecular epidemiological studies of tuberculosis (TB). However, the utility of these methods depends on their capacity to appropriately classify isolates. To determine the operating parameters of spoligotyping and MIRU-VNTR typing, we have compared results generated by these newer tests to the standard typing method, IS6110 restriction fragment length polymorphism, in analyses restricted to high-copy-number IS6110 isolates. Sensitivities of the newer tests were estimated as the percentages of isolates with identical IS6110 fingerprints that had identical spoligotypes and MIRU-VNTR types. The specificities of these tests were estimated as the percentages of isolates with unique IS6110 fingerprints that had unique spoligotypes and MIRU-VNTR types. The sensitivity of MIRU-VNTR typing was 52% (95% confidence interval [CI], 31 to 72%), and the sensitivity of spoligotyping was 83% (95% CI, 63 to 95%). The specificity of MIRU-VNTR typing was 56% (95% CI, 51 to 62%), and the specificity of spoligotyping was 40% (95% CI, 35 to 46%). The proportion of isolates estimated to be due to recent transmission was 4% by identical IS6110 patterns, 19% by near-identical IS6110 patterns, 33% by MIRU-VNTR typing, and 53% by spoligotyping. The low calculated specificities of spoligotyping and MIRU-VNTR typing led to misclassification of cases, inflated estimates of TB transmission, and low positive predictive values, suggesting that these techniques have unsuitable operating parameters for population-based molecular epidemiology studies.