Incidence Rate Ratio Estimates Research Articles

Childhood Cancer Incidence and Survival in South Australia and the Northern Territory, 1990-2017, with Emphasis on Indigenous Peoples.

Reports of a rise in childhood cancer incidence in Australia and globally prompted the investigation of cancer incidence and survival in South Australia (SA) and the Northern Territory (NT) over a 28-year period, with emphasis on Indigenous peoples. This cross-sectional analysis of two prospective longitudinal databases, the SA and NT Cancer Registries (1990-2017), included all reported cases of childhood cancers. Poisson regression provided estimates of incidence rate ratios and survival was modelled using Cox proportional hazard models for children aged <5 and ≥5 years. A total of 895 patients across SA (N = 753) and the NT (N = 142) were ascertained. Overall and in the NT, childhood cancer incidence was higher in males compared with females (IRR 1.19 [1.04-1.35] and 1.43 [1.02-2.01], respectively). Lymphocytic leukemia was the most reported cancer type across all locations. With reference to the 1990-1999 era (181.67/100,000), cancer incidence remained unchanged across subsequent eras in the combined cohort (SA and NT) (2000-2009: 190.55/100,000; 1.06 [0.91-1.25]; 2010-2017: 210.00/100,000; 1.15 [0.98-1.35]); similar outcomes were reflected in SA and NT cohorts. Cancer incidence amongst non-Indigenous children significantly decreased from the 1990-1999 era (278.32/100,000) to the 2000-2009 era (162.92/100,000; 0.58 [0.35-0.97]). Amongst 39 Indigenous children in the NT, incidence rates remained unchanged across eras (p > 0.05). With reference to the 1990-1999 era, overall survival improved in subsequent eras in SA (2000-2009: HR 0.53 [0.38-0.73]; 2010-2017: 0.44 [0.28-0.68]); however, remained unchanged in the NT (2000-2009: 0.78 [0.40-1.51]; 2010-2017: 0.50 [0.24-1.05]). In the NT, overall survival of Indigenous patients was significantly lower compared with the non-Indigenous cohort (3.42 [1.92-6.10]). While the survival of Indigenous children with cancer significantly improved in the last two eras (p < 0.05), compared to the 1990-1999 era, no change was noted amongst non-Indigenous children in the NT (p > 0.05). The incidence of childhood cancers has remained unchanged over 28-years in SA and the NT. Encouragingly, improved survival rates over time were observed in SA and amongst Indigenous children of the NT. Nevertheless, survival rates in Indigenous children remain lower than non-Indigenous children.

Performing up to Nordic principles? Geographic and socioeconomic equity in ambulatory care sensitive conditions among older adults in capital areas of Denmark, Finland and Sweden in 2000–2015

BackgroundDenmark, Finland and Sweden pursue equity in health for their citizens through universal health care. However, it is unclear if these services reach the older adult population equally across different socioeconomic positions or living areas. Thus, we assessed geographic and socioeconomic equity in primary health care (PHC) performance among the older adults in the capital areas of Denmark (Copenhagen), Finland (Helsinki) and Sweden (Stockholm) in 2000–2015. Hospitalisations for ambulatory care sensitive conditions (ACSC) were applied as a proxy for PHC performance.MethodsWe acquired individual level ACSCs for those aged ≥ 45 in 2000–2015 from national hospitalisation registers. To identify whether the disparities varied by age, we applied three age groups (those aged 45–64, 65–75 and ≥ 75). Socioeconomic disparities in ACSCs were described with incidence rate ratios (IRR) and annual rates by education, income and living-alone; and then analysed with biennial concentration indices by income. Geographic disparities were described with biennial ACSC rates by small areas and analysed with two-level Poisson multilevel models. These models provided small area estimates of IRRs of ACSCs in 2000 and their slopes for development over time, between which Pearson correlations were calculated within each capital area. Finally, these models were adjusted for income to distinguish between geographic and socioeconomic disparities.ResultsCopenhagen had the highest IRR of ACSCs among those aged 45–64, and Helsinki among those aged ≥ 75. Over time IRRs decreased among those aged ≥ 45, but only in Helsinki among those aged ≥ 75. All concentration indices slightly favoured the affluent population but in Stockholm were mainly non-significant. Among those aged ≥ 75, Pearson correlations were low in Copenhagen (-0.14; p = 0.424) but high in both Helsinki (-0.74; < 0.001) and Stockholm (-0.62; < 0.001) – with only little change when adjusted for income. Among those aged ≥ 45 the respective correlations were rather similar, except for a strong correlation in Copenhagen (-0.51, 0.001) after income adjustment.ConclusionsWhile socioeconomic disparities in PHC performance persisted among older adults in the three Nordic capital areas, geographic disparities narrowed in both Helsinki and Stockholm but persisted in Copenhagen. Our findings suggest that the Danish PHC incorporated the negative effects of socio-economic segregation to a lesser degree.

Multiple Imputation of Missing Race and Ethnicity in CDC COVID-19 Case-Level Surveillance Data.

The COVID-19 pandemic has resulted in a disproportionate burden on racial and ethnic minority groups, but incompleteness in surveillance data limits understanding of disparities. CDC's case-based surveillance system contains case-level information on most COVID-19 cases in the United States. Data analyzed in this paper contain COVID-19 cases with case-level information through September 25, 2020, which represent 70.9% of all COVID-19 cases reported to CDC during the period. Case-level surveillance data are used to investigate COVID-19 disparities by race/ethnicity, sex, and age. However, demographic information on race and ethnicity is missing for a substantial percentage of COVID-19 cases (e.g., 35.8% and 47.2% of cases analyzed were missing race and ethnicity information, respectively). Our goal in this study was to impute missing race and ethnicity to derive more accurate incidence and incidence rate ratio (IRR) estimates for different racial and ethnic groups, and evaluate the results from imputation compared to complete case analysis, which involves removing cases with missing race/ethnicity information from the analysis. Two multiple imputation (MI) models were developed. Model 1 imputes race using six binary race variables, and Model 2 imputes race as a composite multinomial variable. Our evaluation found that compared with complete case analysis, MI reduced biases and improved coverage on incidence and IRR estimates for all race/ethnicity groups, except for the Non-Hispanic Multiple/other group. Our research highlights the importance of supplementing complete case analysis with additional methods of analysis to better describe racial and ethnic disparities. When race and ethnicity data are missing, multiple imputation may provide more accurate incidence and IRR estimates to monitor these disparities in tandem with efforts to improve the collection of race and ethnicity information for pandemic surveillance.

Exposure\u2013lag response of smoking prevalence on lung cancer incidence using a distributed lag non-linear model

The prevalence of smokers is a major driver of lung cancer incidence in a population, though the “exposure–lag” effects are ill-defined. Here we present a multi-country ecological modelling study using a 30-year smoking prevalence history to quantify the exposure–lag response. To model the temporal dependency between smoking prevalence and lung cancer incidence, we used a distributed lag non-linear model (DLNM), controlling for gender, age group, country, outcome year, and population at risk, and presented the effects as the incidence rate ratio (IRR) and cumulative incidence rate ratio (IRRcum). The exposure–response varied by lag period, whilst the lag–response varied according to the magnitude and direction of changes in smoking prevalence in the population. For the cumulative lag–response, increments above and below the reference level was associated with an increased and decreased IRRcum respectively, with the magnitude of the effect varying across the lag period. Though caution should be exercised in interpretation of the IRR and IRRcum estimates reported herein, we hope our work constitutes a preliminary step towards providing policy makers with meaningful indicators to inform national screening programme developments. To that end, we have implemented our statistical model a shiny app and provide an example of its use.

A Decline and Age Shift in Malaria Incidence in Rural Mali following Implementation of Seasonal Malaria Chemoprevention and Indoor Residual Spraying.

ABSTRACTMany African countries have reported declines in malaria incidence, attributed to the implementation of control strategies. In Mali, artemisinin-based combination therapy (ACT) was introduced in 2004, and long-lasting insecticide-treated nets (LLINs) have been partially distributed free of charge since 2007. In the Malian town of Bandiagara, a study conducted from 2009 to 2013 showed a stable incidence of malaria compared with 1999, despite the implementation of ACTs and LLINs. Since 2016, seasonal malaria chemoprevention has been scaled up across the country. In addition to these strategies, the population of Bandiagara benefited from indoor residual spray implementation in 2017 and 2018 and continued universal bed net coverage. This study aimed to measure the incidence of malaria in Bandiagara, given this recent scaling up of control strategies. A cohort of 300 children aged 6 months to 15 years was followed up from October 2017 to December 2018. We performed monthly cross-sectional surveys to measure anemia and the prevalence of malaria infection by microscopy. The overall incidence of symptomatic malaria was 0.5 episodes/person-year. Malaria incidence in children up to 5 years old significantly declined since 2012 and since 1999 (incidence rate ratio estimates: 6.7 [95% CI: 4.2–11.4] and 13.5 [95% CI: 8.4–22.7]), respectively. The average prevalence of malaria parasitemia was 6.7%. Malaria incidence was higher in children older than 5 years than in those younger than 5 years, highlighting the need to extend malaria control efforts to these older children.

Ammonia, ammonium, and the risk of asthma: A register-based case–control study in Danish children

Background: Evidence concerning the health effects of exposure to fine particulate matter less than 2.5 μm in diameter (PM2.5) and particulate components such as ammonium (NH4 +) is limited. We aimed to investigate gaseous ammonia (NH3), particulate ammonium (NH4 +), the total concentration of these pollutants (NHx), and PM2.5 and their association with asthma in Danish preschool children. Methods: Estimates of incidence rate ratios associated with exposure to air pollution were calculated for children followed for the development of childhood asthma (n = 335,629) from their first to their sixth birthday during the time period 2006–2012 using a time-matched case–control design by combining register information. Modeled average concentrations of NH3, NH4 +, NHx, and PM2.5 (5.56 km × 5.56 km grid resolution) during the past 3 months prior to the date of first diagnosis of asthma for cases. Results: We identified 12,935 incident cases of asthma. The base adjusted hazard ratio of incident asthma was 1.74 for children exposed to the highest decile of NH3 exposure (95% confidence interval = 1.60, 1.89) compared to children exposed to the lowest decile of NH3 exposure during the past 3 months prior to first diagnosis. Similar hazard ratios were found for the highest levels of NH4 + (2.33; 2.04, 2.65) and NHx (1.82; 1.68, 1.96). The positive associations were slightly attenuated, when adjusting for socioeconomic status, but disappeared when additionally adjusting for region, except in a two-pollutant model including NH4 + and PM2.5. No association was found with PM2.5. The findings were robust when using different exposure time windows. Conclusions: Increased NH4 + exposure may be a risk factor for the onset of asthma in preschool children.

The risk of developing major depression among individuals with subthreshold depression: a systematic review and meta-analysis of longitudinal cohort studies.

Studies have consistently shown that subthreshold depression is associated with an increased risk of developing major depression. However, no study has yet calculated a pooled estimate that quantifies the magnitude of this risk across multiple studies. We conducted a systematic review to identify longitudinal cohort studies containing data on the association between subthreshold depression and future major depression. A baseline meta-analysis was conducted using the inverse variance heterogeneity method to calculate the incidence rate ratio (IRR) of major depression among people with subthreshold depression relative to non-depressed controls. Subgroup analyses were conducted to investigate whether IRR estimates differed between studies categorised by age group or sample type. Sensitivity analyses were also conducted to test the robustness of baseline results to several sources of study heterogeneity, such as the case definition for subthreshold depression. Data from 16 studies (n = 67 318) revealed that people with subthreshold depression had an increased risk of developing major depression (IRR = 1.95, 95% confidence interval 1.28-2.97). Subgroup analyses estimated similar IRRs for different age groups (youth, adults and the elderly) and sample types (community-based and primary care). Sensitivity analyses demonstrated that baseline results were robust to different sources of study heterogeneity. The results of this study support the scaling up of effective indicated prevention interventions for people with subthreshold depression, regardless of age group or setting.

Signal Detection for Recently Approved Products: Adapting and Evaluating Self-Controlled Case Series Method Using a US Claims and UK Electronic Medical Records Database.

The Self-Controlled Case Series (SCCS) method has been widely used for hypothesis testing, but there is limited evidence of its performance for safety signal detection. The objective of this study was to evaluate SCCS for signal detection on recently approved products. A retrospective study covered the period after three recently marketed drugs were launched through to 31 December 2010 using The Health Improvement Network, a UK primary care database, and Optum, a US claims database. The SCCS method was applied to examine five heterogenous outcomes with desvenlafaxine and escitalopram and six outcomes with adalimumab for Signals of Disproportional Recording (SDRs); a positive finding was determined to be when the lower bound of 95% Confidence Interval ofthe incidence rate ratio (IRR) estimate was> 1. Multiple design choices were tested and the trend in IRR estimates over calendar time for one drug event pair was examined. All six outcomes with adalimumab, three of five outcomes with desvenlafaxine, and four of five outcomes with escitalopram had SDRs. SCCS highlighted all acute events in the primary analysis but was less successful with slower-onset outcomes. Performance varied by risk period definition. Changes in IRR estimates over quarterly intervals for adalimumabwith herpes zoster showed marked higher SDR within 9months of drug launch. SCCS shows promise for signal detection: it may highlight known associations for recent marketed products and has potential for early signal identification. SCCS performance varied by design choice and the nature of both exposure and event pair. Future work is needed to determine how effective the approach is in prospective testing and determining the performance characteristics of the approach.

HOST Study — HIV in Orthopaedic Skeletal Trauma Study: protocol for a multicentre case-cohort study

Background: Human immunodeficiency virus (HIV) and antiretroviral therapy (ART) have both been shown to reduce bone mineral density, mineralisation and bone turnover. Our study group and other researchers have suggested that HIV may impair fracture healing, based on extrapolation from basic science. These observations prompted this study as the true effect of HIV and highly active antiretroviral therapy (HAART) on bone healing is very poorly understood and has not previously been investigated. Methods: HOST Study is a multicentre case-cohort study being undertaken at two orthopaedic trauma centres in Cape Town, South Africa. All adult patients older than 18 years with fresh (within 2 weeks of injury), closed and open, tibia and femur fractures who undergo intramedullary (IM) nailing for fracture fixation will be eligible or the study. Participants will be recruited over 24 months and undergo a baseline questionnaire, HIV testing and assessment of their bone mineral density (BMD). They will be followed up at 2 and 6 weeks, and at 3, 6, 9 and 12 months. All adult patients who develop delayed bone union at the 6-month follow-up will be considered cases. Adult patients who show evidence of radiological union at 6 months or less will be considered controls. We will then determine if HIV is a risk factor for the development of delayed bone union. HIV prevalence levels in the cases and controls will be summarised using IRR (incidence rate ratio) statistics with their 95% confidence intervals. Negative binomial regression methodswill be used to adjust the IRR estimates for the possible effects of confounding factors and/or important covariates. Results: Outcomes from the primary manuscript will be disseminated through publications in academic journals and presentations at relevant orthopaedic conferences. We will communicate trial results to all participating sites. Participating sites will communicate results with patients who have indicated an interest in knowing the results. Trial registration number: ClinicalTrials.gov - NCT03131947 Site of study: Groote Schuur Hospital and Tygerberg Hospital, Cape Town, South Africa

Antibiotic Exposure and Risk of Community-associated Clostridium difficile infection (CA-CDI): A Self-Controlled Case Series Analysis

BackgroundCA-CDI accounts for up to 50% of all CDIs. Case–control studies (CCS) have been used to estimate the odds ratio (OR) of CA-CDI associated with antibiotic exposure. These ORs demonstrate significant heterogeneity across studies. Unlike CCS, a self-controlled case series (SCCS) design can be used to control for all time-invariant confounders leading to less biased effect estimates.MethodsAdults (≥18 years) registered (N = 139,670) with the Barrie and Community Family Health Team (BCFHT) were included in the study. Cases were defined as any patient with an incident case of CA-CDI and ≥1 antibiotic exposure occurring between January 1, 2011 and December 31, 2016. The SCCS model was used to estimate the association between antibiotic exposure and CA-CDI. The SCCS model yields estimates of the relative incidence rate of CA-CDI in exposure periods relative to non-exposure periods within a case. Exposure periods were defined as starting two days after any antibiotic prescription and ending 60 days later. Multiple exposure periods and time-varying confounders due to calendar year were included in the final model. The relative incidence rate ratio (IRR) was estimated using conditional poisson regression analysis. Proton pump inhibitor (PPI) use was included as an effect modifier. Antibiotics were divided into high-risk (fluoroquinolone, clindamycin, and cephalosporin) and low-risk exposures. Research ethics approval was obtained from the BFCHT research ethics board.ResultsAmong 544 total CDI cases, N = 189 CA-CDI cases met the inclusion criteria. Any antibiotic exposure increases the risk by ≥2-fold, with no difference observed between high and low-risk groups (IRR=1.11, 95% CI 0.53–2.36) (Table 1).ConclusionAntibiotic exposure increases the risk of CA-CDI, with IRR estimates similar to those observed for healthcare-associated-CDI. This, along with the control of all time-invariant confounders by the SCCS method suggests a less biased effect estimate previously reported from CCS.Table 1 Variable IRR 95% Confidence Interval P-value Antibiotic Exposure GroupPPINoneYes0.80(0.62–1.03)0.09Low riskYes1.95(0.09–4.24)0.09High riskYes1.20(0.42–3.40)0.73OverallLow risk2.03(1.19–3.47)0.009High risk2.26(1.29–3.98)0.005Disclosures All authors: No reported disclosures.

The impact of 10-valent and 13-valent pneumococcal conjugate vaccines on hospitalization for pneumonia in children: A systematic review and meta-analysis.

This systematic review and meta-analysis aimed at summarizing available data on the impact of PCV10 and PCV13 in reducing the incidence of CAP hospitalizations in children aged <5years. A systematic search of the literature was conducted. We included time-series analyses and before-after studies, reporting the incidence of hospitalization for pneumonia in the periods before and after the introduction of PCV10 or PCV13 into the immunization program. Pooled estimates of Incidence Rate Ratio (IRR) were calculated by using a random-effects meta-analytic model. Results were stratified according to age-groups (<24months and 24-59months) and case definitions of pneumonia (clinically and radiologically confirmed pneumonia). A total of 1533 potentially relevant articles were identified. Of these, 12 articles were included in the analysis. In children aged <24months, the meta-analysis showed a reduction of 17% (95%CI: 11-22%, p-value<0.001) an of 31% (95%CI: 26-35%, p-value<0.001) in the hospitalization rates respectively for clinically and radiologically confirmed pneumonia, respectively, after the introduction of the novel PCVs. In children aged 24-59months, the meta-analysis showed a reduction of 9% (95%CI: 5-14%, p-value<0.001) and of 24% (95%CI: 12-33%, p-value<0.001) in the hospitalization rates for clinically and radiologically confirmed pneumonia, respectively, after the introduction of the novel PCVs. High heterogeneity was detected among studies evaluating the hospitalization rate for clinically and radiologically confirmed pneumonia. The results of this study revealed a significant impact of PCV10 and PCV13 in reducing the hospitalizations for pneumonia, particularly in children aged <24months and for radiologically confirmed disease. Further appropriately designed studies, comparing the impact of PCV10 and PCV13, are needed in order to obtain solid data on which to establish future immunization strategies.

Matching the Statistical Model to the Research Question for Dental Caries Indices with Many Zero Counts

Marginalized zero-inflated count regression models have recently been introduced for the statistical analysis of dental caries indices and other zero-inflated count data as alternatives to traditional zero-inflated and hurdle models. Unlike the standard approaches, the marginalized models directly estimate overall exposure or treatment effects by relating covariates to the marginal mean count. This article discusses model interpretation and model class choice according to the research question being addressed in caries research. Two data sets, one consisting of fictional dmft counts in 2 groups and the other on DMFS among schoolchildren from a randomized clinical trial comparing 3 toothpaste formulations to prevent incident dental caries, are analyzed with negative binomial hurdle, zero-inflated negative binomial, and marginalized zero-inflated negative binomial models. In the first example, estimates of treatment effects vary according to the type of incidence rate ratio (IRR) estimated by the model. Estimates of IRRs in the analysis of the randomized clinical trial were similar despite their distinctive interpretations. The choice of statistical model class should match the study's purpose, while accounting for the broad decline in children's caries experience, such that dmft and DMFS indices more frequently generate zero counts. Marginalized (marginal mean) models for zero-inflated count data should be considered for direct assessment of exposure effects on the marginal mean dental caries count in the presence of high frequencies of zero counts.

Venous thromboembolism in patients with implantable cardioverter-defibrillators.

To examine the incidence of venous thromboembolism (VTE) and its risk factors among patients with implantable cardioverter-defibrillators (ICDs). All first-time ICD recipients in Denmark during 2000-12 were identified from medical databases. Incident VTEs were ascertained, overall and according to gender, age, Charlson Comorbidity Index score (no, moderate, or severe comorbidity), prior pacemaker or cardiac resynchronization therapy (CRT-D) implantation, and ICD type (single-chamber, dual-chamber, or CRT-D). We computed the risk of VTE within 3 months and 5 years of implantation, taking death into account as a competing risk. We used Cox proportional hazards regression to compute hazard ratios as estimates of incidence rate ratios (IRRs). Among 8132 ICD recipients, 136 VTEs were diagnosed during up to 13 years of follow-up (median = 3.0 years). The VTE incidence rate was thus 4.5 per 1000 person-years [95% confidence interval (CI): 3.7-5.2]. Venous thromboembolism risk was 0.3% (95% CIs ranging from 0.1 to 0.7%) within 3 months following ICD implantation regardless of comorbidity level. Within 5 years following implantation it was 1.4% (95% CI: 0.8-2.3%), 1.3% (1.0-1.8%), and 3.2% (95% CI: 2.4-4.1%) for patients with no, moderate, and severe comorbidity, respectively. Overall, severe comorbidity conferred a 2.7-fold higher incidence rate ratio than no comorbidity (95% CI: 1.6-4.6). Incidence rate ratios did not differ by gender, age, or ICD type. Three-month risk of VTE following ICD implantation was 0.3% regardless of comorbidity level. Five-year risk of VTE following ICD implantation was 1.9% and more than twice as high for patients with severe comorbidity as for patients without comorbidity.

Short-term weight gain after antiretroviral therapy initiation and subsequent risk of cardiovascular disease and diabetes: the D:A:D study.

The aim of the study was to assess the impact of the gain in body mass index (BMI) observed immediately after antiretroviral therapy (ART) initiation on the subsequent risk of cardiovascular disease (CVD) and diabetes. We analysed data from the Data Collection on Adverse Events of Anti-HIV Drugs (D:A:D) cohort study. Outcomes were development of (i) CVD (composite of myocardial infarction/stroke/coronary procedure) and (ii) diabetes. The main exposure variable was change in BMI from ART initiation (pre-ART) to 1 year after initiation (continuous variable) in treatment-naïve individuals initiating ART with no history of CVD or diabetes (for respective outcomes). BMI [weight (kg)/(height (m))(2)] was categorized as underweight (< 18.5), normal (18.5-25), overweight (25-30) and obese (> 30). Poisson regression models were fitted stratified for each pre-ART BMI category to allow for category-specific estimates of incidence rate ratio (IRR). Models were adjusted for pre-ART BMI and CD4 count, key known risk factors (time-updated where possible) and calendar year. A total of 97 CVD events occurred in 43,982 person-years (n = 9321) and 125 diabetes events in 43,278 person-years (n = 9193). In fully adjusted analyses for CVD, the IRR/unit gain in BMI (95% confidence interval) in the first year of ART, by pre-ART BMI category, was: underweight, 0.90 (0.60-1.37); normal, 1.18 (1.05-1.33); overweight, 0.87 (0.70-1.10), and obese, 0.95 (0.71-1.28) (P for interaction = 0.04). For diabetes, the IRR/unit gain in BMI was 1.11 (95% confidence interval 1.03 to 1.21), regardless of pre-ART BMI (P for interaction > 0.05). Short-term gain in BMI following ART initiation appeared to increase the longer term risk of CVD, but only in those with pre-ART BMI in the normal range. It was also associated with increased risk of diabetes regardless of pre-ART BMI.

Influence of different drugs on HIV risk in people who inject: systematic review and meta-analysis.

To assess systematically the risk of HIV acquisition by type of drug injected across different settings. A systematic review and meta-analysis were conducted. Databases were searched for studies of HIV incidence in people who inject different drugs (PWID). Pooled HIV incidence rate ratio (IRR) was used to compare HIV risk between injecting and not-injecting a given drug, when possible, or otherwise with those reported not to have injected the substance. Pooled estimates of crude IRR were derived using random-effects models. Variations in IRR were assessed in subgroup analyses, by drug and geographical region. Of 5779 studies screened, 15 were included. HIV incidence was reported for people injecting cocaine (eight: North America, Europe), amphetamine-type stimulants (ATS) (four: Western and Eastern Europe, Asia), heroin (11: all settings), opiate-stimulants (four: North America, Western and Eastern Europe) and opiates-sedatives (five: Europe, Asia). HIV risk in cocaine injectors was 3.6 times 95% confidence interval (CI) = 2.8-4.7, I(2) = 0%; n = 4) that of non-injectors and 3.0 for ATS injectors (95% CI = 2.2-4.1, I(2) = 0%; n = 2). Higher sexual risk was reported in cohorts injecting stimulants. Compared to not-injecting, HIV IRR was 2.8 (95% CI = 1.7-4.7, I(2) = 77%; n=6) for all heroin injectors and 3.5 (95% CI = 2.3-5.2, I(2) = 40%; n=5) for heroin injectors in Asia and Europe. The risk of HIV acquisition in people who inject drugs appears to vary by drug type but differences are not statistically significant, precluding conclusive grading of risk.

Signal detection of adverse events with imperfect confirmation rates in vaccine safety studies using self‐controlled case series design

The Vaccine Safety Datalink project captures electronic health record data including vaccinations and medically attended adverse events on 8.8 million enrollees annually from participating managed care organizations in the United States. While the automated vaccination data are generally of high quality, a presumptive adverse event based on diagnosis codes in automated health care data may not be true (misclassification). Consequently, analyses using automated health care data can generate false positive results, where an association between the vaccine and outcome is incorrectly identified, as well as false negative findings, where a true association or signal is missed. We developed novel conditional Poisson regression models and fixed effects models that accommodate misclassification of adverse event outcome for self-controlled case series design. We conducted simulation studies to evaluate their performance in signal detection in vaccine safety hypotheses generating (screening) studies. We also reanalyzed four previously identified signals in a recent vaccine safety study using the newly proposed models. Our simulation studies demonstrated that (i) outcome misclassification resulted in both false positive and false negative signals in screening studies; (ii) the newly proposed models reduced both the rates of false positive and false negative signals. In reanalyses of four previously identified signals using the novel statistical models, the incidence rate ratio estimates and statistical significances were similar to those using conventional models and including only medical record review confirmed cases.

Incident fractures in HIV-infected individuals

Some but not all studies indicate that individuals with HIV infection are at an increased risk of fracture. We systematically reviewed the literature to investigate whether incidence of fracture (both overall and fragility) differs between individuals with and without HIV. A systematic review and meta-analysis. Medline, Scopus and the Cochrane Library databases for all studies ever published up to 28 September 2012 and electronically available conference abstracts from CROI, ASBMR, IAS and AIDS were searched. All studies reporting incidence of all fracture and fragility fracture in HIV-infected adults were included. A random effects model was used to calculate pooled estimates of incidence rate ratios (IRRs) for studies that presented data for HIV-infected and controls. For all studies, incidence rates of fracture and predictors of fracture among HIV-infected individuals were summarized. Thirteen eligible studies were analysed, of which seven included controls. Nine studies reported all incident fractures and 10 presented incident fragility fractures. The pooled IRR was 1.58 [95% confidence interval (CI) 1.25-2.00] for all fracture and 1.35 (95% CI 1.10-1.65) for fragility fracture. Smoking, white race and older age were consistent predictors for fragility fractures. Our results indicate that HIV infection is associated with a modest increase in incident fracture. Future research should focus on clarifying risk factors, designing appropriate interventions and the long-term implications of this increased risk for an ageing HIV-infected population.

P3.102 HIV Risk in People Who Inject Different Drugs: Systematic Review and Meta-Analysis

ObjectiveTo compare and understand how the risk of HIV infection in people who inject drugs (PWID) varies by substances and regions.MethodsA systematic review and meta-analysis was conducted. EMBASE and Medline...

Risk of childhood injuries after prenatal exposure to maternal bereavement: a Danish National Cohort Study

ObjectivesThe aim of this study was to assess the risk of injuries among children exposed to a stressful life exposure (defined as bereavement) before conception or during fetal life.DesignPopulation-based cohort...

Nonsteroidal anti‐inflammatory drugs and the risk of skin cancer

Nonsteroidal anti-inflammatory drugs (NSAIDs) may prevent the development of cancer by inhibiting cyclooxygenase (COX) enzymes, which are involved in carcinogenesis. Therefore, the authors of this report examined the association between NSAID use and the risk of squamous cell carcinoma (SCC), basal cell carcinoma (BCC), and malignant melanoma (MM). From 1991 through 2009, all incident cases of SCC (n = 1974), BCC (n = 13,316), and MM (n = 3242) in northern Denmark were identified. Approximately 10 population controls (n = 178,655) were matched to each case by age, gender, and county of residence. The use of aspirin, other nonselective NSAIDs, or selective COX-2 inhibitors was ascertained through a prescription database. Conditional logistic regression analyses adjusted for potential confounders were used to compute odds ratios as estimates of incidence rate ratios (IRRs). For NSAIDs overall, ever use (>2 prescriptions) compared with nonuse (≤2 prescriptions) was associated with a decreased risk of SCC (IRR, 0.85; 95% confidence interval [CI], 0.76-0.94) and MM (IRR, 0.87; 95% CI, 0.80-0.95), especially for long-term use (≥7 years) and high-intensity use (>25% prescription coverage during the total duration of use). NSAID use was not associated with a reduced risk of BCC overall (IRR, 0.97; 95% CI, 0.93-1.01), but the risk of BCC at sites other than the head and neck was reduced in association with long-term use (IRR, 0.85; 95% CI, 0.76-0.95) and high-intensity use (IRR, 0.79; 95% CI, 0.69-0.91). All estimates of reduced risk were driven primarily by the use of nonselective NSAIDs and older COX-2 inhibitors (diclofenac, etodolac, and meloxicam). The current results indicated that NSAID use may decrease the risk of SCC and MM.