Lessons learned from a multi-data source research collaboration: The mirabegron post-authorization safety study program.

BACKGROUND: Many factors contribute to developing and conducting a successful multi-data source, non-interventional, post-authorization safety study (NI-PASS) for submission to multiple health authorities. Such studies are often large undertakings; evaluating and sharing lessons learned can provide useful insights to others considering similar studies. OBJECTIVES: We discuss challenges and key methodological and organizational factors that led to the delivery of a successful post-marketing requirement (PMR)/PASS program investigating the risk of cardiovascular and cancer events among users of mirabegron, an oral medication for the treatment of overactive bladder. RESULTS: We provide context and share learnings, including sections on research program collaboration, scientific transparency, organizational approach, mitigation of uncertainty around potential delays, validity of study outcomes, selection of data sources and optimizing patient numbers, choice of comparator groups and enhancing precision of estimates of associations, potential confounding and generalizability of study findings, and interpretation of results. CONCLUSIONS: This large PMR/PASS program was a long-term commitment from all parties and benefited from an effective coordinating center and extensive scientific interactions across research partners, scientific advisory board, study sponsor, and health authorities, and delivered useful learnings related to the design and organization of multi-data source NI-PASS.

Mortality in Patients with Parkinson's Disease-Related Psychosis Treated with Pimavanserin Compared with Other Atypical Antipsychotics: A Cohort Study.

INTRODUCTION: Pimavanserin is approved in the USA to treat hallucinations and delusions associated with Parkinson's disease psychosis (PDP). OBJECTIVES: We evaluated mortality in patients with PDP after initiation of pimavanserin or comparator atypical antipsychotics, overall, over time, and across subgroups. METHODS: A cohort of patients aged ≥65 years in the USA with PDP newly initiating pimavanserin or a comparator atypical antipsychotic (clozapine, quetiapine, risperidone, olanzapine, aripiprazole, brexpiprazole) was identified in 2016-2019 Medicare claims data. All-cause mortality in the propensity score-matched treatment groups was compared with hazard ratios (HRs) and 95% confidence intervals (CIs) estimated with Cox-proportional hazards models. Cumulative incidence curves and time period-specific models evaluated risk over time. Subgroup and sensitivity analyses were performed, including a sub-cohort of long-term care (LTC) or skilled nursing facility (SNF) residents. RESULTS: We identified 2892 pimavanserin initiators and 19,083 comparator initiators (overall 47% female, mean age = 80.9 years, LTC/SNF residents = 30%). Before matching, pimavanserin users had fewer severe comorbidities and more anti-Parkinson medication use than comparators. Matching resulted in 2891 patients in both groups, and all covariates were well balanced. In the matched cohort, the HR for mortality for pimavanserin versus comparator was 0.78 (95% CI 0.67-0.91), with the lowest time period-specific HRs in the first 180 days. Hazard ratios were similar across sensitivity analyses and subgroups. In LTC/SNF residents, the HR was 0.78 (95% CI 0.60-1.01). CONCLUSION: The observed mortality rates were lower among patients treated with pimavanserin compared with those treated with other atypical antipsychotics. STUDY REGISTRATION: European Union Post-authorization Study (EU PAS) register number 46331.

Clinical outcomes and treatment patterns of older adults with dementia-related psychosis by dementia type in the United States.

BACKGROUND: Little is known about the incidence of clinical events and treatment patterns among older adults with dementia-related psychosis. Given that dementia-related psychosis comprises various dementia types, this study describes the incidence of clinical events and treatment patterns by dementia type after patients with dementia are diagnosed with psychosis. METHODS: Adults aged ≥ 65 years with dementia and newly diagnosed with psychosis were identified in US Medicare claims during 2013-2018. Baseline characteristics were evaluated at the time of the psychosis diagnosis. After the initial psychosis diagnosis, incidence rates (IRs) of clinical events (e.g., falls/fractures, infections, healthcare utilization), mortality, and patterns of antipsychotic treatment were described for each dementia type (Alzheimer's disease [AD], Parkinson's disease dementia [PDD], dementia with Lewy bodies [DLB], frontotemporal dementia [FTD], vascular dementia [VD], and unspecified dementia). Daily mean cumulative counts were estimated to describe the incidence of recurrent events over time. Mortality was described using Kaplan-Meier survival curves. RESULTS: We identified 484,520 patients with dementia-related psychosis: mean age, 84 years (standard deviation, 7.8); female, 66%. At the time of psychosis diagnosis, the most prevalent type of dementia was unspecified dementia (56%), followed by AD (31%), VD (12%), PDD (10%), DLB (3%), and FTD (< 1%), and most patients had scores indicating severe illness on the Charlson Comorbidity Index (71%) and frailty index (62%). Across all dementia types, IRs (per 100 person-years) were high for emergency department visits, oral anti-infective use, and urinary tract infections after the initial psychosis diagnosis. Patients with DLB had the highest incidence of most clinical outcomes. After 1 year of follow-up, the cumulative probability of death was about 30% for all dementia types, and after 5 years, was about 80% among patients with DLB, VD, AD, or PDD and about 60%-65% among patients with FTD or unspecified dementia. CONCLUSIONS: Patients with dementia-related psychosis had a high burden of comorbidities, frailty, emergency department visits, infections, and death. Specifically, after DRP diagnosis, patients with DLB and VD had the highest burden of clinical events of interest.

Validation of cardiovascular outcomes and risk factors in the Clinical Practice Research Datalink in the United Kingdom.

PURPOSE: Strategies to identify and validate acute myocardial infarction (AMI) and stroke in primary-care electronic records may impact effect measures, but to an unknown extent. Additionally, the validity of cardiovascular risk factors that could act as confounders in studies on those endpoints has not been thoroughly assessed in the United Kingdom Clinical Practice Research Datalink's (CPRD's) GOLD database. We explored the validity of algorithms to identify cardiovascular outcomes and risk factors and evaluated different outcome-identification strategies using these algorithms for estimation of adjusted incidence rate ratios (IRRs). METHODS: First, we identified AMI, stroke, smoking, obesity, and menopausal status in a cohort treated for overactive bladder by applying computerized algorithms to primary care medical records (2004-2012). We validated these cardiovascular outcomes and risk factors with physician questionnaires (gold standard for this analysis). Second, we estimated IRRs for AMI and stroke using algorithm-identified and questionnaire-confirmed cases, comparing these with IRRs from cases identified through linkage with hospitalization/mortality data (best estimate). RESULTS: For AMI, the algorithm's positive predictive value (PPV) was >90%. Initial algorithms for stroke performed less well because of inclusion of codes for prevalent stroke; algorithm refinement increased PPV to 80% but decreased sensitivity by 20%. Algorithms for smoking and obesity were considered valid. IRRs based on questionnaire-confirmed cases only were closer to IRRs estimated from hospitalization/mortality data than IRRs from algorithm-identified cases. CONCLUSIONS: AMI, stroke, smoking, obesity, and postmenopausal status can be accurately identified in CPRD. Physician questionnaire-validated AMI and stroke cases yield IRRs closest to the best estimate.

Patterns of use of antimuscarinic drugs to treat overactive bladder in Denmark, Sweden, and the United Kingdom.

PURPOSE: To describe the use of antimuscarinic drugs to treat overactive bladder (OAB) in Denmark, Sweden, and the United Kingdom (UK). METHODS: We identified new users of darifenacin, fesoterodine, oxybutynin, solifenacin, tolterodine, and trospium aged 18 years or older from the Danish National Registers (2004-2012), the Swedish National Registers (2006-2012), and UK Clinical Practice Research Datalink (2004-2012). Users were followed until disenrollment, cancer diagnosis, death, or study end. Treatment episodes, identified by linking consecutive prescriptions, were described with respect to duration, drug switch, and drug add-on. RESULTS: Mean age of OAB drug users was 66 years in Denmark (n = 72,917) and Sweden (n = 130,944), and 62 years in the UK (n = 119,912); 60% of Danish and Swedish patients and 70% of UK patients were female. In Denmark, of 224,680 treatment episodes, 39% were with solifenacin, and 35% with tolterodine; 2% were with oxybutynin. In Sweden, of 240,141 therapy episodes, 37% were with tolterodine and 35% with solifenacin; 5% were with oxybutynin. In the UK, of 245,800 treatment episodes, 28% were with oxybutynin, 27% with solifenacin, and 26% with tolterodine. In the three countries, 49%-52% of treatment episodes comprised one prescription and over 80% of episodes ended because of no refill; less than 20% ended because of a switch to another antimuscarinic. During the study years, we observed a change in OAB treatment preference from tolterodine to solifenacin. CONCLUSIONS: In these cohorts, persistence with antimuscarinic drugs was low. By 2012, the preferred drug was solifenacin; oxybutynin use was marginal in Nordic countries compared with the UK.

Variation in Cardiovascular Risk Related to Individual Antimuscarinic Drugs Used to Treat Overactive Bladder: A UK Cohort Study.

BACKGROUND: Blocking muscarinic receptors could have an effect on cardiac function, especially among elderly patients with overactive bladder (OAB). STUDY OBJECTIVE: To investigate the risk of cardiovascular (CV) events in users of antimuscarinic drugs to treat OAB. DESIGN, SETTING, AND PARTICIPANTS: Cohort study of new users of darifenacin, fesoterodine, oxybutynin, solifenacin, tolterodine, or trospium, 18 years or older, in the United Kingdom's Clinical Practice Research Datalink (CPRD), 2004-2012. OUTCOME MEASUREMENTS AND MAIN RESULTS: Using tolterodine as the reference, we estimated propensity-score-stratified incidence rate ratios (IRRs) for acute myocardial infarction, stroke, CV mortality, major adverse cardiac events (MACE, a combined end point of the previous three), and all-cause death for individual antimuscarinic drugs. The study cohort included 119,912 new users of OAB drugs. The mean age at cohort entry was 62 years, 70% were female, and the mean follow-up was 3.3 years. The adjusted IRR for MACE and current use of oxybutynin compared with current use of tolterodine was 1.14 (95% confidence interval [CI] 1.01-1.30). In contrast, the IRR was 0.65 (CI 0.56-0.76) for current use of solifenacin compared with tolterodine. In this study, performed with health care data, the distribution of risk factors was relatively similar across users of different OAB drugs and, although our analyses controlled for a range of measured potential confounders, residual confounding cannot be ruled out. CONCLUSIONS: In an observational comparative study of users of medications to treat OAB conducted in routine clinical practice, the risk for CV side effects was increased in users of oxybutynin and decreased in users of solifenacin compared with users of tolterodine.

Validation of Cancer Cases Using Primary Care, Cancer Registry, and Hospitalization Data in the United Kingdom.

BACKGROUND: In the United Kingdom, hospital or cancer registry data can be linked to electronic medical records for a subset of general practices and years. METHODS: We used Clinical Practice Research Datalink data (2004-2012) from patients treated for overactive bladder. We electronically identified provisional cases of 10 common cancers in General Practitioner Online Database data and validated them by medical profile review. In practices with linkage to Hospital Episodes Statistics and National Cancer Data Repository (2004-2010), we validated provisional cancer cases against these data sources. This linkage also let us identify additional cancer diagnoses in individuals without cancer diagnosis records in the General Practitioner Online Database. RESULTS: Among 50,840 patients, 1,486 provisional cancer cases were identified in the General Practitioner Online Database for 2004-2012. Medical profile review confirmed 93% of 661 cases in nonlinked practices (range, 100% of non-Hodgkin lymphomas and uterine cancer to 77% of skin melanomas) and 96% of 825 cases in linked practices (100% of kidney and uterine cancers to 92% of melanomas). In the subset of linked practices, for 2004-2010, 720 cases were confirmed, of which 68% were identifiable in the General Practitioner Online Database (range, 90% of breast to 36% of kidney cancers). CONCLUSIONS: Most cases of cancer identified electronically in the General Practitioner Online Database were confirmed. A substantial proportion of cases, especially of cancer types not typically managed by general practitioners, would be missed without Hospital Episodes Statistics and National Cancer Data Repository data (and are likely missed in nonlinked practices). See video abstract at, http://links.lww.com/EDE/B315. REGISTRATION (BEFORE STUDY CONDUCT): European Union electronic Register of Post-Authorisation Studies (EU PAS Registry) number EUPAS5529, http://www.encepp.eu/encepp/viewResource.htm?id=11107.

Cancer Incidence after Initiation of Antimuscarinic Medications for Overactive Bladder in the United Kingdom: Evidence for Protopathic Bias.

STUDY OBJECTIVE: To estimate the incidence of 10 common cancers among patients treated with antimuscarinic medications for overactive bladder (AMOABs). DESIGN: Retrospective cohort study. DATA SOURCE: United Kingdom's Clinical Practice Research Datalink. PATIENTS: A total of 119,912 adults with no previous cancer diagnosis who were new users of AMOABs-darifenacin, fesoterodine, oxybutynin, solifenacin, tolterodine, or trospium-between January 2004 and December 2012. MEASUREMENTS AND MAIN RESULTS: Sex-specific incidence rates per 1000 person-years and 95% confidence intervals (CIs) were estimated for each study cancer (bladder, breast, colorectal, lung, melanoma, non-Hodgkin lymphoma, pancreatic, prostate, renal, and uterine cancer) overall and stratified by time since cohort entry and by cumulative AMOAB dose. Among the 119,912 patients followed for 399,365 person-years, 4117 incident study cancers occurred. The incidence rate of prostate cancer was 14.2 (95% CI 12.9-15.5) in the year after cohort entry and decreased markedly thereafter. The incidence rate of bladder cancer was also higher in the year after cohort entry than subsequently (men: 5.5, 95% CI 4.8-6.4; women: 1.2, 95% CI 1.0-1.5). The incidence rates of both prostate and bladder cancer decreased with increasing cumulative dose of AMOAB. We observed no similar relations between incidence rates of other study cancers and time since cohort entry. CONCLUSION: High incidence rates of bladder and prostate cancer soon after AMOAB initiation and a negative correlation between incidence and cumulative AMOAB dose suggest that protopathic bias is a more likely explanation for these findings than causality. (Protopathic bias in this context means patients' urinary symptoms prompted treatment with an AMOAB, but the symptoms were actually due to a cancer that was already present, although not yet diagnosed or not yet recorded.) To avoid unnecessary delays in the diagnosis of prostate and bladder cancer, physicians should consider these diseases in patients for whom treatment with AMOABs is indicated.

The feasibility of using multiple databases to study rare outcomes: the potential effect of long-acting beta agonists with inhaled corticosteroid therapy on asthma mortality.

PURPOSE: Long-acting beta agonists (LABAs) when used without concomitant inhaled corticosteroids (ICS) increase the risk of asthma-related deaths, but the effect on asthma-related death of LABA used in combination with ICS therapy is unknown. To address this question, we explored the feasibility of conducting an observational study using multiple US health care data sources. METHODS: Retrospective cohort study to evaluate the likelihood of getting an upper 95% confidence limit ≤1.4 for the asthma mortality rate ratio and ≤0.40 per 10 000 person-years for the mortality rate difference, assuming no effect of new use of combined LABA + ICS (versus non-LABA maintenance therapy) on asthma mortality. Ten research institutions executed centrally distributed analytic code based on a standard protocol using an extracted (2000-2010) persistent asthma cohort (asthma diagnosis and ≥4 asthma medications in 12 months). Pooled results were analyzed by the coordinating center. Asthma deaths were ascertained by linkage with the National Death Index. RESULTS: In a cohort of 994 627 persistent asthma patients (2.4 million person-years; 278 asthma deaths), probabilities of the upper 95% confidence limit for effect estimates being less than targeted values, assuming a null relation, were about 0.05. Modifications in cohort and exposure definitions increased exposed person-time and outcome events, but study size remained insufficient to attain study goals. CONCLUSIONS: Even with 10 data sources and a 10-year study period, the rarity of asthma deaths among patients using certain medications made it infeasible to study the association between combined LABA + ICS and asthma mortality with our targeted level of study precision. Copyright © 2016 John Wiley & Sons, Ltd.

Improving the Identification of Out-of-Hospital Sudden Cardiac Deaths in a General Practice Research Database.

BACKGROUND: The ascertainment of sudden cardiac death (SCD) in electronic health databases is challenging. OBJECTIVES: Our objective was to evaluate the applicability of the validated computer definition of SCD developed by Chung et al. in a retrospective study of SCD and domperidone exposure in the Clinical Practice Research Datalink (CPRD). METHODS: We assessed out-of-hospital SCD by applying the validated computer definition and linking data with Hospital Episode Statistics and death certificates. We developed a separate algorithm to identify end-of-life care in noninstitutionalized patients and excluded associated deaths from the analysis to address their misclassification as SCD. RESULTS: Of the 681,104 patients in the study cohort, 3444 were initially classified as out-of-hospital SCD. Next, 163 deaths were identified as expected deaths by our algorithm for end-of-life home care. After review of patient profiles, 162 were classified as expected deaths because of evidence that the patient received palliative or end-of-life care, but one was a false negative. The exclusion of such cases appreciably changed the odds ratio for current exposure to domperidone compared with non-use of study medications from 2.09 (95 % confidence interval [CI] 1.16-3.74) to 1.71 (95 % CI 0.92-3.18). A similar effect on the odds ratio was observed for current exposure to metoclopramide but not to proton pump inhibitors. CONCLUSIONS: Our algorithm to identify end-of-life care at home in the CPRD performed well, with only one false negative. The exclusion of misclassified cases of SCD reduced the magnitude of the odds ratios for SCD associated with domperidone and metoclopramide exposure by controlling protopathic bias.

Risk of Out-of-Hospital Sudden Cardiac Death in Users of Domperidone, Proton Pump Inhibitors, or Metoclopramide: A Population-Based Nested Case-Control Study.

INTRODUCTION: Epidemiological studies have linked domperidone use with serious cardiac arrhythmias, including sudden cardiac death, but data on age, dose, and duration of use are limited. OBJECTIVES: The aim of this study was to assess the risk of out-of-hospital sudden cardiac death associated with domperidone use versus proton pump inhibitors (PPIs), metoclopramide, or non-use of all three medications, and to evaluate the risk of sudden cardiac death in relation to age and domperidone dose. METHODS: This was a population-based case-control study nested in a cohort of subjects aged ≥2 years in the Clinical Practice Research Datalink with one or more prescriptions for domperidone, any PPI, or metoclopramide from 2005 to 2011. Out-of-hospital sudden cardiac death was assessed by linkage with Hospital Episode Statistics and death certificates. Controls were matched on age, sex, and medical practice. The risk of sudden cardiac death in domperidone users versus risk in users of PPIs or metoclopramide was evaluated with multivariable conditional logistic regression; case-crossover analysis addressed possible residual confounding. RESULTS: From the study cohort (n = 681,104), 3239 sudden cardiac death cases were matched to 12,572 controls. The adjusted odds ratio (95 % confidence interval) for sudden cardiac death with current use of domperidone alone was 1.71 (0.92-3.18) versus non-use of study medications, 1.26 (0.68-2.34) versus current PPI use, and 0.40 (0.17-0.94) current metoclopramide use. The adjusted odds ratio (95 % confidence interval) relative to exposure to no study drug for domperidone >30 mg/day (eight cases, five controls) was 3.20 (0.59-17.3) and 1.65 (0.89-3.07) for age ≥61 years (27 cases, 49 controls). The odds ratio (95 % confidence interval) was 3.17 (1.72-5.83) for within-person periods of domperidone use versus non-use in the case-crossover analysis. CONCLUSIONS: Compared with non-use of any study drug, current domperidone use was associated with sudden cardiac death in nested case-control and case-crossover analyses, with a suggestion of higher risk in older persons and users of higher daily doses.

Risk of acute liver injury associated with the use of moxifloxacin and other oral antimicrobials: a retrospective, population-based cohort study.

STUDY OBJECTIVE: To estimate the incidence and relative risk of a hospitalization or emergency visit for noninfectious liver injury in users of eight oral antimicrobials-amoxicillin, amoxicillin-clavulanic acid, clarithromycin, cefuroxime, doxycycline, levofloxacin, moxifloxacin, telithromycin-compared with nonusers of these antimicrobials. DESIGN: Retrospective, observational cohort study with a nested case-control analysis. DATA SOURCE: HealthCore Integrated Research Database. PATIENTS: Adults with continuous health plan enrollment for at least 6 months before study entry who had a new dispensing of a study antimicrobial between July 1, 2001, and March 31, 2009. Cases had diagnoses indicating noninfectious liver injury during follow-up. To control for potentially confounding risk factors, 10 controls at risk for liver injury during follow-up were matched to each case by age, sex, and event date (liver injury date of the case), and analyses were adjusted for medical history, concomitant drugs, and health care service use. MEASUREMENTS AND MAIN RESULTS: Two physician reviewers (blind to exposure) validated the cases. Among 1.3 million antimicrobial users, we identified 607 cases of liver injury, including 82 cases of severe hepatocellular injury and 11 cases of liver failure. Liver injury incidence in nonusers of study antimicrobials was 35/100,000 person-years (95% confidence interval [CI] 29-42/100,000 person-years). For valid cases, the adjusted relative risk among current users of multiple antimicrobials was 3.2 (95% CI 1.6-6.7). Levofloxacin had the highest relative risk for current single use (3.2, 95% CI 1.8-5.8). Relative risks were also elevated for amoxicillin-clavulanic acid (2.5, 95% CI 1.3-5.0), doxycycline (2.5, 95% CI 1.2-5.2), moxifloxacin (2.3, 95% CI 1.1-4.7), and amoxicillin (2.3, 95% CI 1.1-4.7). CONCLUSION: The results support a comparatively high adjusted relative risk of liver injury among patients exposed concurrently to multiple antimicrobials and modest elevations in the risk for several antimicrobials used alone; however, we found little evidence of any strong effect of commonly used antimicrobials on the risk of liver injury.

Validation of acute liver injury cases in a population-based cohort study of oral antimicrobial users.

We conducted a cohort study of acute, noninfectious liver injury among oral antimicrobial users. Potential cases were identified in the HealthCore Integrated Research Database (HIRD(SM)) population between July 1, 2001, and March 31, 2009, using ICD-9-CM codes primarily for acute and subacute necrosis of the liver, hepatic coma, and unspecified hepatitis. Liver test results were used to confirm case status according to published criteria. Two physician reviewers experienced in studying acute liver injury (blinded to study drug exposures) evaluated data abstracted from hospital and emergency department records to validate potential cases. Of 715 potential cases having claims associated with any of the primary screening codes, 312 (44%) were valid cases, 108 (15%) were not cases, and 295 (41%) were of uncertain status (records inadequate for validation). Among potential cases with adequate medical records, the PPV for presence of any of the primary codes was 74% (95% CI, 70%-78%). The highest PPV for a single code was for acute and subacute necrosis of the liver (84%; 95% CI, 77%-90%). Evaluation of cases of noninfectious liver injury using hospital and emergency department medical records continues to represent the preferred approach in studies using insurance claims data.

Use of depot medroxyprogesterone acetate contraception and incidence of bone fracture.

OBJECTIVE: Depot medroxyprogesterone acetate (DMPA) reversibly reduces bone mineral density. To estimate the extent to which DMPA might increase fracture risk, we undertook a retrospective cohort study of fractures in DMPA users and users of non-DMPA contraceptives, using the General Practice Research Database. METHODS: Eligible women were aged younger than 50 years at the qualifying first contraceptive prescription. The DMPA users were classified by DMPA exposure (cumulative and time of last dose) based on prescription records. All incident fractures were included; fracture incidence and risk factors before starting contraceptive use (DMPA or other) also were estimated. RESULTS: We identified 11,822 fractures in 312,395 women during 1,722,356 person-years of follow-up. Before contraceptive use started, DMPA users had higher fracture risk than nonusers (incidence rate ratio 1.28, 95% confidence interval [CI] 1.07-1.53). After DMPA started, crude fracture incidence was 9.1 per 1,000 person-years for DMPA users and 7.3 for nonusers (crude incidence rate ratio 1.23, 95% CI 1.16-1.30). Fracture risk in DMPA users did not increase after starting DMPA (incidence rate ratio after or before 1.08, 95% CI 0.92-1.26). There was little confounding by age or other factors that could be measured. Fracture incidence was 9.4 per 1,000 person-years in low-exposure DMPA users, and 7.8 per 1,000 in high-exposure DMPA users. The DMPA users had higher fracture risk than nonusers at the start of contraceptive use, with no discernible induction period. CONCLUSION: Although DMPA users experienced more fractures than nonusers, this association may be the result of confounding by a pre-existing higher risk for fractures in women who chose DMPA for contraception.

A US postmarketing evaluation of the frequency and safety of live attenuated influenza vaccine use in nonrecommended children younger than 5 years: 2009-2010 season.

The 2007 US approval for use of Ann Arbor strain live attenuated influenza vaccine (LAIV) in children aged 24 through 59 months included precautions against use in (1) children <24 months and children aged 24 through 59 months with (2) asthma, (3) recurrent wheezing, and (4) altered immunocompetence. Results from the third season (2009-2010) of a 3-year study postmarketing commitment to monitor LAIV vaccination rates and frequency of hospitalizations or emergency department visits within 42 days after LAIV are reported here. As in the first 2 seasons, LAIV usage in cohorts 1, 2, and 4 were low relative to those in LAIV-recommended populations. The only numerically increased risk observed was for respiratory events in children aged <24 months administered LAIV, compared to those administered trivalent inactivated influenza vaccine (TIV). The number of children vaccinated with LAIV was small and precluded precise quantification of rare event.

Predicting protein function by multi-label correlated semi-supervised learning.

Assigning biological functions to uncharacterized proteins is a fundamental problem in the postgenomic era. The increasing availability of large amounts of data on protein-protein interactions (PPIs) has led to the emergence of a considerable number of computational methods for determining protein function in the context of a network. These algorithms, however, treat each functional class in isolation and thereby often suffer from the difficulty of the scarcity of labeled data. In reality, different functional classes are naturally dependent on one another. We propose a new algorithm, Multi-label Correlated Semi-supervised Learning (MCSL), to incorporate the intrinsic correlations among functional classes into protein function prediction by leveraging the relationships provided by the PPI network and the functional class network. The guiding intuition is that the classification function should be sufficiently smooth on subgraphs where the respective topologies of these two networks are a good match. We encode this intuition as regularized learning with intraclass and interclass consistency, which can be understood as an extension of the graph-based learning with local and global consistency (LGC) method. Cross validation on the yeast proteome illustrates that MCSL consistently outperforms several state-of-the-art methods. Most notably, it effectively overcomes the problem associated with scarcity of label data. The supplementary files are freely available at http://sites.google.com/site/csaijiang/MCSL.

A postmarketing evaluation of the frequency of use and safety of live attenuated influenza vaccine use in nonrecommended children younger than 5 years.

The 2007 US approval for use of live attenuated influenza vaccine (LAIV) in children aged 24-59 months included precautions against use in (1) children <24 months and children aged 24-59 months with (2) asthma, (3) recurrent wheezing, and (4) altered immunocompetence. A postmarketing commitment was initiated to monitor LAIV use and the frequency of select safety outcomes in these cohorts. Vaccination rates and the frequency of hospitalizations or emergency department visits within 42 days after LAIV and trivalent inactivated influenza vaccine (TIV) administration were estimated from 2007 to 2009 claims data from a health insurance database. Rates of LAIV use per 10,000 child-days among cohorts 1, 2, and 4 were low relative to rates among the LAIV-recommended population (2007-2008; 0.03-0.78 vs. 1.32, 2008-2009; 0.08-3.26 vs. 5.94). However, rates of LAIV use per 10,000 child-days in cohort 3 were similar to rates among the LAIV-recommended population (2007-2008; 1.55 vs. 1.32, 2008-2009; 5.01 vs. 5.94). The rate of emergency department visits/hospitalizations within 42 days of vaccination with LAIV was the same as or less than the rate within 42 days of vaccination with TIV. Less restricted LAIV use in children with past wheezing may be related to the broad definition of recurrent wheezing used in national guidelines and the current study. In the small number of nonrecommended children receiving LAIV, no safety signals were identified.