Predicting Seasonal Influenza Hospitalizations Using an Ensemble Super Learner: A Simulation Study.

Accurate forecasts can inform response to outbreaks. Most efforts in influenza forecasting have focused on predicting influenza-like activity, with fewer on influenza-related hospitalizations. We conducted a simulation study to evaluate a super learner's predictions of 3 seasonal measures of influenza hospitalizations in the United States: peak hospitalization rate, peak hospitalization week, and cumulative hospitalization rate. We trained an ensemble machine learning algorithm on 15,000 simulated hospitalization curves and generated weekly predictions. We compared the performance of the ensemble (weighted combination of predictions from multiple prediction algorithms), the best-performing individual prediction algorithm, and a naive prediction (median of a simulated outcome distribution). Ensemble predictions performed similarly to the naive predictions early in the season but consistently improved as the season progressed for all prediction targets. The best-performing prediction algorithm in each week typically had similar predictive accuracy compared with the ensemble, but the specific prediction algorithm selected varied by week. An ensemble super learner improved predictions of influenza-related hospitalizations, relative to a naive prediction. Future work should examine the super learner's performance using additional empirical data on influenza-related predictors (e.g., influenza-like illness). The algorithm should also be tailored to produce prospective probabilistic forecasts of selected prediction targets.

What explains racial/ethnic inequities in the uptake of differentiated influenza vaccines?

We investigated the role of individual, community and vaccinator characteristics in mediating racial/ethnic disparities in the uptake of differentiated influenza vaccines (DIVs; including high-dose, adjuvanted, recombinant and cell-based vaccines). We included privately-insured (commercial and Medicare Advantage) ≥65 years-old community-dwelling health plan beneficiaries in the US with >1 year of continuous coverage and who received ≥1 influenza vaccine during the study period (July 2014-June 2018). Of 2.8 million distinct vaccination claims, 60% were for DIVs; lower if received in physician offices (49%) compared to pharmacies/facilities (74%). Among those vaccinated in physician offices, non-whites had lower odds of receiving a DIV if they lived in a non-minority county (0.77;95%CI 0.75-0.80) and even lower odds if they lived in a minority county (0.62;0.60-0.63). Differences in education, household income, medical history, community and vaccinator characteristics did not fully explain the disparities. Similar patterns emerged for vaccinations in pharmacies/facilities, although disparities disappeared altogether after controlling for socio-economic and vaccinator characteristics. When vaccinated in physician offices, minority county residents were less likely to receive a DIV, especially for non-whites (0.72;0.67-0.78). These disparities disappeared for whites, but not for non-whites, after controlling for community and vaccinator characteristics. We found an alarming level of inequity in DIV vaccine uptake among fully insured older adults that could not be fully explained by differences in sociodemographic, medical, community, and vaccinator characteristics. New strategies are urgently needed to address these inequities.

Correlation of long-term care facility vaccination practices between seasons and resident types.

BACKGROUND: Influenza vaccination varies widely across long-term care facilities (LTCFs) due to staff behaviors, LTCF practices, and patient factors. It is unclear how seasonal LTCF vaccination varies between cohabitating but distinct short-stay and long-stay residents. Thus, we assessed the correlation of LTCF vaccination between these populations and across seasons. METHODS: The study design is a national retrospective cohort using Medicare and Minimum Data Set (MDS) data. Participants include U.S. LTCFs. Short-stay and long-stay Medicare-enrolled residents age ≥ 65 in U.S. LTCFs from a source population of residents during October 1st-March 31st in 2013-2014 (3,042,881 residents; 15,683 LTCFs) and 2014-2015 (3,143,174, residents; 15,667 LTCFs). MDS-assessed influenza vaccination was the outcome. Pearson correlation coefficients were estimated to assess seasonal correlations between short-stay and long-stay resident vaccination within LTCFs. RESULTS: The median proportion of short-stay residents vaccinated across LTCFs was 70.4% (IQR, 50.0-82.7%) in 2013-2014 and 69.6% (IQR, 50.0-81.6%) in 2014-2015. The median proportion of long-stay residents vaccinated across LTCFs was 85.5% (IQR, 78.0-90.9%) in 2013-2014 and 84.6% (IQR, 76.6-90.3%) in 2014-2015. Within LTCFs, there was a moderate correlation between short-stay and long-stay vaccination in 2013-2014 (r = 0.50, 95%CI: 0.49-0.51) and 2014-2015 (r = 0.53, 95%CI: 0.51-0.54). Across seasons, there was a moderate correlation for LTCFs with short-stay residents (r = 0.54, 95%CI: 0.53-0.55) and a strong correlation for those with long-stay residents (r = 0.68, 95%CI: 0.67-0.69). CONCLUSIONS: In LTCFs with inconsistent influenza vaccination across seasons or between populations, targeted vaccination protocols for all residents, regardless of stay type, may improve successful vaccination in this vulnerable patient population.

Persistence of Racial Inequities in Receipt of Influenza Vaccination Among Nursing Home Residents in the United States.

BACKGROUND: We sought to determine if racial differences in influenza vaccination among nursing home (NH) residents during the 2008-2009 influenza season persisted in 2018-2019. METHODS: We conducted a cross-sectional study of NHs certified by the Centers for Medicare & Medicaid Services during the 2018-2019 influenza season in US states with ≥1% Black NH residents and a White-Black gap in influenza vaccination of NH residents (N = 2 233 392) of at least 1 percentage point (N = 40 states). NH residents during 1 October 2018 through 31 March 2019 aged ≥18 years and self-identified as being of Black or White race were included. Residents' influenza vaccination status (vaccinated, refused, and not offered) was assessed. Multilevel modeling was used to estimate facility-level vaccination status and inequities by state. RESULTS: The White-Black gap in influenza vaccination was 9.9 percentage points. In adjusted analyses, racial inequities in vaccination were more prominent at the facility level than at the state level. Black residents disproportionately lived in NHs that had a majority of Blacks residents, which generally had the lowest vaccination. Inequities were most concentrated in the Midwestern region, also the most segregated. Not being offered the vaccine was negligible in absolute percentage points between White residents (2.6%) and Black residents (4.8%), whereas refusals were higher among Black (28.7%) than White residents (21.0%). CONCLUSIONS: The increase in the White-Black vaccination gap among NH residents is occurring at the facility level in more states, especially those with the most segregation.

On the Causal Interpretation of Rate-Change Methods: The Prior Event Rate Ratio and Rate Difference.

A growing number of studies use data before and after treatment initiation in groups exposed to different treatment strategies to estimate "causal effects" using a ratio measure called the prior event rate ratio (PERR). Here, we offer a causal interpretation for PERR and its additive scale analog, the prior event rate difference (PERD). We show that causal interpretation of these measures requires untestable rate-change assumptions about the relationship between 1) the change of the counterfactual rate before and after treatment initiation in the treated group under hypothetical intervention to implement the control strategy; and 2) the change of the factual rate before and after treatment initiation in the control group. The rate-change assumption is on the multiplicative scale for PERR but on the additive scale for PERD; the 2 assumptions hold simultaneously under testable, but unlikely, conditions. Even if investigators can pick the most appropriate scale, the relevant rate-change assumption might not hold exactly, so we describe sensitivity analysis methods to examine how assumption violations of different magnitudes would affect study results. We illustrate the methods using data from a published study of proton pump inhibitors and pneumonia.

Interventions on health care providers to improve seasonal influenza vaccination rates among patients: a systematic review and meta-analysis of the evidence since 2000.

BACKGROUND: Seasonal influenza vaccination (SIV) rates remain suboptimal in many populations, even in those with universal SIV. OBJECTIVE: To summarize the evidence on interventions on health care providers (physicians/nurses/pharmacists) to increase SIV rates. METHODS: We systematically searched/selected full-text English publications from January 2000 to July 2019 (PROSPERO-CRD42019147199). Our outcome was the difference in SIV rates between patients in intervention and non-intervention groups. We calculated pooled difference using an inverse variance, random-effects model. RESULTS: We included 39 studies from 8370 retrieved citations. Compared with no intervention, team-based training/education of physicians significantly increased SIV rates in adult patients: 20.1% [7.5-32.7%; I2 = 0%; two randomized controlled trials (RCTs)] and 13.4% [8.6-18.1%; I2 = 0%; two non-randomized intervention studies (NRS)]. A smaller increase was observed in paediatric patients: 7% (0.1-14%; I2 = 0%; two NRS), and in adult patients with team-based training/education of physicians and nurses together: 0.9% (0.2-1.5%; I2 = 30.6%; four NRS). One-off provision of guidelines/information to physicians, and to both physicians and nurses, increased SIV rates in adult patients: 23.8% (15.7-31.8%; I2 = 45.8%; three NRS) and paediatric patients: 24% (8.1-39.9%; I2 = 0%; two NRS), respectively. Use of reminders (prompts) by physicians and nurses slightly increased SIV rates in paediatric patients: 2.3% (0.5-4.2%; I2 = 0%; two RCTs). A larger increase was observed in adult patients: 18.5% (14.8-22.1%; I2 = 0%; two NRS). Evidence from both RCTs and NRS showed significant increases in SIV rates with varied combinations of interventions. CONCLUSIONS: Limited evidence suggests various forms of physicians' and nurses' education and use of reminders may be effective for increasing SIV rates among patients.

Impact of a peer comparison intervention on seasonal influenza vaccine uptake in community pharmacy: A national cluster randomized study.

BACKGROUND: Seasonal influenza vaccine (SIV) uptake in the United States remains suboptimal, requiring new and innovative strategies. OBJECTIVE: To evaluate the impact of a behavioral peer comparison (PC) intervention on SIV uptake in community pharmacies across the United States. METHODS: A cluster randomized study was conducted across a national network of Walmart community pharmacies (> 4500 sites) during the 2019-2020 influenza season. The clusters consisted of 416 markets, each containing an average of 11 pharmacies. All pharmacies in a market were randomly assigned to either no intervention or the PC intervention, a software-delivered communication informing on-site staff, including pharmacists and pharmacy technicians, of their pharmacy's weekly performance, measured as SIV doses administered, compared with that of peer pharmacies within their market. The outcome was the pharmacy-level cumulative SIV doses administered during the intervention period (September 1, 2019,-February 29, 2020). Linear regression models were used to estimate the PC impact, with multiway cluster-robust SEs estimated by market and state. RESULTS: A total of 4589 pharmacies were enrolled in the study, with 2297 (50.1%) randomized to the control group and 2292 (49.9%) randomized to the PC intervention group. Overall, compared with the control pharmacies, the PC pharmacies administered 3.7% (95% CI -0.3% to 7.9%) additional SIV doses. Among large-format pharmacies, the PC pharmacies administered 4.1% (95% CI 0.1%-8.3%) additional SIV doses compared with the controls. Historically low-performing large-format PC pharmacies administered 6.1% (95% CI 0.5%-11.9%) additional SIV doses compared with the controls. No statistically significant treatment effects were observed among small-format pharmacies. CONCLUSION: Our findings demonstrate that PCs can improve SIV uptake among large-format community pharmacies, with historically low-performing pharmacies potentially exhibiting the greatest relative impact. Wide-scale implementation of PCs in community pharmacies may help to further improve SIV uptake in these settings.

Geographic Variation in Pneumonia and Influenza in Long-Term Care Facilities: A National Study.

There is large county-level geographic variation in pneumonia and influenza hospitalizations among short-stay and long-stay long-term care facility residents in the United States. Long-term care facilities in counties in the Southern and Midwestern regions had the highest rates of pneumonia and influenza from 2013 to 2015. Future research should identify reasons for these geographic differences.

Assessing the prior event rate ratio method via probabilistic bias analysis on a Bayesian network.

BACKGROUND: Unmeasured confounders are commonplace in observational studies conducted using real-world data. Prior event rate ratio (PERR) adjustment is a technique shown to perform well in addressing such confounding. However, it has been demonstrated that, in some circumstances, the PERR method actually increases rather than decreases bias. In this work, we seek to better understand the robustness of PERR adjustment. METHODS: We begin with a Bayesian network representation of a generalized observational study, which is subject to unmeasured confounding. Previous work evaluating PERR performance used Monte Carlo simulation to calculate joint probabilities of interest within the study population. Here, we instead use a Bayesian networks framework. RESULTS: Using this streamlined analytic approach, we are able to conduct probabilistic bias analysis (PBA) using large numbers of combinations of parameters and thus obtain a comprehensive picture of PERR performance. We apply our methodology to a recent study that used the PERR in evaluating elderly-specific high-dose (HD) influenza vaccine in the US Veterans Affairs population. That study obtained an HD relative effectiveness of 25% (95% CI: 2%-43%) against influenza- and pneumonia-associated hospitalization, relative to standard-dose influenza vaccine. In this instance, we find that the PERR-adjusted result is more like to underestimate rather than to overestimate the relative effectiveness of the intervention. CONCLUSIONS: Although the PERR is a powerful tool for mitigating the effects of unmeasured confounders, it is not infallible. Here, we develop some general guidance for when a PERR approach is appropriate and when PBA is a safer option.

Quantifying the annual incidence and underestimation of seasonal influenza: A modelling approach.

BACKGROUND: Seasonal influenza poses a significant public health and economic burden, associated with the outcome of infection and resulting complications. The true burden of the disease is difficult to capture due to the wide range of presentation, from asymptomatic cases to non-respiratory complications such as cardiovascular events, and its seasonal variability. An understanding of the magnitude of the true annual incidence of influenza is important to support prevention and control policy development and to evaluate the impact of preventative measures such as vaccination. METHODS: We use a dynamic disease transmission model, laboratory-confirmed influenza surveillance data, and randomized-controlled trial (RCT) data to quantify the underestimation factor, expansion factor, and symptomatic influenza illnesses in the US and Canada during the 2011-2012 and 2012-2013 influenza seasons. RESULTS: Based on 2 case definitions, we estimate between 0.42-3.2% and 0.33-1.2% of symptomatic influenza illnesses were laboratory-confirmed in Canada during the 2011-2012 and 2012-2013 seasons, respectively. In the US, we estimate between 0.08-0.61% and 0.07-0.33% of symptomatic influenza illnesses were laboratory-confirmed in the 2011-2012 and 2012-2013 seasons, respectively. We estimated the symptomatic influenza illnesses in Canada to be 0.32-2.4 million in 2011-2012 and 1.8-8.2 million in 2012-2013. In the US, we estimate the number of symptomatic influenza illnesses to be 4.4-34 million in 2011-2012 and 23-102 million in 2012-2013. CONCLUSIONS: We illustrate that monitoring a representative group within a population may aid in effectively modelling the transmission of infectious diseases such as influenza. In particular, the utilization of RCTs in models may enhance the accuracy of epidemiological parameter estimation.

Risk factors for pneumonia and influenza hospitalizations in long-term care facility residents: a retrospective cohort study.

BACKGROUND: Older adults who reside in long-term care facilities (LTCFs) are at particularly high risk for infection, morbidity and mortality from pneumonia and influenza (P&I) compared to individuals of younger age and those living outside institutional settings. The risk factors for P&I hospitalizations that are specific to LTCFs remain poorly understood. Our objective was to evaluate the incidence of P&I hospitalization and associated person- and facility-level factors among post-acute (short-stay) and long-term (long-stay) care residents residing in LTCFs from 2013 to 2015. METHODS: In this retrospective cohort study, we used Medicare administrative claims linked to Minimum Data Set and LTCF-level data to identify short-stay (< 100 days, index = admission date) and long-stay (100+ days, index = day 100) residents who were followed from the index date until the first of hospitalization, LTCF discharge, Medicare disenrollment, or death. We measured incidence rates (IRs) for P&I hospitalization per 100,000 person-days, and estimated associations with baseline demographics, geriatric syndromes, clinical characteristics, and medication use using Cox regression models. RESULTS: We analyzed data from 1,118,054 short-stay and 593,443 long-stay residents. The crude 30-day IRs (95% CI) of hospitalizations with P&I in the principal position were 26.0 (25.4, 26.6) and 34.5 (33.6, 35.4) among short- and long-stay residents, respectively. The variables associated with P&I varied between short and long-stay residents, and common risk factors included: advanced age (85+ years), admission from an acute hospital, select cardiovascular and respiratory conditions, impaired functional status, and receipt of antibiotics or Beers criteria medications. Facility staffing and care quality measures were important risk factors among long-stay residents but not in short-stay residents. CONCLUSIONS: Short-stay residents had lower crude 30- and 90-day incidence rates of P&I hospitalizations than long-stay LTCF residents. Differences in risk factors for P&I between short- and long-stay populations suggest the importance of considering distinct profiles of post-acute and long-term care residents in infection prevention and control strategies in LTCFs. These findings can help clinicians target interventions to subgroups of LTCF residents at highest P&I risk.

Revisiting the epidemiology of pertussis in Canada, 1924-2015: a literature review, evidence synthesis, and modeling study.

BACKGROUND: Disease surveillance is central to the public health understanding of pertussis epidemiology. In Canada, public reporting practices have significantly changed over time, creating challenges in accurately characterizing pertussis epidemiology. Debate has emerged over whether pertussis resurged after the introduction of adsorbed pertussis vaccines (1981-1985), and if the incidence fell to its pre-1985 after the introduction of acellular pertussis vaccines (1997-1998). Here, we aim to assemble a unified picture of pertussis disease incidence in Canada. METHODS: Using publicly available pertussis surveillance reports, we collected, analyzed and presented Canadian pertussis data for the period (1924-2015), encompassing the pre-vaccine era, introduction of vaccine, changes to vaccine technology, and the introduction of booster doses. Information on age began to be reported since 1952, but age reporting practices (full, partial or no ages) have evolved over time, and varied across provinces/territories. For those cases reported without age each year, we impute an age distribution by assuming it follows that of the age-reported cases. RESULTS: Below the age of 20 years, the adjusted age-specific incidence from 1969 to 1988 is substantially higher than existing estimates. In children < 1 year, the incidence in some years was comparable to that during the 1988-1999 resurgence. CONCLUSIONS: The results presented here suggest that the surge in the average yearly incidence of pertussis that began in 1988 was weaker than previously inferred, and in contrary to the past findings, below age 5, the average yearly incidence of pertussis from 1999 to 2015 (when the incidence dropped again) has been lower than it was from 1969 to 1988.

High-dose influenza vaccination and mortality among predominantly male, white, senior veterans, United States, 2012/13 to 2014/15.

IntroductionIt is unclear whether high-dose influenza vaccine (HD) is more effective at reducing mortality among seniors.AimThis study aimed to evaluate the relative vaccine effectiveness (rVE) of HD. MethodsWe linked electronic medical record databases in the Veterans Health Administration (VHA) and Medicare administrative files to examine the rVE of HD vs standard-dose influenza vaccines (SD) in preventing influenza/pneumonia-associated and cardiorespiratory mortality among VHA-enrolled veterans 65 years or older during the 2012/13, 2013/14 and 2014/15 influenza seasons. A multivariable Cox proportional hazards model was performed on matched recipients of HD vs SD, based on vaccination time, location, age, sex, ethnicity and VHA priority level. ResultsAmong 569,552 person-seasons of observation, 207,574 (36%) were HD recipients and 361,978 (64%) were SD recipients, predominantly male (99%) and white (82%). Pooling findings from all three seasons, the adjusted rVE estimate of HD vs SD during the high influenza periods was 42% (95% confidence interval (CI): 24-59) against influenza/pneumonia-associated mortality and 27% (95% CI: 23-32) against cardiorespiratory mortality. Residual confounding was evident in both early and late influenza periods despite matching and multivariable adjustment. Excluding individuals with high 1-year predicted mortality at baseline reduced the residual confounding and yielded rVE of 36% (95% CI: 10-62) and 25% (95% CI: 12-38) against influenza/pneumonia-associated and cardiorespiratory mortality, respectively. These were confirmed by results from two-stage residual inclusion estimations.DiscussionThe HD was associated with a lower risk of influenza/pneumonia-associated and cardiorespiratory death in men during the high influenza period.

Relative Vaccine Effectiveness of High-Dose Versus Standard-Dose Influenza Vaccines Among Veterans Health Administration Patients.

Background: We examined whether a high-dose inactivated influenza vaccine was more efficacious in preventing hospitalizations than a standard-dose vaccine in the Veterans Health Administration (VHA) senior population. Methods: This study estimated the relative vaccine effectiveness (rVE) of high dose versus standard dose using a retrospective cohort of VHA patients 65 years of age or older in the 2015-2016 influenza season. To adjust for measured confounders, we matched each high-dose recipient with up to 4 standard-dose recipients vaccinated at the same location within a 2-week period and having 2 or more pre-existing medical comorbidities. We used the previous event rate ratio method (PERR), a type of difference-in-differences analysis, to adjust for unmeasured confounders. Results: We evaluated 104965 standard-dose and 125776 high-dose recipients; matching decreased the population to 49091 standard-dose and 24682 high-dose recipients. The matched, PERR-adjusted rVE was 25% (95% confidence interval [CI], 2%-43%) against influenza- or pneumonia-associated hospitalization, 7% (95% CI, -2% to 14%) against all-cause hospitalization, 14% (95% CI, -8% to 32%) against influenza- or pneumonia-associated outpatient visit, 5% (95% CI, 2%-8%) against all-cause outpatient visit, and 38% (95% CI, -5% to 65%) against laboratory-confirmed influenza. Conclusions: In protecting senior VHA patients against influenza- or pneumonia-associated hospitalization, a high-dose influenza vaccine is more effective than a standard-dose vaccine.

The Importance of Frailty in the Assessment of Influenza Vaccine Effectiveness Against Influenza-Related Hospitalization in Elderly People.

Background: Influenza is an important cause of morbidity and mortality among older adults. Even so, effectiveness of influenza vaccine for older adults has been reported to be lower than for younger adults, and the impact of frailty on vaccine effectiveness (VE) and outcomes is uncertain. We aimed to study VE against influenza hospitalization in older adults, focusing on the impact of frailty. Methods: We report VE of trivalent influenza vaccine (TIV) in people ≥65 years of age hospitalized during the 2011-2012 influenza season using a multicenter, prospective, test-negative case-control design. A validated frailty index (FI) was used to measure frailty. Results: Three hundred twenty cases and 564 controls (mean age, 80.6 and 78.7 years, respectively) were enrolled. Cases had higher baseline frailty than controls (P = .006). In the fully adjusted model, VE against influenza hospitalization was 58.0% (95% confidence interval [CI], 34.2%-73.2%). The contribution of frailty was important; adjusting for frailty alone yielded a VE estimate of 58.7% (95% CI, 36.2%-73.2%). VE was 77.6% among nonfrail older adults and declined as frailty increased. Conclusions: Despite commonly held views that VE is poor in older adults, we found that TIV provided good protection against influenza hospitalization in older adults who were not frail, though VE diminished as frailty increased. Clinical Trials Registration: NCT01517191.

Influenza vaccine effectiveness against influenza-related hospitalization during a season with mixed outbreaks of four influenza viruses: a test-negative case-control study in adults in Canada.

BACKGROUND: The Serious Outcomes Surveillance (SOS) Network was established to monitor seasonal influenza complications among hospitalized Canadian adults and to assess the effectiveness of influenza vaccination against severe outcomes. Here we report age- and strain-specific vaccine effectiveness (VE) in preventing severe outcomes during a season characterized by mixed outbreaks of four different influenza strains. METHODS: This prospective, multicentre, test-negative case-control study evaluated the VE of trivalent influenza vaccine (TIV) in the prevention of laboratory-confirmed influenza-hospitalization in adults aged ≥16 years (all adults) and adults aged 16-64 years (younger adults). The SOS Network identified hospitalized patients with diagnoses potentially attributable to influenza during the 2011/12 influenza season. Swabs collected at admission were tested by reverse transcriptase polymerase chain reaction (RT PCR) or viral culture to discriminate influenza cases (positive) from controls (negative). VE was calculated as 1-odds ratio (OR) of vaccination in cases versus controls × 100. RESULTS: Overall, in all adults, the unadjusted and adjusted VEs of TIV against influenza-hospitalization were 41.8% (95% Confidence Interval [CI]: 26.0, 54.3), and 42.8% (95% CI: 23.8, 57.0), respectively. In younger adults (16-64 years), the unadjusted and adjusted VEs of TIV against influenza-hospitalization were 35.8% (95% CI: 4.5, 56.8) and 33.2% (95% CI: -6.7, 58.2), respectively. In the all adults group, adjusted VE against influenza A/H1N1 was 72.5% (95% CI: 30.5, 89.1), against A/H3N2 was 86.1% (95% CI: 40.1, 96.8), against B/Victoria was 40.5% (95% CI: -28.9, 72.6), and against B/Yamagata was 32.3% (95% CI: -8.3, 57.7). The adjusted estimate of early season VE (from November 1 to March 11) was 54.4% (95% CI: 29.7-70.4), which was higher than late season (from March 11 to May 25) VE estimate (VE: 29.7%, 95% CI: -5.3, 53.1). CONCLUSIONS: These results suggest that TIV was highly effective against A viruses and moderately effective against B viruses during a mild season characterised by co-circulation of four influenza strains in Canada. Findings underscore the need to provide VE assessment by subtype/lineage as well as the timing of vaccination (early season vs late season) to accurately evaluate vaccine performance and thus guide public health decision-making. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT01517191. Registration was retrospective and the date of registration was January 17, 2012.

Cost-Effectiveness of Quadrivalent versus Trivalent Influenza Vaccine in the United States.

BACKGROUND: Designed to overcome influenza B mismatch, new quadrivalent influenza vaccines (QIVs) contain one additional B strain compared with trivalent influenza vaccines (TIVs). OBJECTIVE: To examine the expected public health impact, budget impact, and incremental cost-effectiveness of QIV versus TIV in the United States. METHODS: A dynamic transmission model was used to predict the annual incidence of influenza over the 20-year-period of 2014 to 2034 under either a TIV program or a QIV program. A decision tree model was interfaced with the transmission model to estimate the public health impact and the cost-effectiveness of replacing TIV with QIV from a societal perspective. Our models were informed by published data from the United States on influenza complication probabilities and relevant costs. The incremental vaccine price of QIV as compared with that of TIV was set at US $5.40 per dose. RESULTS: Over the next 20 years, replacing TIV with QIV may reduce the number of influenza B cases by 27.2% (16.0 million cases), resulting in the prevention of 137,600 hospitalizations and 16,100 deaths and a gain of 212,000 quality-adjusted life-years (QALYs). The net societal budget impact would be US $5.8 billion and the incremental cost-effectiveness ratio US $27,411/QALY gained. In the probabilistic sensitivity analysis, 100% and 96.5% of the simulations fell below US $100,000/QALY and US $50,000/QALY, respectively. CONCLUSIONS: Introducing QIV into the US immunization program may prevent a substantial number of hospitalizations and deaths. QIV is also expected to be a cost-effective alternative option to TIV.

Cost of hospital management of Clostridium difficile infection in United States-a meta-analysis and modelling study.

BACKGROUND: Clostridium difficile infection (CDI) is the leading cause of infectious nosocomial diarrhoea but the economic costs of CDI on healthcare systems in the US remain uncertain. METHODS: We conducted a systematic search for published studies investigating the direct medical cost associated with CDI hospital management in the past 10 years (2005-2015) and included 42 studies to the final data analysis to estimate the financial impact of CDI in the US. We also conducted a meta-analysis of all costs using Monte Carlo simulation. RESULTS: The average cost for CDI case management and average CDI-attributable costs per case were $42,316 (90 % CI: $39,886, $44,765) and $21,448 (90 % CI: $21,152, $21,744) in 2015 US dollars. Hospital-onset CDI-attributable cost per case was $34,157 (90 % CI: $33,134, $35,180), which was 1.5 times the cost of community-onset CDI ($20,095 [90 % CI: $4991, $35,204]). The average and incremental length of stay (LOS) for CDI inpatient treatment were 11.1 (90 % CI: 8.7-13.6) and 9.7 (90 % CI: 9.6-9.8) days respectively. Total annual CDI-attributable cost in the US is estimated US$6.3 (Range: $1.9-$7.0) billion. Total annual CDI hospital management required nearly 2.4 million days of inpatient stay. CONCLUSIONS: This review indicates that CDI places a significant financial burden on the US healthcare system. This review adds strong evidence to aid policy-making on adequate resource allocation to CDI prevention and treatment in the US. Future studies should focus on recurrent CDI, CDI in long-term care facilities and persons with comorbidities and indirect cost from a societal perspective. Health-economic studies for CDI preventive intervention are needed.

Cost-effectiveness evaluation of quadrivalent influenza vaccines for seasonal influenza prevention: a dynamic modeling study of Canada and the United Kingdom.

BACKGROUND: The adoption of quadrivalent influenza vaccine (QIV) to replace trivalent influenza vaccine (TIV) in immunization programs is growing worldwide, thus helping to address the problem of influenza B lineage mismatch. However, the price per dose of QIV is higher than that of TIV. In such circumstances, cost-effectiveness analyses provide important and relevant information to inform national health recommendations and implementation decisions. This analysis assessed potential vaccine impacts and cost-effectiveness of a country-wide switch from TIV to QIV, in Canada and the UK, from a third-party payer perspective. METHODS: An age-stratified, dynamic four-strain transmission model which incorporates strain interaction, transmission-rate seasonality and age-specific mixing in the population was used. Model input data were obtained from published literature and online databases. In Canada, we evaluated a switch from TIV to QIV in the entire population. For the UK, we considered two strategies: Children aged 2-17 years who receive the live-attenuated influenza vaccine (LAIV) switch to the quadrivalent formulation (QLAIV), while individuals aged > 18 years switch from TIV to QIV. Two different vaccination uptake scenarios in children (UK1 and UK2, which differ in the vaccine uptake level) were considered. Health and cost outcomes for both vaccination strategies, and the cost-effectiveness of switching from TIV/LAIV to QIV/QLAIV, were estimated from the payer perspective. For Canada and the UK, cost and outcomes were discounted using 5 % and 3.5 % per year, respectively. RESULTS: Overall, in an average influenza season, our model predicts that a nationwide switch from TIV to QIV would prevent 4.6 % influenza cases, 4.9 % general practitioner (GP) visits, 5.7 % each of emergency room (ER) visits and hospitalizations, and 6.8 % deaths in Canada. In the UK (UK1/UK2), implementing QIV would prevent 1.4 %/1.8 % of influenza cases, 1.6 %/2.0 % each of GP and ER visits, 1.5 %/1.9 % of hospitalizations and 4.3 %/4.9 % of deaths. Discounted incremental cost-utility ratios of $7,961 and £7,989/£7,234 per quality-adjusted life-year (QALY) gained are estimated for Canada and the UK (UK1/UK2), both of which are well within their respective cost-effectiveness threshold values. CONCLUSIONS: Switching from TIV to QIV is expected to be a cost-effective strategy to further reduce the burden of influenza in both countries.

Natural attack rate of influenza in unvaccinated children and adults: a meta-regression analysis.

BACKGROUND: The natural (i.e. unvaccinated population) attack rate of an infectious disease is an important parameter required for understanding disease transmission. As such, it is an input parameter in infectious disease mathematical models. Influenza is an infectious disease that poses a major health concern worldwide and the natural attack rate of this disease is crucial in determining the effectiveness and cost-effectiveness of public health interventions and informing surveillance program design. We estimated age-stratified, strain-specific natural attack rates of laboratory-confirmed influenza in unvaccinated individuals. METHODS: Utilizing an existing systematic review, we calculated the attack rates in the trial placebo arms using a random effects model and a meta-regression analysis (GSK study identifier: 117102). RESULTS: This post-hoc analysis included 34 RCTs (Randomized Control Trials) contributing to 47 influenza seasons from 1970 to 2009. Meta-regression analyses showed that age and type of influenza were important covariates. The attack rates (95% CI (Confidence Interval)) in adults for all influenza, type A and type B were 3.50% (2.30%, 4.60%), 2.32% (1.47%, 3.17%) and 0.59% (0.28%, 0.91%) respectively. For children, they were 15.20% (11.40%, 18.90%), 12.27% (8.56%, 15.97%) and 5.50% (3.49%, 7.51%) respectively. CONCLUSIONS: This analysis demonstrated that unvaccinated children have considerably higher exposure risk than adults and influenza A can cause more disease than influenza B. Moreover, a higher ratio of influenza B:A in children than adults was observed. This study provides a new, stratified and up to-date natural attack rates that can be used in influenza infectious disease models and are consistent with previous published work in the field.