Impact of neuraminidase inhibitor treatment on hospitalisation in patients with pandemic influenza: a multi-method modelling approach

Venkatesan, Sudhir (2019) Impact of neuraminidase inhibitor treatment on hospitalisation in patients with pandemic influenza: a multi-method modelling approach. PhD thesis, University of Nottingham.

[img] PDF (Thesis - as examined) - Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (2MB)

Abstract

Background: Influenza pandemics are public health emergencies that can have potentially serious consequences. The 2009-10 influenza A(H1N1) pandemic was the first pandemic of the 21st century, and neuraminidase inhibitor (NAI) treatment was an important component of pandemic influenza mitigation plans around the world. While NAIs were licensed on the basis of reducing duration of illness in infected patients, their effectiveness in reducing complications, particularly in patients with pandemic influenza is less clear. Evaluating the impact of NAI treatment on outcomes of public health interest in patients with pandemic influenza is, therefore, crucial to informing future pandemic influenza preparedness plans. The main aim of this thesis was to evaluate the impact of NAI treatment on hospitalisation in patients with pandemic influenza.

Methods: This thesis used an interdisciplinary approach to address the thesis aim. The first part of the thesis used an Individual Participant Data (IPD) meta-analysis of clinical data to derive NAI effectiveness estimates, and the second part used computational and health economic modelling methods to extrapolate the NAI effectiveness estimates to assess public health impact.

A total of five studies were performed to address the thesis aim. In the first two studies, IPD meta-analyses were performed using a mixed-effects logistic regression, and mixed-effects negative binomial regression models, respectively, to estimate the effectiveness of community-based NAI treatment on the risk of subsequent hospital admission (Study 1) the effectiveness of in-hospital NAI treatment on the risk of subsequent length of hospital stay (LoS) (Study 2). In the third study, a novel behavioural model was developed (based on the COM-B model of behaviour change) and integrated with an agent-based epidemic model (ABM) to explore the impact of potential public health interventions on improving NAI treatment collection rates in patients (not at high-risk) during a pandemic. In the fourth study, this coupled behavioural-epidemic model was extended to include a discrete event simulation (DES) to assess the impact of community-based NAI treatment on hospital admissions, and NAI treatment given on hospital admission on hospital bed occupancy during a pandemic scenario. In the final study, a health economic decision tree model was used to evaluate the cost-effectiveness of community-based NAI treatment in a range of pandemic influenza scenarios.

Results: Using a global pooled dataset with clinical data on patients with laboratory-confirmed and clinically suspected A(H1N1)pdm09, the first study found that in a population of patients with severe influenza, community-based NAI treatment was associated with a reduction in the likelihood of hospital admission, compared to no NAI treatment [odds ratio: 0.24; 95% confidence interval (95% CI): 0.20-0.30). The second study found that an NAI ‘treatment-at-admission’ policy would reduce the subsequent LoS by 19% (incidence rate ratio: 0.81; 95% CI: 0.78-0.85), when compared to later/no NAI treatment. Using an empirically validated behavioural model, in the third study, a coupled behavioural-epidemic ABM indicated that public health interventions geared towards improving people’s ‘Capability’ and ‘Opportunity’ are likely to improve NAI collection rates during an influenza pandemic. Study 4 found that public health interventions that seek to improve people’s ‘Capability’ to collect NAIs could in turn result in a reduction in peak hospitalisation during a pandemic by 37.5% (through improved NAI collection rate). Further, when improved community-based NAI treatment collection was combined with a ‘treat-at-admission’ (with NAI) policy for hospitalised patients, peak hospital bed occupancy could be reduced by up to 41%, when compared to no NAI treatment. Finally, the decision tree model suggested that, across pandemic scenarios, community-based NAI treatment would be cost-saving for patients at high-risk of complications from pandemic influenza. Community-based NAI treatment was only found to be cost-saving for the whole population in higher severity influenza pandemics.

Conclusions: When interpreted in the context of the limited previous evidence on this topic, the findings across studies indicate that NAI treatment, both community-based and in-hospital, are beneficial to high-risk patients with severe influenza, and also tend to be cost-saving in this group of patients. A combination NAI treatment strategy that combines improved NAI treatment compliance and a ‘treat-at-admission’ strategy could result in significantly reduced hospital bed occupancy, particularly during severe influenza pandemics. The findings of this thesis support current advice on pandemic influenza treatment and shed valuable insight for future pandemic preparedness planning.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Siebers, Peer-Olaf
Nguyen-Van-Tam, Jonathan
Keywords: Pandemic influenza; Neuraminidase inhibitors; Agent-based modelling; IPD meta-analysis; Social simulation
Subjects: W Medicine and related subjects (NLM Classification) > WC Communicable diseases
Faculties/Schools: UK Campuses > Faculty of Medicine and Health Sciences > School of Medicine
Item ID: 56626
Depositing User: Venkatesan, Sudhir
Date Deposited: 11 Sep 2019 09:32
Last Modified: 31 May 2020 04:30
URI: http://eprints.nottingham.ac.uk/id/eprint/56626

Actions (Archive Staff Only)

Edit View Edit View