Timing of vaccination campaigns against pandemic influenza in a population dynamical model of Vancouver, Canada

Timing of vaccination campaigns against pandemic influenza in a population dynamical model of Vancouver, Canada Jessica M. Conway 1,2, Rafael Meza 2, Bahman Davoudi-Dehagi 2, Ashleigh Tuite 3, Babak Pourbohloul 2, and the CanPan Team 1 Department of Mathematics University of British Columbia 2 Division of Mathematical Modeling University of British Columbia Centre for Disease Control 3 Dalla Lana School of Public Health University of Toronto January 11, 211 Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 1 / 19

Introduction - ph1n1 influenza In BC: First cases in April/May 29... but epidemic began in earnest at the start of September. 18 hospitalizations / 57 deaths in BC. Characterized by higher attack rates in younger individuals. BC ph1n1 Surveillance Update, Feb 25 21. Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 2 / 19

Introduction - Immunization BC ph1n1 vaccination campaign began Oct. 26th, 29. Goal: Model pandemic influenza in an urban setting (we use GVRD as representative). Examine role of vaccination campaign timing and targeting strategies in mitigating impact of pandemic. Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 3 / 19

Simplified Model Schematic Complicating details: Split exposed, infectious compartments so distribution of time spent in them is a more realistic gamma. Heterogeneous model: each compartment represents a combination of different age and activity levels... Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 4 / 19

Age and activity levels Age Address different vulnerability to infection, mortality rates, etc and assess different vaccination coverage strategies. Activity Capture contribution of individuals with very high contact rates. Population fraction (%) 4 35 3 25 2 15 1 5 5 6 15 16 3 31 1 1+ Pourbohloul et al. (25) 2 2 4 5 17 18 2425 5455 64 65+65+(care) Age groups Note: We focus on GVRD as a representative urban area. Results shown for broader age groups (to compare with data). Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 5 / 19

Simplified Model Schematic Each compartment represents 4 sub-compartments (8 age groups 5 activity levels). Further key detail (before you ask!): Carefully derive contact probabilities for transmissibility from Vancouver contact network e.g. probability that someone age 25-54 with low activity level contacts someone age 5-17 with high activity levels, etc. Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 6 / 19

Baseline Parameters Variable Value Latent period (mean) 3 days Initial asymptomatic infectious period (mean) 1 day Total duration of infectiousness (mean) 7 days Basic reproduction number R 1.4 Vaccine Efficacy.9 Proportion of pop. with pre-existing immunity.5 (age 55+ only) Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 7 / 19

Vaccination Coverages Consider 3 vaccination coverage strategies: strategy coverages in GVRD. UC strategy Uniform coverage across all age groups PC strategy Immunizing school-aged children & their parents only Note: overall coverage 45% for /UC, 36% for PC. Age specific coverage (%) 1 9 8 7 6 5 4 3 2 1 UC PC 2 2 4 5 17 18 24 25 54 55 64 65+ 65+ (care) Age Group Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 8 / 19

Model predictions - Vancouver ph1n1 epidemic Age Spec. Inc. per 1, 2 15 1 5 Age Groups 4 5 17 18 54 55+ Oct 1 Nov 1 Dec 1 Jan 1 Feb 1 Mar 1 ICs: 1 cases on Sept 6 (educated guess, based on 1 confirmed cases). 8-week vaccination campaign starting October 26th, 21. Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 9 / 19

Vaccination rates Vaccinations concurrent with epidemic What kind of distribution rates should we consider? Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 1 / 19

Model predictions - Vancouver ph1n1 epidemic Age Spec. Inc. per 1, 2 15 1 5 Age Groups 4 5 17 18 54 55+ Oct 1 Nov 1 Dec 1 Jan 1 Feb 1 Mar 1 ICs: 1 cases on Sept 6 (educated guess, based on 1 confirmed cases). 8-week vaccination campaign starting October 26th, 21. Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 11 / 19

Model predictions - Vancouver ph1n1 epidemic A B September October Fraction of total cases C.5.4.3.2.1 Reported Model 4 5 17 18 54 55+ Age groups November Fraction of total cases.5.4.3.2.1 4 5 17 18 54 55+ Age groups D Cumulative cases Sept.1 to Dec.1 Fraction of total cases.5.4.3.2.1 Fraction of total cases.5.4.3.2.1 4 5 17 18 54 55+ Age groups 4 5 17 18 54 55+ Age groups Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 12 / 19

Model predictions - Vancouver ph1n1 epidemic Age Spec. Inc. per 1, 2 15 1 5 Age Groups 4 5 17 18 54 55+ Oct 1 Nov 1 Dec 1 Jan 1 Feb 1 Mar 1 ICs: 1 cases on Sept 6 (educated guess, based on 1 confirmed cases). 8-week vaccination campaign starting October 26th, 21. Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 13 / 19

Different Campaign Start Dates % reduction in final attack rate 8 6 4 2 Campaign Length 12 wks 1 wks 8 wks 6 wks 4 wks Jul Aug Sept Oct Nov Dec Start date of vaccination campaign Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 14 / 19

Different coverage strategies - attack rate % reduction in final attack rate 8 6 4 Start of epidemic Targeting strategy campaign start date 2 Parents & children Uniform Jul Aug Sept Oct Nov Dec Start date of vaccination campaign Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 15 / 19

Different coverage strategies - mortality % reduction in mortality 8 6 4 Start of epidemic Targeting strategy campaign start date 2 Parents & children Uniform Jul Aug Sept Oct Nov Dec Start date of vaccination campaign Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 16 / 19

Different coverage strategies Attack Rate: Mortality: % reduction in final attack rate 8 6 4 Start of epidemic Targeting strategy campaign start date 2 Parents & children Uniform Jul Aug Sept Oct Nov Dec Start date of vaccination campaign % reduction in mortality 8 6 4 Start of epidemic Targeting strategy campaign start date 2 Parents & children Uniform Jul Aug Sept Oct Nov Dec Start date of vaccination campaign Notice: Better strategy depends on goals. As campaign is increasingly delayed, strategy becomes less important. Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 17 / 19

Vaccine Efficacy ( strategy) Vaccine Efficacy 1.9.8.7.6 Start.9of epidemic.85.75.65.55.25.15 start date of vaccination campaign.5.45.35.5 Jul Aug Sept Oct Nov Dec Start date of vaccination campaign Impact of efficacy diminished as epidemic progresses. A lesser vaccine distributed early is preferable. Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 18 / 19

Discussion Developed a model of pandemic influenza in an urban setting that I I captures variability in age and behaviour includes vaccinations concurrent with epidemic validating it through comparison with ph1n1 influenza in GVRD. Different campaign start times: the sooner the better. Different coverage strategies: effective for campaigns started before/early in epidemic, less so for campaigns started later. Are the logistics worth the cost? Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 19 / 19

Role of mortality parameters Consider different mortality profiles: log 1 (Mortality per 1, pop.) 3 2.5 2 1.5 1.5 Source of mortality parameters Donaldson et al. (21) Hadler et al. (21), method 1 Hadler et al. (21), method 2 Tuite et al. (21) 2 2 4 5 17 18 24 25 54 55 64 65+ 65+, care Age Groups Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 2 / 19

Role of mortality parameters Resulting mortality reductions: % reduction in mortality 8 6 4 Targeting strategy campaign start date 2 Parents & children Uniform Jul Aug Sept Oct Nov Dec Start date of vaccination campaign % reduction in mortality 8 6 4 Targeting strategy campaign start date 2 Parents & children Uniform Jul Aug Sept Oct Nov Dec Start date of vaccination campaign % reduction in mortality 8 6 4 Targeting strategy campaign start date 2 Parents & children Uniform Jul Aug Sept Oct Nov Dec Start date of vaccination campaign % reduction in mortality 8 6 4 Targeting strategy campaign start date 2 Parents & children Uniform Jul Aug Sept Oct Nov Dec Start date of vaccination campaign Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 21 / 19

Time course of Vancouver epidemic Model prediction: Incident cases per 1 population 5 45 4 35 3 25 2 15 1 5 4 5 17 18 54 55+ Sept 1 Oct 1 Nov 1 Dec 1 Jan 1 Feb 1 Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 22 / 19

Sensitivity Analysis Vary latent period (2-4 days) and infectious period (5-7 days). Cumulative Attack Rate (%) 6 5 4 3 R =1.8 R 2 =1.6 Start of R 1 =1.4 vaccination R =1.2 campaign Sept Oct Nov Dec Jan Feb Mar Apr May Date Conway et al. (UBC) Campaign Timing in Vancouver, CAN 11 Jan 211 23 / 19