Pandemic influenza in Papua New Guinea: a modelling study comparison with pandemic spread in a developed country

OBJECTIVES The possible occurrence of a highly pathogenic influenza strain is of concern to health authorities worldwide. It is known that during past influenza pandemics developing countries have experienced considerably higher death rates compared with developed countries. Furthermore, many developing countries lack appropriate pandemic preparedness plans. Mathematical modelling studies to guide the development of such plans are largely focused on predicting pandemic influenza spread in developed nations. However, intervention strategies shown by modelling studies to be highly effective for developed countries give limited guidance as to the impact which an influenza pandemic may have on low-income countries given different demographics and resource constraints. To address this, an individual-based model of a Papua New Guinean (PNG) community was created and used to simulate the spread of a novel influenza strain. The results were compared with those obtained from a comparable Australian model. DESIGN A modelling study. SETTING The towns of Madang in PNG (population ∼35 000) and Albany (population ∼30 000) in Australia. OUTCOME MEASURES Daily and cumulative illness attack rates in both models following introduction of a novel influenza strain into a naive population, for an unmitigated scenario and two social distancing intervention scenarios. RESULTS The unmitigated scenario indicated an approximately 50% higher attack rate in PNG compared with the Australian model. The two social distancing-based interventions strategies were 60-70% less effective in a PNG setting compared with an Australian setting. CONCLUSIONS This study provides further evidence that an influenza pandemic occurring in a low-income country such as PNG may have a greater impact than one occurring in a developed country, and that PNG-feasible interventions may be substantially less effective. The larger average household size in PNG, the larger proportion of the population under 18 and greater community-wide contact all contribute to this feature.