Population mobility and dengue fever transmission in a major city in Southeastern Brazil, 2007-2015

Background Around 14% of world dengue virus (DENV) cases occur in the Americas, the majority of them in Brazil. Although socioeconomic, environmental and behavioral correlates of dengue have been analyzed for different contexts, the role played by population mobility on DENV epidemics, especially at the local level, remains scant. This study assesses whether the daily pattern of population mobility is associated with DENV transmission in Campinas, a Brazilian major city with over 1.2 million inhabitants in São Paulo state. Methodology/Principal Findings DENV notifications from 2007 to 2015 were geocoded at street level (n=114,884) and combined with sociodemographic and environmental data from the 2010 Population Census. Population mobility was extracted from the Origin-Destination Survey (ODS), carried out in 2011, and daily precipitation was obtained from satellite imagery. Zero-Inflated Negative Binomial (ZINB) regression models controlled by demographic and environmental factors revealed that high population mobility had a substantial positive effect on higher risk for DENV transmission. High income and residence in apartments were found to be protective against the disease, while unpaved streets, number of strategic points (such as scrapyards and tire repair shops), and precipitation were consistently risk factors for DENV infection. Conclusions/Significance The use of fine-scale geographical data can unravel transmission idiosyncrasies not evident from a coarse spatial analysis. Even in a major city like Campinas, the vast majority of population daily mobility occurs at short distances. Based on our results, public policies on DENV transmission control should dedicate special attention to local hubs of population mobility, especially during high transmission weeks and in high dengue incidence areas. Author Summary Currently, about half of the world population is at risk of a dengue infection. Numerous studies have addressed the socioeconomic and environmental determinants of the disease. However, little is known about the role played by population mobility on dengue transmission, particularly at the local scale. This study aims at investigating this issue. Our hypothesis was that population movements are a prominent driving force for dengue diffusion locally. We investigated the case of Campinas, a municipality with over 1.2 million inhabitants in Brazil that recorded dengue epidemics in 2007, 2014 and 2015. Our study focused on the years 2007 to 2015, comprising more than 114 thousand cases, geocoded to the household address, and combined with socioeconomic, environmental and daily population mobility data. Our results showed that even controlling for demographic and environmental factors, population mobility was the most important predictor for dengue fever incidence.


Introduction
Despite the growing concern about other infectious diseases transmitted by Aedes 67 aegypti, such as Zika virus and chikungunya, dengue virus (DENV) remains a global 68 threat [1-3]. The incidence of this disease has grown dramatically in recent decades. 69 Economically, the annual global cost of DENV is estimated at around US$8.9 billion  In Brazil's most populous state, São Paulo, DENV transmission began to be reported, 86 with clinical and laboratory diagnosis, in March 1984 [12]. Campinas is the third most  (v) Campinas has collected unique data on regular daily mobility.   The most affluent groups live in the center and north of the municipality, where urban 131 infrastructure is of better quality. In contrast, the south concentrates the impoverished 132 people, with less access to urban services. The urban configuration of Campinas 133 encourages population mobility, given that the territory is abundantly crossed by 134 major roads and highways (Fig 1).  rates. In all of these models, the variable population mobility was exactly the same, 245 and statistical significance was evaluated at a p-value < 0.01.   information to be geocoded at the household level and included in the analysis.

258
Geocoding success was high, and varied from 90.0% in 2014 to 99.4% in 2009 (Table   259 1).     With regards to the intensity of transmission and of mobility (Table 3)   The spatial distribution of the total number of epidemiological weeks with high 296 incidence rates (Fig 3)  Jardim Paulicéia, and Jardim Campos Elísios neighborhoods. These areas are the 302 ones with the highest historical concentration of dengue cases in the city during the 303 study period, and where epidemics recurrently occur (Fig 1 and S1   shades of blue, from areas with no epidemiological week with high dengue incidence rates 310 to areas that registered the peak of 16 to 42 epidemiological weeks with high incidence rates 311 during the study period. Figure created  As for the intensity of mobility, while the "most mobile" populations were apparently 320 widespread in Campinas, they mostly overlapped with "High Areas" distribution, mobility. Dengue incidence rates classified in 5 categories ranging from 49 to more than 330 56,000 cases per 100,000 persons, colored from light blue (lowest) to dark purple (highest 331 incidence rates). Figure created

339
In assessing the role of population mobility, the complete model (Model 1,  Tables 5 and 6 show stratified models by high and low epidemiological weeks 352 (Models 2 and 3), and Tables 7 and 8 present stratified models by high and low 353 incidence areas (Models 4 and 5). Results indicate that precipitation was positively 354 associated with dengue incidence rates in periods of low transmission weeks and low 355 incidence areas (Models 3 and 5). The variable households in unpaved streets was 356 significantly correlated with dengue incidence rates in high transmission weeks and 357 areas (Models 2 and 4). Living in high mobility areas was consistently a risk a factor 358 for dengue incidence rates across all models. On average, living in a highly mobile 359 area increased dengue incidence rate in 46% and 69% during high transmission 360 weeks and in high transmission areas, respectively (Models 2 and 4). During low 361 transmission weeks and in low transmission areas, being part of a highly mobile 362 population increased by 26% and 35% dengue incidence rates, respectively (Models   In Model 3, income per capita was protective against dengue fever, although not 378 significant. Sex ratio was a protection factor against the disease in all stratified 379 models, suggesting that more male residents are associated with a lower dengue 380 incidence rate -in Models 2 and 4, the increase of one male per 100 female residents 381 was accompanied by a decrease of 7% and 9% in the incidence rates, respectively.

382
With respect to SPs, Models 1 to 5 showed that the addition of one SP contributed to 383 an average increase of 3% in dengue incidence rates, with low variation across the 384 different models. Similarly, the increase in 1% of population living in apartments 385 reduced approximately 1% to 2% of dengue incidence rates on average across the 386 distinct models' stratification.  Living in a high incidence area increased dengue rates by 3 times in high 392 transmission weeks (Model 2), while in low transmission weeks it increased rates by 393 8 times (Model 3). Conversely, during high transmission weeks living in high incidence areas increased dengue rates by 2 times (Model 4), while living in low 395 incidence areas the dengue incidence was increased by more than 15 times (Model 396 5). Consequently, living in high transmission areas was associated with a higher 397 increase in dengue incidence rates during low transmission weeks. On the other 398 hand, high transmission weeks were most positively associated with increases in 399 dengue incidence rates in low incidence areas.  Survey showed that women tended to have higher mobility within their area of 459 residence, and therefore travel smaller distances than men (S2 Table). The extent to 460 which this pattern implies lower exposure to an infection depends on the 461 characteristics of the areas where they live and where they usually go.

462
Regarding Strategic Points, we found that the addition of one SP in an area tended 463 to increase the dengue incidence rate by 3%, a result consistent with previous