Risk analysis for the reintroduction and transmission of measles in the post-elimination period in the Americas

ABSTRACT Objective. To propose and test a model for analyzing municipalities’ level of risk of reintroduction and transmission of the measles virus in the post-elimination period in the Americas. Methods. An ecological-analytical study was conducted using data on the measles epidemic that occurred in 2013–2015 in northeastern Brazil. The variables for analysis were selected after an extensive review of scientific literature on the risk of importation of measles cases. A univariate analysis considering the presence or absence of confirmed cases of measles in 184 municipalities in the state of Ceará, Brazil, was carried out to evaluate the association between the dependent variable and 23 independent variables, grouped into four categories: 1) characteristics of the municipalities; 2) quality indicators for immunization programs and epidemiological surveillance; 3) organizational structure for the public health response; and 4) selected impact indicators. A P value < 0.05 was considered significant. All variables with P < 0.200 were analyzed using multivariate logistic regression. Based on the results, the municipalities were categorized by four levels of risk (“low,” “medium,” “high,” and “very high”). Results. The model sensitivity was 95% for concordance between municipalities classified as “high risk” and “very high risk” and those that had an epidemic between 2013 and 2015 in Ceará. Of the 38 municipalities that had an epidemic, 76% (29/38) were classified as “high risk” and “very high risk”; 146 municipalities did not report cases (P < 0.0002). Conclusions. Given the imminent risk of reintroduction of measles circulation in the post-elimination period in the Americas, this model may be useful in identifying areas at greater risk for reintroduction and continued transmission of measles. Knowledge of vulnerable areas could trigger appropriate surveillance and monitoring to prevent sustained transmission.

confirmed in Brazil, Canada, Ecuador, and the United States (6). In 2013, the measles virus was reintroduced in Brazil, with about 1 200 confirmed cases in Permanbuco and Ceará, two northeastern states (7,8).
Given the need for sustainable elimination of the disease in the Americas, and taking into account the new context and challenges of the post-elimination period, two strategic plans were developed and implemented and remain in force until 2020-a global plan, and a Regional plan (9)(10)(11).
A recent study has developed a tool to measure the risk of dissemination of measles in regions close to elimination a (12). The objective of this study was to propose and test the model for analyzing municipalities' level of risk of reintroduction and transmission of the measles virus in the post-elimination period in the Americas. The model analyzed data on the measles epidemic in northeastern Brazil between 2013 and 2015.

Description of study and the selection and analysis of variables
This study was ecological and analytical (13). The variables were selected after an extensive review of scientific literature on the risk of importation of measles cases. The municipality was the unit of analysis, according to the International Classification of Diseases, 10 th revision (ICD-10) and PAHO/WHO recommendations. The scenario of the year before the occurrence of a measles epidemic in Brazil's northeastern state of Ceará was analyzed. The occurrence or nonoccurrence of measles cases in 184 municipalities in that state was the dependent (outcome) variable. The selected independent variables (a total of 23) were grouped into four categories: 1) municipality characteristics; 2) quality indicators for immunization programs and epidemiologic surveillance; 3) organizational structure for the public health response; and 4) impact indicators ( Table 1). The independent variables were dichotomized ("presence" versus a The tool does not allow for measuring the risk of reintroduction of the virus in areas that have already eliminated the disease because the data measured in the study include the public health response to cases and the population profile affected by measles. "absence"), using medians and values defined as adequate by the Brazilian Ministry of Health and Ministry of Tourism, and PAHO/WHO, as cutoff points.
A univariate analysis was carried out to evaluate the association between the 23 independent variables and the dependent variable (occurrence or  nonoccurrence of measles. In this first step of the analysis, variables with a P value < 0.05 based on the chi-square test or Fisher's exact test-a total of sixwere considered statistically significant indicators of risk and were incorporated into the model for multivariate analysis. All variables with a P value < 0.200 were analyzed using multivariate logistic regression. In the second step of the analysis, three additional variables were added: 1) coverage with dose 1 of the measles-mumps-rubella vaccine (MMR); 2) the notification rate for febrile eruptive (exanthematic) diseases; and 3) the presence of vulnerabilities (border with other countries, favelas (shanty towns), violence, indigenous communities, population resistant to vaccination, difficult geographic access, and areas with trade fairs and mass events). Each municipality received a total score of 0 to 100. After weighting the variables and scores, the municipalities were classified as "low risk," "medium risk," "high risk," or "very high risk," using the 20th, 60th, and 90th percentiles to establish cutoff points. Thirteen points were given for each variable in the first extract (for the variables with statistical significance) except "coverage by Community Health Agents (CHAs)," which had a higher odds ratio (7.22) so was worth 14 points. Municipalities with a total score of 28 points or less were classified as "low risk," whereas those with 29-46 points were classified as "medium risk," those with 47-67 points were classified as "high risk," and those with scores of 68 or higher were classified as "very high risk" (Figure 1).

Ethical considerations
The study complied with all ethical requirements of Brazilian National Health Council (Conselho Nacional de Saúde, CNS) resolution #466/2012 and was approved by the Ethics and Research Committee of Christus University Center (approval #43405315.3.0000.5049).

RESULTS
Based on the analysis, six variables were statistically significant (P < 0.05) and were thus selected as indicators for the model: tourism, population density, percentage of urbanization, dropout rate between MMR dose 1 and 2, proportion of population covered by CHAs, and proportion of population covered by Family Health Strategy (FHS) teams (Table 1). Table 2 provides details about the scores and criteria for each indicator.
The model showed a sensitivity of 94.7% for concordance between municipalities classified as "high risk" and "very high risk" and those that had an epidemic between 2013 and 2015 in Ceará. Two of the 184 municipalities (5.3%) were classified as "low risk," and had one confirmed case each, but the source of infection was not the municipality of residence, so both cases were considered imported. Among the 38 municipalities that had an epidemic, 76% (29/38) were classified as "high risk" and "very high risk"; the 146 municipalities that did not report cases 42.5% (62/146) had a "medium risk" classification (P < 0.0002).

DISCUSSION
The model presented in this study used data on the pre-epidemic scenario in Ceará, Brazil, in an effort to reproduce the characteristics of areas with measles cases after elimination had been achieved. The risk analysis generated by the model was validated using data from the epidemic that occurred in Ceará between 2014 and 2015. The objective was to establish a relationship between certain characteristics of the municipalities and health systems and occurrence of the epidemic. The model presented here could be used as a tool for strategic planning to maintain, monitor, and sustain measles elimination efforts in the Americas.  All 184 municipalities included in the analysis scored poorly for the health system quality indicators related to immunization programs and epidemiological surveillance. If specific actions were implemented to improve the quality of these health system indicators, risk of measles reintroduction and transmission in these municipalities may decrease, even if deficits in all of the health-related indicators studied here can not be addressed through interventions due to resource constraints. Specific actions may also be taken to address some of the indicators unrelated to the health system, such as tourism and population density. For example, the authors suggest that 1) municipalities with a strong potential for tourism ensure that all professionals working in the tourism sector are vaccinated with two doses of the MMR; 2) municipalities with high rates of urbanization consider an alternative schedule for vaccinating the target population, given the dynamics of large urban centers; and 3) municipalities with high demographic densities consider microscenarios (community plans) for immunization to help prevent the accumulation of susceptible populations.
Various tools have been developed to predict the future occurrence of diseases in a population, including prediction models developed for breast cancer, in which screening or chemoprophylaxis may be considered useful for those at high risk, some of which are relevant not only for clinical decision-making but also for estimating overall health costs (14)(15)(16)(17)(18)(19)(20). Studies describing the application of risk stratification and prediction models suggest that there is strong evidence for the use of these predictive models with administrative and clinical data for patients with chronic disease (21)(22)(23).
In Ireland, using predictive models and risk stratification, certain actions were incorporated in primary care and hospital admissions were reduced (24). Other studies have shown efficacy in customizing and developing models or tools that meet the needs of a given health systemincluding predictive models for communicable diseases that consider the infectivity potential of the etiological agent and the level of population immunity-and are predominantly based on scientific evidence drawn from models of dynamic disease transmission, and designed to support global efforts to control and eliminate immune-preventable diseases (10,(25)(26)(27)(28)(29)(30).
Endemic transmission of measles virus in other parts of the world remains a risk for regions that have eliminated the disease, and until there is a disruption of virus transmission worldwide, the possibility of importation of measles cases and occurrence of outbreaks remains. Although the elimination of measles in the Americas was certified in 2016, sporadic reintroductions may result in new transmission chains (13) that spread according to the level of immunity of the resident population (21). Therefore, the main challenges to maintaining the elimination of measles are ensuring 1) sensitive surveillance; 2) an effective response to the importation of the wild virus; 3) homogeneous vaccination coverage (≥ 95%) in the municipalities; and 4) the elaboration of an integrated action plan, with periodic risk analysis (15,31).
The risk of transmission of the measles virus is associated with vaccine coverage and characteristics of the municipality such as response capacity and epidemiologic surveillance. The ideal scenario is that all measles cases be readily identified by the health service and secondary cases avoided by implementing strategies that contain transmission in the community. Therefore, the variables related to the public health response and structure of the health services of the municipalities, along with the operational indicators, allowed for analysis of the risk of transmission of the measles virus in Ceará.

Limitations
The limitations of this study were mainly due to the inability to validate the study method in other countries in the Americas that have achieved measles elimination (including those that had epidemics, such as Ecuador and Venezuela) and in other states in Brazil (e.g., Bahia, Paraíba, and Pernambuco). This limitation was due to lack of access to databases from the abovementioned countries and other states in Brazil. The limitation was not related to the study method, and the variables studied here are available for other countries of the Americas. Future validation studies could address this gap. Another limitation may be the quality of the analyzed data, which were secondary and collected for other purposes.

Conclusions
Given the imminent risk of reintroduction of measles circulation in the post-elimination period in the Americas, this model may be useful in identifying areas at greater risk for reintroduction and continued transmission of measles. Knowledge of vulnerable areas could trigger appropriate surveillance and monitoring to prevent sustained transmission.

Conflicts of interest. None.
Disclaimer. Authors hold sole responsibility for the views expressed in the manuscript, which may not necessarily reflect the opinion or policy of the RPSP/ PAJPH or the Pan American Health Organization (PAHO).