The Traffic Light Approach: Indicators and Algorithms to Identify Covid-19 Epidemic Risk Across Italian Regions

With the beginning of the autumn-winter season, Italy experienced an increase of SARS-CoV-2 cases, requiring the Government to adopt new restrictive measures. The national surveillance system in place defines 21 key process and performance indicators addressing for each Region/Autonomous Province: (i) the monitoring capacity, (ii) the degree of diagnostic capability, investigation and contact tracing, and (iii) the characteristics of the transmission dynamics as well as the resilience of health services. Overall, the traffic light approach shows a collective effort by the Italian Government to define strategies to both contain the spread of COVID-19 and to minimize the economic and social impact of the epidemic. Nonetheless, on what principles color-labeled risk levels are assigned on a regional level, it remains rather unclear or difficult to track.

With the beginning of the autumn-winter season, Italy experienced an increase of SARS-CoV-2 cases, requiring the Government to adopt new restrictive measures (1). The national surveillance system has been implemented according to the Prime Ministerial Decree (DPCM) of November 3, 2020 (2), the previous strategic documents for monitoring for the second wave (3,4) and the DPCM of April 26, 2020 (5). The latter document defines 21 key process and performance indicators addressing for each Region/Autonomous Province: (i) the monitoring capacity, (ii) the degree of diagnostic capability, investigation and contact tracing, and (iii) the characteristics of the transmission dynamics as well as the resilience of health services ( Table 1) (5). Based on a weekly assessment, each region is assigned to different levels of risk (high, medium, and low), which are mapped using a "traffic light" color code (red, orange, and yellow). The criteria are compliant with the European Commission's decision to adopt a common color code to provide strategy to restrict free movement across the European Union (6) and with the WHO frameworks (7).
The complete algorithm is presented in Figure 1. The overall risk assessment is based on 21 key process and performance indicators ( Table 1). Either thresholds or comparative evaluation are defined and monitored on a weekly basis by each region as part of the national integrated surveillance system. Indicators 1.1-1.6 are used for an initial assessment of the quality of the information collected. If a region has those indicators above threshold, it is automatically labeled as a high-risk region.   Non-increase the number of active transmission outbreaks in the Region; Absence of transmission outbreaks on the regional territory for which a risk assessment was not quickly carried out and the opportunity to establish a sub-regional "red zone" evaluated  If a region complies with the threshold, two algorithms are used in order to evaluate the epidemic probability risk and epidemic impact risk ( Figure 1A). Indicators used to evaluate the probability risk range from 3.1 to 3.7, while indicators used to evaluate the impact risk range from 3.1 to 3.9 (5). A risk assessment matrix (RAM) based on the probability and impact risk defines the risk level as: very low, low, moderate, high, and very high ( Figure 1B). An additional classification of the level of risk has been defined assuming four possible scenarios based on the reproductive number (Rt) ( Table 1-indicator 3.2) as described in Figure 1C (3). For each region, the combination of RAM and the four possible scenarios defines the overall high, medium and low risk, which is color coded as red, orange, and yellow (2) (Figure 1D). A region is considered red when the RAM is equal to very high risk or high risk and complies with Scenario 4; orange when the RAM is equal to very high risk or high risk and complies with Scenario 3; and yellow when it is in Scenario 1 or Scenario 2, independently from the RAM results. From December 6 until December 12, the overall distribution of the Regions and Autonomous Provinces was as follows: 12 color-coded as yellow areas, eight color-coded as orange areas, and one as red area.
The risk assessment method outlined by the Italian Government is in line with the WHO and European Commission frameworks (6,7). Although detailed indicators and the risk assessment process were defined, we draw attention to the following critical issues: (1) some indicators are not clearly outlined either in their meaning or in the proposed threshold (indicators 2.1, 2.4, 2.5, 2.6, 3.1-3.6); (2) the two algorithms of probability and impact risk assessment do not report which specific indicators are used to assess the four questions ( Figure 1A); (3) overall, the information needed to define the entire algorithm is fragmented in different documents (2)(3)(4)(5); (4) although the 21 indicators have been defined in the DPCM of April 26 (5), the related weekly reports started being described on a regional basis only from October 26 (8); (5) potential enforcement of the risk assessment algorithm by local regional regulation. As an example, the administration of the Abruzzo Region, which was on a Red risk level until December 5, decided to unilaterally self-declare on an orange risk level despite the Italian central Government did not allow the shift until December 12 (9).
The traffic light approach shows a collective effort by the Italian Government to define strategies to both contain the spread of COVID-19 and to minimize the economic and social impact of the epidemic. Nonetheless, on what principles color-labeled risk levels are assigned on a regional level, it remains rather unclear or difficult to track.

AUTHOR CONTRIBUTIONS
All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.