sign in sign up
<<
Home , Hermosillo

CASE STUDY: Hermosillo, , Captures Heat-Related Illnesses at Medical Facilities Using New Database

The Problem: The high rates of mortality prompted all levels of govern- ment and health authorities to support measures to pro- The threat of climate change in the state of raises tect and prevent heat exposure in the region. Although serious concerns regarding the protection of human reportable diseases are included in routine epidemiolog- health, as daily maximum temperatures over 44°C are ical surveillance in the region, information on HRI is not often recorded and conditions are expected to worsen. collated nor reported to central health authorities. Fur- Sonora State has a desert climate, characterized by low thermore, electronic medical records are collected in precipitation, high sun exposure, and extreme heat. hospitals or other medical facilities in Hermosillo, the Approximately 60 percent of all heat mortality observed capital city of Sonora, but for epidemiological analy- in Mexico for 2015 occurred in that state (Figure 10). ses, paper records prevail.

Figure 10. Mortality due to extreme heat in Mexican states for 2015

Sonora Yucatan San Luis Potosi 0 50 100 150 200 250 300 350 400 Number of Heat-Related Deaths, 2014 Source: Sistema de Vigilancia Epidemiológica en México, 2015

18 Commission for Environmental Cooperation The Solution The common occurrence of extreme heat in this region makes it difficult to generate adherence to pro- Working with Cofepris, the Ministry of Health, and tective measures for extreme heat; messaging needs the CEC, Sonora’s regional health authority (Comis- to be targeted to the most vulnerable populations to ión Estatal de Protección contra Riesgos Sanitarios del reduce “alert fatigue.” Estado de Sonora—Coesprisson) established several objectives with the goal of creating a real-time SyS Four main presenting HRI were identified by the system for the city of Hermosillo in a 2016 pilot SyS medical facilities in Hermosillo: (1) dehydration, (2) project that would enable timely identification of heat exhaustion, (3) sunstroke and (4) sunburn. The health impacts due to extreme temperature and evi- causes of the heat exposure in most of the cases iden- dence-based policy development to reduce mortality tified were attributed to occupational exposures (e.g., and morbidity rates. These objectives included the farm and mine workers). The existing epidemiologi- following: cal surveillance consisted of weekly bulletins provided by local medical facilities of reportable diseases, but • conducting an analysis of HRI rates in the Coesprisson wanted to design a real-time system that region, included SyS of HRI. The team approached the prob- • designing and implementing a computerized lem by hiring six staff (two medical students, three platform to receive and store real-time data nurses, and one paramedic) to be employed within related to the health effects and correlating these two Hermosillo hospitals to actively seek out HRI data with climate and demographic information, cases and record their details into a newly constructed • promoting coordinated work by data owners database (Figure 11). Using appropriate security and (i.e., meteorology and health), and privacy features, the team established the basis of a • implementing coordinated measures for health heat-specific SyS system that was implemented in the protection and prevention strategy during two hospitals, and can be expanded to more hospi- extreme heat events. tals and/or health outcomes as resources permit. For

Figure 11. The dashboard for Hermosillo’s syndromic surveillance for heat-related illnesses Source: Hugo Medina 2016

A Guide for Syndromic Surveillance for Heat-Related Health Outcomes in North America 19 example, one measure to protect data security and should ensure that messaging and adaptive strate- patient privacy allows only system administrators gies are appropriate to the population they target. For access to all medical records; staff that input data example, the initial data from this project indicate have access to records only as they are entered into that most HRI are due to occupational exposures. the database. Prevention and protective strategies need to include input from employers, workers, and occupational The SyS system aims to collect information regarding health authorities. Safety specialists could be put in both case details and the causes of the HRI. Data are place for specific occupational sectors to ensure that collected in real time for HRI cases, including the fol- policies and education strategies are properly imple- lowing data elements: mented and effective. The data collected with the SyS system in Hermosillo should be used to demonstrate • reporting medical facility, the increased risk experienced by certain local occu- • address of medical facility, pational groups to create protective policy. Data col- • effect of injury (i.e., illnesses or death), lected during the 2016 pilot project indicate that the • basic cause of injury/death (i.e., dehydration, typical patient for HRI presenting at the participating heat exhaustion, heat stroke, sunburn), medical facilities was a male (42 of 58 cases), exposed • demographic information (patient name, age, to extreme heat at work (35 of 58 cases) or dehydra- sex, address of residence) tion (44 of 58 cases). • address of exposure/incident, • date of exposure/date of notification, There are often circumstances where electronic data • name of staff member reporting case, sources are unavailable and researchers need to create • SyS study site, opportunities to collect the information. In Her- • environmental temperature at time of exposure, mosillo, the database created by Coesprisson to col- and lect HRI information allows researchers and public • date and time of information capture. health authorities to analyze and display health out- comes in near-real time in a circumstance where elec- At present, temperature is entered into the database at tronic medical records are not readily available for the same time as patient information, but automatic this purpose. This system is an example of active sur- weather information input is being pursued with the veillance, where the accuracy and effectiveness of the collaboration of Mexico’s meteorological system. system are dependent upon the input practices of the staff responsible for reporting the HRI cases. Addi- tional staff would be required to expand the pilot SyS Lessons Learned to other medical facilities, as well as additional staff training and enhancements to the current database One of the main challenges faced by Coesprisson in to support multiple simultaneous users. It is possible their pilot project was the issue of alert fatigue. The that further simplification of the application to run community is used to daily life in a consistently hot on operating systems for cell phones and tablet-based environment, and lives with the daily risk of expo- technologies could be employed to enhance the pilot sure to extreme temperatures. Education strategies SyS reporting protocols. This SyS system could easily and the implementation of policies for different sec- be expanded to other reportable HRI, as well to other tors (e.g., workplaces, schools, etc.) that are sensi- syndromes of medical interest (e.g., occupational tive to the adaptive capacities of each specific sector injury, chemical exposures), thereby increasing the need to be developed in coordination with the at-risk value of this database to public health surveillance populations. Such collaborative policy development strategies in Hermosillo.

20 Commission for Environmental Cooperation 3 Defining a Heat Syndrome email application, a syndrome classifier can recognize text in a healthcare record that has varying probabili- The pre-diagnostic data collected from the various ties of being related to specific symptoms, which can sources must be processed and classified into med- in turn indicate certain medical conditions (or syn- ically relevant syndromes in order to derive epi- dromes). To allow the system to recognize the various demiological information. To generate syndromic syndromes of interest, the system needs to be tested groupings, the electronic record needs to be analyzed with a large dataset consisting of healthcare records of and classified using a statistical syndrome classifier; known disposition. For a detailed description of these the classifier chosen depends on the data that is used. processes, various approaches are described in Using For example, if the SyS system is based on free-text chief complaints for syndromic surveillance: A review chief complaints from acute care records, the classi- of chief complaint-based classifiers in North America fier will be based on groupings of keywords and/or (Conway, Dowling, and Chapman 2013). phrases. If the data consist of pre-diagnostic codes, the classifier will be based on those specific codes or Applying these methods to HRI poses challenges; the groups of codes. This is relevant for SyS systems using etiology of HRI are generally not specific, and pre- data from hospitals that employ drop-down, pre-de- senting symptoms may be misinterpreted and/or mis- fined coding for recording the reason for hospital classified into other syndromes. Several approaches to visits instead of free-text chief complaints. this problem are available: for example, some systems may classify a single record into several different syn- Different approaches are taken by different SyS sys- dromes, and priority could be given to an HRI syn- tems to define syndromes. Some are based on machine drome when temperatures are above predetermined learning and natural language processing (NLP) tech- thresholds. The approach used depends on the goals niques, where classification algorithms are applied to of the SyS system, as well as the quality of the data textual data to effectively teach the system to correctly sources. The Michigan case study illustrates both the classify words and phrases (or codes) into groupings of methodology and the challenges of defining a heat medical interest. Like a spam detector operating in an syndrome to capture HRI.

A Guide for Syndromic Surveillance for Heat-Related Health Outcomes in North America 21