Clark County has a population of about 430,000. The target audience includes populations potentially exposed to communicable diseases and other hazards. With reductions in an already diminishing workforce, public health must explore new, more efficient ways of reaching populations exposed to potential hazards. We have done this with the use of the Emergency Community Notification System (ECNS), an automated system with polling capability and real time data available via secure web-based portal.
The objectives of this practice are:
1. To identify disease outbreaks more quickly and efficiently.
2. To increase the accuracy of clinical and exposure data entry.
3. To reduce the cost of outbreak investigations.
Clark County began developing this practice in March 2011 by exploring the use of ECNS, which is owned by the county emergency management agency. Public Health negotiated an agreement outlining cost and procedures in April and in June public health received log in credentials allowing use of the system. Public Health tested the system with test calls to public health staff and subsequent analysis of the responses. After the test, department leadership authorized use of the system for real events. Since ECNS was in place in our emergency management department, Public Health did not incur any start up costs. The only cost to Public Health is a nominal per use fee.
Our use of this system met all the objectives listed above. In the last use of the practice, investigation of high absentee rate at a local school, led to rapid identification of a disease outbreak. The practice reached 90% (49) households within 5 minutes of launching the automated survey, and cost $7.70. The initial real time data characterized the scope of the outbreak, and allowed Public Health to provide control measures by the end of the school day. Planning and collaboration among Public Health, local emergency management and the school system allowed for successful deployment of the practice. The practice met the needs of the school system, the parents contacted, the community and our health department.
Identifying potential outbreaks can be resource and time intensive. Delays in notifying potentially exposed populations and obtaining data necessary to identify disease outbreaks and prevent disease transmission can lead to increased morbidity and mortality.
By Washington State law and administrative rule, investigations and control of disease outbreaks are the responsibility of local health jurisdictions and a public expectation. During a measles outbreak investigation, we had to contact over 150 people potentially exposed. We were concerned by the time it took to conduct the investigation, including call backs and the time and resources it took to enter and clean the data. Consequently, we approached our emergency management colleagues to determine whether we could use their Emergency Community Notification System (ECNS) to contact exposed populations and conduct automated surveys.
The practice allows to us to select and reach predetermined, specific populations by automated phone within minutes; we can reach the entire county population, if necessary, within 30 minutes.
The practice allows us to survey those contacted through an automated polling feature using keypads on any phone.
The practice automatically, immediately populates an electronic database allowing us to analyze the data in real time and consequently respond to the emerging outbreak rapidly and efficiently.
The practice significantly decreases the number of exposed or ill people requiring individual interviews by staff. For example, in a measles outbreak, the automated system would allow us to select out for personal interviews only those who had indicated through the automated system that they were not immunized.
A literature review and search of public health practice databases revealed only one reference to use of the kind of system this practice describes, the use of "reverse 9-1-1" to reach vulnerable populations in Calvert County. However, that practice only involved notifying individuals about potential hazards and did not include use of a polling feature to gather data from those potentially exposed.
The current approach is for public health staff to conduct serial phone interviews with every person on a line list who has been potentially exposed to disease. Our innovative practice involves an automated, web-based phone survey that reaches each respondent nearly simultaneously, without intensive use of public health staff. In addition, this practice allows respondents to provide data through phone keypads, uploads the data into an electronic database and provides a continually updated (every 60 seconds) display for public health staff.
The stakeholders in the practice include the local health departments, local emergency managers, the vendor of the ECNS system, any residents of the county who may have been exposed to disease or institutions that require support from public health during a disease outbreak.
Role of Stakeholders/Partners
Local emergency managers train public health staff in the use of the system and authorize public health to use the system. The agreement between public health and emergency managers specifies the limits of the system’s use and its cost to public health. Public health provides after action reports on its use of the system to emergency managers. The vendor for the system provides technical support for use of the more sophisticated polling capabilities. Stakeholders representing populations potentially affected, such as schools, provide the phone numbers/contact information used in the automated phone system.
Clark County Public Health employees create the automated interview tool (tailored to the specific situation), determine the population to be reached through the automated system, launch the automated interview system, and analyze the data acquired. In addition, Clark County Public Health staff follow-up with individual phone calls for respondents whose previous automated responses require additional clarification or respondents who require specific interventions, such as prophylaxis or immunizations.
The community shareholders are supportive of the practice because it allows Clark County Public Health (CCPH) to identify potentially affected constituents and it allows CCPH to share control measures rapidly with the shareholders, often within the same day, such as a school day. For example, in cases involving increased school absentee rates, the practice can immediately identify whether a gastrointestinal or respiratory disease outbreak is occurring. Basically, it supports shareholders in their efforts to work with CCPH to control potential outbreaks. Previously, similar efforts could take up to several days. The practice also helps emergency managers understand how epidemiologists investigate and control communicable disease outbreaks.
The system allows slots for only five questions, each of which can be programmed for up to nine responses. Surveys must be carefully crafted to elicit the most important information within the constraint of the five question limit.
During investigations of school absenteeism, some households might provide survey responses for more than one child. The system allows only one survey per household and only allows five questions per survey. Although public health staff can program the system to allow respondents to indicate whether they had more one child in school, two limitations prevent the system from obtaining meaningful data for two or more children per household. These include the inability to send more than one automated survey to each household and the five-question limitation. Instead, public health staff would have to identify multiple-child households and follow-up with personal phone calls as necessary.
Since respondents are offered an option for inbound calling if they wish to talk to a staff person, the unit handling those calls must be prepared. The number of inbound calls is small, but the LHD unit that will receive the calls requires a thorough briefing on the problem. We experienced no barriers to collaboration. Our partners were eager to work with us in developing and implementing the practice.
For all objectives, staff required orientation to and training in system use.
Developed policies and procedures on using ECNS in disease investigations.
Required after action reports that specified time intervals related to the speed of detection of outbreaks.
Implemented the automated system and reported results.
Staff learned to construct surveys carefully to elicit the most important information within the constraint of the five question limit.
Limited the use of our call center to investigations requiring detailed information (not amenable to five-question surveys) up front. Used the automated system for identifying a potential outbreak - will continue to do so in the future. The after action reports provided documentation on the cost for each investigation. Since the data and results of each use of the system are archived on the vendor's servers, the data and results are perpetually available for examination.
Process & Outcome
Time from notification to beginning the interview process.
Time from initial survey or interview to development of the list of respondents requiring follow-up, including obtaining additional data.
Time between initial survey or interviews and identification/characterization of an outbreak.
Performance measures were derived from the database generated from the automated system. We compared these measures with similar data described in after action reports from previous outbreaks in which this practice was not used. In a measles investigation, using a call center, up to 12 staff surveyed 151 potentially exposed people in 48 hours. Initial interviews took an average of twenty minutes each. In a recent instance of high absenteeism at an elementary school, the automated system (requiring no staff) contacted 49 households within 5 minutes, and response data to survey questions was available in real time. Although the situations were different, we believe that the system could provide rapid initial information triage for more serious outbreaks, including measles. ECNS provided much quicker characterization of the situation than the traditional call center method. Use of the practice revealed opportunities for system operators to refine the process of constructing the survey questions and defining possible responses.
Amount of time used for data clarification and cleaning, including correcting errors in the database.
Time between completion of data entry and the production of final reports by epidemiologists. Performance measures were derived from the data developed by the automated system. We compared these measures with similar data described in after action reports from previous outbreaks in which this practice was not used.
Data from a traditional call center during a measles outbreak requires collation and aggregation of twelve handwritten interview forms that staff later enter into an electronic database (double data entry). The database then had to be used by epidemiologists to display the data in a form useful for Operations Section personnel.
In comparison, the automated ECNS system avoids double data entry by having respondents enter data by phone, which then populates the electronic database. In the school absentee investigation, ECNS automatically populated the data set, eliminating the need for separate data entry, and created a display that characterized the source of the high absenteeism as a viral gastrointestinal illness. With this practice, opportunities for data entry error were reduced, and no clarification of the data from a call center staff was necessary. Since the data and results of each use of the system are archived on the vendor's servers, the data and results are perpetually available for examination. The health officer and communicable disease staff used the data to formulate appropriate control measures and communicate them to the affected school by the end of the school day.
Difference in cost of staff time between use of ECNS and traditional call centers.
The ECNS system generates easily analyzable electronic data for time used. For traditional call centers, call center log sheets and Operations Section records provide time use data in paper form. For both systems, cost is determined by applying the hourly rate for public health staff working in the different systems to the number of hours they worked.
The cost in staff time for the traditional call center in the measles outbreak was approximately $2,160.00. The cost for staff time for ECNS use in the high absentee rate investigation was $120.00. The maximum system use cost for ECNS would have been $7.70 ($0.16 per minute per call) (in this case, the cost was zero because the emergency management contract offers a bank of free minutes for calls).
All stakeholders are enthusiastic about using the practice. The only stakeholders whose resources are used specifically for ECNS are emergency management and public health. Emergency management uses the system for many kinds of targeted emergency notifications to the community, and is committed to maintaining the system. Public health, even in the current fiscal climate, is committed to using this low-cost practice, which realizes fiscal and human resource savings. No resources beyond those already committed are necessary. The use of the practice is enshrined in policy.