Communicating with At-Risk Populations
Although the Commercial Mobile Alert Service (CMAS) will send alerts to people’s cellular telephones and thus reach a large fraction of the population, this approach will also present special challenges for certain at-risk segments of the population. In the workshop session on communicating with at-risk populations, Judy Harkins, Gallaudet University, discussed research performed at Gallaudet University on how best to provide alerts to people who are deaf and hard of hearing. Ed Price, Georgia Institute of Technology, discussed research completed at the Georgia Institute of Technology’s Rehabilitation Engineering Research Center for Wireless Technologies (Wireless RERC) on how people with hearing, sight, physical, and cognitive disabilities interact with cellular telephones and alerts. Chris Mayhorn, North Carolina State University, shared insights on reaching older adults. Brenda Phillips, Oklahoma State University, served as moderator and highlighted the interactions of gender and ethnicity on emergency preparation, education, and response. The rest of this chapter provides an integrated summary of these presentations and the discussions that followed, organized by topic.
USE OF WIRELESS DEVICES BY PEOPLE WITH DISABILITIES
The Wireless RERC uses ongoing surveys, which have over the past few years involved more than 1,600 people with disabilities, to collect data on use trends and user needs. People with disabilities have important communications needs, often using mobile devices as a lifeline, not
only during medical emergencies but also daily to further independence and access to services. For people with disabilities who are unable to drive themselves, mobile devices can be imperative in arranging for transportation. Indeed, the penetration of wireless devices in communities of people who have disabilities is the same, if not slightly higher, than in the community at large.
This finding, and the observation that text messages are increasingly used to deliver alerts and warnings, prompted the Wireless RERC to conduct a series of field trials, focus groups, and simulations to investigate a variety of approaches to delivering text alerts. The study population was primarily individuals who are blind or have low vision, or who are deaf or hard of hearing, along with a few participants with cognitive disabilities; the study population included people with varying levels of technical savvy. The tests used two different devices: (1) a Windows mobile device with custom software to provide a tone alert similar to that used in the Emergency Alert System (EAS), other customized attention signals, and text-to-speech processing; and (2) standard BlackBerry devices that support text messaging and are commonly used in the community of people who are deaf and hard of hearing.
The Wireless RERC studies were completed before an initial set of requirements for the CMAS program was set forth by the 2007 Commercial Mobile Service Alert Advisory Committee (CMSAAC) report.1 The studies used a standard short message service (SMS) text message and a Web page, putting essential information in the SMS body and including a hyperlink to the full alert on the Web page. The first test group, composed of people who are blind and vision-impaired, used the custom device. These test subjects found that the tone alerts and speech synthesis were a significant improvement over the phones that they normally used. The second test group, composed primarily of people who are deaf and hard of hearing, used the BlackBerries. They found the text message alerts to be useful but not a big improvement over the systems that they normally used. For example, many in this group already subscribed to alerts from third-party providers such as the Weather Channel. For them, the EAS-like alert was slightly preferred because it provided somewhat more detailed information and did not contain advertisements. Notably, in post–field surveys, 83 percent of people with sensory limitations said that receiving emergency alerts by way of wireless devices was highly desirable.
After the 2008 adoption of the CMSAAC recommendations by the FCC, Wireless RERC conducted a second series of tests using CMAS-like messages (90-character messages that did not contain links to second-
ary sources) and alerting by means of a distinctive tone and vibration cadence. About three-quarters of those who had participated in previous tests thought that this CMAS-like alert was an improvement over the third-party alerts with which they were familiar. They observed that CMAS-like messages were simple and that they clearly indicated what action to take, but these people suggested that it would be useful if links to supplemental information were included.
SPECIAL CONSIDERATIONS FOR PEOPLE WHO ARE BLIND OR HAVE LOW VISION
Text alerts are inherently ill suited for people who are blind unless the telephone that they are using has the ability to read the text—that is, has text-to-speech capabilities. Without the inclusion of text-to-speech capabilities, CMAS-like messages simply cannot be used by those who are blind. Moreover, there are questions about the more general usability of cell phones by people who are blind—the phones that they commonly employ fall into two categories. The first category consists of much older telephones with fewer features and settings and thus very simple interfaces that can readily be memorized. However, these older phones cannot receive CMAS messages, and it is unclear whether new phones that do support CMAS will be simple enough to be used by those who are blind. The second category consists of new smartphones that have built-in voice commands and voice menus. Today, these capabilities are only found on high-end smartphones, which may be too expensive for many.
SPECIAL CONSIDERATIONS FOR THOSE WITH IMPAIRED HEARING
Fortunately, there are several resources for emergency alerting that already exist for those who have a hearing impairment. E-mail and SMS alerting systems are already understood to work well. Television coverage of emergencies provides a great deal of text and visual graphics, so that even those who cannot hear what is being said can glean at least some of the needed information from the media coverage. There are National Oceanic and Atmospheric Administration (NOAA) weather radios with text displays, flashing lights to garner attention, and vibration functions to awaken a person. However, the content of the text component is much more limited than the audio (speech) information that is broadcast to the radios. Telephone emergency notification systems are not particularly useful for those with impaired hearing, as fewer people with impaired hearing have a land line, and most telephone alerting systems do not provide a TTY option, which would allow a person who is deaf to receive the calls
as text. One of the greatest challenges is the “eye-busy” situation in which the attention of a person with impaired hearing is focused elsewhere. The attention-getting mechanism must be disruptive enough to cause the user to change focus immediately.
Once the existence of an emergency is known by a person who has a hearing impairment, the Web is extremely useful for that person. As discussed previously, social and new video can play an important role. However, any video posted by emergency response or management needs to be captioned or it is not particularly useful for those with a hearing impairment. During the San Diego wildfires, a local organization posted video communicating in American Sign Language (ASL) for those with deafness, indicating that there were resources as well interpreters at a local shelter.
When people who are deaf are not at home, it becomes even more difficult to alert them to potential crises and disasters. Conventional sources of emergency warnings—car radios, sirens, and public-address systems—are inaccessible to them. Communication with bystanders and often even communication with responders are not possible. CMAS will play an important role in reaching people who have impaired hearing when they are not at home.
Gallaudet University’s experience with opt-in emergency notification highlights the importance of CMAS participation’s being opt-out. Gallaudet had an opt-in e-mail and mobile alerting system; however, only 15 percent of those under the age of 25 registered. The university has since moved to an opt-out e-mail alert. Students have their phones with them constantly, but if they cannot be induced to opt in, the system is useless. Students claim that they choose not to participate because too-frequent alerts would be annoying.
Tone and Vibration
An important component of CMAS is the unique tone to alert cellular telephone users that the message being received is different from other messages. The current specified tone is the EAS two-tone signal at 960 and 853 hertz. The cell phone industry supported these tones because most ring tones are in the high range. However, these frequencies can be difficult for those who are hearing-impaired, as their hearing capability usually resides in the lower frequencies. This challenge will need to be considered during future iterations of CMSAAC recommendations.
CMAS will also provide a distinct vibration cadence for people with impaired hearing. A recent study of a small sample of Gallaudet students and staff was done to determine what the best vibration cadence might
be.2 Participants rated the following four vibratory temporal patterns presented on a mobile: (1) no pattern/constant, (2) even on-off, (3) long and short pattern similar to the one selected by CMAS, and (4) long and buzzy short pulses. Three different durations of each pattern were tried: 4 seconds, 8 seconds, and 12 seconds. Participants evaluated the effectiveness of the patterns and the similarity to their own devices’ default tone, responding to the question of whether the patterns would be sufficiently distinguishable to get their attention in an emergency. Based on the results of this study, the signal needs to be long. The best ratings were for single signals at about 12.5 seconds. The even-on-off pattern, which is currently specified by CMAS, was somewhat more preferred. Although a temporal and unique pattern is important, the nature of the specific pattern is secondary to length in importance, according to the study.
The study at Gallaudet University did not examine the strength of the vibration, as that cannot be changed on current mobile devices. However, it is important to note that as devices become smaller, vibrations are becoming weaker. This could pose a challenge for future generations of alerting systems.
American Sign Language
Wireless RERC completed some research with people who are deaf who use American Sign Language (ASL) as their primary language. For most people who are born with deafness or who are early deafened, ASL, not English, is likely to be their primary language, and some may have very limited proficiency in English. (ASL and English are distinct languages, with quite different grammatical structures.) Wireless RERC simulated ASL alerts on smartphones. Test subjects received a CMAS-type 90-character message followed by a video of someone signing the alert. Most people thought that the combination of text and ASL was better than either one alone, suggesting that ASL may be a desirable capability in future iterations of CMAS.
The research conducted by Wireless RERC also found that the common terminology used in the National Weather Service alerts, such as “Take cover” or “Low-lying area,” do not translate well into ASL and that not all people understood those terms. People in different parts of the country may use different ASL terms, but the advantage here at least is that National Weather Service alerts are also regionalized. The
National Weather Service should work more closely with the community of people who are deaf to identify more easily understood words and descriptions.
SPECIAL CONSIDERATIONS WITH RESPECT TO DISABILITIES IN THE ELDERLY POPULATION
The overall population of the world is aging. Based on projections, by 2050 at least a full 20 percent of the U.S. population, or 70 million people, will be 65 years of age or older.3 During disasters or crises, older adults are vulnerable for several reasons. They are much more likely to become casualties during a disaster, to suffer long-term psychological distress, and to recover economically more slowly.
Every sensory channel can become less sensitive as individuals age. For example, reduced fine motor control makes cellular telephone buttons and keys difficult to manipulate, a problem that is compounded for those with arthritis of the fingers. Research looking at the usability of several alternatives—touch screens, larger keyboards, and voice input—has found that touch screens are the easiest for elderly persons to use and can enable older adults to achieve performance comparable to that of younger adults. However, these solutions can present their own challenges, including accidental activation and arm fatigue.
Declines in vision can also impair effective response to warnings and alerts in older adults. Older adults generally have greater susceptibility to glare and difficulty distinguishing between certain colors, for example, blue and green. For CMAS and mobile devices, text for older persons needs to be made more readable through the use of 12- or 14-point, sans serif fonts and the avoidance of colors that are difficult to distinguish between. Older adults also tend to have difficulty hearing higher frequencies. As discussed earlier, most ring tones are in the higher frequencies, and this could prevent older adults from hearing the alerts. The preference of a particular vibratory cadence may differ by age.
In addition to an increase in visual and auditory disabilities or impairments that often occurs in older persons, there are comorbidities associated with age. Older adults thus often face multiple hindrances in receiving and responding to alerts and warnings delivered by mobile devices.
There are also significant challenges for those with diminished cognitive abilities. Steps such as comprehending an alert, seeking additional information, and deciding on an appropriate decision to take protective action rely heavily on cognitive abilities such as attention and memory.
As people age, these abilities may diminish as a result of changes in perception, motor abilities, and cognition as well as changes in memory and reading comprehension skills. Research on cognitive aging indicates that there are often deficits in the selective attention of older adults, meaning that they may have difficulty identifying what information is important. To alleviate such limitations caused by cognitive challenges, it is useful to direct attention to the specific parts of a message that are the most important. Older adults may also experience memory deficit and may have problems simultaneously processing information while reading text. Cognitive overload or information overload can tax working memory, which suggests the importance of not sending multiple messages in rapid succession and of avoiding overly complex instructions and jargon.
All of the factors described above can decrease the usability of cellular telephones by older people, which is a major factor in reducing the adoption of cell phones by this group. Fifteen percent of Americans do not access the Internet on a regular basis; most of these are older adults. Only 25 percent of adults 65 years of age or older have cell phones. Furthermore, owning a cell phone does not necessarily mean the owner can use the device. Older adults are not generally viewed as early adopters of new technologies, but it is a misconception that they are technophobes. Their choice of whether or not to use cell phones or other information technology depends on the technology’s utility and ease of use. Persons who are elderly will probably need to be trained not only to use the mobile devices but also educated about how useful the alerts and warnings system would be.
The importance that gender can have in people’s behavior during disasters was underscored by the 2004 Indian Ocean tsunami, in which a large majority of fatalities in some communities were women and children.4 Many women drowned because they had not been taught to swim, in part because their customary role in the culture is to bring fish caught by the men to market. Another cultural factor involved in these drownings was the clothing customarily worn by women, in which they became entangled as they tried to escape the flood waters or search for their children. Children, who were most likely to be looked after by women, were also placed at higher risk.
In the United States, an important gender-based social pattern is that women are more likely than men to be the caregivers for children and elderly relatives, who are more likely to be at risk in a disaster. When
issuing messages, consideration needs to be given to the recipients of the message and also to those who surround the people at risk. Another significant gender-based factor is that women tend to act sooner than men when a warning or alert is issued, which places men at greater risk. During Hurricane Mitch in 1998, a greater number of men than of women died.5 This is generally attributed to the desire of men to protect their resources and family homes. Coinciding with studies highlighting the isolation of older populations, particularly of older males, data on Hurricane Katrina show that older African-American men tended to die disproportionately compared to other populations.6
CONSIDERATIONS RELATED TO RACE AND ETHNICITY
As discussed in Chapter 4, educational initiatives need to take into account an ethnic community’s usual information channel. This information should inform the development of public education campaigns.
Preferred information sources (television, radio, or online) can vary across racial and ethnic lines. For example, it has been reported that many Mexican-Americans prefer to get information on community initiatives and programs at neighborhood meetings. Race and ethnicity also play a role in the public’s response to alerts.7 It is important to identify where credibility and trust lie and to use those avenues. Furthermore, racial and ethnic minorities are more likely to seek multiple confirmations from informal sources and to delay taking protective action. This was particularly apparent during Hurricane Katrina, when affected populations first gathered multiple generations living within the area before making a decision on protective action.
Language can also create a challenge in receiving messages among populations with limited proficiency in English. A tornado hit Saragosa, Texas, in 1987. Unfortunately an English-language warning was translated incorrectly, and Spanish-speaking people thought that they were getting
The World Bank. “Hurricane Mitch—The Gender Effects of Coping and Crises.” PREMnotes, No. 57, August 2001. Available at http://www1.worldbank.org/prem/PREMNotes/premnote57.pdf. Accessed December 21, 2010.
Sebastian N. Jonkman, Bob Maaskant, Ezra Boyd, and Marc Loyd Levitan. “Loss of Life Caused by the Flooding of New Orleans After Hurricane Katrina: Analysis of the Relationship Between Flood Characteristics and Mortality.” Risk Analysis 29:676-698 (2009).
For additional information, see Ronald W. Perry and Lisa S. Nelson, “Ethnicity and Hazard Information Dissemination,” Environmental Management 15(4):581-587 (1991); Ronald Perry and A. Mushkatel, Minority Citizens in Disasters, University of Georgia Press, Athens, Ga., 1996; Ronald Perry and M. Lindell, Communicating Environmental Risk in Multiethnic Communities, Sage, Thousand Oaks, Calif., 2004; or A. Fothergill, E.G.M. Maestas, and J.D. Darlington, “Race, Ethnicity and Disasters in the United States: A Review of the Literature,” Disasters 23:156-173 (1999).
news of a tornado, not a warning of a tornado. People were gathered at community centers, and so they did not have access to radio. Others were home watching Spanish-language television, which was not broadcasting the message. These gaps in message receipt led to a large number of people not taking the appropriate protective action.
OBSERVATIONS OF WORKSHOP PARTICIPANTS
The panelists and participants in the discussion following the panel offered the following observations regarding communicating with at-risk populations:
Affordability and accessibility have to be considered in developing warning systems and designing new technologies. A technology that is out of reach for a large segment of the population loses a great deal of its usefulness.
Poor literacy is another challenge, which suggests that message testing needs to be done with a diverse set of test users.
In addition to taking into account the challenges faced by elderly persons, it is also important to consider the use of alerts and warnings by children. For example, how will a child who is home alone respond to an alert?
Credibility of the person or system conveying an alert or warning message is critical to ensuring that people take appropriate action, and people tend to trust “people like themselves.” This suggests the need for attention to diversity in educational campaigns and message formulation.