This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
One of the essential services provided by the US local health departments is informing and educating constituents about health. Communication with constituents about public health issues and health risks is among the standards required of local health departments for accreditation. Past research found that only 61% of local health departments met standards for informing and educating constituents, suggesting a considerable gap between current practices and best practice.
Social media platforms, such as Twitter, may aid local health departments in informing and educating their constituents by reaching large numbers of people with real-time messages at relatively low cost. Little is known about the followers of local health departments on Twitter. The aim of this study was to examine characteristics of local health department Twitter followers and the relationship between local health department characteristics and follower characteristics.
In 2013, we collected (using NodeXL) and analyzed a sample of 4779 Twitter followers from 59 randomly selected local health departments in the United States with Twitter accounts. We coded each Twitter follower for type (individual, organization), location, health focus, and industry (eg, media, government). Local health department characteristics were adopted from the 2010 National Association of City and County Health Officials Profile Study data.
Local health department Twitter accounts were followed by more organizations than individual users. Organizations tended to be health-focused, located outside the state from the local health department being followed, and from the education, government, and non-profit sectors. Individuals were likely to be local and not health-focused. Having a public information officer on staff, serving a larger population, and “tweeting” more frequently were associated with having a higher percentage of local followers.
Social media has the potential to reach a wide and diverse audience. Understanding audience characteristics can help public health organizations use this new tool more effectively by tailoring tweet content and dissemination strategies for their audience.
In the United States, local health departments are governmental agencies providing essential public health services in cities, counties, metropolitan areas, districts, and tribal areas [
Approximately 80% of US adult Internet users have searched for health information online, making the Internet second only to health care providers as a source for health information [
Understanding the audience intentionally receiving health information is key to successful health communication [
As of July 30, 2012, we identified 217 local health departments nationwide using Twitter (identification process described elsewhere [
Location: Where is the Twitter follower located?
In-state where local health department is located
Out-of-state (in United States)
Outside United States
Unable to determine (in United States)
Unable to determine
Type: Based on the username and description, does the follower account appear to be for an...?
Individual
Organization/business
Unable to determine
Industry: Which of the following industries best describes the follower based on description or linked website?
Spam
Unable to determine
Health-focused
Private user
Educational institution
US government
Local
State
National
Campaign/program
Not-for-profit (nongovernmental)
Local
State
National
International
Campaign/program
For profit
Hospital or hospital system
Private physician or physician offices
Drug company or representative
Managed care
Patient advocate
Medical device maker, seller
Fitness center/gym/personal training
Diet/nutrition (eg, Weight Watchers, Jenny Craig)
Assisted living
Media
TV
Newspaper/magazine (print media)
Radio
Social media or website (not affiliated with TV, radio, newspaper)
Public relations firm
Other (make note)
Not health-focused
Private user
Educational institution
US government
Local
State
National
Campaign/program
Not-for-profit (nongovernmental)
Local
State
National
International
Campaign/program
For profit
Media
TV
Newspaper/magazine (print media)
Radio
Social media or website (not affiliated with TV, radio, newspaper)
Public relations firm
Other (make note)
Project team members reviewed the Twitter follower information to develop a coding scheme with 3 broad categories: location, type, and industry. Location was an indicator of whether the Twitter follower was in the same state with the local health department, out-of-state (in the United States), outside the United States, unable to tell but within the United States, and unable to determine. Type of follower included 3 categories: individual, organization/business, and unable to determine. To discern individual followers, we looked for information written in the first person using “I” or “my” or other descriptors or wording indicating the user was an individual. Organizational accounts, on the other hand, often included a statement of organizational purpose. Industry was divided into 4 subgroups: health-related, non-health-related, spam, and unable to determine. Classification of an account as spam occurred when the account did not appear to be a legitimate person or business. For example, one spam account was following 2002 others, had 67 followers, and had never tweeted. Another spam-classified account had no user description, 49 followers, and had tweeted over 1000 times with nearly all the tweets being retweeted from another user, suggesting automated retweeting by a spambot, or a program that is designed to send out spam. Within industry were several specific types of organizations. If the industry was not easy to glean from the Twitter profile (eg, “St. Mary’s is a nonprofit hospital”), we searched the user’s website and often found the type on the About Us page. Consistent with previous research [
To test the reliability of the coding scheme, 4 coders coded data from the same 100 Twitter followers (2.09%) randomly selected from the overall sample of 4779. Krippendorff’s alpha for nominal data was computed for each of the 3 broad coding categories (type, location, industry). For follower type, alpha was .70 (95% CI .63-.77), for follower location alpha was .88 (95% CI .84-.91), and for industry alpha was .68 (95% CI .64-.72). Given these acceptable reliability scores, the full dataset was divided and coded independently by the 4 coders.
For the first aim of the study, understanding the general characteristics of local health department Twitter followers, frequencies and percentages were used to examine the distribution of follower types. Chi-square tests were used to determine whether certain types of followers were more or less likely than expected given the overall distribution of followers. For example, were individual users more or less likely than expected to be local? Standardized residuals were calculated for significant chi-square test results; standardized residuals greater than 1.96 indicated significantly more followers than expected fell into a given category, whereas standardized residuals less than –1.96 indicated significantly fewer followers than expected fell into a category. Followers classified as “unable to determine” were omitted from analyses.
The second and third aims were to understand the relationship between local health department characteristics, Twitter use, and characteristics of their Twitter followers. For these aims, follower data was aggregated by local health department to compare the proportions of follower types in local health departments varying by (1) Twitter use (number of followers, number of tweets) and (2) resources (population size, staffing, funding per capita). Twitter usage was obtained through NodeXL in April 2013. Local health department resource information was obtained from the NACCHO 2010 Profile Study.
Because local health departments provide services to their local constituents, it is important to know what is associated with reaching local individuals. Local health departments with a public information officer may have more, and more organized, information-sharing efforts in the local community given this specialized staffing. Larger jurisdiction population and a higher number of tweets are associated with more followers [
Local health departments with a public information officer have a higher proportion of local followers and individual followers than local health departments without a public information officer.
The more Twitter followers and tweets a local health department has, the higher the percentages of local and individual Twitter followers there will be.
The larger the jurisdiction population, the higher the percentages of local Twitter followers and individual Twitter followers there will be.
Other ways for local health departments to reach and inform local constituents could be through local media and local government. Journalists have adopted social media as sources of information, with more than 30% of print journalists deeming social media as important or very important as of 2009 [
Local health departments with a public information officer have a higher proportion of local media followers (TV, radio, and newspaper) and local government followers than local health departments without a public information officer.
The larger the jurisdiction population, the higher the percentages of local media (TV, radio, and newspaper) followers and local government followers there will be.
Hypotheses 1, 3, 4, and 5 aid in addressing aim 2 (understanding the relationship between local health department characteristics and Twitter use), whereas hypothesis 2 aids in addressing aim 3 (understanding the relationship between local health department characteristics and characteristics of their Twitter followers).
The 59 local health departments in the sample had between 9700 and 3 million constituents in their local jurisdictions according to the 2010 NACCHO Profile Study. In all, 29 (49%) of the departments reported having a public information officer, whereas 23 (39%) reported not having one (7/59, 12% were missing data on this staffing). The median number of Twitter followers was 218 (range 7-11,827; mean 770.1, SD 1688.0); the median number of sent tweets per health department since adopting Twitter was 324 (range 0-5849; mean 667.9, SD 1083.1). Health departments in the sample joined Twitter between June 2008 and January 2012; more than half (34/59, 58%) joined in 2009.
Overall, we found that local health departments had more Twitter followers that were organizations (2591/4434, 58.43%) than individuals (1843/4434, 41.57%). Of the 1843 individuals, 1267 (68.75%) were private personal accounts (private users not affiliated with a specific organization or business). The 1267 private individuals comprised 29.07% of the follower industry classifications for the 4359 classified by industry (
Omitting the private user category because it only applied to individuals, follower industry was significantly associated with follower type (χ2
6=112.6,
Considering all cases (including private users), industry was associated with health focus (χ2
7=783.3,
Characteristics of local health department Twitter followers.
Characteristic | All | Individuals | Organizations |
|
||||
|
n | % | n | % | n | % |
|
|
|
4434 |
|
|
|
|
|
|
|
|
Individual | 1843 | 41.57 | — | — | — | — |
|
|
Organization | 2591 | 58.43 | — | — | — | — |
|
|
4359 |
|
477 |
|
1551 |
|
<.001a | |
|
Private user | 1267 | 29.07 | — | — | — | — |
|
|
Education | 160 | 3.67 | 24 | 5.03 | 136 | 5.33 |
|
|
Government | 497 | 11.40 | 71 | 14.88 | 426 | 16.70 |
|
|
Nonprofit | 556 | 12.76 | 31 | 6.50b | 525 | 20.58c |
|
|
For-profit | 1053 | 24.16 | 174 | 36.48 | 868 | 34.03 |
|
|
Media | 599 | 13.74 | 160 | 33.54c | 422 | 16.54b |
|
|
Other | 192 | 4.40 | 16 | 3.35b | 170 | 6.67 |
|
|
Spam | 35 | 0.80 | 1 | 0.21 | 4 | 0.16 |
|
|
4340 |
|
1738 |
|
2563 |
|
<.001 | |
|
Yes | 1748 | 40.28 | 419 | 24.11b | 1303 | 50.83c |
|
|
No | 2592 | 59.72 | 1319 | 75.89c | 1260 | 49.16b |
|
|
3878 |
|
1371 |
|
2417 |
|
<.001 | |
|
In-state | 2149 | 55.42 | 840 | 61.27c | 1281 | 53.00b |
|
|
Out-of-state | 1195 | 30.81 | 363 | 26.48b | 815 | 33.72c |
|
|
In US state unknown | 337 | 8.69 | 41 | 2.99b | 147 | 6.08b |
|
|
Outside US | 197 | 5.08 | 127 | 9.26 | 174 | 7.20 |
|
aPrivate person was omitted for the purposes of bivariate analysis.
bMore followers than expected fell into this category (standardized residuals >1.96).
cFewer followers than expected fell into this category (standardized residuals <–1.96).
Examples of common categories of local health department Twitter followers.
Follower type | Health focus | Twitter user description |
Private person | No | I love music and travel. Watching movies and sunsets. I like to play World of Warcraft. |
|
Yes | Huggable Health Educator |
Nonprofit organizations | No | IBA is a nonprofit agency dedicated to empowering individuals and families through education, economic development, technology and the arts. |
|
Yes | Advocating for the health and dignity of Denver’s injection drug users in accordance with #harmreduction principles. Syringe exchange and Naloxone Save Lives! |
Media individuals | No | San Diego reporter at The Daily Transcript. Freelance sports writer for Southwest Riverside News Network. Associate Producer for KUSI Prep Pigskin Report. |
|
Yes | Journalist covering medical/health & fitness and writing features for the Tyler Morning Telegraph. |
On average, more than half of a local health department’s Twitter followers were from within the state (5%, range 10%-91%). Consistent with the overall composition of the follower sample, 70% (41/59) of local health departments were followed by a higher proportion of organizations than individuals (range 22%-83% organizational followers). An even greater number of departments had a majority of non–health-focused followers (compared to health) with 44 of 59 (75%) departments having more than 50% non–health-focused followers (range 17%-86%). The average percentage of in-state media followers (TV, newspaper, radio) was 6% (SD 5.8) with a range from 0 to 25%. The average percentage of in-state government followers was 7% (SD 8.1) with a range from 0 to 43%.
We hypothesized that the local health departments with a public information officer have a higher proportion of local followers and individual followers
than local health departments without a public information officer. A
We hypothesized that, the more Twitter followers and tweets a local health department had, the higher the percentages of local and individual Twitter
followers there was. There was a strong positive correlation between the number of followers and the number of tweets for a local health department (
We hypothesized that, the larger the jurisdiction population, the higher the percentages of local Twitter followers and individual Twitter followers. A correlation coefficient indicated a positive and significant, although weak, association between jurisdiction population and percent of followers who were local (
We also hypothesized that local health departments with a public information officer have a higher proportion of local media followers (TV, radio, and
newspaper) and local government followers than local health departments without a public information officer. The
Finally, we hypothesized that, the larger the jurisdiction population, the higher the percentages of local media (TV, radio, and newspaper) followers and local government followers. The size of the jurisdiction population was not associated with the percent of followers who were local government or local media.
The results of our research reveal that local health departments on Twitter in the United States are followed by more organizations than individuals. Many of the organizations are health-focused, out-of-state, and from the education, government, and nonprofit sectors, suggesting that there may be a communication network comprised of organizations with a health-related mission developing on Twitter. Individual followers of local health departments, on the other hand, tended to be locally based and largely did not have a health focus, indicating that, at least where individuals are concerned, local health departments may not be just tweeting to members of the public who are already health-focused (ie, the choir).
Evidence of a network of public health organizations developing on Twitter is consistent with at least 2 recent studies [
Consistent with other studies [
Although tweeting more often and serving larger populations has been associated with having more followers overall for local health departments [
Our findings may be useful for local health departments in at least two ways. First, we identified the characteristics (eg, jurisdiction population size, employing a public information officer) and practices (eg, tweet frequency) associated with local health department Twitter follower characteristics. This information could inform strategic planning for local health departments using or considering using Twitter. For example, if a goal of a local health department is to reach greater numbers of individuals rather than organizations, their Twitter strategy could include a regular daily or weekly tweeting schedule.
Second, understanding who the Twitter followers are could help local health departments better target tweets to diverse audiences. For example, local media and policymakers may be important followers for a local health department. Standard strategies (eg, tweeting more, developing an easy-to-find user profile [
Some local health departments are already focused on reaching specific individual and organizational audiences. For example, the Chicago Department of Public Health has begun conducting campaigns using social media as a dissemination channel and making explicit efforts to interact with local individuals through Twitter activities, such as Twitter live chats. One of these events took place in early 2013, when Dr Julie Morita of the Chicago Department of Public Health answered questions from Chicagoans in a Twitter live chat about flu (using hashtag #FluChicago) just as local news coverage of flu was increasing. Through their official Twitter account (@ChiPublicHealth), she answered questions ranging from the Chicago mayor asking about prevention when in close contact with many people (
Other health departments have adopted Twitter with the purpose of sharing information with public health organizations. For example, local health departments across Utah made a statewide effort to adopt and use social media. A local health department practitioner at Bear River Health (@BearRiverHealth) in Utah described this strategy as follows, “Not only does it allow us an opportunity to share information, it allows us to communicate in a new way with the communities that we serve together as a state. For example, when we launch an immunization campaign we now have the ability to share the same message seamlessly across our entire state through Twitter and Facebook. We share one another’s posts, comment on status, and generally connect” (Jill Parker, personal communication, November 2012). This coordinated effort and active use of Twitter has resulted in Utah communities such as the jurisdiction of Bear River Health with a 2010 population of 163,836 to reach more than 3200 followers, more than 5 times the average number of followers for a local health department [
Chicago mayor Rahm Emanual participating in the #FluChicago 2013 Twitter chat about flu prevention with @ChiPublicHealth during flu season (Photo from Chicago Mayor’s Office, @ChicagosMayor [
Limitations to this study include cross-sectional data, reliance on self-reported information, and a lack of information on follower engagement. For example, because many Twitter users do not include geographically specific information about their location in the user profile, the coding of local was limited to identifying whether a follower was in the same state as the local health department. Likewise, a Twitter user may have had a health focus, but did not include health-related language in their profile and was coded as nonhealth. In addition, without additional information about Twitter follower engagement, such as mentions and retweets, it is impossible to know the extent to which the followers were actively engaged with the health departments through their Twitter accounts [
National Association of County and City Health Officials
Support for this research was provided by a Mentored Research Scientist Development Award from the Robert Wood Johnson Foundation through the National Coordinating Center for Public Health Services and Systems Research.
None declared.