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 JMIR Public Health and Surveillance, is properly cited. The complete bibliographic information, a link to the original publication on http://publichealth.jmir.org, as well as this copyright and license information must be included.
Twitter is a free social networking and microblogging service on the Internet. Medical professionals and patients have started to use Twitter in medicine. Twitter use by medical institutions can interactively and efficiently provide public health information and education for laypeople.
This study examined Twitter usage by medical institutions.
We reviewed all Japanese user accounts in which the names of medical institutions were described in the user’s Twitter profile. We then classified medical institutions’ tweets by content.
We extracted 168 accounts for medical institutions with ≥500 followers. The medical specialties of those accounts were dentistry and oral surgery (n=73), dermatology (n=12), cosmetic surgery (n=10), internal medicine (n=10), ophthalmology (n=6), obstetrics and gynecology (n=5), plastic surgery (n=2), and others (n=50). Of these, 21 accounts tweeted medical knowledge and 45 accounts tweeted guidance about medical practice and consultation hours, including advertisements. In the dentistry and oral surgery accounts, individual behavior or thinking was the most frequent (22/71, 31%) content. On the other hand, consultation including advertisements was the most frequent (14/23, 61%) in cosmetic surgery, plastic surgery, and dermatology.
Some medical specialties used Twitter for disseminating medical knowledge or guidance including advertisements. This indicates that Twitter potentially can be used for various purposes by different medical specialties.
Twitter is a free social networking and microblogging service on the Internet. Twitter has grown rapidly in popularity [
Medical professions have started to use Twitter [
Information about Twitter use by medical institutions is highly limited and the optimal methods of use remain unclear. In this study, we examined Twitter usage by medical institutions and Twitter’s role in public medical information. This should assist medical institutions that use Twitter in making significant changes from noninteractive, 1-way publicity to interactive information exchange, wherein personal opinions, needs, and requests are disseminated and collected in the future.
We searched Twitter profiles using a method detailed elsewhere [
We collected the latest tweets of the top 5 accounts in terms of number of followers for measuring longitudinal changes. We extracted tweets from 2012 to 2014. The website we used for collecting the latest tweets was TwimeMachine [
We searched the accounts of medical institutions in English-speaking countries for comparison with the accounts of medical institutions in Japan. The search words were “hospital” and “clinic” in English. We extracted the accounts with the top 10 number of followers, reading medical institution profiles. The website used for collecting the accounts of medical institutions in English-speaking countries was followerwonk [
We classified the medical institution tweets into 6 categories using the 100 most recent tweets. The categories were medical knowledge (category A); consultation guidance (including advertisements and newsletters from medical institutions; category B); suggestions from patients (category C); links to other home pages only (category D); individual behavior or thinking (category E); and tweets with multiple kinds of content (eg, tweets with a URL link to a home page and medical knowledge tweets from the same account, classified by YS, a medical informatics researcher; category F). In the most recent 100 tweets, the most categorized tweets from categories A to E were the account classifications. When categories A to E were almost equal, we classified the accounts as category F. The categories for tweets that the primary investigator (YS) could not categorize were decided by discussion with the secondary investigator (NH).
Tweet morphemes in Japanese were analyzed using the method previously developed for frequency comparison of nouns by medical specialties [
We identified 1211 Twitter accounts with ≥500 followers in which the keywords “hospital” and “clinic” were included in the profile. From those, we extracted 168 accounts by medical institutions using the following keywords in Japanese:
Categorized Twitter accounts (relevant Twitter accounts with ≥500 followers on July 15, 2013).
Type of Twitter account | Search words | Total no. of accounts | ||||
|
|
|
|
|||
|
|
|
|
|
|
|
|
Medical institutions | 30 | 6 | 67 | 65 | 168 |
|
Pharmacies | 2 | 0 | 0 | 0 | 2 |
|
Doctors, dentists | 86 | 12 | 39 | 53 | 190 |
|
Nurses, midwives | 35 | 1 | 0 | 6 | 42 |
|
Paramedics (except nurses) | 71 | 1 | 10 | 5 | 87 |
|
Other hospital personnel | 68 | 4 | 3 | 7 | 82 |
|
Tweets about medical knowledgea | 18 | 0 | 6 | 5 | 29 |
|
Patients | 11 | 0 | 1 | 1 | 13 |
|
Total | 321 | 24 | 126 | 142 | 613 |
Accounts unrelated to medicine | 356 | 31 | 58 | 153 | 598 | |
Total no. of accounts | 677 | 55 | 184 | 295 | 1211 |
aThe accounts’ proprietors were unknown.
The medical specialties of those accounts were dentistry and oral surgery (n=73), dermatology (n=12), cosmetic surgery (n=10), internal medicine (n=10), ophthalmology (n=6), obstetrics and gynecology (n=5), plastic surgery (n=2), and others (n=50;
We identified 190 Twitter accounts of doctors or dentists in this search, including 10 pediatricians. We searched “pediatrician” using 16 Profile Search beta version for Twitter on March 12, 2014 and identified 12 pediatrician accounts with ≥500 followers.
The top 10 accounts by number of followers of Japanese medical institutions.
No. | Account name | Hospital or clinic | Medical specialty | Prefecture | Date joined | Total tweets | Followers | Tweets/daya |
1 | @team_nakagawa | Hospital | General hospital |
Tokyo | March 15, 2011 | 285 | 169,407 | 0.33 |
2 | @Sanhujinka | Clinic | Obstetrics and gynecology | Kanagawa | April 7, 2010 | 81,753 | 90,906 | 68.41 |
3 | @biyoudoctor | Clinic | Cosmetic surgery | Tokyo | September 20, 2010 | 4603 | 48,151 | 4.47 |
4 | @SBCLASIK | Clinic | Ophthalmology | Tokyo | March 8, 2010 | 6303 | 45,063 | 5.15 |
5 | @sbcmatsuoka | Clinic | Dermatology | Tokyo | September 24, 2010 | 1559 | 32,609 | 1.52 |
6 | @GVBDO | Clinic | Dentistry | Gifu | May 21, 2010 | 31,568 | 27,121 | 27.43 |
7 | @ryouki4181 | Clinic | Dentistry | Osaka | July 6, 2010 | 1422 | 21,365 | 1.29 |
8 | @suzuki1855 | Clinic | Dentistry | Yamagata | November 9, 2009 | 1656 | 18,914 | 1.23 |
9 | @Drponchi | Clinic | Internal medicine | Kanagawa | May 19, 2011 | 7370 | 16,852 | 9.35 |
10 | @moriyukabc | Clinic | Cosmetic surgery | Mie | December 5, 2010 | 359 | 13,916 | 0.38 |
Median |
|
|
|
|
3129.5 | 29,865 | 3.00 |
aUp to July 15, 2013.
The top 10 accounts by number of followers of medical institutions in English-speaking countries.
No. | Account name | Hospital or clinic | Medical specialty | Country | Date joined | Total tweets | Followers | Tweets/daya |
1 | @MayoClinic | Hospital | General hospital | US | April 30, 2008 | 25,859 | 1,233,111 | 9.44 |
2 | @HarvardHealth | Hospital | General hospital | US | March 4, 2009 | 2,898 | 998,726 | 1.19 |
3 | @ClevelandClinic | Hospital | General hospital | US | March 14, 2009 | 27,644 | 542,697 | 11.42 |
4 | @StJude | Hospital | Children’s hospital | US | October 23, 2007 | 6,567 | 388,754 | 2.24 |
5 | @HopkinsMedicine | Hospital | General hospital | US | February 7, 2009 | 11,469 | 329,343 | 4.67 |
6 | @FUEHairClinics | Clinic | Cosmetic surgery | UK | August 10, 2013 | 738 | 97,286 | 0.91 |
7 | @GreatOrmondSt | Hospital | Children’s hospital | UK | October 22, 2009 | 20,703 | 82,405 | 9.41 |
8 | @HospitalsApollo | Hospital | General hospital | India | October 25, 2010 | 26,494 | 77,120 | 14.47 |
9 | @seattlechildren | Hospital | Children’s hospital | US | September 4, 2008 | 6,675 | 55,531 | 2.56 |
10 | @NIHClinicalCntr | Hospital | General hospital | US | February 6, 2009 | 3,111 | 52,221 | 1.27 |
Median |
|
|
|
|
9072 | 213,314.5 | 3.61 |
aUp to October 30, 2015.
Year of Twitter registration by 168 medical institutions, showing the number of Twitter registrations from 2009 to 2011 per half year. The first registration was in July 2, 2009. The latest registration was on November 27, 2011.
Number of Twitter accounts in each medical specialty. The medical specialties of general hospitals were included in the specialty accounts and not in “hospitals”.
Number of tweets from medical institutions in Japan from 2012 to 2014 with content on therapeutic radiology (account 1), cosmetic surgery (account 3), and ophthalmology (account 4).
We classified tweets by medical institutions into 6 categories by content (see
Number of tweets by categories of contents.
Category | Content | No. of tweetsa |
A | Medical knowledge | 21 |
B | Consultation guidance, including advertisements and newsletters from medical institutions | 45 |
C | Patients’ suggestions | 3 |
D | Links only to home pages, blogs, etc | 37 |
E | Individual behavior or thinking | 37 |
F | Tweets with multiple types of content | 20 |
aA total of 5 accounts could not be accessed to obtain tweets on January 11, 2014.
Sample tweets by content category.
Category | Account |
Tweet samplea |
A | 1 | Internal exposure to radioactive cesium through foods reminds us of “Minamata disease” caused by concentrated organic mercury in the food chain from Chisso Minamata factory. Those who ate fish polluted by organic mercury in high densities accumulated “liposoluble” organic mercury in their brain and frequently experienced neuropathy. |
A |
|
On the other hand, radioactive cesium seldom causes biological concentrations like organic mercury. Cesium is an alkali metal close to potassium. Taken into the body, cesium distributes evenly to all cells and is excreted in urine. This was established by analysis of euthanized domestic animals in Fukushima prefecture. |
A |
|
Cesium is excreted in urine like potassium. The amount of cesium in the body is halved at nine days in infants and at three months in adults. Unlike organic mercury, cesium is not accumulated in the body. Even when cesium is excreted, it exists on Earth because the half-life of cesium-137 is approximately 30 years. |
A | 2 | Defining blood–brain barrier (BBB) disruption as a hyperdense lesion on the CT scan after endovascular therapy in acute ischemic stroke patients, BBB disruption is independently associated with lower rates of early major neurologic improvement, higher rates of mortality, and hemorrhagic complications. http://**** |
A | 3 | We call it a wedge-shaped defect when the sides of the teeth are sharpened. Clenching one’s teeth is the most common cause of wedge-shaped defects. Images→http://**** |
B | 4 | ****clinic for liposuction and fat infusion! ****clinic is an expert in liposuction with more than 30 thousand cases. We have been chosen for techniques and trust. |
B | 5 | We will start flu shots on October 15. We are accepting reservations from companies for flu shots. Call ** in charge. |
C | 6 | [In patient's voice]: “I found this dentist. This dentist described the treatment I wanted and how to brush well. This dentist explained other things besides treatment, too. If I had brushed my teeth earlier, I would have better teeth. I regret that now.” |
D | 7 | Our blog was renewed. [Report of cardiology training] at http://**** |
E | 8 | Let’s drink water because it’s hot. |
E | 9 | I feel that it’s the end of the year due to increasing numbers of patients with hand injuries. |
I feel that it’s spring because of the number of infectious atheroma cases. | ||
F | 10 | Anticancer activity of extract from grape seeds: Many studies suggest that an extract from grape seeds has anticancer properties. http://**** |
Watching and doing dance performances are fun, even if we cannot do them. Everyone says, “If I was XX years younger, I would dance.” Let’s go dancing! Throw away your assumptions and prejudices. |
aJapanese conversations were translated into English. The number of characters could be more than 140 because the tweets were translated from Japanese sentences.
Number of accounts by medical specialty tweeting content categorized as follows: medical knowledge; consultation guidance including advertisements and newsletters from medical institutions; suggestions from patients; links only to homepages, blogs, and others; individual behavior or thinking; and tweets including multiple types of content.
The frequency of nouns per medical specialty is shown in
In this study, we classified Twitter accounts by medical practice, including dentistry and oral surgery, cosmetic surgery, plastic surgery, dermatology, other surgery, and internal medicine including pediatrics, and by tweet contents. Powerful accounts with ≥500 followers varied by clinical specialties, as did Twitter use indicators such as amount of active Twitter use. There were 179,857 medical institutions in Japan [
Many of the top 10 Japanese accounts were medical institutions of cosmetic surgery and dental clinics, which frequently performed medical treatment not covered by health insurance. This suggested that the accounts of cosmetic surgery clinics possibly used Twitter for advertising. Few of the accounts were from general hospitals, while many were from clinics. In comparison, many of the top 10 accounts from English-speaking countries were from general hospitals. The British Medical Association has published guidelines about social media use among doctors and medical students in the United Kingdom [
Tweet contents were also markedly different between the specialties. For example, cosmetic surgery accounts, where most practices are not covered by insurance, tended to tweet consultation guidance, including advertisements. We also found advertisement tweets covering introductions to clinic operations, consultation hours, and working hours of the doctor in charge of operations. Of the tweets by cosmetic surgery accounts, 38% (9/24) were categorized as advertisements, defined as tweets with concrete prices or discounts. Most practices in cosmetic surgery are not covered by insurance and it is necessary to attract customers or collect patients. Thus, Twitter might be used as an advertising medium to attract cosmetic surgery patients. This supported our inference that medical institution management affected Twitter use.
On the other hand, the Twitter accounts of some specialties were used for medical knowledge and education. Words about pregnancy frequently appeared in the other surgery category, such as “uterus,” “spermatozoon,” and “infertility” (
Some Twitter accounts used Twitter for both attracting patients and disseminating medical knowledge. Words about dental practice frequently appeared in the dentistry and oral surgery accounts, such as “decayed tooth,” “implant” (
In some specialties, such as internal medicine, surgery, and pediatrics, Twitter was seldom used. Of the 168 accounts, 83 (49.4%) were dentistry and cosmetic surgery accounts. In contrast, only 11 accounts (6.5%) were in internal medicine, pediatrics, and surgery. In those specialties, most medical practices in Japan are covered by medical insurance. Thus, the institutions likely do not need to attract customers or collect patients. In particular, there was only 1 pediatric Twitter account, which was significantly lower than in other specialties. Parents with small children often use Twitter because they have high Internet literacy [
In this study, we first showed how medical institutions use Twitter and what they tweet. Nevertheless, several issues remain for discussion.
First, we conducted this study with power accounts (having ≥500 followers) as the focus. Thus, we might not have obtained the same results if we had included accounts with fewer followers. About 98% of Twitter accounts have fewer than 500 followers in Japan [
Second, we did not examine the medical accuracy of tweets in this study. Several studies have examined deleterious and inaccurate information on websites and social media [
Third, the results of this study depended on the search site; however, it remains unclear what Web search tools are sufficiently accurate and adequate. Some search sites for Twitter accounts exist, such as twpro, Twitter Search, and Twitter Profile Search [
Fourth, we did not examine whether Twitter advertisements conformed to guidelines about medical institution advertising. Advertising by medical institutions is regulated by the Japanese Medical Care Act [
The number of Twitter accounts by medical institutions and the tweet contents differed by medical specialty. Some Twitter accounts attracted customers or collected patients. On the other hand, some accounts provided medical knowledge. Thus, Twitter potentially can be used for various purposes by different medical specialties.
Twenty most frequently tweeted Japanese nouns in each medical specialty.
This work was supported by a research grant (H24-26, 24200701 to Dr Narimatsu) from the Ministry of Health, Labour and Welfare of Japan.
We would like to thank Editage for English-language editing of an earlier version of the manuscript.
None declared.