We established the AHRN on Facebook in 2014 to provide an online repository of summarized AIH literature, discussion of study results and application to patient care, and opportunities to participate in our ongoing AIH research studies. As of May 2017, the AHRN Facebook group has 1640 members and is led and moderated by a hepatologist (CL) from Indiana University (IU). New membership has grown by individuals finding the page via Facebook search results, referrals shared in other AIH Facebook groups by our members, and direct physician-to-clinical-patient marketing. The AHRN is a private Facebook group; individual users wishing to join must be granted approval from the group moderator. Once approved, members are able to view and share content with other group members. A private group setting has the benefit of protecting the privacy of members by hiding their group affiliation from nonmember friends and allows group moderators to screen potential members for illegitimate accounts.

We posted a GRACE study recruitment advertisement, consisting of a recruitment message, study summary sheet, and representative study image, on the AHRN Facebook group after obtaining institutional review board approval. The study summary sheet outlined study aims, procedures, inclusion criteria, and the study coordinator’s contact information (phone and email). Inclusion criteria consisted of individuals 18 years and older, diagnosis of AIH or AIH with features of overlap disease (primary biliary cholangitis [PBC] or primary sclerosing cholangitis [PSC]) established by a medical doctor, and no provision of medical care at IU or surrounding affiliate institutions. The study advertisement was posted on the AHRN Facebook group at quarterly intervals over 24 months by the group moderator. The study advertisement directed interested individuals to contact the study coordinator to discuss study objectives and subsequently give consent by phone.

After verbal consent, a study packet containing epidemiologic surveys [14,15], an Oragene DNA saliva kit (DNA Genotek Inc), study instructions, consent documents, health information release form, and prepaid return envelope were mailed to the participant’s home address. Self-reported demographic parameters and medical history were collected from a single survey [14]. Participants were then assigned a study identification number linking all completed materials. Upon receipt of study materials, the participant’s health information release was faxed to the physician treating the AIH in order to collect supportive data related to diagnostic testing. Requested medical records included procedure notes and reports of liver biopsy and pathology and antinuclear antibody, antismooth muscle antibody, antiliver kidney microsomal antibody, antisoluble liver antigen, antimitochondrial antibody (AMA), both cytoplasmic and perinuclear antineutrophil cytoplasmic, immunoglobulin G, and viral hepatitis serologies. Failure to receive a participant’s study materials or outside medical records within 1 month of distribution resulted in repeat contact to each entity up to 3 times. Outside medical record nonresponse was further targeted with a follow-up phone call to the appropriate clerk on the day of medical release transmission.

Once received, outside medical records were reviewed and tabulated into the local AIH database by study personnel. Missing clinical and laboratory support data for AIH diagnosis from outside records were identified and subsequently targeted with repeat release of information requests. Collected results (including saliva kit, epidemiologic surveys, and outside diagnostic reports) were stored in a secure location.

Saliva was collected from the subjects using an Oragene collection kit and stored according to kit instructions. DNA was isolated from 500 µL of saliva in PrepIT-L2P Purifier Reagent (Lot PT150219-A, DNA Genotek Inc), and remaining collected saliva sample was saved for future study.

We enrolled 29 participants in 2 years between June 2014 and June 2016. Participants were from 20 different states, with 27 (93%) indicating a primary residence outside of Indiana and a median distance of 727 miles from our center. Of 29 participants, 28 (97%) successfully completed and returned study instruments within 3 months of the coordinator-led phone consent. All participants were female, 90% (26/29) were white, and median age at enrollment was 52 years.

Of 28 participants, 20 (71%) subjectively reported a diagnosis of AIH alone, while 29% (8/28) reported AIH with features of PBC (Table 1). Among all included participants, 46% (13/28) reported concurrent, extrahepatic autoimmune disease (data not shown). Other frequent medical conditions and symptoms reported included history of contraceptive use (26/28, 93%), history of multiple urinary tract infections (19/28, 68%), frequent fatigue (19/28, 68%), occasional to constant itching (13/28, 46%), and right upper abdominal discomfort (17/28, 61%).

All participants recorded their current medication list within an epidemiologic survey provided at enrollment (Table 2). Discrepancies were found between the participant surveys and medical records for immunosuppressant medication usage. Additionally, 3 participants reported no active immunosuppressant use, which was contradicted by their outside medical records. One participant had no documented evidence of immunosuppression in the physician notes and also had an incomplete simplified score, high AMA titer, and a biopsy consistent with PBC alone.

Appropriate outside medical records were received after the first request along with the medical information release form for 79% (23/29) of participants (Table 1). A total of 50% (13/26) of available liver biopsies were typical for AIH, 1 (4%) indicated PBC only, and 12 (46%) were identified as compatible with AIH according to the simplified criteria histologic findings [16]. Of the participants with definite AIH per the simplified criteria, 25% (1/4) had a clinically reported a positive AMA titer as well as findings on biopsy of AIH with features of PBC.

Three-quarters of the participants (22/29, 76%) had available autoimmune markers for the simplified score assessment. However, there were a number of missing values for the simplified criteria given the nature of data collection. Therefore, in order to enroll AIH patients into the GRACE study, we sought to categorize the participants most likely to have AIH based on meeting at least 2 of 3 lab- or clinician-reported criteria: positive autoimmune markers (simplified criteria), liver biopsy with typical findings of AIH, and/or reported treatment with immunosuppressant medications. Of 29 participants, 24 (83%) met at least 2 of 3 criteria, with 19 of these participants (79%) satisfying all 3.

Saliva samples were obtained from all 29 study participants. All collected samples were at least 2 mL in volume. DNA isolation from representative saliva samples yielded sufficient quantity and quality of DNA for genetic investigations. Median value of DNA yield was 41 µg per 500 µL saliva sample (interquartile range [IQR] 41.86) and 260/280 ratio was 1.84 (IQR 0.125).

Sustained AIH research advances are continually impeded by low participant enrollment numbers, limited and disconnected investigators, and poor engagement of geographically spread patients wanting to contribute to scientific progress [9]. However, social media presents a unique platform to address these significant investigational barriers. Our research engagement with the AHRN Facebook group supports patient-oriented research and facilitates participation in research studies even from a distance. Furthermore, we assert that our research team correctly phenotyped and recruited 24 AIH patients and 4 AIH patients with overlap features of PBC to the GRACE study from outside our institutional reach.

The retrieval of supporting outside records and eventual clinical diagnosis of AIH can be challenging for participants recruited both locally and remotely. Remote access was impeded most frequently by delayed response from corresponding medical record departments. Despite the obstacles inherently imposed by the study design, we were able to successfully collect epidemiologic questionnaires and biologic samples from all 29 participants using Facebook. While the effort spent on retrieval of outside records was time intensive, the passive approach to recruitment required less time spent identifying and recruiting patients enrolled locally through traditional methods. The successful recruitment of 29 subjects was substantial for the GRACE study, as our traditional approach at a large academic center provided only 120 participants (24% more cases obtained with social media) in the same duration of study.

SNS have evolved rapidly over the past decade to support our need for social connectedness and instant digital knowledge. Application of these tools in rare disease research is important, as they transcend well-established limitations such as cost, prolonged study courses, and geographic barriers. Our study was cost- and time-efficient and effective at recruiting patients up to 2200 miles away from our medical center. Furthermore, the cohort was similar to prior work according to gender (female predominant), ethnicity, concurrent autoimmune illnesses, fatigue, and itch [2,17]. Participants in similar studies at our institution are often reimbursed US $25 for expenses related to travel, parking, or time burden. Those expenses were not incurred by the GRACE study participants and thus reduced overall cost of study conduct by approximately US $725. Furthermore, our approach to research also reduces time burden for study participants. Our Web-based approach allows participants to complete the study procedures at their convenience rather than on the schedule of the study team and eliminates the additional time spent traveling to and from the study center.

In our study, prospective participants quickly engaged our study coordinator for recruitment. However, self-attestation of AIH diagnosis and current treatment strategies were not always congruent with their physician-composed medical records. Furthermore, we found certain medical records had insufficient documentation or lacked specific workup components. Our assessment of at least 2 of 3 key criteria, as outlined above, suggests AIH diagnosis in 24 out of 29 study participants (83%). Other studies have shown similar levels of agreement between self-reported diagnosis data and medical record documentation using traditional research methods, suggesting this challenge is not unique to social media–based research [18,19].

Successful implementation of a social media–based methodology comes with its own unique set of challenges. While SNS can supplement traditional recruitment, the process of sending and receiving medical records and study materials from affiliated institutions and participants proved to be time-intensive for the study team. Nonetheless, the time cost associated with this methodology is considerably less compared to the effort required to initiate recruitment at multiple study sites to achieve a similar geographical reach. Additionally, the demographics of AIH patients support a social media approach as they align well with the SNS user demographics [20]. This group may be better able to provide accurate self-reported data, as it has been found that female gender and higher educational attainment, characteristics of SNS users, are associated with increased uniformity between patient reports and medical records [19]. The demographic similarity between AIH patients and SNS users may not exist for some rare diseases and must be considered when evaluating the applicability of social media methods to specific patient populations.

Finally, a notable challenge of implementing SNS in medical research is simply its acceptability by providers and health systems. The medical community has been hesitant to embrace the potential of this approach to research recruitment, in part due to the risk of Health Insurance Portability and Accountability Act violations and the potential for demographic biases [9,21]. Furthermore, low acceptance may stem from slow adoption of new technologies, particularly in fast-paced clinical environments, and research regarding the use of SNS in medical research is still limited [22]. The rise of mobile health technologies using mobile phones and wearable devices will create new opportunities for conducting remote clinical studies drawing participants from SNS. We believe as these technologies become more prevalent in clinical and research settings, acceptance of new research methodologies such as SNS will also increase. Failure to integrate this model represents a missed opportunity, as many of the over 2 billion daily Facebook users have actively sought medical information or participated in health-related digital support groups [23,24].

We foresee opportunities for social media to help connect like-minded investigators for research collaborations [25]. As we continue to supplement site-based recruitment for GRACE with Facebook-based participant enrollment, we anticipate additional opportunities for social media to be used to complement traditional studies through study advertisements, recruitment opportunities, collaborations with patient groups, or study-specific groups or pages designed for easy communication with study participants. Using SNS for recruiting participants for research regarding rare diseases such as AIH serves as a viable bridge to connect patients to investigators despite geographical barriers and a lack of local opportunities. We are confident our method could become a frontrunner in the improvement of low study numbers in rare disease research. In overcoming recruitment limitations, we now seek a way to efficiently expedite the time between recruiting individuals, obtaining their biologic samples and epidemiological data, and confirming their diagnosis.

In summary, SNS can be an effective tool for facilitating patient-oriented research in AIH. We have successfully implemented a Facebook-based methodology to engage individuals unable to participate in research through traditional methods due to geographic or financial barriers. Mindful use of SNS can transcend established limitations in rare disease research and further cultivate powerful and engaged research communities.