Chatbot 1
Chatbot 1
a r t i c l e i n f o a b s t r a c t
Article history: Objective: The AAP, AFP, and ACP have authored statements and recommendations to clinicians about the impor-
Received 16 December 2018 tance of the transition from pediatric to adult care. The Got Transition program provides a framework and re-
Revised 4 September 2019 sources based on AAP, AFP and ACP recommendations to promote skill attainment in self-care. Engaging
Accepted 4 September 2019 adolescents along the transition journey has proven challenging. Use of smartphones, text messaging, and social
Available online xxxx
media are prevalent among teenagers, offering a unique opportunity to engage teenagers in their preferred chan-
nel to provide tools and resources to help them successfully transition to adult focused care.
Methods: A multidisciplinary team of clinicians, quality improvement facilitators, and human-centered designers
at the University of Vermont (UVM) Children's Hospital designed tools for teens with chronic conditions that
support the Got Transition recommendations. Using a co-creative design process, we created a novel tool to in-
crease engagement among teenagers. We conducted a pilot study of 13 teenagers with a chronic medical condi-
tion using a text messaging platform (chatbot) with scripted interactions to increase engagement and deliver
educational content according to Got Transition.
Results: Mean engagement was 97% during the study period. Qualitative feedback from study participants sug-
gests our chatbot should be extended and shows promise to help teenagers attain self-care skills on the transition
journey.
Conclusions: A scripted text messaging platform is feasible and appears to be well-received by patients and care-
givers. Furthermore, our approach emphasizes the need to engage teenagers through multiple platforms to effec-
tively serve as “coaches” during the transition to adult care.
© 2018 Elsevier Inc. All rights reserved.
https://doi.org/10.1016/j.pedn.2019.09.004
0882-5963/© 2018 Elsevier Inc. All rights reserved.
86 J. Beaudry et al. / Journal of Pediatric Nursing 49 (2019) 85–91
Fig. 1. A patient creating their own visual map to a successful transition of care journey.
adolescent transition to adult care has been shown to lead to self-care Advisory Council, and the VCHIP Youth Health Advisory Council. Our de-
skill attainment (Sebastian, 2014). sign process created a clear structure for meaningful participation with
The use of smartphones, text messaging, and social media are nearly the work that was action-oriented, energizing, and productive.
ubiquitous among teenagers, with the Pew Research Center reporting in With greater frequency, health care organizations are building inter-
2018 that access to a smartphone among teens was at 95% (Pew nal human-centered design teams to improve existing services and sys-
Research Center, 2018). The Associate Press-NORC Center for Public Af- tems as well as create new models for health care delivery (Molloy,
fairs Research, a partnership between the Associated Press and NORC at 2018; Kachirskaia et al., 2018). Human-centered design (HCD) is a col-
the University of Chicago, presented a report in 2017 that found 91% of laborative and creative process focused on understanding the implicit
teens using native text messaging to communicate (The Associated and explicit needs of end users (Brown, 2008). HCD engages patients,
Press-NORC, 2017). Given the saturation of this technology, it is not sur- families, providers, and staff to actively participate in the co-creation
prising that health care providers have attempted to use these channels of healthcare services. Our project united clinicians trained in quality
to engage teens in promoting healthy behaviors, from smoking cessa- improvement and the Institute for Healthcare Improvement (IHI)
tion to obesity reduction (Hingle, 2013; Mason, Campbell, Way, & Model for Improvement and designers with expertise in the human-
Keyser-Marcus, 2015). These trends present a unique opportunity to centered design process championed by hiCOlab. Both methodologies
engage teenagers in their preferred communication channel to provide (quality improvement and human-centered design) involve developing
tools and resources to help them successfully transition to adult focused clear and measurable aims using an iterative process for brainstorming
care. different ideas, designing and prototyping the strongest ones, and test-
The primary objective of this pilot study was to test the feasibility of ing them with patients and clinicians. Our team facilitated a participa-
a novel texting platform aimed at increasing engagement –measured by tory research and co-design process to explore how we might better
the response rate to text messages- among teenagers while teaching es- help adolescents with chronic conditions transition successfully to inde-
sential self-care skills. We also sought to develop unique partnerships pendence in managing their own health care. This highly collaborative
with quality improvement and human-centered design specialists to approach brought several key partners closer together, catalyzing the
create educational and coaching tools to prepare adolescents and fami- creation of innovative responses to address the problem of transition
lies for the transition of care process in selected pediatric specialty readiness in our patient population.
clinics. Both objectives were inspired by the joint consensus statements The project team conducted semi-structured interviews with and
and the Got Transition framework to address Transition Readiness, one observations of over 60 stakeholders including patients, caregivers, pro-
of Got Transition's Six Core Elements. viders, and other clinical staff. (Fig. 1) We reviewed relevant scholarly
research and case studies to establish a broader context for pediatric
Methods healthcare transitions. These research activities yielded nuanced detail
and richness to aid our understanding of the complex issues adolescent
Design approach patients face as they transition, while engaging our stakeholders as par-
ticipants in the design process.
Vermont Child Health Improvement Program (VCHIP) is the oldest Working with the principles and methods of co-design, we facili-
running national improvement partnership (Shaw, Norlin, Gillespie, tated several co-creative workshops with patients. In small peer groups,
Weissman, & McGrath, 2013). VCHIP established the organizational in- patients created large, wall-sized timelines of their individual transition
frastructure that brought together several stakeholders including the of care journeys, sharing their experiences, past frustrations and suc-
Vermont Department of Health (VDH), the University of Vermont cesses, as well as hopes and fears about the future. Building off the
(UVM) Medical Center, UVM Children's Hospital, and hiCOlab
(Healthcare Innovation Collaboratory), a design and innovation lab em- Table 1
bedded in the UVM Health Network Medical Group. We built important Age breakdown of participants at time of enrollment.
relationships between these different organizations that included 14 years 15 years 16 years 17 years
leaders and partners from pediatric and adult specialty clinics and pri-
2 1 9 1
mary care practices, hospital operations, the Patient and Family
J. Beaudry et al. / Journal of Pediatric Nursing 49 (2019) 85–91 87
Intervention
Table 2
Schedule of message topics.
of care timelines with each other, learned about some of the most im- to additional online resources. An example of one of these interactions
portant skills needed for a successful transition, and created PSA-style is given in Fig. 3.
posters advocating for these skills. The workshop was also used to intro-
duce teens to the text messaging chatbot and enroll them in the Results
program.
The chatbot (Fig. 2) was designed to send weekly text messages on Data collected included analytics from the text messaging software,
Wednesday evenings to adolescent patients to begin to prepare them and surveys designed to measure user behavior, including knowledge
for the transition of care journey. A platform using basic text messaging gained, satisfaction of the materials and workshop, and usability of the
was selected because it did not require an app to download – it is porta- technology. Surveys were deployed to adolescents at the following in-
ble across virtually every mobile phone – and it is one of the most dom- tervals: at the beginning of the study, before and after the introductory
inant communication channels among teens. The messages invite workshop, and at the close of the study. Thirteen adolescents completed
interaction by asking patients about different aspects of their health the initial survey and twelve completed the final survey; all 13 partici-
care each week (Table 2), including topics like understanding one's pants were counted when measuring engagement throughout the dura-
chronic condition and medical history, ordering prescription medica- tion of the text messaging program.
tion refills, contacting a doctor's office, and preparing for doctor's ap- Quality improvement methodology using statistical process control
pointments. Based on how the user responds to these questions, the charts evaluated the engagement rate. Over the course of the six-
chatbot then replies with appropriate encouragement, tips, and links month period of the chatbot, the overall weekly response rate to the
messages (rate of engagement) showed a mean of 97% on a statistical
process control chart (Fig. 4). This corresponded with what patients re-
ported in the final survey: they overwhelmingly agreed that text mes-
saging was the right channel for them, and they indicated that the
rate of one message per week was preferred. They reported being moti-
vated to respond to the texts because of its “ease of use” and “friendly”
persona. Participants stated that they would more likely than not rec-
ommend the text messaging program to their peers.
Analysis of the pre and post project surveys revealed an increased
patient participation in many important activities needed for successful
health care transitions, like managing medications, ordering prescrip-
tion refills, and proactively contacting providers about health-related
concerns (Figs. 5-7). However, study participants did not feel more con-
fident with the “softer” skill of self-advocacy. Participants reported that
they do not understand health insurance and would benefit from more
education on that topic. Study participants reported that they appreci-
ate the peer-to-peer interactions of the transition of care workshop
and they indicated a strong interest in more opportunities to connect
with young adults who can provide mentorship along the transition of
care journey. Two thirds of the participants reported that the workshop
was helpful in teaching them transition related topics.
Total number of participants: n = 12.
Total number of participants: n = 12.
Total number of participants: n = 12.
Discussion
Fig. 4. A statistical process control chart (P chart) of weekly teen engagement with the texting platform. The percent of responses to the text messages is on the Y-axis.
retained information from the modules and were able to affect change. staff, patients and families. Consistent with existing literature, we
We recognize that the timing of our post pilot survey deployment may would like to engage and develop roles for nurse clinicians, social
not have allowed the amount of time necessary to practice these new work case managers and physicians. Young adults indicated on their
skills. Additionally, skill attainment should be assessed longitudinally post pilot surveys that they value their health care clinician's role in pro-
which would require a system and data collection tool. (Baker, viding education and counseling on health care topics and view them as
Riekert, Sawicki, & Eakin, 2015; Wood, 2014) Our interpretation of the a key partner. Moving forward, we envision nurse clinicians taking a
pilot, even with its limitations, is that the digital formats are appealing more active role in facilitating similar workshops.
to adolescents and young adults and provide another vehicle to coach Finally, our project combined two different but similar methodolo-
skills and convey important information in the transition journey. This gies (quality improvement and human-centered design) to develop
format may be considered when programs address Transition Readiness and test a novel intervention. These methodologies generated a deeper
as part of the Got Transitions Six Core Elements framework. understanding of how clinicians approach the process of transitions, the
We speculate that the workshop was an equally important tactic for patient and family experience of preparing for the transition, engage-
creating high engagement throughout the duration of the pilot because ment with key stakeholders, change management, and rapid
it created an energizing social experience among a group of peers, and it prototyping/rapid cycle improvement. In our experience, the combined
provided a structured onboarding experience. Unfortunately, we are not methodology resulted in an innovation aimed at solving a notoriously
able to continue the workshop due to resources required from clinic challenging problem while delivering a richer experience for the project
Fig. 5. Pre and post results of medication tracking after our text messaging pilot. Gray icons represent patients not performing the task. Orange icons represent patients completing the task.
90 J. Beaudry et al. / Journal of Pediatric Nursing 49 (2019) 85–91
Fig. 6. Pre and post results of medication refills after our text messaging pilot. Gray icons represent patients not performing the task. Orange icons represent patients completing the task.
team. We plan to sustain our partnership to encourage innovation and a participation in a transition program, we do not think that relying solely
deeper understanding of complex systems that impact our patients and on digital formats is sufficient to impart knowledge and teach skills nec-
families. essary for a successful transition to adult care. Future work must inte-
An important limitation of our study is the use of a digital format re- grate clinical workflows, clinical champions and peer interactions as
lying on text messaging. Our project required study participants to have outlines in the joint AAP, ACP and AAFP statements. As stated in the
a cellular phone with reliable and affordable service, which introduces 2018 consensus statement, encouraging innovation and improvement
selection bias. Furthermore, our study period was relatively short and methodology are essential to advancing the challenging field of adoles-
might not be long enough to detect changes in cellular phone numbers cent transition (AAP, AAFP, & ACP, 2018).
or lapses in cellular service that would be a barrier to this tool. Future
studies should investigate how differences in geography, race, gender, CRediT authorship contribution statement
socio-economic status impact engagement and effectiveness of the
program. Jeremy Beaudry: Conceptualization, Software, Investigation,
The transition from pediatric to adult care is fraught with numer- Resources, Writing - original draft. Alyssa Consigli: Conceptualization,
ous challenges to adolescents, parents, caregivers and clinicians. Ide- Investigation, Resources, Writing - review & editing, Project administra-
ally, the transition process would evolve over time with the focus on tion. Colleen Clark: Software, Investigation. Conceptualization. Keith J.
developing self-care skills for the adolescent (Tuchman, Schwartz, Robinson: Conceptualization, Methodology, Formal analysis, Investiga-
Sawicki, & Britto, 2010). In order for this to occur, the adolescent re- tion, Supervision, Writing - review & editing.
quires a strong support network with effective engagement to
“coach” the development of self-care skills. As digital media use con- Acknowledgments
tinues to increase among teenagers, we believe healthcare teams
should leverage technology such as text messaging and related for- Vermont Child Health Improvement Program (VCHIP) provided
mats to increase engagement. oversight of this research, and this work is funded in part with monies
provided by or through the State of Vermont. The State does not neces-
Conclusion sarily endorse the researchers’ findings and/or conclusions. The findings
and/or conclusions may be inconsistent with the State’s policies, pro-
Our pilot study tested the feasibility of a text messaging platform to grams, and objectives. University of Vermont Medical Center Marketing
increase engagement. While we showed promise to increase and Communication.
Fig. 7. Pre and post results of contacting your doctor's office after our text messaging pilot. Gray icons represent patients not performing the task. Orange icons represent patients
completing the task.
J. Beaudry et al. / Journal of Pediatric Nursing 49 (2019) 85–91 91
References Lemke, M., Kappel, R., McCarter, R., D'Angelo, L., & Tuchman, L. K. (2018). Perceptions of
health care transition care coordination in patients with chronic illness. Pediatrics,
American Academy of Pediatrics, American Academy of Family Physicians, American Col- 141(5), e20173168. https://doi.org/10.1542/peds.2017-3168.
lege of Physicians, Transitions Clinical Report Authoring Group, Cooley, W. C., & Mason, M. J., Campbell, L., Way, T., & Keyser-Marcus, L. (2015). Development and out-
Sagerman, P. J. (2011). Supporting the health care transition from adolescence to comes of a text messaging tobacco cessation intervention with urban adolescents.
adulthood in the medical home. Pediatrics, 128(1), 182–200. https://doi.org/10. Substance Abuse, 36(4), 500.
1542/peds.2011-0969. McManus, M. A., Pollack, L. R., Cooley, W. C., McAllister, J. W., Lotstein, D., Strickland, B., &
American Academy of Pediatrics, American Academy of Family Physicians, American Col- Mann, M. Y. (2013). Current status of transition preparation among youth with spe-
lege of Physicians, Transitions Clinical Report Authoring Group, White PH, Cooley cial needs in the United States. PEDIATRICS, 131(6), 1090–1097. https://doi.org/10.
WC. Supporting the health care transition from adolescence to adulthood in the med- 1542/peds.2012-3050.
ical home. Pediatrics . 2018; 142(5):e20182587. (2019). Pediatrics, 143(2), Molloy, S. (2018, April 20). Innovation labs in healthcare - a review of design labs as a
e20183610.doi:https://doi.org/10.1542/peds.2018-3610. model for healthcare innovation [MRP]. Retrieved August 11, 2019, from http://
Baker, A. M., Riekert, K. A., Sawicki, G. S., & Eakin, M. N. (2015). CF RISE: Implementing a openresearch.ocadu.ca/id/eprint/2364/
clinic-based transition program. Pediatric Allergy, Immunology, and Pulmonology, 28 Sebastian, R. A. (2014). Measuring youth health engagement: Development of the youth
(4), 25–254. https://doi.org/10.1089/ped.2015.0594. engagement with health services survey. Journal of Adolescent Health, 55(3), 334–340.
Brown, T. (2008). Design thinking. Harvard Business Review, 86(6), 84 Retrieved from https://doi.org/10.1016/j.jadohealth.2014.02.008.
https://hbr.org/2008/06/design-thinking. Shaw, J. S., Norlin, C., Gillespie, R. J., Weissman, M., & McGrath, J. (2013). The National Im-
Hingle, M. (2013). Texting for health: The use of participatory methods to develop provement Partnership Network: State-based partnerships that improve primary
healthy lifestyle messages for teens. Journal of Nutrition Education and Behavior, 45 care quality. Academic Pediatrics, 13(6), S84–S94. https://doi.org/10.1016/j.acap.
(1), 12–19. https://doi.org/10.1016/j.jneb.2012.05.001. 2013.04.001.
Kachirskaia, I., Mate, K. S., & Neuwirth, E. (2018, June 28). Human-centered design and per- Teens, Social Media & Technology 2018 | Pew Research Center. Retrieved August 9, 2019,
formance improvement: Better together. NEJM Catalyst Retrieved August 11, 2019, from https://www.pewinternet.org/2018/05/31/teens-social-media-technology-
from https://catalyst.nejm.org/hcd-human-centered-design-performance- 2018/(2018, May 31)..
improvement/. Tuchman, L. K., Schwartz, L. A., Sawicki, G. S., & Britto, M. T. (2010). Cystic fibrosis and
Instagram and Snapchat are Most Popular Social Networks for Teens; Black Teens are transition to adult medical care. PEDIATRICS, 125(3), 566–573. https://doi.org/10.
Most Active on Social Media, Messaging Apps | APNORC.org | APNORC.org (Ed.). 1542/peds.2009-2791.
(2017). Retrieved August 9, 2019, from apnorc website. http://apnorc.org/PDFs/Teen% Wood, D. L. (2014). The transition readiness assessment questionnaire (TRAQ): Its factor
20Social%20Media%20Messaging/APNORC_Teens_SocialMedia_Messaging_2017_ structure, reliability, and validity. Academic Pediatrics, 14(4), 415–422. https://doi.
FINAL.pdf. org/10.1016/j.acap.2014.03.008.
Lebrun-Harris, L. A., McManus, M. A., Ilango, S. M., et al. (2011). Transition planning
among US youth with and without special health care needs. Pediatrics(4), 142
(doi:e20180194).