Article Text
PDF
Statistics from Altmetric.com
Summary box
What are the new findings?
Healthcare hackathons are celebrated for bringing cross-disciplinary teams together to do innovative work, but the typical multiday schedule is a barrier to entry for many clinician trainees.
An abbreviated version of the traditional hackathon, which we term a ‘microhack’, may be an effective way to include clinician trainees in innovation training.
We hosted two microhacks over 2 years and at least half of our 98 participants were clinicians, demonstrating interest in these types of collaborative, problem-solving events.
How might it impact on healthcare in the future?
We share our experiences and lessons learnt in an event toolkit, in the hope that our work can be replicated and improved on at other institutions.
Introduction
Background
As healthcare becomes increasingly complex, team oriented and practised beyond the traditional bounds of hospitals and clinics, doctors need explicit training in and exposure to innovation.1 We describe training in innovation as the skills needed to turn insights gained from the practice of medicine into creative action in order to move the field of healthcare forward. These skills are broad and diverse, from clinical informatics2 to design thinking,3 making them difficult to teach in a typical classroom setting. Healthcare hackathons provide an opportunity to introduce these skills in a problem-based, cross-disciplinary format.4 Modelled after events from the technology sector, healthcare hackathons bring individuals from different backgrounds together to tackle healthcare-specific challenges over a short amount of time. Similarly, a healthcare datathon brings data scientists, statisticians, engineers and clinicians together to analyse a dataset and investigate data-driven solutions to common clinical problems.5 These types of events have been celebrated for their ability to create diverse teams and involve learners in the earlier stages of problem solving and innovation.6 Studies have shown that hackathons also enhance the knowledge and skills associated with being an effective team player and leader.7 Despite the tremendous potential these events hold in medical education, they typically span multiple days, making it nearly impossible for clinician trainees with demanding schedules to participate. Shorter versions of healthcare hackathons have previously been described8 as a means to expand access to innovation, but we believe ours are the first to focus on clinician trainee recruitment and involvement.
Objectives
We condensed the traditional multiday healthcare hackathon into shorter ‘microhack’ events with the goals of determining:
Whether we could recruit clinician trainees to participate.
Whether clinician trainees found the events valuable.
Methods
At our large academic institution, our resident-run interest group, the Beth Israel Deaconess Innovation Group, heard from clinician trainees and colleagues at the Massachusetts Institute of Technology (MIT) Sloan School of Management of interest in collaborating to address challenges in healthcare. We felt a healthcare hackathon, with its emphasis on cross-disciplinary teamwork and rapid innovation, was the ideal setting to accelerate partnerships, but clinician trainees voiced concerns about their ability to engage in a multiday event. We; therefore, restructured the traditional hackathon event into a 1-day ‘microhack’ that would preserve the values of collaboration and creative problem solving but take place in a compressed time frame on a weeknight or weekend day when trainees were more likely to have time off from clinical responsibilities. Our first microhack, hosted in April 2019, partnered Masters of Business Administration (MBA) students from MIT with clinician trainees from Beth Israel Deaconess Medical Center (BIDMC) to design solutions to healthcare challenges proposed by the clinicians (table 1).
Event details for the BIG microhack and the BIG microdatathon
We invited resident physicians to submit challenges they encountered in their everyday work and preselected six questions for teams to tackle, focusing on challenges that were ideal for multidisciplinary problem-solving and that could inspire diverse ideas. We recruited medical students, resident physicians and fellows (a group termed ‘clinician trainees’ from here on) from all specialty training programmes at BIDMC as well as MBA students from MIT via email. At the event, participants were given colour-coded name badges indicating their role and were instructed to form teams of approximately 50% clinicians. Each team had a facilitator who had undergone a 1-hour training on IDEO’s design thinking methodology,9 using ‘How might we’ statements and rapid brainstorming with coloured Post-It notes. The facilitators led each team in multiple rounds of ideation and discussion, ultimately guiding them to select and develop one idea. At the end of the event, each team presented its idea and organisers invited feedback via an anonymous paper survey.
From this event, we saw great enthusiasm and potential, and desire to create more tangible outputs. We partnered with MIT Critical Data for our next event and built on their experience hosting datathons using a clinical database called the Medical Information Mart for Intensive Care (MIMIC)5 to create an event that fit our microhack model. We again prioritised recruiting clinician trainees and paired them with data scientists from MIT. Organisers invited participants via email and created two questions for teams to investigate, focusing on questions that would impact clinical practice and could be answered using MIMIC data without requiring advanced machine learning techniques. We termed this shortened datathon a ‘microdatathon’ and hosted it in February 2020. Each team was composed of approximately 50% clinicians and included a data scientist mentor familiar with using MIMIC. Participants were provided a colour-coded name badge on arrival where colour reflected their role as a clinician, data scientist, or other. They were instructed to sit at a colour-matched chair; each table had a predetermined number of colour-coded chairs in order to ensure diversity of teams. At the conclusion of the event, each team presented preliminary data. Organisers displayed a QR code that linked to an anonymous online survey and invited participants to provide feedback on the event. We invite other institutions to iterate and improve on our microhack events by using our attached toolkit (online supplemental appendix A).
Supplemental material
Data analysis
On registration, participants provided demographic information including organisational affiliation and role via a Google Form (Alphabet, Mountain View, California, USA). We recorded attendance on arrival and extracted deidentified registration data of those who attended into a Microsoft Excel for Office 365 database (Microsoft, Redmond, Washington). After each event, we asked participants to share anonymous feedback on what they enjoyed about the event and what could be improved. These data were stored in a Microsoft Excel for Office 365 document. We used descriptive statistics to report the number and types of participants at each event. Given the small number of post-event feedback responses, we did not code the comments, and rather reviewed them to synthesise general themes.
Results
Participants
Seventy-five participants registered for the microhack and 38 attended the event. There were 21 clinicians, all of whom were trainees. Ninety-nine participants registered for the microdatathon. We accepted 50 participants due to space limitations and 60 participants attended the actual event. Twenty-four were clinicians (38% trainees) and 27 were data scientists, most of whom were students from local universities (table 2). Clinicians across both events represented over eight different specialties.
Participant breakdown for the BIG microhack and the BIG microdatathon
Survey data
Nine of 38 participants in the microhack and 34/60 participants in the microdatathon completed the postevent survey. Responses were overall positive with many participants reporting that they enjoyed meeting people from diverse backgrounds and having the opportunity to experience how data science is used in the real world. Brief tutorials covering design thinking and introductions to data science and dataset usage were well received. Suggestions for improvement included making the event longer in order to accomplish more work and building in more structure. There were many requests for more time to network. Some participants felt that prework would be helpful for them to better understand concepts covered during the event and improve day-of efficiency.
Discussion
Through the inaugural microhack, and its second iteration, the microdatathon, we launched a novel event model to engage clinician trainees in innovative cross-disciplinary teamwork to address today’s healthcare challenges. With the identified limitation being time, we distilled the most generative elements from hackathons and datathons into a micro model. Among the 98 participants at our two events, at least half were clinicians and one-third were clinician trainees, and many found their experience meaningful.
Limitations of our analysis include lack of data collected on participant expectations, suboptimal response rate from the participants surveyed, and lack of a validated measure of interdisciplinary collaboration and other skills to test for development. We had an improved survey response rate at our second event, perhaps because of ease of access to an electronic survey, and plan to administer electronic surveys moving forward. We also intend to measure the educational value of future events by using a validated survey such as the Affective Learning Scale,10 which has been used in prior datathons7 to measure participant attitudes and potential behavioural changes.
We will continue to refine our event model and expand the diversity of our participants in the future. Long format hackathons and datathons have previously shown to be productive events, sparking partnerships, research collaborations and even intellectual property.11 While we plan to prioritise the single-day model for in-person events, a virtual format may help us address some of the feedback we received from participants, and will likely be a necessity in the future due to social distancing requirements.
A virtual format would allow teams to work asynchronously, and preparatory work such as an introduction to programming could empower clinician trainees to take a more hands-on role in data analysis and improve team effectiveness. Research has shown that many residency programmes have used asynchronous learning for years now to help address challenging resident schedules,12 with promising feedback in interdisciplinary settings as well.13 A virtual format may also facilitate recruitment of content experts, who could serve as mentors to refine problem solving and expand networking opportunities. This advantage was seen in a recent week-long virtual datathon hosted by MIT, where the virtual nature and flexible time commitment allowed a large group of content experts to mentor participants throughout the event.14 Other areas of future investigation include how to move projects beyond the ideation phase and how to facilitate the acquisition of resources and faculty to implement proposed solutions.
Conclusion
Healthcare hackathons provide an opportunity to teach important problem-solving skills in multidisciplinary teams, but the typical multiday model of these events make it difficult for clinician trainees to participate. We created a novel event format termed the ‘microhack’ that shortens the traditional hackathon while preserving its innovative features. Our microhacks successfully included clinician trainees and may be a valuable tool for mentoring them in healthcare innovation and cross-disciplinary collaboration. Further research is needed to determine how these events can be formatted to maximise educational value.
Ethics statements
Ethics approval
Participants consented to the collection and deidentified reporting of their professional affiliation and role, and aggregated event feedback. The Beth Israel Deaconess Medical Centre Committee on Clinical Investigations determined our protocol involving human subjects to meet criteria for exempt status (IRB Protocol #: 2020P000825, 8/18/20). Our IRB approved the protocol with a waiver of informed consent.
Acknowledgments
We would like to thank the Beth Israel Deaconess Medical Centre Internal Medicine residency programme, who supported our efforts to introduce this new educational model for trainees. We also give appreciation to our partners including the MIT Sloan School of Business, MIT Biotechnology Group, MIT Critical Data, Google Cloud, the Beth Israel Deaconess Graduate Medical Education Office, the Beth Israel Deaconess Innovation Group and the Beth Israel Deaconess Surgical Programme in Innovation. In particular, we would like to acknowledge C. Christopher Smith, MD, Ginger Jiang, MD, and Osaze C. Udeagbala, BS/MBA for their roles in realising the events.
Footnotes
Contributors All authors designed the study. MA, ERD, BDL and ML collected and analysed the data and wrote the manuscript. They contributed equally to this paper and are considered cofirst authors. LAC and BKP reviewed the data and manuscript and gave final approval.
Funding Funding for event supplies (posters, Post-It notes, pens) and refreshments provided by the Beth Israel Deaconess Department of Medicine, the Beth Israel Deaconess Graduate Medical Education Office and the MIT Biotechnology Group. Cloud storage space provided by Google Cloud for the microdatathon. LAC is funded by the National Institute of Health through NIBIB R01 EB017205.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.