Transitioning To Value-Based Care? A Chatbot May Help. Session 214, March 8, 2018 Brett Swenson, MD, Swenson MD Premier Care

Transcription

1 Transitioning To Value-Based Care? A Chatbot May Help. Session 214, March 8, 2018 Brett Swenson, MD, Swenson MD Premier Care 1

2 Affiliation Disclosure Brett Swenson, MD Advisor: SimplifiMed, Inc. 2

3 Agenda My background and practice Value Based Care Engaging patients with a chatbot Questions 3

4 Learning Objectives The need for automation in Value-Based Care What is a Chatbot? How to successfully deploy a Chatbot in your practice? 4

5 5

6 Once upon a time 6

7 Value Based Care Value is achieved with happy, healthy patients. 7

8 Metrics Happy patients = High patient satisfaction Responsiveness Communication Healthy patients = Outcomes 8

9 The need for new solutions Responsiveness Old way Phone calls and voice messages New way immediate response Communication Phone, , portal Interactive SMS Outcomes Patient billing Quality measures 9

10 Chatbot A computer program designed to simulate conversation with human users, especially over the internet *. * Oxford dictionary 10

11 Engaging Patients with a Chatbot Phase 1/3 11

12 Patient initiated communication First phase focused on patient initiated communication for: Making an appointment (e.g. sick visit) Cancelling/rescheduling an appointment Medication refills Instead of calling the clinic, patients can simply text and interact with the BOT to secure these services. The BOT was not integrated with the EHR Required staff to regularly check SimplifiMed dashboard 12

13 The process Patients received a text, informing them availability of the BOT to secure services: Acquired consent for our modified HIPAA policy and added language describing use of SMS. Being conservative with the new technology, engaged patients in batches of 200/day The entire process took nearly 2 weeks Patient s were given option to opt out: <1% patients opted out Majority of patients opting out were residing in other cities (e.g. Snowbirds) 13

14 Teaching point: behavior change is hard Usage of the new system was low Most of the patients continued to call the clinic, instead of using the BOT: Patients are in the habit of calling the clinic Medication refill was the most widely used feature. A lack of integration with the EHR 14

15 Phase-1 result Measure Note Positives Patients received notice of new service available. Minimal opt-out. Chatbot successfully delivered notice of request from patient to staff to refill meds or schedule visits. Negatives Bulk of the patients still called the clinic due to long-established behavior. Office staff still had to manually enter appointments due to lack of EHR integration. 15

16 Engaging Patients with a Chatbot Phase 2/3 16

17 Polling Question How did your doctor remind you to get a flu shot this year? 1) Postcard/Call/ 2) Patient portal 3) Other 4) My doctor didn't remind 17

18 Clinic initiated communication Second phase focused at clinic initiating communication for preventive care (Flu vaccination). The BOT was not yet integrated with the EHR 18

19 The process Staff uploaded two CSV files: A file containing patients cell number A file containing available time slots. Patients received a text, nudging them to book an appointment to have the vaccination: Patient could decline the service. Patient could schedule an appointment for the vaccination. 19

20 Teaching point: 1 in 3 patients responded An overwhelming majority of the patients responded within the first 60 seconds of getting the first text. Almost all of the the patients who decided to get care, scheduled the appointment within 5 minutes of getting the first text. A tiny fraction (<5%) responded significantly later but the same day. 1 patient responded after 15+ days. 20

21 Phase-2 result Measure Note Positives Patients were notified of a preventative health recommendation (quality measure). Response rate was very high compared to other methods. Response was much quicker than other methods. Negatives Due to a lack of EMR integration, staff still had to manually enter appointments into slots after receiving the list of booked appointments. 21

22 Engaging Patients with a Chatbot Phase 3/3 22

23 Clinic Initiated Communication with Integration Being rolled out in Feb 18 Scheduling of preventive care (annual, quality measures) As opposed to focus on the external interface, in this phase we are internally focused The bot works directly with the EHR All the appointments directly made in the PM suite 23

24 The process System configuration Identify the list of CPT/ICD codes for preventive care measures Establish criteria to identify the patients The system is expected to run in the background without requiring any manual input from clinic staff. 24

25 Implementation Summary: EHR is the key Regulatory: Ensure that your HIPAA policy covers SMS/texting. EHR: EHR should have Single Sign On (SSO) capability EHR should have extensive set of APIs for two-way data exchange with the chatbot. HL-7 is insufficient. Operations: Start with approximately 500 patients/md. Proactively complete OK to text field in the PM suite. 25

26 Future Directions All preventative care and chronic care guidelines available for implementation Custom follow up based on procedure or condition Completing medical record/history Payment 26

27 The Big Picture Healthcare is moving toward value based care The healthcare systems and practices that thrive going forward will adapt to this new approach To provide value we must design the system for the benefit of the patient rather than the benefit of the provider organization 27

28 The current suite of solutions has a hole Phones are answered only when staff is available Paging after hours is only for emergencies s are mostly unread or not acted upon Portal usage is low Adoption/installation of Apps is low. 28

29 Chatbot is the way to go 24/7 access SMS based interface Interactive Designed to provide nudges that support preventative care and chronic care management to meet quality measures and improve outcomes. 29

30 Questions Please complete the online session evaluation Contact info : Brett Swenson, MD healthtechdoc@gmail.com 30