Operationalizing & Automating Patient Workflows by Howard Rosen (CEO), William D. Harms (Director of Research)

Background

LifeWIRE’s interactive communication platform was initially implemented in small clinical settings where one clinician managed 20-30 patients. In that environment it was relatively easy to check in with each patient, at least a few times per week, and to manually make any changes to their treatment plans. As the utilization has become more robust, and it is being depended upon on for the primary monitoring, the work environment demanded that routine tasks be automated, even using patient/member responses to automatically modify the patient’s treatment environment. Transitioning into a world of automated interaction requires levels of expertise and understanding not heretofore required in business.  Our several years of experience with multiple corporate clients has provided great insight into these unique skillsets.  Automated interactions are such a new discipline that in many cases it has been difficult for our clients to know what they needed until they actually saw it in action. Key to it all is marrying the need with the ease of use.  This effective interaction then allows the focus to turn to away from an operational disruption to deriving new insights from their patients.

Specific Example – Women’s Maternity Care

A great example of automation is within the Veterans’ Administration (“VA”) Women’s Maternal Care program. In this instance, MCC administers 14 interactions (Edinburg Depression, Psychosocial, Benefits, Breastfeeding, Domestic Violence, Medications, etc.) along with six sets of Reminders and Motivations to each patient. Interactions vary depending on where the patient has progressed in their pregnancy or post-partum care.  Each set of interactions are scheduled at various intervals.  Based on the responses to the questions referrals are made and other interactions or assessments initiated. Previously the clinician/administrator would have to access over 50 screens to simply enroll an end-user. It wasn’t practical. LifeWIRE’s solution was to ask each veteran’s delivery date. From this we calculate each occurrence of each interaction. When the clinician interviews the end-user they simply click on a ‘clock’ icon and the entire scheduling process is finished. The next challenge for them was monitoring all the data – there was too much time involved in going to each interaction for each user. Again, LifeWIRE provided a solution. The clinician enters an end-user name/number within a date range, and the system provides a sequential report of all scheduled and completed interactions. This becomes the interview dataset. It is then either electronically or manually (depending on the client), “pasted” into the end-user record and saved permanently. To complement this automation, LifeWIRE provided an interactive chat that the customer, clinician or end-user can initiate. Similar to Facebook, the clinician gets a chat window at the bottom of their screen they can use to interactively communicate with the end-user. They can have as many chat encounters open as they need.  Most critically, however, is that these interactive conversations then become inserted into the patient case notes as treatment documentation.

Automation Utilization: More Examples

The key message is: whatever process you find yourself doing more than three times per week, it should be automated. Here are some bullet points that may help:  

•  Any repetitive processing task, such as scheduling of an interaction or reminder, will be consolidated to a single keystroke.
•  Any routine reporting can be provided on demand to use as interview notes and then pasted to the end-user record.
•  Specific answers can be tied to a notification routine where the clinician is notified, and external program managers receive emails notifying them of an end-user’s needs (i.e. housing, financial, employment, etc.) and contact information.
•  End-users can be segmented into subgroups with differing treatment levels, schedules, or needs, allowing the clinician to readily access and track end-users in need.
•  Clinicians only have to sort through the groups they are focusing on.
•  Smaller groups result in faster processing times.

This provides a highly responsive platform. However, as a customer populates the platform with a large number of end-user/members the amount of data can easily become overwhelming. To manage that LifeWIRE lets administrators use groups to keep track of end-users following certain patterns. For instance:

•  The first automated question often asks ‘Would you be more comfortable responding in English or Spanish?
• •  Based on the response the system splits the end-users into two groups so they can be directed to the properly translated interactions and questions.
  •  Again, based on their responses end-users/members will receive their interactions with different frequencies.
•  Finally, LifeWIRE uses groups for various follow-up levels. End-users/members who respond in ways that indicate a particular situation can be moved to the appropriate group.

Each environment is different, but once the data is obtained it is used to minimize the manual interaction with the clinician/administrator. Once implemented, we discovered another unanticipated aspect. Employees of our customers were already tasked with a full workload, and implementing, monitoring, and making changes based on received data were above and beyond their full-time duties. In these cases, automation of these services addressed two critical needs. First, we were able to create a 'zero footprint' where daily reports and processes were automatically routed to the employee's mail and IT infrastructure. This created an environment where the interactions became part of the daily workflow. Next, but perhaps more importantly, automating incoming interactions allowed employees to prioritize tasks, focusing first on the approximately 8% that required acute human interactions. We were able to assist our clients with fully automated interactions between the system and their clients that addressed the remaining 92% of requests while delivering personalized support and interactions with their patients. Key to all this we have found is ease of use, and that it can never be TOO easy to use, nor TOO easy to accommodate changing needs. So where we have found that customized programming is required to meet every Customers’ unique program needs, it is crucial that they can do so themselves through simple pull down menus, in real time. We have found that within a client group, there are many similarities however, which can be grouped in various permutations to create new operations. If a program is used in several treatment facilities, the automation should transfer between facilities. Fundamentally, the more control of the platform they have, the more routine you can alleviate and the more insight they can derive, the better the value to Customers and end-users.