83bar is a marketing agency specializing in healthcare.
We pair ad and social media services with a call center of nurses who educate, qualify, and schedule patients for appointments. To remain HIPAA (Health Insurance Portability and Accountability Act) compliant, most 3rd party marketing services are not an option. We developed a dashboard web app to integrate information about patients across platforms. The dashboard displays survey, email, and appointment information for various clients, clinics, and our internal strategy team.
See a quick video on the 83bar services here.
The dashboard is in continuous development.
Initially, the dashboard was primarily for internal use to manage campaigns. As marketers, our priority was to paint a picture of the patient journey and create campaigns to educate and guide patients to the rights health services. As we took on bigger enterprise level clients, it was clear we needed to paint another story for our clients about our process and results.
We created new permission levels and user types for specialized roles in the health care industry:
coordinate the physician and patient appointments.
Typically, this would involve multiple doctors at a single office. At large hospitals or networks, a manager can be in charge of more than one office location and more than one doctor.
manage specific procedures and medical devices for many clinics and regions.
They often evaluate campaigns and performance of specific clinics. They provide resources for the physicians and additional follow up for patients.
Our internal team
involves strategists, account managers, and call center managers.
Our internal team analyzes campaign success metrics and produces reports for clients.
Initially, 83bar was something of an assembly line. We ran campaigns, handed off patients to clinics and never heard back from them. We were severely lacking information about the intent of patients scheduling appointments.
We wanted to make sure that the people we contacted were truly in need of the health services. Our relationships with clinics and clients became a conversation back and forth to manage patients before, during, and after appointments. This follow up information from clinics made it easier to understand patient needs and optimize campaigns.
I was the sole designer for client facing tools.
Building off our internal suite of tools, I reimagined or helped develop new tools for clients. I fulfilled UX, UI, and some front end development roles for each of these projects. I visualized databases for external users, created user flows and wireframes, coordinated work with the development team, and in some cases aided in or provided CSS for the development team.
Each project was approached as an isolated tool to solve specific client problems. We are currently working towards conceptualizing our dashboard as a unified piece of software with a suite of tools.
This was the first tool designed for clinic managers to manage various appointment types and statuses at any given in time.
Problem: The assembly line hand off wasn't enough to manage appointments. Our nurse call center, as the first point of contact for these patients, had earned their trust and credibility. Often times, patients would call our nurses rather than the clinic to manage their appointments or ask questions. The clinics needed the information we had on patients to provide better care and earn that patient trust.
Objective: Allow internal users and clinic managers to manage appointments, track outcomes, and follow up with patients before and after appointments.
We chose a week-long view so clinic managers could plan ahead and prepare for incoming patients.
Each day includes a total count of the appointments and their statuses. Appointment types, such as an initial consult or a follow up appointment, can be filtered.
Each appointment is color coded and given icons to indicate a specific status and outcome. Appointments with bright red indicators required follow up from the clinic managers.
These appointment outcome statuses allow our team to understand the effectiveness of our campaigns--which patients showed and truly needed these services?
"Appointment Types" are preconfigured by internal strategists to set time spans to accommodate short, long, or (infinitely) double booked appointments. This allows us to track a patient's journey form an initial consult (often 5-30 minutes) to full procedures (hour+ in length).
Each appointment slot is tied to a patient profile listing their appointments, notes from call center nurses and a time line of previous touch points of engagement.
Appointments and other data can be edited or added from these popup modules. Upon hitting "Save All Changes", automated messages and confirmation emails tied to specific patient statuses or appointment statuses (ie a cancelation) are sent out to patients.
The more our clients used the app, the better our metrics became to direct marketing strategies. As reporting got easier for clients, we could see details about which patients were showing up and why. We also began to identify where campaigns were going wrong or if there were problems occurring due to clinic behaviors.
Meticulous clinic managers were using the calendar on a daily basis to update the appointments in their electronic health records systems (EHR). However, we struggled to on board other clinic managers.
Problem: Many clinic managers were resistant to adopting new tools. They were often stressed by busy clinics or were not very tech savvy (some clinics are still paper only).
Objective: Create a single stop source for appointment information. Use call to actions and a simple tool that would require the minimum engagement necessary for clinic managers. If a manager only engaged once per day and failed to answer emails, this would cover any appointments coming in.
Appointments that have changed/affect the schedule appear as a notification or task on a list. Each type of appointment change (cancelation, booking, etc) has a specific call to action. Each table contains the information needed to complete that call to action.
As clinic managers complete each task, it can be cleared so that any other schedule coordinators at the clinic do not duplicate the work.
A count of appointments for the next 7 days is always present at the top to provide a quick reference to the calendar.
All future appointment changes (new bookings, cancellations, reschedules) are focused on tasking the user to transfer changes to the clinic's electronic health records (EHR).
All past appointment notifications are focused on updating the 83bar dashboard.
On-boarding clients and clinics to the new home page should increase engagement with the 83bar Dashboard. This simple tool should serve as an easy introduction to the dashboard. We hope to lead users to become more comfortable with other dashboard capabilities, such as the calendar and search tool.
As users engage, the products will be refined to fit their emerging needs.
Our current search tool is only designed to seek out single individuals, typically for troubleshooting. Account managers and strategists need to regularly generate reports and lists of patients for clients. That data can only be accessed through our development team.
Problem: Account managers and strategists need to regularly generate reports and lists of patients for clients. That data can only be accessed through our development team due to the interface limitations. Complex reports are often collected in parts and then recomposed through excel.
Project: Create a tool that can search and display all data sets and types in our backend. Allow the user to specify the display and sort of the information. The scope is limited to searching lists of individuals; aggregate data must be handled by a different developer tool for data visualization.
Our first priority was to take into account all the types of backend data tables that were available for the search. Our data tables are generally divided into 4 types:
Lead/Patient information such as name, email, and phase (where they are in the funnel, such as 'booked' or 'rescue')
Appointment data such as appoint time, type, and status (ie canceled)
Interaction data or call center data. An interaction can be a set of calls or emails within a very short period of time with a call center agent. During calls, an agent can send patients a specific email or information about a service.
Client information is applied to all these data tables, so it became a ubiquitous category. A user will typically specify this first in their search process.
Much of the user needs could be indentified by looking at reports currently being generated. Qualitative interviews help shed light on needs that weren't yet being met by any tools. Account managers walked through different client calls. Call center managers showed me their troubleshooting process for campaigns and the call center.
Identifying when and why leads are being disqualified
Optimize time of campaign materials and booking
See lead demographics across clients, identify cross-client lead patterns/problems
Review appointment and interaction outcomes
Call Center Agents
Troubleshoot specific patient issues
Determine call positioning and timing (identify major barriers to scheduling, misconceptions, best frequency for call backs)
Understand lead patterns of behavior (non-responsiveness, causes for cancellations, bookings/success)
Identify agent behavior that need correction
- Generate regular reports for clients (all patient updates and new appointments)
Prove ROI and value for clients
Identify poor clinic practices or patterns of behavior (ie review patient feedback, consult on sales process)
Explain problem appointments and cancelation reasons
The search tools was distilled into 4 major steps
Search specify the parameters to narrow your list of leads
Column and sort selection control the data to display and how to sequence the results
Result display editing the parameters, columns, or sort sequence
Saving reports for future use or to create template reports
The UX patterns we looked at fell on a spectrum between a step-by-step guided method and a complex display of all the possible options.
Guided search methods are typically better, but many of our users are power users and know exactly what they're searching for. We wanted to reduce the number of clicks to achieve their goals.
Showing all the data options was too overwhelming. Each search parameter has different field types, such as a date picker or number range, as well as an operator, such as include, starts with, etc).
Certain key data points are dependent upon other search parameters, so a guided process was necessary to accommodate search dependencies.
guided Advanced Search
We settled on a guided search process. Data categories are easy to tab between so that a user can view all the possible parameters.
Cards are generated for each parameter chosen. Each card contains a set of operators to choose from and fields to specify the parameter.
Each card is an independent parameter that adds a layer to the filter. Each card only functions as an 'and' statement. We limited this project's minimum viable product to avoid 'or' statements for the sake of simplicity.
The only use case where 'or' statements were necessary were for clinic update reports for account management. In this case, we create a specific filter card that could include all the known parameters that would be needed.
Column and Sort Selection