AI Analysis on How One Medical Delights Patients and Improves the Patient Experience
The data science on One Medical’s glowing reviews and how their delivery model drives their patient experience.
How One Medical Siphons Market Share
When Healthcare Dive asked One Medical’s CFO, Bjorn Thaler, on why they focus on primary care:
“Frankly, because it’s ripe for disruption and nobody else does it.”
Given this past decade’s billions of dollars in health system mergers and consolidations, “nobody else does it,” is strong language. However, where’s the lie? Patients still wait long hours in uncomfortable environments filling frustrating forms, and it took a pandemic to normalize virtual visits.
In the words of a One Medical patient from the 1,000+ reviews we’ve analyzed, “They had the whole ‘video visit’ thing down before we were all forced into it!” One Medical’s two-prong approach:
- Mend primary care’s weaknesses with modern technology
- Refer patients to the biggest brands in their markets (Axios)
What does it take to contend with a two-fisted juggernaut?
Your Patient Experience is in Your Reviews
Patient surveys such as HCAHPS can’t capture the reasons for fast-changing patient sentiment. For example, access your latest HCAHPS results and list the top reasons for great patient experiences after COVID-19.
You can’t. Surveys are limited to the questions organizations thought to ask, so the flood of new paradigms such as video visits and curbside service don’t appear in HCAHPS. Additionally, numerical rankings remove crucial nuance about patients’ experiences. These rankings may provide overall ideas on quality and satisfaction, but there isn’t enough nuance to make the voice of the patient actionable.
Online reviews are great at capturing specifics! However, unstructured data is the most challenging to deal with. Often, teams will read reviews and attempt to perform qualitative analysis based on trends they manually identify. These approaches are time consuming and lack statistical rigor. Most often, then, this valuable patient feedback can’t influence the strategic decisions needed to improve patient experiences.
And Health Systems Can Finally Listen to that Feedback
Natural Language Processing (NLP) is a subset of data science concerned with understanding or even generating free-form text. NLP helps identify speaker sentiment, summarize documents, and write news articles, but can NLP help us understand large amounts of feedback on the patient experience?
Scattertext, a Natural Language Processing approach by Amazon’s Jason Kessler, is powerful for understanding sentiments across thousands of reviews. Below, we use this approach on 1,500+ One Medical Yelp reviews to learn what patients are saying and why:
The view below can take up to two minutes to load:
It’s pretty, but what does this tell me about my patient experience?
We chart phrases by how likely they are to occur in “Great!” reviews ( >3 stars) or “Bad” reviews (<4 stars). The farther right, the more frequently words occur in bad reviews, and the farther up, the more frequently words occur in great reviews.
We focus heavily on terms in the upper left (“desk staff”, “good things”, “video visit”, etc.). This represents where great associations are the highest and negative associations are the lowest. Conversely, phrases in the lower right quadrant (“insurance company”, “billing department”, “nurse practitioner”) are associated with poor reviews. The upper right phrases (“onemedical group”, “same day”, “primary care”) appear evenly in great and bad reviews.
Click on any of these terms to explore exactly what patients are saying about them. To answer the above question about what’s helped with great experiences post-COVID19, we start with video visits.
Winning with Video Visits
If you click or search “video visit,” you’ll see how enthusiastic patients are about telemedicine. There are only three “video visit” mentions in bad reviews, and those are associated with either timing or insurance.
What patients are saying they love:
- 24/7 access
- Speed of being seen “I did a video visit with a PA… and they booked me an appt for an hour later”
- Skipping office visits
Video visits aren’t a one-size-fits-all approach, so delivering care according to patient preferences is key for a good patient experience.
Timing is Everything, Speed is Crucial
Speed is one of One Medical’s major brand promises. If you click “appointment time,” you’ll note that they deliver on this promise. A Chicago reviewer noted:
“At exactly my appointment time (love it), I was called back for my appointment”
The excitement over on-time appointments in several reviews has poor implications on overall healthcare experiences. For organizations that are delivering on this promise, however, it’s great!
Furthermore, many key phrases in great reviews referred to staff, rooms, waiting area, and other components of service & facilities. You can explore these easily by looking for related phrases under “Top Great!”
Bill Me Up Just to Let Me Down
Although the overall experience tends to be great, patients are extremely negative about their billing practices. Clicking into phrases such as “billing department” paints a grim picture. Many references say “Incompetent,” and patients note that their requests either get lost or ignored.
“Insurance company” isn’t too far off from the negative perceptions associated with the billing department. Many reviews fixate on a poor relationship between One Medical and their insurers. Reviews highlight surprise drops in coverage or prices much higher than what insurers will pay.
Billing tends to be a weak point for most organizations. There’s an opportunity to provide a better experience for patients as they work through the most difficult times in their lives.
How to Take Action
These overall analytics are available to help your organization strategize on patient growth and loyalty. If you’re in a competitive market, exploring here opens the door on what patients are saying without expensive and awkward surveys.
These overall topics can translate into Next Best Actions. Some options are directing video visits to people who would have a great experience with them or better coordinating billing efforts.
Next Best Actions use AI to surface important steps for patients and drive your organization to fill them. If you’re interested in how Next Best Actions drive strategic goals, our Voice-of-Patient Benchmark performs this analysis on your health system and up to three competitors.
Chris Hemphill leads market facing data science projects for SymphonyRM. By combining experience in operations, sales, and marketing, Chris's goal is to improve the quality of decisions that health systems make in order to better patient care.
In addition to duties at SymphonyRM, Chris speaks and teaches on data science regularly.
