Details

Review Date
11/13/2023
Purchase Date
Q2'23
Implementation Time
One week
Product Still in Use
Yes
Purchase Amount
Per visit completed using Ambience
Intent to Renew
90%
Sourced by

Product Rating

Product Overall
4.0
Use Case Fit
4.0
Ease of Use
4.5
API
N/A
Integrations
4.0
Support
4.0
Value
4.0

About the Reviewer

Purchasing Team
Implementation Team
Product Oversight

Reviewer Organization

N/A

Reviewer Tech Stack

athenahealth

Other Products Considered

Nabla

Summary

  • Product Usage: Ambience, an AI scribe, is integrated with athenahealth and is used for patient consultation and note-taking, enabling the clinicians to focus more on patient care.

  • Strengths: Ambience records every consultation, automates note-taking, and is a significant time-saver; it takes only a couple of minutes to produce a note that requires minimal editing.

  • Weaknesses: The system may experience outages, disrupting workflows, and sometimes producing inaccurate medical notes.

  • Overall Judgment: Despite its weaknesses, Ambience has generally met the companys key requirements, and the majority of their clinicians appreciate it.

Review

So today we’re chatting about Ambience and how it’s used at your company. Before we jump into that, could you give a brief overview of the company and your role there?

We are a digital health provider organization. Our team of clinicians specialize in women’s health, hormonal changes, and the body changes that occur during different stages of life. We provide telehealth services, and all of our clinicians are advanced practitioners.

As part of the product team, I focus on clinical care delivery, taking care of everything from the moment a patient books a visit to the completion of the visit and follow-up care.

What was the need that drove you to look for a product like Ambience?

Our main goal was to ensure that we could consistently provide high-quality care to every patient, regardless of their level of complexity. With telehealth, it was important for us to find a way to truly connect with our patients and show that we were present and engaged. We wanted to go beyond just typing on a computer and instead have eye-to-eye contact and create a sense of connection and support.

Another important aspect was addressing the challenges that come with telehealth. Clinicians were spending long hours in front of their computers, not just counseling patients, but also documenting, charting, and carrying out various tasks. We wanted to find a way to provide excellent care while also being efficient and minimizing day-to-day work stress. Our goal was to create a positive work environment for our clinicians.

What were the key requirements that you were looking for as you evaluated these types of solutions?

First, we wanted to ensure that the patient’s perception of the solution was positive and that they were able to maintain trust in our clinicians. Safety was a top priority for us, so we conducted thorough analysis to ensure there were no safety concerns.

The next requirement was clinician trust and adoption. We wanted our clinicians to adopt the solution as soon as possible, especially if we were going to integrate it with Athena. We also wanted the solution to have a high level of accuracy and minimize the need for additional editing. This was crucial to avoid any quality concerns or patient safety risks.

Efficiency was another important requirement. We wanted the solution to be a true time-saver for our clinicians, helping them streamline their workflow. Finally, we required seamless tech-to-tech integration. We emphasized to their team that the solution must autofill into our note template, with no copying and pasting or manual work involved. This integration was key in simplifying our clinicians’ interaction with the EHR and aligning with our goal of making their lives as simple as possible.

What do you mean when you refer to safety?

I mean ensuring that what gets documented is completely accurate, which is critical to ensuring high quality patient care. While our clinicians edit all notes, there is always a potential for human error. If something false is documented, it becomes a permanent part of the patient’s medical record. This is a major concern. Additionally, in our type of care, medications may be prescribed at doses or for uses that differ from the most common ones. We don’t want medications to be taken out of context or misrepresented in the notes generated by our AI model.

What other vendors have you investigated, and how did they compare to Ambience?

We initially chose Ambience at a time when there weren’t many big players in the market. We recently have considered Nabla, who stood out due to their willingness to provide personalized services to each user. Their team is attentive and works closely with clients to optimize processes for healthcare providers.

We were promised similar levels of personalization with Ambience. However, as they and we have grown, and we’ve had less time and fewer available resources, we’ve noticed less focus on customization. There are also noticeable differences in user interface between the two platforms (though both are Chrome extensions). They’re both great, but they differ in dashboard design and editing capabilities within the dashboard.

Although we’ve only used Ambience for a few months, it wouldn’t make sense to switch to another product and have our clinicians adapt to a new platform again. While we might contemplate a change again in the future, at present, we are staying with Ambience.

Have you found a difference in the quality of the note between Ambience and Nabla?

Nabla originally built their product on women’s health, so their model appears to accurately handle women’s health aspects from the get-go, based on my experience with a simple demo and independent usage without any assistance from their product team. It might have taken us less time to reach our current stage with Nabla as opposed to Ambience, given that we work in women’s health, but without having used Nabla in any substantial way, it’s difficult to make a definitive comparison.

How was the sales process with Ambience?

The sales process was smooth. The team focused on our goals, requirements, and desires, driving a successful sales process. They provided significant support during our pilot phase, going so far as to meet with our clinicians to ensure full adoption. The level of support experienced during the initial sales process and the pilot phase unfortunately hasn’t been maintained, given their growth strategy.

How was the onboarding and setup process?

Our clinicians initially met with Ambience representatives for one-on-one or small-group training sessions. The pilot phase included our CMO initially, and then we added five other clinicians, each of whom received training at the start of their shift and provided feedback after using Ambience for a full day. We found the most effective training method was the 30-minute live Zoom sessions, with optional 15-minute afternoon check-ins via Zoom or a quick Slack check-in. We did keep weekly group drop-in Zoom sessions available, but they weren’t highly utilized.

On the technical setup side, we focused on ensuring the autofill functionality was accurate, so it matched our encounter note templates and integrated properly with Athena. The process of fine-tuning Ambience for our requirements included reviewing patient recordings and template accuracy with clinicians individually. The initial work in training the model was not extensive, but we’re now revisiting this with their engineering team to prevent unwanted automation behaviors and ensure the system meets our risk assessment criteria. Our use cases have been presented to the engineers, who now seem more responsive to our feedback and willing to make the desired adjustments.

How are you using Ambience in your clinical workflow?

Clinicians at our facility utilize Ambience in two ways. Ideally, patients complete a detailed health questionnaire before their visit, allowing clinicians to prepare verbally. They access the patient’s chart in Athena, enable the Ambience Chrome extension, and confirm the captured details like patient name, date of birth, gender, and visit type. Our visit types are linked to templates in Ambience.

Our clinicians first use Ambience for AI-based dictation, pre-reading the health questionnaire and documenting key information in the HPI. They pause the recording when they finish this step. When they start the visit with the patient, they obtain patient consent to continue recording the encounter with Ambience in the background. Some clinicians enter diagnosis codes during the visit, which generates the problem list, while others do it post visit. That’s up to their personal preference. After the patient departs, clinicians finalize their diagnosis codes in Athena, check the problem list, and proceed to the Ambience dashboard for note generation and editing. This editing step within Ambience helps to train the model.

Once edits are complete, clinicians use an autofill feature to insert the transcription into the Athena note, which usually takes less than a minute. They add any prescriptions, lab orders, or additional diagnosis codes not captured by Ambience, then send the chart for clinic leadership review, concluding the process.

Is Ambience automatically generating codes out of the conversation, or are the clinicians listing the codes themselves?

We ask our clinicians to enter their own diagnosis codes as they’re having the conversation with the patient. This helps them recall the patient’s concerns, presentation, and medical history. Ambience recently introduced a billing dashboard feature, displaying a billing card on their main dashboard. This card reviews the recording and suggests additional diagnosis codes that clinicians can consider adding to the chart. While this feature wasn’t widely used initially due to a one-way flow issue—information pre-populated in Athena reflected in Ambience, but not vice versa—they resolved it last week. Now our clinicians can add new diagnoses in the Ambience dashboard and it syncs back to Athena without autofill issues. We’re monitoring its usage to see if clinicians adopt it more, but it’s not mandatory since our clinicians are adept at manually adding diagnosis codes based on patient problem lists and medical history.

Are you having any issues with Ambience?

Background noise isn’t a big problem. However, one issue we face is, if a clinician wears headphones, it can disrupt the system. The alignment between the headphones and the Chrome tab is crucial; if it’s not right, the recording isn’t captured. This poses a challenge, especially since many clinicians prefer headphones for privacy while working from home. It’s a patient safety measure when they don’t have a closed room.

Another hiccup occurs when patients encounter tech issues and can’t join through Athena’s telehealth platform. In such cases, the provider calls the patient, maintains the video, and uses Doximity dialer for audio. If the Doximity dialer isn’t close enough to the computer, we face audio capture problems. Clinicians are trained on this, and Ambience now has its own dialer, making things a bit smoother. Some clinicians use the Doximity app on their computer to avoid issues, but it still happens occasionally. The constant movement of the sound bars on Ambience serves as our visual indicator, showing if audio is being captured on both ends.

How is the reliability and stability of the platform?

In terms of stability, Ambience does experience outages and issues. Their team is quite understanding and will rerun all the notes, but it significantly impacts our clinical team. If our clinicians can’t close the chart from one patient before they move on to the next one, they’ll have to do extra work after hours, which is far from ideal for them or us. So, stability can be a concern.

Occasionally, the note isn’t ready or the autofill breaks. There are instances where part of the encounter note doesn’t generate from Ambience, like the discussion note sections or the assessment and plan. You end up with fragments instead of the full note. However, the recording is still captured. As long as that’s in place, they can run it through the model later, and eventually, we get the note.

Does Ambience provide patient-facing notes?

They provide the discussion note template, which integrates into our discussion note in Athena. However, they don’t generate our care summary template. We handle that aspect through Athena. I’m not sure if they offer it as a product feature, but even if they did, we likely wouldn’t take advantage of it. We’ve stuck with our custom template built with Athena from the early days.

What would you characterize as the key strengths and weaknesses of Ambience?

Ambience’s strengths lie in the high certainty that the recording will be captured, allowing clinicians to focus on one-on-one patient conversations without constant typing. The significant time savings is another plus, with a consistent note generally produced in just a couple of minutes and requiring minimal editing, compared to the 15-20 minutes of traditional charting.

However, weaknesses come into play with product stability. When there’s an issue, it disrupts clinician flow, and in cases of patient safety concerns, it has a substantial impact, such as with errors in medications and dosing or incorrect patient health history. Initially, there were more instances of inaccuracies, with transcriptions sometimes stating the opposite of what was discussed. This led to the need for extensive editing, actually increasing charting time. With increased familiarity with Ambience and adjustments on the Athena side, these issues have decreased, but they still pop up occasionally.

How much customization is available?

All of our providers use the same interface. We can customize by visit types, but building out new templates requires negotiation with Ambience.

How was the integration process with Athena?

It was quite smooth. They handled the entire integration for us, initially using whatever they could with the Chrome extension. For us, I shared our ideal user flow and the best notes from our clinicians, indicating how we wanted elements to flow into different sections of the chart. We had several meetings, I provided documentation, and they took notes and took care of it. I continue to meet with them for new releases to ensure they align with our clinicians’ needs and fit seamlessly into our workflow. The heavy lifting was taken care of by their engineering team, making it one of the most seamless parts of our collaboration.

How has your experience been with Ambience’s support?

It’s been a bit of a rollercoaster. It started off strong during the pilot phase, which is typical when a company is showcasing their capabilities. After signing with them, we encountered a dip in the support quality. We provided feedback, and subsequently we were reassigned to a new support person. Since then, the support has improved significantly.

We’ve also made some changes on our end. Now, we have a medical liaison, and we both engage with Ambience together. Some points are better conveyed from a product standpoint, while others are more effectively communicated from a clinical perspective. Teaming up allows us to address concerns comprehensively. It took some time to establish effective ongoing communication and get our concerns addressed promptly, but it’s good now. There was a definite positive change after the initial three weeks of less-than-ideal service.

How do you typically engage with them from a support perspective?

We meet weekly with our account managers, who also handle clinician onboarding. As needed, we have meetings with their engineers or different team members. Their sales team has someone dedicated to updating us on new features and gauging our interest in piloting them. Additionally, we’ve set up ongoing Slack channels between our teams and an extra channel specifically for clinician support. The Slack setup works effectively because our clinicians are comfortable with it, providing them with immediate, on-demand support from the Ambience team.

Do you feel like you made the right decision in going with Ambience?

I think we did, although it might have been beneficial to spend a bit more time exploring what other competitors were offering at the time of adoption. They’re meeting our key requirements, and the support team is great about solving issues when they arise. The majority of our clinicians really appreciate Ambience, and to some extent, they’ve become dependent on it. So removing it without a suitable replacement wouldn’t be feasible. We’d need to have something similar in place.

What would lead you to look elsewhere?

We’d reassess if there were persistent safety and quality issues or if the product stability compromised the timeliness and efficiency goals we initially set. The safety concerns revolve around our clinicians’ trust, while the efficiency aspect is more aligned with our business and operational goals.

Do you have any advice for people who are in the process of evaluating these solutions right now?

Definitely make sure to define and contract the ongoing support structure you’ll need from them in advance. Since we’ve had our new account manager on board, the support has been excellent. A key piece of advice is to ensure that the partner from Ambience has a clinical background. In our case, our partner is a PA, so she understands our challenges and their impact on clinicians, which has been incredibly beneficial.

Another important aspect is understanding the QA processes, both internally within your company and with the vendor. Establishing a strong mutual understanding between the internal clinical leadership team and the vendor’s clinical leadership team is vital. This is something we’ve learned over time because different entities have varying levels of risk tolerance and comfort regarding accuracy. Aligning these aspects ensures a smoother collaboration.