The healthcare establishment which won this year’s first prize in the European Private Hospitalization Awards in this category is:
Cuf Hospital in Lisbon, Portugal
for its Outpatient Surgery Follow-up contact by a virtual assistant
Nurses represent a scarce and immensely valuable asset within the healthcare sector. It is imperative for healthcare providers to meticulously consider the role of these professionals and optimize their effectiveness by assigning them primarily to direct patient care, thereby reducing administrative and non-specialised tasks.
The integration of artificial intelligence (AI) serves as a pivotal tool in alleviating administrative burdens of the nurses’ teams. Hence, CUF embarked on a venture to assess the acceptance of a Virtual Assistant for outpatient surgery follow-ups among our customers. This decision stemmed from the recognition that our nursing staff, across all units, invested significant hours daily in repetitive phone calls.
Initiating with the translation of these predefined contact protocols into a voice script, in collaboration with the startup Tucuvi, we commenced the training of the virtual assistant to proficiently communicate, comprehend, and write in Portuguese. The successful implementation of this project required the collaborative efforts of a multidisciplinary team encompassing technical personnel such as engineers, clinical experts including nurses and doctors, and project managers.
Our goal was to deploy a solution that clients would readily embrace, comprehend, and interact with, capable of succinctly summarising conversations while highlighting crucial topics. Moreover, the system was designed to alert for any abnormal responses from clients, thereby enabling prompt intervention by the clinical team.
This project yielded notable outcomes, with client satisfaction in engaging with the virtual assistant scoring an impressive 8.9 out of 10. The detection of clinical risks among clients was markedly enhanced, facilitating expedited responses from the clinical team. Importantly, it has allowed our nurses to devote more time to direct patient care within the hospital setting, thereby fostering increased job satisfaction as they engage in tasks aligned with their professional training.
Furthermore, the reports generated by the virtual assistant offer comprehensive, standardised, and accurate documentation, facilitated by the utilization of speech-to-text technology, thus surpassing traditional handwritten notes.
Initially, we identified a critical use case requiring an innovative solution: clinical follow-up interactions. Collaborating with a Spanish startup equipped with class I medical device certification and expertise in this domain, we embarked on the development of an outpatient surgery follow-up project to be piloted within one of our medium-sized hospitals.
By pooling our resources, we built the scripts and trained the virtual assistant to effectively communicate, comprehend, and generate written content in Portuguese, bridging the language gap as the startup had primarily operated in Spain. This step demanded substantial investment in time and effort to refine the natural language processing capabilities of the virtual assistant and optimize the speech-to-text functionality.
Concurrently, enhancements were made to the dashboard features and alert configurations to ensure that all alerts received thorough review by the clinical team. The virtual assistant was named Clara, and a set of key performance indicators (KPIs) were established to assess the pilot’s efficacy, including the percentage of follow-ups conducted by Clara, the percentage of contacts deemed clinically appropriate, the percentage of clients comfortable with Clara’s interactions, the percentage of alerts routed to nursing staff, and the average call duration.
To ascertain the effectiveness of our solution, we conducted during the initial month duplicate calls, with both the virtual assistant and the nursing team reaching out following outpatient surgeries, ensuring seamless integration and adherence to established protocols. Encouraged by the resounding success observed during the trial period, the pilot progressed to random audits, culminating in a very positive evaluation overall. Subsequently, following a thorough examination by our Data Protection Officer (DPO), IT, and legal teams, a unanimous decision was reached to extend the implementation of the solution across all twelve of our healthcare units.
The virtual assistant showcases remarkable proficiency in both verbal communication and speech analysis, coupled with a remarkable capacity to adapt to the intricacies of healthcare contexts. Capable of comprehending inquiries, drug names (both commercial and generic), and even offering rephrased questions upon client request, its linguistic and contextual aptitude is noteworthy. Its linguistic and contextual capabilities are impressive to the point where, despite lacking a human voice, some clients are oblivious to conversing with a machine, attesting to its adherence to a real human interaction.
The ability to triage alerts in levels of severity based on the answers given by the patients is also a crucial AI feature that helps the nursing teams to focus on the most critical patients first.
Moreover, the reports generated by the virtual assistant surpass those traditionally produced by nurses, exhibiting a comprehensive inclusion of pertinent details. This enhanced quality is the result of rigorous quality controls implemented throughout the process, ensuring accuracy and completeness.
The tool has a very robust analytics component, which essentially provides an overview of key performance indicators such as the percentage of people with fever, the percentage of people experiencing pain, etc. This provides a general overview of the post-operative recovery progress, enabling nursing teams to take concrete actions to improve it.
The benefits for healthcare professionals are better time management and efficiency:
- 80% less time spent in the follow-up task;
- The follow-up coverage is now 100%;
- Nurses have more time to spend on in person care;
- Better prioritization of the most urgent cases;
- Increased efficiency in patient contact (since whenever they need to take any action, they know the reason beforehand);
- Better reports (more complete, standard and accurate).
For the clients:
- Secure follow-up – the virtual assistant ensures all the outpatient surgeries have a follow-up call, to guarantee that the patients are recovering as they should (in some cases the nurses could be overload in work, and miss the follow-ups; in other cases the patient would not answer the first time, and the virtual assistant will, undoubtedly, call back and receive the answers needed) .
- Comprehensive follow-up – an assistant that takes into account every word said and alerts the clinical team to provide support in a short period of time.
- Accurate follow-up – a complete report from the call assures the clinical team has all the information organised and stored, vanishing second contacts with the client to confirm notes.
- Greater comfort in the following post-operative period, as technology always adheres to the established schedules and never misses any contact.
Operational benefits and efficiency gains from the implementation of the AI integration project
Having all the regulatory and security measures into account, the technology can be applied in many different scenarios:
- Post-surgical follow-ups for conventional surgeries;
- Pre-surgical questionnaires;
- Daily monitoring of patients receiving home care;
- Daily monitoring of convalescent patients with complex or multimorbid conditions;
- Monitoring of chronic patients;
- Pre-anesthesia questionnaire;
- Post-discharge follow-up;
- Follow-up after emergency room visit;
- Exams preparation;
- Among others.
