Health center chatbot

AI systems > Health center chatbot
Social services and health care

Overview

Health center chatbot

Tags

chatbothealth centercare need assessment

The chatbot provides health and illness-related advice easily without queuing. Chatbot directs the users to the right digital health services and advises on questions related to dental, mental health, substance abuse and social services. The service answers the most frequently asked questions and direct to the sources of reliable information based on users’ needs. The chatbot is integrated with other digital services of health centres. Through it, customers may be directed to, for example, the city’s digital services, Omaolo, Omakanta or social counselling chat. The service enables transactions around the clock anonymously and securely. All health centers use the service in Helsinki, and it covers all services of the health centers.

The clinical guidance of the service is based on national practices of evidence-based treatments, the Current Care guidelines and the emergency care evaluation criteria. The service utilises natural language processing based artificial intelligence to analyse the searches entered by the customer and to identify the intended service need, as well as to find the respective instructions.

The service expands the accessibility of health center services by enabling remote transactions around the clock. The service enhances the customer experience of digital services for health centers by providing a single channel for all digital services. The service streamlines the service production of health centers by automating the service on the frequently asked questions and by the automated collection of background information. The service also has positive effects on harmonising the quality of the service and on improving the employee experience of counselling services.


Link to service

Contact Information


  • Department

  • Helsinki City Hospital
  • Contact person for inquiries

  • Timo Lukkarinen
  • Contact email

  • sosiaalijaterveys@hel.fi
  • Contact phone

  • 09 310 5015
  • External vendors

  • Kuura Health Oy


More detailed information on the system

Here you can get acquainted with the information used by the system, the operating logic, and its governance in the areas that interest you.


DatasetsShow MoreShow Less

Key data sources utilised in the development and use of the system, their content and utilisation methods. The different data sources are separated by subheadings.

Contact reasons of health center digital services

The topics covered by the service have been determined in the early stages of the service’s development by utilising the anonymous statistical data on the contact reasons for the digital services of Helsinki health centers. In this way, it has been ensured that the themes that are of the most interest and concern to customers are included in the service from the very beginning. The topics covered by the chatbot are continuously updated based on current knowledge of clinical experts, as well as analysis based on discussion data.

Question-answer pairs and structured discussion paths

Question-answer pairs and discussion paths are determined using the Current Care guidelines, the national emergency care criteria, and the medical expertise of the City of Helsinki and Kuura Health staff. Discussion paths are continuously developed by using discussion analytics of the chatbot, and by leveraging the insights of health center clinical work experts on current themes and necessary guidelines.

Synonym database

The synonym database is a database developed by experts from Kuura Health Oy and the City of Helsinki. The database contains the alternative Finnish expressions to the subject words covered by the chatbot. The synonym database improves the chatbot’s ability to identify the user’s transaction needs from diverse use of the Finnish language.

Data processingShow MoreShow Less

The operational logic of the automatic data processing and reasoning performed by the system and the models used.

Model architecture

The search function of the service uses natural language analysis to identify the customer’s needs and to find the corresponding instructions. In this process, the keyword entered by the customer is automatically reduced into its base form by removing special characters and placeholders (stemming). The search terms are then automatically expanded by identifying similar search terms from the synonym database. The base forms of expanded search terms are then used to identify the matching instructions from the question-and-answer database by using the Term Frequency – Inverse Document Frequency (TFIDF) search algorithm.

Content

Model architecture

Non-discriminationShow MoreShow Less

Promotion and realisation of equality in the use of the service.

The service is currently only available in Finnish, and therefore only Finnish-speaking customers can benefit from the extended availability of remote health center services. The Swedish and English language versions of the service have been identified as development items. To further expand the accessibility, the use of speech recognition technology could be explored in future.

Human oversightShow MoreShow Less

Human oversight during the use of the service.

The service is monitored and further developed by appointed experts. Monitoring is done by reviewing the feedback and development ideas accumulated from the service on a daily basis, as well as by analysing the discussion statistics of the service. Based on these, experts identify new topics of interest to customers, as well as any potential needs to change the existing question-answer paths. The task of the experts operating the service is to create new discussion paths based on the analysis, as well as to improve the findability of instructions, for example, with the help of new keywords. The expert team operating and developing the service has the assigned back-up arrangements to ensure continuity in case of absences.

RisksShow MoreShow Less

Risks related to the system and its use and their management methods.

The risk level of the service is assessed as low. The service is a tool for self-assessment of care needs. The service is not a medical device, but it can direct customers to other medical devices.

The data protection risks of the service have been assessed by conducting a data protection impact analysis. The privacy risks relate in particular to the possibility of entering personal data into the service, even if the service itself does not request it. The risk has been managed with clear instructions and functionality, where the service warns if the search term entered appears to be a personal ID. All information entered by the customer is automatically deleted from the system within a week.

The service instructions supported by the chatbot and the referrals to other services require constant monitoring and maintenance for possible changes and to retain quality. For this, designated experts have been assigned to monitor the service quality, to maintain and to develop the service further.

The impact assessment and risk analysis have been carried out on 10/2019.

Was this information useful?