Unlocking human-like conversations: Scoping review of automation techniques for personalized healthcare interventions using conversational agents

BACKGROUND

Conversational agents (CAs) offer a sustainable approach to deliver personalized interventions and improve health outcomes.

OBJECTIVES

To review how human-like communication and automation techniques of CAs in personalized healthcare interventions have been implemented. It is intended for designers and developers, computational scientists, behavior scientists, and biomedical engineers who aim at developing CAs for healthcare interventions.

METHODOLOGY

A scoping review was conducted in accordance with PRISMA Extension for Scoping Review. A search was performed in May 2023 in Web of Science, Pubmed, Scopus and IEEE databases. Search results were extracted, duplicates removed, and the remaining results were screened. Studies that contained personalized and automated CAs within the healthcare domain were included. Information regarding study characterization, and human-like communication and automation techniques was extracted from articles that met the eligibility criteria.

RESULTS

Twenty-three studies were selected. These articles described the development of CAs designed for patients to either self-manage their diseases (such as diabetes, mental health issues, cancer, asthma, COVID-19, and other chronic conditions) or to enhance healthy habits. The human-like communication characteristics studied encompassed aspects like system flexibility, personalization, and affective characteristics. Seven studies used rule-based models, eleven applied retrieval-based techniques for content delivery, five used AI models, and six integrated affective computing.

CONCLUSIONS

The increasing interest in employing CAs for personalized healthcare interventions is noteworthy. The adaptability of dialogue structures and personalization features is still limited. Unlocking human-like conversations may encompass the use of affective computing and generative AI to help improve user engagement. Future research should focus on the integration of holistic methods to describe the end-user, and the safe use of generative models.

Predicting post-discharge complications in cardiothoracic surgery: A clinical decision support system to optimize remote patient monitoring resources

Cardiac surgery patients are highly prone to severe complications post-discharge. Close follow-up through remote patient monitoring can help detect adverse outcomes earlier or prevent them, closing the gap between hospital and home care. However, equipment is limited due to economic and human resource constraints. This issue raises the need for efficient risk estimation to provide clinicians with insights into the potential benefit of remote monitoring for each patient. Standard models, such as the EuroSCORE, predict the mortality risk before the surgery. While these are used and validated in real settings, the models lack information collected during or following the surgery, determinant to predict adverse outcomes occurring further in the future. This paper proposes a Clinical Decision Support System based on Machine Learning to estimate the risk of severe complications within 90 days following cardiothoracic surgery discharge, an innovative objective underexplored in the literature. Health records from a cardiothoracic surgery department regarding 5 045 patients (60.8% male) collected throughout ten years were used to train predictive models. Clinicians' insights contributed to improving data preparation and extending traditional pipeline optimization techniques, addressing medical Artificial Intelligence requirements. Two separate test sets were used to evaluate the generalizability, one derived from a patient-grouped 70/30 split and another including all surgeries from the last available year. The achieved Area Under the Receiver Operating Characteristic curve on these test sets was 69.5% and 65.3%, respectively. Also, additional testing was implemented to simulate a real-world use case considering the weekly distribution of remote patient monitoring resources post-discharge. Compared to the random resource allocation, the selection of patients with respect to the outputs of the proposed model was proven beneficial, as it led to a higher number of high-risk patients receiving remote monitoring equipment.

Scaling-up digital follow-up care services: collaborative development and implementation of Remote Patient Monitoring pilot initiatives to increase access to follow-up care

Background

COVID-19 increased the demand for Remote Patient Monitoring (RPM) services as a rapid solution for safe patient follow-up in a lockdown context. Time and resource constraints resulted in unplanned scaled-up RPM pilot initiatives posing risks to the access and quality of care. Scalability and rapid implementation of RPM services require social change and active collaboration between stakeholders. Therefore, a participatory action research (PAR) approach is needed to support the collaborative development of the technological component while simultaneously implementing and evaluating the RPM service through critical action-reflection cycles.

Objective

This study aims to demonstrate how PAR can be used to guide the scalability design of RPM pilot initiatives and the implementation of RPM-based follow-up services.

Methods

Using a case study strategy, we described the PAR team's (nurses, physicians, developers, and researchers) activities within and across the four phases of the research process (problem definition, planning, action, and reflection). Team meetings were analyzed through content analysis and descriptive statistics. The PAR team selected ex-ante pilot initiatives to reflect upon features feedback and participatory level assessment. Pilot initiatives were investigated using semi-structured interviews transcribed and coded into themes following the principles of grounded theory and pilot meetings minutes and reports through content analysis. The PAR team used the MoSCoW prioritization method to define the set of features and descriptive statistics to reflect on the performance of the PAR approach.

Results

The approach involved two action-reflection cycles. From the 15 features identified, the team classified 11 as must-haves in the scaled-up version. The participation was similar among researchers (52.9%), developers (47.5%), and physicians (46.7%), who focused on suggesting and planning actions. Nurses with the lowest participation (5.8%) focused on knowledge sharing and generation. The top three meeting outcomes were: improved research and development system (35.0%), socio-technical-economic constraints characterization (25.2%), and understanding of end-user technology utilization (22.0%).

Conclusion

The scalability and implementation of RPM services must consider contextual factors, such as individuals' and organizations' interests and needs. The PAR approach supports simultaneously designing, developing, testing, and evaluating the RPM technological features, in a real-world context, with the participation of healthcare professionals, developers, and researchers.

Machine Learning Approaches for the Frailty Screening: A Narrative Review

Frailty characterizes a state of impairments that increases the risk of adverse health outcomes such as physical limitation, lower quality of life, and premature death. Frailty prevention, early screening, and management of potential existing conditions are essential and impact the elderly population positively and on society. Advanced machine learning (ML) processing methods are one of healthcare’s fastest developing scientific and technical areas. Although research studies are being conducted in a controlled environment, their translation into the real world (clinical setting, which is often dynamic) is challenging. This paper presents a narrative review of the procedures for the frailty screening applied to the innovative tools, focusing on indicators and ML approaches. It results in six selected studies. Support vector machine was the most often used ML method. These methods apparently can identify several risk factors to predict pre-frail or frailty. Even so, there are some limitations (e.g., quality data), but they have enormous potential to detect frailty early.

Developing and validating high-value patient digital follow-up services: a pilot study in cardiac surgery

Background

The existing digital healthcare solutions demand a service development approach that assesses needs, experience, and outcomes, to develop high-value digital healthcare services. The objective of this study was to develop a digital transformation of the patients’ follow-up service after cardiac surgery, based on a remote patient monitoring service that would respond to the real context challenges.

Methods

The study followed the Design Science Research methodology framework and incorporated concepts from the Lean startup method to start designing a minimal viable product (MVP) from the available resources. The service was implemented in a pilot study with 29 patients in 4 iterative develop-test-learn cycles, with the engagement of developers, researchers, clinical teams, and patients.

Results

Patients reported outcomes daily for 30 days after surgery through Internet-of-Things (IoT) devices and a mobile app. The service’s evaluation considered experience, feasibility, and effectiveness. It generated high satisfaction and high adherence among users, fewer readmissions, with an average of 7 ± 4.5 clinical actions per patient, primarily due to abnormal systolic blood pressure or wound-related issues.

Conclusions

We propose a 6-step methodology to design and validate a high-value digital health care service based on collaborative learning, real-time development, iterative testing, and value assessment.

Which outcomes are most important to measure in patients with COVID-19 and how and when should these be measured? Development of an international standard set of outcomes measures for clinical use in patients with COVID-19: a report of the International Consortium for Health Outcomes Measurement (ICHOM) COVID-19 Working Group

OBJECTIVES

The COVID-19 pandemic has resulted in widespread morbidity and mortality with the consequences expected to be felt for many years. Significant variation exists in the care even of similar patients with COVID-19, including treatment practices within and between institutions. Outcome measures vary among clinical trials on the same therapies. Understanding which therapies are of most value is not possible unless consensus can be reached on which outcomes are most important to measure. Furthermore, consensus on the most important outcomes may enable patients to monitor and track their care, and may help providers to improve the care they offer through quality improvement. To develop a standardised minimum set of outcomes for clinical care, the International Consortium for Health Outcomes Measurement (ICHOM) assembled a working group (WG) of 28 volunteers, including health professionals, patients and patient representatives.

DESIGN

A list of outcomes important to patients and professionals was generated from a systematic review of the published literature using the MEDLINE database, from review of outcomes being measured in ongoing clinical trials, from a survey distributed to patients and patient networks, and from previously published ICHOM standard sets in other disease areas. Using an online-modified Delphi process, the WG selected outcomes of greatest importance.

RESULTS

The outcomes considered by the WG to be most important were selected and categorised into five domains: (1) functional status and quality of life, (2) mental functioning, (3) social functioning, (4) clinical outcomes and (5) symptoms. The WG identified demographic and clinical variables for use as case-mix risk adjusters. These included baseline demographics, clinical factors and treatment-related factors.

CONCLUSION

Implementation of these consensus recommendations could help institutions to monitor, compare and improve the quality and delivery of care to patients with COVID-19. Their consistent definition and collection could also broaden the implementation of more patient-centric clinical outcomes research.

Markers for upper limb dysfunction in Amyotrophic Lateral Sclerosis using analysis of typing activity

OBJECTIVE

Assistive devices based on keyboard access support communication and control tools for patients with Amyotrophic Lateral Sclerosis (ALS). The aim of this work was to explore movement activity in the use of keyboards and identify markers for upper limb (UL) dysfunction.

METHODS

We present a longitudinal study including 19 ALS patients, followed for 2-20 months. Typing activity was recorded with an accelerometer placed on the posterior part of patients' index finger. Participants performed the same 10-word typing task (2-6 assessments). Time and acceleration during keystroke were the main outcomes of this study. Patients were compared with 20 healthy subjects and 6 patients with other neuromuscular disorders.

RESULTS

During disease progression, mean time in holding down a key increased and was longer than in control subjects. Acceleration at key press and key release decreased with progression of UL dysfunction. Delay between tapping and pressing down each key increased with UL dysfunction.

CONCLUSIONS

Delay in pressing and releasing keys are markers of UL dysfunction in ALS. The decrease in the acceleration of movements related to keystroke can contribute to monitor disease progression.

SIGNIFICANCE

Typing activity can be explored to access remotely and continuously to ALS progression by patients who use assistive communication devices.