Proof of Concept of Remote Management of Chronic Inflammatory Airway Diseases for Patient Empowerment

Chronic rhinosinusitis (CRS) is a multifactorial disease characterized by persistent symptomatic inflammation of the mucosa of the nose and paranasal sinuses, with (CRSwNP) or without (CRSsNP) the presence of nasal polyps. It affects 5 to 12% of the general population. CRS is often associated with asthma, which has a prevalence of 4% in the general population, reaching 30%-70% among patients with CRS. The current standard clinical evaluation of patients for both diseases has two main components: a subjective one (self-assessment provided by the patient), based mainly on the PROMs (Patient-Reported Outcome Measures) questionnaire, and an objective one (formulated by the clinician). Questionnaires present accuracy and response rate problems that have been investigated in the literature, finding that short questionnaires, incentives, personalization of questionnaires as well as repeat sending strategies or telephone reminders have a beneficial impact on the quantity and quality of responses. Today there are many new channels provided by technology. Among them, AI chatbots have been used in a variety of healthcare domains such as medical consultations, disease diagnosis, mental health support and, more recently, risk communications for the COVID-19 pandemic, and can offer a better way to collect questionnaires. At the same time, the recent technical solution of new non-invasive techniques for RFID radio frequency identification devices allows subjective reports to be accompanied with objective reports. The PRECISION project aims to evaluate systems for home monitoring of chronic rhinosinusitis (CRS) and asthma, two highly prevalent chronic diseases. The frequent association between the two pathological conditions is a further argument in support of the rationality of a common approach. As regards the collection of PROMs, three acquisition channels will be compared: i) AI Chatbots; ii) PhoneBot; iii) Mobile application. Data will be analyzed in relation to patient profiles to define the quality and quality of response. Regarding objective evaluation, the project will investigate the efficiency of objective remote airflow measurements for both upper (CRS) and lower (asthma) airways using dedicated non-invasive systems based on RFID technology.

The primary objective of the study is to create a Proof of Concept for the remote acquisition of subjective/objective data for patients with Asthma and CRS. In the current standard clinical assessment of patients with chronic diseases such as asthma and CRS there are two main components, one subjective (provided by the patient) and one objective (provided by the doctor). In particular, subjective data are acquired via standardized PROMs (such as SNOT-22 for CRS and ACT for asthma). The collection of objective data is obtained through clinical evaluations, such as endoscopic evaluation, olfactometry, radiological evaluation (with the definition of Lund-Mackey score), measurement of nasal flow (usually through PNIF or rhinomanometry), cytology, blood tests (in particular eosinophil count and IgE), FeNO (fractionated exhaled nitric oxide). Both types of data are needed simultaneously for the correct treatment and follow-up of the disease and, above all, the quality of the measurements is crucial. The primary objective of the project is to quantify how much the patient's experience in managing part of their illness remotely improves compared to a hospital visit. This will be quantitatively assessed through the administration of "PREMs" (Patient-Reported Experience Measures) questionnaires. In particular, in accordance with the state of the art, the investigators will evaluate patient satisfaction based on two questionnaires: * CSQ8 (Client Satisfaction Questionnaire): provides a quantitative evaluation of the treatment * PSSUQ (Post-Study System Usability Questionnaire): provides a quantitative assessment of the appropriateness of the technology used Finally, the investigastors will evaluate patients' satisfaction in being followed partly from home through an ad-hoc questionnaire. Secondary objective: Evaluation of the best channels for the administration of PROMs based on age group and cultural level. PROMs capture a person's perception of their health through questionnaires, have a decidedly preponderant role in the clinical evaluation of patients with CRS and asthma according to the new classification systems and are the fulcrum of the subjective part (e.g. SinoNasal Outcome Test (SNOT), Asthma Control Test (ACT), Chemosensory Complaint Score (CCS)). Given the time-consuming nature of completing PROMs and the ease with which they can be performed remotely, more and more PROMs are being delivered online (e.g., with website forms). The issue with such a modality in the real-world context is patient compliance over time, as the low morbidity of controlled disease may be associated with decreased motivation. Furthermore, patients may need clarification on specific questions from healthcare professionals, which contribute to the causes for which PROMs are still administered within many hospitals today. In this project three different channels will be evaluated: Social Network Chatbots: Chatbots are "natural language online human-computer dialogue systems." The technology behind chatbots today is quite complex and spans the fields of natural language processing, response generation and dialogue management. Our goal is to evaluate the use of chatbots to increase engagement and implement explanations of PROMs. The chatbot will also be used for medical remains, which helps improve follow-up adherence. By using existing social networks (e.g. Whatsapp, Telegram), patients do not need specific training as they are already used to using instant messaging applications. Phonebot is a phone system that handles incoming and outgoing phone calls, as well as an organization's internal communications. The investigators will evaluate the effectiveness of using telephone calls to complete the survey required for PROMs. The system evaluated uses synthetic voice to utter questions to patients and speech/tone recognition to capture responses. Mobile App: The availability of dedicated apps on smartphones and the possibilities offered by this technology (push notifications, personalized graphics and usability) is another possible channel for administering PROMs. The investigators will evaluate the effectiveness of this channel to acquire the responses of the PROMs. The channels will be provided to different patients with the aim of: * Understand which channel provides better compliance, such as persistence of data acquisition over time * Understand whether the channel shows a difference in the quality of data acquired through baseline data acquisition and clinical monitoring * Understanding which channel is most suitable for a specific patient profile, analyzing the correlation of the activity with the age and education of the individual * Evaluation of the patient empowerment provided by the use of the channel Techniques for acquiring objective airflow data using non-invasive techniques will also be evaluated. Breath monitoring is an essential tool in the early diagnosis of respiratory and cardiovascular diseases. In particular, respiratory rate is crucial for monitoring the evolution of respiratory disorders, such as Chronic Obstructive Pulmonary Disease (COPD), Chronic Rhinosinusitis (CRS), Asthma or the recent COVID-19 disease. Breathing is typically monitored within a spirometry test that measures respiratory flows and, consequently, estimates respiratory volumes and rates. However, it can only be performed in a hospital with the supervision of a provider and requires the patient to breathe in a controlled manner while wearing nasal or oral probes. Wearable and skin-friendly technologies may offer interesting alternatives to mitigate such discomfort. In particular, battery-free ultra-high frequency (UHF) radio frequency identification (RFID) devices could provide minimally invasive, easy-to-use and low-cost diagnostic procedures. Therefore, the project will consider techniques for the objective assessment of airflow, and in particular: * in asthma the forced expiratory volume in one second (FEV1) and the peak expiratory flow (PEF) * in CRS the nasal peak inspiratory flow (PNIF) appears more reliable than (anterior active) rhinomanometry (AAR) and acoustic rhinometry (AR) i) a commercially available portable spirometer will be used to measure PEF, FEV remotely, with an intuitive APP that assists patients in providing data, receiving it and forwarding it to the hospital ii) the same hardware can be used for measuring the PNIF, through an adaptation of the software and connectable nasal masks available on the market. As a result, a portable dual-function device suitable for remote assessment of upper and lower airway function will be available by reusing and adapting available technology. The expected results of this specific objective are: * definition of the patient's breathing pattern using available tools and reliability for the early diagnosis of anomalous events; * independent analysis of the performance of the nostrils and assessment of the reliability of obtaining an analogue of PNIF; * evaluation of the reliability of the instruments in the definition of equivalent spirometric parameters. Data management will take place according to the rules defined by the Health Information Portability and Accountability Act (HIPAA) and in Europe by the General Data Protection Regulation (GDPR). The remote approaches tested in this project will generate a huge amount of real data for analysis. A specific data management plan will be developed to define in advance the methods for collecting, managing and protecting clinical research data. The plan will define specific data collection objectives, data type and location, data access, roles and responsibilities, allowing the research team a thorough understanding of the requirements. A key role will be played by the REDCap (Research Electronic Data Capture) platform, which will be used to correctly collect and manage data. This is a secure website application, compliant with many standards such as FISMA and HIPAA, developed to manage data for clinical research. During the experiments, the data will be collected in 2 ways: one will be standard with acquisition of subjective information (PROM) and objective information (airflows, clinical and endoscopic findings) and subsequent data entry. The other mode will benefit from specially created pipelines and will allow data to be acquired directly via MHT (Mobile Health Technology) and, in the case of phonebots, directly via the cloud, and will therefore be totally automated. This will drastically increase the amount of data potentially acquired through this second modality, without logistical limitations and greater possibilities, working with big data, to build reliable predictive models for the diseases under investigation. The project will develop dedicated data pipelines to be automatically filled with PROMs and airflow data acquired via MHT into the Redcap database. The phonebot and chatbot will use cloud solutions such as Twilio (Communication APIs for SMS, Voice, Video \& Authentication n.d.) or Amazon Connect (which, in turn, is based on Twilio). These systems began as Private Branch Exchanges (i.e. business telephone systems that manage incoming and outgoing calls) and have evolved towards multi-channel communication systems. They now integrate telephony, SMS, chatbots with different Instant Messaging channels, real-time video streaming and email systems. Furthermore, they also integrate several related technologies such as artificial intelligence systems to manage conversation, speech synthesis systems, data flow and dialogue composition tools. The investigators will make use of such systems as convenient from different perspectives. First, due to their company policy, Whatsapp forces their customers to use resellers (like Twilio) to access Whatsapp Business APIs that otherwise cannot be reached directly by us. Secondly, such cloud systems have the ability to scale to millions of users in minutes through the use of cloud computing, which is very convenient as the project has the ambition to be easily replicated. Finally, solutions like Amazon Connect make use of cutting-edge Artificial Intelligence (natural language processing) solutions that are shared and then tested for effectiveness with the popular 'Amazon Alexa' product. Customized mobile applications will be developed to provide questionnaires to patients. These applications will be released on Android and iOS platforms. All these applications will be designed together with the project partners, to derive the specifications of the finite state machine and clarify the usage flows. Statistical plan The planning of the three clinical studies was done considering the desired statistical power, the short time for completion of the project (only 2 years) and the fact that the three clinical studies are 'de facto' also Proof of concept for a very innovative approach with the ambition of changing the daily management of 2 very common chronic diseases and achieving a turning point towards real patient empowerment. For clinical trial 1, to achieve a confidence level of 95% and a confidence interval of 15, the required sample size was estimated at 40. As regards clinical study 2,the investigators wanted to set up a study capable of detecting a 20% difference between the arms, for which, being a four-arm study, 12 patients for each arm would be sufficient. But also, since it is a four-arm study, some correction should be applied (even a Bonferroni correction) and, if applied on the alpha value, with six possible comparisons (not only of the treatment groups with the control group, but also between treatment groups) the alpha value would go from 0.05 to approximately 0.009, so the investigators considered it useful to increase each arm to 20 subjects (total number of subjects 80). Finally, for clinical trial 3, the investigators wanted to be able to detect 25% differences in airflow measurements between the standard in-office tests and the 2 home systems, for which a sample size of 15 patients is required. To evaluate statistically significant differences for the primary endpoints the investigators will use: For trial 1, where the primary endpoint is patient satisfaction R-Square, Chi-Square, Pearson test and Fisher exact test to evaluate satisfaction of patients in the treatment group compared to the control group. The alpha value will be set to 0.05. The analysis will be applied to the questionnaires (PREMs) mentioned previously, in order to quantify the improvement in terms of patient experience treated using MHT (spirometers and PROMs) compared to the current scenario which only involves periodic in-person visits to the hospital. For trial 2, to evaluate the primary endpoint which is the completion rate of the questionnaires in the four arms, the investigators will use ANOVA followed by t-tests to perform the 6 possible comparisons between the 6 groups. For ANOVA an alpha value of 0.05 will be used, while for single comparisons a Bonferroni adjustment will be applied, bringing the alpha to 0.009. The following figures of merit will be evaluated: * completion rate: how many questionnaires have been completed, drop-off rate; * quality of the answers: also carried out through the analysis of meta-information (e.g. "gold questions") and using the compilation of the questionnaires in the hospital as ground truth; * best channel for the administration of PROMs with respect to the patient's technological profile: Phonebot, Instant Messaging and dedicated Apps; * patient empowerment through questionnaires. For trial 3, the non-inferiority hypothesis will be tested by comparing remote measurements with standard ambulatory tests via paired t-tests. Alpha will be set to 0.05. The secondary endpoints will be evaluated with the same software (JMP from the SAS institute) with appropriate statistical tools. Finally, the aggregated data found will be the subject of scientific publications, enriching the general know-how on the topic and therefore allowing secondary statistical uses of the data.