PROACT Design and Development
PROACT was a password-protected mobile application developed with ease of use and security in mind. It had a help function that contains a video on how to use the application, and a link to a technical support team for help with password resets and any technical issues. The application allowed patients to record messages (audio/video) and for these to be streamed to a server and made instantly viewable to the patient’s onsite clinical team who were then able to respond appropriately. Patients received a return video/audio message in a chronologically arranged ‘message list’ that they were able to view at will.
After a delay of 32 days, the patient’s message was sent through to a small group of independent analysts. This delay was to ensure that any adverse event information was discussed between the patient and their onsite clinical team and that all relevant updates were entered into the clinical study database for reporting as per the clinical study protocol. The analysts removed personally-identifiable information while capturing the patient message which was then relayed to the sponsor’s study team for further analysis (see Fig. 1).
Patients and medical team members had the ability to trigger deletion of any of their own videos recorded and uploaded to the PROACT system by pressing the ‘delete’ button within the application. This triggered total deletion of that video and associated information (but not analyzed/anonymized data).
Data Storage and Protection
Rigorous controls were implemented to ensure data protection and patient confidentiality. All connections to the PROACT application were secured via https using a digital certificate. Within the PROACT environment, encrypted data were hosted by the Safe Harbor Certified Provider (Rackspace) in accordance with national and international standards on data protection [8–10]. User access to decrypted data was granted at different levels dependent on role (patient, investigator, administrator, sponsor).
Pilot Study Objectives
The pilot study was conducted using a clinical protocol, which had been developed, written and submitted to an Ethics Committee for approval. The objectives were to investigate the uptake and use of PROACT when made available to patients undergoing treatment for cancer in a clinical trial setting and to characterize the information that patients volunteered. The principal focus in this respect was to understand the technicalities and practicalities of the application. The study was also designed to preliminarily assess the potential benefit of receiving direct feedback from patients for the sponsor, and to explore analysis methods that would convert the information into an interpretable dataset while also protecting patient privacy. The objectives were developed with the long-term aim of PROACT being used in a way that fits with patients’ day-to-day lives.
Study Design, Participants, and Data Collection
Patients aged ≥18 years recruited to Phase I AstraZeneca (the sponsor) oncology studies at the Sarah Cannon Research Institute, London, UK were eligible to participate. Patients were also required to have a “smart” phone or tablet computer that they were willing to use (without reimbursement) for the transmission of audio/video recordings. Provision of informed written consent was also required.
Following enrolment, patients downloaded the PROACT application to their “smart” phone or computer tablet and were instructed on its technical use. Beyond technical advice, patients were provided minimal guidance on use of the system. This allowed the sponsor to gather insight into the nature and level of support that PROACT use may require in the future, albeit from a small sample of patients already comfortable with using mobile devices. Patients’ messages were created, transmitted and handled as described previously.
Data evaluation focused on identifying contextual key words and topics, using manual text analysis of transcribed data. Patient-identifiable data was removed while capturing the message a patient wanted to convey in an interpretable dataset. To enable this, a hierarchical dictionary and flexible coding database were created to allow for aggregation of data, while linking back to the patient’s own words (see Fig. 2). The database comprised a series of entries, each comprising a coded patient identity, date of communication, data category (e.g., safety, efficacy, convenience), data group (e.g., system affected), topic (e.g., specific treatment burden), the patient quote, status of the communicated event/issue, and response/outcome. Also, a simple color scale was included to visualize the impact of the reported topic upon that patient’s wellbeing or daily activities [green, minimal impact on quality of life (also used to describe if an event had resolved or was improving); amber, moderate impact on quality of life, able to generally continue daily living activities with some limitation (also used to describe an ongoing event); red, significant impact on quality of life (also used to describe a new or worsening event)]. The nature and content of the messages, therefore, drove the development of the dictionary and coding database in an iterative process. These outputs were then assessed to see whether they were able to capture the core elements of the patient contribution in a format that would be useful for contributing to drug programme decisions.
The study received Ethical Research Committee approval (with positive comments regarding steps taken to ensure patient confidentiality and data protection) and it was conducted in accordance with applicable regulatory requirements and other relevant guidelines, such as patient data privacy requirements previously described [8–10].
In parallel with the pilot study, we conducted an interactive meeting with patient and carer representatives of the National Cancer Research Institute’s Consumer Forum (in the UK). Participants in a patient focus group were invited to provide insights into the needs of cancer patients during treatment with a focus on communications in the clinical trial environment. Open forum discussions initially concentrated on three main areas: (1) current clinical practice, (2) patient needs, and (3) carer needs and contributions. The first half of the discussion focused on people’s experiences and opinions before the PROACT concept was unveiled (to mitigate against any bias that proposing PROACT may have introduced). The second half of the session focused on discussion and input into PROACT itself: the concept, approach, use, and guidance. Information gathered was used to assess whether any improvements could be made to the use of PROACT during the pilot study, and as well as aiding development of an enhanced version if further evaluation of the application was deemed appropriate.