Mapping to the results of the survey
The survey results provided an insight to various treatments undertaken by the advanced prostate cancer patients. The analysis of the survey results suggested a self-care pattern among advanced prostate cancer patients and showed a practice of self-reporting about the complementary and OTC medications. The survey results showed that the self-care interventions undertaken by the patients are an integral part of understanding their journey. The patient reported data should be integrated with appropriate clinical treatment interventions undertaken and such integration will enable a holistic understanding of advanced prostate cancer patient’s clinical journey.
Rationale – arbitrary
A patient journey is commonly represented as a work flow of patients from various health service providers. Our research into existing methods and tools shows that the existing systems mainly focus on work flow and data flow arising out of the work flow related to a particular health care setting.10
These systems are appropriate for enhancing business processes involved in providing care to the prostate cancer patients. However, clinical decisions can be made using a more granular level of patient data that integrates their clinical treatment information with self-administered intervention. The information presented to the decision makers including carers as well as patients can be enhanced at a finer level of disease-specific details. In the context of self-management of the disease conditions, the patient data should be aggregated and presented to empower the patients. A self-management solution for advanced prostate cancer patients can provide them with a better understanding of their disease progression. There is a need for a new perspective towards modelling the patient journey. A patient’s journey can be described in various health states. The health states can be represented by specific stages of advanced prostate cancer progression. The journey of men diagnosed with prostate cancer can be described in various health states (HS) such as localised prostate cancer, locally advanced prostate cancer and other advanced stages of the disease. Each health state can be characterised by specific clinical treatments, health outcomes achieved and any prostate cancer-specific side effects including mental health disorders such as anxiety and depression. Our research about existing online information resources for self-managed care of advanced prostate cancer has found that the information is mainly static and is of generic nature. The prostate cancer survivors need to process vast amount of generic information themselves and they frequently have to depend on their care providers for their personalised information. There is no clear evidence to suggest these existing methods and ways of distributing information to prostate cancer survivors are effective from health outcomes and health economics perspectives. Therefore, there is a need for information solutions that are more adaptable and personalised to empower prostate cancer survivors to take action for managing their own health especially during their rehabilitation. The proposed solution will add value to Prostate cancer survivors by providing personalised information to take actions for their self-managed care especially during their rehabilitation. Based on the findings of our survey, a research prototype was developed to describe sharing of complementary medication outcomes among prostate cancer patients. The purpose of the research prototype is to demonstrate efficient use of an online Personal Health Record (PHR) solution and self-reported data by the patients. It is assumed that the online PHR solution will allow patients to record data into their clinical record. The proposed system retrieves the clinical records from the various source systems. The unique contribution of our system is the collaborative access to self-care solutions through a CBR engine that retrieves the most similar problems-solution-outcome triple from the most similar patient record. The design of a case to be represented and stored in the CBR engine was given a lot of consideration. A case that models specific features or attributes of a patient’s health state can be stored in different representational formats.12
The case features were chosen using the data items that are available or can be obtained. In the PHR environment, a patient can record data about their disease stage, complementary medication and health outcomes. The data structures for the features have to be available or be capable of being types that are available.16
The features were chosen to facilitate an effective and efficient retrieval of cases stored in the case repository. The case-base was built using features: (i) current disease stage; (ii) complementary medication; (iii) outcome; (iv) outcome method. The case-base was developed using MySQL and the interface was developed using motion charts in Java Script. Figure 1 shows the interface of our research prototype.
Proposed solution prototype with a CBR engine
The computation of case similarity is a key aspect in CBR. The similarity is computed using Manhattan or city block distance measure.17
The similarity between two patient cases can be measured on an absolute scale. Manhattan or city-block distance measure is based on the absolute values of the case features. The similarity measure is derived from the method of computing the walking distance between two points in a city like New York’s Manhattan district where each component is the number of blocks in north-south and east-west directions. The formulation for the similarity between two patient cases is modelled as below:
Where dij is the distance between data points in the case-base related to patient case. Each data point corresponds to the value of the features. The similarity computation determines the most similar case and the CBR engine retrieves it for sharing it with end-user. The end-user of our proposed system could be a patient searching for outcomes of complementary medication reported by patients on the same disease age. The cases were matched using the feature values of current disease stage and complementary medication. The disease stage and complementary medication are ordinal variables.
The features used in our approach are equally weighted for the simplicity of illustration of our prototype. The weights associated with the features can be determined by several methods and it is a separate research question. There can be various approaches for determining weights associated with the features in data aggregation problems especially in the context of case-based reasoning.
The described CBR approach facilitates the retrieval of the most similar patient case for sharing it with other patients at the similar disease stage. The visual data presentation is based on the changing values of prostate-specific antigen (PSA) levels over a given period in a line graph on logarithmic scale. The PSA test is a commonly used diagnostic measure for determining the disease progression of prostate cancer. The other key information such as current treatments, medications is displayed on the patient journey visual. The data visualisation assists patients in understanding their individual disease progression measured in PSA values. The data visualisation also assists patients in understanding the impact of treatments at specific disease stage. The care providers can also use the proposed visualisation in their clinical decision-making process.
The proposed approach could benefit patients in the self-management of their disease conditions. The proposed system may provide evidence of effectiveness of OTC complementary medication used by patients at similar disease stages. This evidence will improve patient education. The patients may learn about effective interventions by knowing about similar patient journeys. The patients may also use the proposed system to undertaken preventative actions in consultation with their primary clinicians. The proposed solution can also be integrated with existing social networks as patients or users with chronic conditions will be able to share their specific journey with other users of their choice. The proposed system is a research prototype with the potential to generate new evidence of effectiveness of specific interventions at each health state. A full scale commercial grade development and evaluation will be undertaken in future. The proposed approach presents a new perspective of modelling journey of patients with chronic diseases using Al techniques and data driven methods.