Understanding the patient experience of healthcare is an essential element in providing patient-centric care. It is also a fundamental pillar of ensuring care quality. However, consistent application and utilisation of patient feedback is often inhibited by the unstructured nature of the data. For this reason a manual review is needed to extract insights. Our healthcare industry has invested heavily in recording, analysing and monitoring clinical outcomes data. However, relatively little emphasis has been placed on innovation in assessing patient experience to improve care delivery.
This research addresses how can healthcare organisations can leverage real-time patient experience data to improve care delivery?
In order to contribute a clear body of knowledge in response, this research ascertains the root cause of limited use of patient experience data; tests how novel analytic techniques can enhance data utliltiy; and explores how free-text experience data can be used to improve organisational responsiveness to patient feedback in real-time.
First of all, this research systematically appraises the evidence on which designs of effective real-time feedback (RTF) systems are based. Four strategic themes are identified – namely, capacity/resource, usage, interoperability, and impact. These themes heavily influence the readiness and digital maturity of RTF systems, and it is recommended that healthcare services address the four themes individually in the design and implementation of RTF systems. Next, process mapping is employed to identify barriers and enablers of RTF and their effect on staff enagagement. Through interviews with healthcare stakeholders, opportunities for RTF to drive localised improvement are then identified.
Secondly, this research leverages text analytics, namely Natural Language Processing (NLP), with supervised learning methods to analysed patient experience free-text data. Text (or theme) classification and sentiment analysis of the Friends and Family Test (FFT) free-text data, to a large extent, confirms the accuracy of the machine learning model. Furthermore, a co-design approach enables the FFT text analytics output to be displayed in a format in line with the interests of frontline staff and key stakeholders. A visualisation tool is deployed in combination with improvement models tested in various healthcare settings, selected by the most common themes and associated sentiment. This revitalised the benefits of hearing from patients in their own words by holding narrative data to the same standard of scientific rigour already applied to quantitative data. In addition, it reinforced staff to take action and address the most important patient concerns in a timely manner.
Individually, these findings convey a series of policy recommendations, while cumulatively, they showcase the possibilities for a more patient-centric healthcare service.