An evaluation of the feasibility to use digital solutions to support the provision of healthcare in the NHS

Naicker, Kevin (2021) An evaluation of the feasibility to use digital solutions to support the provision of healthcare in the NHS. (PhD thesis), Kingston University, .


In England, there is an increase in prevalence of patients with long-term conditions (LTC) with approximately 26 million people having at least one LTC, utilising 70% of the total health and social budget. The Five Year Forward View and the NHS Long Term plan have identified that digital healthcare is a cost-effective technology that has the potential to integrate systems, improve efficiencies and have better clinical and social outcomes for patients. Whilst the successful implementation of digital healthcare in the NHS has been challenging, the severe acute respiratory syndrome coronavirus-2 (COVID-19) has been a major driver to support the rapid implementation of digital technology to safely maintain services and public health. This thesis aimed to evaluate the feasibility of using digital solutions to support the provision of healthcare in the borough of Croydon. Overall, a mixed-method approach consisting of quantitative and qualitative techniques was used to investigate the usability and acceptability of three technology enabled care solutions: a telehealth monitoring system in care home with residents with dementia, a follow up telephonic solution post-discharge and a digital stethoscope to screen for congenital murmurs. The study that assessed the potential role of telehealth system for monitoring residents with dementia living in a care home took place over three six-month phases [control phase (CP), active monitoring phase (AMP) and active monitoring with text alerts (AMTAP)]. The solution involved carers recording vital signs and completing health assessment questions. In AMTAP, the early warning triage system generated text alerts when abnormal responses or vital signs were detected. Twenty-seven residents participated during the CP and AMP whilst only fourteen residents participated during AMTAP. The quantitative section of this study calculated the frequency effect of the telehealth solution on the general practitioner visits (GPV), antibiotic prescribing (AP), emergency department (ED) visits and hospital inpatient (IP) events whilst the qualitative section of this study explored multidisciplinary healthcare professionals (HCPs) awareness and experiences of the telehealth system and reported on their perceptions pre and post implementation. The digital healthcare solution resulted in more frequent monitoring of residents’ vital signs (weekly vs monthly) resulting in an increase in frequency of GP (p=0.009) and AP (p<0.001) and a decrease in ED, IP and duration of an admission. Multidisciplinary HCPs were interviewed pre-AMP (n=33) and post-AMP (n=28). HCPs working with the digital solution reported increase knowledge and awareness after using the digital solution whereas HCPs who did not have direct contact reported a lack of understanding and awareness. Barriers to implementation included: lack of resources, training, inadequate staffing, equipment failure and poor system implementation. Indirect beneficial outcomes included: improved medication changes for residents and improved professional satisfaction and therefore HCPs wanted to use digital healthcare in the future. The second study evaluated a post discharge telephonic intervention by two nurses in the role of discharge advocates (DA) to ensure that the patients received the required post discharge care. The intervention was part of a funded project aiming to design a readmission prediction risk score system (OPTIMAL) to identify patients in need of a post discharge intervention to prevent a readmission within 30 days of discharge. The DA recruited eligible patients into the intervention (n=740) and control arms (n=730). It was determined that a sample size of 265 patients’ evaluations per a group (total 530, [intervention (n=265) and control (n=265)]) would be needed to determine patients’ satisfaction levels at 95% confidence interval. The sample extracted for evaluation had a statistically significant (p=0.001) higher mean OPTIMAL 30-day readmission risk score than the study arms. The OPTIMAL risk score of admission was reflective of the actual rate of re-admission with an average predictive score of 16.28% and actual 30-day readmission rate of 15.12% (n=223) for the whole sample. In the sample tested, the OPTIMAL predictive percentage readmission scores were 20.27% and 20.91% in the evaluation and control arms compared to (n=61, 23%) actual 30 days readmission rate across both arms. There was no significant difference in 30-day readmission rate between the study control and intervention arms. However, the percentage actual readmission rate was statistically significantly lower for the intervention evaluation group (9.4%) as compared to the control evaluation group (13.5%) (p<0.001). The DA call was perceived positively by patients as they felt that the DA understood their health status (81.1%) and was quite helpful (82.2%). Based on the findings, perhaps the intervention should not have been offered to all patients but to those that will most benefit from it, so targeted intervention based on the OPTIMAL readmission risk score, as the intervention did statistically significantly lower readmission rate for those patients. Lastly, an investigation was conducted to evaluate both the feasibility of an electronic stethoscope for the detection of congenital heart murmurs and its usability and associated software amongst clinicians with different levels of experience. Paediatric patients (n=72) with suspected murmurs attending a Paediatric Cardiologist led outpatient clinic and forty age-matched participants with no discernible murmurs consented to 30-second heart sound Consultant Paediatrician verified recordings using a 3M™ Littmann® Electronic Stethoscope Model 3200 to create a heart sound database. MATLAB (The MathWorks Inc., USA) was used to create sound waveforms and the 3M™ Littmann Steth Assist Heart and Lung Sound Visualization Software® was used to record and playback heart sounds. For the recordings without murmurs (n=6), the waveform between heart sounds appeared regular and smooth whereas in the recordings with murmurs (n=6) the waveforms between heart sounds had varying frequency with some higher frequency components. This was apparent with grade 3 and above murmurs, but this was less apparent in grade one and two murmurs, hence not proving a useful tool for screening. Clinicians (n=38) retrospective assessment of heart sounds played through a Bluetooth speaker resulted in system sensitivity of 77% and a specificity of 69%. The ability to distinguish between a normal and abnormal heart sound when listening to the audio samples was related to the experience of the clinician, with consultants scoring the highest. Unfortunately, clinicians (61%) reported that they would not be comfortable to confirm a diagnosis remotely using the system. All clinicians were able to acquire heart sounds using the electronic stethoscope. However, only Consultants (n=11), Senior House Officers (SHOs) and Registrars (n=4) ranked the electronic stethoscope with an acceptable System Usability Scale (SUS) score (≥70). Clinicians identified advantages for the system, with potential use as an educational tool and for the retrospective review of heart sounds. The three studies above evaluated the feasibility of using digital solutions to support the provision of healthcare. The evaluation has proven that digital solutions have the potential to support HCPs in healthcare provision, but the technology, organisation and patients need to be considered so that the proper ‘digital fit’ can be achieved to ensure that digital solutions are adopted by HCPs and that patients can experience the full benefits from them for both their healthcare and social outcomes.

Actions (Repository Editors)

Item Control Page Item Control Page