Limits of data anonymity: lack of public awareness risks trust in health system activities

A systematic review of networks for prognostic prediction of health outcomes and diagnostic prediction of health conditions within Electronic Health Records

BACKGROUND AND OBJECTIVE

Using graph theory, Electronic Health Records (EHRs) can be represented graphically to exploit the relational dependencies of the multiple information formats to improve Machine Learning (ML) prediction models. In this systematic qualitative review, we explore the question: How are graphs used on EHRs, to predict diagnosis and health outcomes?

METHODOLOGY

The search strategy identified studies that used patient-level graph representations of EHRs to utilise ML to predict health outcomes and diagnoses. We conducted our search on MEDLINE, Web of Science and Scopus.

RESULTS

832 studies were identified by the search strategy, of which 27 studies were selected for data extraction. Following data extraction, 18 studies used ML with patient-level graph-based representations of EHRs to predict health outcomes and diagnoses. Models ranged from traditional ML to neural network-based models. MIMIC-III was the most used dataset (n = 6, where n is the number of occurrences), followed by National Health Insurance Research Database (NHIRD) (n = 4) and eICU Collaborative Research Database (eICU) (n = 4). The most predicted health outcomes were mortality (n = 9; 21%), hospital readmission (n = 9; 21%), and treatment success (n = 4; 9%). Model performances ranged across outcomes, mortality prediction (Area Under the Receiver Operating Characteristic (AUROC): 72.1 - 91.6; Area Under Precision-Recall Curve (AUPRC): 34.8 - 81.3) and readmission prediction (AUROC: 63.7 - 85.8; AUPRC 39.86 - 84.7). Only one paper had a low Risk of Bias (RoB) that applied to our research question (4%).

CONCLUSION

Graph-based representations using EHRs, for individual health outcomes and diagnoses requires further research before we can see the results applied clinically. The use of graph representations appears to improve EHR representation and predictive performance compared to baseline ML methods in multiple fields of medicine.

Video Recording Patients for Direct Care Purposes: Systematic Review and Narrative Synthesis of International Empirical Studies and UK Professional Guidance

BACKGROUND

Video recordings of patients may offer advantages to supplement patient assessment and clinical decision-making. However, little is known about the practice of video recording patients for direct care purposes.

OBJECTIVE

We aimed to synthesize empirical studies published internationally to explore the extent to which video recording patients is acceptable and effective in supporting direct care and, for the United Kingdom, to summarize the relevant guidance of professional and regulatory bodies.

METHODS

Five electronic databases (MEDLINE, Embase, APA PsycINFO, CENTRAL, and HMIC) were searched from 2012 to 2022. Eligible studies evaluated an intervention involving video recording of adult patients (≥18 years) to support diagnosis, care, or treatment. All study designs and countries of publication were included. Websites of UK professional and regulatory bodies were searched to identify relevant guidance. The acceptability of video recording patients was evaluated using study recruitment and retention rates and a framework synthesis of patients' and clinical staff's perspectives based on the Theoretical Framework of Acceptability by Sekhon. Clinically relevant measures of impact were extracted and tabulated according to the study design. The framework approach was used to synthesize the reported ethico-legal considerations, and recommendations of professional and regulatory bodies were extracted and tabulated.

RESULTS

Of the 14,221 abstracts screened, 27 studies met the inclusion criteria. Overall, 13 guidance documents were retrieved, of which 7 were retained for review. The views of patients and clinical staff (16 studies) were predominantly positive, although concerns were expressed about privacy, technical considerations, and integrating video recording into clinical workflows; some patients were anxious about their physical appearance. The mean recruitment rate was 68.2% (SD 22.5%; range 34.2%-100%; 12 studies), and the mean retention rate was 73.3% (SD 28.6%; range 16.7%-100%; 17 studies). Regarding effectiveness (10 studies), patients and clinical staff considered video recordings to be valuable in supporting assessment, care, and treatment; in promoting patient engagement; and in enhancing communication and recall of information. Observational studies (n=5) favored video recording, but randomized controlled trials (n=5) did not demonstrate that video recording was superior to the controls. UK guidelines are consistent in their recommendations around consent, privacy, and storage of recordings but lack detailed guidance on how to operationalize these recommendations in clinical practice.

CONCLUSIONS

Video recording patients for direct care purposes appears to be acceptable, despite concerns about privacy, technical considerations, and how to incorporate recording into clinical workflows. Methodological quality prevents firm conclusions from being drawn; therefore, pragmatic trials (particularly in older adult care and the movement disorders field) should evaluate the impact of video recording on diagnosis, treatment monitoring, patient-clinician communication, and patient safety. Professional and regulatory documents should signpost to practical guidance on the implementation of video recording in routine practice.

TRIAL REGISTRATION

PROSPERO CRD42022331825: https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=331825.

Evidence-based guiding principles to build public trust in personal data use in health systems

Objective

Public trust in health systems is pivotal for their effective and efficient functioning. In particular, public trust is essential for personal data use, as demonstrated in debates in many countries, for example, about whether data from COVID-19 contact tracing apps should be pooled or remain on individuals' smartphones. Low levels of public trust pose a risk not only to health system legitimacy but can also harm population health.

Methods

Synthesising our previous qualitative and theoretical research in the English National Health Service which enabled us to conceptualise the nature of public trust in health systems, we present guiding principles designed to rebuild public trust, if lost, and to maintain high levels of public trust in personal data use within the health system, if not.

Results

To build public trust, health system actors need to not rush trust building; engage with the public; keep the public safe; offer autonomy to the public; plan for diverse trust relationships; recognise that trust is shaped by both emotion and rational thought; represent the public interest; and work towards realising a net benefit for the health system and the public.

Conclusions

Beyond policymakers and government officials, the guiding principles address a wide range of actors within health systems so that they can work collectively to build public trust. The guiding principles can be used to inform policymaking in health and health care and to analyse the performance of different governments to see if those governments that operate in greater conformity with the guiding principles perform better.