Can we identify the prevalence of perinatal mental health using routinely collected health data?: A review of publicly available perinatal mental health data sources in England.

Introduction: Perinatal mental health (PMH) conditions affect around one in four women, and may be even higher in women from some ethnic minority groups and those living in low socioeconomic circumstances. Poor PMH causes significant distress and can have lifelong adverse impacts for some children. In England, current prevalence rates are estimated using mental health data of the general population and do not take sociodemographic variance of geographical areas into account. Services cannot plan their capacity and ensure appropriate and timely support using these estimates. Our aim was to see if PMH prevalence rates could be identified using existing publicly available sources of routine health data. Methods: A review of data sources was completed by searching NHS Digital (now NHS England), Public Health England and other national PMH resources, performing keyword searches online, and research team knowledge of the field. The sources were screened for routine data that could be used to produce prevalence of PMH conditions by sociodemographic variation. Included sources were reviewed for their utility in accessibility, data relevance and technical specification relating to PMH and sociodemographic data items. Results: We found a PMH data 'blind spot' with significant inadequacies in the utility of all identified data sources, making it impossible to provide information on the prevalence of PMH in England and understand variation by sociodemographic differences. Conclusions: To enhance the utility of publicly available routine data to provide PMH prevalence rates requires improved mandatory PMH data capture in universal services, available publicly via one platform and including assessment outcomes and sociodemographic data.

Liaison psychiatry-measurement and evaluation of service types, referral patterns and outcomes (LP-MAESTRO): a protocol.

INTRODUCTION: We describe the protocol for a project that will use linkage of routinely collected NHS data to answer a question about the nature and effectiveness of liaison psychiatry services in acute hospitals in England. METHODS AND ANALYSIS: The project will use three data sources: (1) Hospital Episode Statistics (HES), a database controlled by NHS Digital that contains patient data relating to emergency department (ED), inpatient and outpatient episodes at hospitals in England; (2) ResearchOne, a research database controlled by The Phoenix Partnership (TPP) that contains patient data relating to primary care provided by organisations using the SystmOne clinical information system and (3) clinical databases controlled by mental health trusts that contain patient data relating to care provided by liaison psychiatry services. We will link patient data from these sources to construct care pathways for patients who have been admitted to a particular hospital and determine those patients who have been seen by a liaison psychiatry service during their admission.Patient care pathways will form the basis of a matched cohort design to test the effectiveness of liaison intervention. We will combine healthcare utilisation within care pathways using cost figures from national databases. We will compare the cost of each care pathway and the impact of a broad set of health-related outcomes to obtain preliminary estimates of cost-effectiveness for liaison psychiatry services. We will carry out an exploratory incremental cost-effectiveness analysis from a whole system perspective. ETHICS AND DISSEMINATION: Individual patient consent will not be feasible for this study. Favourable ethical opinion has been obtained from the NHS Research Ethics Committee (North of Scotland) (REF: 16/NS/0025) for Work Stream 2 (phase 1) of the Liaison psychiatry-measurement and evaluation of service types, referral patterns and outcomes study. The Confidentiality Advisory Group at the Health Research Authority determined that Section 251 approval under Regulation 5 of the Health Service (Control of Patient Information) Regulations 2002 was not required for the study 'on the basis that there is no disclosure of patient identifiable data without consent' (REF: 16/CAG/0037).Results of the study will be published in academic journals in health services research and mental health. Details of the study methodology will also be published in an academic journal. Discussion papers will be authored for health service commissioners.

Machine-learning based identification of undiagnosed dementia in primary care: a feasibility study.

BACKGROUND: Up to half of patients with dementia may not receive a formal diagnosis, limiting access to appropriate services. It is hypothesised that it may be possible to identify undiagnosed dementia from a profile of symptoms recorded in routine clinical practice. AIM: The aim of this study is to develop a machine learning-based model that could be used in general practice to detect dementia from routinely collected NHS data. The model would be a useful tool for identifying people who may be living with dementia but have not been formally diagnosed. DESIGN & SETTING: The study involved a case-control design and analysis of primary care data routinely collected over a 2-year period. Dementia diagnosed during the study period was compared to no diagnosis of dementia during the same period using pseudonymised routinely collected primary care clinical data. METHOD: Routinely collected Read-encoded data were obtained from 18 consenting GP surgeries across Devon, for 26 483 patients aged >65 years. The authors determined Read codes assigned to patients that may contribute to dementia risk. These codes were used as features to train a machine-learning classification model to identify patients that may have underlying dementia. RESULTS: The model obtained sensitivity and specificity values of 84.47% and 86.67%, respectively. CONCLUSION: The results show that routinely collected primary care data may be used to identify undiagnosed dementia. The methodology is promising and, if successfully developed and deployed, may help to increase dementia diagnosis in primary care.