Empowerment of care home staff through effective collaboration with healthcare.

Integrative local health delivery models in the UK, under the framework of Enhanced Health in Care Homes (EHICH), have been developed to improve joint working between health and social care to benefit the patient. Despite this drive toward health and social care integration, research on the barriers, facilitators, and impact of partnership working on role of care home staff is underdeveloped. This study set out to explore views on how closer working between health and social care can impact on the role of care home staff and any barriers to effective integration. Staff from 25 care homes and GPs from their partnered practices were interviewed to explore the impact of the partnership. Homes receiving regular visits from the same health professional found the relationship between the two sectors had benefitted both residents and staff. The development of trusting relationships, access to support and information, and recognition and respect were all seen as facilitating the partnership and enhancing patient care. Regular and effective interactions with health-care professionals were key and had the potential to empower and increase confidence of care home staff in their role around health care. Factors negatively impacting on strength of relationship such as visits by inconsistent professional and high turnover of care home staff were a barrier to successful partnerships. Experiences of poor interactions with those from health-care services where there was an absence of a trusting relationship were disempowering to care home staff and remain a barrier to effective wider health and social care collaboration.

ChemPix: automated recognition of hand-drawn hydrocarbon structures using deep learning.

Inputting molecules into chemistry software, such as quantum chemistry packages, currently requires domain expertise, expensive software and/or cumbersome procedures. Leveraging recent breakthroughs in machine learning, we develop ChemPix: an offline, hand-drawn hydrocarbon structure recognition tool designed to remove these barriers. A neural image captioning approach consisting of a convolutional neural network (CNN) encoder and a long short-term memory (LSTM) decoder learned a mapping from photographs of hand-drawn hydrocarbon structures to machine-readable SMILES representations. We generated a large auxiliary training dataset, based on RDKit molecular images, by combining image augmentation, image degradation and background addition. Additionally, a small dataset of ∼600 hand-drawn hydrocarbon chemical structures was crowd-sourced using a phone web application. These datasets were used to train the image-to-SMILES neural network with the goal of maximizing the hand-drawn hydrocarbon recognition accuracy. By forming a committee of the trained neural networks where each network casts one vote for the predicted molecule, we achieved a nearly 10 percentage point improvement of the molecule recognition accuracy and were able to assign a confidence value for the prediction based on the number of agreeing votes. The ensemble model achieved an accuracy of 76% on hand-drawn hydrocarbons, increasing to 86% if the top 3 predictions were considered.