Lessons from the Literature

2019; Volume 20, No 11, November

Lessons from the Literature

Under My Microscope

Dr. Tania Pastrana

IAHPC research advisor Dr. Tania Pastrana selects one article from recently published medical literature and describes why it is worthwhile.

Development and Validation of a Deep Learning Algorithm for Mortality Prediction in Selecting Patients with Dementia for Earlier Palliative Care Interventions

Liqin Wang, Long Sha, Joshua R. Lakin, Julie Bynum, David W. Bates, Pengyu Hong, Li Zhou. JAMA Netw Open 2019; 2(7): e196972.


Dementia is a progressive, life-limiting condition that causes a patient to have complex needs. Palliative care benefits the patient as well as formal and informal carers.1 The incidence of dementia increases with age, and the population is aging. Worldwide, 47 million persons are estimated to live with dementia in 2015, and the figure is expected to reach 75.63 million in 2030 and 135.46 million in 2050.2 But when is the right time for palliative care intervention?

Wang et al. develop an algorithm that uses patient demographics and clinical notes to predict mortality risk, as a proxy indicator for identifying patients with dementia who need palliative care.

They conducted a retrospective cohort study and review for a data set of 2,692 adult patients with dementia.

Mortality prediction models reached an area under the receiver operating characteristic curve of 0.978 for predicting death in six months, 0.956 for one year, and 0.943 for two years.


1. Hughes J C, Robinson L, Volicer L. Specialist palliative care in dementia. BMJ 2005; 330(7482): 57-58.
2. Alzheimer's Disease International. Policy Brief for G8 Heads of Government. The Global Impact of Dementia 2013-2050. London: Alzheimer's Disease International; 2013.

Importance: Early palliative care interventions drive high-value care but currently are underused. Health care professionals face challenges in identifying patients who may benefit from palliative care.

Objective: To develop a deep learning algorithm using longitudinal electronic health records to predict mortality risk as a proxy indicator for identifying patients with dementia who may benefit from palliative care.

Design, Setting, and Participants: In this retrospective cohort study, 6-month, 1-year, and 2-year mortality prediction models with recurrent neural networks used patient demographic information and topics generated from clinical notes within Partners HealthCare System, an integrated health care delivery system in Boston, Massachusetts. This study included 26 921 adult patients with dementia who visited the health care system from January 1, 2011, through December 31, 2017. The models were trained using a data set of 24 229 patients and validated using another data set of 2692 patients. Data were analyzed from September 18, 2018, to May 15, 2019.

Main Outcomes and Measures:The area under the receiver operating characteristic curve (AUC) for 6-month and 1- and 2-year mortality prediction models and the factors contributing to the predictions.

Results:The study cohort included 26 921 patients (16 263 women [60.4%]; mean [SD] age, 74.6 [13.5] years). For the 24 229 patients in the training data set, mean (SD) age was 74.8 (13.2) years and 14 632 (60.4%) were women. For the 2692 patients in the validation data set, mean (SD) age was 75.0 (12.6) years and 1631 (60.6%) were women. The 6-month model reached an AUC of 0.978 (95% CI, 0.977-0.978); the 1-year model, 0.956 (95% CI, 0.955-0.956); and the 2-year model, 0.943 (95% CI, 0.942-0.944). The top-ranked latent topics associated with 6-month and 1- and 2-year mortality in patients with dementia include palliative and end-of-life care, cognitive function, delirium, testing of cholesterol levels, cancer, pain, use of health care services, arthritis, nutritional status, skin care, family meeting, shock, respiratory failure, and swallowing function.

Conclusions and Relevance:A deep learning algorithm based on patient demographic information and longitudinal clinical notes appeared to show promising results in predicting mortality among patients with dementia in different time frames. Further research is necessary to determine the feasibility of applying this algorithm in clinical settings for identifying unmet palliative care needs earlier.

Do you have any questions regarding Under My Microscope?

Contact Tania Pastrana

Media Watch: GPs & Quality Care

Some initiatives gaining attention on best practices for general practitioners and their key role in palliative care are described in the articles below, selected from Barry R. Ashpole’s weekly report, Media Watch.

End-of-life care in rural general practice: How best to support commitment and meet challenges?

BMC Palliative Care | Online – 25 June 2019 – This study explored the perspectives of GPs and other stakeholders on rural GPs’ involvement and challenges in providing palliative and end-of-life care in regional Australia. The rural GPs’ central role in end-of-life care (EoLC) was recognized, but multiple challenges were identified. Some more pronounced than in urban settings included resource limitations and lack of training. Inappropriate payment models discouraged GPs’ involvement in some aspects of EoLC, such as case conferences and home visits. GPs in rural/regional communities often reported closer doctor-patient relationships and better care integration and collaboration. These positive aspects of care could be further developed to enhance service provision. This study highlights the importance of regular interactions with other professionals and patients in providing EoLC, but many GPs and other stakeholders found such interactions more challenging than the more “technical” aspects of care.

Full text

General practitioners’ perceptions of best practice care at the end of life: A qualitative study

BJGP Open | Online – 3 September 2019 – The findings of this explorative study, as elsewhere, suggest that GPs define best practice palliative care (PC) as being accessible, holistic, dynamic, integrated, patient-centered, and a natural extension of primary care. The findings also indicate that there is a need for advocacy work to help educate clinicians, patients, and the public about the role GPs can play in PC delivery. The authors provide preliminary suggestions for how identified features of care could be implemented in day-to-day practice. These suggestions include establishing PC pathways that involve multidisciplinary teamwork, and ensuring continuous and flexible care delivery tailored to the changing needs of patients and their families.

DOI: 10.3399/bjgpopen19X101660

Full text

Experiences with approaches to advance care planning with older people: A qualitative study among Dutch general practitioners

BJGP Open | Online – 25 November 2018 – Approaches to advance care planning (ACP) with older patients can be divided into two categories: systematic and ad hoc. Respondents had positive experiences with both approaches. With systematic approaches, respondents discussed a fixed combination of topics with systematically selected community-dwelling patients and patients living in residential care facilities. Ad hoc approaches consisted of discussing one or two topics, which were often related to the near future and varied from treatment limitations to care, preferred place of living, and views on life, dying and death. Respondents reported these topics were discussed when patients took the initiative or seemed to be open to ACP, when a patient’s situation deteriorated, or if respondents felt the provided care was not appropriate. Different systematic and ad hoc approaches complemented each other in the ACP process; they could be used simultaneously or sequentially, and were used for both initiating and following up on ACP.

DOI: 10.1136/bmjopen-2018-024762

Full text

Recruiting general practitioners for palliative care research in primary care: Real-life barriers explained

BMC Family Practice | Online – 5 March 2019 – Recruitment in primary palliative care research is difficult for many reasons. This study shows a high level of interest from the GPs to be involved, but also produced high GP dropout rates and low data completion. Suggestions for future recruitment and quality improvement efforts in this field are formulated. The key is to offer a well-defined intervention that clearly benefits the GPs and/or the patients, and a well-designed research protocol that lightens the research burden on GPs and patients.

Full text

Australian GPs’ perceptions of barriers and enablers to best practice palliative care: A qualitative study

BMC PALLIATIVE CARE | Online – 23 July 2019 – Providing quality palliative care (PC) in residential aged care facilities is a high priority for ageing populations worldwide. Nursing assistants (NAs) – however termed – are the least qualified staff and provide most of the direct care. They have an important role at the frontline of care, spending more time with residents than any other care provider, but have been found to lack the necessary knowledge and skills to provide PC. The level of competence of this workforce to provide PC requires evaluation using a valid and reliable instrument designed for their level of education and the responsibilities and practices of their role. This study provides preliminary evidence for the validity and reliability of three new questionnaires that demonstrate sensitivity for NAs’ level of education and required knowledge, skills, and attitudes for providing a palliative approach. Implications for practice include the development of PC competencies through structured education and training across this workforce, and ongoing professional development opportunities for NAs, especially for those with the longest tenure.

DOI: 10.1186/s12904-019-0478-6

Full text

Media Watch, published by Barry R. Ashpole, a Canadian communications consultant and educator, monitors the literature and the lay press on issues specific to the quality of end-of-life care. It is international in scope and distribution. View current and back issues here.

Previous Page News Index Next Page