黑料网

In advanced, incurable cancer patients, the kind and severity of psychological and physical symptoms change with time. The sick and informal carers both have to deal with social obligations. Palliative care is described as "using a team-oriented approach to improve the quality of life of patients and their families who are dealing with challenges associated to a life-threatening disease" in the World Health Organization's (WHO) definition of palliative care [1]. It prevents and lessens suffering by the early diagnosis, correct evaluation, and treatment of pain and other disorders, whether they are physical, psychological, or spiritual. It offers a support system to help people live as actively as they can till death. PC respects patients' rights to adequate, culturally acceptable, and personally pertinent information about their health status is order for them to participate in the decisionmaking process for their care [2]. All health care levels from primary to tertiary care and settings including homes and institutions can benefit from palliative care [3]. Primary care is offered by general practitioners, oncologists, hospitals, and outpatient centres [4,5]. Palliative-care specialists act as consultants in secondary palliative care, which is given to all patients with a symptomatic advanced, progressive lifethreatening disease and a lack of curative options [6]. Additionally, most standards make reference to this collective. PC has evolved over the past 50 years from caring for patients who were at the end of their lives to become a highly specialized discipline that focuses on providing supportive care to patients with life-limiting conditions across the disease trajectory. The difficulty of a reliable and timely screening is still seen differently, as is the timing of palliative care throughout an illness [7]. To the best of our knowledge, there is no review that combines the ideas of creating synthetic data and considering its potential use to the screening of PC requirements in the literature. We provide an introduction to both disciplines here for an initial conjunction and motivation for using this rapidly developing computational field within an important medical domain in order to increase public awareness and start a discussion about such cutting-edge PC or related disciplines in personalized medicine.

Two screening techniques are typically used to determine if palliative care is appropriate: one is based on the patient's prognosis, and the other is limited to PC requirements. Focusing on prognosis is necessary since palliative care has to be intensified in the last two months of life for the majority of patients with advanced cancer symptoms and other diseases. The main warning signs of this stage include ill general health, weight loss, clinical symptoms like anorexia, breathlessness, or confusion, and abnormal laboratory findings like a high white cell count, lymphopenia, hyperalbuminemia, elevated lactate dehydrogenase, or C-reactive protein and Vitamin B12 levels. The prognosis can also be determined by scores measuring physical impairments and patient mortality based on comorbidities or the frequency of symptoms, as well as other specific criteria [8]. The accuracy of principally five methodologies was evaluated throughout eight studies in research that employed prognostic tools for identification, according to a detailed study of that research. The specificity varied from 26% to 99%, whereas the sensitivity ranged from 3% to 94%. The authors come to the conclusion that the ability of current screening tools to identify individuals with severe progressive diseases who are likely to need palliative care is limited. The current gold standard for identifying patient needs is the Patient Reported Outcome Measurement (PROM). The Distress Thermometer (DT) and other instruments, including MIDOS, ESAS, and IPOS, are now recommended for use in symptom evaluation. The Minimal Documentation System for Patients in Palliative Care includes ten questions on unpleasant physical symptoms as well as inquiries about worry and despair (MIDOS). The Edmonton Symptom Assessment System (ESAS) includes eight physically distressing symptoms as well as questions regarding mood and general wellbeing. The distress thermometer allows patients to assess their psychological suffering on a scale of zero to ten [9,10]. Physical symptoms and those from the psychosocial category are combined on the Integrated Palliative Care Outcome Scale (IPOS). According to a recent study, AI-based Clinical Decision Supports Systems (CDSS) may considerably help in detecting PC needs if the findings from these PROMs are included into the systems. Numerous studies have been conducted to date to assess screening methods that incorporate prognostic variables with symptoms and needs, including diagnosis, functional status, sequelae, and comorbidities (such as symptom management, distress, and support from family).

Discussion

Building artificial intelligence models for synthetic data creation to improve Machine Learning (ML) technique is a crucial part of computer science research and is now used to allied medical fields like Systems Medicine and Medical Informatics. Medical researchers have been driven for over 10 years.

Due to the overall unavailability and sparsity of data, there are still major limitations imposed by the idea of personalized decision-making help based on patient data. Compare this to current technology, which enables us to generate and analyse patient data in a number of formats, including tabular data on health records, medical images, genetic data, and even audio and video. One method of overcoming these data limitations with regard to medical records is the generation of synthetic tabular data based on real-world data. Consequently,ML-assisted decision-support may be easier to understand with more important patient data accessible. Modern ML algorithms generate and make decisions based on such data, from a methodological standpoint. A widespread practical implementation in actual clinical settings is however still hindered by severe issues. In this study, we will offer the first insights into existing opinions and potential effects of adopting synthetic data generation in palliative care screening using a tough primary example of highly customised, hardly available patient information. Together, the reader and I will learn some fundamental concepts and practical options that are relevant to producing and utilising synthetic data for ML-based screening in palliative care and other domains.

Conclusion

Palliative care is now a highly specialised specialty devoted to providing supportive care to patients with life-limiting diseases at every stage of their illness rather than only caring for people at the very end of their lives. This extremely particular track demands specialised attention and awareness in order to conduct a proper and timely screening, which is a time-consuming and domain-expertise-driven approach that is difficult to carry out in clinical practise at all times. A physician-independent automated screening supporting the doctor's evaluation would be required to improve the referral of patients with unmet palliative care requirements. Although existing AI solutions provide a valuable toolbox, they are still limited in terms of data accessibility, which limits their ability to be used in a wide variety of clinical settings.

Acknowledgement

Not applicable.

Conflict of Interest

Author declares no conflict of interest.

1. Radbruch L, De Lima L, Knaul F, Wenk R, Ali Z, et al. (2020) J Pain Symptom Manag 60: 754-764.

, ,

2. Lopes IM, Guarda T, Oliveira P (2020) J Med Syst 44: 1-9.

, ,

3. Kamal AH, Bausewein C, Casarett DJ, Currow DC, Dudgeon DJ, et al. (2020) J Clin Oncol 38: 987-994.

, ,

4. Hui D, Bruera E (2016) Nat Rev Clin Oncol 13: 159-171.

, ,

5. Rangachari D, Smith TJ, Kimmel S (2013) Cancer J 19: 373.

, ,

6. Schenker Y, Arnold R (2015) JAMA 314: 1565.

, ,

7. Schenker Y, Crowley-Matoka M, Dohan D, Rabow MW, Smith CB, et al. (2014) J Oncol Pract 10: 37.

, ,

8. Naghavi M, Abajobir AA, Abbafati C, Abbas KM, Abd-Allah F, et al. (2017) Lancet 390: 151-210.

, ,

9. Ellis J, Cobb M, O’Connor T, Dunn L, Irving G, et al. (2015) Chronic Illn 11: 198-209.

, ,

10. Goldsbury DE, O’Connell DL, Girgis A, Wilkinson A, Phillips JL, et al. (2015) BMC Health Serv Res 15: 1-14.

, ,