Tasman Medical Journal

ISSN:  2652-1881

Volume 2021 - Issue 2.

Australia’s contribution to Artificial Intelligence research in medicine

Jonathon E Stewart, Girish Dwivedi, Kevin Chai and Frank Sanfilippo

Abstract

Abstract

Objectives: Artificial intelligence (AI) research is undergoing a renaissance, brought about by advances in both machine learning (ML) and deep learning (DL). There is currently no data on Australia’s contribution to AI research in medicine. We performed a bibliometric analysis to assess the quantity and quality of scientific publications concerning AI in medicine over the last ten years, and compare results from Australia to nine other countries publishing in this domain.

Methods: The SCOPUS database was searched for all journal articles or conference papers containing the words “artificial intelligence”, “machine learning” or “deep learning” published in subject areas related to medicine from 1 January 2008 to 31 December 2017.

Results: Australia was the source of 470 papers from 2008 to 2017 and accounted for 3.45% of world’s AI, ML, and DL in medicine publications over this period. Worldwide, publications increased significantly, especially in the last 2 years of the period studied. The increase in output was notable from China (254%), India (179%), Australia (149%), and the United States (142%). When 2016 publication output is adjusted by population Australia ranks third (behind Canada and the United Kingdom), but the trend is not statistically significant. We found no significant variation in the average number of citations received per paper in any of the top publishing countries.

Conclusion: The publication rate for AI, ML, and DL related to medicine has increased over the last 10 years. Australia’s contribution to AI research in this area, measured as both average citation rate or adjusted by country population is comparable in quantity and quality to other leading countries.

Tasman Medical Journal 2021; 3(2): 87-93

ABSTRACT | FULL TEXT | PDF

Treatment outcomes in a 10-year geographic cohort of patients with adult-onset Still’s Disease: Case series and literature review

Pauline Habib, Graeme J Carroll and Angela Jacques

Abstract

Introduction: Adult-onset Stills Disease (AOSD) is a rare immuno-inflammatory disease with an unpredictable course. Contemporary treatment with biological agents is commonplace but the efficacy of treatment and effects on long-term outcomes are poorly defined. We investigated prognostic markers and evaluated outcomes in an Australian cohort of hospitalized patients with AOSD diagnosed and treated over 10 years.

Methods: Patients greater than 18 years of age with a discharge ICD-10 code for AOSD were selected from medical records at 3 tertiary hospitals in Perth, Western Australia from 2009-2019. Demographic data, and clinical and serological outcomes after 12 months of follow-up were collected. Relationships between plasma ferritin and clinical remission, flare rate, need for conventional treatment with or without conventional or biological disease-modifying agents (csDMARD and bDMARD) and Physician Global Assessment score (PGA) were evaluated, as were the effect of treatments on serologic markers of remission and PGA. Fishers exact tests for a categorical comparisons and Mann-Whitney U-tests for comparison of continuous outcomes were performed. P < 0.05 was considered statistically significant.

Results: All patients received corticosteroid therapy and 18 of 24 also received csDMARDs. Of these, 8 progressed to bDMARDs. Lower baseline ferritin concentrations were observed in those who achieved complete remission, required csDMARDs +/- bDMARDs, or had more flares per 100 patient-years and lower PGA scores, but the results were not statistically significant. Serological markers of remission decreased markedly with all treatments, but no significant differences between treatments or in patients achieving remission was observed.

Conclusions: Long term follow-up was limited but remission rates appeared low. No unequivocal predictors of treatment response or outcome were identified. A trend towards greater use of csDMARDs and bDMARDs, and low PGA scores (poor outcome at 12 months), implying worse outcomes was observed with hyperferritinemia. Accordingly, marked hyperferritinemia may be a biomarker for more sustained or severe AOSD manifestations and predispose to more relapses. Further research is required to examine this possibility in more detail.

Tasman Medical Journal 2021; 3(2): 94-101

ABSTRACT | FULL TEXT | PDF