Computing and Information Systems - Research Publications
Permanent URI for this collection
Search Results
1 - 10 of 2479
-
ItemModelling mass drug administration strategies for reducing scabies burden in Monrovia, Liberia(CAMBRIDGE UNIV PRESS, 2023-08-18)Scabies is a parasitic infestation with high global burden. Mass drug administrations (MDAs) are recommended for communities with a scabies prevalence of >10%. Quantitative analyses are needed to demonstrate the likely effectiveness of MDA recommendations. In this study, we developed an agent-based model of scabies transmission calibrated to demographic and epidemiological data from Monrovia. We used this model to compare the effectiveness of MDA scenarios for achieving scabies elimination and reducing scabies burden, as measured by time until recrudescence following delivery of an MDA and disability-adjusted-life-years (DALYs) averted. Our model showed that three rounds of MDA delivered at six-month intervals and reaching 80% of the population could reduce prevalence below 2% for three years following the final round, before recrudescence. When MDAs were followed by increased treatment uptake, prevalence was maintained below 2% indefinitely. Increasing the number of and coverage of MDA rounds increased the probability of achieving elimination and the number of DALYs averted. Our results suggest that acute reduction of scabies prevalence by MDA can support a transition to improved treatment access. This study demonstrates how modelling can be used to estimate the expected impact of MDAs by projecting future epidemiological dynamics and health gains under alternative scenarios.
-
ItemNo Preview AvailableAnomaly Detection in Endemic Disease Surveillance Data Using Machine Learning Techniques(MDPI, 2023-07)Disease surveillance is used to monitor ongoing control activities, detect early outbreaks, and inform intervention priorities and policies. However, data from disease surveillance that could be used to support real-time decisionmaking remain largely underutilised. Using the Brazilian Amazon malaria surveillance dataset as a case study, in this paper we explore the potential for unsupervised anomaly detection machine learning techniques to discover signals of epidemiological interest. We found that our models were able to provide an early indication of outbreak onset, outbreak peaks, and change points in the proportion of positive malaria cases. Specifically, the sustained rise in malaria in the Brazilian Amazon in 2016 was flagged by several models. We found that no single model detected all anomalies across all health regions. Because of this, we provide the minimum number of machine learning models top-k models) to maximise the number of anomalies detected across different health regions. We discovered that the top three models that maximise the coverage of the number and types of anomalies detected across the thirteen health regions are principal component analysis, stochastic outlier selection, and the minimum covariance determinant. Anomaly detection is a potentially valuable approach to discovering patterns of epidemiological importance when confronted with a large volume of data across space and time. Our exploratory approach can be replicated for other diseases and locations to inform monitoring, timely interventions, and actions towards the goal of controlling endemic disease.
-
ItemNo Preview AvailableSelf-Rated Oral Health as a Valid Measure of Oral Health Status in Adults Living in Rural Australia(MDPI, 2023-06)It is unclear how well self-rated oral health (SROH) reflects actual oral health status in the rural Australian population. Therefore, this study aimed to compare the clinically assessed oral health status and SROH of adults living in rural Australia. The data were from 574 participants who took part in the Crossroads II cross-sectional study. Three trained and calibrated dentists evaluated the oral health status of participants based on WHO criteria. SROH was assessed with the question 'Overall, how would you rate the health of your teeth and gums?', with a score ranging from excellent = 5 to poor = 1. A logistic regression analysis (LRA) was performed, allowing us to assess factors associated with SROH. The mean age of participants was 59.2 years (SD 16.3), and 55.3% were female. The key results from the LRA show poorer SROH in those with more missing teeth (OR = 1.05; 95% CI; 1.01-1.08), more decayed teeth (OR = 1.28; 95% CI: 1.11-1.46), and more significant clinical attachment loss of periodontal tissue (6mm or more) (OR = 2.63; 95% CI: 1.29-5.38). This study found an association between negative SROH and clinical indicators used to measure poor oral health status, suggesting that self-rated oral health is an indicator of oral health status. When planning dental healthcare programs, self-reported oral health should be considered a proxy measure for oral health status.
-
ItemNo Preview AvailableModerated Online Social Therapy for Carers of Early Psychosis Clients in Real-World Settings: Cluster Randomized Controlled Trial(JMIR PUBLICATIONS, INC, 2023)BACKGROUND: Family carers of youth recovering from early psychosis experience significant stress; however, access to effective family interventions is poor. Digital interventions provide a promising solution. OBJECTIVE: Our objective was to evaluate across multiple Australian early psychosis services the effectiveness of a novel, web-based early psychosis intervention for carers. METHODS: In this cluster randomized controlled trial conducted across multiple Australian early psychosis services, our digital moderated online social therapy for carers (Altitudes) plus enhanced family treatment as usual (TAU) was compared with TAU alone on the primary outcome of perceived stress and secondary outcomes including mental health symptoms and family variables at the 6-month follow-up. RESULTS: Eighty-six caregivers were randomized and data were available for 74 young people in their care. Our primary hypothesis that carers randomized to Altitudes+TAU would report greater improvements in perceived stress at follow-up compared with carers randomized to TAU alone was not supported, with the TAU alone group showing more improvement. For secondary outcomes, the TAU alone group showed improved mindfulness over time. Regardless of group assignment, we observed improvements in satisfaction with life, quality of life, emotional overinvolvement, and burden of care. In contrast, hair cortisol concentration increased. Post hoc analyses revealed more contact with early psychosis services in the intervention group compared to TAU alone and that improvements in perceived stress and social support were associated with use of the intervention in the Altitudes+TAU group. In this study, 80% (12/15) reported a positive experience with Altitudes and 93% (14/15) would recommend it to others. CONCLUSIONS: Our trial did not show a treatment effect for Altitudes in perceived stress. However, our post hoc analysis indicated that the amount of use of Altitudes related to improvements in stress and social support. Additional design work is indicated to continue users' engagement and to significantly improve outcomes in problem-solving, communication, and self-care. TRIAL REGISTRATION: Australian and New Zealand Clinical Trials Registry ACTRN12617000942358; https://trialsearch.who.int/Trial2.aspx?TrialID=ACTRN12617000942358.
-
ItemNo Preview AvailableWeb-Based STI/HIV Testing Services Available for Access in Australia: Systematic Search and Analysis(JMIR PUBLICATIONS, INC, 2023-09-22)BACKGROUND: Sexually transmitted infection (STI) rates continue to rise in Australia, and timely access to testing and treatment is crucial to reduce transmission. Web-based services have been viewed as a way to improve timely access to STI/HIV testing and have proliferated in recent years. However, the regulation of these services in Australia is minimal, leading to concerns about their quality. The purpose of this review was to systematically identify web-based STI/HIV testing services available in Australia and assess them on aspects of quality, reliability, and accessibility. OBJECTIVE: We aim to systematically identify and assess web-based STI/HIV testing services available in Australia. METHODS: A Google search of Australian web-based services was conducted in March 2022 and repeated in September 2022 using Boolean operators and search terms related to test services (eg, on the internet or home), STIs (eg, chlamydia or gonorrhea), and test type (eg, self-test). The first 10 pages were assessed, and services were categorized as self-testing (ST; test at home), self-sampling (SS; sample at home and return to laboratory), or self-navigated pathology (SNP; specimens collected at pathology center). Website reliability was assessed against the Health on the Net Foundation code of conduct, and service quality was assessed using a scorecard that was developed based on similar reviews, Australian guidelines for in-person services, and UK standards. Additionally, we looked at measures of accessibility including cost, rural access, and time to test results. RESULTS: Seventeen services were identified (8 ST, 2 SS, and 7 SNP). Only 4 services offered recommended testing for all 4 infections (chlamydia, gonorrhea, syphilis, and HIV) including genital, anorectal, and oropharyngeal sites, and 5 offered tests other than those recommended by Australian testing guidelines (eg, Ureaplasma). Nine services (1 SNP, 8 self-test) had no minimum age requirements for access. Reliability scores (scale 0-8) were similar between all services (range 4.75-8.0). Quality weighted scores (scale 0-58) were similar between SNP and SS services (average 44.89, SD 5.56 and 44.75, SD 1.77, respectively) but lower for ST services (22.66, SD 8.93; P=.002). Government-funded services were of higher quality than private services (43.54, SD 6.71 vs 29.43, SD 13.55; P=.03). The cost for services varied between SNP (Aus $0-$595; ie, US $0-$381.96), self-sample (Aus $0; ie, US $0), and ST (Aus $0-$135; ie, US $0-$86.66). The time to test results was much shorter for SNP services (~4 days) than for SS (~12 days) and ST (~14 days). CONCLUSIONS: This review identified considerable variability in the quality and reliability of the web-based STI/HIV testing services in Australia. Given the proliferation and use of these services will likely increase, it is imperative that Australia develops national standards to ensure the standard-of-care offered by web-based STI/HIV testing services is appropriate to protect Australian users from the impact of poorly performing and inappropriate tests.
-
ItemNo Preview AvailableAn Evaluation of the Boys Do Cry Suicide Prevention Media Campaign on Twitter: Mixed Methods Approach.(JMIR Publications Inc., 2023-09-07)BACKGROUND: In most countries, men are more likely to die by suicide than women. Adherence to dominant masculine norms, such as being self-reliant, is linked to suicide in men in Western cultures. We created a suicide prevention media campaign, "Boys Do Cry," designed to challenge the "self-reliance" norm and encourage help-seeking in men. A music video was at the core of the campaign, which was an adapted version of the "Boys Don't Cry" song from "The Cure." There is evidence that suicide prevention media campaigns can encourage help-seeking for mental health difficulties. OBJECTIVE: We aimed to explore the reach, engagement, and themes of discussion prompted by the Boys Do Cry campaign on Twitter. METHODS: We used Twitter analytics data to investigate the reach and engagement of the Boys Do Cry campaign, including analyzing the characteristics of tweets posted by the campaign's hosts. Throughout the campaign and immediately after, we also used Twitter data derived from the Twitter Application Programming Interface to analyze the tweeting patterns of users related to the campaign. In addition, we qualitatively analyzed the content of Boys Do Cry-related tweets during the campaign period. RESULTS: During the campaign, Twitter users saw the tweets posted by the hosts of the campaign a total of 140,650 times and engaged with its content a total of 4477 times. The 10 highest-performing tweets by the campaign hosts involved either a video or an image. Among the 10 highest-performing tweets, the first was one that included the campaign's core video; the second was a screenshot of the tweet posted by Robert Smith, the lead singer of The Cure, sharing the Boys Do Cry campaign's video and tagging the campaign's hosts. In addition, the pattern of Twitter activity for the campaign-related tweets was considerably higher during the campaign than in the immediate postcampaign period, with half of the activity occurring during the first week of the campaign when Robert Smith promoted the campaign. Some of the key topics of discussions prompted by the Boys Do Cry campaign on Twitter involved users supporting the campaign; referencing the original song, band, or lead singer; reiterating the campaign's messages; and having emotional responses to the campaign. CONCLUSIONS: This study demonstrates that a brief media campaign such as Boys Do Cry can achieve good reach and engagement and can prompt discussions on Twitter about masculinity and suicide. Such discussions may lead to greater awareness about the importance of seeking help and providing support to those with mental health difficulties. However, this study suggests that longer, more intensive campaigns may be needed in order to amplify and sustain these results.
-
ItemNo Preview AvailablePatients' Views on AI for Risk Prediction in Shared Decision-Making for Knee Replacement Surgery: Qualitative Interview Study(JMIR PUBLICATIONS, INC, 2023-09-18)BACKGROUND: The use of artificial intelligence (AI) in decision-making around knee replacement surgery is increasing, and this technology holds promise to improve the prediction of patient outcomes. Ambiguity surrounds the definition of AI, and there are mixed views on its application in clinical settings. OBJECTIVE: In this study, we aimed to explore the understanding and attitudes of patients who underwent knee replacement surgery regarding AI in the context of risk prediction for shared clinical decision-making. METHODS: This qualitative study involved patients who underwent knee replacement surgery at a tertiary referral center for joint replacement surgery. The participants were selected based on their age and sex. Semistructured interviews explored the participants' understanding of AI and their opinions on its use in shared clinical decision-making. Data collection and reflexive thematic analyses were conducted concurrently. Recruitment continued until thematic saturation was achieved. RESULTS: Thematic saturation was achieved with 19 interviews and confirmed with 1 additional interview, resulting in 20 participants being interviewed (female participants: n=11, 55%; male participants: n=9, 45%; median age: 66 years). A total of 11 (55%) participants had a substantial postoperative complication. Three themes captured the participants' understanding of AI and their perceptions of its use in shared clinical decision-making. The theme Expectations captured the participants' views of themselves as individuals with the right to self-determination as they sought therapeutic solutions tailored to their circumstances, needs, and desires, including whether to use AI at all. The theme Empowerment highlighted the potential of AI to enable patients to develop realistic expectations and equip them with personalized risk information to discuss in shared decision-making conversations with the surgeon. The theme Partnership captured the importance of symbiosis between AI and clinicians because AI has varied levels of interpretability and understanding of human emotions and empathy. CONCLUSIONS: Patients who underwent knee replacement surgery in this study had varied levels of familiarity with AI and diverse conceptualizations of its definitions and capabilities. Educating patients about AI through nontechnical explanations and illustrative scenarios could help inform their decision to use it for risk prediction in the shared decision-making process with their surgeon. These findings could be used in the process of developing a questionnaire to ascertain the views of patients undergoing knee replacement surgery on the acceptability of AI in shared clinical decision-making. Future work could investigate the accuracy of this patient group's understanding of AI, beyond their familiarity with it, and how this influences their acceptance of its use. Surgeons may play a key role in finding a place for AI in the clinical setting as the uptake of this technology in health care continues to grow.
-
ItemNo Preview AvailableVideo Calls as a Replacement for Family Visits During Lockdowns in Aged Care: Interview Study With Family Members(JMIR PUBLICATIONS, INC, 2023)BACKGROUND: Lockdowns have been used to prevent the spread of transmissible illnesses such as influenza, norovirus, and COVID-19 in care homes. However, lockdowns deny care home residents supplemental care and the socioemotional enrichment that comes from seeing family members. Video calling has the potential to enable ongoing contact between residents and family members during lockdowns. However, video calls can be considered by some as a poor substitute for in-person visits. It is important to understand family members' experiences with video calling during lockdowns to ensure the effective use of this technology in the future. OBJECTIVE: This study aimed to understand how family members use video calls to communicate with relatives living in aged care during lockdowns. We focused on experiences during the COVID-19 pandemic, which involved extensive lockdowns in aged care homes. METHODS: We conducted semistructured interviews with 18 adults who had been using video calls with relatives living in aged care during pandemic lockdowns. The interviews focused on how participants had been using video calls, what benefits they gained from video-based interactions, and what challenges they encountered when using the technology. We analyzed the data using the 6-phase reflexive approach to thematic analysis by Braun and Clarke. RESULTS: We developed 4 themes through our analysis. Theme 1 interprets video calling as a medium for the continuation of care during lockdowns. Using video calls, family members were able to provide social enrichment for residents and engaged in health monitoring to uphold residents' welfare. Theme 2 highlights how video calling extended care by supporting frequent contact, transmitting nonverbal cues that were essential for communication, and negating the need for face masks. Theme 3 interprets organizational issues such as the lack of technology and staff time as impediments to the continuation of familial care through video. Finally, theme 4 highlights the need for 2-way communication, interpreting residents' unfamiliarity with video calling and their health conditions as further barriers to the continuation of care. CONCLUSIONS: This study suggests that, during restrictions arising from the COVID-19 pandemic, video calls became a medium for enabling family members to continue participating in the care of their relatives. The use of video calls to continue care illustrates their value for families during times of mandatory lockdown and supports the use of video to complement face-to-face visits at other times. However, better support is needed for video calling in aged care homes. This study also revealed a need for video calling systems that are designed for the aged care context.
-
ItemNo Preview Available
-
ItemNo Preview AvailableUpdatable Learned Indexes Meet Disk-Resident DBMS - From Evaluations to Design Choices(Association for Computing Machinery (ACM), 2023-06-13)Although many updatable learned indexes have been proposed in recent years, whether they can outperform traditional approaches on disk remains unknown. In this study, we revisit and implement four state-of-the-art updatable learned indexes on disk, and compare them against the B+-tree under a wide range of settings. Through our evaluation, we make some key observations: 1) Overall, the B+-tree performs well across a range of workload types and datasets. 2) A learned index could outperform B+-tree or other learned indexes on disk for a specific workload. For example, PGM achieves the best performance in write-only workloads while LIPP significantly outperforms others in lookup-only workloads. We further conduct a detailed performance analysis to reveal the strengths and weaknesses of these learned indexes on disk. Moreover, we summarize the observed common shortcomings in five categories and propose four design principles to guide future design of on-disk, updatable learned indexes: (1) reducing the index's tree height, (2) better data structures to lower operation overheads, (3) improving the efficiency of scan operations, and (4) more efficient storage layout.