CATS: Customizable Abstractive Topic-based Summarization

Neural sequence-to-sequence models are the state-of-the-art approach used in abstractive summarization of textual documents, useful for producing condensed versions of source text narratives without being restricted to using only words from the original text. Despite the advances in abstractive summarization, custom generation of summaries (e.g., towards a user’s preference) remains unexplored. In this article, we present CATS, an abstractive neural summarization model that summarizes content in a sequence-to-sequence fashion while also introducing a new mechanism to control the underlying latent topic distribution of the produced summaries. We empirically illustrate the efficacy of our model in producing customized summaries and present findings that facilitate the design of such systems. We use the well-known CNN/DailyMail dataset to evaluate our model. Furthermore, we present a transfer-learning method and demonstrate the effectiveness of our approach in a low resource setting, i.e., abstractive summarization of meetings minutes, where combining the main available meetings’ transcripts datasets, AMI and International Computer Science Institute(ICSI) , results in merely a few hundred training documents.

Denoising EEG Signals for Real-World BCI Applications Using GANs

As a measure of the brain's electrical activity, electroencephalography (EEG) is the primary signal of interest for brain-computer-interfaces (BCI). BCIs offer a communication pathway between a brain and an external device, translating thought into action with suitable processing. EEG data is the most common signal source for such technologies. However, artefacts induced in BCIs in the real-world context can severely degrade their performance relative to their in-laboratory performance. In most cases, the recorded signals are so heavily corrupted by noise that they are unusable and restrict BCI's broader applicability. To realise the use of portable BCIs capable of high-quality performance in a real-world setting, we use Generative Adversarial Networks (GANs) that can adopt both supervised and unsupervised learning approaches. Although our approach is supervised, the same model can be used for unsupervised tasks such as data augmentation/imputation in the low resource setting. Exploiting recent advancements in Generative Adversarial Networks (GAN), we construct a pipeline capable of denoising artefacts from EEG time series data. In the case of denoising data, it maps noisy EEG signals to clean EEG signals, given the nature of the respective artefact. We demonstrate the capability of our network on a toy dataset and a benchmark EEG dataset developed explicitly for deep learning denoising techniques. Our datasets consist of an artificially added mains noise (50/60 Hz) artefact dataset and an open-source EEG benchmark dataset with two artificially added artefacts. Artificially inducing myogenic and ocular artefacts for the benchmark dataset allows us to present qualitative and quantitative evidence of the GANs denoising capabilities and rank it among the current gold standard deep learning EEG denoising techniques. We show the power spectral density (PSD), signal-to-noise ratio (SNR), and other classical time series similarity measures for quantitative metrics and compare our model to those previously used in the literature. To our knowledge, this framework is the first example of a GAN capable of EEG artefact removal and generalisable to more than one artefact type. Our model has provided a competitive performance in advancing the state-of-the-art deep learning EEG denoising techniques. Furthermore, given the integration of AI into wearable technology, our method would allow for portable EEG devices with less noisy and more stable brain signals.

Advancement, Collaboration and Effective Combat in Low Resource Setting: Public Health Overview After One year of the Corona Virus Pandemic in Bangladesh

Background: Current Coronavirus pandemic causing millions of deaths and unfathomable damage of nations worldwide, especially in health sector. Bangladesh is dealing with the biggest catastrophic public health event of the history in a courageous and effective way. An evidence based narrative review has been undergone to scientifically describe Bangladesh government’s measures to encounter the Corona pandemic, so far. The aim of this study is to document the collaborative action of different ministries of Bangladesh government during this pandemic to understand the in-depth steps of the healthcare provision and disaster preparedness of the public-private-international association in a low-resource setting. Methods: A literature review over five months has been conducted to write down the evidential narration of the activities against the pandemic damage in Bangladesh. Keyword and result based literatures and current media reports searched has been done. Selection criteria: Both online and offline reports, descriptive articles, governmental portal and ministerial websites were reviewed. The description is reported specifically based on the documents directed by government to fight against COVID-19 from the beginning of the pandemic till the writing period. Findings and discussion: In spite of the resource constraints, government of Bangladesh has been able to limit the damage in an optimal level. The inter- and- inter ministerial functional proposition and collaboration in national and international stakeholders initiated and sustained by the government strengthen the shield against the Coronavirus invasion. Conclusion: The sufferings brought by the pandemic knows no bound. The pandemic damage and ruin are unspeakable and undeniable at the same time. It is time to observe the positivity and critically appreciate the efforts taken by the current governmental authority to make a constructive remark for present situation, and be prepare for future building of the nation. JOPSOM 2021; 40(1): 66-71

Operating room efficiency in a low resource setting: a pilot study from a large tertiary referral center in Ethiopia

Background The operating room (OR) is one of the most expensive areas of a hospital, requiring large capital and recurring investments, and necessitating efficient throughput to reduce costs per patient encounter. On top of increasing costs, inefficient utilization of operating rooms results in prolonged waiting lists, high rate of cancellation, frustration of OR personnel as well as increased anxiety that negatively impacts the health of patients. This problem is magnified in developing countries, where there is a high unmet surgical need. However, no system currently exists to assess operating room utilization in Ethiopia. Methodology A prospective study was conducted over a period of 3 months (May 1 to July 31, 2019) in a tertiary hospital. Surgical case start time, end time, room turnover time, cancellations and reason for cancellation were observed to evaluate the efficiency of eight operating rooms. Results A total of 933 elective procedures were observed during the study period. Of these, 246 were cancelled, yielding a cancellation rate of 35.8%. The most common reasons for cancellation were related to lack of OR time and patient preparation (8.7% and 7.7% respectively). Shortage of facilities (instrument, blood, ICU bed) were causes of cancelation in 7.7%. Start time was delayed in 93.4% (mean 8:56 am ± 52 min) of cases. Last case completion time was early in 47.9% and delayed in 20.6% (mean 2:54 pm ± 156 min). Turnover time was prolonged in 34.5% (mean 25 min ± 49 min). Total operating room utilization ranged from 10.5% to 174%. Operating rooms were underutilized in 42.7% while overutilization was found in 14.6%. Conclusion We found a high cancellation rate, most attributable to late start times leading to delays for the remainder of cases, and lack of preoperative patient preparation. In a setting with a high unmet burden of surgical disease, OR efficiency must be maximized with improved patient evaluation workflows, adequate OR staffing and commitment to punctual start times. We recommend future quality improvement projects focusing on these areas to increase OR efficiency.

Nursing Care of Low-Risk Newborns in Low Resource Setting: Nurses' Aides May Bridge the Gap at Community Hospitals

Objective: To study the role of a nurses' aide in the care for newborns weighing between 1500 and 2000 g at birth in a low resource setting. Study Design: Observational. Setting: The General hospital in 1994-95, in a public sector, located in a remote area in India Intervention: A female ward assistant with seven years of schooling trained, on-the-job, to keep babies warm, initiate maternal breastfeeding, and to detect rapid breathing. The nursing staff from the pediatric ward supervised her performance. A separate "warm room" appropriately heated for preterm and sick babies became a makeshift nursery. The nursing staff administered enteral feeding, oxygen, and antibiotics. Services of the resident doctors or general duty medical officers were not available. Results: The survival rate was nearly 100% for babies with birthweights between 1,500 and 2,000 g (none referred out). Conclusions: A nurses' aide may facilitate the delivery of special care for newborns where nursing personnel are grossly inadequate and saving babies weighing between 1,500 and 2,000 g may need minimal inputs. It may be worthwhile to target 1,500 and 2,000 g birthweight categories even when resources are meager. What is already known about this subject? Low resource settings face staff shortages, especially nursing staff. Health workers with midwifery skills can deliver nearly 90% of essential care services for maternal and neonatal health. A substantial proportion of neonatal deaths occur among moderately low birth weight babies. What does this study add? It is possible to train a semi-literate person to facilitate early breastfeeding and to keep a baby warm. A large proportion of deaths among babies with birthweight ranging from 1500 to 2000 g are preventable with meager resources. How might this impact on clinical practice or future developments? The facilities facing shortage of nursing staff in low resource settings, may employ nurses’ aide to deliver basic newborn care.

A study on outcome of controlled ovarian hyperstimulation and intrauterine insemination (husband) in a medical college in Kolkata

Background: A combination of controlled ovarian hyperstimulation and intrauterine insemination (IUI) remains an important treatment option for couple having infertility. Success rate of IUI with ovulation induction ranges from 8-20% depending on many factors. Aims and Objectives: To assess the factors affecting the success rate of IUI and to evaluate the success of ovulation Induction by different methods of controlled stimulation protocol. Materials and Methods: It is a duration-based prospective cross-sectional study where total of 67 couples were included by inclusion and exclusion criteria. They underwent 90 cycles of IUI with each couple having a maximum of three cycles. Ovulation induction was done by clomiphene citrate or letrozole or gonadotrophins. Semen preparation was done by density gradient method. The outcomewas measured by positive urine pregnancy test. Range, percentage, confidence interval, mean with standard deviation, median, range, and P-value were calculated. P<0.05 was taken as statistically significant. Results: Out of 90 IUI cycles 8 were successful resulting in a success rate of 8.8% per cycle and 11.9% per couple. Factors that had a positive impact were follicle size >21 sqmm, endometrial thickness >9 mm, post wash count >15 million/ml, >2 cycles of IUI and on the number of follicles 2 or more on the day of trigger. Conclusion: The study concluded that IUI after ovulation induction can be a simple and safe cost-effective procedure in selected group of infertile couple. Clinical significance IUI following ovulation induction can be a successful approach for specific indications in a low-resource setting where options for other ART interventions are absent or limited.

Tibial plateau fractures in a limited-resource setting: A prospective study of surgical techniques and outcome over three years

Introduction: Tibial plateau fractures form a wide spectrum of injuries accounting for 1.2% of all fractures and a prevalence of 10 cases per 100,000 inhabitants. Methodology: A prospective consecutive multicentre study from May 2018 to May 2021 was carried out in Yaounde. All consenting cases of tibial plateau fracture underwent surgical treatment while patients with pathologic fractures, previous knee osteoarthritis, medically unfit for surgery, and discharging against medical advice were excluded. Data was analysed with SPSS 26.0 and the level of significance set at p<0.05. Results:Eighty-four (84) cases of tibial plateau fractures were sampled and 68 consented to surgery. The mean age was 42 ±13.6 years and sex ratio 2.4. Estimated prevalence was 2.2 cases per 100,000 inhabitants. Schatzker type II fractures were most represented (33.3%). The left leg was affected in 57.1%. Motorbike accidents were the main cause of injury (66.7%). Of the 68 operated, 63.3% by plating osteosynthesis, 32.4% by external fixation, and 4.4% by screws fixation. Tricortical iliac bone graft was realised in 4 cases. The minimum follow-up was 6 months, with a median of 18 months (5 to 37 months). Plating osteosynthesis (p<0.001), operative time between 60 to 120 minutes (p<0.02) and a good radiologic fracture healing (p<0.04) were associated with a satisfactory outcome. Poor prognosis was seen with open fractures (p<0.001), bridging external fixation (p<0.001), and Schatzker VI fractures (p<0.02). Complications included post-traumatic osteoarthritis (64.7%), post-traumatic osteomyelitis (29.4%), knee ankylosis (5.9%), and limb malalignment (30.9%). Conclusion:The prevalence of tibial plateau fractures remains lower than reported in literature but it is projected to rise. Plating remains a viable treatment option. A larger scale study will establish the burden of this entity in our context.

Participatory health research under COVID-19 restrictions in Bauchi State, Nigeria: Feasibility of cellular teleconferencing for virtual discussions with community groups in a low-resource setting

Introduction During the COVID-19 pandemic, researchers have used Internet-based applications to conduct virtual group meetings, but this is not feasible in low-resource settings. In a community health research project in Bauchi State, Nigeria, COVID-19 restrictions precluded planned face-to-face meetings with community groups. We tested the feasibility of using cellular teleconferencing for these meetings. Methods In an initial exercise, we used cellular teleconferencing to conduct six male and six female community focus group discussions. Informed by this experience, we conducted cellular teleconferences with 10 male and 10 female groups of community leaders, in different communities, to discuss progress with previously formulated action plans. Ahead of each teleconference call, a call coordinator contacted individual participants to seek consent and confirm availability. The coordinator connected the facilitator, the reporter, and the participants on each conference call, and audio-recorded the call. Each call lasted less than 1 h. Field notes and debriefing meetings with field teams supported the assessment of feasibility of the teleconference meetings. Results Cellular teleconferencing was feasible and inexpensive. Using multiple handsets at the base allowed more participants in a call. Guidelines for facilitators and participants developed after the initial meetings were helpful, as were reminder calls ahead of the meeting. Connecting women participants was challenging. Facilitators needed extra practice to support group interactions without eye contact and body language signals. Conclusions With careful preparation and training, cellular teleconferencing can be a feasible and inexpensive method of conducting group discussions in a low-resource setting.