No abstract

The ability to automatically detect stuttering events in speech could help speech pathologists track an individual's fluency over time or help improve speech recognition systems for people with atypical speech patterns. Despite increasing interest in this area, existing public datasets are too small to build generalizable dysfluency detection systems and lack sufficient annotations. In this work, we introduce Stuttering Events in Podcasts (SEP-28k), a dataset containing over 28k clips labeled with five event types including blocks, prolongations, sound repetitions, word repetitions, and interjections. Audio comes from public podcasts largely consisting of people who stutter interviewing other people who stutter. We benchmark a set of acoustic models on SEP-28k and the public FluencyBank dataset and highlight how simply increasing the amount of training data improves relative detection performance by 28% and 24% F1 on each. Annotations from over 32k clips across both datasets will be publicly released.

Factors affecting the eligibility and acceptability of voluntary counselling and rapid HIV testing (VCT) were examined among pregnant women presenting in labour in Pune, India. Of the 6702 total women appearing at the delivery room from April 2001 to March 2002, 4638 (69%) were admitted for normal delivery. The remaining women presented with obstetrical complications, delivered immediately or were detected to be in false labour. Overall, 2818 (61%) of the admitted women had been previously tested for HIV during their pregnancy. If previously seen in the hospital's affiliated antenatal clinic, the likelihood of being previously tested was 89%, in contrast to 27% of women having prenatal care elsewhere. Of the admitted women, 3436 (74.3%) were assessed for their eligibility for rapid HIV VCT in the delivery room. Only 1322 (38%) of these women were found to be in early labour and without severe pain or complications, and therefore eligible for rapid HIV screening in the delivery room (DR). Of those 1322 eligible women, only 582 (44%) consented and were tested for HIV, of whom nine (1.6%) were found to be HIV-infected. Of the 1674 women arriving in the DR with no evidence of previous HIV testing, through this DR screening programme, we identified four women with HIV who could now benefit from treatment with ART. Given the high rates of HIV testing in the antenatal clinic at this site and the challenges inherent to conducting DR screening, alternatives such as post-partum testing should be considered to help reduce maternal to infant transmission in this population.

This study examined acceptability among pregnant women and their husbands for HIV testing within the antenatal clinic (ANC) and delivery room (DR) of a government hospital in Pune, India from September 2000 to November 2001. Acceptance of HIV counselling and testing was high with 83% of eligible women in the antenatal clinic (851 of 1025) and 68% of eligible women in the delivery room (417 of 613) getting tested on the same day. Structured interviews were conducted on 94 pregnant women in the ANC 50 women in the DR, and 100 husbands who accompanied their wives in the ANC. These data indicated that the majority of women agreed to be tested independently without the need for further consultation with family members, a view that was strongly supported in this sub-sample of accompanying husbands. For delivering women who were not progressing in their labour, counselling in the DR allowed for individual attention to questions and concerns thereby making counselling in the DR feasible.

Our objective was to determine the level of HIV/AIDS knowledge of pregnant women in India. In a sub-sample of these women, we documented the extent to which they experienced adverse social and physical difficulties within their home. The study was performed at an urban antenatal hospital clinic in Maharastra, India. From April to September 2001, structured interviews were conducted on 707 randomly selected antenatal clinic patients related to HIV/AIDS knowledge. Of these, 283 were further interviewed to document any social or physical difficulties they experienced. Over 75% of women displayed knowledge of primary transmission routes. Nearly 70% of women demonstrated knowledge of maternal to child transmission, however, only 8% knew of any methods of prevention. TV and written material were more strongly related to knowledge than access to radio messages or conversations with individuals. Thirty per cent of the women experienced physical or mental abuse or their spouse's alcohol and/or drug problems. Women reporting such abuse were more than twice as likely to have adequate HIV/AIDS knowledge compared with women reporting no such abuse. We found no relationship between reported household abuse and educational level of woman, husband, occupation of either partner, language or religion. We found no relationship between HIV status and knowledge of HIV and no relationship between HIV status and risk of abuse in the household. However, the total number of HIV patients in our sample was very small.