Using machine learning techniques to characterize sleep-deprived driving behavior.

OBJECTIVE: Sleep deprivation is known to affect driving behavior and may lead to serious car accidents similar to the effects from e.g., alcohol. In a previous study, we have demonstrated that the use of machine learning techniques allows adequate characterization of abnormal driving behavior after alprazolam and/or alcohol intake. In the present study, we extend this approach to sleep deprivation and test the model for characterization of new interventions. We aimed to classify abnormal driving behavior after sleep deprivation, and, by using a machine learning model, we tested if this model could also pick up abnormal driving behavior resulting from other interventions. METHODS: Data were collected during a previous study, in which 24 subjects were tested after being sleep-deprived and after a well-rested night. Features were calculated from several driving parameters, such as the lateral position, speed of the car, and steering speed. In the present study, we used a gradient boosting model to classify sleep deprivation. The model was validated using a 5-fold cross validation technique. Next, probability scores were used to identify the overlap of driving behavior after sleep deprivation and driving behavior affected by other interventions. In the current study alprazolam, alcohol, and placebo are used to test/validate the approach. RESULTS: The sleep deprivation model detected abnormal driving behavior in the simulator with an accuracy of 77 ± 9%. Abnormal driving behavior after alprazolam, and to a lesser extent also after alcohol intake, showed remarkably similar characteristics to sleep deprivation. The average probability score for alprazolam and alcohol measurements was 0.79, for alcohol 0.63, and for placebo only 0.27 and 0.30, matching the expected relative drowsiness. CONCLUSION: We developed a model detecting abnormal driving induced by sleep deprivation. The model shows the similarities in driving characteristics between sleep deprivation and other interventions, i.e., alcohol and alprazolam. Consequently, our model for sleep deprivation may serve as a next reference point for a driving test battery of newly developed drugs.

The use of machine learning improves the assessment of drug-induced driving behaviour.

RATIONALE: Car-driving performance is negatively affected by the intake of alcohol, tranquillizers, sedatives and sleep deprivation. Although several studies have shown that the standard deviation of the lateral position on the road (SDLP) is sensitive to drug-induced changes in simulated and real driving performance tests, this parameter alone might not fully assess and quantify deviant or unsafe driving. OBJECTIVE: Using machine learning we investigated if including multiple simulator-derived parameters, rather than the SDLP alone would provide a more accurate assessment of the effect of substances affecting driving performance. We specifically analysed the effects of alcohol and alprazolam. METHODS: The data used in the present study were collected during a previous study on driving effects of alcohol and alprazolam in 24 healthy subjects (12 M, 12 F, mean age 26 years, range 20-43 years). Various driving features, such as speed and steering variations, were quantified and the influence of administration of alcohol or alprazolam was assessed to assist in designing a predictive model for abnormal driving behaviour. RESULTS: Adding additional features besides the SDLP increased the model performance for prediction of drug-induced abnormal driving behaviour (from an accuracy of 65 %-83 % after alprazolam intake and from 50 % to 76 % after alcohol ingestion). Driving behaviour influenced by alcohol and alprazolam was characterised by different feature importance, indicating that the two interventions influenced driving behaviour in a different way. CONCLUSION: Machine learning using multiple driving features in addition to the state-of-the-art SDLP improves the assessment of drug-induced abnormal driving behaviour. The created models may facilitate quantitative description of abnormal driving behaviour in the development and application of psychopharmacological medicines. Our models require further validation using similar and unknown interventions.

Development of Novel, Value-Based, Digital Endpoints for Clinical Trials: A Structured Approach Toward Fit-for-Purpose Validation.

Novel digital endpoints gathered via wearables, small devices, or algorithms hold great promise for clinical trials. However, implementation has been slow because of a lack of guidelines regarding the validation process of these new measurements. In this paper, we propose a pragmatic approach toward selection and fit-for-purpose validation of digital endpoints. Measurements should be value-based, meaning the measurements should directly measure or be associated with meaningful outcomes for patients. Devices should be assessed regarding technological validity. Most importantly, a rigorous clinical validation process should appraise the tolerability, difference between patients and controls, repeatability, detection of clinical events, and correlation with traditional endpoints. When technically and clinically fit-for-purpose, case building in interventional clinical trials starts to generate evidence regarding the response to new or existing health-care interventions. This process may lead to the digital endpoint replacing traditional endpoints, such as clinical rating scales or questionnaires in clinical trials. We recommend initiating more data-sharing collaborations to prevent unnecessary duplication of research and integration of value-based measurements in clinical care to enhance acceptance by health-care professionals. Finally, we invite researchers and regulators to adopt this approach to ensure a timely implementation of digital measurements and value-based thinking in clinical trial design and health care. SIGNIFICANCE STATEMENT: Novel digital endpoints are often cited as promising for the clinical trial of the future. However, clear validation guidelines are lacking in the literature. This paper contains pragmatic criteria for the selection, technical validation, and clinical validation of novel digital endpoints and provides recommendations for future work and collaboration.

Microhardness of composite materials with different organic phases in deep class II cavities: an in vitro study.

OBJECTIVE: This study compared the microhardness of three composite resins with different organic matrices in deep class II cavities. MATERIALS AND METHOD: A total of 36 extracted molars were randomly assigned to six groups and standardized class II cavities were prepared. The cavity design comprised three steps in a mesiodistal direction with an increasing depth (2, 4, and 6 mm). Twelve cavities each were restored using Filtek Supreme (FS), Quixfil (QF), and Filtek Silorane (SI). The materials were applied in incremental layers of 2 mm and cured either with Halogen Translux Energy (HTE) (n=18) or LED Bluephase C8 (LED) (n=18). Subsequently, the specimens were cross-sectioned, and microhardness was determined in various depths and at two different distances from the matrix. RESULTS: QF yielded the highest KHN microhardness values (92.67 ± 12.77), followed by FS (65.53 ± 19.52) and SI (57.67 ± 8.33). Composites cured with LED achieved higher KHN values. All materials showed the highest microhardness values within the superficial increments and at a distance of 1000 µm from the matrix.

Ovarian cancer and oral contraceptives: collaborative reanalysis of data from 45 epidemiological studies including 23,257 women with ovarian cancer and 87,303 controls.

BACKGROUND: Oral contraceptives were introduced almost 50 years ago, and over 100 million women currently use them. Oral contraceptives can reduce the risk of ovarian cancer, but the eventual public-health effects of this reduction will depend on how long the protection lasts after use ceases. We aimed to assess these effects. METHODS: Individual data for 23,257 women with ovarian cancer (cases) and 87,303 without ovarian cancer (controls) from 45 epidemiological studies in 21 countries were checked and analysed centrally. The relative risk of ovarian cancer in relation to oral contraceptive use was estimated, stratifying by study, age, parity, and hysterectomy. FINDINGS: Overall 7308 (31%) cases and 32,717 (37%) controls had ever used oral contraceptives, for average durations among users of 4.4 and 5.0 years, respectively. The median year of cancer diagnosis was 1993, when cases were aged an average of 56 years. The longer that women had used oral contraceptives, the greater the reduction in ovarian cancer risk (p<0.0001). This reduction in risk persisted for more than 30 years after oral contraceptive use had ceased but became somewhat attenuated over time-the proportional risk reductions per 5 years of use were 29% (95% CI 23-34%) for use that had ceased less than 10 years previously, 19% (14-24%) for use that had ceased 10-19 years previously, and 15% (9-21%) for use that had ceased 20-29 years previously. Use during the 1960s, 1970s, and 1980s was associated with similar proportional risk reductions, although typical oestrogen doses in the 1960s were more than double those in the 1980s. The incidence of mucinous tumours (12% of the total) seemed little affected by oral contraceptives, but otherwise the proportional risk reduction did not vary much between different histological types. In high-income countries, 10 years use of oral contraceptives was estimated to reduce ovarian cancer incidence before age 75 from 1.2 to 0.8 per 100 users and mortality from 0.7 to 0.5 per 100; for every 5000 woman-years of use, about two ovarian cancers and one death from the disease before age 75 are prevented. INTERPRETATION: Use of oral contraceptives confers long-term protection against ovarian cancer. These findings suggest that oral contraceptives have already prevented some 200,000 ovarian cancers and 100,000 deaths from the disease, and that over the next few decades the number of cancers prevented will rise to at least 30,000 per year.

Leukaemia and aplastic anaemia in patients irradiated for ankylosing spondylitis. 1957.

Reported here is a seminal study of leukaemia among patients who had been treated with x-rays for ankylosing spondylitis. The findings were first published in a Medical Research Council (MRC) report in 1957. The report is now published in a scientific journal for the first time, 50 years after it first appeared.

Recent trends in cancer mortality in the UK.

Comparisons of survival rates of given diseases with different treatments or in different places often gave misleading results until the introduction of controlled trials. Recent reports of relatively low survival rates following the treatment of cancer in the UK compared to the rates in other countries, not based on controlled trials, may consequently be misleading. Their validity has, therefore, been tested by comparing the levels and trends in mortality--the ultimate criterion by which the success or failure of any system of care can be judged. For this purpose, rates and trends in rates over 20-50 years have been compared in five European countries of similar economic status (France, Italy, the Netherlands, Sweden, and the UK). The UK rates are not generally worse than those in the other countries and are sometimes better. Exceptions were cancer of the lung, large bowel, and breast, the first of which is explained by differences in the prevalence of smoking.

Mortality experience of male workers at a UK tin smelter.

BACKGROUND: Between 1937 and 1991, Capper Pass and Sons Limited operated a tin smelter complex in North Humberside, UK, at which employees were potentially exposed to a number of substances, including lead, arsenic, cadmium and natural series radionuclides. Decommissioning and site clearance continued until 1995. Between 1967 and 1995 the company was a subsidiary of Rio Tinto plc. AIMS: The aim was to identify any significant excess, or deficits, in mortality among former employees that might be attributable to factors associated with occupation. METHODS: We defined a cohort of 1462 males who had been employed for at least 12 months between 1/11/1967 and 28/7/1995, followed-up through to 31/12/2001. The mortality of the cohort was compared against that expected for both national and regional populations. RESULTS: Mortality from all causes and all cancers did not differ from that expected. Mortality from ischaemic heart disease showed a deficit and mortality from lung cancer showed a statistically significant excess. Mortality from smoking related diseases other than lung cancer showed a non-significant deficit. CONCLUSIONS: The pattern of lung cancer mortality is consistent with the hypothesis that the risk of lung cancer has been enhanced by occupational exposure to one or more carcinogens, the effect of which diminishes with time since exposure. The deficit in ischaemic heart disease may be attributed to a protective effect from manual labour. The results provide no evidence for attribution of other excess or deficits in mortality to factors associated with employment.