Tag: statistics

AI Ethics. 2021 Nov 29:1-9. doi: 10.1007/s43681-021-00118-4. Online ahead of print.

ABSTRACT

Today Artificial Intelligence (AI) supports difficult decisions about policy, health, and our personal lives. The AI algorithms we develop and deploy to make sense of information, are informed by data, and based on models that capture and use pertinent details of the population or phenomenon being analyzed. For any application area, more importantly in precision medicine which directly impacts human lives, the data upon which algorithms are run must be procured, cleaned, and organized well to assure reliable and interpretable results, and to assure that they do not perpetrate or amplify human prejudices. This must be done without violating basic assumptions of the algorithms in use. Algorithmic results need to be clearly communicated to stakeholders and domain experts to enable sound conclusions. Our position is that AI holds great promise for supporting precision medicine, but we need to move forward with great care, with consideration for possible ethical implications. We make the case that a no-boundary or convergent approach is essential to support sound and ethical decisions. No-boundary thinking supports problem definition and solving with teams of experts possessing diverse perspectives. When dealing with AI and the data needed to use AI, there is a spectrum of activities that needs the attention of a no-boundary team. This is necessary if we are to draw viable conclusions and develop actions and policies based on the AI, the data, and the scientific foundations of the domain in question.

PMID:34870283 | PMC:PMC8628283 | DOI:10.1007/s43681-021-00118-4

Commun Med (London). 2021;1(1):4. doi: 10.1038/s43856-021-00007-1. Epub 2021 Jun 30.

ABSTRACT

BACKGROUND: In early March 2020, a SARS-CoV-2 outbreak in the ski resort Ischgl in Austria initiated the spread of SARS-CoV-2 throughout Austria and Northern Europe.

METHODS: Between April 21^st and 27^th 2020, a cross-sectional epidemiologic study targeting the full population of Ischgl (n = 1867), of which 79% could be included (n = 1473, incl. 214 children), was performed. For each individual, the study involved a SARS-CoV-2 PCR, antibody testing and structured questionnaires. A mathematical model was used to help understand the influence of the determined seroprevalence on virus transmission.

RESULTS: The seroprevalence was 42.4% (95% confidence interval (CI) 39.8-44.7). Individuals under 18 showed a significantly lower seroprevalence of 27.1% (95% CI 21.3-33.6) than adults (45%; 95% CI 42.2-47.7; OR of 0.455, 95% CI 0.356-0.682, p < 0.001). Of the seropositive individuals, 83.7% had not been diagnosed to have had SARS-CoV-2 infection previously. The clinical course was generally mild. Over the previous two months, two COVID-19-related deaths had been recorded, corresponding to an infection fatality rate of 0.25% (95% CI 0.03-0.91). Only 8 (0.5 %) individuals were newly diagnosed to be infected with SARS-CoV-2 during this study.

CONCLUSIONS: Ischgl was hit early and hard by SARS-CoV-2 leading to a high local seroprevalence of 42.4%, which was lower in individuals below the age of 18 than in adults. Mathematical modeling suggests that a drastic decline of newly infected individuals in Ischgl by the end of April occurred due to the dual impact from the non-pharmacological interventions and a high immunization of the Ischgl population.

PMID:34870284 | PMC:PMC8633917 | DOI:10.1038/s43856-021-00007-1

Lancet Reg Health Am. 2022 Feb;6:100102. doi: 10.1016/j.lana.2021.100102. Epub 2021 Nov 3.

ABSTRACT

BACKGROUND: Brazil has faced two simultaneous problems related to respiratory health: forest fires and the high mortality rate due to COVID-19 pandemics. The Amazon rain forest is one of the Brazilian biomes that suffers the most with fires caused by droughts and illegal deforestation. These fires can bring respiratory diseases associated with air pollution, and the State of Pará in Brazil is the most affected. COVID-19 pandemics associated with air pollution can potentially increase hospitalizations and deaths related to respiratory diseases. Here, we aimed to evaluate the association of fire occurrences with the COVID-19 mortality rates and general respiratory diseases hospitalizations in the State of Pará, Brazil.

METHODS: We employed machine learning technique for clustering k-means accompanied with the elbow method used to identify the ideal quantity of clusters for the k-means algorithm, clustering 10 groups of cities in the State of Pará where we selected the clusters with the highest and lowest fires occurrence from the 2015 to 2019. Next, an Auto-regressive Integrated Moving Average Exogenous (ARIMAX) model was proposed to study the serial correlation of respiratory diseases hospitalizations and their associations with fire occurrences. Regarding the COVID-19 analysis, we computed the mortality risk and its confidence level considering the quarterly incidence rate ratio in clusters with high and low exposure to fires.

FINDINGS: Using the k-means algorithm we identified two clusters with similar DHI (Development Human Index) and GDP (Gross Domestic Product) from a group of ten clusters that divided the State of Pará but with diverse behavior considering the hospitalizations and forest fires in the Amazon biome. From the auto-regressive and moving average model (ARIMAX), it was possible to show that besides the serial correlation, the fires occurrences contribute to the respiratory diseases increase, with an observed lag of six months after the fires for the case with high exposure to fires. A highlight that deserves attention concerns the relationship between fire occurrences and deaths. Historically, the risk of mortality by respiratory diseases is higher (about the double) in regions and periods with high exposure to fires than the ones with low exposure to fires. The same pattern remains in the period of the COVID-19 pandemic, where the risk of mortality for COVID-19 was 80% higher in the region and period with high exposure to fires. Regarding the SARS-COV-2 analysis, the risk of mortality related to COVID-19 is higher in the period with high exposure to fires than in the period with low exposure to fires. Another highlight concerns the relationship between fire occurrences and COVID-19 deaths. The results show that regions with high fire occurrences are associated with more cases of COVID deaths.

INTERPRETATION: The decision-make process is a critical problem mainly when it involves environmental and health control policies. Environmental policies are often more cost-effective as health measures than the use of public health services. This highlight the importance of data analyses to support the decision making and to identify population in need of better infrastructure due to historical environmental factors and the knowledge of associated health risk. The results suggest that The fires occurrences contribute to the increase of the respiratory diseases hospitalization. The mortality rate related to COVID-19 was higher for the period with high exposure to fires than the period with low exposure to fires. The regions with high fire occurrences is associated with more COVID-19 deaths, mainly in the months with high number of fires.

FUNDING: No additional funding source was required for this study.

PMID:34870262 | PMC:PMC8632600 | DOI:10.1016/j.lana.2021.100102

Front Digit Health. 2021 Nov 16;3:727504. doi: 10.3389/fdgth.2021.727504. eCollection 2021.

ABSTRACT

Mobile measures of human circadian rhythms (CR) are needed in the age of chronotherapy. Two wearable measures of CR have recently been validated: one that uses heart rate to extract circadian rhythms that originate in the sinoatrial node of the heart, and another that uses activity to predict the laboratory gold standard and central circadian pacemaker marker, dim light melatonin onset (DLMO). We first find that the heart rate markers of normal real-world individuals align with laboratory DLMO measurements when we account for heart rate phase error. Next, we expand upon previous work that has examined sleep patterns or chronotypes during the COVID-19 lockdown by studying the effects of social distancing on circadian rhythms. In particular, using data collected from the Social Rhythms app, a mobile application where individuals upload their wearable data and receive reports on their circadian rhythms, we compared the two circadian phase estimates before and after social distancing. Interestingly, we found that the lockdown had different effects on the two ambulatory measurements. Before the lockdown, the two measures aligned, as predicted by laboratory data. After the lockdown, when circadian timekeeping signals were blunted, these measures diverged in 70% of subjects (with circadian rhythms in heart rate, or CRHR, becoming delayed). Thus, while either approach can measure circadian rhythms, both are needed to understand internal desynchrony. We also argue that interventions may be needed in future lockdowns to better align separate circadian rhythms in the body.

PMID:34870267 | PMC:PMC8634937 | DOI:10.3389/fdgth.2021.727504

Lancet Reg Health Eur. 2022 Jan;12:100267. doi: 10.1016/j.lanepe.2021.100267. Epub 2021 Nov 30.

ABSTRACT

BACKGROUND: Countries in the World Health Organization (WHO) European Region differ in terms of the COVID-19 vaccine supply conditions. We evaluated the health and economic impact of different age-based vaccine prioritisation strategies across this demographically and socio-economically diverse region.

METHODS: We fitted age-specific compartmental models to the reported daily COVID-19 mortality in 2020 to inform the immunity level before vaccine roll-out. Models capture country-specific differences in population structures, contact patterns, epidemic history, life expectancy, and GDP per capita.We examined four strategies that prioritise: all adults (V+), younger (20-59 year-olds) followed by older adults (60+) (V20), older followed by younger adults (V60), and the oldest adults (75+) (V75) followed by incrementally younger age groups. We explored four roll-out scenarios (R1-4) – the slowest scenario (R1) reached 30% coverage by December 2022 and the fastest (R4) 80% by December 2021. Five decision-making metrics were summarised over 2021-22: mortality, morbidity, and losses in comorbidity-adjusted life expectancy, comorbidity- and quality-adjusted life years, and human capital. Six vaccine profiles were tested – the highest performing vaccine has 95% efficacy against both infection and disease, and the lowest 50% against diseases and 0% against infection.

FINDINGS: Of the 20 decision-making metrics and roll-out scenario combinations, the same optimal strategy applied to all countries in only one combination; V60 was more or similarly desirable than V75 in 19 combinations. Of the 38 countries with fitted models, 11-37 countries had variable optimal strategies by decision-making metrics or roll-out scenarios. There are greater benefits in prioritising older adults when roll-out is slow and when vaccine profiles are less favourable.

INTERPRETATION: The optimal age-based vaccine prioritisation strategies were sensitive to country characteristics, decision-making metrics, and roll-out speeds. A prioritisation strategy involving more age-based stages (V75) does not necessarily lead to better health and economic outcomes than targeting broad age groups (V60). Countries expecting a slow vaccine roll-out may particularly benefit from prioritising older adults.

FUNDING: World Health Organization, Bill and Melinda Gates Foundation, the Medical Research Council (United Kingdom), the National Institute of Health Research (United Kingdom), the European Commission, the Foreign, Commonwealth and Development Office (United Kingdom), Wellcome Trust.

PMID:34870256 | PMC:PMC8629724 | DOI:10.1016/j.lanepe.2021.100267

Cell Genom. 2021 Oct 13;1(1):100005. doi: 10.1016/j.xgen.2021.100005.

ABSTRACT

Genome sequencing has recently become a viable genotyping technology for use in genome-wide association studies (GWASs), offering the potential to analyze a broader range of genome-wide variation, including rare variants. To survey current standards, we assessed the content and quality of reporting of statistical methods, analyses, results, and datasets in 167 exome- or genome-wide-sequencing-based GWAS publications published from 2014 to 2020; 81% of publications included tests of aggregate association across multiple variants, with multiple test models frequently used. We observed a lack of standardized terms and incomplete reporting of datasets, particularly for variants analyzed in aggregate tests. We also find a lower frequency of sharing of summary statistics compared with array-based GWASs. Reporting standards and increased data sharing are required to ensure sequencing-based association study data are findable, interoperable, accessible, and reusable (FAIR). To support that, we recommend adopting the standard terminology of sequencing-based GWAS (seqGWAS). Further, we recommend that single-variant analyses be reported following the same standards and conventions as standard array-based GWASs and be shared in the GWAS Catalog. We also provide initial recommended standards for aggregate analyses metadata and summary statistics.

PMID:34870259 | PMC:PMC8637874 | DOI:10.1016/j.xgen.2021.100005

Lancet Reg Health Eur. 2022 Jan;12:100265. doi: 10.1016/j.lanepe.2021.100265. Epub 2021 Nov 30.

ABSTRACT

BACKGROUND: To limit the spread of COVID-19 in March 2020, the population of England was instructed to stay home, leaving only for essential shopping, health-care, work, or exercise. The impact on population activity behaviours is not clear. We describe changes in duration and types of activity undertaken by adults ≥16 years in England between March and May 2016-19 and 2020, by socio-demographic strata.

METHODS: Using nationally representative data collected between November 2015 and May 2020 by the Sport England Active Lives Surveys (n=726,257) we assessed trends in amount and type of non-occupational moderate-to-vigorous physical activity. Using data from n=74,430 mid-April to mid-May respondents, we then estimated the odds ratios of reporting any activity in the four-week recall period in 2020 compared to 2016-19. Gamma regressions estimated the mean ratios (MR) of duration amongst those reporting any activity in 2020 compared to 2016-19.

FINDINGS: Population activity declined substantially after the restrictions were introduced. Compared to 2016-19 levels, the odds of reporting any activity in 2020 were 30% lower (95% confidence interval (CI) 26-34%). The largest declines were amongst non-white ethnicities, the youngest and oldest age groups, and the unemployed; no socio-demographic subgroup had higher odds. Amongst those undertaking activity, weekly duration was similar in the two periods (MR 0.99, 95%CI (0.96-1.01%)). The odds of participating in walking for leisure and gardening were 11% (6-16%) and 15% (9-21%) higher, respectively, whereas the odds for team and racket sport and walking for travel participation were 76% (73-79%) and 66% (64-68%) lower, respectively.

INTERPRETATION: Restrictions introduced in Spring 2020 likely reduced physical activity levels in England. The magnitude of the declines were not uniform by demographic groups or by activity type, which future policies should consider.

FUNDING: TS, KW, SJS, and SB are supported by UK Medical Research Council [grant numbers MC_UU_00006/4 and MC_UU_12015/3] and SB is supported by the NIHR Biomedical Research Centre in Cambridge (IS-BRC-1215-20014).

PMID:34870255 | PMC:PMC8629728 | DOI:10.1016/j.lanepe.2021.100265

Tech Innov Gastrointest Endosc. 2021;23(4):297-303. doi: 10.1016/j.tige.2021.06.006. Epub 2021 Jun 24.

ABSTRACT

BACKGROUND: Excess post-operative opioid medication use can delay recovery and is associated with long-term misuse, addiction, and overdose. We aimed to explore the effect of pre-procedural thoracic paravertebral nerve block (PNB) on pain-related outcomes after POEM.

METHODS: In this retrospective cohort study, consecutive patients who did and did not receive a PNB prior to POEM were compared. The outcomes were peak and cumulative pain scores, total opioid use during hospitalization, and length of stay. After adjusting for confounders, the associations between nerve block and the outcomes of interest were explored.

RESULTS: Forty-nine consecutive patients were enrolled; 25 patients received a block whereas the subsequent 24 did not. There were no differences in baseline characteristics between the study groups. In unadjusted analyses, there was no significant difference between patients who did and did not undergo PNB in peak pain score (7.8 vs. 8.7, p=0.14), cumulative pain score in the first 12 hours (area under curve 66.5 vs. 75.8, p=0.22), median total opioid use (38.9 mg morphine equivalent dosing vs. 42, p=1.00), and median length of hospitalization (26.5 hours vs. 24, p=0.35). In multivariable regression models, PNB was not associated with a reduction in pain scores, opioid use, or hospitalization. There were no adverse events related to the block.

CONCLUSIONS: In this exploratory, observational study, paravertebral nerve block immediately before POEM did not result in a statistically significant reduction in pain-related outcomes or hospitalization. Additional observational studies may elucidate whether higher anesthetic doses or longer acting formulations would be of value.

PMID:34870251 | PMC:PMC8635293 | DOI:10.1016/j.tige.2021.06.006

Public Health Pract (Oxf). 2021 Nov;2:100219. doi: 10.1016/j.puhip.2021.100219. Epub 2021 Nov 27.

ABSTRACT

OBJECTIVES: In the past months, many countries have adopted varying degrees of lockdown restrictions to control the spread of the COVID-19 virus. According to the existing literature, some consequences of lockdown restrictions on people’s lives are beginning to emerge yet the evolution of such consequences in relation to the time spent in lockdown is understudied. To inform policies involving lockdown restrictions, this study adopted a data-driven Machine Learning approach to uncover the short-term time-related effects of lockdown on people’s physical and mental health.

STUDY DESIGN: An online questionnaire was launched on 17 April 2020, distributed through convenience sampling and was self-completed by 2,276 people from 66 different countries.

METHODS: Focusing on the UK sample (N = 325), 12 aggregated variables representing the participant’s living environment, physical and mental health were used to train a RandomForest model to estimate the week of survey completion.

RESULTS: Using an index of importance, Self-Perceived Loneliness was identified as the most influential variable for estimating the time spent in lockdown. A significant U-shaped curve emerged for loneliness levels, with lower scores reported by participants who took part in the study during the 6th lockdown week (p = 0.009). The same pattern was replicated in the Greek sample (N = 137) for week 4 (p = 0.012) and 6 (p = 0.009) of lockdown.

CONCLUSIONS: From the trained Machine Learning model and the subsequent statistical analysis, Self-Perceived Loneliness varied across time in lockdown in the UK and Greek populations, with lower symptoms reported during the 4th and 6th lockdown weeks. This supports the dissociation between social support and loneliness, and suggests that social support strategies could be effective even in times of social isolation.

PMID:34870253 | PMC:PMC8626633 | DOI:10.1016/j.puhip.2021.100219

Arch Acad Emerg Med. 2021 Oct 13;9(1):e67. doi: 10.22037/aaem.v9i1.1325. eCollection 2021.

ABSTRACT

INTRODUCTION: Red blood cell distribution width (RDW) has been introduced as a predictive factor for mortality in several critical illnesses and infectious diseases. This study aimed to assess the possible relationship between RDW on admission and COVID-19 in-hospital mortality.

METHOD: This cross-sectional study was performed using the Isfahan COVID-19 registry. Adult confirmed cases of COVID-19 admitted to four hospitals affiliated with Isfahan University of Medical Sciences in Iran were included. Age, sex, O2 saturation, RDW on admission, Intensive Care Unit admission, laboratory data, history of comorbidities, and hospital outcome were extracted from the registry. Cox proportional hazard regression was used to study the independent association of RDW with mortality.

RESULTS: 4152 patients with the mean age of 61.1 ± 16.97 years were included (56.2% male). 597 (14.4%) cases were admitted to intensive care unit (ICU) and 477 (11.5%) cases died. The mortality rate of patients with normal and elevated RDW was 7.8% and 21.2%, respectively (OR= 3.1, 95%CI: 2.6-3.8), which remained statistically significant after adjusting for age, O2 saturation, comorbidities, and ICU admission (2.03, 95% CI: 1.68-2.44). Moreover, elevated RDW mortality Hazard Ratio in patients who were not admitted to ICU was higher than ICU-admitted patients (3.10, 95% CI: 2.35-4.09 vs. 1.47, 95% CI: 1.15-1.88, respectively).

CONCLUSION: The results support the presence of an association between elevated RDW and mortality in patients with COVID-19, especially those who were not admitted to ICU. It seems that elevated RDW can be used as a predictor of mortality in COVID-19 cases.

PMID:34870233 | PMC:PMC8628640 | DOI:10.22037/aaem.v9i1.1325