Tag: statistics

J Stroke Cerebrovasc Dis. 2021 Sep 18;30(12):106108. doi: 10.1016/j.jstrokecerebrovasdis.2021.106108. Online ahead of print.

ABSTRACT

BACKGROUND AND PURPOSE: Post-stroke cognitive impairment is a common and well-known consequence of supra-tentorial infarct, but its prevalence and severity after infra-tentorial infarct is unclear. We compared the frequencies and prognostic value of domain-specific cognitive deficits after supra-tentorial and infra-tentorial infarct.

METHODS: In a consecutive cohort of patients with first-ever stroke (N = 244) admitted to Helsinki University Hospital, 37 patients had an infra-tentorial infarct. Patients were assessed with a comprehensive neuropsychological examination 3 months post-stroke covering 9 cognitive domains and functional disability was assessed at 15 months with the modified Rankin Scale.

RESULTS: There were no statistically significant differences between the frequencies of cognitive deficits in patients with infra-tentorial vs supra-tentorial infarct. Altogether 73% of patients with infra-tentorial infarct and 82% of patients with supra-tentorial infarct had impairment in at least one cognitive domain. Further 42% of patients with infra-tentorial infarct and 47% of those with supra-tentorial infarct had deficits in 3 or more cognitive domains. In patients with infra-tentorial infarct, visuo-constructional deficits were significantly associated with functional disability at 15 months (OR 9.0, 95%CI 1.3-62.5, p = 0.027). In patients with supratentorial infarct, executive deficits (OR 2.9, 95%CI 1.5-5.8, p = 0.002) and visuo-constructional deficits (OR 2.9, 95%CI 1.5-5.7, p = 0.001) showed associations with functional disability at 15 months.

CONCLUSION: Cognitive deficits are as common in patients with infra-tentorial infarct as in those with supra-tentorial infarct, and it is important to recognize them to meet the needs of rehabilitation.

PMID:34547675 | DOI:10.1016/j.jstrokecerebrovasdis.2021.106108

Int J Med Inform. 2021 Sep 13;155:104572. doi: 10.1016/j.ijmedinf.2021.104572. Online ahead of print.

ABSTRACT

PURPOSE: Femoral neck fracture is a frequent cause of hospitalization, and length of stay is an important marker of hospital cost and quality of care provided. As an extension of traditional statistical methods, machine learning provides the possibility of accurately predicting the length of hospital stay. The aim of this paper is to retrospectively identify predictive factors of the length of hospital stay (LOS) and predict the postoperative LOS by using machine learning algorithms.

METHOD: Based on the admission and perioperative data of the patients, linear regression was used to analyze the predictive factors of the LOS. Multiple machine learning models were developed, and the performance of different models was compared.

RESULT: Stepwise linear regression showed that preoperative calcium level (P = 0.017) and preoperative lymphocyte percentage (P = 0.007), in addition to intraoperative bleeding (p = 0.041), glucose and sodium chloride infusion after surgery (P = 0.019), Charlson Comorbidity Index (p = 0.007) and BMI (P = 0.031), were significant predictors of LOS. The best performing model was the principal component regression (PCR) with an optimal MAE (1.525) and a proportion of prediction error within 3 days of 90.91%.

CONCLUSION: Excessive intravenous glucose and sodium chloride infusion after surgery, preoperative hypocalcemia, preoperative high percentages of lymphocytes, excessive intraoperative bleeding, lower BMI and higher CCI scores were related to prolonged LOS by using linear regression. Machine learning could accurately predict the postoperative LOS. This information allows hospital administrators to plan reasonable resource allocation to fulfill demand, leading to direct care quality improvement and more reasonable use of scarce resources.

PMID:34547625 | DOI:10.1016/j.ijmedinf.2021.104572

Clin Neurol Neurosurg. 2021 Sep 10;209:106942. doi: 10.1016/j.clineuro.2021.106942. Online ahead of print.

ABSTRACT

BACKGROUND: Pituitary adenomas (PAs) are among the most common brain tumors which characteristically become symptomatic due to the mass effect on surrounding structures and/or hormonal imbalances. This study describes 28 cases of PAs with an extrasellar invasive growth (EIG) at the early stage of tumor growth with normal sellar cavity size.

METHODS: 1200 cases of PAs either treated medically or via Endoscopic transsphenoidal surgery were reviewed during a 10-year period. Pre-operative imaging was analyzed to evaluate the tumor expansion pattern, tumor invasion sites and other relevant tumor properties. A comprehensive preoperative endocrinological along with postoperative histopathological studies were conducted to evaluate the subjects’ homeostasis and further identify the lesions characteristics.

RESULTS: Of the 28 patients, patients 19, 2, 1 and 6 had a growth hormone (GH)-secreting PA, an adrenocorticotropic hormone (ACTH)-secreting PA, a nonfunctional PA (NFPA) and a mix-hormones secreting PA, respectively. There was a statistically significant difference between patients with and without EIG regarding types of PAs (P = 0.000). Post-hoc tests demonstrated that GH-secreting PAs (P = 0.0003) and mix-hormones secreting PAs (P = 0.0000) are significantly more invasive, while NFPAs (P = 0.0000) are less invasive. There was not a statistically significant difference between the invasion sites and different types of PAs (P = 0.122).

CONCLUSION: Among different histological subtypes of PAs, GH-secreting PAs revealed a remarkable tendency for EIG with normal sellar cavity size. The extra-sellar tumor extension with a normal sella did not correlate with atypical histology. Considering EIG patterns, surgeons should be vigilant to investigate and follow the tumor spreading to its enclosing boundary during surgery, the clival part of sphenoid bone should be more exposed, and both inferomedial and lateral borders of the sphenoid sinus should be carefully explored in order to remove the tumor.

PMID:34547643 | DOI:10.1016/j.clineuro.2021.106942

Curr Opin Chem Biol. 2021 Sep 18;65:136-144. doi: 10.1016/j.cbpa.2021.08.004. Online ahead of print.

ABSTRACT

Since the first revelation of proteins functioning as macromolecular machines through their three dimensional structures, researchers have been intrigued by the marvelous ways the biochemical processes are carried out by proteins. The aspiration to understand protein structures has fueled extensive efforts across different scientific disciplines. In recent years, it has been demonstrated that proteins with new functionality or shapes can be designed via structure-based modeling methods, and the design strategies have combined all available information – but largely piece-by-piece – from sequence derived statistics to the detailed atomic-level modeling of chemical interactions. Despite the significant progress, incorporating data-derived approaches through the use of deep learning methods can be a game changer. In this review, we summarize current progress, compare the arc of developing the deep learning approaches with the conventional methods, and describe the motivation and concepts behind current strategies that may lead to potential future opportunities.

PMID:34547592 | DOI:10.1016/j.cbpa.2021.08.004

Int J Hyg Environ Health. 2021 Sep 18;238:113849. doi: 10.1016/j.ijheh.2021.113849. Online ahead of print.

ABSTRACT

To improve our understanding of internal exposure to multiple chemicals, the concept exposure load (EL) was used on human biomonitoring (HBM) data of the 4th FLEHS (Flemish Environment and Health Study; 2016-2020). The investigated chemicals were per- and polyfluoroalkyl substances (PFASs), bisphenols, phthalates and alternative plasticizers, flame retardants, pesticides, toxic metals, organochlorine compounds and polycyclic aromatic hydrocarbons (PAHs). The EL calculates “the number of chemicals to which individuals are internally exposed above a predefined threshold”. In this study, the 50th and 90th percentile of each of the 45 chemicals were applied as thresholds for the EL calculations for 387 study participants. Around 20% of the participants were exposed to >27 chemicals above the P50 and to >6 chemicals above the P90 level. This shows that participants can be internally exposed to multiple chemicals in relatively high concentrations. When the chemical composition of the EL was considered, the variability between individuals was driven by some chemicals more than others. The variability of the chemical profiles at high exposure loads (EL-P90) was somewhat dominated by e.g. organochlorine chemicals, PFASs, phthalates, PAHs, organophosphate flame retardants, bisphenols (A & F), pesticides, metals, but to a lesser extent by brominated flame retardants, the organophosphorus flame retardants TCIPP & TBOEP, naphthalene and benzene, bisphenols S, B & Z, the pesticide 2,4-D, the phthalate DEP and alternative plasticizer DINCH. Associations between the EL and exposure determinants suggested determinants formerly associated with fat soluble chemicals, PFASs, bisphenols, and PAHs. This information adds to the knowledge needed to reduce the exposure by policymakers and citizens. However, a more in depth study is necessary to explore in detail the causes for the higher EL in some individuals. Some limitations in the EL concept are that a binary number is used for exposure above or below a threshold, while toxicity and residence time in the body are not accounted for and the sequence of exposure in different life stages is unknown. However, EL is a first useful step to get more insight in multiple chemical exposure in higher exposed subpopulations (relative to the rest of the sampled population).

PMID:34547602 | DOI:10.1016/j.ijheh.2021.113849

World Neurosurg. 2021 Sep 18:S1878-8750(21)01376-0. doi: 10.1016/j.wneu.2021.09.038. Online ahead of print.

ABSTRACT

BACKGROUND: Superior semicircular canal dehiscence (SSCD) is the appearance of a third mobile window between the middle fossa and the superior semicircular canal. Surgical management is indicated in patients with persistent and debilitating symptoms. The purpose of this study was to evaluate the association between preoperative variables that may impact postoperative symptomatic resolution.

METHODS: A single-institution retrospective analysis was performed on patients who were surgically treated for SSCD. Patients were divided to different cohorts based on unilateral or bilateral nature of the disease. A p-value <0.05 was considered statistically significant.

RESULTS: A total of 229 surgical repairs were analyzed. Mean age was 51 years (± 7.8 years), and females made up 55.9% of patients. All cohorts were similar with respect to baseline demographics. The most commonly reported preoperative symptoms were tinnitus, dizziness and autophony. The greatest symptomatic resolution was seen in autophony, internal sound amplification, hyperacusis, and oscillopsia. The unilateral SSCD cohort had significantly higher improvement of autophony (p=0.003), aural fullness (p=0.05), tinnitus (p=0.006), hearing loss (p=0.02), dizziness (p=0.006), and headache (p=0.007), compared to the bilateral SSCD cohorts. Among patients with bilateral disease, those with unilateral surgery reported greater symptomatic resolution with respect to hyperacusis (p=0.03), hearing loss (p=0.02), dizziness (p=0.03), and disequilibrium (p<0.001), than those with bilateral operations.

CONCLUSION: Surgical management of SSCD leads to high rates of postoperative symptomatic improvement. Patients with unilateral SSCD benefit greater symptomatic resolution compared to those with bilateral pathology.

PMID:34547526 | DOI:10.1016/j.wneu.2021.09.038

Soc Sci Med. 2021 Sep 17;289:114406. doi: 10.1016/j.socscimed.2021.114406. Online ahead of print.

ABSTRACT

The United States (U.S.) has one of the highest cesarean rates in the world yet little research considers structural factors, like racism and sexism, associated with the higher than recommended cesarean rate. New research operationalizes and quantifies structural sexism across U.S. states, which allows for consideration of how social norms and values around women and their bodies relate to the overmedicalization of birth through cesarean sections. We obtained restricted natality data for 2018 from the U.S. National Center for Health Statistics. In 2018, among people 15-49 years, 987,187 births fit the criteria for low-risk of cesarean section. Structural sexism scores were derived from 6 elements covering economic, political, cultural, and physical arenas that were totaled and standardized to create an aggregate index for each state and DC (scores range from -1.06 to 1.4). Using multivariable logistic and multilevel mixed effects logistic regression models, we examined the associations between structural sexism and low-risk cesarean section for all fifty states and the District of Columbia, controlling for relevant confounders. We found that structural sexism in 2018 was highest in historically religious mountain states and the South. Nationally, the low-risk cesarean rate was 25.1%. Multilevel models show that people living in states with higher structural sexism scores were more likely to have a cesarean section (OR = 1.22, 95% CI: 1.07-1.39). Structural sexism is related to low-risk cesarean rates in U.S., providing evidence that social ideas and norms about women and their bodies are related to overmedicalization of birth. Health policymakers, providers and scholars should pay attention to structural drivers, including structural sexism, as a factor that affects overmedicalization of birth and subsequent health outcomes for pregnant people and their infants.

PMID:34547543 | DOI:10.1016/j.socscimed.2021.114406

J Sport Health Sci. 2021 Sep 18:S2095-2546(21)00103-4. doi: 10.1016/j.jshs.2021.09.006. Online ahead of print.

ABSTRACT

OBJECTIVE: Participation in sports is associated with many benefits to all aspects of health; however, it also come with the risk of injury, particularly concussions. Self-disclosure and care-seeking following a concussion are especially important due to the lack of outwardly visible signs and/or symptoms. While recent research has explored factors affecting concussion disclosure, utilization of isolated methodologies limits the ability to contextualize how disclosure or non-disclosure occurs. Therefore, the purpose of this study was to describe the factors and expectations of National Collegiate Athletic Association (NCAA) athletes that may influence concussion disclosure.

METHODS: This mixed-methods convergent parallel research study included 25 NCAA Division I athletes representing 13 sports, all of whom completed a concussion education session with pre-/post-test surveys and a semi-structured interview. Eligible athletes were at least 18 years old and on an NCAA roster. The surveys focused on previous concussion-related disclosure behaviors, knowledge, attitudes, beliefs, norms, and intentions toward disclosing concussion. Interviews focused on the athletes’ experiences related to concussion disclosure. Survey data were analyzed using descriptive statistics and Mann-Whitney U tests. Interviews were analyzed using a Consensual Qualitative Research (CQR) tradition.

RESULTS: Participants had high concussion knowledge (median = 46), positive attitudes (median = 38), strong beliefs (median = 13), and high intentions to disclose concussion symptoms (median = 7). None of the constructs differed by participant gender. Although quantitative findings were mostly positive, interview data highlighted factors that may explain why some participants are successful in disclosing concussions and why others may find disclosure difficult. Educational efforts, sport culture, and medical professional presence were the primary facilitators discussed by participants. Stigma, pressure, and a lack of team support were perceived as disclosure barriers.

CONCLUSION: The context in which concussion disclosure occurs, or does not occur, is vital to the success of educational interventions. Interventions must prioritize stakeholder and team-based perspectives on concussion to establish a network supportive to disclosure.

PMID:34547482 | DOI:10.1016/j.jshs.2021.09.006

Bone. 2021 Sep 18:116208. doi: 10.1016/j.bone.2021.116208. Online ahead of print.

ABSTRACT

BACKGROUND: Bone mineral density (BMD) loss and fat gain is common in people living with HIV (PLWH), particularly after initiating combination antiretroviral therapy (cART). Given the close metabolic interaction between bone and fat, we tested the hypotheses that changes in bone-derived hormones are associated with fat accumulation and changes in fat-derived hormones are associated with BMD loss following cART initiation.

METHODS: HIV-seropositive subjects (n=15) initiating fixed dose cART of tenofovir disoproxil fumarate/emtricitabine/efavirenz (TDF/FTC/EFV) underwent dual x-ray absorptiometry (DXA) assessment pre-cART and again 12-months post-cART initiation. DXA-derived measurements included BMD at the lumbar spine, femoral neck, total hip, and trochanter and the trunk and total fat. Serum undercarboxylated osteocalcin (ucOCN), sclerostin, lipocalin-2, leptin, and adiponectin were measured pre and post-cART. Spearman’s rank-order correlations assessed the cross-sectional associations between hormones and bone and fat mass pre- and post-cART. Linear regression models adjusting for baseline bone or fat mass assessed the association between hormone change and BMD/fat changes following cART initiation.

RESULTS: ucOCN (p=0.04) and lipocalin-2 (p=0.03) increased post-cART while sclerostin, leptin, and adiponectin remained unchanged. BMD significantly decreased post-cART at all skeletal sites. Trunk and total fat increased post-cART but not significantly, while weight and BMI remained unchanged. In models adjusting for baseline BMD and fat mass, change in ucOCN was negatively associated with change in trunk (p=0.008) and total fat (p=0.01) and the change in leptin was positively associated with change in total hip (p=0.03) and trochanteric BMD (p=0.02).

CONCLUSION: The current study demonstrates bone-fat crosstalk in cART initiating PLWH.

PMID:34547525 | DOI:10.1016/j.bone.2021.116208

Biol Psychol. 2021 Sep 18:108190. doi: 10.1016/j.biopsycho.2021.108190. Online ahead of print.

ABSTRACT

Interoception and homeostatic/allostatic control are intertwined branches of closed-loop brain-body interactions (BBI). Given their importance in mental and psychosomatic disorders, establishing computational assays of BBI represents a clinically important but methodologically challenging endeavor. This technical note presents a novel approach, derived from a generic computational model of homeostatic/allostatic control that underpins (meta)cognitive theories of affective and psychosomatic disorders. This model views homeostatic setpoints as probability distributions (“homeostatic beliefs”) whose parameters determine regulatory efforts and change dynamically under allostatic predictions. In particular, changes in homeostatic belief precision, triggered by anticipated threats to homeostasis, are thought to alter cerebral regulation of bodily states. Here, we present statistical procedures for inferring homeostatic belief precision from measured bodily states and/or regulatory (action) signals. We analyze the inference problem, derive two alternative estimators of homeostatic belief precision, and apply our method to simulated data. Our proposed approach may prove useful for assessing BBI in individual subjects.

PMID:34547398 | DOI:10.1016/j.biopsycho.2021.108190