Refining outcome prediction after traumatic brain injury with machine learning algorithms.

Outcome after traumatic brain injury (TBI) is typically assessed using the Glasgow outcome scale extended (GOSE) with levels from 1 (death) to 8 (upper good recovery). Outcome prediction has classically been dichotomized into either dead/alive or favorable/unfavorable outcome. Binary outcome prediction models limit the possibility of detecting subtle yet significant improvements. We set out to explore different machine learning methods with the purpose of mapping their predictions to the full 8 grade scale GOSE following TBI. The models were set up using the variables: age, GCS-motor score, pupillary reaction, and Marshall CT score. For model setup and internal validation, a total of 866 patients could be included. For external validation, a cohort of 369 patients were included from Leuven, Belgium, and a cohort of 573 patients from the US multi-center ProTECT III study. Our findings indicate that proportional odds logistic regression (POLR), random forest regression, and a neural network model achieved accuracy values of 0.3-0.35 when applied to internal data, compared to the random baseline which is 0.125 for eight categories. The models demonstrated satisfactory performance during external validation in the data from Leuven, however, their performance were not satisfactory when applied to the ProTECT III dataset.

Internet Use as a Moderator of the Relationship between Age and Depression: Implications for Older Adults.

Technology-based interventions have been suggested to prevent depression and loneliness. However, it is not yet well-established if internet use is a protective factor for older adults, allowing them to combat depression and loneliness. Addressing this neglected area of research, this paper examines whether frequent internet use moderated the relationship between depression and age. Using 2016 General Social Survey (n = 869), a series of ordinary least squares regression models were employed. The findings revealed that older adults who frequently used the internet were more depressed than those who did not. This may imply that internet use by older individuals may not be an adequate source of social engagement and may even heighten existing levels of depression. Other interventions may need to be developed to compensate for the loss of physical contact and reduce the level of depression.

Individual risk and familial liability for suicide attempt and suicide in autism: a population-based study.

BACKGROUND: Studies on the individual gender-specific risk and familial co-aggregation of suicidal behaviour in autism spectrum disorder (ASD) are lacking. METHODS: We conducted a matched case-cohort study applying conditional logistic regression models on 54 168 individuals recorded in 1987-2013 with ASD in Swedish national registers: ASD without ID n = 43 570 (out of which n = 19035, 43.69% with ADHD); ASD + ID n = 10 598 (out of which n = 2894 individuals, 27.31% with ADHD), and 270 840 controls, as well as 347 155 relatives of individuals with ASD and 1 735 775 control relatives. RESULTS: The risk for suicidal behaviours [reported as odds ratio OR (95% confidence interval CI)] was most increased in the ASD without ID group with comorbid ADHD [suicide attempt 7.25 (6.79-7.73); most severe attempts i.e. requiring inpatient stay 12.37 (11.33-13.52); suicide 13.09 (8.54-20.08)]. The risk was also increased in ASD + ID group [all suicide attempts 2.60 (2.31-2.92); inpatient only 3.45 (2.96-4.02); suicide 2.31 (1.16-4.57)]. Females with ASD without ID had generally higher risk for suicidal behaviours than males, while both genders had highest risk in the case of comorbid ADHD [females, suicide attempts 10.27 (9.27-11.37); inpatient only 13.42 (11.87-15.18); suicide 14.26 (6.03-33.72); males, suicide attempts 5.55 (5.10-6.05); inpatient only 11.33 (9.98-12.86); suicide 12.72 (7.77-20.82)]. Adjustment for psychiatric comorbidity attenuated the risk estimates. In comparison to controls, relatives of individuals with ASD also had an increased risk of suicidal behaviour. CONCLUSIONS: Clinicians treating patients with ASD should be vigilant for suicidal behaviour and consider treatment of psychiatric comorbidity.

Population-based, multi-generational family clustering study of social anxiety disorder and avoidant personality disorder.

BACKGROUND: We aimed to provide unbiased estimates of familial risk and heritability of social anxiety disorder (SAD) and avoidant personality disorder (AVPD). METHOD: We identified 18 399 individuals diagnosed with SAD and 2673 with AVPD in the Swedish National Patient Register between 1997 and 2009. Risks (odds ratios; OR) for SAD in all biological and non-biological relatives of probands, compared to relatives of unaffected individuals were calculated. We also estimated the risks for AVPD in relatives of probands with SAD. RESULTS: The risk for SAD among relatives of SAD probands increased proportionally to the degree of genetic relatedness. The risks for first-degree relatives [OR 4.74, 95% confidence interval (CI) 4.28-5.25] were significantly higher than for second-degree and third-degree relatives. Second-degree relatives (OR 2.30, 95% CI 2.01-2.63) had significantly higher risk than third-degree relatives (OR 1.72, 95% CI 1.52-1.94). Relatives at similar genetic distances had similar risks for SAD, despite different degrees of shared environment. Heritability was estimated to be approximately 56%. There were no significant sex differences in the familial patterns. The risk of AVPD in relatives of SAD probands was significantly elevated, even after excluding individuals with both diagnoses (first-degree OR 3.54, second-degree OR 2.20, third-degree OR 1.62). Non-biological relatives (spouses/partners) also had elevated risks for both SAD (OR 4.01) and AVPD (OR 3.85). CONCLUSIONS: SAD clusters in families primarily due to genetic factors. SAD and AVPD are aetiologically related and may represent different expressions of the same vulnerability. The strong marital concordance observed in SAD/AVPD may indicate assortative mating but the exact mechanisms and implications require further investigation.

Bipolar disorder and leadership--a total population study.

OBJECTIVE: To investigate whether persons with bipolar disorder and their siblings have leadership traits and are overrepresented in executive professions. METHOD: A nested case-control study based on longitudinal Swedish total population registries. Data from officer suitability interviews (n=1,126,519), and information on occupations were collected. Bipolar patients (n=68,915) and their healthy siblings were compared with controls. RESULTS: Bipolar patients without comorbidity (pure; n=22,980) were overrepresented in both the highest and lowest strata of officer suitability; their healthy siblings in the highest strata only. Patients with pure bipolar disorder were underrepresented in executive professions, whereas their siblings were overrepresented in these professions (particularly political professions). Patients with general bipolar disorder (including those with comorbidities) and their healthy siblings were overrepresented only in the lowest strata of officer suitability ratings. General bipolar patients were underrepresented in executive professions, whereas their siblings had similar rates of executive professions as controls. Adjusting results for IQ slightly attenuated point estimates, but resulted in pure bipolar patients and their siblings no longer being significantly overrepresented in superior strata of officer suitability, and siblings no longer being overrepresented in executive professions. CONCLUSION: Results support that traits associated with bipolar disorder are linked to superior leadership qualities.

Micro-simulation of a smallpox outbreak using official register data.

To explore the efficacy of four vaccine-based policy strategies (ring vaccination, targeted vaccination, mass vaccination, and pre-vaccination of healthcare personnel combined with ring vaccination) for controlling smallpox outbreaks in Sweden, disease transmission on a spatially explicit social network was simulated. The mixing network was formed from high-coverage official register data of the entire Swedish population, building on the Swedish Total Population Register, the Swedish Employment Register, and the Geographic Database of Sweden. The largest reduction measured in the number of infections was achieved when combining ring vaccination with a pre-vaccination of healthcare personnel. In terms of per dose effectiveness, ring vaccination was by far the most effective strategy. The results can to some extent be adapted to other diseases and environments, including other countries, and the methods used can be analysed in their own right.

Economic consequences to society of pandemic H1N1 influenza 2009 - preliminary results for Sweden.

Experiments using a microsimulation platform show that vaccination against pandemic H1N1 influenza is highly cost-effective. Swedish society may reduce the costs of pandemic by about SEK 2.5 billion (approximately EUR 250 million) if at least 60 per cent of the population is vaccinated, even if costs related to death cases are excluded. The cost reduction primarily results from reduced absenteeism. These results are preliminary and based on comprehensive assumptions about the infectiousness and morbidity of the pandemic, which are uncertain in the current situation.

Nanoporous aluminum oxide affects neutrophil behaviour.

This study evaluates neutrophil responses on aluminum oxide membranes. Using an in vitro cell culture system, we have found that the pore size (20 and 200 nm in diameter) of alumina membranes have a significant effect on leukocyte morphology and activation. Specifically, our results show that 20-nm pore-size membranes were more potent in triggering PMN spreading and extending of pseudopodia than 200-nm pore-size membranes. The morphological changes are also associated with cell activation. In fact, adherent neutrophils on 20-nm pore-size membranes elicit much stronger initial oxygen free radical production. Overall, our results point out that membrane pore size significantly affects the extent of cellular responses of adherent neutrophils.

Psychosocial histories, social environment, and HIV risk behaviors of injection and noninjection drug using homeless youths.

Injection drug use is a common risk behavior for HIV infection among homeless, runaway and street youths. However, the psychosocial histories and current social environment of these youths are not well understood. The authors recruited 186 homeless, runaway and street youths using systematic street-based sampling methods, and assessed psychosocial histories, current daily activities, and sexual and drug-related risk behaviors using qualitative and quantitative techniques. Youths reported high lifetime rates of injection drug use (45%), recent drug and alcohol use (100%), and current homelessness (84%). Injection drug using youths were more likely than noninjection drug using youths to report traumatic psychosocial histories, including parental substance use and forced institutionalization, use of alcohol and other noninjection drugs, a history of survival sex, and the use of squats or abandoned buildings as shelter. These findings underscore the need for multifaceted service and prevention programs to address the varied needs of these high-risk youths.