Thinking beyond the mean: a practical guide for using quantile regression methods for health services research

Accuracy and Consistency of Confidence Limits for Monosyllable Identification Scores Derived Using Simulation, the Harrell-Davis Estimator, and Nonlinear Quantile Regression

BACKGROUND

Audiological diagnosis and rehabilitation often involve the assessment of whether the maximum speech identification score (PBmax) is poorer than expected from the pure-tone average (PTA) threshold. This requires the estimation of the lower boundary of the PBmax values expected for a given PTA (one-tailed 95% confidence limit, CL). This study compares the accuracy and consistency of three methods for estimating the 95% CL.

METHOD

The 95% CL values were estimated using a simulation method, the Harrell-Davis (HD) estimator, and non-linear quantile regression (nQR); the latter two are both distribution-free methods. The first two methods require the formation of sub-groups with different PTAs. Accuracy and consistency in the estimation of the 95% CL were assessed by applying each method to many random samples of 50% of the available data and using the fitted parameters to predict the data for the remaining 50%.

STUDY SAMPLE

A total of 642 participants aged 17 to 84 years with sensorineural hearing loss were recruited from audiology clinics. Pure-tone audiograms were obtained and PBmax scores were measured using monosyllables at 40 dB above the speech recognition threshold or at the most comfortable level.

RESULTS

For the simulation method, 6.7 to 8.2% of the PBmax values fell below the 95% CL for both ears, exceeding the target value of 5%. For the HD and nQR methods, the PBmax values fell below the estimated 95% CL for approximately 5% of the ears, indicating good accuracy. Consistency, estimated from the standard deviation of the deviations from the target value of 5%, was similar for all the methods.

CONCLUSIONS

The nQR method is recommended because it has good accuracy and consistency, and it does not require the formation of arbitrary PTA sub-groups.

Health factors that influence sustainable behaviour in a single-player resource management game

Encouraging sustainable use of limited natural, social, and economic resources requires understanding the variety of ways in which people think about how resources work and how they adjust their behaviour (or not) as available resources fluctuate. Previous investigations which have focused on understanding how individuals navigate erodible resources, have tended to use group-based, common pool games. However, such social games make it difficult to disentangle whether resource erosion is linked to difficulty navigating the dynamics of the resource or caused by social factors. Here, in two experiments, we recruited 781 participants to play a single-player resource management game in which individuals were invited to harvest monetary rewards from a fully depletable but stochastically replenishing resource over time. We find that the ability to sustain a resource over successive harvesting opportunities (in order to maximize the total harvested rewards) is reliably worse in individuals reporting elevated psychological distress, the often cooccurring hazardous alcohol use, and elevated rates of delay discounting. The associations between resource outcomes, harmful alcohol use, and psychological distress remained substantial even once we had controlled for elevated discounting rates (as a form of impulsivity and a strong risk factor for these health challenges). By contrast, individuals who reported higher levels of financial literacy and general well-being achieved better resource outcomes. Our observations demonstrate that the capacity to respond effectively to the dynamics of a resource are compromised in individuals at risk of psychological and alcohol-related disorders.

Examining Racial/Ethnic Differences in Patterns of Opioid Prescribing: Results from an Urban Safety-Net Healthcare System

Prescription opioids still account for a large proportion of overdose deaths and contribute to opioid use dependence (OUD). Studies earlier in the epidemic suggest clinicians were less likely to prescribe opioids to racial/ethnic minorities. As OUD-related deaths have increased disproportionately amongst minority populations, it is essential to understand racial/ethnic differences in opioid prescribing patterns to inform culturally sensitive mitigation efforts. The purpose of this study is to estimate racial/ethnic differences in opioid medication use among patients prescribed opioids. Using electronic health records and a retrospective cohort study design, we estimated multivariable hazard models and generalized linear models, assessing racial/ethnic differences in OUD diagnosis, number of opioid prescriptions, receiving only one opioid prescription, and receiving ≥18 opioid prescriptions. Study population (N=22,201) consisted of adult patients (≥18years), with ≥3 primary care visits (ensuring healthcare system linkage), ≥1 opioid prescription, who did not have an OUD diagnoses prior to the first opioid prescription during the 32-month study period. Relative to racial/ethnic minority patients, White patients, in both unadjusted and adjusted analyses, had a greater number of opioid prescriptions filled, a higher proportion received ≥18 opioid prescriptions, and a greater hazard of having an OUD diagnosis subsequent to receiving an opioid prescription (all groups p<0.001). Although opioid prescribing rates have declined nationally, our findings suggest White patients still experience a high volume of opioid prescriptions and greater risk of OUD diagnosis. Racial/ethnic minorities are less likely to receive follow-up pain medications, which may signal low care quality. Identifying provider bias in pain management of racial/ethnic minorities could inform interventions seeking balance between adequate pain treatment and risk of opioid misuse/abuse.

National Cancer Diagnosis Audits for England 2018 versus 2014: a comparative analysis

BACKGROUND

Timely diagnosis of cancer in patients who present with symptoms in primary care is a quality-improvement priority.

AIM

To examine possible changes to aspects of the diagnostic process, and its timeliness, before and after publication of the National Institute for Health and Care Excellence's (2015) guidance on the referral of suspected cancer in primary care.

DESIGN AND SETTING

Comparison of findings from population-based clinical audits of cancer diagnosis in general practices in England for patients diagnosed in 2018 or 2014.

METHOD

GPs in 1878 (2018) and 439 (2014) practices collected primary care information on the diagnostic pathway of cancer patients. Key measures including patient characteristics, place of presentation, number of pre-referral consultations, use of primary care investigations, and referral type were compared between the two audits by descriptive analysis and regression models.

RESULTS

Among 64 489 (2018) and 17 042 (2014) records of a new cancer diagnosis, the percentage of patients with same-day referral (denoted by a primary care interval of 0 days) was higher in 2018 (42.7% versus 37.7%) than in 2014, with similar improvements in median diagnostic interval (36 days versus 40 days). Compared with 2014, in 2018: fewer patients had ≥3 pre-referral consultations (18.8% versus 26.2%); use of primary care investigations increased (47.9% versus 45.4%); urgent cancer referrals increased (54.8% versus 51.8%); emergency referrals decreased (13.4% versus 16.5%); and recorded use of safety netting decreased (40.0% versus 44.4%).

CONCLUSION

In the 5-year period, including the year when national guidelines were updated (that is, 2015), there were substantial improvements to the diagnostic process of patients who present to general practice in England with symptoms of a subsequently diagnosed cancer.

Human milk unmetabolized folic acid is increased following supplementation with synthetic folic acid as compared to (6S)-5-methyltetrahydrofolic acid

Folic acid supplementation is recommended perinatally, but may increase unmetabolized folic acid (UMFA) in human milk; this is concerning as it is an inactive form which may be less bioavailable for the infant. "Natural" (6S)-5-methyltetrahydrofolic acid [(6S)-5-MTHF] is available as an alternative to folic acid, and may prevent the accumulation of UMFA in human milk. Pregnant women (n = 60) were enrolled at 8-21 weeks of gestation and randomized to 0.6 mg/day folic acid or (6S)-5-MTHF. At ~ 1-week postpartum, participants provided a human milk specimen. Total human milk folate (nmol/L) and concentrations of UMFA (nmol/L) were quantified via LC-MS/MS. Differences between groups were evaluated using multivariable quantile/linear regression, adjusting for dietary folate, weeks supplementing, and milk collection methods. No significant difference in total milk folate was found; however, the median milk UMFA concentration was 11 nmol/L higher in those receiving folic acid versus (6S)-5-MTHF (95% CI = 6.4-17 nmol/L), with UMFA representing 28% and 2% of total milk folate. In conclusion, the form of supplemental folate had markedly differential effects on the human milk folate profile, with folic acid increasing the mean proportion of milk UMFA by ~ 14-fold. Investigation of whether increased UMFA impacts folate-related metabolism and infant health outcomes is required.

The level of countries' preparedness to health risks during Covid-19 and pre-pandemic: the differential response to health systems building blocks and socioeconomic indicators

The global health security (GHS) Index assesses countries' level of preparedness to health risks. However, there is no evidence on how and whether the effects of health systems building blocks and socioeconomic indicators on the level of preparedness differ for low and high prepared countries. The aim of this study was to examine the contributions of health systems building blocks and socioeconomic indicators to show differences in the level of preparedness to health risks. The study also aimed to examine trends in the level of preparedness and the World Health Organization (WHO) regional differences before and during the Covid-19 pandemic. We used the 2021 GHS index report data and employed quantile regression, log-linear, double-logarithmic, and time-fixed effects models. As robustness checks, these functional form specifications corroborated with one another, and interval validity tests confirmed. The results show that increases in effective governance, supply chain capacity in terms of medicines and technologies, and health financing had positive effects on countries' level of preparedness to health risks. These effects were considerably larger for countries with higher levels of preparedness to health risks. The positive gradient trends signaled a sense of capacity on the part of countries with higher global health security. However, the health workforce including doctors, and health services including hospital beds, were not statistically significant in explaining variations in countries' level of preparedness. While economic factors had positive effects on the level of preparedness to health risks, their impacts across the distribution of countries' level of preparedness to health risks were mixed. The effects of Social Development Goals (SDGs) were greater for countries with higher levels of preparedness to health risks. The effect of the Human Development Index (HDI) was greatest for countries whose overall GHS index lies at the midpoint of the distribution of countries' level of preparedness. High-income levels were associated with a negative effect on the level of preparedness, especially if countries were in the lower quantiles across the distributions of preparedness. Relative to poor countries, middle- and high-income groups had lower levels of preparedness to health risks, an indication of a sense of complacency. We find the pandemic period (year 2021) was associated with a decrease in the level of preparedness to health risks in comparison to the pre-pandemic period. There were significant WHO regional differences. Apart from the Eastern Mediterranean, the rest of the regions were more prepared to health risks compared to Africa. There was a negative trend in the level of preparedness to health risks from 2019 to 2021 although regional differences in changes over time were not statistically significant. In conclusion, attempts to strengthen countries' level of preparedness to health shocks should be more focused on enhancing essentials such as supply chain capacity in terms of medicines and technologies; health financing, and communication infrastructure. Countries should also strengthen their already existing health workforce and health services. Together, strengthening these health systems essentials will be beneficial to less prepared countries where their impact we find to be weaker. Similarly, boosting SDGs, particularly health-related sub-scales, will be helpful to less prepared countries. Moreover, there is a need to curb complacency in preparedness to health risks during pandemics by high-income countries. The negative trend in the level of preparedness to health risks would suggest that there is a need for better preparedness during pandemics by conflating national health with global health risks. This will ensure the imperative of having a synergistic response to global health risks, which is understood by and communicated to all countries and regions.

Changes in limiting factors for forager population dynamics in Europe across the last glacial-interglacial transition

Population dynamics set the framework for human genetic and cultural evolution. For foragers, demographic and environmental changes correlate strongly, although the causal relations between different environmental variables and human responses through time and space likely varied. Building on the notion of limiting factors, namely that at any one time, the scarcest resource caps population size, we present a statistical approach to identify the dominant climatic constraints for hunter-gatherer population densities and then hindcast their changing dynamics in Europe for the period between 21,000 to 8000 years ago. Limiting factors shifted from temperature-related variables (effective temperature) during the Pleistocene to a regional mosaic of limiting factors in the Holocene dominated by temperature seasonality and annual precipitation. This spatiotemporal variation suggests that hunter-gatherers needed to overcome very different adaptive challenges in different parts of Europe and that these challenges varied over time. The signatures of these changing adaptations may be visible archaeologically. In addition, the spatial disaggregation of limiting factors from the Pleistocene to the Holocene coincided with and may partly explain the diversification of the cultural geography at this time.

Here, the authors use climate and resource availability, to statistically model the limiting factors in the dynamics of hunter-gatherer population densities in Europe between 21,000 and 8,000 years ago. They find that limiting factors varied spatiotemporally and the effects of these may be visible in the archaeological record.

How Responsive is Mortality to Locally Administered Healthcare Expenditure? Estimates for England for 2014/15

BACKGROUND

Research using local English data from 2003 to 2012 suggests that a 1% increase in healthcare expenditure causes a 0.78% reduction in mortality, and that it costs the NHS £10,000 to generate an additional quality-adjusted life year (QALY). In 2013, the existing 151 local health authorities (Primary Care Trusts) were abolished and replaced with 212 Clinical Commissioning Groups (CCGs). CCGs retained responsibility for secondary care and pharmaceuticals, but responsibility for primary care and specialised commissioning returned to central administrators.

OBJECTIVES

The aim was to extend and apply existing methods to more recent data using a new geography and expenditure base, while improving covariate selection and examining the responsiveness of mortality to expenditure across the mortality distribution.

METHODS

Instrumental variable regression is used to quantify the relationship between mortality and local expenditure. Backward selection and regularised regression are used to identify parsimonious specifications. These results are combined with information about survival and morbidity disease burden to calculate the marginal cost per QALY. Unconditional quantile regression (UQR) is used to examine the response of mortality to expenditure across the mortality distribution.

RESULTS

Backward selection and regularised regression both suggest that the marginal cost per QALY in 2014/15 was about £7000 for locally commissioned services. The UQR results suggest that additional expenditure generates larger health benefits in high-mortality areas and that, if anything, the average size of this heterogeneous response is larger than the response at the mean.

CONCLUSIONS

The new healthcare geography and expenditure base can be used to update estimates of the health opportunity costs associated with additional expenditure. The variation in the mortality response across the mortality distribution suggests that the use of the response at the mean will, if anything, underestimate the health opportunity costs associated with a national policy or nationally mandated guidance on the use of new technologies. The health opportunity costs of such policies are likely to be greater (lower) in areas of higher (lower) mortality, increasing health inequalities.

Testing the heterogeneous effect of air transport intensity on CO2 emissions in G20 countries: An advanced empirical analysis

This study investigates the heterogeneous impact of air transport intensity, air passenger transport, and air freight transport on air transport carbon emissions in G20 countries for the period of 1990-2016. The paper employs a robust and advanced fixed-effect panel quantile regression model that considers unobserved discrete and distributional heterogeneity. Our empirical results show that the impact of the independent variables on air transport carbon emissions is quite heterogeneous across various quantiles. More specifically, the effect of air transport intensity, air passenger transport, and air freight transport on carbon emissions is positive and becomes more assertive with the increasing trend at upper quantiles and is quite heterogeneous across all quantiles. Economic growth, urbanization, and tourism are significant contributing factors in enhancing air transport CO₂ emissions, while crude oil price significantly reduces CO₂ emissions. The Dumitrescu and Hurlin causality test estimates indicate that a bidirectional relationship extends from air transport intensity, air passenger transport, and air freight transport to air transport CO₂ emissions. The findings underline the need for cleaner, renewable, and environmentally sustainable energy sources for air transport operations.

Quantile regression to examine the association of air pollution with subclinical atherosclerosis in an adolescent population

BACKGROUND

Air pollution has been associated with carotid intima-media thickness test (CIMT), a marker of subclinical atherosclerosis. To our knowledge, this is the first study to report an association between ambient air pollution and CIMT in a younger adolescent population.

OBJECTIVE

To investigate the associations beyond standard mean regression by using quantile regression to explore if associations occur at different percentiles of the CIMT distribution.

METHODS

We measured CIMT cross-sectionally at the age of 16 years in 363 adolescents participating in the Dutch PIAMA birth cohort. We fit separate quantile regressions to examine whether the associations of annual averages of nitrogen dioxide (NO₂), fine particulate matter (PM_2.5), PM_2.5 absorbance (a marker for black carbon), PM_coarse and ultrafine particles up to age 14 assigned at residential addresses with CIMT varied across deciles of CIMT. False discovery rate corrections (FDR, p < 0.05 for statistical significance) were applied for multiple comparisons. We report quantile regression coefficients that correspond to an average change in CIMT (µm) associated with an interquartile range increase in the exposure.

RESULTS

PM_2.5 absorbance exposure at birth was statistically significantly (FDR < 0.05) associated with a 6.23 µm (95% CI: 0.15, 12.3) higher CIMT per IQR increment in PM_2.5 absorbance in the 10th quantile of CIMT but was not significantly related to other deciles within the CIMT distribution. For NO₂ exposure we found similar effect sizes to PM_2.5 absorbance, but with much wider confidence intervals. PM_2.5 exposure was weakly positively associated with CIMT while PM_coarse and ultrafine did not display any consistent patterns.

CONCLUSIONS

Early childhood exposure to ambient air pollution was suggestively associated with the CIMT distribution during adolescence. Since CIMT increases with age, mitigation strategies to reduce traffic-related air pollution early in life could possibly delay atherosclerosis and subsequently CVD development later in life.

Socioeconomic and demographic correlates of child nutritional status in Nepal: an investigation of heterogeneous effects using quantile regression

Background

Child undernutrition continues to be a major public health problem in many countries, including Nepal. The repercussions of undernutrition are not only limited to the affected children and families but also transcend to the national and global economy. Earlier studies from Nepal have predominantly used either ordinary least squares (OLS) regression or binary regression to analyse the socioeconomic and demographic correlates of the nutritional outcome. In this study, quantile regression was used to understand a complete and more precise estimate of the effects of the covariates on the child nutritional status.

Methods

This study was based on the most recent nationally representative Nepal Multiple Indicator Cluster Survey (MICS) 2019. Height-for-age z scores (HAZ) were used as an indicator for assessing the nutritional status of under-five children. Quantile regression was used to examine the heterogeneous association of covariates with conditional HAZ distribution across the different quantiles (0.10, 0.30, 0.50, 0.85). As a comparison, the effects of covariates at conditional mean of HAZ using OLS regression was also analysed. The graphs were plotted to visualize the changes in the coefficients for each regressor across the entire conditional HAZ distribution.

Results

Age of children, sex of children, province and wealth had a consistent and statistically significant association with HAZ in both OLS and quantile regression. Improved toilet facility was positively correlated with HAZ at the lower tails (tenth and thirtieth percentiles). Ethnicity (Janajati and Newer) was positively correlated with HAZ at the lower tail (thirtieth percentile) and mean (OLS regression). Maternal education was a significant predictor of improved height-for-age across conditional quantiles, except at the tenth percentile. Maternal age, number of under-five children in household, number of household members, and improved source of drinking water showed heterogeneous effects across different quantiles of conditional HAZ distribution.

Conclusion

Use of quantile regression approach showed that the effect of different factors differed across the conditional distribution of HAZ. Policymakers should consider the heterogeneous effect of different factors on HAZ so that the targeted intervention could be implemented to maximize the nutritional benefits to children.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12992-022-00834-4.

Quantile Treatment Effect of Zinc Lozenges on Common Cold Duration: A Novel Approach to Analyze the Effect of Treatment on Illness Duration

Calculation of the difference of means is the most common approach when analyzing treatment effects on continuous outcomes. Nevertheless, it is possible that the treatment has a different effect on patients who have a lower value of the outcome compared with patients who have a greater value of the outcome. The estimation of quantile treatment effects (QTEs) allows the analysis of treatment effects over the entire distribution of a continuous outcome, such as the duration of illness or the duration of hospital stay. Furthermore, most of these outcomes have asymmetric distributions with fat tails, and censored observations are not uncommon. These features can be accounted for in the analysis of the QTE. In this paper, we use the QTE approach to analyze the effect of zinc lozenges on common cold duration. We use the data set of the Mossad (1996) trial with zinc gluconate lozenges, and three data sets of trials with zinc acetate lozenges. In the Mossad (1996) trial, zinc gluconate lozenges shortened common cold duration on average by 4.0 days (95% CI 2.3-5.7 days). However, the QTE analysis indicates that 15- to 17-day colds were shortened by 8 days, and 2-day colds by just 1 day, for the group taking zinc lozenges. Thus, the overall 4.0-day average effect of zinc gluconate lozenges in the Mossad (1996) trial is inconsistent with our QTE findings for both short and long colds. Similar results were found in our QTE analysis of the pooled data sets of the three zinc acetate lozenge trials. The average effect of 2.7 days (95% CI 1.8-3.3 days) was inconsistent with the effects on short and long colds. The QTE approach may have broad usefulness for examining treatment effects on the duration of illness and hospital stay, and on other similar outcomes.

A Quantile Regression Analysis of Factors Associated with First-Time Maternal Fatigue in Korea

The aim of this cross-sectional study was to identify the factors associated with different percentiles of first-time maternal fatigue. A total of 123 first-time healthy mothers aged 18 years or older participated through an online survey. The fatigue was measured by the Korean version of the fatigue severity scale. Main variables were constructed based on the integrated fatigue model, which included mothers' sleep quality, parenting stress, the amount of free time mothers have, the number of the child's night wakings, general characteristics including socioeconomic status, and working status. Quantile regression was used to analyze the associated factors according to the fatigue level of first-time mothers with a young child. The mean age of the mothers and children were 32.11 years and 20.81 months, respectively. Mean fatigue score was 6.16 among the 75% quantile with high fatigue score. Lack of adequate free time in mothers, advanced maternal age, being a housewife, having a moderate income, and frequent night wakings of their child significantly increased fatigue among mothers in the third quantile of fatigue. To reduce fatigue, healthcare providers should focus on exploring ways to reduce maternal sleep disturbance and improve maternal sleep quality.

Polygenic risk scores across the extended psychosis spectrum

As early detection of symptoms in the subclinical to clinical psychosis spectrum may improve health outcomes, knowing the probabilistic susceptibility of developing a disorder could guide mitigation measures and clinical intervention. In this context, polygenic risk scores (PRSs) quantifying the additive effects of multiple common genetic variants hold the potential to predict complex diseases and index severity gradients. PRSs for schizophrenia (SZ) and bipolar disorder (BD) were computed using Bayesian regression and continuous shrinkage priors based on the latest SZ and BD genome-wide association studies (Psychiatric Genomics Consortium, third release). Eight well-phenotyped groups (n = 1580; 56% males) were assessed: control (n = 305), lower (n = 117) and higher (n = 113) schizotypy (both groups of healthy individuals), at-risk for psychosis (n = 120), BD type-I (n = 359), BD type-II (n = 96), schizoaffective disorder (n = 86), and SZ groups (n = 384). PRS differences were investigated for binary traits and the quantitative Positive and Negative Syndrome Scale. Both BD-PRS and SZ-PRS significantly differentiated controls from at-risk and clinical groups (Nagelkerke's pseudo-R2: 1.3-7.7%), except for BD type-II for SZ-PRS. Out of 28 pairwise comparisons for SZ-PRS and BD-PRS, 9 and 12, respectively, reached the Bonferroni-corrected significance. BD-PRS differed between control and at-risk groups, but not between at-risk and BD type-I groups. There was no difference between controls and schizotypy. SZ-PRSs, but not BD-PRSs, were positively associated with transdiagnostic symptomology. Overall, PRSs support the continuum model across the psychosis spectrum at the genomic level with possible irregularities for schizotypy. The at-risk state demands heightened clinical attention and research addressing symptom course specifiers. Continued efforts are needed to refine the diagnostic and prognostic accuracy of PRSs in mental healthcare.

Effectiveness of a nurse-led hospital-to-home transitional care intervention for older adults with multimorbidity and depressive symptoms: A pragmatic randomized controlled trial

OBJECTIVE

To evaluate the effectiveness of a nurse-led hospital-to-home transitional care intervention versus usual care on mental functioning (primary outcome), physical functioning, depressive symptoms, anxiety, perceived social support, patient experience, and health service use costs in older adults with multimorbidity (≥ 2 comorbidities) and depressive symptoms.

DESIGN AND SETTING

Pragmatic multi-site randomized controlled trial conducted in three communities in Ontario, Canada. Participants were allocated into two groups of intervention and usual care (control).

PARTICIPANTS

127 older adults (≥ 65 years) discharged from hospital to the community with multimorbidity and depressive symptoms.

INTERVENTION

This evidence-based, patient-centred intervention consisted of individually tailored care delivery by a Registered Nurse comprising in-home visits, telephone follow-up and system navigation support over 6-months.

OUTCOME MEASURES

The primary outcome was the change in mental functioning, from baseline to 6-months. Secondary outcomes were the change in physical functioning, depressive symptoms, anxiety, perceived social support, patient experience, and health service use cost, from baseline to 6-months. Intention-to-treat analysis was performed using ANCOVA modeling.

RESULTS

Of 127 enrolled participants (63-intervention, 64-control), 85% had six or more chronic conditions. 28 participants were lost to follow-up, leaving 99 (47 -intervention, 52-control) participants for the complete case analysis. No significant group differences were seen for the baseline to six-month change in mental functioning or other secondary outcomes. Older adults in the intervention group reported receiving more information about health and social services (p = 0.03) compared with the usual care group.

CONCLUSIONS

Although no significant group differences were seen for the primary or secondary outcomes, the intervention resulted in improvements in one aspect of patient experience (information about health and social services). The study sample fell below the target sample (enrolled 127, targeted 216), which can account for the non-significant findings. Further research on the impact of the intervention and factors that contribute to the results is recommended.

TRIAL REGISTRATION

clinicaltrials.gov Identifier: NCT03157999.

Demographic Disparities in Proximity to Certified Stroke Care in the United States

[Figure: see text].

A cross-sectional study measuring contact patterns using diaries in an urban and a rural community in South Africa, 2018

BACKGROUND

Describing contact patterns is crucial to understanding infectious disease transmission dynamics and guiding targeted transmission mitigation interventions. Data on contact patterns in Africa, especially South Africa, are limited. We measured and compared contact patterns in a rural and urban community, South Africa. We assessed participant and contact characteristics associated with differences in contact rates.

METHODS

We conducted a cross-sectional study nested in a prospective household cohort study. We interviewed participants to collect information on persons in contact with for one day. We described self-reported contact rates as median number people contacted per day, assessed differences in contact rates based on participant characteristics using quantile regression, and used a Poisson model to assess differences in contact rates based on contact characteristics within age groups. We also calculated cumulative person hours in contact within age groups at different locations.

RESULTS

We conducted 535 interviews (269 rural, 266 urban), with 17,252 contacts reported. The overall contact rate was 14 (interquartile range (IQR) 9-33) contacts per day. Those ≤18 years had higher contact rates at the rural site (coefficient 17, 95% confidence interval (95%CI) 10-23) compared to the urban site, for those aged 14-18 years (13, 95%CI 3-23) compared to < 7 years. No differences were observed for adults. There was a strong age-based mixing, with age groups interacting more with similar age groups, but also interaction of participants of all ages with adults. Children aged 14-18 years had the highest cumulative person hours in contact (116.3 rural and 76.4 urban).

CONCLUSIONS

Age played an important role in the number and duration of contact events, with children at the rural site having almost double the contact rate compared to the urban site. These contact rates can be utilized in mathematical models to assess transmission dynamics of infectious diseases in similar communities.

Oral Oncology Parity Laws, Medication Use, and Out-of-Pocket Spending for Patients With Blood Cancers

BACKGROUND

In this study, we sought to estimate the association between oral oncology parity law adoption and anticancer medication use for patients with chronic myeloid leukemia or multiple myeloma.

METHODS

This was an observational study of administrative claims from 2008 to 2017. Among individuals initiating tyrosine kinase inhibitors (TKIs) for chronic myeloid leukemia or immunomodulatory drugs for multiple myeloma, we compared out-of-pocket spending, adherence, and discontinuation before and after parity among individuals in fully insured plans (subject to parity) vs self-funded plans (exempt from parity) using propensity-score weighted difference-in-differences regression models.

RESULTS

Among patients initiating TKIs (N = 2082) or immunomodulatory drugs (N = 3326) there were no statistically significant differences in adherence or discontinuation associated with parity. The proportion of patients with initial out-of-pocket payments of $0 increased in fully insured plans after parity from 5.7% to 46.1% for TKIs and from 10.9% to 48.8% for immunomodulatory drugs. Relative to changes in self-funded plans, those in fully insured plans were 4.27 (95% CI = 2.20 to 8.27) times as likely to pay nothing for TKIs and 1.96 (95% CI = 1.40 to 2.73) times as likely to pay nothing for immunomodulatory drugs after parity. Similarly, the proportion paying more than $100 decreased from 30.3% to 24.7% for TKIs and 30.6% to 27.5% for immunomodulatory drugs in fully insured plans after parity. Relative to changes in self-funded plans, those in fully insured plans were 0.74 (95% CI = 0.54 to 1.01) times as likely to pay more than $100 for TKIs and 0.85 (95% CI = 0.68 to 1.06) times as likely to pay more than $100 for immunomodulatory drugs after parity.

CONCLUSIONS

Among patients initiating TKIs or immunomodulatory drugs, parity was not associated with better adherence or less discontinuation of therapy but yielded decreased patient out-of-pocket payments for some patients.

Relative Age Effect in Canadian Hockey: Prevalence, Perceived Competence and Performance

The term "relative age effect" (RAE) is used to describe a bias in which participation in sports (and other fields) is higher among people who were born at the beginning of the relevant selection period than would be expected from the distribution of births. In sports, RAEs may affect the psychological experience of players as well as their performance. This article presents 2 studies. Study 1 aims to verify the prevalence of RAEs in minor hockey and test its associations with players' physical self-concept and attitudes toward physical activities in general. Study 2 verifies the prevalence of the RAE and analyzes the performance of Canadian junior elite players as a function of their birth quartile. In study 1, the sample is drawn from 404 minor hockey players who have evolved from a recreational to an elite level. Physical self-concept and attitudes toward different kinds of physical activities were assessed via questionnaires. Results showed that the RAE is prevalent in minor hockey at all competition levels. Minor differences in favor of Q1-born players were observed regarding physical self-concept, but not attitudes. In study 2, data analyses were conducted from the 2018-2019 Canadian Hockey League database. Birth quartiles were compared on different components of performance by using quantile regression on each variable. Results revealed that RAEs are prevalent in the CHL, with Q1 players tending to outperform Q4 players in games played and power-play points. No other significant differences were observed regarding anthropometric measures and other performance outcomes. RAEs are still prevalent in Canadian hockey. Building up perceived competence and providing game-time exposure are examples of aspects that need to be addressed when trying to minimize RAEs in ice hockey.

Exploring Medication Adherence with P2Y12 Inhibitors Using Conditional and Unconditional Quantile Regression Approaches

BACKGROUND

Previous research assessing medication adherence with P2Y₁₂ inhibitors has shown good adherence rates, ranging from 78% to 92%. Studies that used administrative claims data defined adherence using an arbitrary cut point of ≥ 80% medication possession ratio (MPR) or proportion of days covered (PDC). While this method is used frequently, it does not allow the researcher to observe how each factor impacts adherence along the entire distribution. The objective of the study was to use conditional quantile regression (CQR) and unconditional quantile regression (UQR) to assess heterogenous effects of adherence to P2Y₁₂ inhibitors and covariates of interest and compare these results to those from a traditional logistic regression framework.

METHODS AND RESULTS

This study used the commercial claims and encounters databases from IBM^® MarketScan^® from 2010 to 2017. We included patients who had an incident percutaneous coronary intervention, used a drug-eluting stent, and filled an incident prescription for a P2Y₁₂ inhibitor. Adherence was measured for 185 days using PDC. Adherence to branded clopidogrel, generic clopidogrel, branded prasugrel, and branded ticagrelor was assessed, along with factors that could impact adherence, using logistic regression, CQR, and UQR. We found that while adherence to the antiplatelets was generally high, prasugrel and ticagrelor had significantly lower PDC compared to branded clopidogrel, especially around the 30th percentile. Across all quantiles in both the CRQ and UQR frameworks, comorbidities such as diabetes and depression and living in the southern region had significant negative effects on adherence, although the relative impact differed across quantiles.

CONCLUSIONS

Using CQR and UQR allowed for heterogenous assessment of covariates along the adherence distribution, which is not possible with the traditional logistic regression method. The UQR framework revealed patients who initiate prasugrel or ticagrelor generally have lower adherence compared to those treated with branded clopidogrel, especially around the 30th quantile. Using these methods in other types of data sets, such as electronic health records, could help strengthen our results to develop policies to improve antiplatelet adherence in a targeted population.

Confidence Limits of Word Identification Scores Derived Using Nonlinear Quantile Regression

The relation between degree of sensorineural hearing loss and maximum speech identification scores (PBmax) is commonly used in audiological diagnosis and rehabilitation. It is important to consider the relation between the degree of hearing loss and the lower boundary of PBmax, as the PBmax varies largely between subjects at a given degree of hearing loss. The present study determines the lower boundary by estimating the lower limit of the one-tailed 95% confidence limit (CL) for a Dantale I, word list, in a large group of young and older subjects with primarily sensorineural hearing loss. PBmax scores were measured using Dantale I, at 30 dB above the speech reception threshold or at the most comfortable level from 1,961 subjects with a wide range of pure-tone averages. A nonlinear quantile regression approach was applied to determine the lower boundary (95% CL) of PBmax scores. At a specific pure-tone average, if the measured PBmax is poorer than the lower boundary (95% CL) of PBmax, it may be considered disproportionately poor.

Subjective cognitive decline in diabetes: Associations with psychological, sleep, and diabetes-related factors

AIMS

To determine whether psychological, sleep, and diabetes-related factors were associated with subjective cognitive decline (SCD) in adults with type 2 diabetes (T2D).

DESIGN

A secondary analysis with a cross-sectional, correlational design.

METHODS

Data from two parent studies were combined. A total of 105 adults with diabetes were recruited from a Midwestern city in the United States from September 2013-March 2014 and September 2016-September 2017. Subjective cognitive decline was measured with the psychological-cognitive symptom subscale of the Diabetes Symptom Checklist-Revised. Psychological factors (diabetes distress, fatigue, and depressive symptoms) and sleep were measured using the Diabetes Distress Scale and Patient-Reported Outcomes Measurement Information System. Diabetes-related factors were measured with body mass index and glycaemia control. Quantile regression was used to examine the associations.

RESULTS/FINDINGS

The mean age of the participants was 58.10 (SD 7.92) years and 58.1% were women. The participants' mean psychological-cognitive symptom score was 1.41 (SD 1.13). After controlling for age and gender, sleep disturbance, sleep-related impairment, and body mass index were associated with SCD in lower cognitive symptom quantile groups (5th to 30th percentiles). In contrast, fatigue and depressive symptoms were more strongly associated with SCD in higher quantile groups (70th to 95th percentiles).

CONCLUSION

In adults with T2D, SCD was differentially associated with psychological, sleep, and diabetes-related factors depending on cognitive symptom percentiles. Adults with T2D require regular screening for SCD. If they complain of cognitive symptoms at clinical visits, different underlying factors should be assessed according to their symptom severity.

IMPACT

Findings from this study provided evidence for early identification of SCD and its influencing factors, which may help to develop nursing interventions to recognize and/or delay the onset of cognitive impairment in adults with T2D.

Newly graduated dentists' knowledge of temporomandibular disorders compared to specialists in Saudi Arabia

Background

General dentists are often the first healthcare professionals to see patients with orofacial pain (OFP). OFP conditions associated with the temporomandibular joint are often confused with dentoalveolar disorders, which leads to mismanagement. The objective of this study was to evaluate the level of knowledge of temporomandibular disorders (TMD) among newly graduated dentists compared to OFP specialists in Saudi Arabia.

Methods

This was a descriptive cross-sectional study utilizing an anonymous validated questionnaire assessing professional knowledge regarding TMDs in newly qualified dentists and OFP specialists. The questionnaire interrogated four domains including chronic pain/pain behavior, etiology, diagnosis/classification, and treatment/prognosis. OFP specialists were used as the reference group.

Results

A total of 393 dentists participated, a response rate of 67.6% in newly graduated dentists and 77.3% in OFP specialists. The degree of agreement between newly graduated general dentists and OFP specialists was highest for the “chronic pain and pain behavior” domain. The consensus among specialists was highest for the “treatment and prognosis” domain and the least for the “chronic pain and pain behavior” domain.

Conclusion

Newly graduated general dentists have limited knowledge of TMD in almost all domains compared to specialists. Given that a lack of knowledge of TMD can lead to clinical mismanagement, dental school curricula must address this important knowledge gap.

The synergetic effect of alcohol consumption and cigarettes per day on smoking outcomes expectancies among Latinx adult smokers

The present study investigated alcohol consumption and cigarettes per day in relation to smoking outcome expectancies among Spanish-speaking Latinx daily smokers (N = 371). There was a significant interaction between alcohol consumption and number of cigarettes per day on positive smoking expectancies. Specifically, alcohol consumption has a stronger association with positive expectancies for smoking at lower rates of cigarettes per day. No such interaction was evident for negative consequence smoking expectancies. The current study highlights the potential importance of alcohol consumption and smoking rate for better understanding smoking outcome expectancies among Latinx smokers.

Mapping the Relationship between Dysmorphology and Cognitive, Behavioral, and Developmental Outcomes in Children with Autism Spectrum Disorder

Previous studies investigating the association between dysmorphology and cognitive, behavioral, and developmental outcomes among individuals with autism spectrum disorder (ASD) have been limited by the binary classification of dysmorphology and lack of comparison groups. We assessed the association using a continuous measure of dysmorphology severity (DS) in preschool children aged 2-5 years (322 with ASD and intellectual disability [ID], 188 with ASD without ID, and 371 without ASD from the general population [POP]). In bivariate analyses, an inverse association between DS and expressive language, receptive language, fine motor, and visual reception skills was observed in children with ASD and ID. An inverse association of DS with fine motor and visual reception skills, but not expressive language and receptive language, was found in children with ASD without ID. No associations were observed in POP children. These results persisted after exclusion of children with known genetic syndromes or major morphologic anomalies. Quantile regression models showed that the inverse relationships remained significant after adjustment for sex, race/ethnicity, maternal education, family income, study site, and preterm birth. DS was not associated with autistic traits or autism symptom severity, behaviors, or regression among children with ASD with or without ID. Thus, DS was associated with a global impairment of cognitive functioning in children with ASD and ID, but only with fine motor and visual reception deficits in children with ASD without ID. A better understanding is needed for mechanisms that explain the association between DS and cognitive impairment in children with different disorders. Autism Res 2020, 13: 1227-1238. © 2020 International Society for Autism Research, Wiley Periodicals, Inc. LAY SUMMARY: We examined whether having more dysmorphic features (DFs) was related to developmental problems among children with autism spectrum disorder (ASD) with or without intellectual disability (ID), and children without ASD from the general population (POP). Children with ASD and ID had more language, movement, and learning issues as the number of DFs increased. Children with ASD without ID had more movement and learning issues as the number of DFs increased. These relationships were not observed in the POP group. Implications are discussed.

Simultaneous quantile regression and determinants of under-five severe chronic malnutrition in Ghana

BACKGROUND

Under-five malnutrition is a major public health issue contributing to mortality and morbidity, especially in developing countries like Ghana where the rates remain unacceptably high. Identification of critical risk factors of under-five malnutrition using appropriate and advanced statistical methods can help formulate appropriate health programmes and policies aimed at achieving the United Nations SDG Goal 2 target 2. This study attempts to develop a simultaneous quantile regression, an in-depth statistical model to identify critical risk factors of under-five severe chronic malnutrition (severe stunting).

METHODS

Based on the nationally representative data from the 2014 Ghana Demographic and Health Survey, height-for-age z-score (HAZ) was estimated. Multivariable simultaneous quantile regression modelling was employed to identify critical risk factors for severe stunting based on HAZ (a measure of chronic malnutrition in populations). Quantiles of HAZ with focus on severe stunting were modelled and the impact of the risk factors determined. Significant test of the difference between slopes at different selected quantiles of severe stunting and other quantiles were performed. A quantile regression plots of slopes were developed to visually examine the impact of the risk factors across these quantiles.

RESULTS

Data on a total of 2716 children were analysed out of which 144 (5.3%) were severely stunted. The models identified child level factors such as type of birth, sex, age, place of delivery and size at birth as significant risk factors of under-five severe stunting. Maternal and household level factors identified as significant predictors of under-five severe stunting were maternal age and education, maternal national health insurance status, household wealth status, and number of children under-five in households. Highly significant differences exist in the slopes between 0.1 and 0.9 quantiles. The quantile regression plots for the selected quantiles from 0.1 to 0.9 showed substantial differences in the impact of the covariates across the quantiles of HAZ considered.

CONCLUSION

Critical risk factors that can aid formulation of child nutrition and health policies and interventions that will improve child nutritional outcomes and survival were identified. Modelling under-five severe stunting using multivariable simultaneous quantile regression models could be beneficial to addressing the under-five severe stunting.

Lifetime Costs of Surviving Cancer-A Queensland Study (COS-Q): Protocol of a Large Healthcare Data Linkage Study

Australia-wide, there are currently more than one million cancer survivors. There are over 32 million world-wide. A trend of increasing cancer incidence, medical innovations and extended survival places growing pressure on healthcare systems to manage the ongoing and late effects of cancer treatment. There are no published studies of the long-term health service use and cost of cancer survivorship on a population basis in Australia. All residents of the state of Queensland, Australia, diagnosed with a first primary malignancy from 1997–2015 formed the cohort of interest. State and national healthcare databases are linked with cancer registry records to capture all health service utilization and healthcare costs for 20 years (or death, if this occurs first), starting from the date of cancer diagnosis, including hospital admissions, emergency presentations, healthcare costing data, Medicare services and pharmaceuticals. Data analyses include regression and economic modeling. We capture the whole journey of health service contact and estimate long-term costs of all cancer patients diagnosed and treated in Queensland by linking routinely collected state and national healthcare data. Our results may improve the understanding of lifetime health effects faced by cancer survivors and estimate related healthcare costs. Research outcomes may inform policy and facilitate future planning for the allocation of healthcare resources according to the burden of disease.

Racial Comparisons in Timeliness of Colon Cancer Treatment in an Equal-Access Health System

BACKGROUND

Non-Hispanic black (NHB) adults with cancer may have longer time-to-treatment than non-Hispanic whites (NHW) in the United States. Unequal access to medical care may partially account for this racial disparity. This study aimed to investigate whether there were racial differences in time-to-treatment and in treatment delays for patients diagnosed with colon cancer in the equal-access Military Health System (MHS).

METHODS

Patients age 18-79 years diagnosed with colon adenocarcinoma between January 1, 1998, and December 31, 2014, were identified in the Department of Defense Central Cancer Registry and the MHS Data Repository-linked databases. Median time-to-treatment (surgery and chemotherapy) and 95% confidence intervals were compared between NHBs and NHWs in multivariable quantile regression models. Odds ratios and 95% confidence intervals of receiving delayed treatment defined by guidelines for NHBs relative to NHWs were estimated using multivariable logistic regression.

RESULTS

Patients (n = 3067) had a mean age at diagnosis of 58.4 (12.2) years and the racial distribution was 76.7% NHW and 23.3% NHB. Median adjusted time-to-treatment was similar for NHB compared to NHW patients. The likelihood of receiving delayed treatment was similar between NHB and NHW patients.

CONCLUSIONS

In the MHS, there was no evidence of treatment delays for NHBs compared to NHWs, suggesting the role of equal access to medical care and insurance coverage in reducing racial disparities in colon cancer treatment.

In Utero Exposure to Aluminium and Other Neurotoxic Elements in Urban Coastal South African Women at Delivery: An Emerging Concern

Aluminium (Al) is a non-essential neurotoxicant and there is limited information regarding exposure to Al in utero. This study sought to evaluate the in utero exposure to Al in urban South African women, its effects on birth outcomes and possible synergistic effects between Al, essential and neurotoxic elements such as lead (Pb), mercury (Hg) and arsenic (As), as well as a a potential sex-dependent response to these elements in neonates. This study has found elevated levels of Al in urban women at delivery. The Spearman’s rank correlation coefficients (p-value) of the association between maternal serum Al and birth outcomes (gestational age and parity), and between maternal serum Al and Cu, Zn and Se, were statistically significant. However, in the general and the stratified models, no association was found between any of the birth outcomes and maternal serum Al. The association between maternal serum Al and neurotoxic elements at delivery showed a significant positive correlation for Pb only (rho = 0.361; p < 0.001) which was found to be sex-dependent in neonates (males, rho = 0.285; p < 0.004 and females, rho = 0.444, p < 0.001). Our preliminary findings indicate that in utero exposure to Al is an emerging concern requiring further research and directives from public health authorities.

Quantile function modeling with application to salinity tolerance analysis of plant data

BACKGROUND

In plant science, the study of salinity tolerance is crucial to improving plant growth and productivity under saline conditions. Since quantile regression is a more robust, comprehensive and flexible method of statistical analysis than the commonly used mean regression methods, we applied a set of quantile analysis methods to barley field data. We use univariate and bivariate quantile analysis methods to study the effect of plant traits on yield and salinity tolerance at different quantiles.

RESULTS

We evaluate the performance of barley accessions under fresh and saline water using quantile regression with covariates such as flowering time, ear number per plant, and grain number per ear. We identify the traits affecting the accessions with high yields, such as late flowering time has a negative impact on yield. Salinity tolerance indices evaluate plant performance under saline conditions relative to control conditions, so we identify the traits affecting the accessions with high values of indices using quantile regression. It was observed that an increase in ear number per plant and grain number per ear in saline conditions increases the salinity tolerance of plants. In the case of grain number per ear, the rate of increase being higher for plants with high yield than plants with average yield. Bivariate quantile analysis methods were used to link the salinity tolerance index with plant traits, and it was observed that the index remains stable for earlier flowering times but declines as the flowering time decreases.

CONCLUSIONS

This analysis has revealed new dimensions of plant responses to salinity that could be relevant to salinity tolerance. Use of univariate quantile analyses for quantifying yield under both conditions facilitates the identification of traits affecting salinity tolerance and is more informative than mean regression. The bivariate quantile analyses allow linking plant traits to salinity tolerance index directly by predicting the joint distribution of yield and it also allows a nonlinear relationship between the yield and plant traits.

Racial/Ethnic-Specific Differences in the Effects of Inhaled Corticosteroid Use on Bronchodilator Response in Patients With Asthma

American Thoracic Society guidelines recommend inhaled corticosteroid (ICS) therapy, plus a short-acting bronchodilator, in patients with persistent asthma. However, few prior studies have examined the efficacy of this combination in children of all racial/ethnic groups. We evaluated the association between ICS use and bronchodilator response (BDR) in three pediatric populations with persistent asthma (656 African American, 916 Puerto Rican, and 398 Mexican American children). The association was assessed using multivariable quantile regression. After adjusting for baseline forced expiratory volume in one second and use of controller medications, ICS use was significantly associated with increased BDR only among Mexican Americans (1.56%, P = 0.028) but not African Americans (0.49%, P = 0.426) or Puerto Ricans (0.16%, P = 0.813). Our results demonstrate that ICS augmentation is disproportionate across racial/ethnic groups, where improved BDR is observed in Mexican Americans only. This study highlights the complexities of treating asthma in children, and reinforces the importance of investigating the influence of race/ethnicity on pharmacological response.

Timing of respiratory virus molecular testing in emergency departments and its association with patient care outcomes: a retrospective observational study across six Australian hospitals

OBJECTIVE

A rapid molecular diagnostic test (RMDT) offers a fast and accurate detection of respiratory viruses, but its impact on the timeliness of care in the emergency department (ED) may depend on the timing of the test. The aim of the study was to determine if the timing of respiratory virus testing using a RMDT in the ED had an association with patient care outcomes.

DESIGN

Retrospective observational study.

SETTING

Linked ED and laboratory data from six EDs in New South Wales, Australia.

PARTICIPANTS

Adult patients presenting to EDs during the 2017 influenza season and tested for respiratory viruses using a RMDT. The timing of respiratory virus testing was defined as the time from a patient's ED arrival to time of sample receipt at the hospital laboratory.

OUTCOME MEASURES

ED length of stay (LOS), >4 hour ED LOS and having a pending RMDT result at ED disposition.

RESULTS

A total of 2168 patients were included. The median timing of respiratory virus testing was 224 min (IQR, 133-349). Every 30 min increase in the timing of respiratory virus testing was associated with a 24.0 min increase in the median ED LOS (95% CI, 21.8-26.1; p<0.001), a 51% increase in the likelihood of staying >4 hours in ED (OR, 1.51; 95% CI, 1.41 to 1.63; p<0.001) and a 4% increase in the likelihood of having a pending RMDT result at ED disposition (OR, 1.04; 95% CI, 1.02 to 1.05; p<0.001) after adjustment for confounders.

CONCLUSION

The timing of respiratory virus molecular testing in EDs was significantly associated with a range of outcome indicators. Results suggest the potential to maximise the benefits of RMDT by introducing an early diagnostic protocol such as triage-initiated testing.

How did the tobacco ban increase inmates' body weight during incarceration in Canadian federal penitentiaries? A cohort study

OBJECTIVE

This study aimed to determine how inmates' body weight changed during incarceration in Canadian federal penitentiaries, based on their history of tobacco use. Since tobacco was banned from all Canadian federal penitentiaries in 2008, little is known about the unintended health consequences of this ban, especially on inmates' body weight.

DESIGN

Cohort study.

SETTING

Participants were male and female inmates incarcerated for at least 6 months in Canadian federal penitentiaries. We collected data from 10 institutions in two Canadian regions (Ontario and Atlantic).

PARTICIPANTS

We collected data from 754 inmates who volunteered to participate in the study.

INTERVENTION

This study examined weight change in relation to a history of tobacco use. In 2016-2017, anthropometric data were collected and compared with recorded anthropometric data at the beginning of incarceration (mean follow-up of 5.0±8.3 years). Self-reported data on tobacco and substance use were collected. Weight change was compared between inmates with and without a history of tobacco use.

OUTCOMES

The main outcome measures were body weight change (kg), body mass index (BMI) change (kg/m²), annual weight change (kg/year), and BMI and waist circumference (cm) at the time of the interview.

RESULTS

During incarceration, ex-smokers gained more than twice the amount of weight compared with non-smokers (7.5 kg weight gain for smokers vs 3.7 kg weight gain for non-smokers). Once adjusted for covariates in a regression analysis, for inmates who gained the most weight (75th and 90th percentiles), non-smokers had, respectively, 1.64 and 2.3 lower BMI points than ex-smokers.

CONCLUSIONS

During incarceration in Canadian federal penitentiaries, inmates with a history of tobacco use gained significantly more weight than non-smokers. This put them at increased risk of developing obesity-related health problems. This information is important for the prison setting when planning related programmes and regulation.

Heterogeneity in the costs of medical care among people living with HIV/AIDS in the United States

OBJECTIVE

The costs of medical care for people with HIV/AIDS (PWH) vary substantially across demographic groups, stages of disease progression and regionally across the United States. We aimed to estimate medical costs for PWH and examine the heterogeneity in costs within key patient groups typically distinguished in cost-effectiveness analyses.

DESIGN

Retrospective cohort study using health administrative databases for diagnosed PWH in care at 17 HIV Research Network sites across the United States.

METHODS

We estimated mean quarterly costs for key patient groups using multivariable generalized linear mixed effects models. We used quantile regression to highlight differences in the effect of covariates within each patient group (difference between covariate estimates at the mean versus the 90th percentile of quarterly costs), identifying covariates with a larger effect among the highest cost PWH, or generating greater uncertainty in mean cost estimates.

RESULTS

Our sample included 40 022 patients with a median age of 39 years. Mean quarterly costs were highest for people who inject drugs with advanced disease progression and for PWH on antiretroviral treatment (ART). Within patient groups, we found the most heterogeneity at different levels of resource use for PWH on ART and PWH off ART with CD4 cell counts less than 200 cells/μl, people who inject drugs, as well as PWH in the South.

CONCLUSION

The study quantifies heterogeneity in costs both across and within key PWH patient groups. Our results highlight the need for sensitivity analysis on cost estimates and may inform decisions on model structure in cost-effectiveness analyses on HIV/AIDS treatment and prevention strategies.

Quantile regression for survival data in modern cancer research: expanding statistical tools for precision medicine

Quantile regression links the whole distribution of an outcome to the covariates of interest and has become an important alternative to commonly used regression models. However, the presence of censored data such as survival time, often the main endpoint in cancer studies, has hampered the use of quantile regression techniques because of the incompleteness of data. With the advent of the precision medicine era and availability of high throughput data, quantile regression with high-dimensional predictors has attracted much attention and provided added insight compared to traditional regression approaches. This paper provides a practical guide for using quantile regression for right censored outcome data with covariates of low- or high-dimensionality. We frame our discussion using a dataset from the Boston Lung Cancer Survivor Cohort, a hospital-based prospective cohort study, with the goals of broadening the scope of cancer research, maximizing the utility of collected data, and offering useful statistical alternatives. We use quantile regression to identify clinical and molecular predictors, for example CpG methylation sites, associated with high-risk lung cancer patients, for example those with short survival.

Impact of Rapid Molecular Diagnostic Testing of Respiratory Viruses on Outcomes of Adults Hospitalized with Respiratory Illness: a Multicenter Quasi-experimental Study

A standard multiplex PCR offers comprehensive testing for respiratory viruses. However, it has traditionally been performed in a referral laboratory with a lengthy turnaround time, which can reduce patient flow through the hospital. We aimed to determine whether the introduction of a rapid PCR, but with limited targets (Cepheid Xpert Flu/RSV XC), was associated with improved outcomes for adults hospitalized with respiratory illness. A controlled quasi-experimental study was conducted across three hospitals in New South Wales, Australia. Intervention groups received standard multiplex PCR during the preimplementation, July to December 2016 (n = 953), and rapid PCR during the postimplementation, July to December 2017 (n = 1,209). Control groups (preimplementation, n = 937, and postimplementation, n = 1,102) were randomly selected from adults hospitalized with respiratory illness during the same periods. The outcomes were hospital length of stay (LOS) and microbiology test utilization (blood culture, urine culture, sputum culture, and respiratory bacterial and virus serologies). The introduction of rapid PCR was associated with a nonsignificant 8.9-h reduction in median LOS (95% confidence interval [CI], -21.5 h to 3.7 h; P = 0.17) for all patients and a significant 21.5-h reduction in median LOS (95% CI, -36.8 h to -6.2 h; P < 0.01) among patients with positive test results in an adjusted difference-in-differences analysis. For patients receiving test results before disposition, rapid PCR use was associated with a significant reduction in LOS, irrespective of test results. Compared with standard PCR testing, rapid PCR use was significantly associated with fewer blood culture (adjusted odds ratio [aOR], 0.67; 95% CI, 0.5 to 0.82; P < 0.001), sputum culture (aOR, 0.56; 95% CI, 0.47 to 0.68, P < 0.001), bacterial serology (aOR, 0.44; 95% CI, 0.35 to 0.55, P < 0.001) and viral serology (aOR, 0.42; 95% CI, 0.33 to 0.53, P < 0.001) tests, but not with fewer urine culture tests (aOR, 0.94; 95% CI, 0.78 to 1.12, P = 0.48). Rapid PCR testing of adults hospitalized with respiratory illnesses can deliver benefits to patients and reduce resource utilization. Future research should consider a formal economic analysis and assess its potential impacts on clinical decision making.

Racial Differences in Time to Breast Cancer Surgery and Overall Survival in the US Military Health System

Importance

Racial disparities in time to surgery (TTS) after a breast cancer diagnosis and whether these differences account for disparities in overall survival have been understudied in the US population.

Objectives

To compare TTS in non-Hispanic black (NHB) and non-Hispanic white (NHW) women with breast cancer and to examine whether racial differences in TTS may explain possible racial disparities in overall survival in a universal health care system.

Design, Setting, and Participants

Retrospective cohort identified from the Department of Defense Central Cancer Registry and Military Health System Data Repository linked databases containing records between January 1, 1998, and December 31, 2008, of 998 NHB women and 3899 NHW women who received a diagnosis of stages I to III breast cancer and underwent breast-conserving surgery (BCS) or mastectomy in the US Military Health System during the study period. Data analyses were conducted from July 5, 2017, to December 29, 2017.

Main Outcomes and Measures

The main outcome was time to breast cancer surgery. Non-Hispanic black and NHW women were compared at the 25th, 50th (median), 75th, and 90th percentiles of TTS by using multivariable quantile regression. Cox proportional hazards regression models were used to estimate hazard ratios (HRs) and 95% CIs for all-cause death in NHB compared with NHW women after controlling for potential confounders first without and then with TTS.

Results

Among the 4887 NHB and NHW women in the cohort, the mean (SD) age was 50.0 (9.4) years. The median TTS was 21 days (95% CI, 20.6-21.4 days) among NHW women and 22 days (95% CI, 20.6-23.4 days) among NHB women. Non-Hispanic black women had a significantly greater estimated TTS at the 75th (3.6 days; 95% CI, 1.6-5.5 days) and 90th (8.9 days; 95% CI, 5.1-12.6 days) percentiles than NHW women in multivariable models. The estimated differences were similar by surgery type. Non-Hispanic black women had a higher adjusted risk for death (HR, 1.45; 95% CI, 1.06-2.01) compared with NHW women among patients receiving breast-conserving surgery. The risks were similar between races among those receiving mastectomy (HR, 1.06; 95% CI, 0.76-1.48). The HRs remained similar after adding TTS to the Cox proportional hazards regression models.

Conclusions and Relevance

This study's results indicate that time to breast cancer surgery was delayed for NHB compared with NHW women in the Military Health System. However, the racial differences in TTS did not explain the observed racial differences in overall survival among women who received breast-conserving surgery.

An exploration of reported food intake among inmates who gained body weight during incarceration in Canadian federal penitentiaries

Background

Canadian penitentiaries have recently been shown to be obesogenic. However, little is known about the eating habits of inmates who gained weight while living in the prison environment.

Methods

This retrospective cohort study examined the reported food intake of inmates during incarceration in federal penitentiaries. During a face to face interview, anthropometric measures (2016–2017) were taken and compared to anthropometric data at the beginning of incarceration (mean follow-up of 5.0 ± 8.3 years). Self-reported data on food intake were collected via a food frequency questionnaire.

Results

Inmates who gained the most weight (15.7 kg) during incarceration reported not eating vegetables. They were followed by inmates who gained 14.3 kg and reported not eating fruit. Other inmates who gained a significant amount of weight reported not eating cereal, dairy or legumes. Moreover, inmates’ weight gain was also assessed by special diets: inmates following a religious diet (4.5 kg) or a diet of conscience (-0.3 kg) gained less weight than inmates not following a diet (5.8 kg). In comparison to other types of diets, inmates on a medical diet gained the most weight (7.5 kg). Furthermore, inmates who gained significant weight (8.0 kg) also reported not purchasing healthy foods from the commissary store (or “canteen”), whereas inmates who gained less weight (4.8 kg) reported purchasing healthy foods from the commissary store (or “canteen”). The observed weight gain was positively associated with food purchased from the commissary store (or “canteen”), but was not associated with the feeding system of the penitentiary (tray, cafeteria or meal plan).

Discussion

Food intake during incarceration is a modifiable risk factor that could be the target of weight management interventions with inmates. Our findings suggest that inmates who gained the most weight also reported having low intake of foods deemed healthy (vegetables, fruit, cereal, dairy and legumes) from food services and from the commissary store (or “canteen”) purchases.

A quantile regression approach to explain the relationship of Fatigue and Cortisol, Cytokine among Koreans with Hepatitis B

Fatigue is a major symptom among patients with Hepatitis B virus (HBV). However, the physiological mechanisms regarding mediate fatigue and the relationships between fatigue, cortisol and cytokines are unclear in patients with HBV. The purpose of this study was to examine the relationships between perceived fatigue, cortisol, and cytokines in Korean patients with HBV. The mean score for overall fatigue was moderate. In linear regression analyses, TNF-α was only significant as a predictor of fatigue. In quantile regression analyses, the cortisol level was significant in the 70^th, 80^th, 90^th quantiles in the cognitive/mood fatigue subdimension, which correspond to severe levels of fatigue. IL-6 levels were significant in 90^th quantile in overall fatigue and in the cognitive/mood fatigue subdimension. Cortisol, IL-6, and TNF- α were related to perceived fatigue in patients with HBV, particularly in the cognitive/mood dimension. This suggests that the fatigue patterns of patients with HBV are related to their mental and mood states and physical responses, such as cortisol and cytokine levels, are correlated with the nature of the fatigue. In the clinic, interventions to manage fatigue in patients with HBV should be developed based on the characteristics of the fatigue.

Association of stressful life events and psychological problems profile: Results from a large-scale cross-sectional study among Iranian industrial employees using Bayesian quantile structural equation model

Objectives: The current study aimed at evaluating the major domains of stressful life events and their association with psychological problems profile in a large sample of Iranian industrial manufacturing employees. Methods: In a cross-sectional study, 3,063 participants were randomly selected from 16,000 employees working in a big industrial company in Isfahan, Iran. Three common psychological problems i.e. depression, anxiety and psychological distress were evaluated using Persian validated version of Hospital Anxiety and Depression Scale (HADS) and 12-item General Health Questionnaire (GHQ-12), respectively. Self-perceived frequency and intensity of stressful life events were measured by stressful life event (SLE) questionnaire. Bayesian quantile structural equation model in R free statistical software (version 3.4) was used for evaluating the association of stressful life events and levels of psychological problems profile. Results: Using factor analysis, two major domains i.e. socioeconomic and personal stressors were derived from 11 life stressors and a unified measure i.e. psychological problem profile was extracted from three common psychological problems. Financial and daily life stressors had the highest and sexual problems showed lowest intensity. Quantile structural equation model revealed that the psychological problems profile scores had stronger association with personal (β: ranging from 0.45, 1.87) than socioeconomics stressors (β: ranging from 0.11, 0.27). The association of socioeconomic stressors was fairly uniform across quantiles of psychological problems scores, while personal stressors showed stronger association in higher quantiles, meaning that employees with higher mental health problems more experienced personal stressors. Conclusions: Life stressors particularly personal showed direct association with intensity of psychological problems in manufacturing employees. Life stressors are more perceived by employees with higher intensity of psychological problems. The results of current study can be useful in planning occupational health programs in order to improve psychological health and job productivity.

US particulate matter air quality improves except in wildfire-prone areas

Using data from rural monitoring sites across the contiguous United States, we evaluated fine particulate matter (PM2.5) trends for 1988-2016. We calculate trends in the policy-relevant 98th quantile of PM2.5 using Quantile Regression. We use Kriging and Gaussian Geostatistical Simulations to interpolate trends between observed data points. Overall, we found positive trends in 98th quantile PM2.5 at sites within the Northwest United States (average 0.21 ± 0.12 µg·m-3·y-1; ±95% confidence interval). This was in contrast with sites throughout the rest of country, which showed a negative trend in 98th quantile PM2.5, likely due to reductions in anthropogenic emissions (average -0.66 ± 0.10 µg·m-3·y-1). The positive trend in 98th quantile PM2.5 is due to wildfire activity and was supported by positive trends in total carbon and no trend in sulfate across the Northwest. We also evaluated daily moderate resolution imaging spectroradiometer (MODIS) aerosol optical depth (AOD) for 2002-2017 throughout the United States to compare with ground-based trends. For both Interagency Monitoring of Protected Visual Environments (IMPROVE) PM2.5 and MODIS AOD datasets, we found positive 98th quantile trends in the Northwest (1.77 ± 0.68% and 2.12 ± 0.81% per year, respectively) through 2016. The trend in Northwest AOD is even greater if data for the high-fire year of 2017 are included. These results indicate a decrease in PM2.5 over most of the country but a positive trend in the 98th quantile PM2.5 across the Northwest due to wildfires.

US particulate matter air quality improves except in wildfire-prone areas

Using data from rural monitoring sites across the contiguous United States, we evaluated fine particulate matter (PM_2.5) trends for 1988-2016. We calculate trends in the policy-relevant 98th quantile of PM_2.5 using Quantile Regression. We use Kriging and Gaussian Geostatistical Simulations to interpolate trends between observed data points. Overall, we found positive trends in 98th quantile PM_2.5 at sites within the Northwest United States (average 0.21 ± 0.12 µg·m^-3·y^-1; ±95% confidence interval). This was in contrast with sites throughout the rest of country, which showed a negative trend in 98th quantile PM_2.5, likely due to reductions in anthropogenic emissions (average -0.66 ± 0.10 µg·m^-3·y^-1). The positive trend in 98th quantile PM_2.5 is due to wildfire activity and was supported by positive trends in total carbon and no trend in sulfate across the Northwest. We also evaluated daily moderate resolution imaging spectroradiometer (MODIS) aerosol optical depth (AOD) for 2002-2017 throughout the United States to compare with ground-based trends. For both Interagency Monitoring of Protected Visual Environments (IMPROVE) PM_2.5 and MODIS AOD datasets, we found positive 98th quantile trends in the Northwest (1.77 ± 0.68% and 2.12 ± 0.81% per year, respectively) through 2016. The trend in Northwest AOD is even greater if data for the high-fire year of 2017 are included. These results indicate a decrease in PM_2.5 over most of the country but a positive trend in the 98th quantile PM_2.5 across the Northwest due to wildfires.

Suboptimal Iodine Status and Low Iodine Knowledge in Young Norwegian Women

Previous studies have documented mild to moderate iodine deficiency in pregnant and lactating women in Norway. This study focused on non-pregnant young women because their future children may be susceptible to the adverse effects of iodine deficiency. We assessed urinary iodine concentration (UIC), iodine intake from food and supplements, and iodine knowledge in 403 non-pregnant women, mainly students, aged 18–30 years. Iodine concentration was measured in spot urine samples analyzed by inductively coupled plasma mass spectrometry and iodine intake was calculated from a self-reported food frequency questionnaire. Knowledge about iodine was collected through the self-administered, paper-based questionnaire. Median (p25–p75) UIC was 75 (42–130) µg/L and 31% had UIC < 50 µg/L. Habitual iodine intake was 100 (69–136) µg/day. In multiple regression models, supplemental iodine, use of thyroid medication, and iodine intake from food were positively associated with UIC, while vegetarian practice was negatively associated, explaining 16% of the variance. Approximately 40% of the young women had low iodine knowledge score and no differences were found between women in different study programs. Young women in Norway are mild to moderately iodine-deficient, and public health strategies are needed to improve and secure adequate iodine status.

Quantile regression analysis of modifiable and non-modifiable drivers' of blood pressure among urban and rural women in Ghana

High blood pressure is an increasingly problematic public health concern in many developing countries due to the associated cardiovascular and renal complications. This study set out to investigate the drivers of blood pressure among urban and rural women using the 2014 Ghana Demographic and Health Survey data. Diastolic blood pressure (DBP) and systolic blood pressure (SBP) were the outcomes of interest. Our findings showed that body mass index (BMI) had a significant positive effect on DBP and SBP in both urban and rural settings, with the largest effect occurring among women in the 75^th quantile. Arm circumference also had a positive effect on DBP and SBP across all quantiles in both settings. Age had an increasing positive effect along the entire conditional DBP and SBP distribution in both settings. Women who were pregnant had lower DBP and SBP relative to those who were not pregnant in both settings. These results highlight the important drivers of DBP and SBP, and the differential effects of these drivers on blood pressure (BP) among women in urban and rural settings. To increase their effectiveness, interventions to address high BP should take into account these differential effects.

Modeling Linguistic Variables With Regression Models: Addressing Non-Gaussian Distributions, Non-independent Observations, and Non-linear Predictors With Random Effects and Generalized Additive Models for Location, Scale, and Shape

As statistical approaches are getting increasingly used in linguistics, attention must be paid to the choice of methods and algorithms used. This is especially true since they require assumptions to be satisfied to provide valid results, and because scientific articles still often fall short of reporting whether such assumptions are met. Progress is being, however, made in various directions, one of them being the introduction of techniques able to model data that cannot be properly analyzed with simpler linear regression models. We report recent advances in statistical modeling in linguistics. We first describe linear mixed-effects regression models (LMM), which address grouping of observations, and generalized linear mixed-effects models (GLMM), which offer a family of distributions for the dependent variable. Generalized additive models (GAM) are then introduced, which allow modeling non-linear parametric or non-parametric relationships between the dependent variable and the predictors. We then highlight the possibilities offered by generalized additive models for location, scale, and shape (GAMLSS). We explain how they make it possible to go beyond common distributions, such as Gaussian or Poisson, and offer the appropriate inferential framework to account for 'difficult' variables such as count data with strong overdispersion. We also demonstrate how they offer interesting perspectives on data when not only the mean of the dependent variable is modeled, but also its variance, skewness, and kurtosis. As an illustration, the case of phonemic inventory size is analyzed throughout the article. For over 1,500 languages, we consider as predictors the number of speakers, the distance from Africa, an estimation of the intensity of language contact, and linguistic relationships. We discuss the use of random effects to account for genealogical relationships, the choice of appropriate distributions to model count data, and non-linear relationships. Relying on GAMLSS, we assess a range of candidate distributions, including the Sichel, Delaporte, Box-Cox Green and Cole, and Box-Cox t distributions. We find that the Box-Cox t distribution, with appropriate modeling of its parameters, best fits the conditional distribution of phonemic inventory size. We finally discuss the specificities of phoneme counts, weak effects, and how GAMLSS should be considered for other linguistic variables.

Inadequate Iodine Intake in Population Groups Defined by Age, Life Stage and Vegetarian Dietary Practice in a Norwegian Convenience Sample

Inadequate iodine intake has been identified in populations considered iodine replete for decades. The objective of the current study is to evaluate urinary iodine concentration (UIC) and the probability of adequate iodine intake in subgroups of the Norwegian population defined by age, life stage and vegetarian dietary practice. In a cross-sectional survey, we assessed the probability of adequate iodine intake by two 24-h food diaries and UIC from two fasting morning spot urine samples in 276 participants. The participants included children (n = 47), adolescents (n = 46), adults (n = 71), the elderly (n = 23), pregnant women (n = 45), ovo-lacto vegetarians (n = 25), and vegans (n = 19). In all participants combined, the median (95% CI) UIC was 101 (90, 110) µg/L, median (25th, 75th percentile) calculated iodine intake was 112 (77, 175) µg/day and median (25th, 75th percentile) estimated usual iodine intake was 101 (75, 150) µg/day. According to WHOs criteria for evaluation of median UIC, iodine intake was inadequate in the elderly, pregnant women, vegans and non-pregnant women of childbearing age. Children had the highest (82%) and vegans the lowest (14%) probability of adequate iodine intake according to reported food and supplement intakes. This study confirms the need for monitoring iodine intake and status in nationally representative study samples in Norway.

The relative age effect reversal among the National Hockey League elite

Like many sports in adolescence, junior hockey is organized by age groups. Typically, players born after December 31^st are placed in the subsequent age cohort and as a result, will have an age advantage over those players born closer to the end of the year. While this relative age effect (RAE) has been well-established in junior hockey and other professional sports, the long-term impact of this phenomenon is not well understood. Using roster data on North American National Hockey League (NHL) players from the 2008–2009 season to the 2015–2016 season, we document a RAE reversal—players born in the last quarter of the year (October-December) score more and command higher salaries than those born in the first quarter of the year. This reversal is even more pronounced among the NHL “elite.” We find that among players in the 90^th percentile of scoring, those born in the last quarter of the year score about 9 more points per season than those born in the first quarter. Likewise, elite players in the 90^th percentile of salary who are born in the last quarter of the year earn 51% more pay than players born at the start of the year. Surprisingly, compared to players at the lower end of the performance distribution, the RAE reversal is about three to four times greater among elite players.