Bayesian Mixed-effects Model Research Articles

One-Year Mortality Is Not a Reliable Indicator of Lung Transplant Center Performance

BackgroundIn the United States, the Organ Procurement and Transplant Network (OPTN) uses 1-year mortality as the primary measure of transplant center quality. We sought to evaluate the reliability of mortality outcomes in lung transplantation and to compare statistical methods of program performance evaluation. MethodsWe used the Standard Transplant Analysis and Research files from the United Network for Organ Sharing to identify lung transplant recipients from 2013 to 2018 in the United States. We stratified hospitals on the basis of 30-day, 1-year, and 5-year survival by risk adjustment, reliability adjustment with empirical Bayes technique, and hierarchical bayesian mixed effects models currently used by the OPTN. We measured variation in mortality rates and identification of performance outliers between techniques. ResultsWe identified 12,769 recipients in 69 centers. Reliability adjustment reduced variation in hospital outcomes and had a large impact on hospital mortality rankings. For example, with 1-year mortality, 28% (5 hospitals) of the “best” hospitals (top 25%) and 18% (3 hospitals) of the “worst” hospitals (bottom 25%) were reclassified after reliability adjustment. The overall reliability of 1-year mortality was low at 0.42. Compared with the bayesian method used by the OPTN, reliability adjustment identified fewer outliers. The 5-year survival reached a higher reliability plateau with a lower volume of cases required. ConclusionsThe reliability of 1-year mortality in lung transplantation is low, whereas 5-year survival estimates may be more reliable at lower case volumes. Reliability adjustment yielded more conservative measures of center performance and fewer outliers compared with current bayesian methods.

Bayesian nonparametric multivariate spatial mixture mixed effects models with application to American Community Survey special tabulations

Leveraging multivariate spatial dependence to improve the precision of estimates using American Community Survey data and other sample survey data has been a topic of recent interest among data users and federal statistical agencies. One strategy is to use a multivariate spatial mixed effects model with a Gaussian observation model and latent Gaussian process model. In practice, this works well for a wide range of tabulations. Nevertheless, in situations in which the data exhibit heterogeneity within or across geographies, and/or there is sparsity in the data, the Gaussian assumptions may be problematic and lead to underperformance. To remedy these situations, we propose a multivariate hierarchical Bayesian nonparametric mixed effects spatial mixture model to increase model flexibility. The number of clusters is chosen automatically in a data-driven manner. The effectiveness of our approach is demonstrated through a simulation study and motivating application of special tabulations for American Community Survey data.

Allergen immunotherapy in MASK‐air users in real‐life: Results of a Bayesian mixed‐effects model

BackgroundEvidence regarding the effectiveness of allergen immunotherapy (AIT) on allergic rhinitis has been provided mostly by randomised controlled trials, with little data from real‐life studies.ObjectiveTo compare the reported control of allergic rhinitis symptoms in three groups of users of the MASK‐air® app: those receiving sublingual AIT (SLIT), those receiving subcutaneous AIT (SCIT), and those receiving no AIT.MethodsWe assessed the MASK‐air® data of European users with self‐reported grass pollen allergy, comparing the data reported by patients receiving SLIT, SCIT and no AIT. Outcome variables included the daily impact of allergy symptoms globally and on work (measured by visual analogue scales—VASs), and a combined symptom‐medication score (CSMS). We applied Bayesian mixed‐effects models, with clustering by patient, country and pollen season.ResultsWe analysed a total of 42,756 days from 1,093 grass allergy patients, including 18,479 days of users under AIT. Compared to no AIT, SCIT was associated with similar VAS levels and CSMS. Compared to no AIT, SLIT‐tablet was associated with lower values of VAS global allergy symptoms (average difference = 7.5 units out of 100; 95% credible interval [95%CrI] = −12.1;−2.8), lower VAS Work (average difference = 5.0; 95%CrI = −8.5;−1.5), and a lower CSMS (average difference = 3.7; 95%CrI = −9.3;2.2). When compared to SCIT, SLIT‐tablet was associated with lower VAS global allergy symptoms (average difference = 10.2; 95%CrI = −17.2;−2.8), lower VAS Work (average difference = 7.8; 95%CrI = −15.1;0.2), and a lower CSMS (average difference = 9.3; 95%CrI = −18.5;0.2).ConclusionIn patients with grass pollen allergy, SLIT‐tablet, when compared to no AIT and to SCIT, is associated with lower reported symptom severity. Future longitudinal studies following internationally‐harmonised standards for performing and reporting real‐world data in AIT are needed to better understand its ‘real‐world’ effectiveness.

Behavioural reactions to oncoming vehicles as a crucial aspect of wildlife-vehicle collision risk in three common wildlife species

Wildlife-vehicle collisions (WVC) strongly impact road safety. While technical aspects of collision risk and the effects of roads on animal populations are well studied, knowledge about wildlife behaviour prior to and during contact with oncoming vehicles as a crucial aspect of collision risk is still lacking. We analysed 28,400 hours of video data (thermal network cameras at 14 road sections in south-west Germany) with 2,841 animal-vehicle encounters (1,960 roe deer, Capreolus capreolus, 696 red fox, Vulpes vulpes and 185 wild boar, Sus scrofa) and classified animal behaviour before and during contact with a vehicle. We fitted two sets of models to the data. In the first step, we modelled the intensity of the behavioural reaction exhibited by the animals as a function of behavioural and environmental predictors using ordinal Bayesian mixed-effect regression models. In a second step, we modelled the probability of a positive vs. a negative behavioural response in terms of WVC risk using binomial mixed-effect regression models. Both the intensity of behavioural reactions as well as the degree of risk during the interaction with oncoming vehicles differed among the species and as a function of road section layout. Our results showed that animal attentiveness, the behaviour a priori, access to cover, vehicle type and biological seasonality were important predictors of an animal’s response to oncoming vehicles. More specifically, roe deer reacted to oncoming vehicles mostly with short movements away from the road, foxes often reacted unpredictably and wild boar behaviour appeared to be least affected by oncoming vehicles. Thus, we suggest that collision risk for common European mammals is shaped by the interplay of vehicle type, the road layout as well as the species-specific behavioural repertoire including the attentiveness of the animal and the behavioural state prior to an approaching vehicle. In addition, wildlife warning reflectors, a frequently used technique in WVC mitigation, did not alter behavioural reactions and thus failed to reduce WVC risk.

Differential impact of government lockdown policies on reducing air pollution levels and related mortality in Europe

Previous studies have reported a decrease in air pollution levels following the enforcement of lockdown measures during the first wave of the COVID-19 pandemic. However, these investigations were mostly based on simple pre-post comparisons using past years as a reference and did not assess the role of different policy interventions. This study contributes to knowledge by quantifying the association between specific lockdown measures and the decrease in NO2, O3, PM2.5, and PM10 levels across 47 European cities. It also estimated the number of avoided deaths during the period. This paper used new modelled data from the Copernicus Atmosphere Monitoring Service (CAMS) to define business-as-usual and lockdown scenarios of daily air pollution trends. This study applies a spatio-temporal Bayesian non-linear mixed effect model to quantify the changes in pollutant concentrations associated with the stringency indices of individual policy measures. The results indicated non-linear associations with a stronger decrease in NO2 compared to PM2.5 and PM10 concentrations at very strict policy levels. Differences across interventions were also identified, specifically the strong effects of actions linked to school/workplace closure, limitations on gatherings, and stay-at-home requirements. Finally, the observed decrease in pollution potentially resulted in hundreds of avoided deaths across Europe.

The effect of hand hygiene frequency on reducing acute respiratory infections in the community: a meta-analysis.

Hand hygiene is a simple, low-cost intervention that may lead to substantial population-level effects in suppressing acute respiratory infection epidemics. However, quantification of the efficacy of hand hygiene on respiratory infection in the community is lacking. We searched PubMed for randomised controlled trials on the effect of hand hygiene for reducing acute respiratory infections in the community published before 11 March 2021. We performed a meta-regression analysis using a Bayesian mixed-effects model. A total of 105 publications were identified, out of which six studies reported hand hygiene frequencies. Four studies were performed in household settings and two were in schools. The average number of handwashing events per day ranged from one to eight in the control arms, and four to 17 in the intervention arms. We estimated that a single hand hygiene event is associated with a 3% (80% credible interval (-1% to 7%)) decrease in the daily probability of an acute respiratory infection. Three of these six studies were potentially at high risk of bias because the primary outcome depended on self-reporting of upper respiratory tract symptoms. Well-designed trials with an emphasis on monitoring hand hygiene adherence are needed to confirm these findings.

Eye Closure Interacts with Music to Influence Vividness and Content of Directed Imagery

Goal-directed, intentional mental imagery generation supports a range of daily self-regulatory activities, such as planning, decision-making, and recreational escapism. Many clinical interventions for mood and anxiety disorders also use imagery and their effectiveness can be affected by an individual's ability to manipulate vividness and content of mental imagery. Prior literature points towards music as a promising candidate to influence imagination in such settings, but basic questions remain regarding how music affects mental imagery and how it interacts with basic, well-established parameters, such as facilitatory effects of eye closure. One hundred participants listened to music and a silent control whilst performing a guided mental imagery task. Specifically, participants saw a short video of a figure journeying towards a landmark and imagined a continuation of the journey with either closed or open eyes. After each trial, participants reported vividness and content of their imagined journeys. Bayesian Mixed Effects Models obtained strong evidence of greater vividness, duration, as well as distance travelled in music conditions compared to silent conditions. Additionally, interactive effects of music and eye closure were found for both vividness and the emotional valence of imagined content, where music effects were disproportionately amplified by eye closure. Findings further support music's potential to manipulate the perceptual, spatial-temporal, as well as emotional sentiment of deliberately generated mental imagery.

Physical activity is associated with low-level stress among a sample of seniors during COVID-19 outbreak?

This study aimed to verify the association of physical activity, behaviors, and health conditions with stress among a sample of older adults from Manaus, Brazil, during the COVID-19 outbreak. This is a cross-sectional analysis of 79 older adults (76% female; age: 68.24 ± 6.82 years) from Fundação Universidade Aberta da Terceira Idade (FUnATI). The perceived stress (PS) was assessed by the Perceived Stress Scale, while a questionnaire via telephone call. The variables also included physical activity, sleep quality, sociodemographic, and health condition variables. Linear Bayesian mixed-effects regression models were applied to verify the association between these correlates with PS. Older people who met the recommendations before the COVID-19 outbreak but did not meet the same recommendations during the COVID-19 outbreak have a higher PS score compared to those older people who did not meet the PA recommendations at either time. In contrast, those elderly people who met PA recommendations before and during the COVID-19 outbreak had a lower PS score compared to their peers who did not meet physical activity recommendations at either time. Those older people who did not meet the PA recommendations before the COVID-19 outbreak, but did meet the PA recommendations during the COVID-19 outbreak, did not have significant differences in the PS score compared to their peers who did not meet the PA recommendations in either moment. Sleep changes and high BMI were negatively associated with PS. The maintenance of physical activity during the pandemic was associated with lower levels of stress in older people.

134Trends and projections of Treatment as Prevention strategy in Africa and progress towards UNAIDS targets

Abstract Background Africa is lagging behind in progress toward global HIV/AIDS elimination targets. We estimated trends in and projections of annual HIV testing and condom use at last higher-risk sex; and calculated the probability of reaching key UNAIDS’s target. Methods We included 114 nationally-representative datasets in 38 African countries from DHS and MICS with 1456224 sexually active adults age 15-49 from 2003 to 2018. We applied Bayesian mixed effect models to estimate the coverage of annual HIV testing and condom use at last higher-risk sex for every country and year to 2030; and the probability of reaching UNAIDS testing and condom use targets of 95% coverage by 2030. Results Seven countries saw downward trends in annual HIV testing and four saw decreases in condom use at higher-risk sex, while most countries have upward trends in both indicators. The highest coverages of testing in 2030 are predicted in Swaziland, Uganda and Lesotho. Meanwhile, Swaziland, Lesotho, and Namibia will have the highest proportion of condom use in 2030. The probabilities of reaching targets were very low for both HIV testing (0%-28.5%) and condom use (0%-12.1%). Conclusions We observed limited progress on annual HIV testing and condom use at last higher-risk sex in Africa, and little prospect of reaching global targets for HIV/AIDS elimination. Key messages Although some global sponsor agencies are considering withdrawal from supporting Africa, more considerable attention to funding and expanding testing and treatment are needed if HIV/AIDS elimination in Africa is to become more than a distant dream.

Impact of the COVID-19 lockdown policies on reducing air pollution levels and related deaths in Europe

BACKGROUND AND AIM: Previous studies have reported a decrease in air pollution following the enforcement of lockdown measures during the first wave of the COVID-19 pandemic. However, these investigations were mostly based on simple pre-post comparisons using past years as a reference, and did not assess the role of different policy interventions. In this contribution, we quantitatively evaluated the association between various lockdown measures and the decrease in NO2, O3, PM2.5, and PM10 levels across 47 European cities and the associated short-term mortality in the period of February-July 2020. METHODS: We used data from several chemical transport models developed by the Copernicus Atmosphere Monitoring Service (CAMS) to define trends in air pollution under business-as-usual and lockdown scenarios, thus removing differences due to weather conditions and other differences affecting pre-post comparisons. We then applied an advanced spatio-temporal Bayesian non-linear mixed effect model to determine the association with stringency indices of individual policy measures, allowing non-linear relationships and geographical correlations. RESULTS:The findings indicate evidence of non-linear relationships, with a stronger decrease in NO2 and to a lesser extent PMs under very strict lockdown regimes. The effects of lockdown measures vary geographically, with a stronger decline in pollution in Southern and Central Europe. The comparative analysis of separate lockdown policies suggests important differences across interventions. Specifically, actions linked to school/workplace closure, limitations on gatherings, and stay-at-home requirements had strong effects, while restrictions on internal movement and international travels showed little impact. The observed decrease in pollution potentially resulted in hundreds of avoided deaths across the European cities. CONCLUSIONS:This study provides important evidence on the differential impacts of various policies implemented during the COVID-19 pandemic in decreasing the level of pollutants in urban areas across Europe. KEYWORDS: air pollution decline, Covid-19 Government Response, chemical transport model, Bayesian mixed effect model, mortality

The effect of low-frequency equalisation on preference and sensorimotor synchronisation in music

Equalisation, a signal processing technique commonly used to shape the sound of music, is defined as the adjustment of the energy in specific frequency components of a signal. In this work, we investigate the effects of equalisation on preference and sensorimotor synchronisation in music. A total of 21 participants engaged in a goal-directed upper body movements in synchrony with stimuli equalised in three low-frequency sub-bands (0–50, 50–100, and 100–200 Hz). To quantify the effect of equalisation, music features including spectral flux, pulse clarity, and beat confidence were extracted from seven differently equalised versions of music tracks—one original and six manipulated versions for each music track. These music tracks were then used in a movement synchronisation task. Bayesian mixed-effects models revealed different synchronisation behaviours in response to the three sub-bands considered. Boosting energy in the 100–200 Hz sub-band reduced synchronisation performance irrespective of the sub-band energy of the original version. An energy boost in the 0–50 Hz band resulted in increased synchronisation performance only when the sub-band energy of the original version was high. An energy boost in the 50–100 Hz band increased synchronisation performance only when the sub-band energy of the original version was low. Boosting the energy in any of the three sub-bands increased preference regardless of the energy of the original version. Our results provide empirical support for the importance of low-frequency information for sensorimotor synchronisation and suggest that the effects of equalisation on preference and synchronisation are largely independent of one another.

Music influences vividness and content of imagined journeys in a directed visual imagery task

Directed, intentional imagination is pivotal for self-regulation in the form of escapism and therapies for a wide variety of mental health conditions, such anxiety and stress disorders, as well as phobias. Clinical application in particular benefits from increasing our understanding of imagination, as well as non-invasive means of influencing it. To investigate imagination, this study draws from the prior observation that music can influence the imagined content during non-directed mind-wandering, as well as the finding that relative orientation within time and space is retained in imagination. One hundred participants performed a directed imagination task that required watching a video of a figure travelling towards a barely visible landmark, and then closing their eyes and imagining a continuation of the journey. During each imagined journey, participants either listened to music or silence. After the imagined journeys, participants reported vividness, the imagined time passed and distance travelled, as well as the imagined content. Bayesian mixed effects models reveal strong evidence that vividness, sentiment, as well imagined time passed and distances travelled, are influenced by the music, and show that aspects of these effects can be modelled through features such as tempo. The results highlight music’s potential to support therapies such as Exposure Therapy and Imagery Rescripting, which deploy directed imagination as a clinical tool.

Impact of steep slope management system and row orientation on canopy microclimate. Comparing terraces to downslope vineyards

Historically and under cooler climate conditions, steep slope vineyards yielded best quality wines and highest reputation, due to their distinctive microclimate, especially during ripening period. Nevertheless, steep slope vineyard sites primarily suffer from reduced competitiveness leading to abandonment, thus a loss of valuable vineyard sites. The aim of this work was to investigate differences in microclimatic conditions between different steep slope vineyard management systems and row orientations. Records of inner canopy microclimatic parameters were taken over two consecutive vegetation periods including seven vineyard pairs. A Bayesian mixed effect model was used to properly account for the complexity of the conducted experiment. Additionally, irradiation and canopy surface temperature data was compared. Grapevines planted downslope (control) exhibited a more even light distribution on canopy sides. Contrarily, at terraced vineyard sites canopy sides showed big differences regarding light interception, also affecting diurnal canopy surface temperature. Differences in N-S/E-W row orientation comparison were more pronounced compared to vineyard pairs aligned NE-SW/NW-SE. Night temperatures were slightly higher in terraced vineyards, while daily mean and maximum temperature and temperature amplitude were higher in vineyard rows planted in line of the greatest slope. While a treatment effect was not always clearly visible, an impact of row orientation on temperature microclimate was observed. Differences between treatments may become more pronounced under projected climate change conditions with consequences on physiological processes, thus grapevine performance influencing berry composition.

Hospital utilization rates for influenza and RSV: a novel approach and critical assessment

BackgroundInfluenza and respiratory syncytial virus (RSV) contribute significantly to the burden of acute lower respiratory infection (ALRI) inpatient care, but heterogeneous coding practices and availability of inpatient data make it difficult to estimate global hospital utilization for either disease based on coded diagnoses alone.MethodsThis study estimates rates of influenza and RSV hospitalization by calculating the proportion of ALRI due to influenza and RSV and applying this proportion to inpatient admissions with ALRI coded as primary diagnosis. Proportions of ALRI attributed to influenza and RSV were extracted from a meta-analysis of 360 total sources describing inpatient hospital admissions which were input to a Bayesian mixed effects model over age with random effects over location. Results of this model were applied to inpatient admission datasets for 44 countries to produce rates of hospital utilization for influenza and RSV respectively, and rates were compared to raw coded admissions for each disease.ResultsFor most age groups, these methods estimated a higher national admission rate than the rate of directly coded influenza or RSV admissions in the same inpatient sources. In many inpatient sources, International Classification of Disease (ICD) coding detail was insufficient to estimate RSV burden directly. The influenza inpatient burden estimates in older adults appear to be substantially underestimated using this method on primary diagnoses alone. Application of the mixed effects model reduced heterogeneity between countries in influenza and RSV which was biased by coding practices and between-country variation.ConclusionsThis new method presents the opportunity of estimating hospital utilization rates for influenza and RSV using a wide range of clinical databases. Estimates generally seem promising for influenza and RSV associated hospitalization, but influenza estimates from primary diagnosis seem highly underestimated among older adults. Considerable heterogeneity remains between countries in ALRI coding (i.e., primary vs non-primary cause), and in the age profile of proportion positive for influenza and RSV across studies. While this analysis is interesting because of its wide data utilization and applicability in locations without laboratory-confirmed admission data, understanding the sources of variability and data quality will be essential in future applications of these methods.

Elites Do Not Deplete - No Effect of Prior Mental Exertion on Subsequent Shooting Performance in Elite Shooters.

In order to perform at the highest level, elite shooters have to remain focused during the whole course of a tournament, which regularly lasts multiple hours. Investing self-control over extended time periods is often associated with lower levels of perceived self-control strength (i.e., the subjective estimation of how much mental effort one is capable of investing in a given task) and impaired performance in several sports-related domains. However, previous findings on the effects of prior self-control efforts on shooting performance have been mixed, as elite shooters seem to be less affected by preceding self-control demanding tasks than sub-elite athletes. Therefore, the aim of the present study was to investigate the effects of self-control on shooting performance in elite shooters. Hence, we randomly assigned elite shooters to an experimental (n = 12) or a control condition (n = 11) and asked them to perform a series of 40 shots at baseline (T1) and again after a task which either did or did not require self-control (T2). Additionally, we continuously measured the shooters’ level of perceived self-control strength. We assumed that in elite athletes, shooting accuracy as well as the perceived level of self-control strength would not be significantly affected over time from T1 to T2 in both conditions. In line with our assumptions, Bayesian linear mixed effect models revealed that shooting performance remained relatively stable in both conditions over time and the conditions also did not differ significantly in their perceived levels of self-control strength. Contrary to resource-based theories of self-control, these results speak against the idea of a limited self-control resource as previous acts of self-control did not impair subsequent shooting performance in elite athletes.

Gamifying accelerometer use increases physical activity levels of individuals pre-disposed to type II diabetes.

Physical activity is important for preventing obesity and diabetes, but most obese and pre-diabetic patients are not physically active. We developed a Fitbit-based game called MapTrek that promotes walking. We recruited obese and pre-diabetic patients. Half were randomly assigned to the control group and given a Fitbit alone. The others were given a Fitbit plus MapTrek. The MapTrek group participated in 6 months of weekly virtual races. Each week, participants were placed in a race with 9 others who achieved a similar number of steps in the previous week’s race. Participants moved along the virtual route by the steps recorded on their Fitbit and received daily walking challenges via text message. Text messages also had links to the race map and leaderboard. We used a Bayesian mixed effects model to analyze the number of steps taken during the intervention. A total of 192 (89%) participants in the control group and 196 (91%) in the MapTrek group were included in the analyses. MapTrek significantly increased step counts when it began: MapTrek participants walked almost 1,700 steps more than the control group on the first day of the intervention. We estimate that there is a 97% probability that the effect of MapTrek is at least 1,000 additional steps per day throughout the course of the 6-month intervention and that MapTrek participants would have walked an additional 81 miles, on average, before the effect ended. Our MapTrek intervention led to significant extra walking by the MapTrek participants.

Bayesian semiparametric mixed effects models for meta-analysis of the literature data : An application to cadmium toxicity studies.

We propose Bayesian semiparametric mixed effects models with measurement error to analyze the literature data collected from multiple studies in a meta-analytic framework. We explore this methodology for risk assessment in cadmium toxicity studies, where the primary objective is to investigate dose-response relationships between urinary cadmium concentrations and -microglobulin. In the proposed model, a nonlinear association between exposure and response is described by a Gaussian process with shape restrictions, and study-specific random effects are modeled to have either normal or unknown distributions with Dirichlet process mixture priors. In addition, nonparametric Bayesian measurement error models are incorporated to flexibly account for the uncertainty resulting from the usage of a surrogate measurement of a true exposure. We apply the proposed model to analyze cadmium toxicity data imposing shape constraints along with measurement errors and study-specific random effects across varying characteristics, such as population gender, age, or ethnicity.

Bayesian spatial modelling of terrestrial radiation in Switzerland

The geographic variation of terrestrial radiation can be exploited in epidemiological studies of the health effects of protracted low-dose exposure. Various methods have been applied to derive maps of this variation. We aimed to construct a map of terrestrial radiation for Switzerland. We used airborne γ-spectrometry measurements to model the ambient dose rates from terrestrial radiation through a Bayesian mixed-effects model and conducted inference using Integrated Nested Laplace Approximation (INLA). We predicted higher levels of ambient dose rates in the alpine regions and Ticino compared with the western and northern parts of Switzerland. We provide a map that can be used for exposure assessment in epidemiological studies and as a baseline map for assessing potential contamination.

Concord begets concord: A Bayesian model of nominal concord typology

Nominal concord is a phenomenon whereby nominal modifiers (e.g., adjectives, demonstratives, numerals) agree with their nominals along various dimensions (e.g., gender, number, case, definiteness). Here, drawing on a rich and typologically diverse database of nominal concord (Norris 2020), we build a Bayesian mixed effect model of nominal concord. Specifically, we consider two competing hypotheses regarding the statistical relationship between different types of concord within a language: (1) concord begets concord: the presence of some type of concord in a language makes it more likely that it has other types of concord vs. (2) a little concord goes a long way: if a language has some kind of concord, it is less likely to have other types of concord. We present evidence strongly in favor of the first hypothesis, that concord begets concord. Languages with nominal concord tend to have concord in more than one place and of more than one type. Using posterior draws from our model, we also provide quantitative evidence for a number of the tendencies described by Norris (2019a). Future work will build on this model to understand the functional role of nominal concord in language systems, how it evolves, and how it co-evolves with other typological features.

Progress toward HIV elimination goals: trends in and projections of annual HIV testing and condom use in Africa.

To estimate trends in and projections of annual HIV testing and condom use at last higher-risk sex and to calculate the probability of reaching key United Nations Programme on AIDS (UNAIDS)'s target. We included 114 nationally-representative datasets in 38 African countries from Demographic and Health Surveys and Multiple Indicator Cluster Surveys with 1 456 224 sexually active adults age 15-49 from 2003 to 2018. We applied Bayesian mixed effect models to estimate the coverage of annual HIV testing and condom use at last higher-risk sex for every country and year to 2030 and the probability of reaching UNAIDS testing and condom use targets of 95% coverage by 2030. Seven countries saw downward trends in annual HIV testing and four saw decreases in condom use at higher-risk sex, whereas most countries have upward trends in both indicators. The highest coverage of testing in 2030 is predicted in Swaziland with 92.6% (95% credible interval: 74.5-98.1%), Uganda with 90.5% (72.2-97.2%), and Lesotho with 90.5% (69.4%-97.6%). Meanwhile, Swaziland, Lesotho, and Namibia will have the highest proportion of condom use in 2030 at 85.0% (57.8-96.1%), 75.6% (42.3-93.6%), and 75.5% (42.4-93.2%). The probabilities of reaching targets were very low for both HIV testing (0-28.5%) and condom use (0-12.1%). We observed limited progress on annual HIV testing and condom use at last higher-risk sex in Africa and little prospect of reaching global targets for HIV/AIDS elimination. Although some funding agencies are considering withdrawal from supporting Africa, more attention to funding and expanding testing and treatment is needed in this region.