Negative Binomial Mixed Models Research Articles

Impact of four years of annually repeated indoor residual spraying (IRS) with Actellic 300CS on routinely reported malaria cases in an agricultural setting in Malawi.

Indoor residual spraying (IRS) is one of the main vector control tools used in malaria prevention. This study evaluates IRS in the context of a privately run campaign conducted across a low-lying, irrigated, sugarcane estate from Illovo Sugar, in the Chikwawa district of Malawi. The effect of Actellic 300CS annual spraying over four years (2015-2018) was assessed using a negative binomial mixed effects model, in an area where pyrethroid resistance has previously been identified. With an unadjusted incidence rate ratio (IRR) of 0.38 (95% CI: 0.32-0.45) and an adjusted IRR of 0.50 (95% CI: 0.42-0.59), IRS has significantly contributed to a reduction in case incidence rates at Illovo, as compared to control clinics and time points outside of the six month protective period. This study shows how the consistency of a privately run IRS campaign can improve the health of employees. More research is needed on the duration of protection and optimal timing of IRS programmes.

How turtles keep their cool: Seasonal and diel basking patterns in a tropical turtle

Behavioural thermoregulation by ectotherms is an important mechanism for maintaining body temperatures to optimise physiological performance. Experimental studies suggest that nocturnal basking by Krefft's river turtles (Emydura macquarii krefftii) in the tropics may allow them to avoid high water temperatures, however, this hypothesis has yet to be tested in the field. In this study, we examined the influence of environmental temperature on seasonal and diel patterns of basking in E. m. krefftii in tropical north Queensland, Australia. Wildlife cameras were used to document turtle basking events for seven consecutive days and nights for each month over a year (April 2020–March 2021). Air and water temperatures were recorded simultaneously using temperature loggers. We used a negative binomial mixed effects model to compare mean basking durations (min) occurring among four environmental temperature categories based on population thermal preference (26 °C): 1) air temperature above and water temperature below preferred temperature; 2) air temperature below and water temperature above preferred temperature; 3) air and water temperatures both above preferred temperature; and 4) air and water temperatures both below preferred temperature. Basking behaviour was influenced significantly by the relationship between air and water temperature. During the day, turtles spent significantly less time basking when both air and water temperatures were above their preferred temperatures. Conversely, at night, turtles spent significantly more time basking when water temperatures were warm and air temperatures were cool relative to their preferred temperature. This study adds to the growing body of work indicating pronounced heat avoidance as a thermoregulatory strategy among tropical reptile populations.

Subnational inequalities in years of life lost and associations with socioeconomic factors in pre-pandemic Europe, 2009–19: an ecological study

Health inequalities have been associated with shorter lifespans. We aimed to investigate subnational geographical inequalities in all-cause years of life lost (YLLs) and the association between YLLs and socioeconomic factors, such as household income, risk of poverty, and educational attainment, in countries within the European Economic Area (EEA) before the COVID-19 pandemic. In this ecological study, we extracted demographic and socioeconomic data from Eurostat for 1390 small regions and 285 basic regions for 32 countries in the EEA, which was complemented by a time-trend analysis of subnational regions within the EEA. Age-standardised YLL rates per 100 000 population were estimated from 2009 to 2019 based on methods from the Global Burden of Disease study. Geographical inequalities were assessed using the Gini coefficient and slope index of inequality. Socioeconomic inequalities were assessed by investigating the association between socioeconomic factors (educational attainment, household income, and risk of poverty) and YLLs in 2019 using negative binomial mixed models. Between Jan 1, 2009, and Dec 31, 2019, YLLs lowered in almost all subnational regions. The Gini coefficient of YLLs across all EEA regions was 14·2% (95% CI 13·6-14·8) for females and 17·0% (16·3 to 17·7) for males. Relative geographical inequalities in YLLs among women were highest in the UK (Gini coefficient 11·2% [95% CI 10·1-12·3]) and among men were highest in Belgium (10·8% [9·3-12·2]). The highest YLLs were observed in subnational regions with the lowest levels of educational attainment (incident rate ratio [IRR] 1·19 [1·13-1·26] for females; 1·22 [1·16-1·28] for males), household income (1·35 [95% CI 1·19-1·53]), and the highest poverty risk (1·25 [1·18-1·34]). Differences in YLLs remain within, and between, EEA countries and are associated with socioeconomic factors. This evidence can assist stakeholders in addressing health inequities to improve overall disease burden within the EEA. Research Council of Norway; Development, and Innovation Fund of Hungary; Norwegian Institute of Public Medicine; and COST Action 18218 European Burden of Disease Network.

Comparison of factors associated with seasonal influenza and COVID-19 booster vaccination coverage among healthcare personnel working at acute care hospitals during 2021–2022 influenza season, National Healthcare Safety Network, United States

The simultaneous circulation of seasonal influenza virus and SARS-CoV-2 variants will likely pose unique challenges to public health during the future influenza seasons. Persons who are undergoing treatment in healthcare facilities may be particularly at risk. It is important for healthcare personnel to protect themselves and patients by receiving vaccines. The purpose of this study is to assess coverage of the seasonal influenza vaccine and COVID-19 monovalent booster among healthcare personnel working at acute care hospitals in the United States during the 2021–22 influenza season and to examine the demographic and facility characteristics associated with coverage. A total of 3260 acute care hospitals with over 7 million healthcare personnel reported vaccination data to National Healthcare Safety Network (NHSN) during the 2021–22 influenza season. Two separate negative binomial mixed models were developed to explore the factors associated with seasonal influenza coverage and COVID-19 monovalent booster coverage. At the end of the 2021–2022 influenza season, the overall pooled mean seasonal influenza coverage was 80.3%, and the pooled mean COVID-19 booster coverage was 39.5%. Several demographic and facility-level factors, such as employee type, facility ownership, and geographic region, were significantly associated with vaccination against influenza and COVID-19 among healthcare personnel working in acute care hospitals. Our findings highlight the need to increase the uptake of vaccination among healthcare personnel, particularly non-employees, those working in for-profit and non-medical school-affiliated facilities, and those residing in the South.

Single‐cell Transcriptomes Reveal Cell‐type Specific Roles of VEGF Family Genes in the Pathogenesis of Alzheimer’s Disease

AbstractBackgroundThe role of the vascular endothelial growth factors (VEGFs) in the pathogenesis of Alzheimer’s disease (AD) has been recently described, including notable changes along the VEGFB/FLT1 signaling pathway. Yet, cell‐specific alterations are not well characterized. To disentangle VEGFs’ cell‐specific effect on AD‐related traits, we performed association analyses of mRNA expression from 10 VEGF genes with cognitive performance and AD pathology in eight brain cell types using a large single nucleus RNA sequencing dataset from human brain.MethodThis project utilized recently published single‐cell transcriptomes derived from dorsolateral prefrontal cortex tissues of 424 donors collected at Rush Alzheimer’s Disease Center. Processed single cell RNA sequencing data were acquired from AD Knowledge Portal (syn2580853). Mean age at death was 89 years, 68% were females, and 37% were AD cases. Negative binomial lognormal mixed models implemented in the NEBULA R package were used to detect differential expression between AD cases and cognitively normal controls, and to analyze the associations with immunohistochemical amyloid and tangles, and finally to assess cross‐sectional and longitudinal global cognitive scores. Correction for multiple comparisons was completed using the false discovery rate procedure (Pfdr<0.05).ResultPrefrontal cortical FLT1 expression was upregulated in AD participants compared to cognitively normal controls in both endothelial (endo) and microglial (micro) cells. Higher micro‐FLT1, endo‐FLT4, and oligodendrocyte (oligo) VEGFB expression was associated with greater β‐amyloid burden in AD samples, whereas higher VEGFB expression in inhibitory neurons was associated with lower β‐amyloid burden. Higher expression of micro‐FLT1 and oligo‐VEGFB associated with worse cognitive trajectories. Higher micro‐ and endo‐ expression of FLT1 associated with lower cognition scores at the last clinic visit before death. Lastly, higher astrocyte‐expressed NRP1 was associated with lower tangle burden.ConclusionConsistent with our previously reported bulk tissue results, prefrontal cortical expression of FLT1 and FLT4 were associated with global cognition, longitudinal cognitive trajectories, and AD neuropathology, and these associations appear to be limited to endothelial and microglial cells. In contrast, VEGFB expression seems to have opposite effects on amyloid burden depending on the cell types, suggesting its more dynamic role in AD physiopathology.

FoMO predicts alcohol use and harms over and above the big five personality traits among university students

ABSTRACT Background University students are an at-risk group for alcohol use. Fear of Missing Out (FoMO), the feeling that others are having rewarding experiences that one is missing out on, may drive alcohol use and harm. We aimed to investigate whether FoMO could uniquely predict alcohol use over and above known predictors like the big five personality traits. Methods We recruited 392 university students and invited them to take part in a daily diary study where they reported their alcohol use daily over 13 days. Participants also completed a baseline survey and reported personality traits (neuroticism, extraversion, openness, agreeableness, and conscientiousness), and alcohol-related consequences and completed a final follow-up survey where they reported FoMO. Results Using a zero-inflated negative binomial mixed model, we found that those higher in FoMO consumed more alcohol on drinking days (IRR = 1.11) when controlling for personality traits. Similarly, we found that those higher in FoMO experienced more harms (IRR = 1.10), when controlling personality traits. Conclusion We found that FoMO is a distinct construct and not just isomorphic with the big five personality traits. This suggests that reducing FoMO may be a key aim for alcohol interventions and should be considered when aiming to reduce university student alcohol use.

Successful elimination of falciparum malaria following the introduction of community-based health workers in Eastern Myanmar: A retrospective analysis.

Myanmar has a large majority of all malaria in the Greater Mekong Subregion. In the past decade, substantial progress was made in malaria control. The residual burden of malaria is in remote areas where currently recommended malaria elimination approaches are generally not feasible. In such hard-to-reach communities in Mon state, East Myanmar, Medical Action Myanmar introduced community health workers (CHWs) to deliver early diagnosis and treatment for malaria. We conducted a retrospective analysis to assess the impact of this intervention. This retrospective analysis involved data collected routinely from a CHW programme in Mon state conducted between 2011 and 2018. A network of 172 CHWs serving a population of 236,340 was deployed. These CHWs carried out 260,201 malaria rapid diagnostic tests (RDTs) to investigate patients with acute febrile illness. The median blood examination rate was 1.33%; interquartile range (IQR) (0.38 to 3.48%); 95% CI [1.28%, 1.36%] per month. The changes in malaria incidence and prevalence in patients presenting with fever were assessed using negative binomial regression mixed effects models fitted to the observed data. The incidence of Plasmodium falciparum malaria (including mixed infections) declined by 70%; 95% CI [65%, 75%]; p < 0.001 for each year of CHW operation. The incidence of P. vivax malaria declined by 56%; 95% CI [50%, 62%]; p < 0.001 per year. Malaria RDT positivity rates for P. falciparum and P. vivax declined by 69%; 95% CI [62%, 75%]; p < 0.001 and 53%; 95% CI [47%, 59%]; p < 0.001 per year, respectively. Between 2017 and 2018, only 1 imported P. falciparum case was detected in 54,961 RDTs. The main limitations of the study are use of retrospective data with possible unidentified confounders and uncharacterised population movement. The introduction of CHWs providing community-based malaria diagnosis and treatment and basic health care services in remote communities in Mon state was associated with a substantial reduction in malaria. Within 6 years, P. falciparum was eliminated and the incidence of P. vivax fell markedly.

Abstract 18813: Economic Burden of Pregnancy Among Privately Insured Women With Congenital Heart Disease

Background: Women with congenital heart defects (CHD) are increasingly reaching childbearing age. Understanding their healthcare costs in pregnancy may inform resource allocation, financial planning, and pre-conception counseling. Methods: We used 2010-2016 US commercially insured beneficiaries’ data within the IBM ® MarketScan® database to identify pregnancies among 18-55-year-old women with CHD and matched controls without CHD. We assessed baseline characteristics, cardiac and noncardiac conditions, pregnancy related conditions, birth outcomes, and cesarean section. We estimated total and out-of-pocket (OOP) costs, compiled and adjusted to 2016 medical component of the consumer price index. Negative binomial mixed model was used. Results: There were 6385 pregnancies in 5077 women with CHD (672 pregnancies among severe and 5713 among nonsevere) and 9438 pregnancies in 7737 controls. Compared to controls, pregnancies with CHD were more likely to be 35+ years of age at delivery, resided in the Northeast and West regions, had flexible insurance plans, cardiac conditions, noncardiac conditions, birth outcomes, pregnancy related conditions and cesarean sections, while there was no difference in the year of delivery and gestational diabetes rates. After adjusting for all covariates, having CHD was an independent predictor of higher adjusted pregnancy-related cost ratio for total (1.49, 95% confidence interval [CI] 1.40, 1.59) and OOP costs (1.27, 95% CI 1.19, 1.35). The adjusted mean total and OOP cost difference for pregnancy per patient for any CHD, severe CHD, and nonsevere CHD was significantly higher than controls (Figure). Total expenditure and OOP cost of CHD pregnancies during the study period were $162.5 and $15.7 million dollars respectively. Conclusions: We provide novel estimates of economic burden of pregnancy among privately insured CHD patients for clinical and financial counseling and planning for patients, payors and policymakers.

Subnational inequalities in YLLs and associated socioeconomic factors: a disease burden study

Abstract Background Health inequalities are an unjust and avoidable problem. This study examines subnational geographical inequalities in all-cause years of life lost (YLLs) and the association of socioeconomic factors in pre-coronavirus European Economic Area (EEA) countries. Methods In this ecological study complimented with a longitudinal analysis, demographic and socioeconomic data for 1390 small regions and 285 basic regions of 32 EEA countries were extracted from Eurostat. Age-standardised YLL rates per 100,000 population were estimated from 2009 to 2019 based on methods from the Global Burden of Disease Study. Inequalities were assessed using the Gini coefficient (GC) and slope index of inequality (SII). The association between socioeconomic factors by YLLs were assessed using negative binomial mixed models in 2019. Findings Over the period 2009-2019, YLLs have decreased in almost all subnational regions. The GC of YLLs across EEA regions was 14% for females (95% CI = 13·5 to 14·6%) and 17% for males (CI = 16·1 to 17·5%). Greece (GC = 10·1%, CI = 7·8 to 2·5%) and Belgium (GC = 10·8%, CI = 9·5 to 12·0%) had the highest relative inequalities in YLLs for women and men, respectively. Subnational regions with the lowest income (incident rate ratio (IRR)=1·39, CI = 1·23 to 1·58) and levels of educational attainment (IRRfemales=1·19, CI = 1·13 to 1·26; IRRmales=1·22, CI = 1·16 to 1·28), and highest poverty risk (IRR=1·18, CI = 1·12 to 1·25) were associated with increased YLLs, with stronger associations observed in Central and Eastern Europe. Interpretation Differences in YLLs remain within and between EEA countries and are associated with socioeconomic factors. This evidence can assist stakeholders in addressing specific health inequities to improve overall disease burden within the EEA. Key messages • The study highlights the need for public health policies targeted at the subnational level to reduce health inequalities in the EEA. • The study describes the effect of existing public health policies targeting socioeconomic factors.

Longitudinal Microbiome Changes in Supragingival Biofilm Transcriptomes Induced by Orthodontics.

Common oral diseases are known to be associated with dysbiotic shifts in the supragingival microbiome, yet most oral microbiome associations with clinical end points emanate from cross-sectional studies. Orthodontic treatment is an elective procedure that can be exploited to prospectively examine clinically relevant longitudinal changes in the composition and function of the supragingival microbiome. A longitudinal cohort study was conducted among 24 adolescent orthodontic patients who underwent saliva and plaque sampling and clinical examinations at time points: before fixed appliance bonding and at 1, 6, and 12 wk thereafter. Clinical indices included bleeding on probing (BOP), mean gingival index (GI), probing depths (PDs), and plaque index (PI). To study the biologically (i.e., transcriptionally) active microbial communities, RNA was extracted from plaque and saliva for RNA sequencing and microbiome bioinformatics analysis. Longitudinal changes in microbiome beta diversity were examined using PERMANOVA tests, and the relative abundance of microbial taxa was measured using Kruskal-Wallis tests, Wilcoxon rank-sum tests, and negative binomial and zero-inflated mixed models. Clinical measures of oral health deteriorated over time-the proportion of sites with GI and PI ≥1 increased by over 70% between prebonding and 12 wk postbonding while the proportion of sites with PD ≥4 mm increased 2.5-fold. Streptococcus sanguinis, a health-associated species that antagonizes cariogenic pathogens, showed a lasting decrease in relative abundance during orthodontic treatment. Contrarily, caries- and periodontal disease-associated taxa, including Selenomonas sputigena, Leptotrichia wadei, and Lachnoanaerobaculum saburreum, increased in abundance after bonding. Relative abundances of Stomatobaculum longum and Mogibacterium diversum in prebonding saliva predicted elevated BOP 12 wk postbonding, whereas Neisseria subflava was associated with lower BOP. This study offers insights into longitudinal community and species-specific changes in the supragingival microbiome transcriptome during fixed orthodontic treatment, advancing our understanding of microbial dysbioses and identifying targets of future health-promoting clinical investigations. Bonding braces was associated with subsequent changes in the oral microbiome characterized by increases in disease-associated species, decreases in health-associated species, and worsened clinical measures of oral health.

Contact with caregivers is associated with composition of the infant gastrointestinal microbiome in the first 6 months of life.

Little is known about how physical contact at birth and early caregiving environments influence the colonization of the infant gastrointestinal microbiome. We investigated how infant contact with caregivers at birth and within the first 2 weeks of life relates to the composition of the gastrointestinal microbiome in a sample of U.S. infants (n = 60). Skin-to-skin and physical contact with caregivers at birth and early caregiving environments were surveyed at 2 weeks postpartum. Stool samples were collected from infants at 2 weeks, 2, 6, and 12 months of age and underwent 16S rRNA sequencing as a proxy for the gastrointestinal microbiome. Associations between early caregiving environments and alpha and beta diversity, and differential abundance of bacteria at the genus level were assessed using PERMANOVA, and negative binomial mixed models in DEseq2. Time in physical contact with caregivers explained 10% of variation in beta diversity at 2 weeks' age. The number of caregivers in the first few weeks of life explained 9% of variation in beta diversity at 2 weeks and the number of individuals in physical contact at birth explained 11% of variation in beta diversity at 6 months. Skin-to-skin contact on the day of birth was positively associated with the abundance of eight genera. Infants held for by more individuals had greater abundance of eight genera. Results reveal a potential mechanism (skin-to-skin and physical contact) by which caregivers influence the infant gastrointestinal microbiome. Our findings contribute to work exploring the social transmission of microbes.

Comparison of different statistical models for the analysis of fracture events: findings from the Prevention of Falls Injury Trial (PreFIT)

BackgroundFractures are rare events and can occur because of a fall. Fracture counts are distinct from other count data in that these data are positively skewed, inflated by excess zero counts, and events can recur over time. Analytical methods used to assess fracture data and account for these characteristics are limited in the literature.MethodsCommonly used models for count data include Poisson regression, negative binomial regression, hurdle regression, and zero-inflated regression models. In this paper, we compare four alternative statistical models to fit fracture counts using data from a large UK based clinical trial evaluating the clinical and cost-effectiveness of alternative falls prevention interventions in older people (Prevention of Falls Injury Trial; PreFIT).ResultsThe values of Akaike information criterion and Bayesian information criterion, the goodness-of-fit statistics, were the lowest for negative binomial model. The likelihood ratio test of no dispersion in the data showed strong evidence of dispersion (chi-square = 225.68, p-value < 0.001). This indicates that the negative binomial model fits the data better compared to the Poisson regression model. We also compared the standard negative binomial regression and mixed effects negative binomial models. The LR test showed no gain in fitting the data using mixed effects negative binomial model (chi-square = 1.67, p-value = 0.098) compared to standard negative binomial model.ConclusionsThe negative binomial regression model was the most appropriate and optimal fit model for fracture count analyses.Trial registrationThe PreFIT trial was registered as ISRCTN71002650.

Does the built environment influence location- and trip-based sedentary behaviors? Evidence from a GPS-based activity space approach of neighborhood effects on older adults

ObjectiveEvidence on the influence of built environments on sedentary behaviors remains unclear and is often contradictory. The main limitations encompass the use of self-reported proxies of sedentary time (ST), the scarce consideration of the plurality of sedentary behaviors, and environmental exposures limited to the residential neighborhood. We investigated the relationships between GPS-based activity space measures of environmental exposures and accelerometer-based ST measured in total, at the place of residence, at all locations, and during trips. MethodsThis study is part of the CURHA project, based on 471 older adults residing in Luxembourg, who wore a GPS receiver and a tri-axial accelerometer during 7 days. Daily ST was computed in total, at the residence, at all locations and during trips. Environmental exposures included exposure to green spaces, walking, biking, and motorized transportation infrastructures. Associations between environments and ST were examined using linear and negative binomial mixed models, adjusted for demographics, self-rated health, residential self-selection, weather conditions and wear time. ResultsParticipants accumulated, on average, 8 h and 14 min of ST per day excluding sleep time. ST spent at locations accounted for 83 % of the total ST. ST spent at the residence accounted for 87 % of the location-based ST and 71 % of the total ST. Trip-based ST represents 13 % of total ST, and 4 % remained unclassified. Higher street connectivity was negatively associated with total ST, while the density of parking areas correlated positively with total and location-based ST. Stronger associations were observed for sedentary bouts (uninterrupted ST over 20 and 30 min). ConclusionImproving street connectivity and controlling the construction of new parking, while avoiding the spatial segregation of populations with limited access to public transport, may contribute to limit ST. Such urban planning interventions may be especially efficient in limiting the harmful uninterrupted bouts of ST among older adults.

Impact of COVID-19 epidemic on temporal pattern of new HIV diagnoses in Italy, 2021 database.

New HIV diagnoses in Italy decreased drastically in 2020 due to COVID-19 related effects: 50% fewer diagnoses were reported by the National HIV Surveillance System. COVID-19 pandemic impact on HIV surveillance is unclear. We estimated the expected number of new HIV diagnoses in 2020 in order to isolate the impact of the COVID-19 pandemic. We analyzed 29697 new HIV infections diagnosed from 2012 to 2020, reported to the National HIV Surveillance System. We assessed temporal trends of new HIV diagnoses applying negative binomial mixed effects models. We estimated the COVID-19 impact as the difference between the model-estimated slopes from 2012 to 2019 and the change reported in the diagnoses. The expected number of new HIV diagnoses in 2020 was also estimated and compared with the reported count. Based on the historical trend, we expected a 15% (95% CI: 5-25%) decline of new HIV diagnoses in 2020. We reported, however, a 49% decrease, yielding to a 34% net decrease in the number of new diagnoses. The strongest impact was estimated in northern regions (-40%) and MSM (-38%). We estimated 761 (95% prediction interval: 350-1277) missed diagnoses during 2020, the majority of them occurring in the North (465 cases), among MSM (416) and heterosexual males (217). In 2020, when excluding 15% decrease of new diagnoses attributable to the expected reduction, an additional 34% decrease was observed, representing a large decline in new HIV diagnoses associated with the COVID-19 pandemic.

IDESC: identifying differential expression in single-cell RNA sequencing data with multiple subjects

BackgroundSingle-cell RNA sequencing (scRNA-seq) technology has enabled assessment of transcriptome-wide changes at single-cell resolution. Due to the heterogeneity in environmental exposure and genetic background across subjects, subject effect contributes to the major source of variation in scRNA-seq data with multiple subjects, which severely confounds cell type specific differential expression (DE) analysis. Moreover, dropout events are prevalent in scRNA-seq data, leading to excessive number of zeroes in the data, which further aggravates the challenge in DE analysis.ResultsWe developed iDESC to detect cell type specific DE genes between two groups of subjects in scRNA-seq data. iDESC uses a zero-inflated negative binomial mixed model to consider both subject effect and dropouts. The prevalence of dropout events (dropout rate) was demonstrated to be dependent on gene expression level, which is modeled by pooling information across genes. Subject effect is modeled as a random effect in the log-mean of the negative binomial component. We evaluated and compared the performance of iDESC with eleven existing DE analysis methods. Using simulated data, we demonstrated that iDESC had well-controlled type I error and higher power compared to the existing methods. Applications of those methods with well-controlled type I error to three real scRNA-seq datasets from the same tissue and disease showed that the results of iDESC achieved the best consistency between datasets and the best disease relevance.ConclusionsiDESC was able to achieve more accurate and robust DE analysis results by separating subject effect from disease effect with consideration of dropouts to identify DE genes, suggesting the importance of considering subject effect and dropouts in the DE analysis of scRNA-seq data with multiple subjects.

Impact of pharmacist‐provided medication therapy management service on reducing unplanned readmissions in adult patients in Singapore

AbstractIntroductionOlder patients on polypharmacy are predisposed to drug‐related problems (DRPs). While medication therapy management (MTM) with pharmacist involvement can reduce DRP occurrence, few have examined its impact on reducing unplanned admissions. This study was designed to determine whether a pharmacist‐provided MTM service can reduce unplanned admissions through the comparison with patients receiving usual care.MethodsA retrospective case–control study was conducted in Changi General Hospital. Patients enrolled to MTM service from January 2016 to December 2021 were included in the intervention arm (n = 96) while patients who were not enrolled were recruited as control (n = 96). Primary outcome was the incidence rate ratio (IRR) of unplanned admissions within 6 months postindex visit comparing between intervention arm and control arm. Secondary outcomes included number of DRPs identified, types of DRPs, and the potential risks avoided by resolving DRPs. The negative binomial mixed model was used to model the unplanned admissions data.ResultsMTM with pharmacists' involvement was associated with a 39% (IRRs 0.61, 95% confidence interval [CI] 0.37–0.99, p = 0.047) lower rate of admissions in the intervention compared with the control group. There were higher number of DRPs (144 vs. 2) found in the intervention arm compared with control arm, respectively. The most prevalent types of DRPs were “Nonadherence” (80.1%), “Drug omission” (5.6%), and “Inappropriate dose” (2.8%). The most common potential risks avoided were increased cardiovascular risk, n = 29 (22.1%), increased fall risk, n = 18 (13.7%), and increased fracture risk, n = 17 (13.0%).ConclusionThe study suggests that pharmacist‐provided MTM service decreased unplanned admission. It has improved medication safety and quality of care by identifying and resolving more DRPs.

Machine learning and statistical models for analyzing multilevel patent data

A recent surge of patent applications among public hospitals in China has aroused significant research interest. A country’s healthcare innovation capacity can be measured by its number of patents. This paper explores the link between the number of patents and ten independent variables. Multicollinearity was carefully detected and removed by using the variable selection method and LASSO regression, respectively. The Poisson model and the negative binomial model were proposed to analyze the patent data. Three goodness of fit tests, the Pearson test, the deviance test, and the DHARMa non-parametric dispersion test, were conducted to investigate if the model has a good fit. After discovering four clusters by conducting agglomerative hierarchical clustering, these two models were replaced by the negative binomial mixed model. The likelihood ratio test was used to determine which model is more appropriate and the results reveal that the negative binomial mixed model outperforms both the Poisson model and the negative binomial model. Three variables, number of health technicians per 10,000 people, financial expenditure on science and technology as well as number of patent applications per 10,000 health personnel, have a significantly positive relationship with the number of patents in Chinese tertiary public hospitals.

Unravelling the determinants of medical practice variation in referrals among primary care physicians: insights from a retrospective cohort study in Southern Israel

ObjectivesReducing medical practice variation (MPV) is a central theme of system improvement because it is associated with poor health outcomes, increased costs and disparities in care. This study aimed to...

109 Use of the Horse Grimace Scale to evaluate markers of stress in horses involved in equine assisted activities

Much like humans, horses possess the ability to generate a wide range of facial behaviors. These expressions are a form of communication used to convey substantial amounts of information to other members of the herd, including stress. A variety of studies have developed ethograms to quantify the degree of facial behaviors expressed as a means of assessing the welfare of the horse. This is particularly important in working horses, such as those involved in equine assisted activities (EAA). The objective of this study was to use the Horse Grimace Scale (HGS) to assess the stress response of horses involved in EAA by numerically scoring facial muscular tension. The HGS is a scale that focuses on 6 specific facial action units to evaluate the degree of pain and distress experienced by a horse. Each unit was given a score of 0 (not present), 1 (moderately present), or 2 (obviously present). The highest possible score was 12, indicating the greatest degree of pain. Horses trained in EAA (EAA, n = 6) were paired with a veteran with posttraumatic stress disorder (PTSD) for a 30-min weekly lesson or assigned as astanding non-EAA trained control (CON, n = 6) for the duration of 8 wks. Images of the profile of the horses were taken at 3 time points during the lesson: 10 min, 20 min, and 30 min after the start of the session. After the sessions, images were organized and redacted if necessary and scored by 4 independent raters who had received training on the use of the HGS. Score data were analyzed using a negative binomial mixed model, repeated measures ANOVA with fixed effects of treatment (EAA vs CON), session, and time point, and their interactions and a random effect of horse. Intraclass correlation coefficient for inter-rater agreement was 0.446 (P = 2.17e-59). Statistically different means were separated using Tukey's method. Significance was set at P ≤ 0.05. Scores were higher (P = 0.0178) in wk 4 (5.2 ± 0.4, mean ± SE) than wk 1 (3.7 ± 0.4) among CON horses. No other significant differences were found. Participation in EAA did not appear to have an adverse effect on horses as defined by pain and distress scores on the HGS.

130 Investigation of age, status, season, and sex on equine parasitic load in a university program

Strongyles are nematode parasites that can damage the horse's large intestine and cecum and cause symptoms of colic, anemia, and weight loss. Sixty-four horses from the University of Connecticut campus were used to study parasitic shedding patterns, to determine if there was a seasonal, age, sex or work status effect. Horses were assigned to different statuses based on living and exercise situations. The goal of analyses was to assess whether there was a change in parasite count over time, and whether horse age, status and sex had any effect on parasite count. Fecal samples of all horses were collected in March and October for 5 years. Fecal flotation testing was done to evaluate strongyle eggs per gram (EPG). Zero-inflated (ZI), negative binomial mixed effects models were used to evaluate changes in parasite counts over time. Both conditional (nonzero parasite counts) and ZI components were adjusted for status, age, sex, and the interaction of status and year. Horse ID was treated as a random effect, to account for repeated measurements. No sex effect on parasitic load was found with either the conditional or ZI models. Similarly, neither status norseason were significantly associated statistically with parasite counts. Initial age did have an effect on conditional parasite counts, with regression results suggesting that as age increased, parasite counts decreased. Although we have some evidence of declining parasite loads at UConn, this effect was only significant for spring sampling. Fall sampling occurred after the horses had been turned out over the summer, so perhaps they are then at greater risk for getting a heavier parasite load. The decline over the years in the March data could be due to the implementation of a parasitic control protocol by barn management 4 years into the study. This study appears to agree with other studies in that older horses are less affected by internal parasites than their younger counterparts. The parasitic control protocol implemented by barn management appears to have had a positive effect on horses in the spring. More research is needed to determine if this protocol will eventually have an effect on the fall as well.