Behavioral Reactions to Job Insecurity Climate Perceptions: Exit, Voice, Loyalty, and Neglect.

Past work has extensively documented that job insecurity predicts various work- and health-related outcomes. However, limited research has focused on the potential consequences of perceived job insecurity climate. Our objective was to investigate how the psychological climate about losing a job and valuable job features (quantitative and qualitative job insecurity climate, respectively) relate to employees' exit, voice, loyalty, and neglect behaviors, and whether such climate perceptions explain additional variance in these behaviors over individual job insecurity. Data were collected through an online survey using a convenience sample of employees working in different organizations in Türkiye (N = 245). Hierarchical multiple regression analyses showed that quantitative job insecurity climate was associated with higher levels of loyalty and neglect, while qualitative job insecurity climate was related to higher levels of exit and lower levels of loyalty. Importantly, job insecurity climate explained additional variance over individual job insecurity in exit and loyalty. Our findings underscore the importance of addressing job insecurity in a broader context regarding one's situation and the psychological collective climate. This study contributes to addressing the knowledge gap concerning job insecurity climate, an emerging construct in the organizational behavior literature, and its incremental impact beyond individual job insecurity. The foremost implication is that organizations need to pay attention to the evolving climate perceptions about the future of jobs in the work environment, because such perceptions are related to critical employee behaviors.

Face mask recommendations in schools did not impact COVID-19 incidence among 10-12-year-olds in Finland - joinpoint regression analysis.

BACKGROUND: In autumn 2021 in Finland, a recommendation to use face masks was implemented nationwide in schools for pupils ages 12 years and above. While national guidelines were in form of recommendations, cities implemented mandatory masking in schools. Some cities extended this mandate for younger pupils as well. Our aim was to compare COVID-19 incidence among 10-12-year-olds between cities with different recommendations on the use of face masks in schools. METHODS: COVID-19 case numbers, defined as positive laboratory verified SARS-CoV-2 test results, were obtained from the National Infectious Disease Registry (NIDR) of the Finnish Institute for Health and Welfare. Helsinki, Turku and Tampere were selected for comparison since the baseline COVID-19 incidence in the cities had been similar in August and September 2021. Helsinki and Tampere implemented the national recommendation on face mask use at schools, while Turku extended this to include those 10 years old and above, starting from the beginning of semester in early August. Age groups of 7-9-year-olds, 10-12-year-olds and 30-49-year-olds were included in the statistical analysis and moving averages of 14-day incidences per 100 000 inhabitants were used as a dependent variable. Joinpoint regression was used to estimate average percent changes (APC) and average daily percent changes (ADPC) in the 14-day incidences. Differences in the ADPC values between the cities were compared in one-month periods. We also calculated cumulative incidences from the beginning of August to the end of November in the cities by age group. RESULTS: In August, the ADPC was highest in Turku (3.9) and lowest in Tampere (2.0), while in September, the ADPC was highest in Turku (-0.3) and lowest in Helsinki (-3.2) among 10-12-year-olds. In October, the ADPC was highest in Helsinki (2.1) and lowest in Turku (-0.2) and in November, the ADPC was highest in Turku (4.1) and lowest in Tampere (-0.5) among 10-12-year-olds. We also calculated cumulative incidences from the beginning of August to the end of November in the cities by age groups of 7-9 years, 10-12 years, and 30-49 years. The cumulative incidence was highest in Turku in all age groups and lowest in Tampere. CONCLUSIONS: According to our analysis, no additional effect was gained from mandating face masks, based on comparisons between the cities and between the age groups of the unvaccinated children (10-12 years versus 7-9 years).

Linear Regression Analysis of Sleep Quality in People with Insomnia in Wuhan City during the COVID-19 Pandemic.

Objective: COVID-19 has evolved into a major global public health event. The number of people reporting insomnia is growing exponentially during the pandemic. This study aimed to explore the relationship between aggravated insomnia and COVID-19-induced psychological impact on the public, lifestyle changes, and anxiety about the future. Methods: In this cross-sectional study, we used the questionnaires from 400 subjects who were obtained from the Department of Encephalopathy of the Wuhan Hospital of Traditional Chinese Medicine between July 2020 and July 2021. The data collected for the study included demographic characteristics of the participants and psychological scales consisting of the Spiegel Sleep Questionnaire, the Fear of COVID-19 Scale (FCV-19S), the Zung Self-Rating Anxiety Scale (SAS), and the Zung Self-Rating Depression Scale (SDS). The independent sample t-test and one-way ANOVA were used to compare the results. Correlation analysis of variables affecting insomnia was performed using Pearson correlation analysis. The degree of influence of the variables on insomnia was determined using linear regression, and a regression equation was derived. Results: A total of 400 insomnia patients participated in the survey. The median age was 45.75 ± 15.04 years. The average score of the Spiegel Sleep Questionnaire was 17.29 ± 6.36, that of SAS was 52.47 ± 10.39, that of SDS was 65.89 ± 8.72, and that of FCV-19S was 16.09 ± 6.81. The scores of FCV-19S, SAS, and SDS were closely related to insomnia, and the influencing degree was in the following order: fear, depression, and anxiety (OR = 1.30, 0.709, and 0.63, respectively). Conclusion: Fear of COVID-19 can be one of the primary contributors to worsening insomnia.

Real-world effectiveness and factors associated with effectiveness of inactivated SARS-CoV-2 vaccines: a systematic review and meta-regression analysis.

BACKGROUND: The two inactivated SARS-CoV-2 vaccines, CoronaVac and BBIBP-CorV, have been widely used to control the COVID-19 pandemic. The influence of multiple factors on inactivated vaccine effectiveness (VE) during long-term use and against variants is not well understood. METHODS: We selected published or preprinted articles from PubMed, Embase, Scopus, Web of Science, medRxiv, BioRxiv, and the WHO COVID-19 database by 31 August 2022. We included observational studies that assessed the VE of completed primary series or homologous booster against SARS-CoV-2 infection or severe COVID-19. We used DerSimonian and Laird random-effects models to calculate pooled estimates and conducted multiple meta-regression with an information theoretic approach based on Akaike's Information Criterion to select the model and identify the factors associated with VE. RESULTS: Fifty-one eligible studies with 151 estimates were included. For prevention of infection, VE associated with study region, variants, and time since vaccination; VE was significantly decreased against Omicron compared to Alpha (P = 0.021), primary series VE was 52.8% (95% CI, 43.3 to 60.7%) against Delta and 16.4% (95% CI, 9.5 to 22.8%) against Omicron, and booster dose VE was 65.2% (95% CI, 48.3 to 76.6%) against Delta and 20.3% (95% CI, 10.5 to 28.0%) against Omicron; primary VE decreased significantly after 180 days (P = 0.022). For the prevention of severe COVID-19, VE associated with vaccine doses, age, study region, variants, study design, and study population type; booster VE increased significantly (P = 0.001) compared to primary; though VE decreased significantly against Gamma (P = 0.034), Delta (P = 0.001), and Omicron (P = 0.001) compared to Alpha, primary and booster VEs were all above 60% against each variant. CONCLUSIONS: Inactivated vaccine protection against SARS-CoV-2 infection was moderate, decreased significantly after 6 months following primary vaccination, and was restored by booster vaccination. VE against severe COVID-19 was greatest after boosting and did not decrease over time, sustained for over 6 months after the primary series, and more evidence is needed to assess the duration of booster VE. VE varied by variants, most notably against Omicron. It is necessary to ensure booster vaccination of everyone eligible for SARS-CoV-2 vaccines and continue monitoring virus evolution and VE. TRIAL REGISTRATION: PROSPERO, CRD42022353272.

Impact of the COVID-19 pandemic on levels of deep-ocean acoustic noise.

The extraordinary circumstances of the COVID-19 pandemic led to measures to mitigate the spread of the disease, with lockdowns and mobility restrictions at national and international levels. These measures led to sudden and sometimes dramatic reductions in human activity, including significant reductions in ship traffic in the maritime sector. We report on a reduction of deep-ocean acoustic noise in three ocean basins in 2020, based on data acquired by hydroacoustic stations in the International Monitoring System of the Comprehensive Nuclear-Test-Ban Treaty. The noise levels measured in 2020 are compared with predicted levels obtained from modelling data from previous years using Gaussian Process regression. Comparison of the predictions with measured data for 2020 shows reductions of between 1 and 3 dB in the frequency range from 10 to 100 Hz for all but one of the stations.

Influence of geospatial resolution on sociodemographic predictors of COVID-19 in Massachusetts.

PURPOSE: When studying health risks across a large geographic region such as a state or province, researchers often assume that finer-resolution data on health outcomes and risk factors will improve inferences by avoiding ecological bias and other issues associated with geographic aggregation. However, coarser-resolution data (e.g., at the town or county-level) are more commonly publicly available and packaged for easier access, allowing for rapid analyses. The advantages and limitations of using finer-resolution data, which may improve precision at the cost of time spent gaining access and processing data, have not been considered in detail to date. METHODS: We systematically examine the implications of conducting town-level mixed-effect regression analyses versus census-tract-level analyses to study sociodemographic predictors of COVID-19 in Massachusetts. In a series of negative binomial regressions, we vary the spatial resolution of the outcome, the resolution of variable selection, and the resolution of the random effect to allow for more direct comparison across models. RESULTS: We find stability in some estimates across scenarios, changes in magnitude, direction, and significance in others, and tighter confidence intervals on the census-tract level. Conclusions regarding sociodemographic predictors are robust when regions of high concentration remain consistent across town and census-tract resolutions. CONCLUSIONS: Inferences about high-risk populations may be misleading if derived from town- or county-resolution data, especially for covariates that capture small subgroups (e.g., small racial minority populations) or are geographically concentrated or skewed (e.g., % college students). Our analysis can help inform more rapid and efficient use of public health data by identifying when finer-resolution data are truly most informative, or when coarser-resolution data may be misleading.

Factors Influencing SARS-CoV-2 Infection Control Practices of Nurses Caring for COVID-19 Patients in South Korea: Based on Health Belief Model.

This study aimed to verify the level of COVID-19 infection control practices and the factors affecting the COVID-19 infection control practices of Korean nurses based on the health belief model. The participants were 143 nurses experienced in caring for COVID-19 patients in South Korea. Questionnaires were used to measure health beliefs, confidence in practice, knowledge of COVID-19, infection protection environment, and COVID-19 infection control practices. Data were analyzed by performing descriptive statistics, an independent t-test, one-way analysis of variance, the Mann-Whitney test and multiple regression analysis. The mean score for infection control practices related to COVID-19 was 4.76 on a 5-point scale where a higher score indicates superior infection control performance. Multiple regression analysis revealed that the factors that influenced COVID-19 infection control practices were gender, marital status, perceived susceptibility, and confidence in practice related to COVID-19. With COVID-19 approaching endemic and to prevent infectious diseases, it is necessary to emphasize perceived sensitivity by providing accurate information on the risk of infection rather than simply inducing infection control to be divided into individual activities. In addition, nurses' infection control practices should be implemented with confidence with the nurses themselves feeling the need for infection control and not being forced by the social atmosphere or the hospital.

Detection of outlier prefectures on the mortality due to COVID-19 in Japan.

Japan has suffered from COVID-19 with significant mortality, but its prefectural differences are not well investigated. Since the mortality due to COVID-19 was likely to be associated with the number of infected cases, the population density, and the proportion of the elderly population, we tried to detect the outlier prefectures by multiple linear regression analyses. With the use of the Hampel identifier, we found that Hokkaido and Hyogo were the outlier prefectures with higher mortality after adjusting the variables above. Further studies should delineate the causes of these differences.

Effect of continuing the use of renin-angiotensin system inhibitors on mortality in patients hospitalized for coronavirus disease 2019: a systematic review, meta-analysis, and meta-regression analysis.

BACKGROUND: The effect of angiotensin-converting enzyme inhibitors (ACEIs)/angiotensin receptor blockers (ARBs) on mortality was preliminarily explored through the comparison of ACEIs/ARBs with non-ACEIs/ARBs in patients with coronavirus disease 2019 (COVID-19). Reaching a conclusion on whether previous ACEI/ARB treatment should be continued in view of the different ACE2 levels in the comparison groups was not unimpeachable. Therefore, this study aimed to further elucidate the effect of ACEI/ARB continuation on hospital mortality, intensive care unit (ICU) admission, and invasive mechanical ventilation (IMV) in the same patient population. METHODS: We searched PubMed, the Cochrane Library, Ovid, and Embase for relevant articles published between December 1, 2019 and April 30, 2022. Continuation of ACEI/ARB use after hospitalization due to COVID-19 was considered as an exposure and discontinuation of ACEI/ARB considered as a control. The primary outcome was hospital mortality, and the secondary outcomes included 30-day mortality, rate of ICU admission, IMV, and other clinical outcomes. RESULTS: Seven observational studies and four randomized controlled trials involving 2823 patients were included. The pooled hospital mortality in the continuation group (13.04%, 158/1212) was significantly lower than that (22.15%, 278/1255) in the discontinuation group (risk ratio [RR] = 0.45; 95% confidence interval [CI], 0.28-0.72; P = 0.001). Continuation of ACEI/ARB use was associated with lower rates of ICU admission (10.5% versus 16.2%, RR = 0.63; 95% CI 0.5-0.79; P < 0.0001) and IMV (8.2% versus 12.5%, RR = 0.62; 95% CI 0.46-0.83, P = 0.001). Nevertheless, the effect was mainly demonstrated in the observational study subgroup (P < 0.05). Continuing ACEI/ARB had no significant effect on 30-day mortality (P = 0.34), acute myocardial infarction (P = 0.08), heart failure (P = 0.82), and acute kidney injury after hospitalization (P = 0.98). CONCLUSION: Previous ACEI/ARB treatment could be continued since it was associated with lower hospital deaths, ICU admission, and IMV in patients with COVID-19, although the benefits of continuing use were mainly shown in observational studies. More evidence from multicenter RCTs are still needed to increase the robustness of the data. Trial registration PROSPERO (CRD42022341169). Registered 27 June 2022.

Frequent and unpredictable changes in COVID-19 policies and restrictions reduce the accuracy of model forecasts.

Between June and August 2020, an agent-based model was used to project rates of COVID-19 infection incidence and cases diagnosed as positive from 15 September to 31 October 2020 for 72 geographic settings. Five scenarios were modelled: a baseline scenario where no future changes were made to existing restrictions, and four scenarios representing small or moderate changes in restrictions at two intervals. Post hoc, upper and lower bounds for number of diagnosed Covid-19 cases were compared with actual data collected during the prediction window. A regression analysis with 17 covariates was performed to determine correlates of accurate projections. It was found that the actual data fell within the lower and upper bounds in 27 settings and out of bounds in 45 settings. The only statistically significant predictor of actual data within the predicted bounds was correct assumptions about future policy changes (OR 15.04; 95% CI 2.20-208.70; p = 0.016). Frequent changes in restrictions implemented by governments, which the modelling team was not always able to predict, in part explains why the majority of model projections were inaccurate compared with actual outcomes and supports revision of projections when policies are changed as well as the importance of modelling teams collaborating with policy experts.

A panel regression analysis for the COVID-19 epidemic in the United States.

This study explored the roles of epidemic-spread-related behaviors, vaccination status and weather factors during the COVID-19 epidemic in 50 U.S. states since March 2020. Data from March 1, 2020 to February 5, 2022 were incorporated into panel model. The states were clustered by the k-means method. In addition to discussing the whole time period, we also took multiple events nodes into account and analyzed the data in different time periods respectively by panel linear regression method. In addition, influence of cluster grouping and different incubation periods were been discussed. Non-segmented analysis showed the rate of people staying at home and the vaccination dose per capita were significantly negatively correlated with the daily incidence rate, while the number of long-distance trips was positively correlated. Weather indicators also had a negative effect to a certain extent. Most segmental results support the above view. The vaccination dose per capita was unsurprisingly proved to be the most significant factor especially for epidemic dominated by Omicron strains. 7-day was a more robust incubation period with the best model fit while weather had different effects on the epidemic spread in different time period. The implementation of prevention behaviors and the promotion of vaccination may have a successful control effect on COVID-19, including variants' epidemic such as Omicron. The spread of COVID-19 also might be associated with weather, albeit to a lesser extent.

Nonparametric kernel estimation of the probability of cure in a mixture cure model when the cure status is partially observed.

Cure models are a class of time-to-event models where a proportion of individuals will never experience the event of interest. The lifetimes of these so-called cured individuals are always censored. It is usually assumed that one never knows which censored observation is cured and which is uncured, so the cure status is unknown for censored times. In this paper, we develop a method to estimate the probability of cure in the mixture cure model when some censored individuals are known to be cured. A cure probability estimator that incorporates the cure status information is introduced. This estimator is shown to be strongly consistent and asymptotically normally distributed. Two alternative estimators are also presented. The first one considers a competing risks approach with two types of competing events, the event of interest and the cure. The second alternative estimator is based on the fact that the probability of cure can be written as the conditional mean of the cure status. Hence, nonparametric regression methods can be applied to estimate this conditional mean. However, the cure status remains unknown for some censored individuals. Consequently, the application of regression methods in this context requires handling missing data in the response variable (cure status). Simulations are performed to evaluate the finite sample performance of the estimators, and we apply them to the analysis of two datasets related to survival of breast cancer patients and length of hospital stay of COVID-19 patients requiring intensive care.

Effect of epidemic management and control plan on COVID-19 mortality in Iran: an interrupted time series analysis.

Background: Globally, several measures have been taken to decrease COVID-19 mortality. However, the effectiveness of preventive measures on the mortality related to COVID-19 has not been fully assessed. Thus, the present study aimed the present study aimed to assess the success of COVID-19 epidemic management and control plan on the mortality associated with COVID-19 in Iran from February 19, 2020, to February 5, 2021. Methods: In the current quasi experimental study an interrupted time series analysis of daily collected data on confirmed deaths of COVID-19 occurred in Iran and in the world, were performed using Newey ordinary least squares regression-based methods. Results: In Iran, the trend of new deaths increased significantly every day until 24 November 2020 according to pre-intervention slope of [(OR 1.14, 95% CI 0.96-1.32,); P < 0.001]. The occurrence of new deaths had a decreasing trend after November 24, 2020, with a coefficient of [(OR -5.12, 95% CI -6.04 - -4.20), P < 0.001)]. But in the global level daily new deaths was increasing before [(OR 18.66, 95% CI 14.41-2292; P < 0.001)] and after the 24 November 2020 [(OR 57.14, 95% CI 20.80-93.49); P: 0.002]. Conclusions: Iranian COVID-19 epidemic management and control plan effectively reduced the mortality associated to COVID-19. Therefore, it is essential to continue these measures to prevent the increase in the number of deaths.

In Silico Prediction of Intestinal Permeability by Hierarchical Support Vector Regression.

The vast majority of marketed drugs are orally administrated. As such, drug absorption is one of the important drug metabolism and pharmacokinetics parameters that should be assessed in the process of drug discovery and development. A nonlinear quantitative structure-activity relationship (QSAR) model was constructed in this investigation using the novel machine learning-based hierarchical support vector regression (HSVR) scheme to render the extremely complicated relationships between descriptors and intestinal permeability that can take place through various passive diffusion and carrier-mediated active transport routes. The predictions by HSVR were found to be in good agreement with the observed values for the molecules in the training set (n = 53, r2 = 0.93, q CV 2 = 0.84, RMSE = 0.17, s = 0.08), test set (n = 13, q2 = 0.75-0.89, RMSE = 0.26, s = 0.14), and even outlier set (n = 8, q2 = 0.78-0.92, RMSE = 0.19, s = 0.09). The built HSVR model consistently met the most stringent criteria when subjected to various statistical assessments. A mock test also assured the predictivity of HSVR. Consequently, this HSVR model can be adopted to facilitate drug discovery and development.

Multisite learning of high-dimensional heterogeneous data with applications to opioid use disorder study of 15,000 patients across 5 clinical sites.

Integrating data across institutions can improve learning efficiency. To integrate data efficiently while protecting privacy, we propose A one-shot, summary-statistics-based, Distributed Algorithm for fitting Penalized (ADAP) regression models across multiple datasets. ADAP utilizes patient-level data from a lead site and incorporates the first-order (ADAP1) and second-order gradients (ADAP2) of the objective function from collaborating sites to construct a surrogate objective function at the lead site, where model fitting is then completed with proper regularizations applied. We evaluate the performance of the proposed method using both simulation and a real-world application to study risk factors for opioid use disorder (OUD) using 15,000 patient data from the OneFlorida Clinical Research Consortium. Our results show that ADAP performs nearly the same as the pooled estimator but achieves higher estimation accuracy and better variable selection than the local and average estimators. Moreover, ADAP2 successfully handles heterogeneity in covariate distributions.

Worker Characteristics and Measures Associated With Patient and Visitor Violence in the COVID-19 Pandemic: A Multilevel Regression Analysis From China.

Objective: To analyze the patient and visitor workplace violence (PVV) toward health workers (HWs) and identify correlations between worker characteristics, measures against violence and exposure to PVV in COVID-19 pandemic. Methods: A cross-sectional survey utilizing the international questionnaires in six public tertiary hospitals from Beijing in 2020 was conducted, and valid data from 754 respondents were collected. Multilevel logistic regression models were used to determine the association between independents and exposure to PVV. Results: During COVID-19 pandemic and regular epidemic prevention and control, doctors were 5.3 times (95% CI = 1.59~17.90) more likely to suffer from physical PVV than nurses. HWs most frequently work with infants were 7.2 times (95% CI = 2.24~23.19) more likely to suffer from psychological PVV. More than four-fifth of HWs reported that their workplace had implemented security measures in 2020, and the cross-level interactions between the security measures and profession variable indicates that doctors in the workplace without security measures were 11.3 times (95% CI = 1.09~116.39) more likely to suffer from physical PVV compared to nurses in the workplace with security measures. Conclusion: Doctors have higher risk of physical PVV in COVID-19 containment, and the security measures are very important and effective to fight against the physical PVV. Comprehensive measures should be implemented to mitigate hazards and protect the health, safety, and well-being of health workers.

A compelling demonstration of why traditional statistical regression models cannot be used to identify risk factors from case data on infectious diseases: a simulation study.

BACKGROUND: Regression models are often used to explain the relative risk of infectious diseases among groups. For example, overrepresentation of immigrants among COVID-19 cases has been found in multiple countries. Several studies apply regression models to investigate whether different risk factors can explain this overrepresentation among immigrants without considering dependence between the cases. METHODS: We study the appropriateness of traditional statistical regression methods for identifying risk factors for infectious diseases, by a simulation study. We model infectious disease spread by a simple, population-structured version of an SIR (susceptible-infected-recovered)-model, which is one of the most famous and well-established models for infectious disease spread. The population is thus divided into different sub-groups. We vary the contact structure between the sub-groups of the population. We analyse the relation between individual-level risk of infection and group-level relative risk. We analyse whether Poisson regression estimators can capture the true, underlying parameters of transmission. We assess both the quantitative and qualitative accuracy of the estimated regression coefficients. RESULTS: We illustrate that there is no clear relationship between differences in individual characteristics and group-level overrepresentation -small differences on the individual level can result in arbitrarily high overrepresentation. We demonstrate that individual risk of infection cannot be properly defined without simultaneous specification of the infection level of the population. We argue that the estimated regression coefficients are not interpretable and show that it is not possible to adjust for other variables by standard regression methods. Finally, we illustrate that regression models can result in the significance of variables unrelated to infection risk in the constructed simulation example (e.g. ethnicity), particularly when a large proportion of contacts is within the same group. CONCLUSIONS: Traditional regression models which are valid for modelling risk between groups for non-communicable diseases are not valid for infectious diseases. By applying such methods to identify risk factors of infectious diseases, one risks ending up with wrong conclusions. Output from such analyses should therefore be treated with great caution.

The impact of the COVID-19 pandemic lockdown measures on the prescribing trends and utilization of opioids in the English primary care setting: segmented-liner regression analysis.

BACKGROUND: The COVID-19 lockdown has resulted in limited access to most of the conventional chronic pain management services. Subsequently, changes in opioids' utilization could be expected. This study assessed the impact of the first COVID-19 lockdown on opioid utilization using aggregated-level, community dispensing dataset covering the whole English population. RESEARCH DESIGN AND METHODS: A segmented-linear regression analysis was applied to monthly dispensed opioid prescriptions from March 2019 to March 2021. Opioid utilization was measured using the number of opioids' items dispensed/1000 inhabitants and Defined Daily Dose (DDD)/1000 inhabitants/day during 12-months pre/post the lockdown in March 2020 stratified by strong and weak opioids. RESULTS: For all opioids' classes, there were nonsignificant changes in the number of opioids' items dispensed/1000 inhabitants trend pre-lockdown, small increases in their level immediately post-lockdown, and a non-significant decline in the trend post-lockdown. Similarly, a non-significant reduction in the DDD/1000 inhabitant/day baseline trend pre-lockdown, nonsignificant immediate increases in the level post-lockdown, and declines in the trend post-lockdown for all opioids' classes were observed. CONCLUSION: Unexpectedly, opioid utilization does not appear to have been significantly affected by the lockdown measures during the study period. However, patient-level data is needed to determine more accurate estimates of any changes in the opioid prescribing including incident prescribing/use.

Distributed Quasi-Poisson regression algorithm for modeling multi-site count outcomes in distributed data networks.

BACKGROUND: Observational studies incorporating real-world data from multiple institutions facilitate study of rare outcomes or exposures and improve generalizability of results. Due to privacy concerns surrounding patient-level data sharing across institutions, methods for performing regression analyses distributively are desirable. Meta-analysis of institution-specific estimates is commonly used, but has been shown to produce biased estimates in certain settings. While distributed regression methods are increasingly available, methods for analyzing count outcomes are currently limited. Count data in practice are commonly subject to overdispersion, exhibiting greater variability than expected under a given statistical model. OBJECTIVE: We propose a novel computational method, a one-shot distributed algorithm for quasi-Poisson regression (ODAP), to distributively model count outcomes while accounting for overdispersion. METHODS: ODAP incorporates a surrogate likelihood approach to perform distributed quasi-Poisson regression without requiring patient-level data sharing, only requiring sharing of aggregate data from each participating institution. ODAP requires at most three rounds of non-iterative communication among institutions to generate coefficient estimates and corresponding standard errors. In simulations, we evaluate ODAP under several data scenarios possible in multi-site analyses, comparing ODAP and meta-analysis estimates in terms of error relative to pooled regression estimates, considered the gold standard. In a proof-of-concept real-world data analysis, we similarly compare ODAP and meta-analysis in terms of relative error to pooled estimatation using data from the OneFlorida Clinical Research Consortium, modeling length of stay in COVID-19 patients as a function of various patient characteristics. In a second proof-of-concept analysis, using the same outcome and covariates, we incorporate data from the UnitedHealth Group Clinical Discovery Database together with the OneFlorida data in a distributed analysis to compare estimates produced by ODAP and meta-analysis. RESULTS: In simulations, ODAP exhibited negligible error relative to pooled regression estimates across all settings explored. Meta-analysis estimates, while largely unbiased, were increasingly variable as heterogeneity in the outcome increased across institutions. When baseline expected count was 0.2, relative error for meta-analysis was above 5% in 25% of iterations (250/1000), while the largest relative error for ODAP in any iteration was 3.59%. In our proof-of-concept analysis using only OneFlorida data, ODAP estimates were closer to pooled regression estimates than those produced by meta-analysis for all 15 covariates. In our distributed analysis incorporating data from both OneFlorida and the UnitedHealth Group Clinical Discovery Database, ODAP and meta-analysis estimates were largely similar, while some differences in estimates (as large as 13.8%) could be indicative of bias in meta-analytic estimates. CONCLUSIONS: ODAP performs privacy-preserving, communication-efficient distributed quasi-Poisson regression to analyze count outcomes using data stored within multiple institutions. Our method produces estimates nearly matching pooled regression estimates and sometimes more accurate than meta-analysis estimates, most notably in settings with relatively low counts and high outcome heterogeneity across institutions.