Twenty years since Joinpoint 1.0: Two major enhancements, their justification, and impact

Since its release of Version 1.0 in 1998, Joinpoint software developed for cancer trend analysis by a team at the US National Cancer Institute has received a considerable attention in the trend analysis community and it became one of most widely used software for trend analysis. The paper published in Statistics in Medicine in 2000 (a previous study) describes the permutation test procedure to select the number of joinpoints, and Joinpoint Version 1.0 implemented the permutation procedure as the default model selection method and employed parametric methods for the asymptotic inference of the model parameters. Since then, various updates and extensions have been made in Joinpoint software. In this paper, we review basic features of Joinpoint, summarize important updates of Joinpoint software since its first release in 1998, and provide more information on two major enhancements. More specifically, these enhancements overcome prior limitations in both the accuracy and computational efficiency of previously used methods. The enhancements include: (i) data driven model selection methods which are generally more accurate under a broad range of data settings and more computationally efficient than the permutation test and (ii) the use of the empirical quantile method for construction of confidence intervals for the slope parameters and the location of the joinpoints, which generally provides more accurate coverage than the prior parametric methods used. We show the impact of these changes in cancer trend analysis published by the US National Cancer Institute.

Cardiovascular Disease Hospitalizations in Louisiana Parishes' Elderly before, during and after Hurricane Katrina

The research on how health and health care disparities impact response to and recovery from a disaster, especially among diverse and underserved populations is in great need for a thorough evaluation. The time series analysis utilizing most complete national databases of medical records is an indispensable tool in assessing the destruction and health toll brought about by natural disasters. In this study, we demonstrated such an application by evaluating the impact of Hurricane Katrina in 2005 on cardiovascular disease (CVD), a primary cause of mortality among older adults that can be aggravated by natural disasters. We compared CVD hospitalizations before, during and after Katrina between white and black residents of three most populated parishes in Louisiana: Orleans and Jefferson, which were severely affected by the landfall and subsequent floods, and East Baton Rouge, which hosted many of the evacuees. We abstracted 383,552 CVD hospitalization records for Louisiana's patients aged 65+ in 2005⁻2006 from the database maintained by the Center of Medicare &amp; Medicaid Services. Daily time series of CVD-related hospitalization rates at each study parish were compiled, and the changes were characterized using segmented regression. In Orleans Parish, directly affected by the hurricane, hospitalization rates peaked on the 6th day after landfall with an increase (mean ± SD) from 7.25 ± 2.4 to 18.5 ± 17.3 cases/day per 10,000 adults aged 65+ (p < 0.001) and returned to pre-landfall level after ~2 months. Disparities in CVD rates between black and white older adults were exacerbated during and following landfall. In Orleans Parish, a week after landfall, the CVD rates increased to 26.3 ± 23.7 and 16.6 ± 11.7 cases/day per 10,000 people (p < 0.001) for black and white patients, respectively. The abrupt increase in CVDs is likely due to psychosocial and post-traumatic stress caused by the disaster and inadequate response. Inequities in resource allocation and access have to be addressed in disaster preparation and mitigation.

Three Statistical Approaches for Assessment of Intervention Effects: A Primer for Practitioners

INTRODUCTION

Statistical methods to assess the impact of an intervention are increasingly used in clinical research settings. However, a comprehensive review of the methods geared toward practitioners is not yet available.

METHODS AND MATERIALS

We provide a comprehensive review of three methods to assess the impact of an intervention: difference-in-differences (DID), segmented regression of interrupted time series (ITS), and interventional autoregressive integrated moving average (ARIMA). We also compare the methods, and provide illustration of their use through three important healthcare-related applications.

RESULTS

In the first example, the DID estimate of the difference in health insurance coverage rates between expanded states and unexpanded states in the post-Medicaid expansion period compared to the pre-expansion period was 5.93 (95% CI, 3.99 to 7.89) percentage points. In the second example, a comparative segmented regression of ITS analysis showed that the mean imaging order appropriateness score in the emergency department at a tertiary care hospital exceeded that of the inpatient setting with a level change difference of 0.63 (95% CI, 0.53 to 0.73) and a trend change difference of 0.02 (95% CI, 0.01 to 0.03) after the introduction of a clinical decision support tool. In the third example, the results from an interventional ARIMA analysis show that numbers of creatinine clearance tests decreased significantly within months of the start of eGFR reporting, with a magnitude of drop equal to -0.93 (95% CI, -1.22 to -0.64) tests per 100,000 adults and a rate of drop equal to 0.97 (95% CI, 0.95 to 0.99) tests per 100,000 per adults per month.

DISCUSSION

When choosing the appropriate method to model the intervention effect, it is necessary to consider the structure of the data, the study design, availability of an appropriate comparison group, sample size requirements, whether other interventions occur during the study window, and patterns in the data.

Vitamin D and Anti-Müllerian Hormone Levels in Infertility Treatment: The Change-Point Problem

Background: Anti-Müllerian hormone (AMH) is considered to be one of the most significant indicators of women’s fertility. Many studies have shown that vitamin D may modify human reproductive functions; however, the results are conflicting. The composition of follicular fluid (FF) creates the biochemical environment of the oocyte and affects its quality, which later determines the embryo quality. In this study, we aimed to revise with advanced statistical techniques the relationship between AMH and vitamin D in FF. Methods: The study was designed as a prospective single-center study in infertile patients with AMH ≥ 0.7 ng/mL who underwent controlled ovarian hyperstimulation for in vitro fertilization. AMH and vitamin D levels in FF were measured. Next, the standard and advanced statistical (including segmented regression) techniques were applied. Results: We observed a negative linear correlation between levels of AMH in serum and FF and total vitamin D concentrations up to approximately 30 ng/mL; with a statistically significant relationship in FF. Beyond that concentration, the trend was positive but statistically insignificant. Conclusions: As an existing “change-point problem” was noticed, we suggest segmentation in the relationship between vitamin D and AMH during infertility treatment.

Methods, applications, interpretations and challenges of interrupted time series (ITS) data: protocol for a scoping review

OBJECTIVES

Interrupted time series (ITS) design involves collecting data across multiple time points before and after the implementation of an intervention to assess the effect of the intervention on an outcome. ITS designs have become increasingly common in recent times with frequent use in assessing impact of evidence implementation interventions. Several statistical methods are currently available for analysing data from ITS designs; however, there is a lack of guidance on which methods are optimal for different data types and on their implications in interpreting results. Our objective is to conduct a scoping review of existing methods for analysing ITS data, to summarise their characteristics and properties, as well as to examine how the results are reported. We also aim to identify gaps and methodological deficiencies.

METHODS AND ANALYSIS

We will search electronic databases from inception until August 2016 (eg, MEDLINE and JSTOR). Two reviewers will independently screen titles, abstracts and full-text articles and complete the data abstraction. The anticipated outcome will be a summarised description of all the methods that have been used in analysing ITS data in health research, how those methods were applied, their strengths and limitations and the transparency of interpretation/reporting of the results. We will provide summary tables of the characteristics of the included studies. We will also describe the similarities and differences of the various methods.

ETHICS AND DISSEMINATION

Ethical approval is not required for this study since we are just considering the methods used in the analysis and there will not be identifiable patient data. Results will be disseminated through open access peer-reviewed publications.

Trends in seasonal influenza vaccine coverage of target groups in France, 2006/07 to 2015/16: Impact of recommendations and 2009 influenza A(H1N1) pandemic

Background and aimsSeasonal influenza vaccination (SIV) uptake (SIVU) rates in France are below target. We (i) describe trends in French SIVU over 10 consecutive seasons among different target groups and (ii) examine the effects of the 2009 influenza A(H1N1) pandemic and the publication of new SIV recommendations in 2011 and 2013.MethodsOur study was based on records of vaccines delivered in community pharmacies for a permanent, representative sample of 805,000 beneficiaries of the French National Health Insurance Fund. For the first objective, we analysed SIVU rate trends among ≥ 65 year olds as well as among  < 65 year olds with each of the following conditions: diabetes, respiratory, cardiovascular, neuromuscular, or chronic liver disease. For the second goal, we computed segmented log-binomial regression analyses.ResultsAfter the 2009 pandemic, except for the target group with liver diseases, where the difference was not statistically significant, SIVU fell significantly in all groups during the 2010/11 season, remaining relatively stable until 2015/16 in groups not targeted by new recommendations. Crude SIVU rates in 2015/16 were 48% (43,950/91,794) for ≥ 65 year olds and between 16% (407/2,565) and 29% (873/3,056) for  < 65 year olds depending on their condition. SIVU increased modestly after new recommendations were published, but only in patients newly eligible for a free vaccine voucher.ConclusionsOur results suggest: (i) a prolonged confidence crisis in SIV, initially impelled by the 2009 pandemic vaccination campaign; (ii) that new recommendations are ineffective without additional measures. Interventional research in this field is a priority.

Segmented linear modeling of CHO fed-batch culture and its application to large scale production

We describe a systematic approach to model CHO metabolism during biopharmaceutical production across a wide range of cell culture conditions. To this end, we applied the metabolic steady state concept. We analyzed and modeled the production rates of metabolites as a function of the specific growth rate. First, the total number of metabolic steady state phases and the location of the breakpoints were determined by recursive partitioning. For this, the smoothed derivative of the metabolic rates with respect to the growth rate were used followed by hierarchical clustering of the obtained partition. We then applied a piecewise regression to the metabolic rates with the previously determined number of phases. This allowed identifying the growth rates at which the cells underwent a metabolic shift. The resulting model with piecewise linear relationships between metabolic rates and the growth rate did well describe cellular metabolism in the fed‐batch cultures. Using the model structure and parameter values from a small‐scale cell culture (2 L) training dataset, it was possible to predict metabolic rates of new fed‐batch cultures just using the experimental specific growth rates. Such prediction was successful both at the laboratory scale with 2 L bioreactors but also at the production scale of 2000 L. This type of modeling provides a flexible framework to set a solid foundation for metabolic flux analysis and mechanistic type of modeling. Biotechnol. Bioeng. 2017;114: 785–797. © 2016 The Authors. Biotechnology and Bioengineering Published by Wiley Periodicals, Inc.

Design characteristics and statistical methods used in interrupted time series studies evaluating public health interventions: protocol for a review

INTRODUCTION

An interrupted time series (ITS) design is an important observational design used to examine the effects of an intervention or exposure. This design has particular utility in public health where it may be impracticable or infeasible to use a randomised trial to evaluate health system-wide policies, or examine the impact of exposures (such as earthquakes). There have been relatively few studies examining the design characteristics and statistical methods used to analyse ITS designs. Further, there is a lack of guidance to inform the design and analysis of ITS studies.This is the first study in a larger project that aims to provide tools and guidance for researchers in the design and analysis of ITS studies. The objectives of this study are to (1) examine and report the design characteristics and statistical methods used in a random sample of contemporary ITS studies examining public health interventions or exposures that impact on health-related outcomes, and (2) create a repository of time series data extracted from ITS studies. Results from this study will inform the remainder of the project which will investigate the performance of a range of commonly used statistical methods, and create a repository of input parameters required for sample size calculation.

METHODS AND ANALYSIS

We will collate 200 ITS studies evaluating public health interventions or the impact of exposures. ITS studies will be identified from a search of the bibliometric database PubMed between the years 2013 and 2017, combined with stratified random sampling. From eligible studies, we will extract study characteristics, details of the statistical models and estimation methods, effect metrics and parameter estimates. Further, we will extract the time series data when available. We will use systematic review methods in the screening, application of inclusion and exclusion criteria, and extraction of data. Descriptive statistics will be used to summarise the data.

ETHICS AND DISSEMINATION

Ethics approval is not required since information will only be extracted from published studies. Dissemination of the results will be through peer-reviewed publications and presentations at conferences. A repository of data extracted from the published ITS studies will be made publicly available.

Assessing health care interventions via an interrupted time series model: Study power and design considerations

The delivery and assessment of quality health care is complex with many interacting and interdependent components. In terms of research design and statistical analysis, this complexity and interdependency makes it difficult to assess the true impact of interventions designed to improve patient health care outcomes. Interrupted time series (ITS) is a quasi-experimental design developed for inferring the effectiveness of a health policy intervention while accounting for temporal dependence within a single system or unit. Current standardized ITS methods do not simultaneously analyze data for several units nor are there methods to test for the existence of a change point and to assess statistical power for study planning purposes in this context. To address this limitation, we propose the "Robust Multiple ITS" (R-MITS) model, appropriate for multiunit ITS data, that allows for inference regarding the estimation of a global change point across units in the presence of a potentially lagged (or anticipatory) treatment effect. Under the R-MITS model, one can formally test for the existence of a change point and estimate the time delay between the formal intervention implementation and the over-all-unit intervention effect. We conducted empirical simulation studies to assess the type one error rate of the testing procedure, power for detecting specified change-point alternatives, and accuracy of the proposed estimating methodology. R-MITS is illustrated by analyzing patient satisfaction data from a hospital that implemented and evaluated a new care delivery model in multiple units.

Impact of a Multifaceted Pharmacist-Led Intervention on Antimicrobial Stewardship in a Gastroenterology Ward: A Segmented Regression Analysis

BACKGROUND

Irrational use of antimicrobial agents for gastrointestinal diseases deserves attention, but corresponding antimicrobial stewardship programs (ASPs) are generally not a priority for managers. We conducted this study to evaluate the effectiveness of multifaceted pharmacist-led (MPL) interventions in the gastroenterology ward (GW) to provide evidence for the efficacy of ASPs in a non-priority department.

METHODS

This was an interventional, retrospective study implemented in China. The MPL intervention lasting 1.5 years involved daily ward rounds with physicians, regular review of medical orders, monthly indicator feedback, frequent physician training, and necessary patient education. Data on all hospitalized adults receiving antibiotics was extracted from the hospital information system over a 36-month period from January 2016 to December 2018. Segmented regression analysis of interrupted time series was performed to evaluate the effect of the MPL interventions (started in July 2017) on antibiotic use and length of hospital stay, which was calculated monthly as analytical units.

RESULTS

A total of 1763 patients receiving antibiotics were enrolled. Segmented regression models showed descending trends from the baseline in the intensity of antibiotic consumption (coefficient = -0.88, p = 0.01), including a significant decline in the level of change of the proportion of patients receiving combined antibiotics (coefficient = -9.91, p = 0.03) and average length of hospital stay (coefficient = -1.79, p = 0.00), after MPL interventions. The MPL interventions led to a temporary increase in the proportion of patients receiving antibiotics (coefficient = 4.95, p = 0.038), but this was part of a declining secular trend (coefficient = -0.45, p = 0.05).

CONCLUSION

The MPL interventions led a statistically significant decline in the number of patients receiving antibiotics, the antibiotic consumption, and the average hospital stay post-intervention compared to the pre-intervention phase of the study. Health policymakers should actively practice MPL interventions by clinical pharmacists in ASPs in those departments that are not included in priority management.

Belayer: Modeling discrete and continuous spatial variation in gene expression from spatially resolved transcriptomics

Spatially resolved transcriptomics (SRT) technologies measure gene expression at known locations in a tissue slice, enabling the identification of spatially varying genes or cell types. Current approaches for these tasks assume either that gene expression varies continuously across a tissue or that a tissue contains a small number of regions with distinct cellular composition. We propose a model for SRT data from layered tissues that includes both continuous and discrete spatial variation in expression and an algorithm, Belayer, to learn the parameters of this model. Belayer models gene expression as a piecewise linear function of the relative depth of a tissue layer with possible discontinuities at layer boundaries. We use conformal maps to model relative depth and derive a dynamic programming algorithm to infer layer boundaries and gene expression functions. Belayer accurately identifies tissue layers and biologically meaningful spatially varying genes in SRT data from the brain and skin.

Choice of statistical method influences apparent association between structure and function in glaucoma

PURPOSE

The aim of this study was to explore how different statistical methods may lead to inconsistent inferences about the association between structure and function in glaucoma.

METHODS

Two datasets from published studies were selected for their illustrative value. The first consisted of measurements of neuroretinal rim area in the superior-temporal sector paired with the corresponding visual field sensitivity. The second consisted of measurements of average retinal nerve fiber layer thickness over all sectors paired with the corresponding visual field sensitivity. Statistical methods included linear and segmented regression, and a nonparametric local-linear fit known as loess. The analyses were repeated with all measurements expressed as percent of mean normal.

RESULTS

Slopes from linear fits to the data changed by a factor of 10 depending on the linear regression method applied. Inferences about whether structural abnormality precedes functional abnormality varied with the statistical design and the units of measure used.

CONCLUSIONS

The apparent association between structure and function in glaucoma, and consequent interpretation, varies with the statistical method and units of measure. Awareness of the limitations of any statistical analysis is necessary to avoid finding spurious results that ultimately may lead to inadequate clinical recommendations.

Breast cancer among Asian Indian and Pakistani Americans: A surveillance, epidemiology and end results-based study

Breast cancer incidence is increasing among Asian Indian and Pakistani women living in the United States. We examined the characteristics of breast cancer in Asian Indian and Pakistani American (AIPA) and non-Hispanic white (NHW) women using data from the surveillance, epidemiology and end results (SEER) program. Breast cancer incidence rates were estimated via segmented Poisson regression using data between 1990 and 2014 from SEER 9 registries, including New Jersey and California. Disease characteristics, treatment and survival information between 2000 and 2016 for 4900 AIPA and 482 250 NHW cases diagnosed after age 18 were obtained from SEER 18 registries and compared using descriptive analyses and multivariable competing risk proportional hazards regression. Breast cancer incidence was lower in AIPA than NHW women, increased with age and the rate of increase declined after age of 46 years. AIPA women were diagnosed at significantly younger age (mean (SD) = 54.5 (13.3) years) than NHW women (mean (SD) = 62 (14) years, P < .0001) and were more likely than NHW cases (P < .0001) to have regional or distant stage, higher grade, estrogen receptor-negative, progesterone receptor-negative, triple-negative or human epidermal growth factor receptor 2-enriched tumors, subcutaneous or total mastectomy, and lower cumulative incidence of death due to breast cancer (hazard ratio = 0.79, 95% CI: 0.72-0.86, P < .0001). AIPA had shorter median follow-up (52 months) than NHW cases (77 months). Breast cancer in AIPA women has unique characteristics that need to be further studied along with a comprehensive evaluation of their follow-up patterns.

Effects of public health measures during the SARS-CoV-2 pandemic on the winter respiratory syncytial virus epidemic: An interrupted time series analysis

BACKGROUND

Public health measures (PHM) designed to contain the spread of COVID-19 pandemic have influenced the epidemiological characteristics of other viral infections. Its impact on acute RSV bronchiolitis in infants of ≤24 months old has not been systematically studied in our setting.

OBJECTIVES

To describe the monthly pattern of visits to the Paediatric Emergency Department (PED) of patients 0 to 14 years of age, the rate of patients diagnosed with RSV acute bronchiolitis per thousand inhabitants of 0 to 24 months, and the rate of them requiring hospital admission during the winter 2020-2021, in the context of local and national COVID-19 restrictions and compare them to the four previous seasons.

METHODS

Interrupted time series analysis of patients assisted in the PED and diagnosed with or admitted for RSV acute bronchiolitis in a tertiary University Hospital from January 2016 to February 2020 (pre-intervention period) and from March 2020 to June 2021 (post-intervention period).

INTERVENTION

Preventive PHM implemented by the Spanish government weighted by the Containment and Health Index of the Oxford COVID-19 Government Response Tracker.

RESULTS

The intervention was followed by an immediate reduction of the rate of visits to the PED of -19.5 (95% confidence interval [CI] -24.0, -14.9) per thousand, and the rate of diagnoses and admissions for RSV acute bronchiolitis of -44.3 (95% CI -73.8, -14.8) and -1.4 (95% CI -2.7, -0.1) per thousand, respectively, with a delayed rebound.

CONCLUSIONS

After the implementation of PHM to prevent the spread of SARS-CoV-2 infection, an immediate and important decline in the visits to the PED was observed, with an upward change thereafter. There was also an initial reduction in the diagnoses of and admissions by RSV acute bronchiolitis. An upward trend was observed six to nine months after the usual time of the winter RSV epidemic, coinciding with the relaxation of the preventive PHM.

Trend of Smear-positive Pulmonary Tuberculosis in Iran during 1995-2012: A Segmented Regression Model

Background:

Describing trend in tuberculosis (TB) over time can play an important role to assess the disease control strategies and predict the future morbidity and mortality. This study aimed to determine the incidence trend of smear-positive pulmonary tuberculosis (SPPT) in sub-age and sex groups during the years of 1995–2012.

Methods:

This retrospective cohort study was performed in 2015 by using the dataset regarding National Statistics of SPPT reported by World Health Organization during 1995–2012. Annual percent changes (APCs) and average annual percent changes (AAPCs) were estimated to determine the summery statistics of trend using segmented regression model.

Results:

During 1995–2012, there were 96,579 SPPT case notifications in Iran (male to female ratio: 0.99). There was only one change point in 1997 for SPPT incidence in subgroups of age and sex during 1995–2012. The AAPCs for both genders and also all three age groups had a significant descending trend during the time period (P < 0.05).

Conclusions:

Our results showed a downward trend in the SPPT incidence. It seems that to achieve the set goals and high successful in TB control program especially reduction in SPPT, pay more attention to old age and males should be considered. In addition, improvement of clinical and medical care services and notification processes would be imperative.

Timing and temperature thresholds of heat stress effects on fertility performance of different parity sows in Spanish herds

High temperature is an environmental factor that impairs sow fertility. In this study, we identified the critical weeks for heat stress effects on aspects of fertility performance, namely weaning-to-first-service interval (WSI) and farrowing rate (FR). We also examined the threshold temperatures above which the fertility performance deteriorated and whether there were any differences between parities regarding heat stress effects or thresholds. Performance data of sows in 142 herds from 2011 to 2016 were matched to appropriate weekly averaged daily maximum temperatures (Tmax) from weather stations close to the herds. Two types of ratios (i.e., ratio for WSI and odds ratio for FR) were used to identify the critical weeks for heat stress by comparing the respective measures for two sow groups based on Tmax in different weeks around weaning or service events. The ratios for WSI were calculated between groups of sows exposed to Tmax ≥ 27 °C or <27 °C in each week before weaning, with the Tmax cutoff value based on a recent review study. Similarly, the odds ratios for FR for the two groups were calculated in weeks around service. The weeks with the largest differences in the fertility measures between the two Tmax groups (i.e., the highest ratio for WSI and the lowest odds ratio for FR) were considered to be the critical weeks for heat stress. Also, piecewise models with different breakpoints were constructed to identify the threshold Tmax in the critical week. The breakpoint in the best-fit model was considered to be the threshold Tmax. The highest ratios for WSI were obtained at 1 to 3 wk before weaning in parity 1 and 2 or higher sow groups. The threshold Tmax leading to prolonged WSI was 17 °C for parity 1 sows and 25 °C for parity 2 or higher sows. Increasing Tmax by 10 °C above these thresholds increased WSI by 0.65, and 0.33 to 0.35 d, respectively (P < 0.01). For FR, the lowest odds ratios were obtained at 2 to 3 wk before service in parity 0, 1, and 2 or higher sow groups. The threshold Tmax leading to reductions in FR was 20, 21, and 24 to 25 °C for parity 0, 1, and 2 or higher sow groups, respectively. Increasing Tmax by 10 °C above these thresholds decreased FR by 3.0%, 4.3%, and 1.9% to 2.8%, respectively (P < 0.01). These results indicate that the critical weeks for heat stress were 2 to 3 wk before service for FR and 1 to 3 wk before weaning for WSI. The decreases in fertility performance in parity 0 to 1 sows started at temperatures 3 to 8 °C lower than in parity 2 or higher sows.

Fast Bootstrap Confidence Intervals for Continuous Threshold Linear Regression

Continuous threshold regression is a common type of nonlinear regression that is attractive to many practitioners for its easy interpretability. More widespread adoption of thresh-old regression faces two challenges: (i) the computational complexity of fitting threshold regression models and (ii) obtaining correct coverage of confidence intervals under model misspecification. Both challenges result from the non-smooth and non-convex nature of the threshold regression model likelihood function. In this paper we first show that these two issues together make the ideal approach for making model-robust inference in continuous threshold linear regression an impractical one. The need for a faster way of fitting continuous threshold linear models motivated us to develop a fast grid search method. The new method, based on the simple yet powerful dynamic programming principle, improves the performance by several orders of magnitude.

Fast Grid Search and Bootstrap-based Inference for Continuous Two-phase Polynomial Regression Models

Two-phase polynomial regression models (Robison, 1964; Fuller, 1969; Gallant and Fuller, 1973; Zhan et al., 1996) are widely used in ecology, public health, and other applied fields to model nonlinear relationships. These models are characterized by the presence of threshold parameters, across which the mean functions are allowed to change. That the threshold is a parameter of the model to be estimated from the data is an essential feature of two-phase models. It distinguishes them, and more generally, multi-phase models, from the spline models and has profound implications for both computation and inference for the models. Estimation of two-phase polynomial regression models is a non-convex, non-smooth optimization problem. Grid search provides high quality solutions to the estimation problem, but is very slow when done by brute force. Building upon our previous work on piecewise linear two-phase regression models estimation, we develop fast grid search algorithms for two-phase polynomial regression models and demonstrate their performance. Furthermore, we develop bootstrap-based pointwise and simultaneous confidence bands for mean functions. Monte Carlo studies are conducted to demonstrate the computational and statistical properties of the proposed methods. Three real datasets are used to help illustrate the application of two-phase models, with special attention on model choice.

Using Joinpoint regression for drug utilization research: Tutorial and case study of prescription opioid use in the United States

PURPOSE

Drug utilization researchers are often interested in evaluating prescribing and medication use patterns and trends over a specified period of time. Joinpoint regression is a useful methodology to identify any deviations in secular trends without a preconceived notion of where these break points might occur. This article provides a tutorial on the use of joinpoint regression, within Joinpoint software, for the analysis of drug utilization data.

METHODS

The statistical considerations for whether a joinpoint regression analytical technique is a suitable approach are discussed. Then, we offer a tutorial as an introduction on conducting joinpoint regression (within Joinpoint software) through a step-by-step application, which is a case study developed using opioid prescribing data from the United States. Data were obtained from public files available through the Centers for Disease Control and Prevention from 2006 to 2018. The tutorial provides parameters and sample data needed to replicate the case study and it concludes with general considerations for the reporting of results using joinpoint regression in drug utilization research.

RESULTS

The case study evaluated the trend of opioid prescribing in the United States from 2006 to 2018, where time points of significant variation (one in 2012 and another in 2016) are detected and interpreted.

CONCLUSIONS

Joinpoint regression is a helpful methodology for drug utilization for the purposes of conducting descriptive analyses. This tool also assists with corroborating assumptions and identifying parameters for fitting other models such as interrupted time series. The technique and accompanying software are user-friendly; however, researchers interested in using joinpoint regression should exercise caution and follow best practices for correct measurement of drug utilization.

Reduction in antimicrobial use associated with a multifaceted antimicrobial stewardship programme in a tertiary teaching hospital in Shanghai: a segmented regression analysis

BACKGROUND

Rational use of antibiotics received great attention in China, therefore the multifaceted antimicrobial stewardship (MAMS) is urgently required in hospital management. We conducted this study to assess the impact of a MAMS programme on antimicrobial use in a tertiary teaching hospital in Shanghai.

METHODS

This retrospective observational study was conducted at a tertiary teaching hospital in Shanghai. The MAMS programme involved multifaceted interventions consisting of a quality premium with financial incentives, antibiotic restriction, audit and feedback, and education. Data were extracted from the electronic medical records of inpatients to analyse monthly and annual antibiotic consumption and the percentage of antibiotic prescriptions during 2017-2020. Segmented regression analysis of the interrupted time series was used to contrast antimicrobial use during 2019-2020, with non-MAMS data from the 2017-2018 period as the historical control.

RESULTS

With MAMS implementation, antibiotic consumption decreased from 63.3 (59.3, 67.2) defined daily doses (DDDs) per 100 patient-days (PD) to 43.3 (39.0, 49.8) DDDs/100 PD (P<0.001), and the percentage of antibiotic prescriptions decreased from 44.8% (44.1%, 45.4%) to 43.3% (42.2%, 44.3%) (P<0.001). Segmented regression models suggested a reduction in antibiotic consumption (coefficient = -12.537, P<0.001) and indicated a downward trend in the percentage of antibiotic prescriptions (coefficient =-0.165, P=0.049). Neither antibiotic consumption nor the percentage of antibiotic prescriptions was influenced by the coronavirus disease 2019 (COVID-19) pandemic.

CONCLUSIONS

This study suggests that MAMS plays an important role in reducing antibiotic use and is not affected by special circumstances such as the COVID-19 pandemic. This novel intervention, consisting of a quality premium and multidisciplinary cooperation, should be prioritized by policy and decision makers, where rational management of antimicrobial use is urgently needed.

Current Practices in Missing Data Handling for Interrupted Time Series Studies Performed on Individual-Level Data: A Scoping Review in Health Research

OBJECTIVE

Missing data can produce biased estimates in interrupted time series (ITS) analyses. We reviewed recent ITS investigations on health topics for determining 1) the data management strategies and statistical analysis performed, 2) how often missing data were considered and, if so, how they were evaluated, reported and handled.

STUDY DESIGN AND SETTING

This was a scoping review following standard recommendations from the PRISMA Extension for Scoping Reviews. We included a random sample of all ITS studies that assessed any intervention relevant to health care (eg, policies or programmes) with individual-level data, published in 2019, with abstracts indexed on MEDLINE.

RESULTS

From 732 studies identified, we finally reviewed 60. Reporting of missing data was rare. Data aggregation, statistical tools for modelling population-level data and complete case analyses were preferred, but these can lead to bias when data are missing at random. Seasonality and other time-dependent confounders were rarely accounted for and, when they were, missing data implications were typically ignored. Very few studies reflected on the consequences of missing data.

CONCLUSION

Handling and reporting of missing data in recent ITS studies performed for health research have many shortcomings compared with best practice.

Prescribing Variation in General Practices in England Following a Direct Healthcare Professional Communication on Mirabegron

Introduction: Pharmacovigilance may detect safety issues after marketing of medications, and this can result in regulatory action such as direct healthcare professional communications (DHPC). DHPC can be effective in changing prescribing behaviour, however the extent to which prescribers vary in their response to DHPC is unknown. This study aims to explore changes in prescribing and prescribing variation among general practitioner (GP) practices following a DHPC on the safety of mirabegron, a medication to treat overactive bladder (OAB). Methods: This is an interrupted time series study of English GP practices from 2014–2017. National Health Service (NHS) Digital provided monthly statistics on aggregate practice-level prescribing and practice characteristics (practice staff and registered patient profiles, Quality and Outcomes Framework indicators, and deprivation of the practice area). The primary outcome was monthly mirabegron prescriptions as a percentage of all OAB drug prescriptions and we assessed the change following a DHPC issued by the European Medicines Agency in September 2015. The DHPC stated mirabegron use was contraindicated with severe uncontrolled hypertension and cautioned with hypertension. Variation between practices in mirabegron prescribing before and after the DHPC was assessed using the systematic component of variation (SCV). Multilevel segmented regression with random effects quantified the change in level and trend of prescribing after the DHPC. Practice characteristics were assessed for their association with a reduction in prescribing following the DHPC. Results: This study included 7408 practices. During September 2015, 88.9% of practices prescribed mirabegron and mirabegron comprised a mean of 8.2% (SD 6.8) of OAB prescriptions. Variation between practices was classified as very high and the median SCV did not change significantly (p = 0.11) in the six months after the September 2015 DHPC (12.4) compared to before (11.6). Before the DHPC, the share of mirabegron over all OAB drug prescriptions increased by 0.294 (95% confidence interval (CI), 0.287, 0.301) percentage points per month. There was no significant change in the month immediately after the DHPC (−0.023, 95% CI −0.105 to 0.058), however there was a significant reduction in trend (−0.036, 95% CI −0.049 to −0.023). Higher numbers of registered patients, patients aged ≥65 years, and practice area deprivation were associated with having a significant decrease in level and slope of mirabegron prescribing post-DHPC. Conclusion: Variation in mirabegron prescribing was high over the study period and did not change substantively following the DHPC. There was no immediate prescribing change post-DHPC, although the monthly growth did slow. Knowledge of the degree of variation in and determinants of response to safety communications may allow those that do not change prescribing habits to be provided with additional support.

Association Between Accreditation Era, North American Pharmacist Licensure Examination Testing Changes, and First-Time Pass Rates

Objective. To estimate whether first-time pass rates on the North American Pharmacist Licensure Examination (NAPLEX) have been influenced by the number of pharmacy programs founded since 2000, the programs' accreditation era, and the changes to the blueprint as well as changes to the testing conditions and passing standards implemented by the National Association of Boards of Pharmacy (NABP) beginning in 2015.Methods. This was a retrospective, observational cohort study using publicly published data. The number of programs and pass rates were collected from 2008 to 2020. Programs reporting pass rates from 2016 to 2020 were eligible. Accreditation era was defined as programs accredited before or after 2000. Pass rates were categorized into NAPLEX tests administered before or after 2015. Statistical analyses were conducted for comparisons.Results. Pass rates were initially found to decline as the number of programs rose. First-time pass rates of programs accredited before 2000 were higher than pass rates of programs accredited after 2000 every year after 2011. Only 40% of the programs accredited after 2000 exceeded the national average between 2016-2020. Blueprint changes implemented in 2015 and the changes to testing conditions plus passing standards implemented in 2016 had a greater effect on pass rates than the number of programs or applicants.Conclusion. Programs accredited after 2000 generally had lower first-time NAPLEX pass rates. Even so, blueprint changes and changes to the testing conditions plus passing standards instituted by the NABP were more important predictors of the decline of first-time NAPLEX pass rates. Stakeholders should collaborate and embrace best practices for assessing practice-ready competency for licensure.

Design, analysis, power, and sample size calculation for three-phase interrupted time series analysis in evaluation of health policy interventions

OBJECTIVE

To discuss the study design and data analysis for three-phase interrupted time series (ITS) studies to evaluate the impact of health policy, systems, or environmental interventions. Simulation methods are used to conduct power and sample size calculation for these studies.

METHODS

We consider the design and analysis of three-phase ITS studies using a study funded by National Institutes of Health as an exemplar. The design and analysis of both one-arm and two-arm three-phase ITS studies are introduced.

RESULTS

A simulation-based approach, with ready-to-use computer programs, was developed to determine the power for two types of three-phase ITS studies. Simulations were conducted to estimate the power of segmented autoregressive (AR) error models when autocorrelation ranged from -0.9 to 0.9 with various effect sizes. The power increased as the sample size or the effect size increased. The power to detect the same effect sizes varied largely, depending on testing level change, trend changes, or both.

CONCLUSION

This article provides a convenient tool for investigators to generate sample sizes to ensure sufficient statistical power when three-phase ITS study design is implemented.

Effects of Regulation on Carbapenem Prescription in a Large Teaching Hospital in China: An Interrupted Time Series Analysis, 2016-2018

Purpose

Carbapenem resistance due to the overuse of carbapenems has become a public health problem worldwide, particularly in low- and middle-income countries (LMICs). However, there are few policies guiding carbapenem prescription, and their effectiveness is still unclear. A regulation targeting carbapenem prescription was implemented in March 2017 in China. This study aimed to assess the effects of the regulation for providing evidence on the prudent use of carbapenems.

Patients and Methods

This was an interventional, retrospective study started in January 2017. The intervention covered establishing performance appraisal indicators, special authorisation, strict prescribing restrictions, and dedicated supervision, particularly in the intensive care unit (ICU). Data on adult inpatients who received at least one carbapenems were extracted from January 2016 to December 2018. Segmented regression analysis was performed to evaluate the effect of the regulation.

Results

A total of 2005 inpatients received carbapenems. Segmented regression models showed an immediate decline in the intensity of antibiotic consumption (IAC) of carbapenems (coefficient = −9.65, p < 0.001), particularly imipenem (coefficient = −6.82, p = 0.002), and the antibiotic consumption of carbapenems (coefficient = −133.60, p = 0.003) in the ICU. And there is a decreasing trend in the IAC of meropenem (coefficient = −0.03, p = 0.008) in all departments. Furthermore, the IAC of carbapenems and imipenem (coefficient = −0.36, p = 0.035; coefficient = −0.49, p = 0.025, respectively), and the average length of stay (ALoS) (coefficient = −0.73, p < 0.001) showed downward trends in the ICU.

Conclusion

The intervention effectively reduced the IAC of carbapenems and imipenem, carbapenem consumption and the ALoS in the ICU, and the IAC of meropenem in all departments. The effects of the intervention were significant in the ICU, which indicated an urgent need for stronger regulations focusing on critical departments in the future.