A methodological framework for model selection in interrupted time series studies

Motor vehicle collision (MVC) emergency department (ED) visits and hospitalisations in Ontario during the COVID-19 pandemic

BACKGROUND

The COVID-19 pandemic policy response dramatically changed local transportation patterns. This project investigated the impact of COVID-19 policies on motor vehicle collision (MVC)-related emergency department (ED) visits and hospitalisations in Ontario.

METHODS

Data were collected on MVC-related ED visits and hospitalisations in Ontario between March 2016 and December 2022. Using an interrupted time series design, negative binomial regression models were fitted to the pre-pandemic data, including monthly indicator variables for seasonality and accounting for autocorrelation. Extrapolations simulated expected outcome trajectories during the pandemic, which were compared with actual observed outcome counts using the overall per cent change and mean monthly difference. Data were modelled separately for vehicle occupants, pedestrians and cyclists (MVC and non-MVC injuries).

RESULTS

There was a 31.5% decrease in observed ED visits (95% CI -35.4 to -27.3) and a 6.0% decrease in hospitalisations (95% CI -13.2 to 1.6) among vehicle occupants, relative to expected counts during the pandemic. Results were similar for pedestrians. Among cyclist MVCs, there was an increase in ED visits (12.8%, 95% CI -8.2 to 39.4) and hospitalisations (46.0%, 95% CI 11.6 to 93.6). Among non-MVC cyclists, there was also an increase in ED visits (47.0%, 95% CI 12.5 to 86.8) and hospitalisations (50.1%, 95% CI 8.2 to 101.2).

CONCLUSIONS

We observed fewer vehicle occupant and pedestrian collision injuries than expected during the pandemic. By contrast, we observed more cycling injuries than expected, especially in cycling injuries not involving motor vehicles. These observations may be attributable to changes in transportation patterns during the pandemic and increased uptake of recreational cycling.

The influence of the national drug price negotiation policy reform on the medical expenses of patients in Xuzhou City: an interrupted time series analysis

Background

To reduce the burden of patients' medical care, the Xuzhou Municipal Government has initiated an exploratory study on the supply model and categorized management of nationally negotiated drugs. This study aims to understand the extent to which Xuzhou's 2021 reform of the National Drug Price Negotiation (NDPN) policy has had a positive impact on the healthcare costs of individuals with different types of health insurance.

Methods

The Interrupted Time Series Analysis method was adopted, and the changes in average medical expenses per patient, average medical insurance payment cost per patient and actual reimbursement ratio were investigated by using the data of single-drug payments in Xuzhou from October 2020 to October 2022.

Results

Following the implementation of the policy, there was a significant decrease in the average medical expenses per patient of national drug negotiation in Xuzhou, with a reduction of 62.42 yuan per month (p < 0.001). Additionally, the average medical insurance payment cost per patient decreased by 44.13 yuan per month (p = 0.01). Furthermore, the average medical expenses per patient of urban and rural medical insurance participants decreased by 63.45 yuan (p < 0.001), and the average monthly medical insurance payment cost per patient decreased by 57.56 yuan (p < 0.04). However, the mean total medical expenditures for individuals enrolled in employee medical insurance decreased by 63.41 yuan per month (p < 0.001), whereas the monthly decrease was 22.11 yuan per month (p = 0.21). On the other hand, there was no discernible change in the actual reimbursement ratio.

Conclusion

After the adoption of the NDPN policy, a noticeable decline has been observed in the average medical expenses per patient and the mean cost of the average medical insurance payment per patient, although to a limited extent. Notably, the reduction in employee medical insurance surpasses that of urban and rural medical insurance.

Vulnerable road user injury trends following the COVID-19 pandemic in Toronto, Canada: An interrupted time series analysis

BACKGROUND

The COVID-19 pandemic altered traffic patterns worldwide, potentially impacting pedestrian and bicyclists safety in urban areas. In Toronto, Canada, work from home policies, bicycle network expansion, and quiet streets were implemented to support walking and cycling. We examined pedestrian and bicyclist injury trends from 2012 to 2022, utilizing police-reported killed or severely injured (KSI), emergency department (ED) visits and hospitalization data.

METHODS

We used an interrupted time series design, with injury counts aggregated quarterly. We fit a negative binomial regression using a Bayesian modeling approach to data prior to the pandemic that included a secular time trend, quarterly seasonal indicator variables, and autoregressive terms. The differences between observed and expected injury counts based on pre-pandemic trends with 95% credible intervals (CIs) were computed.

RESULTS

There were 38% fewer pedestrian KSI (95%CI: 19%, 52%), 35% fewer ED visits (95%CI: 28%, 42%), and 19% fewer hospitalizations (95%CI: 2%, 32%) since the beginning of the COVID-19 pandemic. A reduction of 35% (95%CI: 7%, 54%) in KSI bicyclist injuries was observed, but However, ED visits and hospitalizations from bicycle-motor vehicle collisions were compatible with pre-pandemic trends. In contrast, for bicycle injuries not involving motor vehicles, large increases were observed for both ED visits, 73% (95% CI: 49%, 103%) and for hospitalization 108% (95% CI: 38%, 208%).

CONCLUSION

New road safety interventions during the pandemic may have improved road safety for vulnerable road users with respect to collisions with motor vehicles; however, further investigation into the risk factors for bicycle injuries not involving motor vehicles is required.

Immediate effects of Ramadan on objective time asleep in male youth football players from the Middle East: an interrupted time-series study

OBJECTIVE

To examine the abrupt effects of Ramadan onset on actigraphy-based time asleep in male youth Muslim football players.

METHODS

We adopted a quasi-experimental, interrupted time-series research design and tracked objective time asleep over a minimum of 12 consecutive nights in the two weeks prior to and immediately after Ramadan onset, respectively. Twenty-two, male academy student-athletes (chronological age range: 12.6 to 16.2 years) participated in the study (464 individual observations). Segmented generalized mixed-effects modelling estimated the effects of Ramadan onset on time asleep during the first period of night sleep only.

RESULTS

Ramadan onset led to an immediate mean reduction of 89 min (95% confidence interval [CI], 54 to 123 min) in time asleep during the first period of night sleep compared to pre-Ramadan sleep patterns. Model-adjusted estimated marginal means for time asleep were ~ 5.7 h (95%CI, 5.1 to 6.2 h) before and ~ 4.2 h (95%CI, 3.6 to 4.7 h) after Ramadan onset. Night sleep interruptions resulting in two or more fragmented periods accounted for 8% (95%CI, 2 to 21%) to 19% (95%, 11 to 29%) of sleep observations before and after Ramadan onset, respectively.

CONCLUSIONS

The onset of Ramadan determined an abrupt reduction in time asleep of ~ 1 h 30 min in the first period of a night cycle and contributed to additional problems of heterogeneous sleep fragmentation that can impact optimal school learning and youth athlete performance development processes.

Speed limits and their effect on air pollution in Mexico City: A quasi-experimental study

Speed limits are an evidence-based intervention to prevent traffic collisions and deaths, yet their impact on air pollution in cities is understudied. The objective of this study was to investigate the association between lower speed limits and air pollution. We leverage the introduction of a new road safety policy in Mexico City in December 2015 which lowered speed limits, increased fines, and installed speed radars to enforce compliance. We tested whether the policy had an impact on particulate matter (PM2.5) and nitrogen dioxide (NO2) at the city level, and whether air-quality monitoring stations' proximity to speed radars moderated this effect due to more acceleration and deceleration around radars. NO2 and PM2.5 concentrations from January 2014 to December 2018 were obtained from the National System of Air Quality Information. Air-quality monitoring stations were classified as in close-proximity or far-from-speed radars. Interrupted time series analyses were conducted for each outcome separately, using linear mixed models and adjusting for seasonality and time-varying confounders: registered vehicles, temperature, wind-speed and relative humidity. The results suggest improvement in both contaminants after the speed limits policy. For NO2, the pre-policy trend was flat, while the post-policy trend showed a decline in concentrations of 0.04 ppb/week. For PM2.5, concentrations were increasing pre-policy by 0.08 μg/m3 per week, then this trend flattened in the post-policy period to a weekly, non-significant, increase of 0.03 μg/m3 (p = 0.08). Air-quality monitors' proximity to speed radars did not moderate the effect of the policy on either of the pollutants. In conclusion, the speed limits policy implemented in Mexico City in 2015 was associated with improvements in air pollution.

The effect of a minimum price per unit of alcohol in Scotland on alcohol-related ambulance call-outs: A controlled interrupted time-series analysis

BACKGROUND AND AIMS

On 1 May 2018, Scotland introduced a minimum unit price (MUP) of £0.50 for alcohol, with one UK unit of alcohol being 10 ml of pure ethanol. This study measured the association between MUP and changes in the volume of alcohol-related ambulance call-outs in the overall population and in call-outs subsets (night-time call-outs and subpopulations with higher incidence of alcohol-related harm).

DESIGN

An interrupted time-series (ITS) was used to measure variations in the daily volume of alcohol-related call-outs. We performed uncontrolled ITS on both the intervention and control group and a controlled ITS built on the difference between the two series. Data were from electronic patient clinical records from the Scottish Ambulance Service.

SETTING AND CASES

Alcohol-related ambulance call-outs (intervention group) and total ambulance call-outs for people aged under 13 years (control group) in Scotland, from December 2017 to March 2020.

MEASUREMENTS

Call-outs were deemed alcohol-related if ambulance clinicians indicated that alcohol was a 'contributing factor' in the call-out and/or a validated Scottish Ambulance Service algorithm determined that the call-out was alcohol-related.

FINDINGS

No statistically significant association in the volume of call-outs was found in both the uncontrolled series [step change = 0.062, 95% confidence interval (CI) = -0.012, 0.0135 P = 0.091; slope change = -0.001, 95% CI = -0.001, 0.1 × 10-3 P = 0.139] and controlled series (step change = -0.01, 95% CI = -0.317, 0.298 P = 0.951; slope change = -0.003, 95% CI = -0.008, 0.002 P = 0.257). Similarly, no significant changes were found for the night-time series or for any population subgroups.

CONCLUSIONS

There appears to be no statistically significant association between the introduction of minimum unit pricing for alcohol in Scotland and the volume of alcohol-related ambulance call-outs. This was observed overall, across subpopulations and at night-time.

Suicide mortality following the implementation of tobacco packaging and pricing policies in Korea: an interrupted time-series analysis

BACKGROUND

To prevent tobacco use in Korea, the national quitline number was added to tobacco packages in December 2012, tobacco prices were raised by 80% in January 2015, and graphic health warning labels were placed on tobacco packages in December 2016. This study evaluated the association of these tobacco packaging and pricing policies with suicide mortality in Korea.

METHODS

Monthly mortality from suicide was obtained from Cause-of-Death Statistics in Korea from December 2007 to December 2019. Interrupted time-series analysis was performed using segmented Poisson regression models. Relative risks (RRs) and 95% confidence intervals (CIs) were calculated adjusted for suicide prevention strategies.

RESULTS

Suicide mortality was 20 per 1,000,000 in December 2007 and showed a downward trend over the study period. After the implementation of tobacco packaging and pricing policies, suicide mortality immediately declined by - 0.09 percent points (95% CI =  - 0.19 to 0.01; P > 0.05) for the national quitline number, - 0.22 percent points (95% CI =  - 0.35 to - 0.09; P < 0.01) for tobacco prices, and - 0.30 percent points (95% CI =  - 0.49 to - 0.11; P < 0.01) for graphic health warning labels. The corresponding RRs for these post-implementation changes compared with the pre-implementation level were 0.91 (95% CI = 0.83 to 1.00), 0.80 (95% CI = 0.70 to 0.91), and 0.74 (95% CI = 0.61 to 0.90), respectively. Significant associations between tobacco control policies and suicide mortality were observed even when stratified by sex and region.

CONCLUSIONS

The findings of this study provide new evidence for an association between tobacco control policies and deaths by suicide. An array of effective tobacco control policies should be considered for prevention programs targeting suicide.

Effect of the COVID-19 pandemic on mental health visits in primary care: an interrupted time series analysis from nine INTRePID countries

Background

The COVID-19 pandemic impacted mental health disorders, affecting both individuals with pre-existing conditions and those with no prior history. However, there is limited evidence regarding the pandemic's impact on mental health visits to primary care physicians. The International Consortium of Primary Care Big Data Researchers (INTRePID) explored primary care visit trends related to mental health conditions in Argentina, Australia, Canada, China, Norway, Peru, Singapore, Sweden, and the USA.

Methods

We conducted an interrupted time series analysis in nine countries to examine changes in rates of monthly mental health visits to primary care settings from January 1st, 2018, to December 31st, 2021. Sub-group analysis considered service type (in-person/virtual) and six categories of mental health conditions (anxiety/depression, bipolar/schizophrenia/other psychotic disorders, sleep disorders, dementia, ADHD/eating disorders, and substance use disorder).

Findings

Mental health visit rates increased after the onset of the pandemic in most countries. In Argentina, Canada, China, Norway, Peru, and Singapore, this increase was immediate ranged from an incidence rate ratio of 1·118 [95% CI 1.053-1.187] to 2.240 [95% CI 2.057-2.439] when comparing the first month of pandemic with the pre-pandemic trend. Increases in the following months varied across countries. Anxiety/depression was the leading reason for mental health visits in most countries. Virtual visits were reported in Australia, Canada, Norway, Peru, Sweden, and the USA, accounting for up to 40% of the total mental health visits.

Interpretation

Findings suggest an overall increase in mental health visits, driven largely by anxiety/depression. During the COVID-19 pandemic, many of the studied countries adopted virtual care in particular for mental health visits. Primary care plays a crucial role in addressing mental ill-health in times of crisis.

Funding

Canadian Institutes of Health Research grant #173094 and the Rathlyn Foundation Primary Care EMR Research and Discovery Fund.

Laboratory Data Timeliness and Completeness Improves Following Implementation of an Electronic Laboratory Information System in Côte d'Ivoire: Quasi-Experimental Study on 21 Clinical Laboratories From 2014 to 2020

BACKGROUND

The Ministry of Health in Côte d'Ivoire and the International Training and Education Center for Health at the University of Washington, funded by the United States President's Emergency Plan for AIDS Relief, have been collaborating to develop and implement the Open-Source Enterprise-Level Laboratory Information System (OpenELIS). The system is designed to improve HIV-related laboratory data management and strengthen quality management and capacity at clinical laboratories across the nation.

OBJECTIVE

This evaluation aimed to quantify the effects of implementing OpenELIS on data quality for laboratory tests related to HIV care and treatment.

METHODS

This evaluation used a quasi-experimental design to perform an interrupted time-series analysis to estimate the changes in the level and slope of 3 data quality indicators (timeliness, completeness, and validity) after OpenELIS implementation. We collected paper and electronic records on clusters of differentiation 4 (CD4) testing for 48 weeks before OpenELIS adoption until 72 weeks after. Data collection took place at 21 laboratories in 13 health regions that started using OpenELIS between 2014 and 2020. We analyzed the data at the laboratory level. We estimated odds ratios (ORs) by comparing the observed outcomes with modeled counterfactual ones when the laboratories did not adopt OpenELIS.

RESULTS

There was an immediate 5-fold increase in timeliness (OR 5.27, 95% CI 4.33-6.41; P<.001) and an immediate 3.6-fold increase in completeness (OR 3.59, 95% CI 2.40-5.37; P<.001). These immediate improvements were observed starting after OpenELIS installation and then maintained until 72 weeks after OpenELIS adoption. The weekly improvement in the postimplementation trend of completeness was significant (OR 1.03, 95% CI 1.02-1.05; P<.001). The improvement in validity was not statistically significant (OR 1.34, 95% CI 0.69-2.60; P=.38), but validity did not fall below pre-OpenELIS levels.

CONCLUSIONS

These results demonstrate the value of electronic laboratory information systems in improving laboratory data quality and supporting evidence-based decision-making in health care. These findings highlight the importance of OpenELIS in Côte d'Ivoire and the potential for adoption in other low- and middle-income countries with similar health systems.

Differences in inpatient performance of public general hospitals following implementation of a points-counting payment based on diagnosis-related group: a robust multiple interrupted time series study in Wenzhou, China

OBJECTIVES

This study measures the differences in inpatient performance after a points-counting payment policy based on diagnosis-related group (DRG) was implemented. The point value is dynamic; its change depends on the annual DRGs' cost settlements and points of the current year, which are calculated at the beginning of the following year.

DESIGN

A longitudinal study using a robust multiple interrupted time series model to evaluate service performance following policy implementation.

SETTING

Twenty-two public general hospitals (8 tertiary institutions and 14 secondary institutions) in Wenzhou, China.

INTERVENTION

The intervention was implemented in January 2020.

OUTCOME MEASURES

The indicators were case mix index (CMI), cost per hospitalisation (CPH), average length of stay (ALOS), cost efficiency index (CEI) and time efficiency index (TEI). The study employed the means of these indicators.

RESULTS

The impact of COVID-19, which reached Zhejiang Province at the end of January 2020, was temporary given rapid containment following strict control measures. After the intervention, except for the ALOS mean, the change-points for the other outcomes (p<0.05) in tertiary and secondary institutions were inconsistent. The CMI mean turned to uptrend in tertiary (p<0.01) and secondary (p<0.0001) institutions compared with before. Although the slope of the CPH mean did not change (p>0.05), the uptrend of the CEI mean in tertiary institutions alleviated (p<0.05) and further increased (p<0.05) in secondary institutions. The slopes of the ALOS and TEI mean in secondary institutions changed (p<0.05), but not in tertiary institutions (p>0.05).

CONCLUSIONS

This study showed a positive effect of the DRG policy in Wenzhou, even during COVID-19. The policy can motivate public general hospitals to improve their comprehensive capacity and mitigate discrepancies in treatment expenses efficiency for similar diseases. Policymakers are interested in whether the reform successfully motivates hospitals to strengthen their internal impetus and improve their performance, and this is supported by this study.

Design and statistical analysis reporting among interrupted time series studies in drug utilization research: a cross-sectional survey

INTRODUCTION

Interrupted time series (ITS) design is a commonly used method for evaluating large-scale interventions in clinical practice or public health. However, improperly using this method can lead to biased results.

OBJECTIVE

To investigate design and statistical analysis characteristics of drug utilization studies using ITS design, and give recommendations for improvements.

METHODS

A literature search was conducted based on PubMed from January 2021 to December 2021. We included original articles that used ITS design to investigate drug utilization without restriction on study population or outcome types. A structured, pilot-tested questionnaire was developed to extract information regarding study characteristics and details about design and statistical analysis.

RESULTS

We included 153 eligible studies. Among those, 28.1% (43/153) clearly explained the rationale for using the ITS design and 13.7% (21/153) clarified the rationale of using the specified ITS model structure. One hundred and forty-nine studies used aggregated data to do ITS analysis, and 20.8% (31/149) clarified the rationale for the number of time points. The consideration of autocorrelation, non-stationary and seasonality was often lacking among those studies, and only 14 studies mentioned all of three methodological issues. Missing data was mentioned in 31 studies. Only 39.22% (60/153) reported the regression models, while 15 studies gave the incorrect interpretation of level change due to time parameterization. Time-varying participant characteristics were considered in 24 studies. In 97 studies containing hierarchical data, 23 studies clarified the heterogeneity among clusters and used statistical methods to address this issue.

CONCLUSION

The quality of design and statistical analyses in ITS studies for drug utilization remains unsatisfactory. Three emerging methodological issues warranted particular attention, including incorrect interpretation of level change due to time parameterization, time-varying participant characteristics and hierarchical data analysis. We offered specific recommendations about the design, analysis and reporting of the ITS study.

Reductions in smoking due to ratification of the Framework Convention for Tobacco Control in 171 countries

Smoking globally kills over half of long-term smokers and causes about 7 million annual deaths. The World Health Organization Framework Convention for Tobacco Control (FCTC) is the main global policy strategy to combat smoking, but its effectiveness is uncertain. Our interrupted time series analyses compared before- and after-FCTC trends in the numbers and prevalence of smokers below the age of 25 years (when smoking initiation occurs and during which response to interventions is greatest) and on cessation at 45-59 years (when quitting probably occurs) in 170 countries, excluding China. Contrasting the 10 years after FCTC ratification with the income-specific before-FCTC trends, we observed cumulative decreases of 15.5% (95% confidence interval = -33.2 to -0.7) for the numbers of current smokers and decreases of -7.5% (95% CI = -10.6 to -4.5) for the prevalence of smoking below age 25 years. The quit ratio (comparing the numbers of former and ever smokers) at 45-59 years increased by 1.8% (1.2 to 2.3) 10 years after FCTC ratification. Countries raising taxes by at least 10 percentage points concurrent with ratification observed steeper decreases in all three outcomes than countries that did not. Over a decade across 170 countries, the FCTC was associated with 24 million fewer young smokers and 2 million more quitters.

Impact of the COVID-19 pandemic on antibiotic prescribing in high-prescribing primary care physicians in Switzerland

OBJECTIVES

The aim of this study was to evaluate antibiotic prescribing of medium-to-high prescribing primary care physicians being followed up after the completion of a Swiss national intervention trial of antibiotic prescription audit and feedback in the first SARS-CoV-2 pandemic year.

METHODS

We used health insurer based claims data to calculate monthly antibiotic prescription rates per 100 consultations (primary endpoint) and applying interrupted time series (ITS) analysis methods, we estimated the immediate (step change) and sustained effects (slope) of the SARS-CoV-2 epidemic in 2020 on antibiotic prescribing compared to the pre-pandemic trial period from 2017-2019.

RESULTS

We analysed data of 2945 of 3426 physicians (86.0%) from the trial with over 4 million consultations annually, who were in 2020 still in practice. Consultations dropped by 43% during the first pandemic year compared with 2017. Median monthly antibiotic prescription rates per 100 consultations in 2017 were 8.44 (Interquartile range [IQ] 6.32-11.50) and 8.35 (6.34-11.74) in the intervention and control groups, respectively, and increased to 15.63 (10.69-23.81) and 16.31 (10.65-24.72) per 100 consultations in 2020. ITS-derived incidence rate ratios for overall antibiotic prescriptions were 2.32 (95% CI 2.07-2.59) for the immediate pandemic effect, and 0.96 (0.95-0.98) for the sustained effect (change in slope in 2020 compared with 2017-2019).

DISCUSSION

The SARS-CoV-2 pandemic had a major impact on antibiotic prescription patterns in primary care in Switzerland. For future viral pandemics, intervention plans with timely activation steps to minimize unjustified antibiotic consumption in primary care should be prepared.

Interpretation of coefficients in segmented regression for interrupted time series analyses

Background Segmented regression, a common model for interrupted time series (ITS) analysis, primarily utilizes two equation parametrizations. Interpretations of coefficients vary between the two segmented regression parametrizations, leading to occasional user misinterpretations. Methods To illustrate differences in coefficient interpretation between two common parametrizations of segmented regression in ITS analysis, we derived analytical results and present an illustration evaluating the impact of a smoking regulation policy in Italy using a publicly accessible dataset. Estimated coefficients and their standard errors were obtained using two commonly used parametrizations for segmented regression with continuous outcomes. We clarified coefficient interpretations and intervention effect calculations. Results Our investigation revealed that both parametrizations represent the same model. However, due to differences in parametrization, the immediate effect of the intervention is estimated differently under the two approaches. The key difference lies in the interpretation of the coefficient related to the binary indicator for intervention implementation, impacting the calculation of the immediate effect. Conclusions Two common parametrizations of segmented regression represent the same model but have different interpretations of a key coefficient. Researchers employing either parametrization should exercise caution when interpreting coefficients and calculating intervention effects.

Prescriptions of antidepressants and anxiolytics in France 2012-2022 and changes with the COVID-19 pandemic: interrupted time series analysis

BACKGROUND

Depression and anxiety have increased in prevalence since the start of the COVID-19 pandemic.

OBJECTIVE

To evaluate the consumption of antidepressants and anxiolytics from 2012 to 2022 and the pandemic's potential impact in France.

METHODS

We conducted an interrupted time series analysis of routine drug sales data (Medic'AM) from all French outpatient pharmacies from 2012 to 2022. We investigated trends in defined daily doses of antidepressants and anxiolytics sold per 1000 inhabitants (DDD/TID) and related expenditures before and after pandemic onset and in relation with stringency of pandemic mitigation measures. Analyses were performed descriptively and using segmented linear regression, autoregressive and autoregressive integrated moving average models.

FINDINGS

From 2012 to 2019, overall monthly antidepressant sales increased (+0.02 DDD/TID) while monthly anxiolytic sales decreased (-0.07 DDD/TID). With pandemic onset, there was a relevant and persisting trend increase (+0.20 DDD/TID per month) for antidepressant sales overall, with an estimated excess of 112.6 DDD/TID sold from May 2020 until December 2022. Anxiolytic sales were elevated from February 2020 throughout the pandemic but returned to expected levels by December 2022, with an estimated excess of 33.8 DDD/TID. There was no evident association between stringency and antidepressant or anxiolytic sales.

CONCLUSIONS

This study showed a protracted trend increase in the consumption of antidepressants since pandemic onset, while increases in anxiolytic consumption were temporary.

CLINICAL IMPLICATIONS

We provide evidence that the COVID-19 pandemic may have had long-lasting consequences on the prevalence and treatment of depression and anxiety disorders, requiring further actions by researchers and policy-makers to address this potential public mental health crisis.

Drug utilisation in children and adolescents before and after the start of the COVID-19 pandemic: Interrupted time-series analyses in three European countries

BACKGROUND

The COVID-19 pandemic has affected children and adolescents in several ways, including worsened mental health, improvement of asthma, and increases in diabetes ketoacidosis. Less is known about how medication use in children and adolescents has been affected by the pandemic.

OBJECTIVES

To explore how the COVID-19 pandemic affected drug utilisation in children and adolescents in Norway, Sweden, and Italy, by child age.

METHODS

We conducted a longitudinal drug utilisation study among all children and adolescents (<18 years old) in Norway and Sweden and a nationwide paediatric database covering 3% of the paediatric population in Italy. We conducted an interrupted time-series analysis from January 2018 to December 2021, with March 2020 as the interruption point. Dispensing or prescription rates of antidepressants, anxiolytics, sleep medications, attention-deficit/hyperactivity disorder (ADHD) medications, insulin, and asthma medications were examined.

RESULTS

The study population in January 2018 consisted of 3,455,521 children and adolescents (136,188 from Italy, 1,160,431 from Norway, and 2,158,902 from Sweden). For sleep medications and insulin, there were only minor changes in level or trend in some age groups after March 2020. For asthma medications, the pandemic was associated with an immediate decrease in dispensing in Norway and Sweden (range of change in level: -19.2 to -3.7 dispensings per 1000 person-months), and an increasing trend in all countries afterward (range of change in trend: 0.3-6.4 dispensings per 1000 person-months), especially for the youngest age groups. Among adolescents, the pandemic was associated with an increased trend for ADHD medications, antidepressants, and anxiolytics in Norway and Sweden, but not in Italy.

CONCLUSIONS

The increasing trend of psychotropic medication dispensing, especially among adolescents after the start of the pandemic, is concerning and should be investigated further. Aside from a temporary effect on asthma medication dispensing, the pandemic did not greatly affect the dispensing of the medications investigated.

Comparing the effects of decreasing prescription opioid shipments and the release of an abuse deterrent OxyContin formulation on opioid overdose fatalities in WV: an interrupted time series study

INTRODUCTION

The 2010 release of an abuse deterrent formulation (ADF) of OxyContin, a brand name prescription opioid, has been cited as a major driver for the reduction in prescription drug misuse and the associated increasing illicit opioid use and overdose rates. However, studies of this topic often do not account for changes in supplies of other prescription opioids that were widely prescribed before and after the ADF OxyContin release, including generic oxycodone formulations and hydrocodone. We therefore sought to compare the impact of the ADF OxyContin release to that of decreasing prescription opioid supplies in West Virginia (WV).

METHODS

Opioid tablet shipment and overdose data were extracted from The Washington Post ARCOS (2006-2014) and the WV Forensic Drug Database (2005-2020), respectively. Locally estimated scatterplot smoothing (LOESS) was used to estimate the point when shipments of prescription opioids to WV began decreasing, measured via dosage units and morphine milligram equivalents (MMEs). Interrupted time series analysis (ITSA) was used to compare the impact LOESS-identified prescription supply changes and the ADF OxyContin release had on prescription (oxycodone and hydrocodone) and illicit (heroin, fentanyl, and fentanyl analogues) opioid overdose deaths in WV. Model fit was compared using Akaike Information Criteria (AIC).

RESULTS

The majority of opioid tablets shipped to WV from 2006 to 2014 were generic oxycodone or hydrocodone, not OxyContin. After accounting for a 6-month lag from ITSA models using the LOESS-identified change in prescription opioid shipments measured via dosage units (2011 Q3) resulted in the lowest AIC for both prescription (AIC = -188.6) and illicit opioid-involved overdoses (AIC = -189.4), indicating this intervention start date resulted in the preferred model. The second lowest AIC was for models using the ADF OxyContin release as an intervention start date.

DISCUSSION

We found that illicit opioid overdoses in WV began increasing closer to when prescription opioid shipments to the state began decreasing, not when the ADF OxyContin release occurred. Similarly, the majority of opioid tablets shipped to the state for 2006-2014 were generic oxycodone or hydrocodone. This may indicate that diminishing prescription supplies had a larger impact on opioid overdose patterns than the ADF OxyContin release in WV.

Impact of Statewide Statute Limiting Days' Supply to Opioid-Naive Patients

INTRODUCTION

To address the ongoing opioid crisis, states use policy enactment to restrict prescribing by licensed healthcare providers and mandate the use of Prescription Drug Monitoring Programs. There have been mixed results regarding the effectiveness of such state policies. The purpose of this study is to evaluate the impact of Colorado Senate Bill 18-022, which limits opioid prescriptions to ≤7-day supply among patients without an opioid prescription in the previous year (i.e., are opioid naive).

METHODS

This is a retrospective interrupted time-series analysis of opioid prescribing to evaluate the weekly percentage of opioid prescriptions consistent with statutory limits for ≤7-day supply among opioid-naive patients before and after enactment using Prescription Drug Monitoring Programs data from May 21, 2017 to May 25, 2019. Statistical analysis was performed in 2021-2022.

RESULTS

The weekly percentage of opioid prescriptions ≤7-day supply increased by an average of 0.12% per week (p<0.0001) from 79.7% to 87.4% in the week before enactment. The week after enactment, the average increased by 0.2% (p=0.67). The year after enactment, the average weekly percentage change was 0.07% per week, a 0.05% decrease (p=0.01).

CONCLUSIONS

Statutory limits on days' supply among opioid-naive patients had little impact on opioid prescribing in Colorado. Legislating limits on opioid prescribing should be evaluated using Prescription Drug Monitoring Program data and considered for deimplementation when not impactful.

Modeling Global Monkeypox Infection Spread Data: A Comparative Study of Time Series Regression and Machine Learning Models

The global impact of COVID-19 has heightened concerns about emerging viral infections, among which monkeypox (MPOX) has become a significant public health threat. To address this, our study employs a comprehensive approach using three statistical techniques: Distribution fitting, ARIMA modeling, and Random Forest machine learning to analyze and predict the spread of MPOX in the top ten countries with high infection rates. We aim to provide a detailed understanding of the disease dynamics and model theoretical distributions using country-specific datasets to accurately assess and forecast the disease's transmission. The data from the considered countries are fitted into ARIMA models to determine the best time series regression model. Additionally, we employ the random forest machine learning approach to predict the future behavior of the disease. Evaluating the Root Mean Square Errors (RMSE) for both models, we find that the random forest outperforms ARIMA in six countries, while ARIMA performs better in the remaining four countries. Based on these findings, robust policy-making should consider the best fitted model for each country to effectively manage and respond to the ongoing public health threat posed by monkeypox. The integration of multiple modeling techniques enhances our understanding of the disease dynamics and aids in devising more informed strategies for containment and control.

Introducing activation functions into segmented regression model to address lag effects of interventions

The interrupted time series (ITS) design is widely used to examine the effects of large-scale public health interventions and has the highest level of evidence validity. However, there is a notable gap regarding methods that account for lag effects of interventions.To address this, we introduced activation functions (ReLU and Sigmoid) to into the classic segmented regression (CSR) of the ITS design during the lag period. This led to the proposal of proposed an optimized segmented regression (OSR), namely, OSR-ReLU and OSR-Sig. To compare the performance of the models, we simulated data under multiple scenarios, including positive or negative impacts of interventions, linear or nonlinear lag patterns, different lag lengths, and different fluctuation degrees of the outcome time series. Based on the simulated data, we examined the bias, mean relative error (MRE), mean square error (MSE), mean width of the 95% confidence interval (CI), and coverage rate of the 95% CI for the long-term impact estimates of interventions among different models.OSR-ReLU and OSR-Sig yielded approximately unbiased estimates of the long-term impacts across all scenarios, whereas CSR did not. In terms of accuracy, OSR-ReLU and OSR-Sig outperformed CSR, exhibiting lower values in MRE and MSE. With increasing lag length, the optimized models provided robust estimates of long-term impacts. Regarding precision, OSR-ReLU and OSR-Sig surpassed CSR, demonstrating narrower mean widths of 95% CI and higher coverage rates.Our optimized models are powerful tools, as they can model the lag effects of interventions and provide more accurate and precise estimates of the long-term impact of interventions. The introduction of an activation function provides new ideas for improving of the CSR model.

Evaluation of statistical methods used to meta-analyse results from interrupted time series studies: A simulation study

Interrupted time series (ITS) are often meta-analysed to inform public health and policy decisions but examination of the statistical methods for ITS analysis and meta-analysis in this context is limited. We simulated meta-analyses of ITS studies with continuous outcome data, analysed the studies using segmented linear regression with two estimation methods [ordinary least squares (OLS) and restricted maximum likelihood (REML)], and meta-analysed the immediate level- and slope-change effect estimates using fixed-effect and (multiple) random-effects meta-analysis methods. Simulation design parameters included varying series length; magnitude of lag-1 autocorrelation; magnitude of level- and slope-changes; number of included studies; and, effect size heterogeneity. All meta-analysis methods yielded unbiased estimates of the interruption effects. All random effects meta-analysis methods yielded coverage close to the nominal level, irrespective of the ITS analysis method used and other design parameters. However, heterogeneity was frequently overestimated in scenarios where the ITS study standard errors were underestimated, which occurred for short series or when the ITS analysis method did not appropriately account for autocorrelation. The performance of meta-analysis methods depends on the design and analysis of the included ITS studies. Although all random effects methods performed well in terms of coverage, irrespective of the ITS analysis method, we recommend the use of effect estimates calculated from ITS methods that adjust for autocorrelation when possible. Doing so will likely to lead to more accurate estimates of the heterogeneity variance.

Impact of the COVID-19 pandemic on acute coronary syndrome hospital admission and management in Slovenia

AIMS

We evaluated the effects of the COVID-19 pandemic on hospital admission and quality of care for acute coronary syndrome.

METHODS AND RESULTS

Data for all patients admitted to hospital care for acute coronary syndromes in Slovenia (nationwide cohort) between 2014 and 2021 were obtained by merging the national hospital database, national medicines reimbursement database and population mortality registry using unique identifying numbers. Using interrupted time series analysis, we assessed the impact of the COVID-19 pandemic on hospital admission rates and quality of care (in-hospital and 30-day mortality, reperfusion and secondary preventive medication uptake). Data were fitted to segmented regression models with March 2020 as the breakpoint. Data on 21 001 patients were included (7057 ST-elevation myocardial infarction (STEMI), 7649 non-ST elevation myocardial infarction (NSTEMI) and 6295 unstable angina). Hospital admissions for STEMI remained stable (92 patients; +1 patient per month, p=0.783), whereas the pandemic was associated with a significant reduction in NSTEMI (81 patients; -21 patients per month, p=0.015) and unstable angina admissions (47 patients; -28 patients per month, p=0.025). In patients with STEMI, the pandemic did not affect reperfusion rates (0.29%, (95% CI) -1.5% to 2.1%, p=0.755) or in-hospital mortality (0.1%, (95% CI) -0.9% to 1.1%, p=0.815), but was associated with a significant negative trend for secondary preventive medication uptake (-0.12%, (95% CI) -0.23% to -0.01%, p=0.034).

CONCLUSION

In Slovenia, hospital admissions for STEMI remained stable throughout the COVID-19 pandemic, but NSTEMI and unstable angina admissions dropped significantly. While mortality and reperfusion rates were not affected, the pandemic was associated with a continual negative time trend for the uptake of secondary preventive medication.

Impact of community mental health services on the adult psychiatric admission through the emergency unit: a 20-year population-based study

Background

There is a lack of real-life population-based study examining the effect of community mental health services on psychiatric emergency admission. In Hong Kong, Integrated Community Center for Mental Wellness (ICCMW) and telecare service were introduced in 2009 and 2012, respectively. We examined the real-life impact of these services on psychiatric emergency admissions over 20 years.

Methods

Number of psychiatric emergency admissions between 2001 and 2020 was retrieved from the Hong Kong Clinical Data Analysis & Reporting System. We used an interrupted time series analysis to examine monthly psychiatric admission trend before and after service implementation, considering socioeconomic and environmental covariates.

Findings

A total of 108,492 psychiatric emergency admissions (47.8% males; 64.9% aged 18-44 years) were identified from the study period, of which 56,858, 12,506, 12,295, 11,791, and 15,051 were that for schizophrenia-spectrum disorders, bipolar affective disorders, unipolar mood disorders, neuroses, and substance use disorders. ICCMW introduction has an immediate effect on psychiatric emergency admission (adjusted estimate per 100,000: -10.576; 95% CI, -16.635 to -4.518, p < 0.001), particularly among adults aged 18-44 years (-8.543; 95% CI, -13.209 to -3.877, p < 0.001), females (-5.843; 95% CI, -9.647 to -2.039, p = 0.003), and with neuroses (-3.373; 95% CI, -5.187 to -1.560, p < 0.001), without a significant long-term effect. Unemployment, seasonality, and infectious disease outbreak were significant covariates.

Interpretation

ICCMW reduced psychiatric emergency admission, but no further reduction following full implementation. Community mental health services should be dynamically tailored for different populations and socioeconomic variations over time.

Funding

None.

Impact of differentiated service delivery models on 12-month retention in HIV treatment in Mozambique: an interrupted time-series analysis

BACKGROUND

HIV treatment has been available in Mozambique since 2004, but coverage of, and retention in, antiretroviral therapy (ART) remain suboptimal. Therefore, to increase health system efficiency and reduce HIV-associated mortality, in November, 2018, the Ministry of Health launched national guidelines on implementing eight differentiated service delivery models (DSDMs) for HIV treatment. We assessed the effect of this implementation on retention in ART 12 months after initiation, and explored the associated effects of COVID-19.

METHODS

In this uncontrolled interrupted time-series analysis, data were extracted from the Mozambique ART database, which contains data on individuals in ART care from 1455 health facilities providing ART in Mozambique. We included individual-level data from facilities that were providing ART at the beginning of the study period (Jan 1, 2016) and at the start of DSDM implementation (Dec 1, 2018). We compared the proportion of individuals retained in ART 12 months after initiation between the periods before (Jan 1, 2017, to Nov 30, 2018) and after (Dec 1, 2019, to June 30, 2021) implementation of the DSDMs, overall and stratified by sex and age. We applied a generalised estimating equation model with a working independence correlation and cluster-robust standard errors to account for clustering at the facility level. In a secondary analysis, we assessed the effect of COVID-19 response measures during the post-intervention period on ART retention.

FINDINGS

The study included 613 facilities and 1 131 118 individuals who started ART during the inclusion period up to June 30, 2020, of whom 79 178 (7·0%) were children (age ≤14 years), 226 224 (20·0%) were adolescents and young adults (age 15-24 years), and 825 716 (73·0%) were adults (age ≥25 years). 731 623 (64·7%) were female and 399 495 (35·3%) were male. Introduction of the DSDMs was associated with an estimated increase of 24·5 percentage points (95% CI 21·1 to 28·0) in 12-month ART retention by the end of the study period, compared with the counterfactual scenario without DSDM implementation. By age, the smallest effect was estimated in children (6·1 percentage points, 1·3 to 10·9) and the largest effect in adolescents and young adults (28·8 percentage points, 24·2 to 33·4); by sex, a larger effect was estimated in males (29·7 percentage points, 25·6 to 33·7). Our analysis showed that COVID-19 had an overall negative effect on 12-month retention in ART compared with a counterfactual scenario based on the post-intervention period without COVID-19 (-10·0 percentage points, -18·2 to -1·8).

INTERPRETATION

The implementation of eight DSDMs for HIV treatment had a positive impact on 12-month retention in ART. COVID-19 negatively influenced this outcome.

FUNDING

None.

TRANSLATION

For the Portuguese translation of the abstract see Supplementary Materials section.

Clinical outcomes of non-COVID-19 orthopaedic patients admitted during the COVID-19 pandemic: a multi-centre interrupted time series analysis across hospitals in six different countries

OBJECTIVES

To assess across seven hospitals from six different countries the extent to which the COVID-19 pandemic affected the volumes of orthopaedic hospital admissions and patient outcomes for non-COVID-19 patients admitted for orthopaedic care.

DESIGN

A multi-centre interrupted time series (ITS) analysis.

SETTING

Seven hospitals from six countries who collaborated within the Global Health Data@Work collaborative.

PARTICIPANTS

Non-COVID-19 patients admitted for orthopaedic care during the pre-pandemic (January/2018-February/2020) and COVID-19 pandemic (March/2020-June/2021) period. Admissions were categorised as: (1) acute admissions (lower limb fractures/neck of femur fractures/pathological fractures/joint dislocations/upper limb fractures); (2) subacute admissions (bone cancer); (3) elective admissions (osteoarthritis).

OUTCOME MEASURES

Monthly observed versus expected ratios (O/E) were calculated for in-hospital mortality, long (upper-decile) length-of-stay and hospital readmissions, with expected rates calculated based on case-mix. An ITS design was used to estimate the change in level and/or trend of the monthly O/E ratio by comparing the COVID-19 pandemic with the pre-pandemic period.

RESULTS

69 221 (pre-pandemic) and 22 940 (COVID-19 pandemic) non-COVID-19 orthopaedic patient admissions were included. Admission volumes were reduced during the COVID-19 pandemic for all admission categories (range: 33%-45%), with more complex patients treated as shown by higher percentages of patients admitted with ≥1 comorbidity (53.8% versus 49.8%, p<0.001). The COVID-19 pandemic was not associated with significant changes in patient outcomes for most diagnostic groups. Only for patients diagnosed with pathological fractures (pre-pandemic n=1671 and pandemic n=749), the COVID-19 pandemic was significantly associated with an immediate mortality reduction (level change of -77.7%, 95% CI -127.9% to -25.7%) and for lower limb fracture patients (pre-pandemic n=9898 and pandemic n=3307) with a significantly reduced trend in readmissions (trend change of -6.3% per month, 95% CI -11.0% to -1.6%).

CONCLUSIONS

Acute, subacute, as well as elective orthopaedic hospital admissions volumes were reduced in all global participating hospitals during the COVID-19 pandemic, while overall patient outcomes for most admitted non-COVID-19 patients remained the same despite the strain caused by the surge of COVID-19 patients.

Impact of pesticide regulations on mortality from suicide by pesticide in China: an interrupted time series analysis

Background

Pesticide bans and regulatory restrictions have been shown to be effective strategies for preventing suicide in several countries. Suicide and suicide by pesticides have decreased significantly in China over the past two decades. However, whether the reduction was associated with pesticide regulation is unknown.

Methods

The monthly data on suicide and suicide by pesticide from 2006 to 2018 were obtained from China's Disease Surveillance Point (DSP) system. Information on China's pesticide regulations since 1970 was obtained from Pesticide Action Network International (PAN International), Joint Meeting on Pesticide Management Highly Hazardous Pesticides (JMPM HHP) lists, the website of the Ministry of Agriculture of China, Pesticide Information Network of China, and the Wan Fang database. Change point detection and policy analysis were combined to identify the time of any trend change breakpoint of suicide and suicide by pesticide. Interrupted time series analysis was used to investigate the pre- and post-breakpoint trends of monthly standardized rates in suicide and suicide by pesticide.

Results

The standardized pesticide suicide rate decreased by 60.5% from 6.50 in 2006 to 2.56 per 100,000 in 2018. Larger declines were evident among people in urban areas (67.3%), female individuals (63.5%), and people aged 15-44 years (68.1%). The effect of policies banning highly hazardous organophosphorus pesticides (HHOP) [rate ratio (RR) = 0.993, 95% CIs (0.991-0.994)] in December 2008 and stopping domestic sales and use of paraquat aqueous solution (RR = 0.992, 95% CIs: 0.990-0.994) in July 2016 were more pronounced than regulating the paraquat-related products (RR = 1.003, 95% CIs: 1.002-1.004) in April 2012.

Conclusion

Declines in suicide by pesticide in China occurred contemporaneously with regulatory bans and restrictions implemented on several pesticides, particularly in urban areas, among female individuals, and the relatively low age profile. These findings indicate the potential influence of these bans on trends of suicide by pesticides.

Impact of a vaccine passport on first-dose SARS-CoV-2 vaccine coverage by age and area-level social determinants of health in the Canadian provinces of Quebec and Ontario: an interrupted time series analysis

BACKGROUND

In Canada, all provinces implemented vaccine passports in 2021 to reduce SARS-CoV-2 transmission in non-essential indoor spaces and increase vaccine uptake (policies active September 2021-March 2022 in Quebec and Ontario). We sought to evaluate the impact of vaccine passport policies on first-dose SARS-CoV-2 vaccination coverage by age, and area-level income and proportion of racialized residents.

METHODS

We performed interrupted time series analyses using data from Quebec's and Ontario's vaccine registries linked to census information (population of 20.5 million people aged ≥ 12 yr; unit of analysis: dissemination area). We fit negative binomial regressions to first-dose vaccinations, using natural splines adjusting for baseline vaccination coverage (start: July 2021; end: October 2021 for Quebec, November 2021 for Ontario). We obtained counterfactual vaccination rates and coverage, and estimated the absolute and relative impacts of vaccine passports.

RESULTS

In both provinces, first-dose vaccination coverage before the announcement of vaccine passports was 82% (age ≥ 12 yr). The announcement resulted in estimated increases in coverage of 0.9 percentage points (95% confidence interval [CI] 0.4-1.2) in Quebec and 0.7 percentage points (95% CI 0.5-0.8) in Ontario. This corresponds to 23% (95% CI 10%-36%) and 19% (95% CI 15%-22%) more vaccinations over 11 weeks. The impact was larger among people aged 12-39 years. Despite lower coverage in lower-income and more-racialized areas, there was little variability in the absolute impact by area-level income or proportion racialized in either province.

INTERPRETATION

In the context of high vaccine coverage across 2 provinces, the announcement of vaccine passports had a small impact on first-dose coverage, with little impact on reducing economic and racial inequities in vaccine coverage. Findings suggest that other policies are needed to improve vaccination coverage among lower-income and racialized neighbourhoods and communities.

A dynamic machine learning model for prediction of NAFLD in a health checkup population: A longitudinal study

Background

Non-alcoholic fatty liver disease (NAFLD) is one of the most common liver diseases worldwide. Currently, most NAFLD prediction models are diagnostic models based on cross-sectional data, which failed to provide early identification or clarify causal relationships. We aimed to use time-series deep learning models with longitudinal health checkup records to predict the onset of NAFLD in the future, and update the model stepwise by incorporating new checkup records to achieve dynamic prediction.

Methods

10,493 participants with over 6 health checkup records from Beijing MJ Health Screening Center were included to conduct a retrospective cohort study, in which the constantly updated initial 5 checkup data were incorporated stepwise to predict the risk of NAFLD at and after their sixth health checkups. A total of 33 variables were considered, consisting of demographic characteristics, medical history, lifestyle, physical examinations, and laboratory tests. L1-penalized logistic regression (LR) was used for feature selection. The long short-term memory (LSTM) algorithm was introduced for model development, and five-fold cross-validation was conducted to tune and choose optimal hyperparameters. Both internal validation and external validation were conducted, using the 20% randomly divided holdout test dataset and previously unseen data from Shanghai MJ Health Screening Center, respectively, to evaluate model performance. The evaluation metrics included area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, Brier score, and decision curve. Bootstrap sampling was implemented to generate 95% confidence intervals of all the metrics. Finally, the Shapley additive explanations (SHAP) algorithm was applied in the holdout test dataset for model interpretability to obtain time-specific and sample-specific contributions of each feature.

Results

Among the 10,493 participants, 1662 (15.84%) were diagnosed with NAFLD at and after their sixth health checkups. The predictive performance of the deep learning model in the internal validation dataset improved over the incorporation of the checkups, with AUROC increasing from 0.729 (95% CI: 0.698,0.760) at baseline to 0.818 (95% CI: 0.798,0.844) when consecutive 5 checkups were included. The external validation dataset, containing 1728 participants, was used to verify the results, in which AUROC increased from 0.700 (95% CI: 0.657,0.740) with only the first checkups to 0.792 (95% CI: 0.758,0.825) with all five. The results of feature significance showed that body fat percentage, alanine transaminase (ALT), and uric acid owned the greatest impact on the outcome, time-specific, individual-specific and dynamic feature contributions were also produced for model interpretability.

Conclusion

A dynamic prediction model was successfully established in our study, and the prediction capability kept improving with the renewal of the latest checkup records. In addition, we identified key features associated with the onset of NAFLD, making it possible to optimize the prevention and control strategies of the disease in the general population.

Optimized segmented regression models for the transition period of intervention effects

BACKGROUND

The interrupted time series (ITS) design is a widely used approach to examine the effects of interventions. However, the classic segmented regression (CSR) method, the most popular statistical technique for analyzing ITS data, may not be adequate when there is a transitional period between the pre- and post-intervention phases.

METHODS

To address this issue and better capture the distribution patterns of intervention effects during the transition period, we propose using different cumulative distribution functions in the CSR model and developing corresponding optimized segmented regression (OSR) models. This study illustrates the application of OSR models to estimate the long-term impact of a national free delivery service policy intervention in Ethiopia.

RESULTS

Regardless of the choice of transition length ([Formula: see text]) and distribution patterns of intervention effects, the OSR models outperformed the CSR model in terms of mean square error (MSE), indicating the existence of a transition period and the validity of our model's assumptions. However, the estimates of long-term impacts using OSR models are sensitive to the selection of L, highlighting the importance of reasonable parameter specification. We propose a data-driven approach to select the transition period length to address this issue.

CONCLUSIONS

Overall, our OSR models provide a powerful tool for modeling intervention effects during the transition period, with a superior model fit and more accurate estimates of long-term impacts. Our study highlights the importance of appropriate statistical methods for analyzing ITS data and provides a useful framework for future research.

Effect estimates can be accurately calculated with data digitally extracted from interrupted time series graphs

Interrupted time series (ITS) studies are frequently used to examine the impact of population-level interventions or exposures. Systematic reviews with meta-analyses including ITS designs may inform public health and policy decision-making. Re-analysis of ITS may be required for inclusion in meta-analysis. While publications of ITS rarely provide raw data for re-analysis, graphs are often included, from which time series data can be digitally extracted. However, the accuracy of effect estimates calculated from data digitally extracted from ITS graphs is currently unknown. Forty-three ITS with available datasets and time series graphs were included. Time series data from each graph was extracted by four researchers using digital data extraction software. Data extraction errors were analysed. Segmented linear regression models were fitted to the extracted and provided datasets, from which estimates of immediate level and slope change (and associated statistics) were calculated and compared across the datasets. Although there were some data extraction errors of time points, primarily due to complications in the original graphs, they did not translate into important differences in estimates of interruption effects (and associated statistics). Using digital data extraction to obtain data from ITS graphs should be considered in reviews including ITS. Including these studies in meta-analyses, even with slight inaccuracy, is likely to outweigh the loss of information from non-inclusion.

Spillover effects of financial incentives for providers onto non-targeted patients: daycase surgery in English hospitals

BACKGROUND

Incentives for healthcare providers may also affect non-targeted patients. These spillover effects have important implications for the full impact and evaluation of incentive schemes. However, there are few studies on the extent of such spillovers in health care. We investigated whether incentives to perform surgical procedures as daycases affected whether other elective procedures in the same specialties were also treated as daycases.

DATA

8,505,754 patients treated for 92 non-targeted procedures in 127 hospital trusts in England between April and March 2016.

METHODS

Interrupted time series analysis of the probability of being treated as a daycase for non-targeted patients treated in six specialties where targeted patients were also treated and three specialties where they were not.

RESULTS

The daycase rate initially increased (1.04 percentage points, SE: 0.30) for patients undergoing a non-targeted procedure in incentivised specialties but then reduced over time. Conversely, the daycase rate gradually decreased over time for patients treated in a non-incentivised specialty.

DISCUSSION

Spillovers from financial incentives have variable effects over different activities and over time. Policymakers and researchers should consider the possibility of spillovers in the design and evaluation of incentive schemes.

The health impact of Scotland's Baby Box Scheme: a natural experiment evaluation using national linked health data

BACKGROUND

Scotland's Baby Box Scheme (SBBS) is a national programme offering a box of essential items to all pregnant women in Scotland intended to improve infant and maternal health. We aimed to evaluate the effect of SBBS on selected infant and maternal health outcomes at population and subgroup levels (maternal age and area deprivation).

METHODS

Our complete-case, intention-to-treat evaluation used national health data (from the Scottish Morbidity Record [SMR] 01, SMR02, and the Child Health Surveillance Programme-Pre School), linking birth records to postnatal hospitalisation and universal health visitor records in Scotland. We considered maternal-infant pairs of all live-singleton births 2 years either side of SBBS introduction (Aug 17, 2015, to Aug 11, 2019). We estimated step-changes and trend-changes in outcomes (hospital admission and self-reported exclusive breastfeeding, tobacco smoke exposure, and infant sleeping position) by week of birth using segmented Poisson regression, adjusting for over-dispersion and seasonality where necessary.

FINDINGS

The analysis comprised 182 122 maternal-infant pairs. The prevalence of tobacco smoke exposure reduced after SBBS introduction: step decrease of 10% (prevalence ratio 0·904 [95% CI 0·865-0·946]; absolute decrease of 1·6% 1 month post-introduction) for infants and 9% (0·905 [0·862-0·950]; absolute decrease of 1·9% 1 month post-introduction) for the primary carer. There was no evidence of changes in infant and maternal all-cause hospital admissions or infant sleeping position. Among mothers younger than 25 years, there was a 10% step-increase in breastfeeding prevalence (1·095 [1·004-1·195]; absolute increase of 2·2% 1 month post-introduction) at 10 days and 17% (1·174 [1·037-1·328]) at 6-8 weeks postnatal. Although associations were robust to most sensitivity analyses, for smoke exposure associations were only observed early in the postnatal period.

INTERPRETATION

SBBS reduced infant and primary carer tobacco smoke exposure, and increased breastfeeding among young mothers in Scotland. However, absolute effects were small.

FUNDING

Medical Research Council, Scottish Government Chief Scientist Office, and National Records of Scotland.

The relationship between publication of high-quality evidence and changes in the volume and trend of subacromial decompression surgery for patients with subacromial pain syndrome in hospitals across Australia, Europe and the United States: a controlled interrupted time series analysis

AIMS

To evaluate the extent to which publication of high-quality randomised controlled trials(RCTs) in 2018 was associated with a change in volume or trend of subacromial decompression(SAD) surgery in patients with subacromial pain syndrome(SAPS) treated in hospitals across various countries.

METHODS

Routinely collected administrative data of the Global Health Data@work collaborative were used to identify SAPS patients who underwent SAD surgery in six hospitals from five countries (Australia, Belgium, Netherlands, United Kingdom, United States) between 01/2016 and 02/2020. Following a controlled interrupted time series design, segmented Poisson regression was used to compare trends in monthly SAD surgeries before(01/2016-01/2018) and after(02/2018-02/2020) publication of the RCTs. The control group consisted of musculoskeletal patients undergoing other procedures.

RESULTS

A total of 3.046 SAD surgeries were performed among SAPS patients treated in five hospitals; one hospital did not perform any SAD surgeries. Overall, publication of trial results was associated with a significant reduction in the trend to use SAD surgery of 2% per month (Incidence rate ratio (IRR) 0.984[0.971-0.998]; P = 0.021), but with large variation between hospitals. No changes in the control group were observed. However, publication of trial results was also associated with a 2% monthly increased trend (IRR 1.019[1.004-1.034]; P = 0.014) towards other procedures performed in SAPS patients.

CONCLUSION

Publication of RCT results was associated with a significantly decreased trend in SAD surgery for SAPS patients, although large variation between participating hospitals existed and a possible shift in coding practices cannot be ruled out. This highlights the complexities of implementing recommendations to change routine clinical practice even if based on high-quality evidence.

Causal effects of mobility intervention policies on intracity flows during the COVID-19 pandemic: The moderating role of zonal locations in the transportation networks

Many studies have investigated the impact of mobility restriction policies on the change of intercity flows during the outbreak of COVID-19, whereas only a few have highlighted intracity flows. By using the mobile phone trajectory data of approximately three months, we develop an interrupted time series quasi-experimental design to estimate the abrupt and gradual effects of mobility intervention policies during the pandemic on intracity flows of 491 neighborhoods in Shenzhen, China, with a focus on the role of urban transport networks. The results show that the highest level of public health emergency response caused an abrupt decline by 4567 trips and a gradually increasing effect by 34 trips per day. The effectiveness of the second return-to-work order (RtW2) was found to be clearly larger than that of the first return-to-work order (RtW1) as a mobility restoration strategy. The causal effects of mobility intervention policies are heterogenous across zonal locations in varying urban transport networks. The declining effect of health emergency response and rebounding effect of RtW2 are considerably large in better-connected neighborhoods with metro transit, as well as in those close to the airport. These findings provide new insights into the identification of pandemic-vulnerable hotspots in the transport network inside the city, as well as of crucial neighborhoods with increased adaptability to mobility interventions during the onset and decline of COVID-19.

Pharmacists' chronic disease management in chronic obstructive pulmonary disease: Effect on health services utilization

BACKGROUND: There is limited real-world evidence on evaluation of chronic disease management initiatives provided by pharmacists to patients with chronic obstructive pulmonary disease (COPD). OBJECTIVE: To evaluate changes in COPD-related health care resource utilization between patients with COPD who had pharmacist-provided chronic disease management (comprehensive annual care plan [CACP]) vs those who did not have CACP. METHODS: Patients with COPD who received a CACP in Alberta between 2012 and 2015 were identified within the Alberta Health administrative data. Each of these patients were matched with 2 control patients with COPD based on age, sex, provider, date of service, and qualifying comorbidities. Controlled interrupted time series analysis was used to evaluate changes in COPD-specific hospitalizations, emergency department (ED) visits, physician visits, and claims for pulmonary function test. Immediate and temporal changes were calculated for the difference in outcomes 1 year before and 1 year after receiving the CACP for the intervention group and matched controls. RESULTS: Eligible patients (N = 74,365), of whom 28,795 (38.7%) had received CACPs, were matched to a total of 45,570 controls. In 1 year after the CACPs implementation, the number of COPD-related hospitalization visits decreased by 174 (95% CI = -270.8 to -76.5) per 10,000 patients per month, COPD-related ED visits decreased by 123 (95% CI = -294.9 to 49.6) per 10,000 per month, general practitioner visits decreased by 153.9 per 10,000 per month (95% CI = -293.3 to -14.5), and pulmonary function test claims decreased by 19.5 per 10,000 per month (95% CI = -70.1 to 31.2) when compared with the matched controls. However, significant difference between the 2 groups was found for COPD-related hospitalizations only, which was not confirmed by the sensitivity analysis. CONCLUSIONS: In patients with COPD who were provided with care plans by their community pharmacists, there was no significant decrease in COPD-related hospitalizations or ED visits over 1 year compared with the matched controls who did not have a pharmacist-provided care plan. Physician visits and pulmonary function tests did not change significantly for those who had CACP compared with those who did not. There is a need to further understand how care plans can better impact other outcomes that are important in COPD management. DISCLOSURES: This study was supported by a grant from the M.S.I. Foundation (Grant#895) based in Alberta, Canada. Dr Bhutani has consulted for Astra Zeneca, GlaxoSmithKline, Boehringer Ingelheim, Valeo, Covis, and Sanofi. The authors declare no other relevant conflicts of interest or financial relationships. This study is based on data provided by Alberta Health. The interpretation and conclusions of the results are those of the researchers and do not necessarily represent the views of the government of Alberta nor the funder (M.S.I. Foundation). All authors meet criteria for authorship as recommended by the International Committee of Medical Journal Editors.

Impacts of unifying urban and rural residents' medical insurance on the hospitalisation expenses of rural patients in eastern China: an interrupted time series analysis

OBJECTIVES

This study evaluated the impact of the Urban and Rural Residents' Basic Medical Insurance scheme on hospitalisation expenses of rural patients in eastern China, which unified separate healthcare systems for urban and rural residents.

DESIGN

Monthly hospitalisation data from municipal and county hospitals were collected from the local Medicare Fund Database, covering the period from January 2018 to December 2021. The unification of insurance between urban and rural patients was implemented at different times for county and municipal hospitals. An interrupted time series analysis was used to assess the immediate and gradual effects of the integrated policy on the total medical expenses, out-of-pocket (OOP) expenses and effective reimbursement rate (ERR) among rural patients.

SETTING AND PARTICIPANTS

This study included 636 155 rural inpatients over 4 years in Xuzhou City, Jiangsu Province, China.

RESULTS

In January 2020, the policy of urban and rural medical insurance was initially integrated in county hospitals, after which the ERR decreased at a monthly rate of 0.23% (p=0.002, 95% CI -0.37% to -0.09%) compared with the preintervention period. After the insurance systems were unified in municipal hospitals in January 2021, OOP expenses decreased by ¥63.54 (p=0.002, 95% CI -102.48 to -24.61) and the ERR increased at a monthly rate of 0.24% (p=0.029, 95% CI 0.03% to 0.045%).

CONCLUSIONS

Our results suggest that the unification of urban and rural medical insurance systems was an effective intervention to reduce the financial burden of illness for rural inpatients, especially OOP expenses for hospitalisation in municipal hospitals.

The Impact of COVID-19 on Mental Healthcare Utilization in Switzerland Was Strongest Among Young Females-Retrospective Study in 2018-2020

Objectives: To provide a thorough assessment of the impact of the COVID-19 pandemic on the utilization of inpatient and outpatient mental healthcare in Switzerland. Methods: Retrospective cohort study using nationwide hospital data (n > 8 million) and claims data from a large Swiss health insurer (n > 1 million) in 2018-2020. Incidence proportions of different types of psychiatric inpatient admissions, psychiatric consultations, and psychotropic medication claims were analyzed using interrupted time series models for the general population and for the vulnerable subgroup of young people. Results: Inpatient psychiatric admissions in the general population decreased by 16.2% (95% confidence interval: -19.2% to -13.2%) during the first and by 3.9% (-6.7% to -0.2%) during the second pandemic shutdown, whereas outpatient mental healthcare utilization was not substantially affected. We observed distinct patterns for young people, most strikingly, an increase in mental healthcare utilization among females aged <20 years. Conclusion: Mental healthcare provision for the majority of the population was largely maintained, but special attention should be paid to young people. Our findings highlight the importance of monitoring mental healthcare utilization among different populations.

Provision of Postpartum Contraception Before and After the Start of the COVID-19 Pandemic in Maine

OBJECTIVE

Preliminary findings from selected health systems revealed interruptions in reproductive health care services due to the COVID-19 pandemic. We estimated changes in postpartum contraceptive provision associated with the start of the COVID-19 pandemic in Maine.

METHODS

We used the Maine Health Data Organization's All Payer Claims Database for deliveries from October 2015 through March 2021 (n = 45 916). Using an interrupted time-series analysis design, we estimated changes in provision rates of long-acting reversible contraception (LARC), permanent contraception, and moderately effective contraception within 3 and 60 days of delivery after the start of the COVID-19 pandemic. We performed 6- and 12-month analyses (April 2020-September 2020, April 2020-March 2021) as compared with the reference period (October 2015-March 2020). We used Poisson regression models to calculate level-change rate ratios (RRs) and 95% CIs.

RESULTS

The 6-month analysis found that provision of LARC (RR = 1.89; 95% CI, 1.76-2.02) and moderately effective contraception (RR = 1.51; 95% CI, 1.33-1.72) within 3 days of delivery increased at the start of the COVID-19 pandemic, while provision of LARC (RR = 0.95; 95% CI, 0.93-0.97) and moderately effective contraception (RR = 1.08; 95% CI, 1.05-1.11) within 60 days of delivery was stable. Rates of provision of permanent contraception within 3 days (RR = 0.70; 95% CI, 0.63-0.78) and 60 days (RR = 0.71; 95% CI, 0.63-0.80) decreased. RRs from the 12-month analysis were generally attenuated.

CONCLUSION

Disruptions in postpartum provision of permanent contraception occurred at the beginning of the COVID-19 pandemic in Maine. Public health policies should include guidance for contraceptive provision during public health emergencies and consider designating permanent contraception as a nonelective procedure.

A quasi-experimental analysis on the causal effects of COVID-19 on urban park visits: The role of park features and the surrounding built environment

Although many studies have explored the correlations between mobility intervention policies and park use during COVID-19, only a few have used causal inference approaches to assessing the policy's treatment effects and how such effects vary across park features and surrounding built environments. In this study, we develop an interrupted time-series quasi-experimental design based on three-month mobile phone big data to infer the causal effects of mobility intervention policies on park visits in Shenzhen, including the first-level response (FLR) and return-to-work (RTW) order. The results show that the FLR caused an abrupt decline of 2.21 daily visits per park, with a gradual reduction rate of 0.54 per day, whereas the RTW order helped recover park visits with an immediate increase of 2.20 daily visits and a gradual growth rate of 0.94 visits per day. The results also show that the impact of COVID-19 on park visits exhibited social and spatial heterogeneities: the mobility-reduction effect was smaller in low-level parks (e.g., community-level parks) with small sizes but without sports facilities and water scenes, whereas parks surrounded by compact neighborhoods and land use were more impacted by the pandemic. These findings provide planners with important insights into resilient green space and sustainable neighborhood planning for the post-COVID era.

Mortality before and after reconfiguration of the Danish hospital-based emergency healthcare system: a nationwide interrupted time series analysis

OBJECTIVES

The study aimed to investigate how the 'natural experiment' of reconfiguring the emergency healthcare system in Denmark affected in-hospital and 30-day mortality on a national level. The reconfiguration included the centralisation of hospitals and the establishment of emergency departments with specialists present around the clock.

DESIGN

Hospital-based cohort study.

SETTING

All public hospitals in Denmark.

PARTICIPANTS

Patients with an unplanned contact from 1 January 2007 until 31 December 2016.

INTERVENTIONS

Stepped-wedge reconfiguration of the Danish emergency healthcare system.

MAIN OUTCOME MEASURES

We determined the adjusted ORs for in-hospital mortality and HRs for 30-day mortality using logistic and Cox regression analysis adjusted for sex, age, Charlson Comorbidity Index, income, education, mandatory referral and the changes in the out of hours system in the Capital Region. The main outcomes were stratified by the time of arrival. We performed subgroup analyses on selected diagnoses: myocardial infarction, stroke, pneumonia, aortic aneurysm, bowel perforation, hip fracture and major trauma.

RESULTS

We included 11 367 655 unplanned hospital contacts. The adjusted OR for overall in-hospital mortality after reconfiguration of the emergency healthcare system was 0.998 (95% CI 0.968 to 1.010; p=0.285), and the adjusted OR for 30-day mortality was 1.004 (95% CI 1.000 to 1.008; p=0.045)). Subgroup analyses showed some possible benefits of the reconfiguration such as a reduction in-hospital and 30-day mortality for myocardial infarction, stroke, aortic aneurysm and major trauma.

CONCLUSIONS

The Danish emergency care reconfiguration programme was not associated with an improvement in overall in-hospital mortality trends and was associated with a slight slowing of prior improvements in 30-day mortality trends.

The possible effect of sociobehavioral factors and public health actions on the mpox epidemic slowdown

OBJECTIVES

A pre-exposure vaccination campaign to prevent the spread of the mpox virus was initiated in Italy in August 2022. We explore the possible factors affecting the trend of mpox cases in an Italian region (Lazio) with a rapid roll-out of the vaccination campaign.

METHODS

We estimated the impact of the communication and vaccination campaign by fitting a Poisson segmented regression model. Results By September 30, 2692, high-risk men who have sex with men had received at least one dose of vaccine, with a vaccination coverage of 37%. The analysis of surveillance data showed a significant decreasing trend in the number of mpox cases starting from the second week after vaccination (incidence rate ratio 0.452 [0.331-0.618]).

CONCLUSION

The reported trend in mpox cases is likely to result from a combination of multiple social and public health factors combined with a vaccination campaign.

Quasi-experimental evaluation of national border closures on COVID-19 transmission

With over 200 pandemic threats emerging every year, the efficacy of closing national borders to control the transmission of disease in the first months of a pandemic remains a critically important question. Previous studies offer conflicting evidence for the potential effects of these closures on COVID-19 transmission and no study has yet empirically evaluated the global impact of border closures using quasi-experimental methods and real-world data. We triangulate results from interrupted time-series analysis, meta-regression, coarsened exact matching, and an extensive series of robustness checks to evaluate the effect of 166 countries' national border closures on the global transmission of COVID-19. Total border closures banning non-essential travel from all countries and (to a lesser extent) targeted border closures banning travel from specific countries had some effect on temporarily slowing COVID-19 transmission in those countries that implemented them. In contrast to these country-level impacts, the global sum of targeted border closures implemented by February 5, 2020 was not sufficient to slow global COVID-19 transmission, but the sum of total border closures implemented by March 19, 2020 did achieve this effect. Country-level results were highly heterogeneous, with early implementation and border closures so broadly targeted that they resemble total border closures improving the likelihood of slowing the pandemic's spread. Governments that can make productive use of extra preparation time and cannot feasibly implement less restrictive alternatives might consider enacting border closures. However, given their moderate and uncertain impacts and their significant harms, border closures are unlikely to be the best policy response for most countries and should only be deployed in rare circumstances and with great caution. All countries would benefit from global mechanisms to coordinate national decisions on border closures during pandemics.

An Interrupted Time Series Analysis Measuring the Impact of Research and Education on Clinical Practice: Decreasing Allograft Use in Young Patients Using a Registry to Track Outcomes

BACKGROUND

We sought to evaluate whether allograft utilization for primary anterior cruciate ligament reconstruction (ACLR) within our health-care system changed following the implementation of an allograft reduction intervention and whether revision rates within the health-care system changed following the initiation of the intervention.

METHODS

We conducted an interrupted time series study using data from Kaiser Permanente's ACL Reconstruction Registry. In our study, we identified 11,808 patients who were ≤21 years of age and underwent primary ACLR from January 1, 2007, through December 31, 2017. The pre-intervention period (15 quarters) was January 1, 2007, through September 30, 2010, and the post-intervention period (29 quarters) was October 1, 2010, through December 31, 2017. Poisson regression was used to evaluate trends over time in 2-year revision rates according to the quarter in which the primary ACLR was performed.

RESULTS

Allograft utilization increased pre-intervention from 21.0% in 2007 Q1 to 24.8% in 2010 Q3. Utilization decreased post-intervention from 29.7% in 2010 Q4 to 2.4% in 2017 Q4. The quarterly 2-year revision rate increased from 3.0 to 7.4 revisions per 100 ACLRs pre-intervention and decreased to 4.1 revisions per 100 ACLRs by the end of the post-intervention period. Poisson regression found an increasing 2-year revision rate over time pre-intervention (rate ratio [RR], 1.03 [95% confidence interval (CI), 1.00 to 1.06] per quarter) and a decreasing rate over time post-intervention (RR, 0.96 [95% CI, 0.92 to 0.99]).

CONCLUSIONS

In our health-care system, we saw a decrease in allograft utilization following the implementation of an allograft reduction program. During the same period, a decrease in the ACLR revision rate was observed.

LEVEL OF EVIDENCE

Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.

Increase in traffic injury crashes following the 2016 Kumamoto earthquake in Japan: A model comparison

OBJECTIVE

Following natural disasters, the number of motor vehicle crashes may increase as drivers are often forced to drive under stressful conditions. This study aims to analyze the changes in motor vehicle crashes that resulted in injury or death (injury crash) following the 2016 Kumamoto earthquake in Japan. An existing study reported that the increased crashes resulted in property damage following the earthquake; however, the effects on injury crashes remain unreported.

METHODS

Interrupted time series analysis is employed to investigate the changes in injury crashes following the earthquake. The results are compared based on several time series models, including negative binomial and autoregressive integrated moving average models. Monthly injury-crash data from 2011 to 2020 in Kumamoto and Fukuoka city is used.

RESULTS

The results reveal a 1,642-count or 20% increase (1.20-times increase, 95% confidence interval: 1.12, 1.27) in injury crashes due to the earthquake in Kumamoto city, where the earthquake damage was heavy. In contrast, statistically significant change is not detected in Fukuoka city, where the earthquake damage is negligible.

CONCLUSION

The results indicate that the earthquake has increased the motor-vehicle-crash risk and that traffic crash alerts are important following disasters.

Implementation of a quality improvement programme using the Active Patient Link call and recall system to improve timeliness and equity of childhood vaccinations: protocol for a mixed-methods evaluation

INTRODUCTION

Call and recall systems provide actionable intelligence to improve equity and timeliness of childhood vaccinations, which have been disrupted during the COVID-19 pandemic. We will evaluate the effectiveness, fidelity and sustainability of a data-enabled quality improvement programme delivered in primary care using an Active Patient Link Immunisation (APL-Imms) call and recall system to improve timeliness and equity of uptake in a multiethnic disadvantaged urban population. We will use qualitative methods to evaluate programme delivery, focusing on uptake and use, implementation barriers and service improvements for clinical and non-clinical primary care staff, its fidelity and sustainability.

METHODS AND ANALYSIS

This is a mixed-methods observational study in 284 general practices in north east London (NEL). The target population will be preschool-aged children eligible to receive diphtheria, tetanus and pertussis (DTaP) or measles, mumps and rubella (MMR) vaccinations and registered with an NEL general practice. The intervention comprises an in-practice call and recall tool, facilitation and training, and financial incentives. The quantitative evaluation will include interrupted time Series analyses and Slope Index of Inequality. The primary outcomes will be the proportion of children receiving at least one dose of a DTaP-containing or MMR vaccination defined, respectively, as administered between age 6 weeks and 6 months or between 12 and 18 months of age. The qualitative evaluation will involve a 'Think Aloud' method and semistructured interviews of stakeholders to assess impact, fidelity and sustainability of the APL-Imms tool, and fidelity of the implementation by facilitators.

ETHICS AND DISSEMINATION

The research team has been granted permission from data controllers in participating practices to use deidentified data for audit purposes. As findings will be specific to the local context, research ethics approval is not required. Results will be disseminated in a peer-reviewed journal and to stakeholders, including parents, health providers and commissioners.

Mental health outcomes in parents of children with a cancer diagnosis in Sweden: A nationwide cohort study

BACKGROUND

The diagnosis of paediatric cancer is a crisis for the parents who are the primary caregivers of the affected child. A comprehensive assessment of the longitudinal impact of childhood cancer on parental mental health and the potential sex differences between the parents is lacking. Thus, we aimed to explore the subsequent short- and long-term mental health outcomes among the parents of children with cancer and examine whether the outcomes vary between the mother and father.

METHODS

By combining several Swedish registers, parents of a child (ages 0-14 years) with a cancer diagnosis between Jan 1, 2006, and Dec 31, 2016 were identified. For each parent of children with cancer, up to five mothers or fathers of cancer-free children were randomly selected and matched, respectively. Hospital contacts for any mental health disorders between 5 years before and 7 years after the diagnosis of childhood cancer were retrieved. An interrupted time series negative binomial regression was performed to assess the short- and long-term impact of a childhood cancer diagnosis on the parents' subsequent mental health outcomes.

FINDINGS

16,199 mothers (2852 with a child with cancer and 13,347 without) and 15,708 fathers (2769 with a child with cancer and 12,939 without) were included in this study. Compared with mothers of children without cancer, mothers of children with cancer had higher risks of mental health disorders in the first year after diagnosis (rate ratio [RR] and 95% Confidence Interval [CI], 1.17 (1.03-1.32)), and notably, the adverse impact became more severe over time (RR and 95% CI, 1.36 (1.07-1.74), in the seventh year). For fathers of children with cancer, the risk of mental health disorders was continuously higher compared to matched comparisons (RR and 95% CI, 1.31 (1.01-1.71)).

INTERPRETATION

Our findings suggested that parental mental health was affected continuously by a diagnosis of childhood cancer in their children. In particular, the mother's mental health was affected more severely. Customised psychological services or interventions are highly needed for the parents of children with cancer.

FUNDING

Swedish Research Council, Allmänna Sjukhusets i Malmö Stiftelsen för bekämpande av cancer, Swedish Heart-Lung Foundation, ALF funding from Region Skåne and China Scholarship Council.

A generalized interrupted time series model for assessing complex health care interventions

Assessing the impact of complex interventions on measurable health outcomes is a growing concern in health care and health policy. Interrupted time series (ITS) designs borrow from traditional case-crossover designs and function as quasi-experimental methodology able to retrospectively analyze the impact of an intervention. Statistical models used to analyze ITS designs primarily focus on continuous-valued outcomes. We propose the "Generalized Robust ITS" (GRITS) model appropriate for outcomes whose underlying distribution belongs to the exponential family of distributions, thereby expanding the available methodology to adequately model binary and count responses. GRITS formally implements a test for the existence of a change point in discrete ITS. The methodology proposed is able to test for the existence of and estimate the change point, borrow information across units in multi-unit settings, and test for differences in the mean function and correlation pre- and post-intervention. The methodology is illustrated by analyzing patient falls from a hospital that implemented and evaluated a new care delivery model in multiple units.

Affordable Care Act (ACA) Implementation and Adolescent Births by Insurance Type: An Interrupted Time Series Analysis of Births between 2009 and 2017 in the United States

BACKGROUND

In 2010, the Affordable Care Act (ACA) was enacted, with full provisions in effect by 2014, including expanded Medicaid coverage, changes to the marketplace, and contraceptive coverage, but its impact on birth trends, particularly adolescent births, is currently unknown.

OBJECTIVES

We sought to determine whether ACA implementation was associated with changes in adolescent births and whether this differed by insurance type (Medicaid or private insurance).

METHODS

We used revised 2009-2017 birth certificate data, restricted to resident women with a Medicaid or privately paid singleton birth (N = 27,748,028). Segmented regression analysis was used to examine births to adolescent mothers (12-19 years old) before and after the ACA.

RESULTS

There were 27,748,028 singleton births (n = 2,013,521 adolescent births) among U.S. residents between 2009 and 2017 in this analytic sample. Adjusted models revealed that the ACA was associated with a 23% significant decrease in odds of an adolescent birth (OR = 0.78; 95% CI, 0.77-0.79) for Medicaid-funded births and a 19% decrease (OR = 0.81; 95% CI, 0.79-0.83) for privately insured births, with a further declining trend. Overall declines in adolescent births among the Medicaid population appear to be driven by states that chose to expand Medicaid.

CONCLUSION

Beyond the declining secular trend already observed in adolescent pregnancy over the last 10 years, the ACA appears to have had a substantial impact on adolescent births, likely due to Medicaid expansion and increased access to affordable contraception. From a population health perspective, efforts to undo the ACA could have important consequences for maternal, infant, and family health in the United States.

The impact of COVID-19 non-pharmaceutical interventions on HIV care continuum in China: An interrupted time series analysis

BACKGROUND

China implemented strict non-pharmaceutical interventions to contain COVID-19 at the early stage. We aimed to evaluate the impact of COVID-19 on HIV care continuum in China.

METHODS

Aggregated data on HIV care continuum between 1 January 2017 and 31 December 2020 were collected from centers for disease control and prevention at different levels and major infectious disease hospitals in various regions in China. We used interrupted time series analysis to characterize temporal trend in weekly numbers of HIV post-exposure prophylaxis (PEP) prescriptions, HIV tests, HIV diagnoses, median time intervals between HIV diagnosis and antiretroviral therapy (ART) initiation (time intervals, days), ART initiations, mean CD4+ T cell counts at ART initiation (CD4 counts, cells/μL), ART collections, and missed visits for ART collection, before and after the implementation of massive NPIs (23 January to 7 April 2020). We used Poisson segmented regression models to estimate the immediate and long-term impact of NPIs on these outcomes.

FINDINGS

A total of 16,780 PEP prescriptions, 1,101,686 HIV tests, 69,659 HIV diagnoses, 63,409 time intervals and ART initiations, 61,518 CD4 counts, 1,528,802 ART collections, and 6656 missed visits were recorded during the study period. The majority of outcomes occurred in males (55·3-87·4%), 21-50 year olds (51·7-90·5%), Southwestern China (38·2-82·0%) and heterosexual transmission (47·9-66·1%). NPIs was associated with 71·5% decrease in PEP prescriptions (IRR 0·285; 95% CI 0·192-0·423), 36·1% decrease in HIV tests (0·639, 0·497-0·822), 32·0% decrease in HIV diagnoses (0·680, 0·511-0·904), 59·3% increase in time intervals (1·593, 1·270-1·997) and 17·4% decrease in CD4 counts (0·826, 0·746-0·915) in the first week during NPIs. There was no marked change in the number of ART initiations, ART collections and missed visits during the NPIs. By the end of 2020, the number of HIV tests, HIV diagnoses, time intervals, ART initiations, and CD4 counts reached expected levels, but the number of PEP prescriptions (0·523, 0·394-0·696), ART collections (0·720, 0·595-0·872), and missed visits (0·137, 0·086-0·220) were still below expected levels. With the ease of restrictions, PEP prescriptions (slope change 1·024/week, 1·012-1·037), HIV tests (1·016/week, 1·008-1·026), and CD4 counts (1·005/week, 1·001-1·009) showed a significant increasing trend.

INTERPRETATION

HIV care continuum in China was affected by the COVID-19 NPIs at various levels. Preparedness and efforts to maintain the HIV care continuum during public health emergencies should leverage collaborations between stakeholders.

FUNDING

Natural Science Foundation of China.

Structured evaluation of a virtual emergency department triage model of care: A study protocol

OBJECTIVE

A new virtual ED service was introduced into a hospital network in the northern suburbs of Melbourne in response to changing needs during the COVID-19 pandemic. The 'virtual ED' utilises a telehealth model as a means of assessment for appropriately selected patients to facilitate either complete care or navigation into streamlined pathways for ongoing care, in some cases bypassing the ED entirely where appropriate. The proposed study aims to evaluate the implementation of the model and identify future improvement opportunities, assess the impact on traditional health service delivery processes and patient experience, and determine the acceptability of the 'virtual ED' model of care.

METHODS

The present study will consist of a pre-post- implementation evaluation using the RE-AIM framework. Routine health service data will be collected for 6 months post-implementation of the virtual ED model and compared to 24 months prior to implementation. Prospective data will be collected using routinely collected and survey data. Interviews and focus groups will be conducted to understand consumer and clinician perspectives on barriers and enablers to implementation and adoption of the virtual ED.

RESULTS

Descriptive statistics will be used to describe the study population and key outcomes, including changes in ED presentations and length of stay. Thematic analysis will be conducted on transcribed interviews and focus group data. This will be triangulated with data collected from patient feedback surveys.

CONCLUSION

This project will support the delivery of care to ED patients by evaluating the 'virtual ED' model of care.