Impact of STROBE Statement Publication on Quality of Observational Study Reporting: Interrupted Time Series versus Before-After Analysis

Background

In uncontrolled before-after studies, CONSORT was shown to improve the reporting of randomised trials. Before-after studies ignore underlying secular trends and may overestimate the impact of interventions. Our aim was to assess the impact of the 2007 STROBE statement publication on the quality of observational study reporting, using both uncontrolled before-after analyses and interrupted time series.

Methods

For this quasi-experimental study, original articles reporting cohort, case-control, and cross-sectional studies published between 2004 and 2010 in the four dermatological journals having the highest 5-year impact factors (≥4) were selected. We compared the proportions of STROBE items (STROBE score) adequately reported in each article during three periods, two pre STROBE period (2004–2005 and 2006–2007) and one post STROBE period (2008–2010). Segmented regression analysis of interrupted time series was also performed.

Results

Of the 456 included articles, 187 (41%) reported cohort studies, 166 (36.4%) cross-sectional studies, and 103 (22.6%) case-control studies. The median STROBE score was 57% (range, 18%–98%). Before-after analysis evidenced significant STROBE score increases between the two pre-STROBE periods and between the earliest pre-STROBE period and the post-STROBE period (median score_2004–05 48% versus median score_2008–10 58%, p<0.001) but not between the immediate pre-STROBE period and the post-STROBE period (median score_2006–07 58% versus median score_2008–10 58%, p = 0.42). In the pre STROBE period, the six-monthly mean STROBE score increased significantly, by 1.19% per six-month period (absolute increase 95%CI, 0.26% to 2.11%, p = 0.016). By segmented analysis, no significant changes in STROBE score trends occurred (−0.40%; 95%CI, −2.20 to 1.41; p = 0.64) in the post STROBE statement publication.

Interpretation

The quality of reports increased over time but was not affected by STROBE. Our findings raise concerns about the relevance of uncontrolled before-after analysis for estimating the impact of guidelines.     

Impact of Pay for Performance on Prescribing of Long-Acting Reversible Contraception in Primary Care: An Interrupted Time Series Study

Background

The aim of this study was to evaluate the impact of Quality and Outcomes Framework (QOF), a major pay-for-performance programme in the United Kingdom, on prescribing of long-acting reversible contraceptives (LARC) in primary care.

Methods

Negative binomial interrupted time series analysis using practice level prescribing data from April 2007 to March 2012. The main outcome measure was the prescribing rate of long-acting reversible contraceptives (LARC), including hormonal and non hormonal intrauterine devices and systems (IUDs and IUSs), injectable contraceptives and hormonal implants.

Results

Prescribing rates of Long-Acting Reversible Contraception (LARC) were stable before the introduction of contraceptive targets to the QOF and increased afterwards by 4% annually (rate ratios  = 1.04, 95% CI = 1.03, 1.06). The increase in LARC prescribing was mainly driven by increases in injectables (increased by 6% annually), which was the most commonly prescribed LARC method. Of other types of LARC, the QOF indicator was associated with a step increase of 20% in implant prescribing (RR =  1.20, 95% CI =  1.09, 1.32). This change is equivalent to an additional 110 thousand women being prescribed with LARC had QOF points not been introduced.

Conclusions

Pay for performance incentives for contraceptive counselling in primary care with women seeking contraceptive advice has increased uptake of LARC methods.     

Impact of a COPD Discharge Care Bundle on Readmissions following Admission with Acute Exacerbation: Interrupted Time Series Analysis

ObjectivesWe evaluated the impact of a COPD discharge care bundle on readmission rates following hospitalisation with an acute exacerbation.DesignInterrupted time series analysis, comparing readmission rates for COPD exacerbations at nine trusts that introduced the bundle, to two comparison groups; (1) other NHS trusts in London and (2) all other NHS trusts in England. Care bundles were implemented at different times for different NHS trusts, ranging from October 2009 to April 2011.SettingNine NHS acute trusts in the London, England.ParticipantsPatients aged 45 years and older admitted to an NHS acute hospital in England for acute exacerbation of COPD. Data come from Hospital Episode Statistics, April 2002 to March 2012.ResultsIn hospitals introducing the bundle readmission rates were rising before implementation and falling afterwards (e.g. readmissions within 28 days +2.13% per annum (pa) pre and -5.32% pa post (p for difference in trends = 0.012)). Following implementation, readmission rates within 7 and 28 day were falling faster than among other trusts in London, although this was not statistically significant (e.g. readmissions within 28 days -4.6% pa vs. -3.2% pa, p = 0.44). Comparisons with a national control group were similar.ConclusionsThe COPD discharge care bundle appeared to be associated with a reduction in readmission rate among hospitals using it. The significance of this is unclear because of changes to background trends in London and nationally.     

Health Facility Utilisation Changes during the Introduction of Community Case Management of Malaria in South Western Uganda: An Interrupted Time Series Approach

Background

Malaria endemic countries have scaled-up community health worker (CHW) interventions, to diagnose and treat malaria in communities with limited access to public health systems. The evaluations of these programmes have centred on CHW’s compliance to guidelines, but the broader changes at public health centres including utilisation and diagnoses made, has received limited attention.

Methods

This analysis was conducted during a CHW–intervention for malaria in Rukungiri District, Western Uganda. Outpatient department (OPD) visit data were collected for children under-5 attending three health centres one year before the CHW-intervention started (pre-intervention period) and for 20 months during the intervention (intervention-period). An interrupted time series analysis with segmented regression models was used to compare the trends in malaria, non-malaria and overall OPD visits during the pre-intervention and intervention-period.

Results

The introduction of a CHW-intervention suggested the frequency of diagnoses of diarrhoeal diseases, pneumonia and helminths increased, whilst the frequency of malaria diagnoses declined at health centres. In May 2010 when the intervention began, overall health centre utilisation decreased by 63% compared to the pre-intervention period and the health centres saw 32 fewer overall visits per month compared to the pre-intervention period (p<0.001). Malaria visits also declined shortly after the intervention began and there were 27 fewer visits per month during the intervention-period compared with the pre-intervention period (p<0.05). The declines in overall and malaria visits were sustained for the entire intervention-period. In contrast, there were no observable changes in trends of non-malarial visits between the pre-intervention and intervention-period.

Conclusions

This analysis suggests introducing a CHW-intervention can reduce the number of child malaria visits and change the profile of cases presenting at health centres. The reduction in workload of health workers may allow them to spend more time with patients or undertake additional curative or preventative roles.     

Evaluating the Impact of Abrupt Changes in Forest Policy and Management Practices on Landscape Dynamics: Analysis of a Landsat Image Time Series in the Atlantic Northern Forest

Sustainable forest management is based on functional relationships between management actions, landscape conditions, and forest values. Changes in management practices make it fundamentally more difficult to study these relationships because the impacts of current practices are difficult to disentangle from the persistent influences of past practices. Within the Atlantic Northern Forest of Maine, U.S.A., forest policy and management practices changed abruptly in the early 1990s. During the 1970s-1980s, a severe insect outbreak stimulated salvage clearcutting of large contiguous tracts of spruce-fir forest. Following clearcut regulation in 1991, management practices shifted abruptly to near complete dependence on partial harvesting. Using a time series of Landsat satellite imagery (1973-2010) we assessed cumulative landscape change caused by these very different management regimes. We modeled predominant temporal patterns of harvesting and segmented a large study area into groups of landscape units with similar harvest histories. Time series of landscape composition and configuration metrics averaged within groups revealed differences in landscape dynamics caused by differences in management history. In some groups (24% of landscape units), salvage caused rapid loss and subdivision of intact mature forest. Persistent landscape change was created by large salvage clearcuts (often averaging > 100 ha) and conversion of spruce-fir to deciduous and mixed forest. In groups that were little affected by salvage (56% of landscape units), contemporary partial harvesting caused loss and subdivision of intact mature forest at even greater rates. Patch shape complexity and edge density reached high levels even where cumulative harvest area was relatively low. Contemporary practices introduced more numerous and much smaller patches of stand-replacing disturbance (typically averaging <15 ha) and a correspondingly large amount of edge. Management regimes impacted different areas to different degrees, producing different trajectories of landscape change that should be recognized when studying the impact of policy and management practices on forest ecology.     

Focus: Preventive Medicine: The Impact of the Social Distancing Policy on COVID-19 Incidence Cases and Deaths in Iran from February 2020 to January 2021: Insights from an Interrupted Time Series Analysis

Background: In December 2019, a viral outbreak occurred in China, and rapidly spread out worldwide. Due to the lack of immediately available vaccines and effective drugs, many policy- and decision-makers have focused on non-pharmacological methods, including social distancing. This study was aimed at assessing the effects of the implementation of this policy in Iran, one of the countries most affected by COVID-19. We conducted a quasi-experimental study, utilizing the interrupted time series analysis (ITSA) approach. Methods: We collected daily data between February 20, 2020 and January 29, 2021, through governmental websites from 954 public hospitals and healthcare settings. The Iranian government launched the social distancing policy on March 27, 2020. Statistical analyses, including ITSA, were carried out with R software Version 3.6.1 (London, UK). Results: During the study period, 1,398,835 confirmed incidence cases and 57,734 deaths occurred. We found a decrease of -179.93 (95% CI: -380.11 to -20.25, P-value=0.078) confirmed incidence cases following the implementation of the social distancing policy, corresponding to a daily decrease in the trend of -31.17 (95% CI: -46.95 to -15.40, P-value=0.08). Moreover, we found a decrease of -28.28 (95% CI: -43.55 to -13.01, P-value=0.05) deaths, corresponding to a daily decrease in the trend of -4.52 (95% CI: -5.25 to -3.78, P-value=0.003). Conclusion: The growth rate of confirmed incidence cases and deaths from COVID-19 in Iran has decreased from March 27, 2020 to January 29, 2021, after the implementation of social distancing. By implementing this policy in all countries, the burden of COVID-19 may be mitigated.     

Changing Use of Surgical Antibiotic Prophylaxis in Thika Hospital,Kenya: A Quality Improvement Intervention with an Interrupted Time Series Design

Introduction

In low-income countries, Surgical Site Infection (SSI) is a common form of hospital-acquired infection. Antibiotic prophylaxis is an effective method of preventing these infections, if given immediately before the start of surgery. Although several studies in Africa have compared pre-operative versus post-operative prophylaxis, there are no studies describing the implementation of policies to improve prescribing of surgical antibiotic prophylaxis in African hospitals.

Methods

We conducted SSI surveillance at a typical Government hospital in Kenya over a 16 month period between August 2010 and December 2011, using standard definitions of SSI and the extent of contamination of surgical wounds. As an intervention, we developed a hospital policy that advised pre-operative antibiotic prophylaxis and discouraged extended post-operative antibiotics use. We measured process, outcome and balancing effects of this intervention in using an interrupted time series design.

Results

From a starting point of near-exclusive post-operative antibiotic use, after policy introduction in February 2011 there was rapid adoption of the use of pre-operative antibiotic prophylaxis (60% of operations at 1 week; 98% at 6 weeks) and a substantial decrease in the use of post-operative antibiotics (40% of operations at 1 week; 10% at 6 weeks) in Clean and Clean-Contaminated surgery. There was no immediate step-change in risk of SSI, but overall, there appeared to be a moderate reduction in the risk of superficial SSI across all levels of wound contamination. There were marked reductions in the costs associated with antibiotic use, the number of intravenous injections performed and nursing time spent administering these.

Conclusion

Implementation of a locally developed policy regarding surgical antibiotic prophylaxis is an achievable quality improvement target for hospitals in low-income countries, and can lead to substantial benefits for individual patients and the institution.     

Dynamical Analysis and Visualization of Tornadoes Time Series

In this paper we analyze the behavior of tornado time-series in the U.S. from the perspective of dynamical systems. A tornado is a violently rotating column of air extending from a cumulonimbus cloud down to the ground. Such phenomena reveal features that are well described by power law functions and unveil characteristics found in systems with long range memory effects. Tornado time series are viewed as the output of a complex system and are interpreted as a manifestation of its dynamics. Tornadoes are modeled as sequences of Dirac impulses with amplitude proportional to the events size. First, a collection of time series involving 64 years is analyzed in the frequency domain by means of the Fourier transform. The amplitude spectra are approximated by power law functions and their parameters are read as an underlying signature of the system dynamics. Second, it is adopted the concept of circular time and the collective behavior of tornadoes analyzed. Clustering techniques are then adopted to identify and visualize the emerging patterns.     

Economic Behavior under the Influence of Alcohol: An Experiment on Time Preferences,Risk-Taking,and Altruism

We report results from an incentivized laboratory experiment undertaken with the purpose of providing controlled evidence on the causal effects of alcohol consumption on risk-taking, time preferences and altruism. Our design disentangles the pharmacological effects of alcohol intoxication from those mediated by expectations, as we compare the behavior of three groups of subjects: those who participated in an experiment with no reference to alcohol, those who were exposed to the possibility of consuming alcohol but were given a placebo and those who effectively consumed alcohol. All subjects participated in a series of economic tasks administered in the same sequence across treatments. After controlling for both the willingness to pay for an object and the potential misperception of probabilities as elicited in the experiment, we detect no effect of alcohol in depleting subjects’ risk tolerance. However, we find that alcohol intoxication increases impatience and makes subjects less altruistic.     

The Impact of Divergence Time on the Nature of Population Structure: An Example from Iceland

The Icelandic population has been sampled in many disease association studies, providing a strong motivation to understand the structure of this population and its ramifications for disease gene mapping. Previous work using 40 microsatellites showed that the Icelandic population is relatively homogeneous, but exhibits subtle population structure that can bias disease association statistics. Here, we show that regional geographic ancestries of individuals from Iceland can be distinguished using 292,289 autosomal single-nucleotide polymorphisms (SNPs). We further show that subpopulation differences are due to genetic drift since the settlement of Iceland 1100 years ago, and not to varying contributions from different ancestral populations. A consequence of the recent origin of Icelandic population structure is that allele frequency differences follow a null distribution devoid of outliers, so that the risk of false positive associations due to stratification is minimal. Our results highlight an important distinction between population differences attributable to recent drift and those arising from more ancient divergence, which has implications both for association studies and for efforts to detect natural selection using population differentiation.     

Efficient Transfer Entropy Analysis of Non-Stationary Neural Time Series

Information theory allows us to investigate information processing in neural systems in terms of information transfer, storage and modification. Especially the measure of information transfer, transfer entropy, has seen a dramatic surge of interest in neuroscience. Estimating transfer entropy from two processes requires the observation of multiple realizations of these processes to estimate associated probability density functions. To obtain these necessary observations, available estimators typically assume stationarity of processes to allow pooling of observations over time. This assumption however, is a major obstacle to the application of these estimators in neuroscience as observed processes are often non-stationary. As a solution, Gomez-Herrero and colleagues theoretically showed that the stationarity assumption may be avoided by estimating transfer entropy from an ensemble of realizations. Such an ensemble of realizations is often readily available in neuroscience experiments in the form of experimental trials. Thus, in this work we combine the ensemble method with a recently proposed transfer entropy estimator to make transfer entropy estimation applicable to non-stationary time series. We present an efficient implementation of the approach that is suitable for the increased computational demand of the ensemble method''s practical application. In particular, we use a massively parallel implementation for a graphics processing unit to handle the computationally most heavy aspects of the ensemble method for transfer entropy estimation. We test the performance and robustness of our implementation on data from numerical simulations of stochastic processes. We also demonstrate the applicability of the ensemble method to magnetoencephalographic data. While we mainly evaluate the proposed method for neuroscience data, we expect it to be applicable in a variety of fields that are concerned with the analysis of information transfer in complex biological, social, and artificial systems.     

Tuberculosis Incidence Correlates with Sunshine: An Ecological 28-Year Time Series Study

Background

Birmingham is the largest UK city after London, and central Birmingham has an annual tuberculosis incidence of 80 per 100,000. We examined seasonality and sunlight as drivers of tuberculosis incidence. Hours of sunshine are seasonal, sunshine exposure is necessary for the production of vitamin D by the body and vitamin D plays a role in the host response to tuberculosis.

Methods

We performed an ecological study that examined tuberculosis incidence in Birmingham from Dec 1981 to Nov 2009, using publicly-available data from statutory tuberculosis notifications, and related this to the seasons and hours of sunshine (UK Meteorological Office data) using unmeasured component models.

Results

There were 9,739 tuberculosis cases over the study period. There was strong evidence for seasonality, with notifications being 24.1% higher in summer than winter (p<0.001). Winter dips in sunshine correlated with peaks in tuberculosis incidence six months later (4.7% increase in incidence for each 100 hours decrease in sunshine, p<0.001).

Discussion and Conclusion

A potential mechanism for these associations includes decreased vitamin D levels with consequent impaired host defence arising from reduced sunshine exposure in winter. This is the longest time series of any published study and our use of statutory notifications means this data is essentially complete. We cannot, however, exclude the possibility that another factor closely correlated with the seasons, other than sunshine, is responsible. Furthermore, exposure to sunlight depends not only on total hours of sunshine but also on multiple individual factors. Our results should therefore be considered hypothesis-generating. Confirmation of a potential causal relationship between winter vitamin D deficiency and summer peaks in tuberculosis incidence would require a randomized-controlled trial of the effect of vitamin D supplementation on future tuberculosis incidence.