Segmented Time-series Analysis Research Articles

Correction: The Impact of Policy Guidelines on Hospital Antibiotic Use over a Decade: A Segmented Time Series Analysis

Correction: The Impact of Policy Guidelines on Hospital Antibiotic Use over a Decade: A Segmented Time Series Analysis

A characteristic pattern in the postural sway of unilateral vestibular impaired patients

The statistical properties of the center of pressure (COP) change over time, and -invariant methods of the COP analysis is not sufficient to monitor the changes. Therefore, dynamic temporal information of the COP signals has been important in assessing the postural stability. The purpose of this study was to evaluate the characteristic pattern of time–frequency dynamics during the upright stance in patients with unilateral vestibular dysfunction (UVD). This study included 20 dizzy patients with UVD and age-matched 20 control subjects without any vestibular disorder. The COP signals were collected using a single standard force platform and a spectral analysis including the time–frequency dynamics was carried out on the basis of the maximum entropy method (MEM) by using a segment time series analysis. The power spectral density (PSD) analysis exhibited an exponential decreasing shape (1/f −β) when plotted on a double logarithmic scale. The average value of β in the low frequency range of less than 1Hz in the UVD group was significantly lower than that in the control group (p<0.05) in the medial-lateral (ML) direction under eyes closed condition. In the segment time series analysis, the peak frequency of the COP signals in the patients with UVD gradually converged to the frequency range of 0.1–0.2Hz in the ML direction under EC condition. Our results suggest that the segment time series analysis of the COP signals can derive a characteristic pattern related to the UVD. It may indicate an increased reflexivity of the COP fluctuation by the UVD, resulting in inducing the instability.

The Impact of Policy Guidelines on Hospital Antibiotic Use over a Decade: A Segmented Time Series Analysis

IntroductionAntibiotic pressure contributes to rising antibiotic resistance. Policy guidelines encourage rational prescribing behavior, but effectiveness in containing antibiotic use needs further assessment. This study therefore assessed the patterns of antibiotic use over a decade and analyzed the impact of different modes of guideline development and dissemination on inpatient antibiotic use.MethodsAntibiotic use was calculated monthly as defined daily doses (DDD) per 100 bed days for nine antibiotic groups and overall. This time series compared trends in antibiotic use in five adjacent time periods identified as ‘Segments,’ divided based on differing modes of guideline development and implementation: Segment 1– Baseline prior to antibiotic guidelines development; Segment 2– During preparation of guidelines and booklet dissemination; Segment 3– Dormant period with no guidelines dissemination; Segment 4– Booklet dissemination of revised guidelines; Segment 5– Booklet dissemination of revised guidelines with intranet access. Regression analysis adapted for segmented time series and adjusted for seasonality assessed changes in antibiotic use trend.ResultsOverall antibiotic use increased at a monthly rate of 0.95 (SE = 0.18), 0.21 (SE = 0.08) and 0.31 (SE = 0.06) for Segments 1, 2 and 3, stabilized in Segment 4 (0.05; SE = 0.10) and declined in Segment 5 (−0.37; SE = 0.11). Segments 1, 2 and 4 exhibited seasonal fluctuations. Pairwise segmented regression adjusted for seasonality revealed a significant drop in monthly antibiotic use of 0.401 (SE = 0.089; p<0.001) for Segment 5 compared to Segment 4. Most antibiotic groups showed similar trends to overall use.ConclusionUse of overall and specific antibiotic groups showed varied patterns and seasonal fluctuations. Containment of rising overall antibiotic use was possible during periods of active guideline dissemination. Wider access through intranet facilitated significant decline in use. Stakeholders and policy makers are urged to develop guidelines, ensure active dissemination and enable accessibility through computer networks to contain antibiotic use and decrease antibiotic pressure.

Impact of a Bronchiolitis Guideline on ED Resource Use and Cost: A Segmented Time-Series Analysis

Bronchiolitis is a major cause of infant morbidity and contributes to millions of dollars in health care costs. Care guidelines may cut costs by reducing unnecessary resource utilization. Through the implementation of a guideline, we sought to reduce unnecessary resource utilization and improve the value of care provided to infants with bronchiolitis in a pediatric emergency department (ED). We conducted an interrupted time series that examined ED visits of 2929 patients with bronchiolitis, aged 1 to 12 months old, seen between November 2007 and April 2013. Outcomes were proportion having a chest radiograph (CXR), respiratory syncytial virus (RSV) testing, albuterol or antibiotic administration, and the total cost of care. Balancing measures included admission rate, returns to the ED resulting in admission within 72 hours of discharge, and ED length of stay (LOS). There were no significant preexisting trends in the outcomes. After guideline implementation, there was an absolute reduction of 23% in CXR (95% confidence interval [CI]: 11% to 34%), 11% in RSV testing (95% CI: 6% to 17%), 7% in albuterol use (95% CI: 0.2% to 13%), and 41 minutes in ED LOS (95% CI: 16 to 65 minutes). Mean cost per patient was reduced by $197 (95% CI: $136 to $259). Total cost savings was $196,409 (95% CI: $135,592 to $258,223) over the 2 bronchiolitis seasons after guideline implementation. There were no significant differences in antibiotic use, admission rates, or returns resulting in admission within 72 hours of discharge. A bronchiolitis guideline was associated with reductions in CXR, RSV testing, albuterol use, ED LOS, and total costs in a pediatric ED.

Analysing policy interventions to prohibit over‐the‐counter antibiotic sales in four Latin American countries

To describe and evaluate policies implemented in Chile, Colombia, Venezuela and Mexico (1995-2009) to prohibit antibiotic OTC sales and explore limitations in available data. We searched and analysed legislation, grey literature and peer-reviewed publications on regulatory interventions and implementation strategies to enforce prohibition of OTC antibiotic sales. We also assessed the impact using private sector retail sales data of antibiotics studying changes in level and consumption trends before and after the policy change using segmented time series analysis. Finally, we assessed the completeness and data quality through an established checklist to test the suitability of the data for analysis of the interventions. Whereas Chile implemented a comprehensive package of interventions to accompany regulation changes, Colombia's reform was limited to the capital district and Venezuela's limited to only some antibiotics and without awareness campaigns. In Mexico, no enforcement was enacted. The data showed a differential effect of the intervention among the countries studied with a significant change in level of consumption in Chile (-5.56 DID) and in Colombia (-1.00DID). In Venezuela and Mexico, no significant change in level and slope was found. Changes in population coverage were identified as principal limitations of using sales data for evaluating the reform impact. Retail sales data can be useful when assessing policy impact but should be supplemented by other data sources such as public sector sales and prescription data. Implementing regulatory enforcement has shown some impact, but a sustainable, concerted approach will be needed to address OTC sales in the future.

False Claims Act Prosecution Did Not Deter Off-Label Drug Use In The Case Of Neurontin

Since 2004 the United States has collected approximately $8 billion from fraud enforcement actions against pharmaceutical manufacturers accused under the federal False Claims Act of illegally promoting drugs for off-label uses. Using the case of gabapentin (Neurontin), a drug approved for epilepsy but prescribed for a variety of conditions, we sought to determine whether the enforcement action also influenced off-label prescribing rates. We conducted a segmented time-series analysis using key legal milestones: the initiation of a sealed investigation, public announcement of the investigation, and settlement of the case. Off-label use grew steadily until settlement, when gabapentin prescriptions declined for both off-label and on-label indications. Because enforcement actions targeting illegal off-label promotion might not have a substantial deterrent effect on prescription rates until after settlement, they should be combined with other efforts to combat off-label promotion. These could include additional resources for enforcement and a steep increase in penalties because settlements to this point have been dwarfed by the financial gains to pharmaceutical companies from engaging in improper off-label marketing.

Time Series Analysis of Incidence Data of Influenza in Japan

BackgroundMuch effort has been expended on interpreting the mechanism of influenza epidemics, so as to better predict them. In addition to the obvious annual cycle of influenza epidemics, longer-term incidence patterns are present. These so-called interepidemic periods have long been a focus of epidemiology. However, there has been less investigation of the interepidemic period of influenza epidemics. In the present study, we used spectral analysis of influenza morbidity records to indentify the interepidemic period of influenza epidemics in Japan.MethodsWe used time series data of the monthly incidence of influenza in Japan from January 1948 through December 1998. To evaluate the incidence data, we conducted maximum entropy method (MEM) spectral analysis, which is useful in investigating the periodicities of shorter time series, such as that of the incidence data used in the present study. We also conducted a segment time series analysis and obtained a 3-dimensional spectral array.ResultsBased on the results of power spectral density (PSD) obtained from MEM spectral analysis, we identified 3 periodic modes as the interepidemic periods of the incidence data. Segment time series analysis revealed that the amount of amplitude of the interepidemic periods increased during the occurrence of influenza pandemics and decreased when vaccine programs were introduced.ConclusionsThe findings suggest that the temporal behavior of the interepidemic periods of influenza epidemics is correlated with the magnitude of cross-reactive immune responses.

Assessment of Changes in Utilization of Health-Care Services after Implementation of a Prior Authorization Policy for Atypical Antipsychotic Agents

In 2004, the Georgia Medicaid program implemented a prior authorization (PA) policy for certain atypical antipsychotic agents, resulting in an average savings of $2.7 million per year. To determine whether implementation of a PA policy for atypical antipsychotic drugs increased health-care utilization in the Georgia Medicaid program from July 2003 to April 2006. A single cohort observational study employing segmented regression and time series analysis was conducted to determine health-care services utilization, including emergency department (ED) visits, outpatient office visits, hospital admissions, and length of stay (LOS). Study subjects included continuously eligible adult Georgia Medicaid recipients with a schizophrenia-related diagnosis and documented use of an atypical antipsychotic medication (N = 12,120). Where applicable, analysis of a noncontinuously eligible population was also performed to investigate disenrollment bias in study results. A significant decline in post-policy trend for the average number of ED visits (absolute difference -0.042 per member per month (PMPM); relative difference -20.92%) and average number of hospital admissions PMPM (absolute difference -0.010 PMPM; relative difference -22.27%) was observed at the end of the study period. Baseline and pre-policy trends were found to be significant predictors for both endpoints. Significant models were not identified for average outpatient office visits PMPM or average LOS per admission. In contrast to other published studies on PA for atypical antipsychotics, patient outcomes improved after the initiation of the policy. To the extent that medical utilization reflects patient health outcomes and health status, the results of this study indicate that the PA program has potentially improved the health of schizophrenic patients in Georgia and lowered program costs.

The impact of nonreferral outpatient co-payment on medical care utilization and expenditures in Taiwan

Background Taiwan's National Health Insurance's (NHI) generous coverage and patients' freedom to access different tiers of medical facilities have resulted in accelerating outpatient care utilization and costs. To deter nonessential visits and encourage initial contact in primary care (physician clinics), a differential co-payment was introduced on 15th July 2005. Under this, patients pay more for outpatient consultations at “higher tiers” of medical facilities (local community hospitals, regional hospitals, medical centers), particularly if accessed without referral. Objectives This study explored the impact of this policy on outpatient medical activities and expenditures, different co-payment groups, and tiers of medical facilities. Methods A segmented time-series analysis on regional weekly outpatient medical claims (January 2004 to July 2006) was conducted. Outcome variables (number of visits, number of outpatients, total cost of outpatient care) and variables for cost structure were stratified by tiers of medical facilities and co-payment groups. Analysis used the auto-regressive integrated moving-average model in STATA 9.0. Results The overall number of outpatient visits significantly decreased after policy implementation due to a reduction in the number of patients using outpatient facilities, but total costs of care remained unchanged. The policy had its greatest impact on the number of visits to regional and local community hospitals but had no influence on those to the medical centers. Medical utilization in physician clinics decreased due to an audit of reimbursement declarations. Overall, the policy failed to encourage referrals from primary care to higher tiers because there was no obvious shifting of medical utilization and costs reversely. Conclusions Differential co-payment policy decreased total medication utilization but not costs to NHI. The results suggest that the increased level of co-payment charge and the strategy of a single cost-sharing policy are not sufficient to promote referrals within the system. To achieve an effective co-payment policy, further research is needed to explore how patients' out-of-pocket payment affects medical utilization and which forces (not susceptible to co-payment) act in tertiary facilities.

Spectral Study of Measles Epidemics: The Dependence of Spectral Gradient on the Population Size of the Community

We have carried out spectral analysis of measles notifications in several communities in Denmark, UK and USA. The results confirm that each power spectral density (PSD) shows exponential characteristics, which are universally observed in the PSD for time series generated from nonlinear dynamical system. The exponential gradient increases with the population size. For almost all communities, many spectral lines observed in each PSD can be fully assigned to linear combinations of several fundamental periods, suggesting that the measles data are substantially noise-free. The optimum least squares fitting curve calculated using these fundamental periods essentially reproduces an underlying variation of the measles data, and an extension of the curve can be used to predict measles epidemics. For the communities with large population sizes, some PSD patterns obtained from segment time series analysis show a close resemblance to the PSD patterns at the initial stages of a period-doubling bifurcation process for the so-called susceptible/exposed/infectious/recovered (SEIR) model with seasonal forcing. The meaning of the relationship between the exponential gradient and the population size is discussed.

A comparison of antihypertensive medication utilization before and after guideline changes using the Department of Defense Prescription Database.

To determine changes in the proportion of antihypertensive medication utilization, distributed by medication classes, associated with the Fifth Report of the Joint National Committee (JNC V) guideline changes and the Department of Defense Pharmacoeconomic Center's recommendations to follow JNC V guidelines in presumed newly treated hypertensive patients. A 43-month, longitudinal, retrospective analysis using data from the Department of Defense Uniformed Services Prescription Database. Seven outpatient US military sites. Eligible patients (n = 7277) included those from seven military sites, aged 20-49 years, who were: (1) active-duty members of the US Armed Forces, (2) active-duty members' family members, (3) retired members of the US Armed Forces, and (4) US Armed Forces retired members' family members. Proportion of antihypertensive medication utilization, distributed by medication class over the 43-month study period. Segmented time series analysis was used for each of the following four medication classes: angiotensin-converting enzyme inhibitors, beta-blockers, calcium-channel blockers, and diuretics. Segmented time series analyses revealed no significant differences in utilization of the four medication classes that corresponded to published guidelines outlining initial antihypertensive therapy. It appeared that JNC V guidelines and recommendations of the Pharmacoeconomic Center to follow JNC V had little effect on the utilization of prescription medication classes studied.

Study on Chaotic Characteristics of Incidence Data of Measles

In this paper, the chaotic characteristics of time series data of monthly number reported cases of measles in New York City from January 1928 to December 1963 were investigated using the computing program “MemCalc”. Power spectral densities (PSDs) were calculated for the measles data, two types of time series numerically calculated from the SEIR epidemic model (the chaotic and periodic time series) and the noisy time series. Exponential characteristics were observed in the PSDs for the measles data and the chaotic and periodic time series, but not observed in the PSDs for the noisy time series. For the measles data and the chaotic time series, the PSDs exhibit the broad continuous peaks peculiar to the chaotic state. The behavior of the three-dimensional spectral array obtained from segment time series analysis is distinguishable from the cases of the periodic and noisy time series. It was concluded that the dynamics of the measles data is chaotic. The present method was used to extract the chaotic characteristics from measles data having a short time period.

New Method of Time Series Analysis and Its Application to Wolf's Sunspot Number Data. II: Periodicities of The Daily Sunspot Numbers

The daily sunspot numbers data from 1979 October 1 to 1993 December 31 were analyzed using a new method of time series analysis proposed by the authors, and the long- and short-term periodicities were investigated. A three-dimensional (3D) spectral array of the normalized power spectral densities (PSD's) obtained by the segment time series analysis revealed the temporal behavior of the 27- and 155-d periodicities in the solar activity minimum. In the normalized 3D spectral array, the solar rotation cycle of 27-d accompanied by its multiplets was confirmed to migrate gradually from the low frequency region toward the high frequency region in the solar activity minimum as the solar activity increases, while the 155-d periodicity exists in the solar activity maximum only. Moreover, the behavior of the daily sunspot numbers variation was conjectured to be deterministic chaos, based on the exponential characteristics of the PSD's and the subharmonic relations of the spectral lines.

Power Spectral Densities of Temporal Variations of Blood Pressures

A comprehensive study of power spectral densities for time series data of blood pressure measured for four individuals, one healthy subject and three diseased subjects (pure autonomic failure (PAF), PAF complicated with supraventricular premature contraction, and multiple cerebral infarction with atrial fibrillation (Af)), was carried out using the computing program “MemCalc.” The whole spectrum for every time series indicates exponential characteristics until it levels off at a limit determined by the accuracy of the present measurement. The three-dimensional spectral array obtained from segment time series analysis showed a fine array of the fundamental and its harmonic modes, and was concluded to be a powerful tool for medical diagnosis. “Homeostasis” in living bodies is discussed on the basis of the present results. A new link between the nonlinear sciences and basic and/or clinical medicine found in the present study is sure to play an important role in medical diagnosis and prognosis.