Feng Chen

The Impact of Media Reporting on the Emergence of Charcoal Burning Suicide in Taiwan

We investigated the association of the intensity of newspaper reporting of charcoal burning suicide with the incidence of such deaths in Taiwan during 1998–2002. A counting process approach was used to estimate the incidence of suicides and intensity of news reporting. Conditional Poisson generalized linear autoregressive models were performed to assess the association of the intensity of newspaper reporting of charcoal burning and non-charcoal burning suicides with the actual number of charcoal burning and non-charcoal burning suicides the following day. We found that increases in the reporting of charcoal burning suicide were associated with increases in the incidence of charcoal burning suicide on the following day, with each reported charcoal burning news item being associated with a 16% increase in next day charcoal burning suicide (p<.0001). However, the reporting of other methods of suicide was not related to their incidence. We conclude that extensive media reporting of charcoal burning suicides appears to have contributed to the rapid rise in the incidence of the novel method in Taiwan during the initial stage of the suicide epidemic. Regulating media reporting of novel suicide methods may prevent an epidemic spread of such new methods.

The impact of media reporting of suicide on actual suicides in Taiwan, 2002–05

Objectives To assess changes in the intensity of suicide news reporting in Taiwans local newspapers after the arrival of a daily tabloid-type newspaper, Apple Daily (AD), and evaluate the impact of suicide news reporting on actual suicides and possible mutual causation. Methods A counting process was used to estimate the intensity of daily suicide news items reported in the China Times (CT) and United Daily (UD) before and after the arrival of AD (2002–05). Poisson regression models were used to assess the impact of the intensity of suicide news reporting on the actual number of next day suicides. Grangers causation model was used to assess mutual causation between suicide news reporting and actual suicides. Results There was a significant increase in reporting intensity of suicide news in the UD soon after the entry of the AD into Taiwans media market, while a delayed increase of approximately 1 year was observed in the CT. After the arrival of the AD, the reporting intensity in the UD was significantly related to the occurrence of actual suicides (p<0.05), even after controlling for social variables, whereas no significant correlation was previously observed. Mutual causation between suicide news reporting and actual suicides was also observed. Conclusions The presence of the AD in Taiwan has fuelled competitive reporting of suicide news among traditional newspapers. This increase in the intensity of suicide news reporting has consequently had an impact on the actual number of suicides. This provides further empirical support for improving media reporting as a key element in suicide prevention.

Nonparametric Estimation of Multiplicative Counting Process Intensity Functions with an Application to the Beijing SARS Epidemic

The local polynomial methods and martingale estimating equations are used to develop closed form estimators of the intensity function and its derivatives for multiplicative counting process models. The consistency and asymptotic normality of the estimators are established. The estimator generalizes that proposed by Ramlau-Hansen (1983) with a smaller bias than the Ramlau-Hansen intensity estimator. The derivative estimators give smoother estimates than the Ramlau-Hansen derivative estimators. The proposed estimators are applied to analyze the infection rate and its derivatives of the 2003 Severe Acute Respiratory Syndrome (SARS) epidemic in Beijing, China.

Assessing the Efficacy of Restricting Access to Barbecue Charcoal for Suicide Prevention in Taiwan: A Community-Based Intervention Trial

Objective Charcoal-burning suicide has recently been spreading to many Asian countries. There have also been several cases involving this new method of suicide in Western countries. Restricting access to suicide means is one of the few suicide-prevention measures that have been supported by empirical evidence. The current study aims to assess the effectiveness of a community intervention program that restricts access to charcoal to prevent suicide in Taiwan. Methods and Findings A quasi-experimental design is used to compare method-specific (charcoal-burning suicide, non-charcoal-burning suicide) and overall suicide rates in New Taipei City (the intervention site, with a population of 3.9 million) with two other cities (Taipei City and Kaohsiung City, the control sites, each with 2.7 million residents) before (Jan 1st 2009- April 30th 2012) and after (May 1st 2012-Dec. 31st 2013) the initiation of a charcoal-restriction program on May 1st 2012. The program mandates the removal of barbecue charcoal from open shelves to locked storage in major retail stores in New Taipei City. No such restriction measure was implemented in the two control sites. Generalized linear regression models incorporating secular trends were used to compare the changes in method-specific and overall suicide rates before and after the initiation of the restriction measure. A simulation approach was used to estimate the number of lives saved by the intervention. Compared with the pre-intervention period, the estimated rate reduction of charcoal-burning suicide in New Taipei City was 37% (95% CI: 17%, 50%) after the intervention. Taking secular trends into account, the reduction was 30% (95% CI: 14%, 44%). No compensatory rise in non-charcoal-burning suicide was observed in New Taipei City. No significant reduction in charcoal-burning suicide was observed in the other two control sites. The simulation approach estimated that 91 (95%CI [55, 128]) lives in New Taipei City were saved during the 20 months of the intervention. Conclusion Our results demonstrate that the charcoal-restriction program reduced method-specific and overall suicides. This study provides strong empirical evidence that restricting the accessibility of common lethal methods of suicide can effectively reduce suicide rates.

Case-cohort analysis of clusters of recurrent events

The case-cohort sampling, first proposed in Prentice (Biometrika 73:1–11, 1986), is one of the most effective cohort designs for analysis of event occurrence, with the regression model being the typical Cox proportional hazards model. This paper extends to consider the case-cohort design for recurrent events with certain specific clustering feature, which is captured by a properly modified Cox-type self-exciting intensity model. We discuss the advantage of using this model and validate the pseudo-likelihood method. Simulation studies are presented in support of the theory. Application is illustrated with analysis of a bladder cancer data.

A study on the mutual causation of suicide reporting and suicide incidences

BACKGROUND Little research has been done on the complex relationships between the effect of news reporting on suicide incidence and vice versa (i.e., mutual causation). Furthermore, few studies have examined whether the entry of a new media outlet into a market changes the media dynamics in that market. METHODS A recursive two-way feedback model was used to test for mutual causation between suicide reporting and suicide incidence on a daily basis. We applied the model to examine the effect of the arrival of the Apple Daily (AD) newspaper in Taiwan and whether its suicide reporting affected the suicide incidence and suicide reporting of two other newspapers, the United Daily (UD) and the China Times (CT). RESULTS The ADs entry into Taiwan led to a major shift in the relationship between suicide incidence and suicide reporting. The AD stimulated more suicide coverage by the UD and the CT the following day; conversely, the UD and the CT had no such impact on the AD. Before the entry of the AD, there was little correlation between daily suicide incidence and suicide reporting, but the suicide reporting of the UD and CT correlated significantly with daily suicide incidence after the entry of the AD. LIMITATIONS Media impact was assessed by number of news items; detailed content analysis of the reporting was not conducted. CONCLUSIONS The vicious business competition facing new dailies in Taiwans media market has changed the mass media ecology. Efforts to prevent suicide by regulating the media should closely monitor not only the behavior of newcomers, but also the established news medias reaction to new competitors.

Nonparametric Estimation for Self-Exciting Point Processes—A Parsimonious Approach

There is ample evidence that in applications of self-exciting point-process models, the intensity of background events is often far from constant. If a constant background is imposed that assumption can reduce significantly the quality of statistical analysis, in problems as diverse as modeling the after-shocks of earthquakes and the study of ultra-high frequency financial data. Parametric models can be used to alleviate this problem, but they run the risk of distorting inference by misspecifying the nature of the background intensity function. On the other hand, a purely nonparametric approach to analysis leads to problems of identifiability; when a nonparametric approach is taken, not every aspect of the model can be identified from data recorded along a single observed sample path. In this article, we suggest overcoming this difficulty by using an approach based on the principle of parsimony, or Occam’s razor. In particular, we suggest taking the point-process intensity to be either a constant or to have maximum differential entropy, in cases where there is not sufficient empirical evidence to suggest that the background intensity function is more complex than those models. This approach is seldom, if ever, used for nonparametric function estimation in other settings, not least because in those cases more data are typically available. However, our “ontological parsimony” argument is appropriate in the context of self-exciting point-process models. Supplementary materials are available online.

Dividend Optimization for Regime Switching General Diffusions

We consider the optimal dividend distribution problem of a financial corporation whose surplus is modeled by a general diffusion process with both the drift and diffusion coefficients depending on the external economic regime as well as the surplus itself through general functions. The aim is to find a dividend payout scheme that maximizes the present value of the total dividends until ruin. We show that, depending on the configuration of the model parameters, there are two exclusive scenarios: (i)the optimal strategy uniquely exists and corresponds to paying out all surpluses in excess of a critical level (barrier) dependent on the economic regime and paying nothing when the surplus is below the critical level;(ii)there are no optimal strategies.

Direct Likelihood Evaluation for the Renewal Hawkes Process

ABSTRACT An interesting extension of the widely applied Hawkes self-exiting point process, the renewal Hawkes (RHawkes) process, was recently proposed by Wheatley, Filimonov, and Sornette, which has the potential to significantly widen the application domains of the self-exciting point processes. However, they claimed that computation of the likelihood of the RHawkes process requires exponential time and therefore is practically impossible. They proposed two expectation–maximization (EM) type algorithms to compute the maximum likelihood estimator (MLE) of the model parameters. Because of the fundamental role of likelihood in statistical inference, a practically feasible method for likelihood evaluation is highly desirable. In this article, we provide an algorithm that evaluates the likelihood of the RHawkes process in quadratic time, a drastic improvement from the exponential time claimed by Wheatley, Filimonov, and Sornette. We demonstrate the superior performance of the resulting MLEs of the model relative to the EM estimators through simulations. We also present a computationally efficient procedure to calculate the Rosenblatt residuals of the process for goodness-of-fit assessment, and a simple yet efficient procedure for future event prediction. The proposed methodologies were applied on real data from seismology and finance. An R package implementing the proposed methodologies is included in the supplementary materials.

Media effects on suicide methods: A case study on Hong Kong 1998-2005.

Background Previous studies have suggested that mass media’s reports of new suicide methods will increase suicides using the same method. The same pattern seems not to apply to a conventional suicide method, unless it was used by a celebrity. Objective 1) to examine media effects on both new and non-new suicide methods during 1998 and 2005 in Hong Kong (HK), when a new method by burning charcoal (CB suicide) was spreading in the region. 2) to examine how CB competed with non-CB methods in terms of media coverage and “recruiting” suicidal persons in the socio-economic context. Methods A self- and mutual- exciting process model was fitted to the data, adjusting for divorce rate, unemployment rate, and property price index. Breaking the whole period into onset, peak, and post-peak stages, the model was fitted again to examine the differences. Results Comparable copycat effects were found on both CB and non-CB suicide news. The only cross-method media effects were found in the onset stage when non-CB suicide news showed suppressing effect on CB suicides. CB suicides reported a significant self-excitation effect. A higher divorce rate and lower property price index were associated with significantly more suicides incidences and more suicide news. Conclusions The emerging of CB suicide method did not substitute media coverage of non-CB suicide in HK. Media effects in this case were not limited to new suicide method or celebrity suicide. The effects were further fueled by adverse socio-economic conditions.