Psychotherapy Brown Bag

How do therapists decide which therapeutic orientation they prefer?

by Michael D. Anestis, M.S.

I came across a study today that looks at something I believe to be an interesting and fairly important question: how do therapists end up deciding upon which therapeutic orientation (e.g., cognitive behavioral therapy, psychodynamic therapy) they prefer?  The study, published in the May 2010 issue of Psychotherapy Research and conducted by Joseph Buckman and Chris Barker of University College London, is not the first to consider this topic, but their findings are interesting and I thought they might prompt some good discussion here on PBB.  Before discussing what Buckman and Barker (2010) found though, let me quickly review some of the prior research findings that the authors discussed. 

Therapists who align themselves with a psychodynamic or experiential therapy style have been shown to score higher on the personality measures of Openness to Experience and Neuroticism than have therapists who consider themselves cognitive-behavioral, behavioral, or systemic (e.g., Arthur, 1998, 2000; Poznanski & McLennan, 2003, 2004).  Openness to Experience involves an appreciation for abstract ideas, art, new experiences, and unusual ideas.  Neuroticism involves the frequent experience of negative emotions, difficulties regulation emotions, an high reactivity to stress. 

In addition to measures of personality, therapists of different orientations have been shown to differ in their philosophical beliefs and world view.  Some researchers have also found that training experiences play a pivotal role in determining how therapists view the proper way to conduct therapy.  For example, Poznanski and McLennan (2003, 2004) found that, whereas nearly all CBT therapists in their sample were heavily influenced by their university training, only a quarter of the psychodynamic and systemic therapists reported the same. 

To look at this further, the Buckman recruited from a pool of 331 trainees in four psychology courses.  Ultimately, 142 participants took part in the study, 82% of whom were female.  At the time of data collection, Buckman was also a student in one of the four classes and, while he was only able to recruit from the other courses by email, he approached his peers in person to discuss participation in his own course.  Perhaps not surprisingly, there was a higher rate of participation in Buckman's course (69%) than in the other three (35%, 28%, and 23%).  This, in my opinion, is a potentially fatal flaw to the study, but the authors noted this issue later, so I will hold off on discussing it until I get to their acknowledgment later on.

Buckman and Barker reported a number of interesting findings.  Importantly, keep in mind that therapeutic orientation was measured dimensionally, meaning that each participant reported the degree to which they related to and espoused the ideas of a number of different orientations rather than describing themselves as one orientation or another.  With respect to personality, they found that CBT orientation was positively correlated with conscientiousness, which involves planning and preparation, and negatively correlated with Openness to Experience.   Psychodynamic orientation resulted in the opposite pattern. With respect to worldview, CBT orientation was positively correlated with maintaining a view that emphasizes testing beliefs through observable phenomena whereas psychodynamic orientation was positively correlated with maintaining a view that emphasizes symbolic interpretations.  With respect to training experiences, both psychodynamic and CBT orientations were correlated with having a training history that emphasized that particular approach, however psychodynamic therapy was also negative correlated with having had supervision in CBT.

The authors did not stop here though.  They were also interested in whether person factors (e.g., personality variables, worldview) or training factors (e.g., course emphasis, supervision) played a larger role in therapists aligning themselves to particular therapeutic orientations.  Person factors accounted for a significant portion of CBT preference, whereas training factors did not.  While person factors also significantly predicted a preference for psychodynamic therapy, training factors accounted for a substantially greater proportion of the variance.  In other words, when individuals preferred CBT, that preference was best explained by stable aspects of their personality whereas when preferred psychodynamic therapy, their personality made a contribution, but their experiences in classes and supervision played a stronger role.

Now, all of these findings are interesting, but they are also quite vague, meaning that several of the findings could be legitimately interpreted in a number of ways, many of which are inconsistent with one another.  The authors themselves mentioned that, because Buckman personally knew many of the participants, there were legitimate questions regarding demand characteristics that likely led some individuals to adjust their answers (e.g., making themselves appear less neurotic).  The fact that the participants were known to the author, in my opinion, is a flaw of epic proportions, but I also understand that as a graduate students, at times your ability to access the sample you need in the manner you'd prefer can be limited.  In that sense, the problem is with the publisher, not the author.  Anyway, let's consider some of the findings themselves quickly.

It does appear, perhaps not surprisingly, that therapists who favor different orientations differ with respect to their personality.  To some degree, I find myself frustrated with the options available regarding orientation though.  Most of the psychologists I know would not indicate that they follow one particular treatment, but rather that they follow what empirical data indicates is the most effective treatment(s) for specific diagnoses.  In other words, sure, I have used CBT in the treatment of binge eating disorder, but I have also used dialectical behavior therapy in the treatment of borderline personality disorder and would gladly use interpersonal psychotherapy in the treatment of depression.  It seems than that the structure of the question made it impossible for some scientifically-minded psychologists to accurately answer. 

The findings regarding training influences were interesting, and seem to indicate that a psychodynamic orientation corresponds to having coursework on that topic and a lack of exposure to CBT, but the cross-sectional nature of the data render the findings a bit difficult to interpret.  In other words, because we only have one time point to consider, we have no idea whether participants' orientations actually changed in response to training and whether training was capable of shifting an individual's orientation from one to another.  Given that students generally have a lot of influence over which graduate programs they attend and which courses they take, it is difficult to know whether training truly had any influence or if people placed themselves in training situations that were consistent with beliefs they already held.  Buckman and Barker (2010) also noted another important issue: because the entire study was based upon self-report questionnaires, we have no idea the degree to which the therapists actually implement the forms of therapy they claimed to espouse (or how well).

Anyway, while I do have some fairly strong issues with the study itself, I'm glad to have read it and think it makes for an interesting conversation piece.  The authors also examined systemic therapy in this study, but for the sake of brevity, I kept the conversation focused on CBT and psychodynamic therapy.  Do you see any particular strengths and/or weaknesses with these data?  Do you think that having a "therapeutic orientation" is a good idea if data indicates that different treatments produce better results for different diagnoses?  How might we be able to improve upon these findings?  Is it a good or a bad thing that personality type influences therapeutic orientation given that data does not support the notion that different treatments are always equally effective for specific diagnoses?

************

If you would like to learn more about this or other topics discussed on PBB, we recommend that you consult our online store for scientifically-based psychological resources.

Mike Anestis is a doctoral candidate in the clinical psychology department at Florida State University and an incoming resident at the University of Mississippi Medical Center.

Psychopathy and risky sexual behavior

by Michael D. Anestis, M.S.

I'm a big fan of research that looks into aspects of our personalities that make us more or less likely to engage in particular types of behaviors.  This form of research gives us clues regarding the mechanisms driving our decision making and offers clues regarding the most effective ways to intervene when behaviors become a problem or to prevent the problems in the first place.  A common misconception many people have when they hear about this type of work is the belief that scientists are proclaiming that certain individuals are fated to end up a certain way or that they had no control over their decision making.  In reality, this could not be further from the truth.  All this type of research is saying is that, to some degree, we are genetically predisposed towards seeing value in particular forms of behaviors.  For instance, individuals with high levels of negative urgency or low levels of distress tolerance, both of which we have discussed at length on PBB, see great value in behaviors that allow them to immediately diminish the intensity of negative emotions.  For them, removing such feeling is of such high value that any behavior capable of immediately offering that outcome (e.g., non-suicidal self-injury, binge eating and purging) will become substantially more likely, regardless of the long term consequences.  For other individuals, an entirely different set of incentives and appraisals are in play and, as such, a different constellation of behaviors in a completely disparate group of contexts are likely.

Today, I would like to discuss a study conducted by Jessica Fulton, David Marcus, and Kathleen Payne of the University of Southern Mississippi (woohoo for Hattiesburg!) that touches on a topic directly relevant to the above discussion.  The study, published in the most recent issue of Personality and Individual Differences (2010), looks at the role of different components of psychopathy in risky sexual behavior (RSB) for both men and women.  We have discussed psychopathy at length as well and I encourage you to click here for a description of what that idea entails.  RSB refers to any behavior that increases the likelihood of harmful outcomes associated with sexual contact.  Such outcomes include contraction of sexually transmitted diseases and unplanned pregnancies and such behaviors include having multiple partners, having unprotected sex, and having casual sex with strangers.

In a sample of 511 undergraduate, the authors found the relationship between psychopathy and RSBs is different for men versus women.  Even when they controlled for sensation seeking - the tendency to seek out activities that involve an element of risk or thrill - Factor 2 of psychopathy, which involves impulsivity, irresponsibility, and aggression, predicted RSBs in both men and women.  In other words, regardless of your sex, on average, the greater degree to which you exhibit components of Factor 2 psychopathy, the more often you will engage in RSBs.  Factor 1 of psychopathy, on the other hand, which involves superficial charm, shallow emotions, pathological lying, and a lack of empathy, guilt, and remorse, only predicted RSBs in men.

The authors anticipated that the data would follow this pattern.  Factor 2, they argued, would increase the opportunity for RSBs regardless of whether an individual is a man or a woman.  The more opportunities an individual has, the greater the odds that person will engage in the behavior.  Factor 1, however, paints a more complicated picture.  Men with the traits that characterize Factor 1, the authors argued, will be socially dominant and likely to pursue sexual gratification in whatever way possible.  Women with these traits, on the other hand, will also be socially dominant, but because these traits also share quite a bit with self-esteem, would also be equipped to refuse the efforts of others to make them engage in RSBs.  As a parallel, they noted that Tevendale, Lightfoot, and Slocum (2009) found that high self-esteem predicted low levels of RSBs in female homeless youth but was unrelated to RSBs in male homeless youth.  It seems that a fearless demeanor accompanied by a positive self-view and a willingness to manipulate others may lead men to view the world and engage with it in one way while women with those same traits might respond entirely differently.

Now, keep in mind that results like these do not apply to every single man or woman.  They represent general trends - our understanding of probability - and that there is variability in any relationship.  Also, keep in mind that, because this study involved only one time point, self-report questionnaires, and an undergraduate population, the work is not perfect.  Still, the information gathered through this work is useful and informative and helps us not only understand RSBs in general, but also why men and women might approach such behaviors differently.

************

If you would like to learn more about this or other topics discussed on PBB, please consult our online store for scientifically-based psychological resources.

Mike Anestis is a doctoral candidate in the clinical psychology department at Florida State University and an incoming resident at the University of Mississippi Medical Center.

PBB Guest Author Brian Thompson on MentalHelp.Net

by Michael D.Anestis, M.S.

PBB guest author Brian Thompson wrote an excellent article on psychotherapy and pseudoscience for the website www.mentalhelp.net.  In the article, Dr.Thompson explains in clear terms why scientific research is essential to our understanding and treatment of mental illness and how consumers can determine whether the treatment they are being offered is the most well-supported option for their particular condition.  Anytime somebody contributes to the cause of educating the public on the many well-supported, effective, affordable treatments for mental illnesses that are available, I find myself extremely pleased.  Well done, Brian!

Click here to read Dr.Thompson's article.

When nothing is something important: Why everyone should know about null findings

by Michael D. Anestis, M.S.

When you open up an article published in a peer reviewed psychology journal, you will encounter a detailed description of a study designed to test a set of hypotheses and a discussion of what worked out.  Sometimes, all hypotheses are supported by data.  Other times, certain hypotheses are supported and the authors discuss what did and did not work out according to plan.  What you will rarely ever see is a paper in which none of the hypotheses were supported by the data.  They are almost never published.

At first glance, this might seem reasonable.  After all, who would want to read about nothing?  That's the tricky issue however: a lack of significant findings is not "nothing."  In reality, when a study's hypotheses are not supported by data, we can actually learn as much as we do when everything works out the way we expected.  In other words, sometimes "nothing" is actually remarkably important.  One of the dangers of a field like clinical psychology is that we often have a philosophical belief that things are a certain way.  Our logic supporting that belief often seems sound and, as a result, we tend not to question our accuracy.

For example, some people might believe that a particular form of therapy is highly effective for a specific mental illness.  They might be able to think about people they know who felt better after using that therapeutic approach or explain why the therapy would theoretically be useful.  Without empirical evidence indicating that this is the case; however, we actually know nothing.  If somebody conducts a trial testing the efficacy of that therapy and it outperforms a waitlist, treatment as usual, or another well established treatment, those results will likely be published and will inform how we think about treatment.  If they do not work out, however, the chances are that none of us will ever see the results.  Journals simply do not often publish "null findings," findings that involve no statistically significant results.  The unfortunate result of this situation is that we do not know how many relationships we believe are significant have been shown not to be in studies that went unpublished.  This is not to say that we should not trust the results of published research.  Instead, what this means is that our knowledge is simply incomplete - that we only think about what works out and miss the rest of the picture.

One idea that I have heard people mention from time to time and which I support wholeheartedly is the need for a Journal of Null Findings.  This journal would be a home for studies in which the results contradicted hypotheses.  By publishing these results, we could theoretically end the "file drawer problem" in which non-significant results are never discussed and the general understanding of a phenomenon is therefore blunted.

Before we get ahead of ourselves here, though, let's discuss a number of reasons why null findings might, in the end, truly represent nothing rather than something important and informative:

***

Bad Design:
Sometimes, when a finding is shown to be non-significant, this is the result of a poorly designed study.  Perhaps the authors used a sample that did not involve enough of the outcome variable (or variability in the amount of the outcome variable) to test the relationship.  In other words, if we want to examine suicidal behavior, but we use an undergraduate sample that involves no history of suicide, we are unlikely to find the results that are truly representative of the phenomenon.  Perhaps the authors used measures that do not adequately capture the variable of interest.  Perhaps the authors did not administer a therapy or test according to proper protocol or even examine whether or not that was the case.

Underpowered Study:
We have discussed statistical power here before, but let me quickly remind you what I mean.  Imagine you flipped a coin five times and you got heads four times.  I suspect you'd find it a little odd, but assume it was likely due to chance.  If you flipped that same coin five hundred times and you got heads four hundred times, however, you'd likely believe that something was off with the coin (and chances are you'd be right).  The thing is, the proportion of times that you got heads was no different in the second example.  You're more certain that the results are legitimate because you had a large enough sample.  In other words, you had enough statistical power to detect a significant effect.  If you do not have a large enough sample - if you are underpowered - it is very difficult to end up with results that are statistically significant, even if the relationship you expected is real.  Remember, the coin was still off when we had only flipped it five times, but we were not yet confident that was the case.  This is a particularly big issue with treatment studies, which are expensive, time-consuming, and difficult to administer and, as such, often end up with very small samples that can understate the results.

Incorrect Analyses:
Think back to our discussions of Bruce Wampold's work and the dodo bird hypothesis.  In his studies, he concluded that all forms of psychotherapy work equally well for all diagnoses.  The problem with his conclusion, however, was that his analyses tested whether all treatments differ from all other treatments for all diagnoses.  Nobody believes this to be true.  Instead, what many psychologists believe (and what the data support as true) is that particular treatments are better than others for particular conditions.  Because Wampold's data did not support the ridiculous notion that everything is different than everything else for everything, he concluded that nothing is different for anything.  In other words, his null findings were the result of asking the wrong question and running the wrong analysis.  Now, in this case, the results were published, but point remains that sometimes we find no significant results because we made them nearly impossible to find.

Chance:
When we published research, we are often asked to explain why we are confident that our findings are not simply type 1 errors - false positives - but there is another equally important idea to consider: false negatives.  Every once in a while, you can produce a negative result simply because the sample you used was unusual.  Unfortunately, we can not really guard against this possibility, which is one of the many reasons we need to consider results in the context of other studies that have examined the same point.  Over time, if a false negative (or positive) is present, this will become apparent because none of the other studies will produce similar results.

***

These are only a few of the reasons why null findings can quite often truly represent nothing of great value.  In order for a Journal of Null Findings to be a useful tool for the scientific community, great care would need to be taken to ensure that included studies are well-designed, well-run, sufficiently powered, analyzed correctly, etc...The goal would not be to publicize bad work, but rather to publicize all good and informative work, regardless of whether or not that work resulted in rejection of the null hypothesis and support for the authors' hypotheses.

What do you think about this?  Do we need to know more about the studies that don't work out in order to contextualize those that do?  Would a Journal of Null Findings be too hard to pull off effectively and ultimately either confuse things or grant a voice to researchers with an agenda to fight the legitimacy of empirical findings not to their liking?

************

If you would like to learn more about this or other topics discussed on PBB, we hope you will browse our online store for scientifically-based psychological resources.

Mike Anestis is a doctoral candidate in the clinical psychology department at Florida State University and an incoming resident at the University of Mississippi Medical Center

PBB Guest Author Craig Bryan in TIME Magazine

by Michael D. Anestis, M.S.

PBB guest author Craig Bryan is quote in a strong article in TIME magazine on suicide in the military.  I highly encourage readers to check out the article by clicking here.

Craig does everyone a service by discussing the idea of the acquired capability for suicide in a media outlet with a broad reach.  The more ideas like that are disseminated to a wide audience, the sooner the world will come to accurately understand the phenomenon of suicide.  Education on a large scale requires that researchers connect with the public through mass media.  Any time a researcher pulls it off like this, it makes me remarkably happy.

Well done, Craig!

Physical activity and depression in female adolescents

by Michael D. Anestis, M.S.

This morning, I came across a very interesting study published in the current issue of the Journal of Consulting and Clinical Psychology by Sarah Jerstad, Kerri Boutelle, Kirsten Ness, and Eric Stice (2010).  Jerstad and her colleagues presented data examining the degree to which physical activity impacts depression vulnerability and depression impacts subsequent physical activity in female adolescents.

The authors noted that a number of previous studies have found a negative correlation between physical activity and depression, meaning that higher levels of depression are generally accompanied by lower levels of physical activity.  Additionally, Jerstad et al (2010) pointed out that psychologists have put forth a number of theories as to why this is the case.  O'Neal, Dunn, and Martinsen (2000) suggested that physical activity might cause an increase in neurotransmitters known as monoamines (e.g., serotonin, norepinephrine, dopamine), which then causes the individual to experience more positive emotions and a lower risk for depression.  Just and Alloy (1997) suggested that physical activity might distract individuals from their negative emotions and thus prevent them from engaging in destructive thinking styles (e.g., rumination) that would otherwise increase depression risk.  Ekeland, Heian, and Hagen (2005) argued that physical activity has been associated with higher levels of self-esteem, which could serve as a protective factor against depression.  These theories are not mutually exclusive - each of them could contribute to the relationship simultaneously - but the data currently available cannot inform us as to which might have the greatest impact.  What's more, cross-sectional data - data collected all at once instead of involving multiple measures over time - cannot even provide any insight into whether decreased physical activity causes higher levels of depression or whether increased depression causes a decrease in physical activity.  Along those lines, certain symptoms of depression such as anhedonia, motor retardation, and decreased energy could theoretically cause individuals to stop engaging in physical activities.

In an effort to clarify these issues, the authors conducted a study in which 496 adolescent females were assessed on physical activity levels, depression symptoms, and a host of other important variables every year for six years.  They chose to focus on females because, beginning in adolescence, females experience depression two to three times as often as men.  They also noted that one earlier study conducted by Birkeland, Torsheim, and Wold (2009) failed to find any relationship between physical activity and depression levels in adolescents over a 10 year period; however, that study used a short self-report measure of depression, which decreases confidence in its findings.  Jerstad et al (2010) hypothesized that low levels of physical activity would predict subsequent increases in depression, even when controlling for the impact of body dissatisfaction, bulimic symptoms, social support, and body mass.  Additionally, they hypothesized that high levels of depression would predict subsequent decreases in physical activity controlling for those same covariates.  Importantly, both depression symptoms and bulimic symptoms were assessed through the use of semi-structured diagnostic interviews.

Overall, 13% of their sample experienced major or minor depression.  The more physically active an individual was at baseline, the less increase in depression symptoms she experienced over the course of the study.  Specifically, for each additional activity that a participant engaged in 10 or more times in a year, risk of later depression symptoms decreased by 1%, risk for major or minor depression decreased by 8%, and risk for major depression decreased by 16%.  This particular set of results supports the notion that physical activity can, in fact, reduce an individual's vulnerability to varying levels of depression in the future. The degree to which physical activity reduced depression risk was modest, meaning that being active is by no means a cure all, but significant mental health benefits are something to take seriously even when the effect size is not large.

Additionally, the authors found that participants who met criteria for major depression were 35% less likely to participate in an additional physical activity 10 or more times in the next year.  Participants who met criteria for minor depression (5 symptoms of depression, at least one of which is at sub-clinical levels) were 18% less likely to participate in an additional physical activity 10 or more times in the next year.  Increases in one depression symptom over a year diminished the likelihood of engaging in an additional physical activity in the subsequent year by 21%.  These findings support the notion that becoming more depressed will impact physical activity in the following year.

Taken together, these findings present us with a scenario in which becoming less physically active makes people more vulnerable to depression and developing more depression makes an individual less likely to become more physically active.  In this sense, a vicious cycle can develop in which one problem leads to the other which leads back to the first problem and so on.  As always, this study is not perfect and readers should be careful not to jump to conclusions about the importance of physical activity in depression.  That being said, the authors should be applauded for developing a methodologically strong study that provided us with data capable of actually addressing the issue of activity and depression.  So what is the take home message?  These data seem to indicate that keeping adolescent girls active is a healthy and potentially fun way to reduce vulnerability to depression.  Physical activity in and of itself will not make an individual immune to depression nor will it serve as a stand alone treatment for those who become depressed, but it is an affordable way to help decrease the likelihood of a young girl struggling with depression.

************

If you would like to learn more about this or other issues discussed on PBB, we hope you will consult our online store for scientifically-based psychological resources.

Mike Anestis is a doctoral candidate in the clinical psychology department at Florida State University and an incoming resident at the University of Mississippi Medical Center

Mobile crisis units: Is there a benefit above and beyond standard law enforcement interventions?

by Michael D. Anestis, M.S.

Some reports estimate that 7-10% of calls made to police departments involve mental illness (Deane, Steadman, Borum, Veysey, & Morrissey, 1999).  Clearly, this means that police departments expend a significant amount of time and resources attempting to manage mental health crises.  These situations not only represent a burden upon law enforcement, however; such situations at times result in arrests and conflict when effective psychological care is a more optimal outcome.  That being said, mobile crisis units (MCUs) have become an increasingly common resource in communities worldwide.  MCUs differ from region to region in terms of hours of availability, speed of services, and the make-up of the team, but they all serve the purpose of attempting to decrease the likelihood that mentally ill individuals will rely on the police and emergency rooms by offering focused psychological care capable of diffusing a crisis and motivating an individual to seek additional community care.

Unfortunately, there is shockingly little research that actually tests the degree to which MCUs offer incremental benefits above and beyond standard care.  Today, I wold like to quickly review one example of research that does, in fact, support their utility.

In 2000, Roger Scott published a study in Psychiatric Services in which he reported data comparing the results of 73 911 calls handled by MCUs and 58 911 calls handled by standard police interventions.  The results of his study, while by no means enough to conclusively support the notion that MCUs are an essential component of psychiatric care, were quite compelling.  The primary variable in this study was the percentage of 911 calls that did not result in hospitalization.  Results indicated that 55% of MCU calls did not result in hospitalization as compared to 28% of standard police calls, a highly significant finding.  Of those who were hospitalized, only 36% of MCU cases were involuntary while 67% of hospitalizations in police cases were made on an involuntary basis.  These results alone are quite important.  First of all, it appears that MCUs are more effective at keeping mentally ill individuals out of expensive inpatient units where beds are limited.  Secondly, when hospitalization occurs in MCU cases, the situation is much more likely to be amiable, with the individual making the decision him or herself to seek impatient care.

The findings from Scott (2000) extend further.  The author also compared the economic impact of the two forms of intervention and found that MCU cases, on average, cost 23% less than police cases, with MCU cases resulting an an average cost of $1,520 and police cases resulting in an average cost of $1,963.

Scott (2000) considered one final outcome variable: perceptions of MCUs by both consumers and police.  The data indicated that both the police and the consumers rated the MCUs highly,indicating that they are not only effective, but also well received.

So what can we take from these findings?  First of all, substantially more empirical research needs to be conducted investigating the effectiveness of MCUs in their various forms.  The Scott (2000) study is by no means that only such study, but there are too few given the proliferation of these units.  That being said, the results here are promising, as they indicate that MCUs are less expensive than standard police interventions, less likely to result in inpatient hospitalization, and very well received both by the police and the consumers.  In other words, a form of intervention that is palatable to all involved exists and it costs less than and achieves better outcomes than standard police interventions.  Should these results generalize across samples, we would thus have ample justification for implementing MCUs on a wider scale, thereby freeing up the police to focus on cases more directly relevant to their own goals.  In other words, everyone wins.

I would love to hear from readers who have worked on or with mobile crisis units.  Did you notice any particularly important assets they bring to the table?  Were there any specific problems that you think might interfere with their effectiveness?  Is there an alternative approach that is better than either of the options outlined in this post?

************

If you would like to learn more about the topics discussed on PBB, we encourage you to explore our online store for scientifically-based psychological resources.

Mike Anestis is a doctoral candidate in the clinical psychology department at Florida State University and an incoming resident at the University of Mississippi Medical Center

"Pulling together" and suicide: Some thoughts on baseball's Opening Day

by Michael D. Anestis, M.S.

In celebration of baseball's Opening Day - and as a Pirates fan, today is Day 1 - I want to take a quick look at some research that has considered sports and mental health.  Specifically, I would like to review a paper published by several of my colleagues - Thomas Joiner, Daniel Hollar, and PBB guest author Kim Van Orden - in the Journal of Social and Clinical Psychology (2006).  In an effort to demonstrate the importance of shared positive experiences in reducing suicide risk, Joiner and colleagues (2006) conducted three small studies.  In each, their goal was to examine the degree to which suicide rates could be predicted by sporting events around which communities "pulled together."

In study 1, the authors wanted to look at the relationship between national rankings of college football teams and local suicide rates.  Joiner and colleagues hypothesized that, in communities in which college football was "the only game in town," meaning that there were not other similar forms of entertainment (e.g., professional sports teams), local suicide rates would be lower in years in which the local college football team achieved a higher national ranking.  To test this, they examined the relationship between suicide rates and national rankings between 1990 and 2002 for Franklin County, Ohio - home of Ohio State University - and Alachua County, Florida - home of the University of Florida.  In an effort to demonstrate that the relationship would be stronger in areas in which college football is the "only game in town," they also examined the relationship between suicide rates and national rankings in Miami-Dade County, Florida - home of the University of Miami as well as a number of other forms of similar entertainment.  In years in which any of the teams was unranked, a ranking of 26 was entered (which is actually likely to suppress any effect, as the actual "ranking" in a poor season might be substantially higher than that).  

Overall, the correlation between suicide rates and national rankings when the two teams were considered together was .55, which was statistically significant (p < .01).  When they controlled for US Gross Domestic Product (GDP); however, the relationship was no longer significant (p = .072).  When the two counties were considered separately, the correlation between local suicide rates and national football rankings was significant for the University of Florida (r = .66, p < .05), but not for Ohio State University (r = .4, p > .10).  These results indicate that national ranking and suicide rates in general are correlated, but that this impact might be better accounted for by the economic situation at the moment.  When considered on the level of individual teams, however, the relationship stands on its own in the smaller area - Gainesville, Florida - indicating that when the area is truly saturated by the team, the impact of "pulling together" around the team might be more pronounced.  The relationship between suicide rates and national ranking was not significant in Miami.  Importantly, the authors were looking at suicide rates for the entire year even though college football season comprises only a small portion of the year.  As such, any relationship between these variables is actually fairly remarkable.

The findings from this first study were admittedly a bit mixed, so Joiner and colleagues (2006) conducted two other studies to enhance their confidence.  In Study 2, the authors looked at the number of suicides in the United States on February 22, 1980 - the day of the "Miracle on Ice" - relative to February 22 of other years during the 1970s and 1980s.  The authors theorized that, in response to the sense of community the developed around the United States hockey team's victory of the Soviet Union in the Olympic Games, deaths by suicide would be reduced, thereby highlighting the importance of feeling as though we belong to a group.  The results of this study supported the authors' hypotheses, with the February 22, 1980 suicide total proving to be lower than that of any other year on the same day.  This was true even in the 1970s, when the US population was smaller (thereby theoretically lowering the number of suicides without lowering the actual suicide rate).  One possible response to these findings could be that individuals became less likely to die by suicide due to "vicarious efficacy," meaning that they felt better about themselves due to the performance of somebody else.  To control for this possibility, the authors designed Study 3 to examine a phenomenon for which this is not a legitimate interpretation.

In Study 3, Joiner and colleagues (2006) examined suicide rates on Super Bowl Sunday from 1984 on, relative to the Sundays immediately preceding and following Super Bowl Sunday.  The authors chose to begin their analyses in the mid-1980s because, at that point, substantial advertising became more prominent and the Super Bowl itself became a much larger phenomenon.  As it turns out, there was, in fact, a time by day interaction, meaning that fewer suicides occurred on Super Bowl Sundays than on comparison Sundays during the same years.

So what can we conclude from these studies?  Obviously, the answer is not simply that everyone needs to follow sports and hope that their team does well.  Instead, these studies provide supportive (although by no means conclusive) evidence that, when people can pull together around a shared event, they feel connected to their community.  A host of variables need to be considered when thinking about suicide risk and we have covered a number of them on PBB, but there is a message to take home from this work: when people can feel connected to others, even through a distant event, their risk levels decrease.  So, as you sit down and watch baseball today and wonder whether rooting passionately for a team is silly, think about the times that you have shared with others directly as a result of being a fan and know that those experiences are, in fact, meaningful.

Enjoy the day everyone - even the Pirates are in contention right now, so reasons for optimism abound.

************

If you would like to learn more about the topics covered on PBB, please consult our online store for scientifically-based psychological resources.

Mike Anestis is a doctoral candidate in the clinical psychology department at Florida State University and an incoming resident at the University of Mississippi Medical Center.

Guns and suicide

by Michael D. Anestis, M.S.

Today's topic is a bit of a sensitive one, as it touches upon a controversial topic that tends to inspire rather strong emotional responses on both sides: gun ownership.  That being said, it is not my goal to have a heated political discussion or to point out the righteousness of one side or the other.  Instead, my intention is to discuss some sobering facts and ask you about your thoughts on the matter.  The following numbers were obtained in large part from a review of the literature in a study published in the most recent issue of Suicide and Life Threatening Behavior by Karen Slovak and Thomas Brewer.

• In 2006, firearms were involved in 47.2% of deaths by suicide in individuals between 15 and 24 years of age and in 72.7% of deaths by suicide in adults over the age of 65 (Centers for Disease Control and Prevention [CDC], 2009).
• In areas in which there are more guns, there are more deaths by suicide (Miller, 2005)
• There are approximately three times as many guns in rural areas as there are in urban areas (Connor, 2005) and suicide rates are higher in rural areas (Branas, Nance, Elliot, Richmond, & Schwab, 2004).
• The presence of a firearm in the home increases suicide risk even when controlling for poverty, urbanization, unemployment, substance dependence and abuse, and mental illness (Miller, Azrael, Hepburn, Hemenway, & Lippman, 2006).

The above numbers paint a fairly clear picture.  Suicide rates are higher in areas in which there are more guns and the predictive utility of gun ownership with respect to suicide is significant above and beyond important variables including whether or not an individual is known to meet criteria for a mental illness.

Part of the reason why these numbers are so powerful is the lethality of guns as a method.  As noted in the Slovak and Brewer (2010) study, guns result in death far more often than other methods.  For instance, one study demonstrated that, while poisoning was involved in 74% of attempts, it only had a case fatality rate of 2%.  Guns, on the other hand, were involved in only 5% of attempts, but had a case fatality rate of 91%.  In other words, guns are not used as frequently as other methods, but when they are used, they almost always result in death.

Given these facts, many psychologists have advocated for gun restriction as an effective means for decreasing suicide rates.  Not surprisingly given the controversial nature of this topic, this approach has been far from universally accepted.  In fact, legislation was proposed in Virginia and West Virginia that would have prevented clinicians from asking patients about gun ownership, possession, or storage.  This legislative effort failed; however, it highlighted the degree to which political positions and empirical data are, at times, at odds with one another.  

Opposition to restricting gun access is not purely political, however.  Some believe that if guns were made less accessible, individuals would simply choose another method, thereby negating any gains made by infringing upon what is considered a constitutional right within the United States (this same argument is often made by individuals opposed to building higher fences along the Golden Gate Bridge, which is the location with the greatest number of suicides worldwide).  There have certainly been some studies that have demonstrated that restricting guns resulted in an immediate jump in deaths by other means that at least partially negated the decrease in suicide deaths by firearms; however, these studies have been very short term examinations that did not include a look at the bigger, longer term picture (I recommend reading Conwell et al., [2002] for a thorough review of the literature on this topic).  For instance, Carrington and Moyer (1994) found that suicide rates decreased significantly over 12 years after stricter gun control legislation was implemented in Canada.  After gun legislation was implemented in Washington, DC in 1968, an immediate and sustained decrease in suicide deaths occurred with no subsequent increase in suicide deaths by other means (Loftin et al., 1991).  A similar pattern emerged in Australia (Rich et al., 1990).

So, the above findings present us with another interesting set of facts:

• Guns are a highly lethal method and account for a disproportionately high number of deaths by suicide given the proportion of attempts involving this method
• Legislation that restricts access to guns does appear, in large part, to decrease suicide rates over a sustained period
• Gun restriction does not generally result in an increase in deaths by suicide through other means that would negate gains made through the legislation

Before turning the conversation over to you, I wanted to quickly present the actual findings of the Slovak and Brewer (2010) study that I mentioned above.  The authors were interested in learning how social workers in various areas within the United States feel about assessing for firearm ownership and counseling clients about the risks.  Here is what they found:

• Social workers who had received training in how to make such assessments and counseling efforts felt more favorably towards using such interventions
• Social workers in urban areas felt more favorable towards such assessments and counseling than did social workers in rural areas.

Now, the first finding could simply reflect that people who already feel favorably towards this approach are more likely to seek training in the method.  The second finding, however, is a bit more interesting as gun-based deaths by suicide are significantly more common in rural areas than in urban areas (homicide rates present an opposite pattern).

All of this leaves us with a tricky situation.  The fact of the matter is that gun ownership increases suicide rates above and beyond the impact of important risk factors like poverty and mental illness.  This presents a particularly troubling issue in rural areas, where guns are more prevalent and clinicians appear to be less willing to assess for their presence.  That being said, gun ownership is a constitutionally protected right, meaning that implementing gun control legislation is going to upset a large number of people. 

I suspect that this conversation could quickly turn ugly, so I will be monitoring comments more closely than normal and will have no reservation in deleting comments likely to prove offensive.  That being said, I am interested in hearing calm and cogent thoughts on how we can resolve this situation.  Is there a way to restrict access to guns as a method for suicide without offending a sizable portion of the country?  What is the best way to navigate this situation?

************

If you would like to learn more about this or other topics discussed on PBB, we encourage you to consult our online store for scientifically-based psychological resources.

Mike Anestis is a doctoral candidate in the clinical psychology department at Florida State University and an incoming resident at the University of Mississippi Medical Center.