Lecture Notes for this Week
Lecture Notes for this Week 76884
Popular in PSYC4201W
verified elite notetaker
Popular in Psychlogy
This 17 page Class Notes was uploaded by Rachel Onefater on Thursday March 31, 2016. The Class Notes belongs to 76884 at George Washington University taught by Dr. George Howe in Spring 2016. Since its upload, it has received 11 views. For similar materials see PSYC4201W in Psychlogy at George Washington University.
Reviews for Lecture Notes for this Week
Report this Material
What is Karma?
Karma is the currency of StudySoup.
Date Created: 03/31/16
Trial Trials: a research method for testing the impact of interventionsa method for testing a program or specific prevention Prevention Programs Treatment Programs Goals of Intervention Trials •Test efficacy: does the program lead to the desired change? –Prevention: reduction in future prevalence of some physical or mental health problem (i.e. may not be able to test it right away because you want to see if it has long term effects.) –Treatment: resolution or cure of an existing problemDo want to take a look at immediate effects of the treatment and/or longterm cure •Test comparative efficacy: if we have two programs for the same condition, which one works better?More focused Trials (Ex: Antidepressant medication has a stronger or weaker impact on depression) •Test differential effica: Does a program work better for some people than for others? •Testing mechanism of effect: Does the program work by changing specific targets (risk, protective, maintenance mechanisms)? Ex: Someone in the midst of a major depressive episode, and what are the things that are maintaining it, and if you do change those things, does that change the effect of the depression? Claims of Efficacy Without Empirical Test ● Before FTC ruling: Peterson and Fung Washington Post Articlereporting on Federal Trade Commission(FTC) Lumosity website makes brain training programs aan in their review, there isn't a lot of evidence that was being claimed to have in the ads. An example of this ● After FTC ruling: After to fine was placed by the FTC, they could not claim that they “Build a better brain..sharper, remember things, [and] learn faster,” but now use “science into delightful brain games”. Causal Inference and Trial Designs •Pretest/ posttest without control group •Pretest/posttest with nonrandomized matched control group •Pretest/posttest with randomized control group Pretest/Posttest Without Control Group •Trials that evaluate change in target or outcome after intervention •But have no comparison group Ex: develop intervention for school system and collect data in the beginning and end of yearstrong at math or mindfulness skills, but you don’t have comparison group Example: Does CBT Change Schemas? → Shimotsu et al (2014) Effectiveness of group cognitivebehavioral therapy in reducing selfstigma in Japanese psychiatric patients. Asian Journal of Psychiatry, 10, 3944. → Group cognitive behavior therapy with 46 Japanese outpatient adults to change schemas → Administered dysfunctional attitudes scale before and after Findings → Reductions in three subscales from pre to post Problems with Pretest/Posttest without Control ● Very weak design ● Cannot rule out alternative causes of naturally occurring change such as: ○ Participants seeking out other treatments or resources during the study ○ Natural recovery in highly symptomatic groups ○ Development Nonrandomized Comparison Group Design •Compare those receiving intervention to other existing groups that do not receive intervention •Select comparison groups to be as similar as possibleto make all other causes equal •But no random assignment to the two conditions Example: Does mindfulness programs for teens reduce negative affect? → Broderick & Metz (2009) Learning to BREATHE: A pilot trial of a mi ndfulness curriculum for adolescents. → Intervention group: 104 seniors in private Catholic school → Comparison group: 17 Juniors in same school → Added mindfulness training to health class (42 sessions) for entire senior class (n=120) → Intervention group showed greater: reductions in negative affect Increases in sense of calmness Problems with nonrandomized comparison group design •Cannot rule out alternative causes thatdiffe between the two groups: –Selection effects (very few juniors opted to provide data) –Naturally occurring changes (juniors may have increased demands compared to seniors) –Development : seniors may be in different developmental period when more change occurs •Can be improved by assessing possible confounds to determine if groups differ on them –In this study, no differences on current or past meditation or yoga practice Randomized Group Comparison Designs •For treatment, referred to as randomized clinical trials (RCT) •Strengths of random assignment: –Greatly increases likelihood that groups are equivalent on plausible confounds, even when those are unknown •Challenges: –May not be able to have control over assignment process Example: Can an intervention program for children of divorce reduce future disorder? → Wolchik et al. (2000)An experimental evaluation of theorybased mother and motherchild programs for children of divorce. → Identified recently divorced families with children aged 812 → Randomly assigned 240 families to: → Group program for custodial mothers → Program for custodial mothers plus group program for children → Selfstudy program with three books for mother and for child Success of Randomization? ● Compared groups on range of pretest measures ○ Demographics ○ Proximal Targets ○ Mental Health Variables ● Only one significant difference: ○ Interparental conflict lower in selfstudy than in combined condition Initial Preventive Effects? ● No effects on child internalizing symptoms (depression, anxiety) at posttest or 6 months ● For externalizing, mother program families different from self study at posttest, 6 month followup Longterm preventative Effects? → Followed up at 6 and 15 years → Combining parents and parent +Child Groups → Rates of any diagnosable mental or emotional disorder (1 year prevalence) substantially reduced: In RCT, what if those assigned to intervention fail to attend? •Possible solutions –Drop them from the study –Replace them with someone new •Problem: These can lead to selection effects –Those who fail to attend may differ on plausible causes of outcome –If we leave them out, there will be fewer in the intervention group than in the control group Solution •Intenttotreat design –Keep random sampling the same –Collect outcome data on everyone, including those who fail to attend intervention –Analyze complete sample •Strengths –Eliminates this source of selection bias, maintaining randomization •Weaknesses –Effects could be diluted because not everyone gets full “dose” of intervention –May therefore require larger samples to detect effect Control Conditions: What are we Controlling For? Designing Both Factual and Counterfactual •If the experimental condition is the factual condition (the cause is present) •Then the control condition is simulating what would have happened if the cause had not been present (thecounterfactual) •So, what is the cause What is present in one condition and absent in the other? •Remembering: all other plausible causes should be equal (across the two groups). Different causes call for different factual/counterfactual conditions •What cause are we interested in? •Some possibilities: –The treatment as a whole –Treatment, but not messages about likely effectiveness –Treatment, but not general social contact and support –Treatment, but not ny other treatments or services Cause of Interest: Treatment as a Whole •Experimental condition: the treatment as a wholeis the cause, including –Specific intervention content (often in specific sessions) –Homework assigned –Messages about the likely impact –The amount and type of social contact (group?, individual therapist?) –Interactions with other participants, including family members •Control condition: relevant counterfactual ino treatment –Control group is offered nothing –Differs on exposure to all elements of the intervention Example: SchoolBased Program to prevent Depression •Gilham et al (2007) SchoolBased Prevention of Depressive Symptoms: A Randomized Controlled Study of the Effectiveness and Specificity of the Penn Resiliency Program. • Randomly assigned middle school children in each of three schools to one of three conditions –Penn Resiliency Program: Focused on cognitive schemas seen as increasing risk for depression –Penn Enhancement Program: Focused on stress management –No intervention control: received no form of intervention •Results: –No group differences in depressive symptoms at posttest or 3 year followup WaitList Control •Variant of no treatment control –Half the sample randomly assigned to wait for treatment –Collect pretest, posttest on full sample during waitlist period •Strengths: –May be more attractive to participants –People consider lotteries for limited resources to be fair, and so may be able to get permission for random assignment more easily •Weaknesses: –Cannot conduct longerterm followup within randomized framework Cause of Interest: Excluding Messages of Likely Effectiveness •Experimental condition: the treatment as a whole, which includes messages that increase expectations of success –Based on evidence that messages increasing hope can themselves have an impact on emotional (and perhaps physical) outcomes –These are an indivisible element in many interventions •Control condition: relevant counterfactual is being exposed only to messages that increase expectations of success (also known as placebo control) –Differs on exposure to all elements of the intervention –Similar on messages of likely effectiveness Example: Medication Trial •Ghanizadeh & Hedayati (2013). Augmentation of fluoxetine with lovastatin for treating major depressive disorder, a randomized doubleblind placebo controlledclinical trial. • Tested whether adding lovastatin (a common medication for reducing cholesterol) to fluoxetine (a common antidepressant) would increase antidepressant effects •Randomly assigned depressed patients to two groups: –Antidepressant with lovastatin –Antidepressant with placebo pill •Doubleblinded (psychiatrist, research staff, patients did not know whether they were receiving lovastatin or placebo) Findings → Both Groups showed Improvement → Those with Lovastatin Improved Faster Cause of Interest: Treatment as a Whole, Excluding General Social Contact and Support •Experimental condition: the treatment as a whole, which includes general social contact and support –Based on evidence that social contact and support by themselves can influence emotional outcomes –These are an indivisible element in many interventions •Control condition: relevant counterfactual is being exposed only to equivalent amount of social contact and support (sometimes called nonspecific factor –Differs on exposure to all elements of the intervention –Similar on amount of contact and support Example: Treatment for PTSD •Foa et al (1991) Treatment of posttraumatic stress disorder in rape victims: A comparison between cognitivebehavioral procedures and counseling. •Assigned rape victims with PTSD to one of four conditions: –Stress inoculation training (breathing exercises, coping, cognitive restructuring) –Prolonged exposure (reliving the scene in imagination, in vivo exposure to avoided contexts) –Supportive counseling –Wait list control •The supportive counseling condition was included to control for nonspecific social support Findings → Posttest: All groups showed improvement → Stress Inoculation Strongest Posttest → Prolonged exposure strongest at 4month followup Cause of Interest: Treatment as a Whole, Excluding Any Others Treatments or Services •Experimental condition: the treatment as a whole, plus other treatments or services –Many participants may seek out or be involved in other services or treatments –Our interest is in whether the new treatment has an impact above and beyond such treatments •Control condition: relevant counterfactual is being exposed only to equivalent amount of other treatments or services:treatmentasusual (TAU) control –Potential confound here: assignment to experimental condition may reduce use of TAU services –If this happens, experimental group differs in two ways (more experimental treatment, less usual treatment) Example: Prevention of Adolescent Depression •Clarke et al (1993). Targeted prevention of unipolar depressive disorder in an atrisk sample of high school adolescents: A randomized trial of group cognitive intervention. • Schoolwide screening of adolescents, identifying those with elevated but subclinical symptoms. •Randomly assigned to –Prevention program plus usual care –Usual care •Assessed use of other treatment over study: –Groups were equivalent on treatment from psychotherapist, antidepressant medication, hospitalization •Prevention program: CBT group Findings •Posttest: –Significantly lower rates of new onset of major depressive disorder in prevention group over 12 months Comparative Efficacy •Testing whether one intervention is better than another •Sometimes referred to as “headtohead” designs •Random assignment to two different interventions –With control group –With no control group Example: Head to Head with Control •Foa et al. (1991) study of treatment for PTSD in rape victims •Random assignment to three intervention and one wait list control group •Allows tests of: –Individual efficacy (each intervention compared to control) –Comparative efficacy (each intervention versus each other intervention) •Comparative efficacy findings: –Stress inoculation had strongest effects at posttest –Prolonged exposure had strongest effects at followup Head to Head Without Control •Design: random assignment to either of two interventions, but no control group •Problem: even if treatment A is better than treatment B, cannot distinguish from: –Treatment A is not effective, and treatment B leads to significant negative impact –Treatment A has some negative impact, and treatment B has even more negative impact •Recommended only when there are enough prior trials demonstrating the both treatments show effects when compared to control Summary •Randomized trials are the gold standard for testing efficacy of interventions •Choice of control groups will depend on rival hypotheses concerning how the intervention has its effect •Comparative efficacy requires direct comparison of interventions Preventing and treating relationship issues InterventionsCan we prevent or treat issues regarding to social relationships • Given that social forces and relationships are strongly associated with stress, can we create interventions that: – Prevent stress reactions through reducing social risk mechanisms – Prevent stress reactions through increasing protective social mechanisms – Treat stressrelated disorders through changing social mechanisms that maintain them Three examples across development • Interventions that target; – Early childhood factors – Factors during adolescence – Intimate partner relationships(def.): marriage, coupling of any kind that involves longterm commitment and connection Risk in infancy and early childhood • David Olds→ Pediatrician Colorado Health Center, and one of the first people to dig into this particular area of risk factors. This work began in the late 1970’s – Noted common risk factors in early development – Young parents(19 or younger) – Single parenthood – First child – Low SES Nursefamily partnership program → If we can get into the homes of women who are at risk, we can focus on biological factors such as smoking, but also relationships factors(i.e. How the parent learns to relate to their infant or child) • The nurse family partnership (NFP) program: – Uses service systems to identify young, single, low SES women pregnant with their first child Explanation: First Child because it is easier to work on the risk with the first pregnancies because after that, women are susceptible for higher risk in later pregnancies – Nurses visit during pregnancy, and regularly for the first year or two after birth • Nurses trained to: – Counsel on tobacco and substance use during pregnancy – Work with mothers on establishing sensitive and responsive parenting – Help mother with their personal development (planning future pregnancies, finishing school, finding employment, childcare) Randomized field trial • Olds et al (1986). Preventing child abuse and neglect:A randomized trial of nurse home visitation. • Small urban area in northern New York • Randomized 400 women to two different interventions – Prenatal screening services – Screening services plus nurse visitation program • Collected data at posttest, and continued to follow for 15 years Findings • Over first four years, NFP group showed: – Significantly fewer subsequent pregnancies – Greater participation in work force Tracked interactions with childprotective services, and did not find any effects here Findings • Over 15 years: NFP group showed: – Lower rates of child protective services reports of abuse or neglect *NOTE: If neglect does exists, it does start at about age 5(age at first neglect) Zielinski et al (2009) Highestrisk group • More pronounced differences in highest risk group (lowest SES, single parent) Zielinski et al (2009) → 14% rate of neglect in the intervention group and 35% of Mothers was reported for neglect in the control group. → Eternalizing Disorder like ADHD, Social Conflict and/or aggressiveness in children. Early intervention can be really helpful → Did a tremendous amount of work to see if the staff was trained in the way that they needed to be, and if they were actually doing what they were supposed to be doing. Family interventions with adolescents • Guillermo Prado, Hilda Pantin • Focused on families of Latino/a adolescents in Miami • Recent immigrant families seen as facing three risks: – Acculturation and discrimination stress (def.): They had to figure out the new landscape and come to understand the American culture, and that culture had a fair amount of discrimination built in. – Parentadolescent conflict, often around different acculturation → Kids are better at parents at learning thingswe have immigrated into a new culture and a new generation. – Negotiating connections with the wider community (school, justice system, work) Familias Unidas program • Family and parent focused intervention that evolved across several studies, emphasizing: – Educating parents about American culture and promoting biculturalism (Szapocznik et al., 1984, 1986) – Fostering family communication skills and reducing conflict (Szapocznik, Santisteban et al., 1989b) – Fostering family connections with other systems, and reducing parent isolation (Mancilla et al., 2002). • All programs showed efficacy in reducing adolescent substance use • Current Familias Unidas program integrates all three intervention components Questions about FU • Does the FU intervention also influence stressrelated problems such as anxiety and depression? • Moderator hypothesis : Does it have different impact for different families? There may be a cause that is more relevant with some groups than with others. • Mediation hypothesis : If so, does it do so by changing social risk and protective mechanisms (family communication)? What did you change early on that brought that final outcome/change about. Combining randomized trials • Perrino et al. (2014) Preventing internalizing symptoms among Hispanic adolescents: A synthesis across Familias Unidas trials. • Combined data from three trials of FU: – Universal intervention: 266 8th grade students from local schools – Targeted intervention: 213 8th grade Hispanic adolescents with behavior problems – Targeted intervention: 242 1217 year old Hispanic students with history of delinquency • Integrative Data Analysis: New methods for analyzing data from combined trials Research question • Do effects differ for families who begin with poor parentadolescent communication? • Moderator Hypothesis: – Those with poor communication will do better because the program is designed to help them – Those who begin with good communication will not benefit as much because they don’t need the intervention • Measured parentadolescent communication prior to random assignment • Internalizing symptoms assessed 24 times, with 12 36 month follow ups Findings • Baseline target levels moderated intervention effects → the better the families did, the less those kids changed. ParentAdolescent communication is actually an active moderator. – Those low on target at baseline showed greatest reduction in internalizing; those high on target showed no effects – Lower is better in this figure: Question of mediation • Why? Changed communication. Changed how parents interact with the broader community. Parent who were isolated had children who were much more at risk • That is, if people do better, is that because the intervention changed what it was targeting? • Mediation: (def.) change in some specific target is what accounts for change in some general outcome • This may involve a cascade of effects: – Intervention influences communication – Communication influences externalizing behavior – Those who don’t develop externalizing behavior are less likely to experience depression or anxiety later Mediational cascades • This may involve a cascade of effects: – Intervention influences communication – Communication influences externalizing behavior – Those who don’t develop externalizing behavior are less likely to experience depression or anxiety later Intervention Communication Externalizing Internalizing Randomized trial to test mediation cascade • Perrino et al. (under review)→ not yet published • Targeted intervention: 213 atrisk adolescent Hispanic youth and families • All in 8th grade • Randomly assigned to Familias Unidas or community practice control • Measures at baseline, 6, 18, and 30 months – Externalizing behavior – Internalizing – Parentadolescent communication Findings: • Lagged cascade using autoregressive model • True developmental model (sameaged cohort) • Tests of opposite direction (I > Int, Int > Ext) Critical thinking minute: Relevant for other ethnicities? Interventions with intimate partners→ Programs to enhance communication • Premarital Enhancement Program (PREP) – Howard Markman, Scott Stanley – Couples groups meeting for 5 sessions • Listening and expressive speaking skills • Problem solving skills (brainstorming, contracting) • Exploration of conscious and unconscious expectations about each other **Examples: Native American Talking Stick, “The Floor” SpeakerListener Rules • Rules for both: – The Speaker is the one with the floor. – Share the floor back and forth. – Explain and seek to understand each other's priorities. – The Speaker can pass the floor at any time to the Listener to hear their side of an issue. a. Speaker Rules • Don't go on and on. • After saying a bit, pause and let the Listener check out what he or she understood. • Speak for yourself; explain why something is important for you. b. Listener Rules • Listen for why things are important for the Speaker. • Every so often check out your understanding of what the Speaker means. • Ask for examples or explanations of what the Speaker said. • Don't offer your own opinions or thoughts until you get the floor. • Concentrate on what the Speaker is saying, and try to "edit out" your own internal responses. Randomized trial of PREP • (Markman et al, 1988) . Prevention of marital distress: A longitudinal investigation. – Randomly assigned 42 couples planning to marry to PREP or no intervention control – PREP involved 5 couples group sessions, focusing on: • Communication skills training • Problem solving training • Clarification of marital expectations • Sexual/sensual education and relationship enhancement – Collected data at posttest and across next 5 years Effectiveness of PREP • Up through 4 year followup, those in PREP showed: • Better communication skills (people rated their discussions as effectiveness of communication). • Better conflict management skills • More support and validation • Better problem solving • Fewer divorces (though difference is not significant) • Evidence of reduced physical violence → Did not put people in the study who were “at risk”, and this was conducted on a normative sample. *NOTE: this study is not relevant today because of increase in access to technology, different gender roles( i.e. homosexual relationships, feminism movement). Critical thinking moment: Continuing relevance? What about treating distressed relationships? Marital conflict and health • KiecoltGlaser & Newton review (2001): Negative or hostile behavior during interactions associated with: – Higher blood pressure and higher levels of epinephrine, norepinephrine (stronger in women) • McShall & Newton (2015) used data from 3 national surveys (10,057 people) – More depression in both males and females (at rates 10 to 25 times higher than in untroubled marriages) – Consistent across 11 ethnic groups *NOTE: says something about the impact of problematic relationships! Brief psychoeducational couples workshop • John Gottman→ Developed the Art and Science of Love (ASL) 2day workshop for distressed couples – Day one: enhancing friendshippowerful protector against stress • Acquiring knowledge of the partner’s inner psychological world • Expressing affection and respect on an everyday basis • Turning toward one’s partner’s bids for emotional connection(def.) knowing what someone needs emotionally(i.e. to be listened to, validated). – Day two: conflict management • Behavioural communication skills training, teaching a gentle approach to conflict management (softened startup, accepting influence, effective repair and compromise). • Dealing with gridlocked content and “existential” differences: honouring each other’s dream → even with those kinds of difference, there needs to be an acceptance with the other rather than the need to change things in the (example: very left wing partner and right wing partner) Evaluating intervention components • Dismantling design: Comparing effects of full intervention with individual components – Tests whether individual components differ in their overall impact – Can also test whether individual components influence different targets – Test whether the full intervention increases efficacy over individual components Brief psychoeducational couples workshop • Babcock et al (2013) A component analysis of a brief psycho‐educational couples' workshop: One‐year followup results. • Tested components of Gottman’s Art and Science of Love (ASL) 2day workshop • Identified significantly distressed couples • Randomly assigned couples to: – Friendship component – Conflict component – Combination of both components – Bibliotherapy (copies of book) • Collected data on satisfaction at posttest, 6 months, 12 months Findings • Combination worked best overall, particularly for women • Men appeared to respond mostly to the friendship module Summary • Interventions targeting social risk factors do appear to have preventive effects, for both early intervention and in adolescence • Interventions targeting intimate partners appear to be useful both preventively and as treatment for more severe distress→ did not look at family therapy interventions. • More complex designs that test moderation, mediation, and dismantling help in understanding what works for whom, and why.