Behavioral Science Research Programs
Goals and objectives of the Department of Behavioral Science include
- Conducting behavioral research on cancer risk behaviors, including the development of a behavioral sciences knowledge base
- Expanding interdisciplinary collaborative research programs between behavioral scientists and other disciplines, including basic science, clinical research and population-based sciences
- Assuring a cohesive, multidisciplinary, critical mass of behavioral science investigators within a well-funded research program
- Building clinical research programs in smoking cessation, genetic counseling and testing for hereditary cancer susceptibility and psychosocial oncology
- Expanding the psychobiology research program for the purpose of conducting research on nicotine dependence, withdrawal and relapse and on basic biobehavioral mechanisms
- Focusing research programs on high-risk, underserved and minority populations wherever possible
Efforts toward these ends include the following research programs:
Relationship Maintenance in Lung Cancer: Associations with Psychological and Marital Functioning
Hoda Badr, Ph.D. Cindy L. Carmack Taylor, Ph.D., Trayce Hall, B.S., Bena Ellickalputhenpura, B.A.
Relationship maintenance behaviors are approaches couples use to sustain or enhance their relationships in the face of a shared stressor. Canary and Stafford (1992) have identified five maintenance strategies involving relationship-oriented discussions (openness, assurances) and interaction behaviors (positivity, network, tasks) that are associated with conflict resolution, commitment and marital satisfaction in non-cancer samples. Use of these strategies may be helpful for couples facing lung cancer. Research indicates lung cancer patients have more problems with communication and interaction with their spouses than patients with other cancers. This may increase their risk for conflict and relationship distress. This study examines the patterns of association of relationship maintenance strategies with psychological distress (Brief Symptom Inventory) and marital functioning (Dyadic Adjustment Scale) among 80 newly diagnosed lung cancer patients and their spouses. For both patients and spouses, multilevel modeling analyses using SAS Proc Mixed showed significant negative associations for all five maintenance strategies with psychological distress (ps >.001). Significant positive associations were also found for positivity, network, and tasks with marital functioning for patients and spouses (ps >.001). However, patients were significantly more likely than their spouses to report greater marital functioning when they also reported greater openness (F(1,72)=5.92, p=.02) and assurances (F(1,72) = 8.0, p=.005), highlighting the potential value of relationship-oriented discussions as a shared coping strategy for couples facing lung cancer. These results underscore the importance of examining the patterns of association of relationship maintenance strategies with both individual psychological and marital functioning.
Sexual Counseling after Prostate Cancer Treatment
Andrea Canada, Ph.D., Leah Neese, M.A., Dawen Sui, M.S., Leslie Schover, Ph.D.
An intervention after treatment for localized prostate cancer (PC) was designed to increase help-seeking for erectile dysfunction (ED) and sexual satisfaction for the man and his wife/partner. Couples were randomized to four sessions of counseling with vs. without the partner present, although all women completed questionnaires and behavioral homework. Sessions covered PC-Baseline, post-treatment, three- and six-month follow-up questionnaires included: International Index of Erectile Functioning; Female Sexual Function Index; Brief Symptom Inventory-18 and utilization of medical treatments for ED. Of 84 survivors recruited, 51 (61%) completed all four sessions, 26 in the “man-alone” and 25 in the “couple” conditions. “Couple” participants were more likely to complete the intervention (66% vs. 57%). No other between-group differences were significant. Regression analyses revealed improvements through three-month follow-up for male emotional distress (p = .0192), global sexual function (p = .0005), sexual satisfaction (p < .0001), erectile function (p = .0002) and intercourse satisfaction (p = .0071), and for female global sexual function (p = .0032), sexual satisfaction (p < .0014) and lubrication (p = .0461). By six-month follow-up, only improvements in male erectile function and sexual satisfaction (p = .0356 and .0021) remained. At baseline, 31% of men were using a medical treatment for ED, compared with 55% at three- and 49% at six-month follow-ups (p = .0032). The intervention increased survivor help-seeking for ED and enhanced sexual function and satisfaction for both partners, with some regression by six months.
The Effects of Smoking Deprivation and Nicotine Administration on Emotional Reactivity
Paul Cinciripini, Ph.D., Brian Carter, Ph.D., Jason Robinson, Ph.D., Cho Lam, Ph.D., Tracy Long, Ph.D., Xifeng Wu, M.D., Ph.D., David Wetter, Ph.D.
The relationship between negative affect and the maintenance of smoking behavior, as well as cessation relapse, has strong empirical support. Although several converging lines of evidence suggest that nicotine and mood are related at a fundamental biological level, this link has not been thoroughly studied. In this study, startle probe methodology was used to examine the effects of nicotine (and nicotine deprivation) on emotional processes associated with drive and motivation. The startle (eye blink) response is an orienting reflex, which occurs during the presentation of an unexpected auditory stimulus (probe). Its strength is measured by the electromyographic (EMG) changes in the orbicularis oculi region of the eye. The magnitude of the blink EMG response to the probe is enhanced when subjects view unpleasant emotional pictures and attenuated when viewing pleasant emotional pictures. The strength of the startle response to these cues is thought to reflect the influence of affective perception on the activation of underlying neurological mechanisms governing motivation and emotion. Negative affect typically activates the brain’s defensive system (avoidance, escape) while positive affect stimulates the appetitive (approach) system. Smokers (n = 115) completed four laboratory sessions crossing (12-hour deprived vs. non-deprived) with nicotine spray (active vs. placebo) during which they viewed affective pictures (positive, negative, neutral) while startle probes were administered. A deprivation by picture valence interaction and a main effect for nicotine spray were found. Nicotine administration suppressed startle responding regardless of picture type and the interaction indicated that startle responses to negative stimuli were enhanced during nicotine deprivation and attenuated under conditions of normal smoking. These findings suggest nicotine deprivation predisposes the smoker to respond to negative emotional cues with enhanced activation of the defensive system while nicotine helps suppress defensive responding. Taken together, this evidence suggests nicotine influences brain structures related to drug motivation and emotion.
Late Relapse in a Community Sample
Ludmila Cofta-Gunn, Ph.D., David Wetter, Ph.D., Rachel Fouladi, Ph.D., Paul M. Cinciripini, Ph.D., Ellen R. Gritz, Ph.D.
Relapse remains the most refractory aspect of nicotine dependence, yet little is known about the prevalence and predictors of late relapse (i.e., relapse after a period of prolonged abstinence). The few data that exist indicate that former smokers may relapse many years after they quit, but these studies generally involve nonrepresentative volunteer samples. Data on late relapse rates in population-based samples are almost nonexistent. The current study investigated the prevalence and predictors of late relapse in a nonvolunteer community-based sample from the Working Well Trial, a worksite-based cancer prevention study. Employees at participating worksites were surveyed at baseline and four years later. Relapse rates over that four-year period were examined in baseline former smokers (N=1,143). Relapse rates significantly differed as a function of the length of abstinence at baseline (p<.001): 29.6% in those reporting abstinence up to one year, 9.4% in those reporting abstinence one to five years, and 1.8% in those reporting abstinence over five years. Relapsers were significantly younger, more likely to live with other smokers, more likely to use smokeless tobacco and less confident they would be nonsmokers in 12 months. However, after controlling for baseline duration of abstinence, there were no significant predictors of relapse. Late relapse rates in this community-based sample were lower than those typically obtained in clinical trials. Variables known to predict short-term relapse did not predict late relapse. The only robust predictor of late relapse was baseline duration of abstinence. Interventions to prevent late relapse will likely need to target specific high risk populations (e.g., former smokers with shorter duration of abstinence) to be effective.
Smoking Cessation Among Hispanic Smokers: ¡Adiós al Fumar!
Patricia Daza, Ph.D., Lynne Nguyen, M.P.H., Carlos Mazas, Ph.D., Luz Mejia, M.A., David Wetter, Ph.D.
¡Adiós al Fumar! is a randomized clinical trial that evaluated the efficacy of delivering an enhanced smoking cessation counseling program to smokers who called the National Cancer Institute’s Cancer Information Service (CIS) requesting help in Spanish. Participants were randomized to either a single, standard CIS counseling call (SC) or to enhanced counseling (the single CIS call plus three additional proactive calls; EC). All assessment and counseling calls were delivered in Spanish. Of 355 callers, 306 agreed to participate (86% participation rate). The sample was 55% male, 92% were immigrants, 60% spoke only Spanish in the home, mean age was 41, cigarettes/day was 10, Fagerstrom score was 6, and 57% had scores on the CES-D indicating possible or probable depression (> 16). The sample was of low socioeconomic status. Over 50% had total annual household incomes <$20,000, 75% had no insurance, and mean years of education was 11. Among EC participants, 93% received at least three of the four counseling calls. The overall efficacy of EC relative to SC across follow-up timepoints approached significance (p = .09). Abstinence rates were significantly different at week 5 (20.3% for EC and 11.7% for SC; p = .05), but were not maintained at week 12 (27.4% for EC and 20.5% for SC; p = .20). In sum, the Adios project recruited a very underserved, low SES, Spanish speaking population who have no or few alternatives for receiving health care; successfully followed up 80% of those individuals; and delivered at least 75% of the treatment dose to 93% of EC participants. Moreover, promising results were found for an enhanced CIS counseling protocol.
Self-Generated Smoking Outcome Expectancies Partially Mediate the Relationship Between Gender and Smoking Behavior in Adolescents
Jennifer Irvin, Ph.D., Cheryl Anderson, Ph.D., Kathryn I. Pollak, Ph.D., David Wetter, Ph.D.
Expectancies serve key roles in smoking initiation. Because gender differences exist in adolescent smoking, the current study examined whether self-generated smoking outcome expectancies mediated the relationship between gender and smoking behavior among 350 female and 315 male high school students. Students were classified as nonsusceptible nonsmokers, susceptible nonsmokers, experimenters and current smokers. Analyses examined the associations between gender and smoking, gender and expectancies and expectancies and smoking, as well as whether expectancies mediated the gender-smoking relationship. All analyses controlled for grade, race/ethnicity and peer smoking. Ordinal logistic regression analyses indicated that boys were more likely to be current smokers and less likely to be nonsusceptible nonsmokers (OR=1.69, p=.0005). Boys were more likely to associate smoking with buzz, pleasure, taste/smell, stimulation, cost and exercise/sport impairment, whereas girls were more likely to associate smoking with weight control, negative aesthetics, addiction and negative mood. After simultaneously controlling for nine positive (pleasure, taste/smell, buzz, enhance self esteem, relieves boredom, escape problems, something to do with hands, coupons/ads, smoke tricks) and five negative (negative social, causes negative mood, cost, endangers/disturbs others, taste/smell impairment) expectancies, the strength of the gender-smoking relationship declined but remained significant (OR=1.38, p=.04). Findings indicate that self-generated expected outcomes for smoking partially mediate the relationship between gender and smoking behavior among high school students. The findings suggest that boys and girls have different motivators for initiating smoking, and prevention interventions might benefit from tailoring based on these differences.
Startle Potentiations and Self-Reported Depressive Symptoms in Smokers
Cho Lam, Ph.D., Paul Cinciripini, Ph.D., Brian Carter, Ph.D., Jason Robinson, Ph.D., David Wetter, Ph.D.
A positive relationship exists between depression and smoking. A significantly larger proportion of smokers than nonsmokers have been diagnosed with major depressive disorder and self-reported depression predicts nicotine withdrawal and relapse during a quit attempt. The startle eye-blink response (potentiated in the presence of aversive affective stimuli and attenuated in the presence of positive stimuli) is an empirically validated measure of emotional reactivity that has previously been shown to be a reliable index of negative affect in nicotine-deprived and non-deprived smokers. This study examined startle eye-blink response in smokers with varying degrees of self-reported depressive symptoms. Smokers were randomly assigned to either quit smoking (n = 70) or to continue smoking regularly (n = 46). Participants in the quit condition received counseling to help them quit. Participants completed four laboratory sessions at identical time points (corresponding to baseline, two, six and 15 days post-quit for the quit smoking group) in which they viewed positive, aversive, and neutral pictures while orbicular EMG was being measured. Results showed that smokers with high pre-quit distress measured at baseline exhibited significantly larger startle responses than those with low or no distress. Furthermore, at two days post-quit, among participants with high levels of pre-quit depressive symptoms, those who quitted smoking at that time but eventually relapsed demonstrated a significant increase in startle potentiation than those who quitted successfully and those who continued smoking. The results suggest that startle potentiation is a sensitive index of pre-quit depressive symptoms and that these symptoms influence the relations between affective processes and quitting.
Validating a Multi-Dimensional Measure of Tobacco Dependence among Spanish-Speaking Smokers
Carlos Mazas, Ph.D., Patricia Daza, Ph.D., Lynne Nguyen, M.P.H., Luz Mejia, M.A., David Wetter, Ph.D.
Most measures of tobacco dependence are based on a unidimensional model of dependence (e.g., physiological addiction). Conversely, the Wisconsin Inventory of Smoking Dependence Motives (WISDM-68) is a new measure that utilizes a multidimensional model of dependence and includes 13 sub-scales assessing motivational factors associated with smoking and relapse (Piper et al., in press, JCCP). The current study evaluated a Spanish language version of the WISDM-68 among Spanish-speaking smokers who called the National Cancer Institute’s Cancer Information Service (CIS) requesting smoking cessation services in Spanish. Approximately 83% of interested callers agreed to participate, yielding 229 participants. To date, 190 participants have completed the 12-week follow-up (83% follow-up rate). Participants were 58% male, 92% were immigrants (67% of Mexican origin), 57% speak only Spanish, mean age was 40, cigarettes/day were 10, Fagerstrom score was 7. The sample was of low socioeconomic status. Over 50% had total annual household incomes <$20,000, 76% had no insurance and mean years of education was 11. All 13 WISDM-68 sub-scales had good internal consistency (coefficient alphas from .74 to .90). WISDM-68 total scores were significantly related to other dependence measures (Fagerstrom Test of Nicotine Dependence, r=.27; cigarettes per day, r=.43; first cigarette within 30 minutes, r=.39). Moreover, four sub-scales prospectively predicted abstinence at 12 weeks post-cessation (Affiliative Attachment, p=.03; Automaticity, p=.04; Craving, p=.04 and Tolerance, p=.01). In sum, our results indicate that the Spanish language version of the WISDM-68 is reliable, valid and predicts outcome.
Colonoscopy Utilization Before and After HNPCC Genetic Testing
Susan Peterson, Ph.D., M.P.H., Ellen Gritz, M.D., Sally Vernon, Ph.D., Catherine Perz, Ph.D., Salma Marani, M.S., Beatty Watts, M.S., Chris Amos, Ph.D., Marsha Frazier, Ph.D., Patrick Lynch, M.D., J.D.
Genetic testing can identify HNPCC-predisposing mutations in persons with a cancer family history suggestive of this syndrome. Mutation carriers have up to an 85% lifetime risk of developing colorectal cancer (CRC), and are advised to have annual or biannual colonoscopy usually beginning at age 25 years. In the absence of genetic testing, persons who are at increased risk for HNPCC by virtue of their family history also are advised to follow the same cancer screening guidelines as mutation carriers. As part of a longitudinal study of psychosocial aspects of HNPCC genetic counseling and testing, we evaluated colonoscopy utilization before and after genetic testing. Subjects were 79 adults from families with a known HNPCC-predisposing mutation and with no prior history of CRC. Prior to genetic counseling, 53% (n=42) had ever had a colonoscopy. After genetic testing, 33% (n=26) were positive for an HNPCC-predisposing mutation. There was no difference in baseline prevalence of colonoscopy use between mutation carriers and non-carriers (50% and 54%, respectively). Within six months of receiving test results, 76% of mutation carriers reported that they had undergone colonoscopy since receiving their results, a significantly greater proportion compared with baseline utilization and with non-carriers (F=18.87, p<0.001). Compared with baseline responses, mutation carriers also expressed significantly greater self-confidence in their ability to follow through with colonoscopy, a greater level of commitment to having colonoscopy and fewer perceived barriers to CRC screening after test results disclosure. Undergoing HNPCC genetic counseling and testing may motivate mutation carriers to have a colonoscopy within a relatively short time period after disclosure of test results. HNPCC genetic counseling and testing also may positively influence mutation carriers’ attitudes toward CRC screening.
Knowledge of Cervical Dysplasia and Human Papillomavirus Among Women Seen in a Colposcopy Clinic
Sandi Pruitt, M.P.H., Patricia Parker, Ph.D., Susan Peterson, Ph.D., M.P.H., Tao Le, B.S., Michele Follen, M.D., Ph.D., Karen Basen-Engquist, Ph.D., M.P.H.
Colposcopy clinic patients with greater knowledge about the range of issues surrounding cervical cancer have higher rates of adherence to follow-up screening and treatment. We report on knowledge of cervical dysplasia and human papilloma virus (HPV) among women (n=175) who attended initial colposcopy appointments to follow-up abnormal Pap smears. Patients included women referred on the basis of an abnormal Pap smear (n=87) and screening clinic patients who had an abnormal Pap smear in the last 12 months (n=88). Demographics, knowledge and psychological distress were assessed in structured interviews before, during and after colposcopy. Respondents had low knowledge scores before and after colposcopy; however, their overall knowledge improved slightly following the exam (t=2.3, df=150, p=.026). When examined by question, respondents demonstrated a significant increase of correct answers to only one question: Does dysplasia, or precancerous cells on the cervix, always go away without treatment? Pre-exam knowledge was positively associated with educational level and was lower among Hispanics and referral patients. Post-exam knowledge was positively associated with pre-exam knowledge, educational level, and embarrassment about the colposcopy. Contrary to our hypotheses, neither pre- nor post-exam knowledge was associated with anxiety in the regression models. These results suggest that routine clinical education during colposcopy can improve patients’ understanding of cervical cancer; however, the low level of knowledge that persisted after colposcopy is a cause for concern.
Technology-Assisted Tobacco Control Programs for High School and Junior College Students
Alexander Prokhorov, M.D., Ph.D.
Smoking prevention and cessation programs based on modern theoretical concepts and computer technologies are in demand. Our research team has recently completed two studies aimed at designing and testing interventions based on transtheoretical model of change (TTM) and utilizing interactive computer technologies. The first project is a CD-ROM-based classroom curriculum (Project ASPIRE). The program provided eight educational “tracks” tailored to student characteristics, such as stages of smoking acquisition or cessation. ASPIRE was evaluated in a group-randomized controlled trial of 16 urban high schools. At 18-month follow-up the primary aims were to examine smoking initiation rates among baseline nonsmokers and smoking cessation rates among baseline smokers. A total of 1160 participants completed the 18-month survey, 610 in the intervention group and 550 in the control group. Among baseline nonsmokers there were significantly fewer students who initiated smoking in the intervention group compared to the control group (2% vs. 6%). An evaluation of subgroups at high risk compared to those at lower risk indicated that the intervention was more effective for those with higher depression, lower resistance skills, higher peer pressure to smoke and prosmoking parental influences. The findings to date reflect promise of the ASPIRE program in positively changing determinants and smoking behaviors among nonsmokers at baseline. The other study is a TTM-guided smoking cessation intervention in 15 junior (two-year) college campuses. A total of 326 student smokers were followed-up, 125 (mean age was 22.6 ± 4.4 years; 62% females) of them were evaluated using spirometry and respiratory symptoms at baseline and at 10-month follow-up assessments. Cotinine validated quit rates between intervention and control were 17% vs. 10%. At 10-month follow-up the intervention was successful in changing determinants of smoking cessation including decisional balance between intervention and control (-2.2 vs. 2.1) and temptations to smoke (36.0 vs. 42.3). Noticeable improvements in respiratory symptoms and pulmonary function were observed in quitters versus continued smokers. Future studies with more subjects and longer follow-up are warranted.
Obesity and Physical Functioning in Endometrial Cancer Survivors
Stacie Scruggs, B.A., Cindy Carmack Taylor, Ph.D., Anuja Jhingran, M.D., Diane Bodurka, M.D., Karen Lu, M.D., Lois Ramondetta, M.D., Eileen Shinn, Ph.D., Karen Basen-Engquist, Ph.D.
Endometrial cancer survivors may experience decreased quality of life (QOL) due to endometrial cancer risk factors (obesity, hypertension) and its sequelae. Physical activity improves QOL in cancer survivors. This study aims to describe physical functioning, body mass index (BMI) and exercise patterns, and investigate the relationship of BMI to physical functioning in an endometrial cancer survivor population. Surveys were mailed to 200 survivors of early stage endometrial cancer diagnosed within the last five years; 126 (63%) were returned. Surveys assessed physical functioning (SF-36 sub-scale), fatigue (BFI), pain (BPI) and physical activity (adapted from BRFSS). Respondents were 72% white, 7% African American, 12% Hispanic, 3% Asian, 6% unknown; average age was 63 (SD=12.6, range=32-89). 21% exercised in the past month at the level of current CDC/ACSM guidelines; 42% reported no physical activity, and 37% reported some activity, but not meeting recommendations. Physical functioning scores (M=63.4; SD=31.3) were significantly lower than that of an age-matched normative sample. Average BMI was 31.5 (SD=9.8; range=17-75). Survivors were categorized as underweight/normal (33%), overweight (16%) or obese (51%). BMI was related to physical functioning (p=.001) and ratings of worst pain (p=.022), but not usual and worst fatigue, or usual pain. Pair-wise comparisons among underweight/normal, overweight and obese survivors showed that obese survivors reported significantly higher pain (M=3.4) and lower physical functioning (M=53.0) than normal BMI survivors (worst pain=1.6; physical functioning=79.1). Results suggest that endometrial cancer survivors, particularly those who are obese, would benefit from a QOL intervention incorporating physical activity, although for some, pain could be a barrier to participation.
Depressive Symptoms and Health Behaviors in Testicular Cancer Survivors
Eileen Shinn, Ph.D., Bob Thornton, M.P.H., Karen Basen-Engquist, Ph.D.
Testis cancer survivors enjoy high long-term survival rates but remain at risk for recurrence and other types of cancers, as well as cardiovascular disease. Past research indicates that survivors of testicular cancer may engage in risk behaviors (smoking, alcohol use) at higher rates than in the general population. The associations between depression and health behaviors in testicular cancer survivors were examined using data from the Testis Cancer Survivors’ Adjustment and Health Behaviors Study, which assessed quality of life and health behaviors in testis cancer survivors compared to their male siblings or relatives. The study sample was comprised of 162 seminomatous, non-seminomatous, or mixed germ cell testicular cancer survivors who were two to seven years from diagnosis. Comparisons of levels of depressive symptoms were based on whether or not survivors were participating in various health behaviors. Depressive symptoms were measured by the Center for Epidemiological Studies Depression Scale (CES-D). Results indicated that 15.4% of the survivors scored in the depressed range using the traditional cutoff of 16 on the CES-D. In general, participants who had unhealthy behaviors had higher levels of depressive symptoms compared to the participants with the healthier behavior: (a) current cigarette smoking M=15.3, SD=11.0 vs. not currently smoking M=8.0, SD=8.7 (p=.000), (b) sedentary lifestyle M=12.2, SD=10.7 vs. active lifestyles M= 8.7, SD= 9.2 (p=.07) (c) eating less than five fruits and vegetables a day M=9.7, SD=9.8 vs. eating five or more M=5.3, SD=4.0 (p=.001). One exception was that participants who reported current use of smokeless and other forms of tobacco had lower levels of depression M=8.3, SD=9.5 compared with those who reported no current use M=9.4, SD= 9.5 (p=.71). These results suggest that depression is a relevant issue in interventions targeting health behaviors in testis cancer survivors.
Using Regression and IRT Models to Detect CES-D Mode Effects: Personal Digital Assistant vs. Pencil and Paper
Richard Swartz, Ph.D., Karon Cook, Ph.D., Carl de Moor, Ph.D., Rachel Fouladi Ph.D., Karen Basen-Engquist, Ph.D.
Due to advances in technology, many researchers are moving towards computer-based administration of psychometric scales typically administered by pencil and paper (P&P). However, changing the mode of administering questionnaires can affect the psychometric properties of the scale. The current study examines the effects of administering the CES-D, a self-report depression scale, using a personal digital assistant (PDA) versus the typical P&P administration. We surveyed 557 patients and non-patients from the waiting areas of a cancer hospital. Each participant received two different administrations, in randomly determined order, of the CES-D (P&P and PDA) sequentially. We regressed the difference between the two scores on several independent variables to test for significant mode effects. The P&P scores were statistically significantly higher than the PDA scores. Also people receiving the P&P version first and people with less education had larger mode effects. We fit an IRT model (Partial Credit Model) and explored item level differences for mode effects and administration order using a DIF analysis. Only one item (out of 20) exhibited statistically significant DIF across modes, and it did not remain significant after adjusting for multiple comparisons. Nine of 40 items (20 for P& P and 20 for the PDA) exhibited statistically significant DIF across administration order, with one item maintaining statistical significance after adjusting for multiple comparisons. The trend for all but one of the items exhibiting DIF implied that items were more difficult (participants need a higher depressive level to give high endorsements) on the second administration. The psychometric properties of the CES-D may not be constant across modes of administration, and differences may not result from DIF. Furthermore, the DIF due to administration order could be suggestive of item exposure issues and may need to be considered when developing computerized adaptive tests for depression.
A Randomized Trial of a Tibetan Yoga Intervention for Cancer Patients
Bob Thornton, M.P.H., Alejandro Chaoul-Reich, M.A., George Perkins, M.D., Kavita Chandwani, M.D., M.P.H., Janet Sterner, Ph.D., M.P.H., Rachel Fouladi, Ph.D., Carla Warneke, M.S., M. Rodriguez, M.D., Lorenzo Cohen, Ph.D.
Research suggests that stress-reduction programs integrated into the cancer setting help patients cope with the effects of treatment and improve quality of life. This study examined the benefits of a Tibetan yoga-based intervention program for patients with lymphoma and breast cancer who were undergoing chemotherapy or radiotherapy, or within one year post-treatment. A seven-week Tibetan yoga program was developed that incorporated stress-reduction techniques including controlled breathing, visualization/meditation and postures from the Tibetan yoga practices of Tsa lung and Trul khor. Two cohorts of 39 lymphoma patients and 58 breast cancer patients were randomly assigned to either the Tibetan yoga program or to a wait-list control group. Patients completed measures of intrusive thoughts and avoidance behaviors (Impact of Events Scale: IES), cancer-related symptoms (MD Anderson Symptom Inventory) and sleep disturbances (Pittsburgh Sleep Quality Index) at baseline, one week and one and three months after the last yoga session. The yoga and control groups did not differ with respect to medical/demographic characteristics or baseline dependent measures. For the breast cancer cohort, a mixed model analysis, controlling for the baseline dependent variable, revealed a group by time effect with respect to IES total scores with the yoga group reporting lower scores than the control group by the three-month assessment (adjusted means: 17.12 vs. 20.14, p=.03). A multivariate linear regression analysis, controlling for baseline, revealed that the yoga group reported lower scores for cancer-related symptoms at the one-week follow-up than the control group (change in adjusted means: -8.05 vs. 3.95, p=.04). The lymphoma cohort revealed significantly lower sleep disturbance scores in the yoga group than the control group (5.8 vs. 8.1, p <.004) including better sleep quality (p<.02), faster sleep latency (p<.01), longer sleep duration (p<.03) and less use of sleep medications (p<.02). The results indicated that the Tibetan yoga program was associated with a reduction in avoidance behaviors, intrusive thoughts and cancer-related symptoms for breast cancer patients and sleep-related outcomes for lymphoma patients.