«ATTENTIONAL BIAS IN PATIENTS WITH IMPLANTABLE CARDIOVERTER DEFIBRILLATORS: EXAMINING MECHANISMS OF HYPERVIGILENCE AND ANXIETY By NEHA K. DIXIT A ...»
Thirty seven VF (ICD) patients (ages 34-80) and 41 AF patients (ages 37-80) participated in the study. Participants were recruited through the Electrophysiology Clinics at Shands Hospital at the University of Florida Health Science Center. Per interview, all participants were right-handed native-English speakers. Our sample consisted of 93% Caucasian, 4% AfricanAmerican, and 3% Hispanic. Potential participants were excluded from the study for the following reasons: 1) Major Axis I psychopathology; 2) dementia or other neurological disease;
3) acute medical illness; 4) current use of antiepileptics or other medication known to significantly affect cognitive functioning; 5) motor deficits that would interfere with the use of the dominant hand for performance of button press associated with the dot-probe task; and
6) a score of less than 30 on the Telephone Interview for Cognitive Status (TICS; Brandt, Spencer, & Folstein, 1988). All participants provided written informed consent according to procedures established by the University of Florida Health Science Center Institutional Review Board. Participants were compensated $10 for their time.
Demographic characteristics of study participants are provided in Table 2-1. ICD and AF patients were well matched for education, t(78) = -.57, p.68, and were screened for reading using the North American Adult Reading Test (NAART; Blair & Spreen, 1989; Nelson, 1982).
ICD patients and AF patients reported similar levels of depressive symptoms on the Beck Depression Inventory, 2nd Edition (BDI-II; Beck, 1996), t(78) = -.03, p.90. ICD patients and AF patients also reported similar levels of state anxiety state, but greater trait anxiety, compared to AF patients t(78) = 1.29, p.04.
Medical data on cardiac diagnoses, current medication, and ICD-related information was also obtained for purposes of characterizing the two groups. Mean ejection fraction was 35.87 (S.D. = 14.36). Respondents’ medical history was significant for ventricular tachycardia (21%), ventricular fibrillation (11%), coronary artery disease (45%), and myocardial infarction (23%).
Seventy two percent of the sample had been diagnosed with congestive heart failure. Medication use was as follows: 58% endorsed taking aspirin, 51% Coumadin, 84% beta-blockers, 15% calcium channel blockers, 30% ACE inhibitors, 20% angiotensin receptor blockers, 48% diuretics, 10% amiodarone, and 5% sotalol.
Participants attended one 1- ½ hour testing session. Prior to the first session, participants were administered the TICS (Brandt et al., 1988) as an initial screen for cognitive impairment.
Potential participants with TICS scores of less than 30 were excluded from the study. Using this cutoff score, the TICS has a reported sensitivity of 94% and a specificity of 100% for distinguishing demented individuals from cognitively intact individuals (Brandt et al., 1988).
Thus, the TICS provided a means to exclude demented individuals from the study. No participants were excluded using this criterion during recruitment for this study.
During the experimental session, all participants received a screening 1 of relevant psychiatric and medical history. Participants were also screened for neurological insult that might be an exclusionary criterion. They were asked whether they have difficulty reading the newspaper to determine visual acuity problems that might interfere with performing the computer task. The presence and severity of depressive symptoms were assessed via the BDI-II.
Participants were screened for psychiatric conditions via clinical interview and review of the medical record.
Participants were also given several psychosocial measures to examine general psychological and emotional functioning. The measures given were: (1) The Florida Shock Anxiety Scale (FSAS) was developed to assess the fear and anxiety that patients commonly have regarding the ICD and its shocks. This 16-item measure examines the cognitive, behavioral, emotional, and social impact of shock anxiety; (2) Spielberger State-Trait Anxiety Inventory (STAI), a clinical measure of anxiety; (3) SF-12, a generalized measure of health related quality of life; (4) The Left Ventricular Dysfunction Questionnaire (LVD-36) a cardiac specific quality of life measure; (5) The Beck Depression Inventory, 2nd Edition (BDI-II) and (6) Telephone Interview for Cognitive Status (TICS). These measures are described in detail below.
Shock Anxiety The Florida Shock Anxiety Scale (FSAS): This scale was developed in the Cardiac Psychology Lab at the University of Florida for a previous study to assess the fear and anxiety that patients may have regarding the ICD and its shocks. This 16-item measure examines the cognitive, behavioral, emotional and social impact of shock anxiety. Full psychometric validation available (Kuhl, Dixit, Wallace, Sears, & Conti, 2005).
General Anxiety State-Trait Anxiety Inventory (STAI): The STAI is a 40-item self-report questionnaire designed to measure both state and trait anxiety (Speilberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983). Trait anxiety is defined as a relatively enduring personality characteristic, or more specifically, as anxiety proneness. State Anxiety is defined by a short-lived anxiety, usually induced by an event or circumstance. Both of these indices of anxiety will be examined to differentiate the extent and level of anxiety.
General Health-Related Quality of Life Short Form-12 (SF-12): This measure was developed to gauge mental and physical functioning and can be separated into two components: physical component summary (PCS-12) and mental component summary (MCS-12). All scores of the SF-12 are comparable and highly correlated with scores from the SF-36, from which it was derived, (ranging from.63-.97; Ware et al., 1995; Ware, Kosinski, & Keller, 1996). The SF-12 reproduced 90% of the variance in the SF-36 PCS and MCS measures in the United States and on cross-validation in the MOS (Ware et al., 1996).
The Left Ventricular Dysfunction Questionnaire (LVD-36): This cardiac-specific measure was designed to assess the impact of left ventricular dysfunction on daily life and wellbeing. Responses are dichotomous (true or false). True responses are summed, which are then calculated as a percentage; higher scores indicate worse functioning (i.e., 0 = best possible score). The measure demonstrated high internal consistency in a sample with chronic left ventricular dysfunction (Kuder-Richardson coefficient =.95) (O’Leary & Jones, 2000).
Depression Beck Depression Inventory-2nd Edition (BDI-II): The BDI-II is a 21-item self-report instrument assessing the presence and severity of depression symptomatology over the preceding two weeks (Beck et al., 1996). Its internal consistency ranges from.91 to.93, its one-week testretest reliability is.93 and moderate to high correlations with other measures of depressive symptomatology supports its convergent validity. BDI-II has been widely used in cardiac populations (Carney, Freedland, Sheline, & Weiss, 1997) and is the gold standard for assessing depressive symptoms in health-related populations (JAMA, 2000).
Cognitive Screener Telephone Interview for Cognitive Status TICS: The TICS is a brief test of cognitive functioning developed. The TICS is similar to the Mini-Mental Status Exam (Folstein, Folstein, & McHugh, 1975), but has a more comprehensive memory assessment, designed for identifying dementia. Potential participants with TICS scores of less than 30 were excluded from the study.
Research has demonstrated that it is as reliable and valid as face-to-face administration. It has a sensitivity of 94% and specificity of 100% for distinguishing normal controls and demented individuals (Brandt et al., 1988) and sensitivity of 82% and specificity of 87% for distinguishing normal controls and amnestic mild cognitively impaired older adults (Cook, Marsiske, & McCoy, 2006).
Reading The NAART (Blair & Spreen, 1989; Nelson, 1982) was used to estimate overall reading abilities.
Experimental Task The computerized task was run on a DELL PC laptop computer using E-Prime software for stimulus presentation and behavioral data collection. To ensure that participants understood task instructions and to increase familiarity with the button-press procedure, participants were pre-practiced on the computerized cognitive task. The task paradigm utilized was a modified version of a classical dot-probe paradigm developed by Williams, Watts and McLeod (1988).
Figure 2-1 illustrates a sample trial of the dot-probe task used in this experiment.
The task comprised a briefly-presented word cue, shortly followed by a target to which participants made a speeded button-press response. Specifically, participants were instructed to focus on the center of the screen where they saw a fixation point. Each trial of the task began with a centrally-located 200-ms duration fixation point followed by a cue word presented to the top or bottom of the fixation cross. After 800 ms, the cue word was immediately replaced by a dot “*” target, which appeared randomly in the same (congruent) or opposing (incongruent) location as the word. Participants were instructed to respond to the presentation of the target by pressing the “h” or “j” keys indicating the location of the dot as quickly and accurately as possible. The dot serving as a target disappeared after the key press or after 4000ms. The intertrial interval from the target offset to the next fixation cross was 1200ms. The participants’ response time (with ms accuracy; RTs) and accuracy to the target were recorded as dependent variables.
Participants performed a total 240 experimental trials, equally and randomly distributed across four word types and two word positions. Each word was repeated four times during the entire task. Fifty percent of trials were congruent, drawing the attention of the participant to the area where the word and asterisk appeared, while the remaining fifty percent of trials were incongruent, drawing participants’ attention to the area opposite the one where the asterisk appeared. Trials were randomized for each word category, with each category presented an equal number of times across the task.
Four different word types were employed as cues, including cardiac-specific threat words (e.g. shock, defibrillator, flutter), non-cardiac-specific threatening words (e.g., fearful, scared, danger), pleasant words (e.g. delighted, confident, happy), and neutral words (e.g. tile, doorknob, bland). Positive, threat and neutral words were chosen from norms of emotional words taken from the Affective Norms for English Words (ANEW; Bradley & Lang, 1999) and matched for frequency of usage in English, average word length, and grammatical structure. Cardiac specific words were chosen from a group amassed and rated by clinicians at the Shands EP clinics and graduate students in the Cardiac Psychology Lab. The cardiac-specific words were selected based on rating for “high” valence and arousal. Valence and arousal ratings were measured separately using a computerized administration of the Self-Assessment Manikin (SAM; Lang, 1980). Both dimensions of valence and arousal were rated on a 9-point Likert scale with 1=least pleasant/arousing and 9=most pleasant/arousing. Forty words comprised the final set of stimuli (Appendix A). As a manipulation check, participants performed valence and arousal ratings for each word seen in the experiment, using the Self-Assessment Manikin (SAM; Lang, 1980) after completion of the dot probe task.
Dot Probe Task Dependant measures for the dot probe will include reaction times and error rates for each of the experimental conditions. For analyses involving RT, we employed median RTs (Ratcliff,
1993) for correct responses. For analyses involving error rates, data were arcsine transformed (Neter, Wasserman, & Kutner, 1985) prior to all analyses. This transformation was used to normalize the distribution of the error data, which is often skewed because the error rates are so low proportionately. Median correct-trial reaction times (RTs) and arcsine errors were calculated for each participant and experimental condition, and subjected to separate Group x 2Cue Validity (Congruent, Incongruent) x 4-Cue Valence (Pleasant, Neutral, General Threat, Cardiac Threat) Analyses of Variance (ANOVAs). Group served as the between-subjects factor, and cue congruency and cue valence served as within-subject factors. To correct for possible violations of sphericity, a Hyundt-felt epsilon adjustment was calculated where appropriate and adjusted p-values and unadjusted degrees of freedom are reported. Effect sizes for ANOVAS
were measured using Eta squared. The following hypotheses were addressed in the analyses:
Hypothesis 1: A main effect of congruency will be seen across groups (slower RTs and greater error rates to incongruent than congruent trials).
Hypothesis 2: ICD patients, compared to arrhythmia controls, will exhibit a specific and disproportionate RT slowing to incongruent- relative to congruent-cue trials specifically involving clinically relevant words.
Hypothesis 3: There will be a significant 3-way interaction, reflecting disproportionate slowing of ICD patients to clinically specific incongruent vs. congruent cues compared to other word types and to AF controls.
Reaction Time Data Overall, there was a significant effect of congruency in the opposite direction than predicted, F(1, 78) = 16.377, p.001, η2 =.98, with longer RTs to the congruent than incongruent condition. There was no significant effect of valence, F(3,228) =.713, p.55, η2 =.15, nor was there a Group x Valence interaction F(3,228) =.478, p.67, η2 =.11. Finally, no Group x Congruency x Valence interaction, was found as hypothesized for RTs F(3,228) = 1.857, p0.14, η2 =.27 (Figure 3-1).
Error Data A main effect of group was observed for error rates, F(1, 78) = 16.099, p.001, η2 =.98, with ICD patients making greater errors overall, than AF patients. Next, we examined the effects of cue type (valence) on dot-probe task performance. There were no significant effects of valence on error rates, F(3,228) =.684, p0.55, η2 =.02, nor was there a Group x Valence interaction F(3,228) =.865, p0.45, η2 =.01. Finally, no significant Group x Congruency x Valence interaction was found for error rates, F(3,78) =.781, p0.50, η2 =.01 (Figure 3-2).