Celestina Barbosa-Leiker

Missing Data in Substance Abuse Treatment Research: Current Methods and Modern Approaches

Two common procedures for the treatment of missing information, listwise deletion and positive urine analysis (UA) imputation (e.g., if the participant fails to provide urine for analysis, then score the UA positive), may result in significant biases during the interpretation of treatment effects. To compare these approaches and to offer a possible alternative, these two procedures were compared to the multiple imputation (MI) procedure with publicly available data from a recent clinical trial. Listwise deletion, single imputation (i.e., positive UA imputation), and MI missing data procedures were used to comparatively examine the effect of two different buprenorphine/naloxone tapering schedules (7- or 28-days) for opioid addiction on the likelihood of a positive UA (Clinical Trial Network 0003; Ling et al., 2009). The listwise deletion of missing data resulted in a nonsignificant effect for the taper while the positive UA imputation procedure resulted in a significant effect, replicating the original findings by Ling et al. (2009). Although the MI procedure also resulted in a significant effect, the effect size was meaningfully smaller and the standard errors meaningfully larger when compared to the positive UA procedure. This study demonstrates that the researcher can obtain markedly different results depending on how the missing data are handled. Missing data theory suggests that listwise deletion and single imputation procedures should not be used to account for missing information, and that MI has advantages with respect to internal and external validity when the assumption of missing at random can be reasonably supported.

Measurement Invariance of the Perceived Stress Scale and Latent Mean Differences across Gender and Time

Measurement invariance of the 2-factor model of the Perceived Stress Scale--10-item version (Cohen & Williamson, 1988) was tested across men and women at two time points and in the combined total sample over a 2-year time frame (n = 871). Measurement invariance results indicated that the scale measured the latent factors, stress and counter-stress, equivalently in men and women and over time. With measurement invariance demonstrated, differences in latent means were tested. Results indicated that men had lower levels of frequencies of stressors, and at one time point, higher levels of counter-stress, when compared with women. When examining change in frequencies of stressors and counter-stress over 2 years with the combined male and female sample, stressors remained stable, yet counter-stress increased over time. These findings may aid in the interpretation of results when examining stressors and counter-stress in clinical samples where one would expect stress to increase, whereas positive psychological states decrease.

Law Enforcement Officer Versus Non–law Enforcement Officer Status as a Longitudinal Predictor of Traditional and Emerging Cardiovascular Risk Factors

Objective: To determine whether law enforcement officer (LEO) status and perceived stress are longitudinal predictors of traditional and inflammatory cardiovascular (CV) risk factors. Method: Linear hierarchical regression was employed to investigate the longitudinal (more than 7 years) relationship between occupational category (LEO vs non-LEO) and perceived stress scale scores, and traditional and inflammatory CV risk factors in an all-male sample of 105 LEOs and 65 non-LEOs. Results: The occupational status of LEOs, compared with that of non-LEOs, predicted higher levels of C-reactive protein, systolic blood pressure, body mass index, and waist circumference. Perceived stress across occupational categories was directly associated with diastolic blood pressure and waist circumference and inversely with fibrinogen. Perceived stress did not interact with occupational category to predict any risk factor. Conclusion: Traditional and inflammatory risk factors, but not perceived stress, appear to contribute to elevated CV risk among LEOs.

Student Perceptions of Stress, Coping, Relationships, and Academic Civility A Longitudinal Study

Academic incivility can increase student stress, jeopardize learning, damage relationships, and negatively impact the academic environment. This 3-year longitudinal study measured a cohort of prelicensure nursing students’ progressive perceptions of stress, coping, student-student and faculty-student relationships, and levels of academic civility. While civility scores remained mild to moderately high overall, there was a slightly declining trend over the 3-year period. Perceived stressors and coping strategies and ways to improve academic civility are identified and discussed.

Longitudinal measurement invariance of the metabolic syndrome: is the assessment of the metabolic syndrome stable over time?

PURPOSE Without verification of longitudinal measurement invariance, researchers cannot be certain whether observed change in the metabolic syndrome reflects true change or changes in assessment or structure of the construct over time. This research tested longitudinal measurement invariance of a 1-factor model of the metabolic syndrome during the course of 6 years. METHODS Tests of longitudinal measurement invariance (configural, metric, and scalar) were conducted on 604 men and women who participated in the Spokane Heart Study from 1996 to 2006. Metabolic syndrome indicators included body mass index, triglycerides, high-density lipoprotein, diastolic blood pressure, and fasting glucose. RESULTS Sequential configural and metric invariance models demonstrated adequate model fit, but the scalar invariance model led to a decrement in fit. Therefore, the theoretical framework of the syndrome and the relationships between the syndrome construct and the indicators appear to be equivalent over time. However, observed values of the metabolic syndrome indicators may differ across time when there is a constant level of the syndrome. CONCLUSIONS Because longitudinal invariance was not fully demonstrated, interpretation of change in the metabolic syndrome over time may be misleading because change may be partly attributable to measurement properties of the indicators. However, a cross-sectional 1-factor model of the metabolic syndrome is supported.

Psychometric properties of the Birth Satisfaction Scale-Revised (BSS-R) for US mothers

Objective: The aims of this study were to continue the scale development process of the Birth Satisfaction Scale-Revised (BSS-R) by refining the scale to make it culturally relevant for US participants, examining the factor structure of the BSS-R, and describing the level of birth satisfaction in a sample of US mothers. Background: The Birth Satisfaction Scale (BSS) was developed in the UK to assess satisfaction of the childbearing women’s experiences of labour and its outcomes. One of the goals of the development of the BSS was to make comparisons across cultures. Methods: One hundred and eighty-one first-time US mothers participated in this study. Confirmatory factor analysis was used to examine a one-factor higher-order model containing three lower-order factors. The higher-order factor was hypothesised to be Experience of childbearing; the lower-order factors were hypothesised as Stress, Quality of Care and Women’s attributes. Results: The results of the higher-order factor model indicated good fit, χ2 = 37.72, p = .22; comparative fit index (CFI) = .99; root mean square error of approximation (RMSEA) = .03; standardised root mean square residual (SRMR) = .04. Cronbach’s α indicated the subscales and total scale were reliable for the US sample (α ranged from .74 to .89). The BSS-R total score was 15.52 (SD = 8.35), and the stress, quality of care, and women’s attributes subscales were 7.15 (SD = 3.80), 4.61 (SD = 3.83) and 3.79 (SD = 2.26), respectively. Conclusion: The BSS-R can provide maternal health professionals and researchers with an instrument to quantify childbearing women’s birthing satisfaction, which in turn can assist in heightening the overall patient care experience.

A ‘missing not at random’ (MNAR) and ‘missing at random’ (MAR) growth model comparison with a buprenorphine/naloxone clinical trial

AIMS To compare three missing data strategies: (i) the latent growth model that assumes the data are missing at random (MAR) model; (ii) the Diggle-Kenward missing not at random (MNAR) model, where dropout is a function of previous/concurrent urinalysis (UA) submissions; and (iii) the Wu-Carroll MNAR model where dropout is a function of the growth factors. DESIGN : Secondary data analysis of a National Drug Abuse Treatment Clinical Trials Network trial that examined a 7-day versus 28-day taper (i.e. stepwise decrease in buprenorphine/naloxone) on the likelihood of submitting an opioid-positive UA during treatment. SETTING 11 out-patient treatment settings in 10 US cities. PARTICIPANTS A total of 516 opioid-dependent participants. MEASUREMENTS Opioid UAs provided across the 4-week treatment period. FINDINGS The MAR model showed a significant effect (B = -0.45, P < 0.05) of trial arm on the opioid-positive UA slope (i.e. 28-day taper participants were less likely to submit a positive UA over time) with a small effect size (d = 0.20). The MNAR Diggle-Kenward model demonstrated a significant (B = -0.64, P < 0.01) effect of trial arm on the slope with a large effect size (d = 0.82). The MNAR Wu-Carroll model showed a significant (B = -0.41, P < 0.05) effect of trial arm on the UA slope that was relatively small (d = 0.31). CONCLUSIONS This performance comparison of three missing data strategies (latent growth model, Diggle-Kenward selection model, Wu-Carrol selection model) on sample data indicates a need for increased use of sensitivity analyses in clinical trial research. Given the potential sensitivity of the trial arm effect to missing data assumptions, it is critical for researchers to consider whether the assumptions associated with each model are defensible.

Empowering Patients with Persistent Pain Using an Internet-based Self-Management Program

New strategies are needed to improve access to cognitive and behavioral therapies for patients with persistent pain. The purpose of this randomized, controlled trial was to determine the effectiveness of the Chronic Pain Management Program, an 8-week online intervention targeting cognitive, emotional, behavioral, and social pain determinants. Program efficacy and engagement was evaluated for 92 individuals with a diagnosis of chronic noncancer pain who had a current opioid prescription. Participants were recruited from primary care practices and Internet sites, then randomly assigned to receive access to the intervention either immediately (treatment group) or after an 8-week delay (wait-list comparison). Biweekly self-report measurements were collected using online surveys on pain, depressive symptoms, pain self-management behaviors, and health care utilization during the 8-week trial. Additional measurements of opioid misuse behaviors, pain self-efficacy, and medicine regimens were completed at baseline and week 8. Engagement was evaluated by examining completion of program learning modules. The results from analysis of variance showed that at week 8, the treatment group had significantly greater improvements on pain self-efficacy and opioid misuse measures than the wait-list comparison group. Engagement level was positively associated with improvements in pain intensity, pain interference, and pain self-efficacy. In conclusion, patients on opioids were able to engage and demonstrate positive outcomes using an Internet-based self-management program. Future efforts toward heightening engagement could further maximize impacts.

Co-occurring mood disorders among hospitalized patients and risk for subsequent medical hospitalization

OBJECTIVE The objective was to determine if patients hospitalized with a primary medical diagnosis and any co-occurring serious mental illness (SMI) were more likely than patients without any co-occurring SMI diagnosis to experience a subsequent medical hospitalization. METHOD This was a longitudinal cohort study of 925,705 adult persons (aged 18+ years). Patients hospitalized in Washington State from 2004 to 2008 were followed through 2009 (for an average of 43 months). RESULTS Compared to patients hospitalized for medical conditions without co-occurring SMI, patients with co-occurring dysthymia, bipolar and major depressive disorders were at an elevated risk for long-term subsequent hospitalization. Patients in the combined co-occurring mood disorders cohort were more likely (hazard ratio=1.13; 99% confidence interval=1.10-1.16; P<.001) than patients in the reference cohort to experience a subsequent medical hospitalization. A significant interaction between substance and mood disorders that increased risk for subsequent hospitalization was also observed. CONCLUSION Hospitalized patients with co-occurring mood disorders are at high risk for repeat hospitalization for a medical reason. This high-risk population, including those with substance abuse, should be a focus of research efforts to identify and address ambulatory-care-sensitive conditions amenable to strategies that decrease complications and illness leading to subsequent hospitalizations.

Development and psychometric evaluation of the genomic nursing concept inventory.

Translation of genome science to improve health outcomes requires nurses to develop genomic competency and literacy, and a robust measure of genomic literacy is needed to advance evidence-based nursing education. This study aimed to develop and evaluate the Genomic Nursing Concept Inventory (GNCI), which is a scale to measure understanding of the genetic/genomic concepts most critical to nursing practice. Applying a multistep process, key concepts were drawn from essential nursing genetic/genomic competencies and validated by expert opinion. Surveys and cognitive interviews of baccalaureate nursing (BSN) students informed item development. A 52-item draft inventory was administered to 238 BSN students. Item analysis informed inventory reduction, and the resulting 31-item inventory was tested with 705 BSN students. Scale difficulty was 47%, item difficulty 13% to 84%, and Cronbach’s alpha 0.77. As scale refinement proceeds, the GNCI provides a useful measure of genomic literacy to inform curriculum design and evaluate outcomes in genomic nursing education.