The Accuracy of Computerized Adaptive Testing in Heterogeneous Populations: A Mixture Item-Response Theory Analysis.

BACKGROUND: Computerized adaptive testing (CAT) utilizes latent variable measurement model parameters that are typically assumed to be equivalently applicable to all people. Biased latent variable scores may be obtained in samples that are heterogeneous with respect to a specified measurement model. We examined the implications of sample heterogeneity with respect to CAT-predicted patient-reported outcomes (PRO) scores for the measurement of pain. METHODS: A latent variable mixture modeling (LVMM) analysis was conducted using data collected from a heterogeneous sample of people in British Columbia, Canada, who were administered the 36 pain domain items of the CAT-5D-QOL. The fitted LVMM was then used to produce data for a simulation analysis. We evaluated bias by comparing the referent PRO scores of the LVMM with PRO scores predicted by a "conventional" CAT (ignoring heterogeneity) and a LVMM-based "mixture" CAT (accommodating heterogeneity). RESULTS: The LVMM analysis indicated support for three latent classes with class proportions of 0.25, 0.30 and 0.45, which suggests that the sample was heterogeneous. The simulation analyses revealed differences between the referent PRO scores and the PRO scores produced by the "conventional" CAT. The "mixture" CAT produced PRO scores that were nearly equivalent to the referent scores. CONCLUSION: Bias in PRO scores based on latent variable models may result when population heterogeneity is ignored. Improved accuracy could be obtained by using CATs that are parameterized using LVMM.

Sex is associated with differences in individual trajectories of change in social health after implantable cardioverter-defibrillator.

BACKGROUND: Social health is a dimension of quality of life, and refers to people's involvement in, and satisfaction with social roles, responsibilities, and activities. The implantable cardioverter-defibrillator is associated with changes in overall quality of life, but little is known about sex differences in individual trajectories of change in social health. METHODS AND RESULTS: We prospectively measured changes in 3 subscales of the SF-36v2 generic health questionnaire (role physical, role emotional, and social functioning), 2 Patient-Reported Outcomes Measurement Information System short forms (satisfaction with participation in social roles and satisfaction with participation in discretionary social activities), and the Florida Patient Acceptance Survey before and at 1, 2, and 6 months after implantation. Individual growth models of temporal change were estimated. The scores of the 6 indicators improved with time. The unconditional model demonstrated significant (fixed effects: P<0.05; covariance parameters: P<0.10) residual variability in the individual trajectories. In the conditional model, men and women differed significantly in their rates of change in the scores of 3 of the 6 measures. Although men's mean scores exceeded women's mean scores on all indicators at baseline (range of relative mean difference: 11.0% to 17.8%), the rate of women's change resulted in a reversal in relative standing at 6 months after implantation, with the mean scores of women exceeding the men's by 4.5% to 5.6%. CONCLUSIONS: Men and women differed in their trajectories of change in social health, both in terms of their starting points (ie, baseline scores) and their rates of change.

Ethnic and gender differences in patient education about heart disease risk and prevention.

OBJECTIVE: To investigate whether there are gender and ethnic disparities in the patient education provided by primary healthcare providers about heart disease (HD) risk and prevention. METHODS: A telephone survey, conducted in four languages, was completed by 976 people, 40+ years of age, in Metro Vancouver, Canada. Questions assessing communication with healthcare providers' provision of HD risk and management education were the focus. RESULTS: Statistically significant gender and ethnic differences were found. Women were less likely to report discussing HD risk and management with their healthcare providers. Chinese-Canadian participants had less likelihood of receiving HD education compared with participants of other ethnic origins. These differences persisted after multivariate adjustment with income, highest level of education attained, age, and other factors. CONCLUSION: Primary healthcare providers should make improved efforts towards education about HD and its risk factors for women in general, and for postmenopausal women especially. PRACTICE IMPLICATIONS: Healthcare providers should be aware that some ethnic populations may not be receiving patient education similar to that received by people of other communities, as found for Chinese-Canadian members of this study community. Further understanding of the barriers faced by ethnic groups must be gained to develop solutions.

The Influence of Web- Versus Paper-based Formats on the Assessment of Tobacco Dependence: Evaluating the Measurement Invariance of the Dimensions of Tobacco Dependence Scale.

The purpose of this study was to examine the influence of mode of administration (internet-based, web survey format versus pencil-and-paper format) on responses to the Dimensions of Tobacco Dependence Scale (DTDS). Responses from 1,484 adolescents that reported using tobacco (mean age 16 years) were examined; 354 (23.9%) participants completed a web-based version and 1,130 (76.1%) completed a paper-based version of the survey. Both surveys were completed in supervised classroom environments. Use of the web-based format was associated with significantly shorter completion times and a small but statistically significant increase in the number of missing responses. Tests of measurement invariance indicated that using a web-based mode of administration did not influence the psychometric functioning of the DTDS. There were no significant differences between the web- and paper-based groups' ratings of the survey's length, their question comprehension, and their response accuracy. Overall, the results of the study support the equivalence of scores obtained from web- and paper-based versions of the DTDS in secondary school settings.

Toward an explanation of observed ethnic differences in youths' tobacco use.

We examined whether differences in smoking rates of Asian and white/Caucasian youth could be explained by personal (gender, employment status, volunteerism, parental education, and income) and social factors including differences in youths' relationships with their parents, extent of enculturation, and exposure to parental or peer smoking. A survey was conducted of a random sample of schools in 2 cities in British Columbia, Canada to obtain data from 3,278 high school students. Results from logistic regression analysis indicated smoking status was explained by place of birth, volunteerism, amount of income received from parents and employers, characteristics of the parental-child relationship, and parental and peer smoking status. Differences in the estimated risks of smoking of Asian youth and white youth were moderated by the youths' willingness to tell their parents about their lives, whether they worked for pay, and whether the Asian youth spoke English at home.

Motives for smoking cessation are associated with stage of readiness to quit smoking and sociodemographics among German industrial employees.

PURPOSE: To test the relationships among particular motives for smoking cessation, stage of readiness to quit (preparation or contemplation), and sociodemographic characteristics. DESIGN: A cross-sectional study to examine attitudes toward and use of health promotion at the worksite, using a self-administered questionnaire. SETTING: Two German metal companies. SUBJECTS: Of 1641 responding employees (response rate 65% in company A and 44% in company B), 360 smokers who intended to quit immediately (n = 105) or in the near future (n = 255) were analyzed. MEASURES: The questionnaire comprised of sociodemographic characteristics, smoking behavior, smoking history, readiness to quit smoking, motives to quit, such as coworkers' complaints and health-related or financial concerns. Chi-squared tests and multiple logistic regression analyses were performed. RESULTS: Health-related reasons (94%) predominated financial (27%) or image-related (14%) reasons for smoking cessation. Participants in the cessation preparation group were more likely to report an awareness of being addicted (79.6% vs. 58.2%; p < .001) and the negative public image (22.5% vs. 11.6%; p < .01) as reasons for quitting compared with those in the contemplation group. In multivariable regression models, the motives for smoking cessation, including reduced performance, family's and coworkers' complaints, pregnancy/children, and negative public image, but not health-related and financial concerns, differed significantly by gender, age, marital status, education, and occupational status. CONCLUSIONS: Motives for smoking cessation vary according to the individual's level of readiness to quit and sociodemographic background.

The role of mediating factors in the association between social deprivation and low birth weight in Germany.

This study examines whether the association between social inequalities and low birth weight (LBW) (occurring in both pre- and full-term births) in Germany can be explained by several potentially confounding factors. These include maternal age, occupational status, marital status, nationality, employment status, smoking, prenatal care, psychosocial stress, obesity, short stature, short inter-pregnancy interval, chronic conditions, and several obstetrical risk factors such as pregnancy induced hypertension. We also examined how the risk for LBW varies over time within each socioeconomic group. We analyzed routinely collected perinatal data on singletons born in the federal state of Lower Saxony, Germany, in 1990, 1995, and 1999 (n = 182,444). After adjustment for all potentially confounding factors in multivariate logistic regression models, working class women, unemployed women, single mothers, and women over 39 years of age were at increased risk for pre- and full-term LBW infants. Migrant status was not related to LBW. We examined variations in the risk for LBW over time within groups, using the 1990 birth cohort as the referent group for the 1995 and 1999 birth cohorts. Compared to 1990, in 1999 women aged 19-34 years, housewives, unemployed women, women of German nationality and women with partners had higher risks for pre- and full-term LBW infants; the eldest subgroup had lower risks for LBW after adjustment for confounding factors. The factors we examined partly explain the social inequalities in LBW occurring in pre- and full-term infants. The subgroups with higher rates of LBW in 1999 compared to 1990, included women experiencing childbirth in an optimal stage of life or in a privileged social context. Public health policies in Germany should target social inequalities contributing to the aetiology of LBW and to the factors that result in increased LBW rates.

Socio-demographic correlates of cigarette smoking among high school students: results from the British Columbia youth survey on smoking and health.

OBJECTIVE: To describe the association between selected socio-demographic factors (age, gender, ethnicity, and region) and the prevalence of smoking among adolescents in two regions of British Columbia, and to report recent findings related to the prevalence of tobacco use in British Columbia. METHODS: A cross-sectional school-based survey was conducted using a random sample of 3,280 students from 13 schools in two regions of British Columbia. Multivariate logistic regression analysis was used to determine the association between age, region, gender and ethnicity and smoking status. RESULTS: Ethnic group membership was strongly associated with smoking status, which changed very little after controlling for the other socio-demographic factors. Controlling for age and ethnicity, the effect of gender on smoking status was moderated by region. CONCLUSION: Reliance on general population tobacco use prevalence rates masks important ethnic and gender differences. To increase the effectiveness of tobacco control policies and programming, greater attention needs to be paid to the socio-demographic correlates associated with adolescents' tobacco use.

Psychometric assessment of the Multidimensional Quality of Life Questionnaire for Persons with HIV/AIDS (MQOL-HIV) in a sample of HIV-infected women.

Since the late 1980s, several HIV-specific quality of life instruments have been developed; however, little testing has been done in terms of their validity and reliability for HIV-infected women. The purpose of this study was to test the content validity, concurrent validity, internal consistency, and test-retest reliability of the Multidimensional Quality of Life Questionnaire for Persons with HIV/AIDS (MQOL-HIV) in a sample of 85 HIV-infected women. The MQOL-HIV is a 40-item scale comprised of 10 dimensions. Most of the items and all of the domains were determined content valid but revision of some of the items and domains is recommended. Concurrent validity was measured between the MQOL-HIV and the MOS-HIV and ranged from 0.51-0.81 between similar domains. Of the 10 domains and the entire instrument, 7 had a Cronbach's alpha over 0.70 (range 0.43-0.92). Eight domains and the entire instrument achieved test-retest correlation coefficients over 0.70 (range 0.60-0.96). Although some revision may make the scale more content-valid for HIV-infected women, given due care in the interpretation of results, the MQOL-HIV can be used with female populations in its current form.

Workload as a determinant of staff injury in intermediate care.

Care aides (CAs) and licensed practical nurses (LPNs) in intermediate care (IC) facilities experience high risks of injury. Workload measures were examined in eight IC facilities to ascertain their relationships with musculoskeletal injury rates, pain, burnout, and self-reported health. Workload was measured using (1) focus groups and telephone interviews to obtain CA/LPN perceptions; (2) systematic observation (numbers of tasks performed in shift); and (3) CA-to-resident staffing ratios. Controlling for resident dependency and facility funding, all workload measures were correlated with staff injury rates and burnout. Facilities with low injury rates had better CA/LPN-to-resident staffing ratios and fewer tasks. The differences in staffing reflected differences in how organizations prioritized and allocated resources. Thus, workload is an important determinant of injuries and increased staffing levels correlate with decreased injuries.

Feasibility of using existing Statistics Canada surveys to describe the health and work of nurses.

Reorganization of nurses' work has raised questions about the effects of working conditions on their health. Nurses, for example, are more likely to miss work because of illness and disability than employees in other occupations. The overall purpose of this descriptive study was to investigate the feasibility of using existing Statistics Canada surveys regularly to describe and monitor the health and working conditions of nurses. Our findings identified significant limitations in existing Statistics Canada surveys, for the study of nurses, including nonspecific or no occupational coding, small samples and partial content related to the work environment. As a result, some estimates would need to be accompanied by statements indicating that the findings do not meet quality standards and that the conclusions would be unreliable and most likely invalid. Additional data are required for a comprehensive assessment of the health status of nurses and the work environment factors that influence their health. These data can be obtained through several vehicles, including using over-sampling strategies for extant and recurring Statistics Canada surveys, adding additional content to those surveys or implementing new surveys specific to nurses and their work. The authors describe the advantages and disadvantages of each of these approaches and conclude that monitoring the health and work environment of nurses in Canada in sufficient detail to inform policy decisions requires a dedicated national survey.

Implementation of provincial/territorial health goals in Canada.

During the past two decades, policy makers in most of Canada's provinces and territories developed broad population-level goal statements about desired health or health and social outcomes. The health goals development process used in each province/territory has been described in government documents and studied by a small number of researchers. However, there is a lack of published research examining the implementation and use of the health goals since they were developed. To begin to fill this gap, we conducted a study between 1998 and 2000 that examined the implementation of provincial/territorial health goals in Canada. Our findings indicate that as the 1990s drew to a close, provincial/territorial health goals were not being used explicitly by policy makers at either provincial/territorial or regional levels in most provinces in Canada to guide health policy and program development, implementation, or evaluation. Instead, the majority of health ministry and regional policy makers were employing strategic/business plans that, at best, reflected or were similar to the original provincial/territorial health goals. Moreover, even though all provinces and the NWT/Nunavut had health goals associated with broad social, economic, and physical environment health determinants, regional-level policy makers were giving priority to health care system goals over all other types of goals. We discuss our findings in relation to studies about health goals in other countries, and we suggest implications that our findings have for both future research and health policy.

Non-consensual sex experienced by men who have sex with men: prevalence and association with mental health.

Little is known about the psychosocial factors associated with sexual assault experienced by males. Men (N=358), 19-35 years of age, recruited by community outreach, completed questionnaires. Eligibility criteria included: being HIV-negative and self-identifying as gay or bisexual. Lifetime prevalence rates of childhood sexual abuse, juvenile prostitution, and adult sexual assault were determined. The mental health of this population was explored including associations between sexual victimization and mental health disorders (alcohol abuse, suicidal ideation and attempts, mood disorders, and poor self-esteem). Almost 1 in 10 of the men had engaged in juvenile prostitution, 14% were forced into sexual activity before 14 years of age, and 14% were sexually victimized after the age of 14. Those exposed to non-consensual sex were 2.9 (95% CI: 1.8-4.7) times more likely to abuse alcohol than those free of victimization. Those who reported childhood sexual abuse were 3.3 (95% CI: 1.7-6.4) times more likely to have attempted suicide. Juvenile prostitution was associated with current depression (OR=6.4; 95% CI: 2.8-14.9). Health professionals have the responsibility to respond competently and sensitively to victims of sexual violence. To do this, many need to recognize the prevalence of male sexual trauma, to deconstruct their personal beliefs about same-sex sexual violence, and to learn to ask sensitive questions in their assessment interviews.