Analysis of count data using poisson regression

Factors affecting the Haemophilia Joint Health Score in children with severe haemophilia

Joint damage from bleeding episodes leads to physical or functional limitations in people with haemophilia. Various factors may influence the frequency and severity of joint damage. This study examined whether age, prophylaxis, history of high-titre inhibitors (HTI) and bleeding events influenced the Haemophilia Joint Health Score (HJHS) in children. Medical and physiotherapy notes of boys with severe haemophilia, aged 4-18 years, were reviewed to identify factors associated with increased HJHS. The HJHS of 83 boys (median age: 11) ranged from 0 to 25, with 44/83 (53%) having a score of zero. The median HJHS was 0 (mean 2.6). In the non-HTI group, the HJHS for boys on late prophylaxis was 2.68 times higher than those who started early and the HJHS was on average 10% higher for every additional recent bleed. In this group the odds of having a zero score fell by 30% for every year increase in age. Boys with a history of HTI had higher HJHS scores than the non-HTI group, and age, number of recent bleeds and tolerized status were positively associated with HJHS. The score rose on average by 28% for every year of age and by 76% for non-tolerized boys. This study provides further evidence supporting early prophylaxis use and the importance of immune tolerance therapy. The HJHS is a useful tool for identifying and tracking changes in joint health with respect to therapy or disease progression. With improvements in haemophilia treatment, the disproportionate number of zero scores will continue to make interpretation of the HJHS challenging.

Prevalence of women's violent and nonviolent offending behavior: a comparison of self-reports, victims' reports, and third-party reports

This study assessed women's violent and nonviolent offending, using data from two online student samples (men and women: n = 344), reporting on either being a perpetrator and witness (women) or being a victim and witness (men). A comprehensive measure of general violence, intimate partner violence (IPV), and nonviolent offending was collected. From women's self-reports, 59.9% reported perpetrating general violent offenses, 58.1% reported perpetrating IPV offenses and 85.6% reported perpetrating nonviolent offenses. Correlations showed that women were involved in a variety of offenses and demonstrated the interrelatedness of general violence and IPV, and of violent and nonviolent offenses. Regression analysis confirmed the close association between partner and general violence, and found that drug offenses were also related to the former and criminal damage to the latter. Overall, the prevalence data demonstrated women's involvement in all types of offending, and a similar pattern of offending was supported across data sources. Limitations of the sampling method and measures are discussed.

The influence of sexually explicit Internet material on sexual risk behavior: a comparison of adolescents and adults

This study had three goals: first, to investigate whether sexually explicit Internet material (SEIM) affects sexual risk behavior; second, to study whether these effects differ between adolescents and adults; and third, to analyze, separately for adolescents and adults, whether gender and age moderate an influence of SEIM on sexual risk behavior. The authors conducted a 2-wave panel survey among nationally representative random samples of 1,445 Dutch adolescents and 833 Dutch adults. SEIM use increased sexual risk behavior among adults, but not among adolescents. More specifically, moderator analyses showed that SEIM use increased sexual risk behavior only among male adults, but not among female adults. In the adolescent sample, no moderating gender effect occurred. Neither among adolescents nor among adults did age moderate the effects. Our study shows that SEIM may influence outcomes related to people's sexual health. It also suggests that male adults may present a potential risk group for adverse effects of SEIM.

Prime Time: 12-month sexual health outcomes of a clinic-based intervention to prevent pregnancy risk behaviors

PURPOSE

Prime Time, a youth development intervention, aims to reduce pregnancy risk among adolescent girls seeking clinic service who are at high risk for pregnancy. This article examines sexual risk behaviors and hypothesized psychosocial mediators after 12 months of the Prime Time intervention.

METHODS

This study was a randomized controlled trial with 253 girls aged 13-17 years meeting specified risk criteria. Intervention participants were involved in Prime Time programming plus usual clinic services for 18 months, control participants received usual clinic services. The intervention used a combination of case management and peer leadership programs. Participants in this interim outcomes study completed self-report surveys at baseline and 12 months after enrollment. Surveys assessed sexual risk behaviors and psychosocial factors targeted for change by Prime Time.

RESULTS

At the 12-month interim, the intervention group reported more consistent use of condoms, hormonal contraception, and dual contraceptive methods with their most recent partner as compared with the control group. The intervention group also reported greater stress management skills with trends toward higher levels of prosocial connectedness at school and with family. No between-group differences were noted in psychosocial measures specific to sex and contraceptive use.

CONCLUSION

Preventing early pregnancy among high-risk adolescents requires multifaceted, sustained approaches. An important research focus involves testing youth development interventions offered through clinic settings, where access to high-risk adolescents is plentiful and few efforts have emphasized a dual approach of building protective factors while addressing risk. Findings suggest that youth development interventions through clinic settings hold promise in reducing pregnancy risk among high-risk youth.

HIV infection as a predictor of methadone maintenance outcomes in Chinese injection drug users

This paper's design is descriptive and correlational based on retrospective self-report survey data collected in Kunming city, China. The study investigated the difference between a group of Chinese HIV positive (N=36) and negative (N=131) opioid dependent adults maintained on methadone treatment. Comparisons were based on their quality of life (QOL), methadone treatment adherence, adverse symptom occurrence related to methadone treatment, and HIV-related behavior changes. No significant differences were found between the two groups in age, methadone maintenance dose, methadone adherence, sex desire, and drug craving level. Participants who were HIV positive reported significantly lower scores on physical health and total health-related qualify of life. They also reported greater engagement in injection related risk behavior before methadone treatment than those who tested HIV negative. For both groups, sexual and injection risk behavior significantly decreased following initiation of methadone treatment. A regression model revealed that those infected with HIV, associated significantly with higher likelihood of reporting constipation and lack of appetite, and higher frequency of reporting abdominal pain and nausea than HIV negative patients. The primary implication of these findings is that HIV positive persons in methadone treatment may require more focused services to meet their special HIV care and substance treatment needs.

Skin Disease and Stigma in Emerging Adulthood: Impact on Healthy Development

Background:

Visible skin disorders can limit healthy psychosocial development in several domains owing to the stigma these disorders create.

Objective:

To assess if emerging adults with acne and eczema perceive stigma in diverse developmental domains resulting from their visible skin conditions.

Methods:

A convenience sample of 336 emerging adults attending a diverse public university in northern California completed a Web-based survey. Multivariate modeling was used to determine if eczema and acne result in perceived stigma.

Results:

The mean (SD) age of respondents was 20.1(1.5) years; most were white (43.2%) and female (62.5%). Almost half (47%) reported acne and 16.4% reported eczema. The likelihood of experiencing perceived stigma was 1.6 (95% CI 1.13–2.27, p < .001) times higher for those with eczema compared to those without eczema, and the likelihood of experiencing perceived stigma was 3.19 (95% CI 2.41–4.22, p < .0001) times higher for those with acne compared to those without acne.

Conclusions:

Our findings indicate that emerging adults perceive stigma as a result of their acne and eczema in several important developmental domains. The results suggest that acne may have a stronger relationship with perceived stigma than eczema, perhaps owing to its greater visibility on the body. University-based activities are recommended to support those with skin disease.

Does cost-benefit analysis or self-control predict involvement in two forms of aggression?

The main aim of this research was to assess the relative association between physical aggression and (1) self-control and (2) cost-benefit assessment, these variables representing the operation of impulsive and reflective processes. Study 1 involved direct and indirect aggression among young Indian men, and Study 2 physical aggression to dating partners among Spanish adolescents. In Study 1, perceived benefits and costs but not self-control were associated with direct aggression at other men, and the association remained when their close association with indirect aggression was controlled. In Study 2, benefits and self-control showed significant and independent associations (positive for benefits, negative for self-control) with physical aggression at other-sex partners. Although being victimized was also correlated in the same direction with self-control and benefits, perpetration and being victimized were highly correlated, and there was no association between being victimized and these variables when perpetration was controlled. These results support the theory that reflective (cost-benefit analyses) processes and impulsive (self-control) processes operate in parallel in affecting aggression. The finding that male adolescents perceived more costs and fewer benefits from physical aggression to a partner than female adolescents did is consistent with findings indicating greater social disapproval of men hitting women than vice versa, rather than with the view that male violence to women is facilitated by internalized patriarchal values.

Is physician engagement with Web-based CME associated with patients' baseline hemoglobin A1c levels? The Rural Diabetes Online Care study

PURPOSE

To investigate the association between physician participants' levels of engagement in a Web-based educational intervention and their patients' baseline diabetes measures.

METHOD

The authors conducted a randomized trial of online CME activities designed to improve diabetes care provided by family, general, and internal medicine physicians in rural areas of 11 southeastern states between September 2006 and July 2008. Using incidence rate ratios derived from negative binomial models, the relationship between physicians' engagement with the study Web site and baseline proportion of their patients having controlled diabetes (hemoglobin A1c < or = 7%) was explored.

RESULTS

One hundred thirty-three participants (intervention = 64; control = 69) provided information for 1,637 patients with diabetes. In the intervention group, physicians in practices in the worst quartiles of A1c control were least engaged with the study Web site in nearly all dimensions. Total number of pages viewed decreased as quartile of A1c control worsened (137, 73, 68, 57; P = .007); similarly, for a given 10% increase in proportion of patients with controlled A1c, participants viewed 1.13 times more pages (95% CI: 1.02-1.26, P = .02). In the control group, engagement was neither correlated with A1c control nor different across quartiles of A1c control.

CONCLUSIONS

Engagement in Web-based interventions is measurable and has important implications for research and education. Because physicians of patients with the greatest need for improvement in A1c control may not use online educational resources as intensely as others, other strategies may be necessary to engage these physicians in professional development activities.

Role of executive dysfunction in predicting frequency and severity of violence

The adverse consequences of violence on society are tremendous. The proportion of offenders incarcerated for violent offenses is large, and the cost of keeping these offenders incarcerated is startling. Understanding and treating the causal underpinnings of violent crime is of utmost importance for individuals and society as a whole. Several factors have been identified as potential contributors to violent crime, including cognitive deficits in executive functioning [Hoaken et al., 2007]. To investigate this further, 77 offenders from Fenbrook Institution, a federal facility, were tested on a battery of executive functioning measures. Offenders were found to have broad and pervasive dysfunction in their executive abilities. In addition, specific scores from the battery were found using regression techniques, to predict the frequency and severity of past violent offending but not nonviolent offending. This speaks of the possibility of a new type of correctional rehabilitation program, one that focuses on the rehabilitation of basic executive functions.

Determinants of the number of antenatal visits in a metropolitan region

BACKGROUND

Antenatal care has a positive effect on pregnancy, both clinically and psychologically, but consensus about the optimal number of antenatal visits is lacking. This study aims to provide insight into the dynamics of the number of antenatal visits a woman receives. Independent effects of predisposing, enabling and pregnancy-related determinants are examined.

METHODS

Women were recruited in nine clinical centres in the Brussels Metropolitan region. Antenatal care use was measured prospectively. A Poisson regression model was applied to measure the independent effect of individual determinants on the number of antenatal visits.

RESULTS

Data on antenatal care trajectories in 333 women were collected. The multivariate analyses showed that women with a Maghreb or Turkish origin had 14% fewer visits compared with European (EU15) women. More highly educated women had 22% more visits compared with those with a low education. Women with a high income had 14% more antenatal visits compared with those with a low income. Fewer antenatal visits were observed in multiparae (15%), women initiating care after 14 weeks of gestation (31%), women without medical risks during the pregnancy (12%) and in women with a continuity of care index of 50% or more (12%). More visits were observed in delivering after week 37 (22% increase).

CONCLUSIONS

Predisposing and enabling factors have to be considered when antenatal care programmes are evaluated in a metropolitan area. Variations in the number of antenatal visits show that socially vulnerable women are more at risk of having fewer visits.

Evidence-based practice implementation: the impact of public versus private sector organization type on organizational support, provider attitudes, and adoption of evidence-based practice

BACKGROUND

The goal of this study is to extend research on evidence-based practice (EBP) implementation by examining the impact of organizational type (public versus private) and organizational support for EBP on provider attitudes toward EBP and EBP use. Both organization theory and theory of innovation uptake and individual adoption of EBP guide the approach and analyses in this study. We anticipated that private sector organizations would provide greater levels of organizational support for EBPs leading to more positive provider attitudes towards EBPs and EBP use. We also expected attitudes toward EBPs to mediate the association of organizational support and EBP use.

METHODS

Participants were mental health service providers from 17 communities in 16 states in the United States (n = 170). Path analyses were conducted to compare three theoretical models of the impact of organization type on organizational support for EBP and of organizational support on provider attitudes toward EBP and EBP use.

RESULTS

Consistent with our predictions, private agencies provided greater support for EBP implementation, and staff working for private agencies reported more positive attitudes toward adopting EBPs. Organizational support for EBP partially mediated the association of organization type on provider attitudes toward EBP. Organizational support was significantly positively associated with attitudes toward EBP and EBP use in practice.

CONCLUSION

This study offers further support for the importance of organizational context as an influence on organizational support for EBP and provider attitudes toward adopting EBP. The study demonstrates the role organizational support in provider use of EBP in practice. This study also suggests that organizational support for innovation is a malleable factor in supporting use of EBP. Greater attention should be paid to organizational influences that can facilitate the dissemination and implementation of EBPs in community settings.

Cognitive-behavioral therapy for panic disorder in patients being treated for alcohol dependence: Moderating effects of alcohol outcome expectancies

Anxiety disorders commonly co-occur with alcohol use disorders and reliably mark a poor response to substance abuse treatment. However, treating a co-occurring anxiety disorder does not reliably improve substance abuse treatment outcomes. Failure to account for individual differences in the functional dynamic between anxiety symptoms and drinking behavior might impede the progress and clarity of this research program. For example, while both theory and research point to the moderating role of tension-reduction alcohol outcome expectancies (TR-AOEs) in the association between anxiety symptoms and alcohol use, relevant treatment studies have not typically modeled TR-AOE effects. We examined the impact of a hybrid cognitive-behavioral therapy (H-CBT) treatment for panic disorder (independent variable) on response to a community-based alcohol dependence treatment program (dependent variable) in patients with higher vs. lower TR-AOEs (moderator). The H-CBT treatment was generally effective in relieving participants' panic symptoms relative to controls. However, TR-AOEs interacted with study cohort (H-CBT vs. control) in predicting response to substance abuse treatment. As expected, the H-CBT was most effective in improving alcohol use outcomes among those with the highest TR-AOEs. The study's primary methodological limitations are related to the quasi-experimental design employed.

In search of the evidentiary foundation of published Canadian economic evaluations (2001–06)

Objectives:The aim of this study was to present a review of economic evaluations conducted from a Canadian perspective and to characterize sources of evidence and statistical methods to analyze effectiveness measures, resource utilization, and uncertainty.

Methods:A search strategy was developed to identify Canadian economic evaluations published between January 2001 and June 2006. A standardized abstraction form was used to extract key data (e.g., study design, data sources, statistical methods).

Results:A total of 153 unique studies were included for review, of which 75 were evaluations of drug therapies and less than half were funded by industry. Cost-effectiveness analysis was the most common type of economic evaluation and 80 percent of the studies used modeling techniques. A single source of evidence for effectiveness measures was used in half of the studies. Statistical methods were commonly reported to compare effectiveness measures when the economic evaluation was conducted alongside a clinical trial but less commonly when determining effectiveness input parameters in model-based economic evaluations, or to analyze resource utilization data. Authors relied mostly on univariate sensitivity analyses to explore uncertainty.

Conclusions:This review identifies the need to improve the conduct and reporting of statistical methods for economic evaluations to improve confidence in the results.

Personality and intimate partner aggression in dating relationships: the role of the "Big Five"

Although personality is shown to predict negative relationship experiences, few researchers have used a structural model of personality to study the ways that personality contributes to intimate partner aggression (IPA). This study investigates the five-factor model of personality and its associations with both the use and receipt of psychological, physical, and sexual IPA in 179 men and 301 women. Each of the five factors of personality was associated with at least one type of IPA perpetration or victimization. The dimensions of neuroticism and agreeableness were the strongest predictors of IPA particularly for women. Results are discussed in terms of why personality should be considered as a predictor for both the use and receipt of IPA, why sex differences emerged, and future research that should be conducted.

Recidivism in HIV-infected incarcerated adults: influence of the lack of a high school education

Recidivism is a pervasive problem facing the incarcerated. Incarcerated persons who are human immunodeficiency virus (HIV)-infected often have multiple risk factors associated with initial incarceration and recidivism, in particular, injection drug use. Yet, some jails provide case management for HIV-infected inmates to provide continuity of health care, which might have positive effects on reentry into the community. We sought to measure recidivism and factors related to recidivism in an HIV-infected cohort in an urban county jail with an active case management program. Fifty-two inmates surveyed in 1999 at the San Francisco County Jail were followed for rearrests through 2006. In follow-up, 73% were re-incarcerated on an average of 6.8 times for 552 days. Risk factors included nonwhite ethnicity, history of homelessness and crack use, common risk factors for incarceration. Less than high school education was associated with recidivism, shorter time to reincarceration, and more incarcerations. HIV-infected inmates spend a high proportion of time in multiple incarcerations, a reflection of the cyclical nature of incarceration despite comprehensive case management. Well-known risk factors for incarceration were associated with recidivism; in addition, lack of high school education played a prominent role. Education should be explored as a way to make further progress on breaking the cycle of incarceration.

Correlates of leisure-time physical activity in Korean immigrant women

This study describes the physical activity behavior of Korean immigrant women and examines the relationships among leisure-time physical activity (LTPA) behavior, background, and intrapersonal correlates of behavior in Korean immigrant women in the United States using a cross-sectional survey design. A convenience sample of Midwestern Korean immigrant women completed the long form of the International Physical Activity Questionnaire in Korean. Among the participants, 78% were physically active, meeting the Healthy People 2010 goal for physical activity when all physical activity behaviors were considered, whereas 23% met the goal when only LTPA was considered. Women who were older, did not have a child younger than 5, used environmental resources for physical activity, had lower depressive symptoms, and had higher confidence for overcoming exercise barriers were spent more time in LTPA. Accommodating those with young children, enhancing environmental resource use, and improving exercise self-efficacy are important considerations for interventions aimed at increasing LTPA in Korean immigrant women.

Feeding skill performance in premature infants during the first year

BACKGROUND

Little is known about premature infants' feeding skill development and the contribution to it of biologic and environmental conditions.

AIMS

Explore the level and variation in feeding skill performance through the first post-term year and examine the contribution to performance of infant neonatal condition and rate of weight gain per day, maternal feeding behavior, and its interaction with neonatal condition.

STUDY DESIGN AND SUBJECTS

In this longitudinal, descriptive study, data sources included observed and videotaped in-home feeding for 45 infants <1250 g birth weight and their mothers (age > or =17 years).

OUTCOME MEASURE

Feeding skill performance (fdgskill): total number of expected skills at 1, 4, 8, and 12 months, post-term age (PTA).

RESULTS

Feeding skill performance varied widely among infants at all four assessments. At 8 and 12 months, fdgskill indicated, for a minority of infants, delay and lack of opportunity to engage in skills associated with new foods and new feeding modalities. Neonatal medical condition contributed significantly to fdgskill at 1 and 4 months, but in the predicted (negative) direction only at four months. Rate of weight gain per day contributed significantly to fdgskill at 1 and 8 months, but in the predicted direction (positive) only at one month. Maternal feeding behavior did not contribute to fdgskill, nor did it interact with infant neonatal conditions to affect fdgskill.

CONCLUSIONS

Although infant neonatal medical status and rate of weight gain per day, before or within the span of time between assessments, accounted for some variance in feeding skill performance within the first three assessments (1, 4, and 8 months), much remains to be explained, including neuro- and oral-motor capacities to manage new foods and feeding modalities and opportunities to practice feeding skills during the last half of the first year.

Methods for Improving Regression Analysis for Skewed Continuous or Counted Responses

Standard inference procedures for regression analysis make assumptions that are rarely satisfied in practice. Adjustments must be made to insure the validity of statistical inference. These adjustments, known for many years, are used routinely by some health researchers but not by others. We review some of these methods and give an example of their use in a health services study for a continuous and a count outcome. For the continuous outcome, we describe re-transformation using the smear factor, accounting for missing cases via multiple imputation and attrition weights and improving results with bootstrap methods. For the count outcome, we describe zero inflated Poisson and negative binomial models and the two-part model to account for overabundance of zero values. Recent advances in computing and software development have produced user-friendly computer programs that enable the data analyst to improve prediction and inference based on regression analysis.

Overdispersion in the Poisson Regression Model

Abstract. This simulation study compares different strategies to solve the problem of underestimating standard errors in the Poisson regression model when overdispersion is present. The study analyses the importance of sample size, Poisson distribution mean, and dispersion parameter in choosing the best index or estimate. Results show that standard error (SE) estimates obtained by resampling (nonparametric bootstrap and jackknife) are the least biased, followed by the direct index based on the χ2, and the so-called robust indexes, in third place. Nevertheless, the inefficiency of resampling estimates is also evident, especially in small samples.

Impact of pneumococcal vaccination on pneumonia rates in patients with COPD and asthma

BACKGROUND

Patients with chronic obstructive pulmonary disease (COPD) are included in several national pneumococcal vaccination recommendations whereas asthma patients are not. The objective of this study was to evaluate pneumonia-related hospitalization risk in patients with COPD or asthma and vaccination impact.

METHODS

We identified patients with documented pneumococcal vaccination from a cohort of veterans with either a diagnosis of asthma or COPD and their matched controls. Patients were identified between October 1, 1997 and September 30, 1998 and followed for 5 years. For each group we identified pneumococcal pneumonia hospitalizations and all pneumonia-related hospitalizations in the periods before and after vaccination. We estimated hospitalization rates and compared rates in the asthma and COPD groups to controls using negative binomial regression models.

RESULTS

We identified 16,074 COPD patients (average age 65.8 years), 14,028 controls for the COPD patients (average age 67.5 years), 2,746 asthma patients (average age 53.0 years), and 1,345 controls for the asthma patients (average age 59.2 years). Compared to controls, the adjusted risk of pneumococcal pneumonia hospitalizations before pneumococcal vaccination was COPD = 8.02 (95% CI, 4.44-14.48) and asthma = 0.76 (0.17-3.53). For any pneumonia-related hospitalization, the adjusted risk was COPD = 3.91 (3.40-4.50) and asthma = 1.45 (0.85-2.46). After vaccination, events decreased in all groups. The adjusted risk for pneumococcal pneumonia hospitalizations postvaccination was COPD = 3.87 (2.55-5.88) and asthma = 0.30 (0.04-1.99). For any pneumonia-related hospitalization the adjusted risk was COPD = 3.71 (3.33-4.13) and asthma = 0.79 (0.50-1.25).

CONCLUSIONS

This study supports the value of vaccinating COPD patients; however, the value of vaccination for asthma patients is less certain.

Human Papillomavirus Type 16 and 18 Variants: Race-Related Distribution and Persistence

BACKGROUND

Analogous to the geographic distribution of variants of human papillomavirus (HPV) types, the distribution and persistence of these variants among infected individuals may be related to the racial composition of a population living in one geographic region.

METHODS

We studied 1114 women in the United States participating in the Atypical Squamous Cells of Undetermined Significance/Low-Grade Squamous Intraepithelial Lesion Triage Study who were positive for HPV16 and/or HPV18 at enrollment. Race was self-reported. HPV samples were characterized by sequencing the E6 gene and part of the long control region and classified as variants according to established lineages. Subjects were examined for HPV every 6 months for 2 years. All statistical tests were two-sided.

RESULTS

HPV18 African variants were predominant in the 97 HPV18-infected African American women (i.e., 62 women or 63.9%, 95% confidence interval [CI] = 53.5% to 73.4%), and European variants were common in the 168 HPV18-infected white women (i.e., 91 women or 54.2%, 95% CI = 46.3% to 61.9%). HPV16 African variants accounted for 43 (26.5%, 95% CI = 19.9% to 34.0%) of the infections in the 162 HPV16-infected African American women but for only 25 (4.3%, 95% CI = 2.8% to 6.3%) in the 584 HPV16-infected white women. The likelihood of remaining HPV18 positive was statistically significantly higher in white women infected with European than in white women infected with African variants (P = .04); the reverse was observed in African American women (P = .03). A similar pattern was observed for persistence of HPV16 variants, with the likelihood of remaining positive being higher for white women, but lower for African American women, infected with an European variant than with an African variant (P = .03 and P = .16, respectively).

CONCLUSIONS

Variants of HPV16 and HPV18 appear to persist longer in a host whose race indicates an ancestral geographic distribution that was once shared with that of the variant-i.e., European variants persist longer in white women, and African variants persist longer in African American women.