Too much ado about two-part models and transformation?

Buntin, Melinda Beeuwkes; Zaslavsky, Alan M.

doi:10.1016/j.jhealeco.2003.10.005

Cited by 609 publications

(226 citation statements)

References 13 publications

Supporting

Mentioning

224

Contrasting

Order By: Relevance

“…However, the results of these models did not differ substantively from the gamma regression models. We only report the results of the gamma regression models because these results can be interpreted directly in terms of the original scale of the dependent variable (Buntin & Zaslavsky, 2004). In the series of gamma regression models, the first model included only demographic characteristics as explanatory variables.…”

Section: Discussionmentioning

confidence: 99%

Cost of health care utilization among homeless frequent emergency department users.

Mitchell

León

Byrne

et al. 2017

Psychological Services

View full text Add to dashboard Cite

Research demonstrates that homelessness is associated with frequent use of emergency department (ED) services, yet prior studies have not adequately examined the relationship between frequent ED use and utilization of non-ED health care services among those experiencing homelessness. There has also been little effort to assess heterogeneity among homeless individuals who make frequent use of ED services. To address these gaps, the present study used Medicaid claims data from 2010 to estimate the association between the number of ED visits and non-ED health care costs for a cohort of 6,338 Boston Health Care for the Homeless Program patients, and to identify distinct subgroups of persons in this cohort who made frequent use of ED services based on their clinical and demographic characteristics. A series of gamma regression models found more frequent ED use to be associated with higher non-ED costs, even after adjusting for demographic and clinical characteristics. Latent class analysis was used to examine heterogeneity among frequent ED users, and the results identified 6 characteristically distinct subgroups among these persons. The subgroup of persons with trimorbid illness had non-ED costs that far exceeded members of all 5 other subgroups. Study findings reinforce the connection between frequent ED use and high health care costs among homeless individuals and suggest that different groups of homeless frequent ED users may benefit from interventions that vary in terms of their composition and intensity.

show abstract

Section: Discussionmentioning

confidence: 99%

Cost of health care utilization among homeless frequent emergency department users.

Mitchell

León

Byrne

et al. 2017

Psychological Services

View full text Add to dashboard Cite

show abstract

“…Costs have thus been estimated using a one-model generalized linear model with link = log and gamma distribution [37]. A Conditional Logit Model was used for estimating OR and significance for comorbidity per added BMI point above 30.…”

Section: Methodsmentioning

confidence: 99%

The Socioeconomic Burden of Obesity

Kjellberg¹,

Larsen

Ibsen

et al. 2017

Obes Facts

View full text Add to dashboard Cite

Objective: To evaluate the socioeconomic impact of obesity by estimating the direct and indirect costs associated with obesity in Denmark, based on individual level data. Methods: Costs were assessed for different BMI groups, and the relative risks for change in direct and indirect costs per BMI point above 30 were estimated. A fourth analysis estimated the odds ratio for comorbidities per BMI point above 30. Individual data on income, social transfer payments, healthcare costs and diagnoses were retrieved from national registries. Results: One BMI point above 30 was associated with a 2% decrease in income, a 3% increase in social transfer payments, and a 4% increase in healthcare costs. In absolute numbers, income contributed to most of the total economic burden. One BMI point above 30 was also associated with increased comorbidity, which explains the increase in both direct and indirect costs. Conclusion: Obesity is associated with increased comorbidity, giving rise to an increase in both direct and indirect costs. Especially income is affected, which emphasizes the importance of including both measures when evaluating the total socioeconomic burden of obesity. Our findings draw attention to the potential for saving public resources and preventing loss of income by preventing obesity.

show abstract

“…We implemented a two-part model (23,24,25), first using simultaneous logistic regressions to determine whether treatment system (mental health or substance abuse treatment) was a significant predictor of utilization. We then conducted zero-truncated negative binomial regressions (26) to determine whether treatment system was a significant predictor of amounts of services received that were measured in number of days (counts).…”

Section: Methodsmentioning

confidence: 99%

Disparities in Use of Mental Health and Substance Abuse Services by Persons With Co-occurring Disorders

Havassy

Alvidrez²,

Mericle³

2009

Psychiatric Services

View full text Add to dashboard Cite

Objectives Individuals with co-occurring mental and substance use disorders require psychiatric and substance abuse treatments. A critical question is whether these individuals are treated for both disorders. Methods This study prospectively examined 24-month service utilization patterns of 224 persons with co-occurring disorders who were recruited from crisis residential programs in the mental health treatment system (N=106) and from crisis residential detoxification programs in the substance abuse treatment system (N=118) in San Francisco. Utilization data were collected from the billing-information systems of both treatment systems. Demographic and clinical data were obtained in interviews with participants. Data were analyzed for group differences with chi square tests and logistic, linear, and zero-truncated negative binomial regression. Results After the analyses controlled for demographic and clinical factors, participants recruited from the substance abuse treatment system were less likely than those from the mental health treatment system to obtain any mental health services, mental health day treatment, transitional residential care, case management, and other outpatient services (p<.001 for all comparisons). They were more likely to obtain crisis residential detoxification (p=.003), had more days of drug residential treatment (p=.028), but received fewer hours of outpatient services (p=.012). Conclusions There were disparities in patterns of service utilization, although there were no significant diagnostic differences between the two groups. These findings should be valuable in considering systems development and modification. Furthermore, they can contribute to research about factors that underlie results. Study replications should be conducted to assess the robustness of these findings in other locales. (Psychiatric Services 60:217–223, 2009)

show abstract

Too much ado about two-part models and transformation?

Cited by 609 publications

References 13 publications

Cost of health care utilization among homeless frequent emergency department users.

Cost of health care utilization among homeless frequent emergency department users.

The Socioeconomic Burden of Obesity

Disparities in Use of Mental Health and Substance Abuse Services by Persons With Co-occurring Disorders

Contact Info

Product

Resources

About