Diagnostics for Multiple Imputation in Stata

Eddings, Wesley; Marchenko, Yulia

doi:10.1177/1536867x1201200301

Cited by 93 publications

(71 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…After an initial 5 imputation to test convergence, we increased imputation to 60 and burn-in iteration by 50. We carried out imputation sensitivity analysis and checked the fit of the imputation model using the Stata command “Midiagplots”[42]. For all analyses using imputed data, estimates were combined across the imputed datasets based to Rubin's rules[43].…”

Section: Methodsmentioning

confidence: 99%

Outcomes of Nigeria's HIV/AIDS Treatment Program for Patients Initiated on Antiretroviral Treatment between 2004-2012

et al. 2016

View full text Add to dashboard Cite

BackgroundThe Nigerian Antiretroviral therapy (ART) program started in 2004 and now ranks among the largest in Africa. However, nationally representative data on outcomes have not been reported.MethodsWe evaluated retrospective cohort data from a nationally representative sample of adults aged ≥15 years who initiated ART during 2004 to 2012. Data were abstracted from 3,496 patient records at 35 sites selected using probability-proportional-to-size (PPS) sampling. Analyses were weighted and controlled for the complex survey design. The main outcome measures were mortality, loss to follow-up (LTFU), and retention (the proportion alive and on ART). Potential predictors of attrition were assessed using competing risk regression models.ResultsAt ART initiation, 66.4 percent (%) were females, median age was 33 years, median weight 56 kg, median CD4 count 161 cells/mm3, and 47.1% had stage III/IV disease. The percentage of patients retained at 12, 24, 36 and 48 months was 81.2%, 74.4%, 67.2%, and 61.7%, respectively. Over 10,088 person-years of ART, mortality, LTFU, and overall attrition (mortality, LTFU, and treatment stop) rates were 1.1 (95% confidence interval (CI): 0.7–1.8), 12.3 (95%CI: 8.9–17.0), and 13.9 (95% CI: 10.4–18.5) per 100 person-years (py) respectively. Highest attrition rates of 55.4/100py were witnessed in the first 3 months on ART. Predictors of LTFU included: lower-than-secondary level education (reference: Tertiary), care in North-East and South-South regions (reference: North-Central), presence of moderate/severe anemia, symptomatic functional status, and baseline weight <45kg. Predictor of mortality was WHO stage higher than stage I. Male sex, severe anemia, and care in a small clinic were associated with both mortality and LTFU.ConclusionModerate/Advanced HIV disease was predictive of attrition; earlier ART initiation could improve program outcomes. Retention interventions targeting men and those with lower levels of education are needed. Further research to understand geographic and clinic size variations with outcome is warranted.

show abstract

Section: Methodsmentioning

confidence: 99%

Outcomes of Nigeria's HIV/AIDS Treatment Program for Patients Initiated on Antiretroviral Treatment between 2004-2012

et al. 2016

View full text Add to dashboard Cite

show abstract

“…In the imputation model, we included all explanatory variables listed above, including the outcome variable and the Nelson‐Aalenestimate of the cumulative hazard to the survival time for each individual outcome assessment . We conducted diagnostics using the midiagplots function in Stata . Within each of the 10 imputed datasets and using all the covariates listed above, we calculated a propensity score, wherein AA was the reference group for drug comparison (Appendix A).…”

Section: Methodsmentioning

confidence: 99%

The comparative effectiveness of fourth‐line drugs in resistant hypertension: An application in electronic health record data

Sinnott

Smeeth

Williamson

et al. 2019

Pharmacoepidemiology and Drug

View full text Add to dashboard Cite

Purpose To examine the utility of electronic health records from a routine care setting in assessing comparative effectiveness of fourth‐line anti‐hypertensive drugs to treat resistant hypertension. Methods We conducted a cohort study using the Clinical Practice Research Datalink: a repository of electronic health records from UK primary care. We identified patients newly prescribed fourth‐line anti‐hypertensive drugs (aldosterone antagonist , beta‐blocker, or alpha‐blocker). Using propensity score–adjusted Cox proportional hazards models, we compared the incidence of the primary outcome (composite of all‐cause mortality, stroke, and myocardial infarction) between patients on different fourth‐line drugs. AA was the reference drug in all comparisons. Secondary outcomes were individual components of the primary outcome, blood pressure changes, and heart failure. We used a negative control outcome, Herpes Zoster, to detect unmeasured confounding. Results Overall, 8639 patients were included. In propensity score–adjusted analyses, the hazard ratio for the primary outcome was 0.81 (95% CI, 0.55‐1.19) for beta‐blockers and 0.68 (95% CI, 0.46‐0.96) for alpha‐blockers versus AA. Findings for individual cardiovascular outcomes trended in a more plausible direction, albeit imprecise. A trend for a protective effect for Herpes Zoster across both comparisons was seen. Conclusions A higher rate of all‐cause death in the AA group was likely due to unmeasured confounding in our analysis of the composite primary outcome, supported by our negative outcome analysis. Results for cardiovascular outcomes were plausible, but imprecise due to small cohort sizes and a low number of observed outcomes.

show abstract

“…Missing data were addressed via multiple imputation in which twenty complete data sets, a number recommended to help ensure stable parameter and standard error estimates (Graham, Olchowski, and Gilreath 2007), were constructed using Stata’s implementation of multiple imputation with chained equations (StataCorp 2013). Diagnostic tests of the imputed data indicated that the values did not appreciably depart from the distributions of the variables in the original dataset (Eddings and Marchenko 2012). …”

Section: Methodsmentioning

confidence: 99%

Parental and adolescent health behaviors and pathways to adulthood

Bauldry

Shanahan

Macmillan

et al. 2016

Social Science Research

View full text Add to dashboard Cite

This paper examines associations among parental and adolescent health behaviors and pathways to adulthood. Using data from the National Longitudinal Study of Adolescent to Adult Health, we identify a set of latent classes describing pathways into adulthood and examine health-related predictors of these pathways. The identified pathways are consistent with prior research using other sources of data. Results also show that both adolescent and parental health behaviors differentiate pathways. Parental and adolescent smoking are associated with lowered probability of the higher education pathway and higher likelihood of the work and the work & family pathways (entry into the workforce soon after high school completion). Adolescent drinking is positively associated with the work pathway and the higher education pathway, but decreases the likelihood of the work & family pathway. Neither parental nor adolescent obesity are associated with any of the pathways to adulthood. When combined, parental/adolescent smoking and adolescent drinking are associated with displacement from the basic institutions of school, work, and family.

show abstract

Diagnostics for Multiple Imputation in Stata

Cited by 93 publications

References 9 publications

Outcomes of Nigeria's HIV/AIDS Treatment Program for Patients Initiated on Antiretroviral Treatment between 2004-2012

Outcomes of Nigeria's HIV/AIDS Treatment Program for Patients Initiated on Antiretroviral Treatment between 2004-2012

The comparative effectiveness of fourth‐line drugs in resistant hypertension: An application in electronic health record data

Parental and adolescent health behaviors and pathways to adulthood

Contact Info

Product

Resources

About