Browsing by UMass Chan Affiliation "New England Research Institute"

Now showing items 1-3 of 3

    • A comparison of anlaytic methods for non-random missingness of outcome data
      Crawford, Sybil L.; Tennstedt, S L; McKinlay, John B. (1995-02-01)
      Missing outcome values occur frequently in survey data and are rarely missing randomly. Depending on the pattern of missingness, the choice of analytic method has implications for accuracy of the estimated outcome distribution as well as multivariate models. Data from a study of patterns of care in disabled elders were used to evaluate several common methods when missingness of the outcome was nonrandom. Results from single and multiple model-based imputation were compared with results from complete-case analysis and mean imputation. By ignoring nonrespondents' covariate information, the latter two methods yielded biased estimates of population means. Mean imputation and single model-based imputation underestimated standard errors by treating imputed values as if they were observed. Mean imputation also distorted the relationship between the outcome and predictors. Multiple model-based imputation provided an easily implemented method of adjustment for non-random non-response in both univariate and multivariate analyses.

    • Do characteristics of informal caregivers affect the length of hospital stay for frail elders
      Skinner, K M; Tennstedt, S L; Crawford, Sybil L. (1994-05-01)
      Studies examining the length of hospital stay for the very old rarely include characteristics of the informal caregiver as potentially explanatory variables. This study expanded a frequently used conceptual model of health services use to include caregiving factors as enabling variables. Data were collected from a representative sample of frail elders who had been hospitalized in the previous 18 months (n = 181) and their primary informal caregivers. The findings indicate that elders whose primary caregiver was a coresiding offspring had shorter hospitalization than elders whose offspring caregiver lived in different households. These data indicate that the amount of variance explained by this conceptual model is increased when social factors are included as explanatory variables.

    • Longitudinal patterns and correlates of hormone replacement therapy use in middle-aged women
      Johannes, Catherine B.; Crawford, Sybil L.; Posner, J G; McKinlay, S M (1994-09-01)
      Patterns of hormone replacement therapy (HRT) use over time and predictors of initiating or discontinuing use were examined in a longitudinal study. A cohort of 2,425 women aged 45-55 years identified from a population-based random survey in Massachusetts in 1981-1982 was followed by six telephone interviews, 9 months apart. Cohort participants were either premenopausal (66.8%) or in early perimenopause (33.2%). During the study, prevalence of use was low overall (12.3%) and was considerably higher for surgical menopause (45%) than for peri- (9.3%), natural (4.5%), or premenopause (1.5%). Predictors of HRT uptake and discontinuation (from time t - 1 to time t) were examined by repeated-measures logistic regression, stratified by surgical status. For surgical menopause, the only significant predictor of HRT uptake was recent surgery (odds ratio = 4.4; 95% confidence interval 2.73-7.22), while for nonsurgical subjects, menopausal status (primarily perimenopause), prior use of HRT, health care utilization, hot flashes, alcohol consumption, regular exercise, and leaner body mass were all significant predictors of uptake. Discontinuing HRT was inversely associated with surgical and perimenopause and positively related to prior short-term use and health care utilization. Nonsurgical HRT users had a somewhat more favorable cardiovascular risk profile than did nonusers.