Longitudinal analysis based on age
- Study: Influence of social network on occurrence of dementia: a community-based longitudinal study.
- Link:https://www.ncbi.nlm.nih.gov/pubmed/10776744
- Research question: Do single social network components and different degrees of the social connections affect dementia incidence?
- Study sample: A community cohort of 1203 non-demented people, living at home in a district of Stockholm, Sweden who had good cognition were derived from a longitudinal population-based study of all inhabitants of this district over age 75 in October 1987.
- Longitudinal design: At the beginning of the study, participants were examined and interviewed. The follow-up examination was completed an average of three-years after this baseline interview. The examination included collection of family and personal history, clinical examination, and psychological tests. If the individual was not able to answer, an informant was interviewed and if an individual had move he or she was traced and asked to participate. For those who had died, hospital records and death certificates were collected.
- Analysis approach: Cox proportional hazard models were used to estimate the relative risks and corresponding 95% CI of developing dementia in relations to different social network components. A series of four steps, each considering the social-network variables in different ways: each category of every indicator, grouping categories in each indicator, grouping indicators into factors, and a summary index. The association was assessed using a univariate Cox proportional hazard model and then with multivariate models. Confounders used as covariates in the models included: age, sex, education, baseline MMSE score, symptomatic depression, ADL index, and vascular disease.
Longitudinal analysis based on time since study enrollment
- Study: Longitudinal study of vision and retinal nerve fiber layer thickness in multiple sclerosis
- Link: http://onlinelibrary.wiley.com/doi/10.1002/ana.22005/full
- Research question: Is there a relationship between visual loss and RNFL thinning over time in MS patients?
- Study sample: The treatment group included 299 patients with MS (593 eyes) with at least one follow-up visit and over 6 months of follow-up time. The control group was recruited from among staff and family of patients and had no history of ocular or neurologic disease.
- Longitudinal design: Following an initial baseline visit, patients were invited to follow-up at 6- to 12-month intervals. A past history of acute optic neuritis was determined by self- and physician-report and confirmed by record review. Types of disease modifying therapies were recorded at study visits.
- Analysis approach: GEE models, accounting for age and adjusting for within patient, intereye correlations, were used to determine the relation between follow-up period and PNFL thickness from baseline. Each one-year follow-up interval was plotted against change in RNFL, with each eye represented only once in the model. Logistic regression models, accounting for age and adjusting for within patient intereye correlations were used to assess the association of RNFL thinning with losses of acuity and examine the effects of disease duration and disease-modifying therapies on the relations of RNFL thinning to visual loss and length of follow-up. To examine whether eyes with longer follow up were more likely to have visual of RNFL loss beyond levels expected on test-retest variability, the chi-square test for trend in binomial proportions was used.
-
Longitudinal analysis based on time since initial treatment
- Study: Alcoholism Treatment and Total Health Care Utilization and Costs
- Link: http://jamanetwork.com/journals/jama/fullarticle/362261
- Research question: How does alcoholism treatment affect overall health care utilization and costs?
- Study sample: The treatment group was made up of 1645 families who filed claims for alcoholism treatment under the Federal Employees Health Benefit Program between 1980 and 1983 through Aetna Life and Casualty Company based on their limited classification system and were continuously covered under the insurance plan for the duration of the study period. The control group was a randomly selected group of 3598 continuously enrolled families that had filed no claims for alcoholism treatment during the study period
- Longitudinal design: Medical care claims for both groups for services rendered during the period of January 1980 through September 1983 were analyzed
- Analysis approach: Four-year average per capita costs were compared for families with and without an alcoholic member. And based on identification of first alcoholism treatment date, costs were calculated and averages compared between pre- and post-treatment periods of similar lengths. No statistically significant differences in the demographics of cohorts compared were found