
Message/Author 


I have access to seven years of data. Each year has a varying number of patients  some are measured each year, while some patients enter or exit the healthcare system at different years and are thus captured for a varying number of years. Within a given year, each patient is measured on the outcome of interest varying amounts of time. So, one patient might be measured 100 times while another 0 times. The outcome is in minutes. Can the outcomes per individual be aggregated per year and treated as continuous? Substantive theory supports heterogeneity of different classes in the population. Can this data be modeled using GMM? 


How you would handle this would depend on the time unit over which development occurs. If you think it is annual, you could aggregate over year. If you think another time unit is appropriate, you could consider that. 


Thank you for the prompt response and taking the time to address such a basic query! The outcome is case management that is measured in increments of 15, 30, or 45 minute sessions, each 15 minute increment coded separately. Since I'm interested in looking at change over a seven year period, what I'd like to do is combine the number of sessions per individual so that I have a total number of minutes per individual per year. 

Back to top 

