-------------------------------------------------------------------------------
help for evhistplot
-------------------------------------------------------------------------------

Plot individual event histories with one or two time scales (Lexis diagram)

        evhistplot event_varlist [if] [in] , id(varname) start(date_string)
             end(date_string) [evtype(varname) birth(varname) nsub(integer
             100) agescale(real 365.25) scatter options]

Description

    evhistplot plots individual event histories with different event types
    indicated by differently colored dots resulting in one string of beads
    for each individual. Different event types may reside in different
    variables in event_varlist, or they may be identified in the
    evtype(typevar) option.

    When birth(varname) is not specified, the individual histories are given
    as horizontal lines with calendar time on the X-axis. When birth(varname)
    is specified, the individual histories are plotted in a Lexis diagram
    with calendar time on the X-axis and age on the Y-axis.


Options

    id(varname[, plot]) specifies the variable holding the identifier of each
        individual. If the option plot is given, then the original ID numbers
        are used on the Y-axis for identifying individuals in plots with
        horizontal lines. If the plot option is omitted, then the n
        individuals in the plot are associated with random integers 1,...,n.

    start(date_string) specifies the starting time of the observation window.
        The date must be specified in the format of the d(l) function, cf.
        the help for d(). All events before this date are omitted from the
        plot.

    end(date_string) specifies the end time of the observation window. The
        date must be specified in the format of the d(l) function, cf. the
        help for d(). All events after this date are omitted from the plot.

    evtype(varname) is optional and specifies a variable holding the type of
        event recorded in the variable given as main argument. If more
        variables are given in varlist, then the event types refer to the
        first variable only of varlist. The remaining variables of varlist
        then each holds a unique type of events.

    birth(varname) is optional and specifies a variable holding birth dates
        of individuals. If specified, a Lexis diagram is plotted with
        calendar time on the X-axis and age on the Y-axis.

    nsub(integer) is optional and limits the number of individuals included
        in the plot to the specified count. Subjects are chosen at random for
        inclusion in the graph. If not specified, all subjects are included.

    agescale(real) is optional and indicates the scale on the age axis of the
        Lexis diagram, ie. the Y-axis. Requires specification of birth(
        varname) to have effect. The default value corresponds to age in
        years, when variables hold date values.

    scatter_options are optional and allow fine tuning of the resulting plot.
        In particular, the msize and msymbol options are useful for modifying
        the size and shape of markers used.


Examples

    Consider a dataset where the variable recdate holds dispensing dates for
    medications for a group of patients identified by a unique key in the
    variable patid. Dates of birth and death are held in variables with
    obvious names.

    . evhistplot recdate birthdate deathdate, id(patid) start(1jan1993)
        end(31dec2000) (births and deaths are only plotted for those born in
        the observation window)

    . evhistplot recdate birthdate deathdate, id(patid) start(1jan1993)
        end(31dec2000) evtype(medictype) (different medications are plotted
        in different colors)

    . evhistplot recdate deathdate, id(patid) start(1jan1993) end(31dec2000)
        evtype(medictype) birth(birthdate) (makes a Lexis-diagram)

    The ancillary file -rrdat1_plot.do- shows how the commands can be used to
    plot individual event histories with respect to jobs and marriage.

Author

    Henrik Stovring (hstovring@health.sdu.dk), Research Unit for General
    Practice, University of Southern Denmark, Denmark.

Also see


    Online: help for twoway, scatter