Interactively plot a given patient's data, allowing the real-time exploration
of the various CMAs and their parameters.
It can use Rstudio's manipulate library or Shiny.
plot_interactive_cma(data = NULL, ID = NULL, cma.class = c("simple",
"per episode", "sliding window")[1], print.full.params = FALSE,
ID.colname = NA, event.date.colname = NA,
event.duration.colname = NA, event.daily.dose.colname = NA,
medication.class.colname = NA, date.format = "%m/%d/%Y",
followup.window.start.max = 5 * 365, followup.window.duration.max = 5
* 365, observation.window.start.max = followup.window.start.max,
observation.window.duration.max = followup.window.duration.max,
align.all.patients = FALSE, align.first.event.at.zero = TRUE,
maximum.permissible.gap.max = 2 * 365,
sliding.window.start.max = followup.window.start.max,
sliding.window.duration.max = 2 * 365,
sliding.window.step.duration.max = 2 * 365, backend = c("shiny",
"rstudio")[1], use.system.browser = FALSE,
get.colnames.fnc = function(d) names(d),
get.patients.fnc = function(d, idcol) unique(d[[idcol]]),
get.data.for.patients.fnc = function(patientid, d, idcol, cols = NA,
maxrows = NA) d[d[[idcol]] %in% patientid, ], ...)Usually a data.frame containing the events (prescribing
or dispensing) used to compute the CMA. Must contain, at a minimum, the patient
unique ID, the event date and duration, and might also contain the daily
dosage and medication type (the actual column names are defined in the
following four parameters). Alternatively, this can be any other data source
(for example, a connection to a database), in which case the user must redefine
the arguments get.colnames.fnc, get.patients.fnc and
get.data.for.patients.fnc appropriately. Currently, this works only when
using Shiny for interactive rendering. For a working example, please see
the vignette describing the interfacing with databases.
The ID (as given in the ID.colname column) of the patient
whose data to interactively plot (if absent, pick the first one); please not
that this an be interactively selected during plotting.
The type of CMAs to plot; can be "simple" (CMA0 to CMA9), "per episode", or "sliding window".
A logical specifying if the values of all the parameters used to generate the current plot should be printed in the console (if TRUE, it can generate extremely verbose output!).
A string, the name of the column in data
containing the unique patient ID, or NA if not defined.
A string, the name of the column in
data containing the start date of the event (in the format given in
the date.format parameter), or NA if not defined.
A string, the name of the column in
data containing the event duration (in days), or NA if not
defined.
A string, the name of the column in
data containing the prescribed daily dose, or NA if not defined.
A string, the name of the column in
data containing the classes/types/groups of medication, or NA
if not defined.
A string giving the format of the dates used in the
data and the other parameters; see the format parameters of the
as.Date function for details (NB, this concerns only the
dates given as strings and not as Date objects).
The maximum number of days when the follow-up window can start.
The maximum duration of the follow-up window in days.
The maximum number of days when the observation window can start.
The maximum duration of the observation window in days.
Should the patients be aligend?
Should the first event be put at zero?
The maximum permissible gap in days.
The maximum number of days when the sliding windows can start.
The maximum duration of the sliding windows in days.
The maximum sliding window step in days.
The plotting backend to use; "shiny" (the default) tries to use the Shiny framework, while "rstudio" uses the manipulate RStudio capability.
For shiny, use the system browser?
A function taking as parameter the data source
and returning the column names. Must be overridden when the data source is
not derived from a data.frame.
A function taking as parameter the data source
and the patient ID column name, and returns the list of all patient IDs.
Must be overridden when the data source is not derived from a data.frame.
A function taking as parameter a (set
of) patient ID(s), the data source, and the patient ID column name, and returns
the list of all patient IDs. Must be overridden when the data source is not
derived from a data.frame.
Extra arguments.
Nothing
The manipulate is kept for backward compatibility only, as it is much
more limited than Shiny and will receive no new development in the
future.
Shiny currently allows the use of any other data source besides a
default (and usual) data.frame (or derived), such a connection to an
SQL database. In this case, the user must redefine the three
argument functions get.colnames.fnc, get.patients.fnc and
get.data.for.patients.fnc which collectively define an interface for
listing the column names, all the patient IDs, and for retreiving the actual
data for a (set of) patient ID(s). A fully worked example is described in
the vignette detailing the access to standard databases storaging the
patient information.
For more info please see the online vignette https://htmlpreview.github.io/?https://github.com/ddediu/AdhereR/blob/master/online-only-doc/adherer_interactive_plots/adherer_interctive_plots.html.
The online vignette https://htmlpreview.github.io/?https://github.com/ddediu/AdhereR/blob/master/online-only-doc/adherer_interactive_plots/adherer_interctive_plots.html.
# NOT RUN {
plot_interactive_cma(med.events,
ID.colname="PATIENT_ID",
event.date.colname="DATE",
event.duration.colname="DURATION",
event.daily.dose.colname="PERDAY",
medication.class.colname="CATEGORY");
# }
Run the code above in your browser using DataLab