There are many instances in healthcare where
the data collected for a performance indicator
is date, time or date/time. For the sake of
simplicity in this discussion, the three possibilities
will be referred to collectively as date/time.
Analyzing date/time values or plotting dates
and time on a control chart must be done with
careful planning. Date/time values cannot be
treated the same as any number. The main issue
is that dates and time are not decimal numbers.
Attempting to perform any type of mathematics
on date/time values produces erroneous results
unless they are used in a way that is consistent
with other numeric values.
Statit has features that simplify the handling
of date/time. For example, if the variable is
defined as a time variable, then the data is
stored and plotted correctly. The following
data set is an example of this. This data is
the time that hospice workers checked in at
the start of the day. The first column of data
is the actual time stamp of the check-in. The
second column is the value that the computer
uses for time calculations. For a time stamp,
the computer translates it to the number of
seconds elapsed from a standard value, which
is midnight. The first row shows 36060 seconds
since midnight. This is the number of seconds
that have elapsed between midnight and 10:01
am. It is not necessary to display the seconds
in Statit. It is done automatically for the
user by defining the Check_In_Time column as
a time variable.
| Check-in Time |
Seconds |
| 10:01:00 |
36060 |
| 09:42:00 |
34920 |
| 07:15:00 |
26100 |
| 07:35:00 |
27300 |
| 08:11:00 |
29460 |
| 07:59:00 |
28740 |
| 08:45:00 |
31500 |
| 10:34:00 |
38040 |
| 08:24:00 |
30240 |
| 08:19:00 |
29940 |
| 09:02:00 |
32520 |
| 08:01:00 |
28860 |
| 08:50:00 |
31800 |
| 07:55:00 |
28500 |
| 08:28:00 |
30480 |
| 09:02:00 |
32520 |
| 10:14:00 |
36840 |
| 09:21:00 |
33660 |
| 08:12:00 |
29520 |
Producing an i-chart on this data is very
straight forward. Since the data are time stamps,
it is appropriate to use variable charts. The
measurement of time is a continuous or parametric
value. Time is typically rounded to minutes
or seconds for convenience. For our purposes,
it is not necessary to work with a more precise
measurement of time, but it is possible.
Selecting the Check_in_Time or Seconds columns
will produce identical charts. The format in
the Check_in_Time column is only for improving
readability. The computer stores it as seconds.
Therefore, it is plotted with the seconds’
values. Notice that the data tips, however,
can be controlled by changing the print format
from the workspace so they display in a familiar
format. (To view data tips, mouse over data
points, summaries, or other hotlinks in the
image.)
If dates and times are entered as character
or string values, Statit will not recognize
them as valid variables for numerical calculations.
They can certainly be used along the X-axis,
however. When selecting a data variable for
the Y-axis, Statit will not include any text
variables in the list of variables from which
to choose. If this happens, it provides a clue
that the data values are not of the proper type.
Verify that the data is what you expect by right-clicking
on the column header, choose print format from
the menu, and check the Output Print Format
Category Type. If the type that is highlighted
by default is not what you expect, then the
data must be converted. It is not sufficient
to merely change the print format from this
menu from text to date or text to numeric.
Another strategy for working with time is to
use the difference of 2 times as the data variable.
Frequently, this is a natural progression as
in the examples of wait time and turn-around
time. The resultant values for wait time or
turn-around time can be expressed in whatever
time units make sense, such as seconds, minutes,
days etc. As long as the beginning and ending
variables are valid dates, dates and times,
or just times, the difference can be calculated
within Statit. Once the time difference has
been calculated and used in a chart, it is important
to include the date/time units in the chart.
This can easily be done by specifying the units
in the appropriate axis label or putting it
in the chart title or tag.
The following data illustrates how time differences
can be charted. The first column is the date
and time stamp of the receipt of the test by
the lab. The second column is the date and time
stamp at the completion of the analysis. These
date values were read into the workspace as
date variables. The compute/transform feature
under the Edit menu, Edit>Compute>Compute/Transform,
was invoked to subtract Received date values
from the Completed date values. The results
were put into a new variable, TotalTime.
| Received |
Completed |
TotalTime |
| 01-May-2005 |
06:01:00 01-May-2005 11:10:00 |
0.214583 |
| 01-May-2005 |
06:10:00 01-May-2005 10:50:00 |
0.194444 |
| 01-May-2005 |
07:14:00 01-May-2005 12:15:00 |
0.209028 |
| 01-May-2005 |
08:30:00 01-May-2005 13:42:00 |
0.216667 |
| 01-May-2005 |
10:13:00 03-May-2005 01:15:00 |
1.626389 |
| 01-May-2005 |
11:11:00 02-May-2005 08:00:00 |
0.867361 |
| 01-May-2005 |
12:14:00 01-May-2005 16:50:00 |
0.191667 |
| 01-May-2005 |
13:30:00 01-May-2005 20:20:00 |
0.284722 |
| 01-May-2005 |
14:40:00 02-May-2005 07:20:00 |
0.694444 |
| 01-May-2005 |
15:10:00 03-May-2005 10:30:00 |
1.805556 |
| 01-May-2005 |
16:35:00 02-May-2005 07:15:00 |
0.611111 |
| 01-May-2005 |
16:51:00 04-May-2005 09:30:00 |
2.69375 |
| 02-May-2005 |
07:05:00 02-May-2005 11:45:00 |
0.194444 |
| 02-May-2005 |
07:25:00 02-May-2005 13:12:00 |
0.240972 |
| 02-May-2005 |
07:51:00 03-May-2005 08:00:00 |
1.00625 |
| 02-May-2005 |
08:01:00 02-May-2005 21:15:00 |
0.551389 |
| 02-May-2005 |
08:25:00 02-May-2005 14:03:00 |
0.234722 |
| 02-May-2005 |
08:45:00 02-May-2005 12:01:00 |
0.136111 |
| 02-May-2005 |
09:20:00 02-May-2005 16:05:00 |
0.28125 |
| 02-May-2005 |
10:10:00 03-May-2005 16:30:00 |
1.263889 |
| 02-May-2005 |
11:01:00 04-May-2005 07:45:00 |
1.863889 |
| 02-May-2005 |
11:37:00 03-May-2005 09:00:00 |
0.890972 |
| 02-May-2005 |
13:27:00 02-May-2005 22:20:00 |
0.370139 |
| 02-May-2005 |
14:24:00 03-May-2005 10:17:00 |
0.828472 |
| 02-May-2005 |
16:10:00 03-May-2005 08:03:00 |
0.661806 |
An i-chart of TotalTime data produces the
following chart:
To view data tips, mouse over data points, summaries,
or other hotlinks in the image.
Hopefully, these examples will help provide
guidance when working with dates and times.
Since clock times, as we commonly use them,
are not decimal numbers, it is important to
allow the software to convert them to a form
that can be accurately used for calculating
the values necessary for reliable charts. Even
though the results of date/time conversions
may not be the same as the conventional formats
to which we are accustomed, be aware of the
importance of using them in this fashion.
