Laboratory Turnaround Time & Control Chart to Evaluate Laboratory Processes

D

D.Salman

Dear Experts,
I have received a set of data (time in minutes format) of 9 days, for example, the first day I had 111 points each point reflect the difference between the time the doctor order laboratory test and the time the lab technician arise the result from system.

Date Number of data per day
05/04/07 111
06/04/07 96
13/04/07 65
14/04/07 63
15/04/07 62
16/04/07 24
17/04/07 35
18/04/07 39
19/04/07 38

My question is, can we use control chart to evaluate the process, because the laboratory test should be done within a specific timeframe?
Any feedback will be highly appreciated, and Many thanks.
 
D

Darius

Re: Laboratory turnaround time & control chart

My question is, can we use control chart to evaluate the process, because the laboratory test should be done within a specific timeframe?

The answer is yes, but....

Theorically you are right, but your data looks a little bit strange, it looks like the difference of the time the doctor order laboratory test and the time the lab technician arise the result from system is time dependient (or day to say so), the time has some efect that in the first days looks like they do the complicated job, and then after finished with the complicated ones, they do the normal job ( the last part of the chart).

Difference in time = -5.46667 * (time in days since the beginning) + 106.6

rsq = 0.882125

Weir, IMHO, I would like to know why such behabiur, remember that the control charts work for the same process condition, you can't mix different things in one chart.

Of course you can use the slope to estimate values and take off the trend, so it could be more meaninfull the control chart (but the differences look to much autocorrelated as intervals for 4 points cicles).
 
Last edited by a moderator:
V

Vassilis

Re: Laboratory turnaround time & control chart

Mr Salman,

For the data you have provided us the only control chart that i think is appropriate is an I-MR ( individual - Moving Range Chart).

As you can see from the attached file your process seems to be out of statistical control (the red points is out of UCL and LCL). You have to validate your process first using more days (at least 20 days), since the given data is not efficient ( just 9 days), and then if you want to use just one sample per day then use the I-MR chart.
If not, then you have to use larger sample size per day ( >2) to use an X-R chart for the time difference which is better i think.:D
 

Attachments

  • I-MR.doc
    55.5 KB · Views: 212
D

D.Salman

Re: Laboratory turnaround time & control chart

Mr Salman,

For the data you have provided us the only control chart that i think is appropriate is an I-MR ( individual - Moving Range Chart).

As you can see from the attached file your process seems to be out of statistical control (the red points is out of UCL and LCL). You have to validate your process first using more days (at least 20 days), since the given data is not efficient ( just 9 days), and then if you want to use just one sample per day then use the I-MR chart.
If not, then you have to use larger sample size per day ( >2) to use an X-R chart for the time difference which is better i think.:D

Dear Vassilis,
Thanks for the information.
The data is not only one sample per days, the second column in my original message shows the number of samples in the 9 days , for example in the first day we 111 points , and in the last day ( day 9) we have 38 points.
Thanks.
 
V

Vassilis

Re: Laboratory turnaround time & control chart

The data is not only one sample per days, the second column in my original message shows the number of samples in the 9 days , for example in the first day we 111 points , and in the last day ( day 9) we have 38 points.
Thanks.

Mr Salman , do you mean that the first day you had 111 defective occasions(delayed orders),the 9th day 38 defective occasions ,or do you mean that the first day total time (delayed) in minutes where 111 and the 9th 38 minutes?

My questions is crucial in which is the appropriate Control Chart that you have to use.
In the first situation (defective occasions /day ) the appropriate Control Chart would be a C-chart. (see attached file).
In the second situation ( total minutes /day) it would be I-MR chart as i have pointed out in my previous post.

I hope that i have helped you...
 

Attachments

  • C Chart.doc
    38 KB · Views: 275
Top Bottom