i am doing a GB project on Turn around time of tests which are received in our laboratory 24/7.The doctors are available from 9a.m to5.30 p.m.The TAT is applicable only when doctors are available & defined as 4 hrs to the patient & my process width is 3.30hrs.The problem is some tests are over in 2hrs & some take 3.30.this is a data where average & S.D varies.

What are the statiscal tools applicable for this type of data so that I can complete my project.Kindly help

Hi Radhika,

Let me see if I understand this correctly...

If a doctor orders a test at 5:20 pm and the test results are ready at 9:00 am in the morning, is the TAT 10 minutes only as the doctor was available until 5:30 pm and your lab works 24/7?

 

SJ 

No it is that the process continues even in the absence of the doctor by the technical staff but the result validation is done by the doctor next day & only then the is relaeased.If there is a problem it is rechecked or reprocessed based on the problem & issued.The patients are also informed that results come only next day if the sample is given after 4.00p.m.

I hope this clarifies your doubt

Dr.radhika Chowdary

Your output data TAT is continuous. Your upper specification limit seems to be 4 hours (though based on your last comment, it seems to be 1 1/2 hours since you mention that test results are due back the same day if the sample is given at 4 pm and the doctor leaves at 5:30 pm).

If your standard deviation is high (lot of variation) in your TAT, then you probably have to collect a lot of data to analyze it (check your required sample size based on sample size calculations).

Based on your data, determine the baseline performance (number of defects, DPMO, Cp/Cpk, and sigma level). If these are lower than where you like it to be and there are significant benefits of improving this performance, then you could potentially work on improving this performance.'You would need to create a project charter and get it approved from the right project sponsor.

The type of tools you would use to analyze this problem depends on what the data type of potential root causes are. I would recommend that you do a fishbone diagram to identify all the possible causes, maybe perform a C&E matrix to narrow down the causes and then depending on the type of X data, use the appropriate type of statistical analysis tool to validate the root cause.

Hope this helps.

SJ.

 

Dear SJ / VK sir,

I have completed green belt training in Nov'2009 at Mumbai conducted by VK and Mr. Shantanu Kumar.

I have worked on a problem and created a project Charter. Request you to kindly review and give your comments so that I can understand if I am moving into right direction.

Manufacturing Cost figures will be available in few days. Please also advise me about the next few steps.

Thanks & regards,

Tarun Shrivastva

 

Hi Radhika,

Although most of the steps are cleared by SJ, I would suggest to you to go for Time and motion study also to find out the time wasted in Non value added things also.

Regards

Navin Rohilla

BB