Data-Driven Solutions to Common Preanalytical Issues Part 3: Leaking Urine Containers [Hot Topic]

Darci Block, Ph.D.

In Part 3 of the series on data-driven solutions to common preanalytical issues, Dr. Block describes the process used at Mayo Clinic to reduce the number of random urine containers that leak during transport. The number of leaking containers was reduced by gathering data in a systematic way, involving all stakeholders, and using open communication.

Presenter and Credentials:
Darci Block, Ph.D., Director Laboratory Services and Co-Director Central Clinical Laboratory in the Department of Laboratory Medicine and Pathology at Mayo Clinic in Rochester, Minnesota


Our speaker for this program is Darci R. Block, PhD, Director Laboratory Services and Co-Director Central Clinical Laboratory in the Department of Laboratory Medicine and Pathology at Mayo Clinic in Rochester, MinnesotaWelcome to Mayo Medical Laboratories Hot Topics. These presentations provide short discussion of current topics and may be helpful to you in your practice. Today our topic is data-driven solutions to improving the error rate for a specimen container used for urinalysis.

Dr. Block, thank you for presenting today.

I have no disclosures

During the last part of this series we will examine a second case example where data was gathered and used to make a process improvement that many institutions face at some point or another, involving urine containers that leak during transit.

This diagram is meant to show all the possible sources of data that may have value for various process improvement projects. Not every piece of data will be needed for every project.

The laboratory noticed an increase in specimens arriving to the lab leaking with insufficient volume to perform testing.  At the same time, the maintenance department had contacted the lab supervisor sharing concern that the increased number of incidents where the pneumatic tube system needed to be taken down for decontamination would adversely affect its lifespan before major maintenance was needed.

Using the 5-why method for root cause analysis, I’ve started with the basic question:  Why do urine containers leak?  The answer is not all that surprising:   The lid does not seal.  Next question:   Why doesn’t the lid seal?  The answer gets a little less defined.  After speaking with primary senders of the container it appears that it is  difficult to tighten properly, and up until recently they were all sent with a piece of parafilm around the lid.  The parafilm was removed from service for most areas and was no longer available for use, and the leaking took a sudden increase.   Our hypothesis was that the parafilm was masking a long-standing problem with the urine containers that are prone to leaking and difficult to teach the diverse number of users to look for problems with the lid sealing properly.  Our proposed solution was to find a container that did not leak.

In order to do so, we needed to gather some data.  We needed to understand the current error rate, identify the stakeholders, define the requirements of a new container, and consider the costs of any proposed containers.

This diagram helps identify stakeholders and understand the process better.  We receive urine specimens from outpatient clinics where containers are handed out to patients and returned to drop stations.  Couriers transport them on foot to the testing laboratory for processing and testing.  Urine specimens are also received from the hospital where they are collected by nurses or in some cases a specialized urinary catheterization team.  They send this container via pneumatic tube to the testing lab for processing and testing.  Whatever the solution was, it needed to be user friendly for patients and a diverse group of clinical staff.  We also needed to keep in mind the validation requirements that the testing laboratories may have for a different container.

The error rate was estimated during a one-week period where a clip board was used in the processing area to document any containers that arrived with evidence of leaking.  The event reporting system was only used when the specimen had to be recollected for insufficient volume due to leaking. The process flow helped to identify the stakeholders which included patients, nursing, cath. team, testing labs, and maintenance. The requirements we looked for in a solution included having an adequate volume for all testing, easy to use, nonleaking, and considered acceptable by the testing laboratories.  We also contacted the purchasing department who identified those vendors we had existing relationships with and the selection of container that were available to consider switching to.

Early on we realized there were 2 possible directions to take where a primary container can be used and sent directly to the lab or a secondary evacuated container can be filled at the point of collection and sent for testing.  This balance summarizes some of the advantages we considered with each system.  However, in our assessment, we ended up deciding to stay with a primary container.  It was the easiest to switch to and could continue to be used across both hospitals and clinics with minimal additional education or training.  The challenge was to find one that did not leak.

Next we collected data in a transport study.  We attempted to mimic the existing process as much as possible.  We asked multiple different people to help fill the container with water and close it to be packaged and sent via one of a number of different pneumatic tube or drop-off stations.  We asked the volunteers what they thought of the container as we needed to assess whether additional or specialized instructions would be needed when containers were handed out.  Upon receipt in the lab, all evidence of leaking was documented.

Initially, we piloted a wide variety of containers including the current one causing problems.  You will note that even with the small number tested, some could easily be excluded from further consideration and 3 were subjected to further studies.

In further transport studies it is worth noting that the current container was operating at 3 sigma, whereas during our baseline assessment period had an error rate closer to 4 sigma.  This may be an artifact of under reporting during the assessment period and potentially the short duration with which the data was collected.  The next important lesson is conducting a study with a sufficiently large “N” so that there are enough opportunities for errors to occur.  We were reasonably happy with the improved performance of the last container with a 1% error rate having the potential to be even less if users could be taught to look for and correct cross-threaded lids prior to sending, though it didn’t happen frequently enough to preclude its approval.

One last consideration when it comes to cost analysis is to recognize that even a few pennies in savings really adds up when you consider the volume used.  For example, if you use a million containers per year, you can save $10,000 annually by reducing the cost per container by a single penny.  It really adds up!

In the post-assessment period, the number of events where specimen was received with insufficient volume for testing was 0.11%, with the cause being lid ajar over a 2-week period.   When we recollected the “evidence of leaking” data in the processing area, we noted it was very similar to the rate we measured during the study period.  While we may not have entirely eliminated leaking, there were at least no surprises after implementation based on the data collected during the assessment period.

This process improvement was challenging due to the diversity of users for this particular container.  We did our best to include them in the assessment and review of data as much as possible.   We had to consider users from the beginning of the process through to the end which is sometimes difficult to see unless you map out the entire process and pull those stakeholders in as needed.   I don’t know anyone who isn’t cost conscious these days so it’s important to remember that in some circumstances pennies may add up.  Communication is key to any successful process improvement implementation!

Ultimately, the leaking error rate improved by 50% with many still being able to be tested, noting lid ajar as the primary cause for leaking.   Fortunately, the leaking that did occur did not leave the biohazard packaging in most cases.   And the container has been standardized across all areas, simplifying purchasing and cost savings.

I think this implementation and process improvement worked in a predictable fashion because the transport studies that were done to choose the best container were statistically powered and well executed.  Additionally, the stakeholders were engaged in the process and notified as decisions were made.

In conclusion, it is recommended that meaningful data from multiple sources is gathered to address the issue at hand.  It is helpful to consult what has already been published on this topic.  It doesn’t make sense to reinvent the wheel when we can learn from our peers.  Guidelines are nice; CLSI are my go-to for most things because they provide tried and true recommendations that are very pertinent to laboratory medicine.  Baseline error rates are so important because it may actually reveal the complaint you are trying to fix isn’t really as big of problem after all.  Finally, flow diagrams are so helpful for seeing a bigger picture and figuring out whom all is impacted by what may seem like a small change that may have large downstream consequences. This provides an example of engaging the stakeholders in a process change and using data to make decisions that probably would not have been made de novo.

There is a long list of people to thank that contributed to this work.

If this topic was of interest to you please join us for the Phlebotomy Conference here in Rochester, Minnesota, April 23rd to the 24th, 2015.


MML Education

This post was developed by our Education and Technical Publications Team.