A 250-bed hospital had implemented total quality management (TQM) to improve customer delight in its diagnostic clinics. After that success, the management team of the hospital was ready to proceed with further process improvement. After completing a brainstorming and prioritization exercise, the team determined to next address the problem: delay in receiving test reports for patients in the intensive care unit (ICU).

A cross-functional group was chosen as the process improvement team; they went through a two-day “quality mindset” training and started meeting every two weeks to address the delays in turnaround time (TAT).

This case study is available for purchase on the iSixSigma Marketplace.

Process improvement efforts at the hospital followed the seven steps of problem solving:

  1. Define the problem and measure it
  2. Research root causes (using the 5 Whys)
  3. Discuss and agree upon countermeasures
  4. Test countermeasures
  5. Check results
  6. Standardize results
  7. Compile quality improvement (QI) story

In addition to these seven steps, the team also incorporated just-in-time (JIT) manufacturing principles by cutting non-value added stages and activities in the process, speeding up and reducing work through automation, and improving workflow.

Step 1: Define the problem and measure it

Problem = (desired state – current state) of “required” TAT

Any TAT exercise starts with identifying the start and end of the process. Each team member was, therefore, asked to state the start and end of their step(s) in the test report delivery process. Their responses are shown in Table 1.

Table 1: Start and End Steps in Report Delivery Process (Before)
Team Member Start of Process End of Process
ICU doctor Makes test request Receives report
ICU nurse Takes blood sample Receives report
Laboratory nurse Receives blood sample Inputs result in computer
Laboratory doctor Receives blood sample Signs report

The ICU doctor requested a test at the start of the process and at the end of the process the same doctor received the results of the test and signed off on the results. The ICU doctor (as the internal customer) set the desired TAT as 20 minutes.

When questioned further, the ICU doctors clarified that their desired TAT time was a maximum of 20 minutes. Statistically speaking, this meant that for 3σ performance, TAT should take fewer than 20 minutes 99.7 percent of the time. Unfortunately, from the start this was known to be impossible as different lab tests take different amounts of time.

Understanding the constraint imposed by the different lab tests, the team then asked the ICU doctors to provide their desired TAT for the most common lab tests. Actual times (times taken for normal operations of tests) were not measured at this stage in the project. It was assumed that the doctors knew the actual times and that their desired times for tests would reflect that knowledge. The desired times are listed in Table 2.

Table 2: Desired TAT of Laboratory Tests

Test Type

Desired Time of Test (minutes)

LFT

45

PT/PTT

20

HBsAg/HIV

120

Blood group typing

15

Serum electrolytes

20

FDP

60

Kt/V

10

Ab4

5

Creatinine

45

The final metric emerged at this point. A delay was defined as the actual TAT minus the desired TAT for each test. The average + 3σ of this delay needed to be less than 0 for 99.7 percent of the tests to be within desired levels.

The delay of 10 random tests per day was recorded. In the first week, the results were:

  • Average delay = 77 minutes
  • Standard deviation (σ) = 58 minutes
  • Average + 3σ = 251 minutes

The goal was now defined: reduce 251 minutes to zero. (Although the overall goal was to reduce to zero minutes, a drop of at least 50 percent in the initial go-round of this improvement project would be considered a success). An X-bar control chart for the TAT delay recorded in the first week is shown in Figure 1.

Figure 1: X-bar Control Chart for TAT Delay (Before)
Figure 1: X-bar Control Chart for TAT Delay (Before)

Step 2: Root cause analysis: 5 Whys

When using root cause analysis for TAT problems, the root causes are typically found by mapping the process and identifying all non-value added activities, stages and work on the one hand, and by batching and correcting imbalanced workloads that are impeding workflow on the other.

The process steps are shown in Table 3 below. Cells with a blue background indicate steps that were automated moving forward. Cells with a yellow background indicate non-value-added (NVA) steps that were either eliminated, modified to save time or reallocated.

Table 3: Sequence Steps for Making and Receiving Laboratory Results – Before

Step Number

Current Step in Process

Starts With or Is Handled By

Ends With (as needed)

Wait Time Involved

1

Requests test

Doctor

ICU nurse

X

2

Takes sample

ICU nurse

3

Inputs samples into computer, which generates barcode tracker

ICU nurse

4

Creates hard copy request form

ICU nurse

5

Informs good housekeeping that a “batch” of tests are ready

ICU nurse

Good housekeeping (GHK)

X

6

Takes samples to sample receiving station, which receives samples from all locations in the hospital and redirects them to appropriate testing department

GHK

Sample receiving station (SRS)

7

Opens computer data entry window

SRS

X

8

Enters identifier in computer system to match patient to barcode

SRS

9

Checks tests, tubes, etc.

SRS

10

Attaches barcodes to samples

SRS

11

Puts orange stickers on tests that must be completed urgently

SRS

12

Sorts trays by department and puts into separate trays

SRS

X

13

Prepares batches of tests

SRS

X

14

Calls GHK to notify them the tests are ready to go to the laboratory

SRS

GHK

X

15

Takes tests to laboratory for processing

GHK

Lab tech

16

Enters identifier data into notebook

Lab tech

X

17

Loads samples into centrifuge

Lab tech

18

Unloads centrifuge

Lab tech

19

Puts samples into test machine trays

Lab tech

X

20

Loads trays into test machines

Lab tech

X

21

Administers tests

Lab tech

22

Informs ICU nurse of test reports verbally

Lab tech

ICU nurse

23

Enters test results into computer system

Lab tech

X

24

Prints reports of test results

Lab tech

X

25

Provides reports to head of department for approval

Lab tech

Head of department (HOD)

X

26

Signs results in batches

HOD

GHK

X

27

Notifies GHS that results are ready

HOD

X

28

Sorts results

GHK

X

29

Delivers reports

GHK

ICU nurse

X

30

Informs ICU doctors of test results

ICU nurse

X

Step 3: Generate countermeasure ideas

A laboratory test sample was followed to map the entire process as shown in Table 3 and to reveal possible areas for improvement.

Steps 5 through 9 and 12 through 15 were NVA activities and could be eliminated by having the test sample go directly from the ICU to the laboratory rather than going through an intermediary desk. Paper entries (steps 4, 16, 24 and 26) could be eliminated by using the existing computer system; this would also eliminate paper transfers and the waiting involved in locating a messenger, transporting and delivering the reports by hand. With automation, the results could be sent directly from the computer system to the doctors’ mobile telephones.

The improvement project was carried out in two phases, each two areas of the process:

  1. Process steps supporting, but occurring largely outside, the laboratory
  2. Process steps occurring in the laboratory

Phase One

The first phase of change proposed is shown alongside the existing process steps in Table 4. (If a proposed step is not noted, no change was made.) Cells with a green background are value-added activities.

Table 4: Sequence Steps for Making and Receiving Laboratory Results – After

Step Number

Current Process Step

Starts With or Is Handled By

Ends With (as needed)

Proposed Step

Starts With or Is Handled By

Ends With (as needed)

1

Makes test request

Doctor

Nurse

Makes test request

Doctor

Nurse

2

Takes sample

Nurse

Takes sample

Nurse

3

Inputs samples into computer, which generates barcode tracker

Nurse

Inputs samples into computer

Nurse

4

Creates hard copy request form

Nurse

Attaches barcodes to samples/attaches orange stickers to relevant samples

Nurse

5

Informs good housekeeping (GHK)(Batch)

Nurse

GHK

Puts samples into trays for laboratory

Nurse

6

GHK takes samples to sample receiving station (SRS) [receives samples from all locations in the hospital and redirects them to appropriate testing department]

GHK

SRS

Comes at fixed intervals to pick up trays and deliver them to laboratory

GHK

Lab tech

7

Opens computer data entry window

SRS

Scans barcode, which confirms receipt of sample automatically

Lab tech

8

Enters identifier in computer system

SRS

Loads samples into centrifuge

Lab tech

9

Checks tests, tubes, etc.

SRS

Unloads samples from centrifuge

Lab tech

10

Barcodes

SRS

Puts samples into test machine trays

Lab tech

11

Puts orange stickers on tests that must be completed urgently

SRS

Loads trays into test machines

Lab tech

12

Sorts trays by department and puts into separate trays

SRS

Administers tests, reports are automatically reported into computer system

Lab tech

13

Batches of tests are prepared

SRS

Checks and approves (as appropriate) reports

Lab tech

14

Calls GHK to notify them the tests are ready to go to the laboratory

SRS

GHK

Alerts doctors and nurse of test results readiness via information technology solutions (e.g., iPad application)

Auto

15

GHK takes tests to laboratory for processing

GHK

Lab tech

Approves hard copy results

HOD

16

Laboratory technician enters information in a notebook

Lab tech

Delivers at routine, defined intervals

GHK

ICU

17

Loads samples into centrifuge

Lab tech

18

Unloads centrifuge

Lab tech

 

19

Puts samples into test machine trays

Lab tech

20

Loads trays into test machines

Lab tech

21

Administers tests

Lab tech

22

Informs ICU nurse of test reports verbally

Lab tech

Nurse

23

Enters test results into computer system

Lab tech

24

Prints test results reports

Lab tech

25

Provides reports to head of department (HOD) for approval

Lab tech

26

HOD signs results in batches

HOD

GHK

27

HOD notifies GHS that results are ready

HOD

28

GHK sorts results

GHK

29

Delivers report

GHK

Nurse

30

Informs ICU doctors of test results

Nurse

Doctor

Step 4: Test the idea

The now 16-step process was then tested – the team was alerted that samples were to be run without delay at any step during this trial. (Steps 7, 12 and 13 would need information technology changes and would be managed with minimum delay for test purposes.) PT/PTT,  lipase and amylase are three tests that are expected to take a similar amount of time (TAT). For nine instances pre-project, PT/PTT had average TATs of 95 to 173 minutes (Table 5). The results of a few of those tests after implementing the new procedure are shown in Table 6.

Table 5: Turnaround Times for PT/PTT Samples – Before

Test

TAT (minutes)

PT/PTT

95

PT/PTT

95

PT/PTT

95

PT/PTT

95

PT/PTT

95

PT/PTT

173

PT/PTT

160

PT/PTT

140

 

Table 6: Turnaround Times for PT/PTT, Lipase and Amylase Samples – After

Test

TAT (minutes)

PT/PTT

54

Lipase

56

Amylase

60

Overall, the TAT delay was reduced by more than 50 percent. Two samples that took longer were incorrectly processed with the existing (more step-intensive) process. All personnel were briefed and a full-scale trial was initiated.

Step 5: Check the result

Following a brief period of getting used to the new system, a dramatic improvement took place as shown in the X-bar control chart in Figure 2.

Figure 2: X-bar Control Chart for TAT Delay (After)
Figure 2: X-bar Control Chart for TAT Delay (After)

The TAT delay had reduced by approximately 50 percent to 60 percent.

TAT

Before

After

Average

90

38

Average + 3σ

296

106

When the processes were further streamlined, the average + 3σ reduced by another 50 percent, from 106 minutes to 53 minutes.

Phase 2

The streamlining of the process steps that occurred outside of the laboratory proceeded, along with the necessary information technology changes and updated. Simultaneously, the project team began to address what was happening inside of the laboratory in order to reduce TAT to less than 56 minutes (which is shown in Table 6).

Sample arrival-pattern mapping indicated that the biochemistry and hematology sections were processing the maximum loads of tests. The biochemistry section was selected as the section for the project team to collaborate with as it worked on a new process.

Table 7, below, shows the TAT as recorded in various test batches.

Table 7: TAT Delays
Test Numbers
(each number is an average of ten tests
[i.e., 20-29 is ten tests])

20-29

30-39

40-49

50-59

60-69

70-79

80-89

Average

59

69

48

90

92

58

38

Std Dev (σ)

50

57

24

67

61

43

23

Avg + 3

209

239

120

296

276

187

106

Steps 2-4: From causes to countermeasures

The process was mapped to identify the reasons for significant TATs and to then develop countermeasures as shown in Table 8.

Table 8: Sequence Steps for Laboratory Tests

Step Number

Current Process Step

Step Starts With or Is Handled By

Batch (when appropriate)

Proposed Step

Step Is Handled By

1

Places samples in “in” tray

GHK

Places samples in “in” tray

Lab tech 1

2

Pickes up samples when time is available

Lab tech 1

X

Picks up samples every 30 minutes (up to 24 samples)

Lab tech 1

3

Loads samples into centrifuge

Lab tech 1

Scans barcode

Lab tech 1

4

Enters information in a notebook

Lab tech 1

Loads samples into centrifuge

Lab tech 1

5

Unloads centrifuge, load back in tray

Lab tech 1

X

Unloads centrifuge

Lab tech 1

6

Picks up samples when time is available

Lab tech 2

Loads trays into biochemical equipment

Lab tech 1

7

Enters information in a notebook

Lab tech 2

Sends results to computer system (automatically by machine)

Lab tech 1

8

Loads samples into biochemical machine trays

Lab tech 2

Approves report electronically

Lab tech 1

9

Loads into biochemical machine

Lab tech 2

HOD approves reports as they are ready

Lab tech 1

10

Cuts and clips printouts of results to initial test requests

Lab tech 2

X

11

Enters test results into computer system

Lab tech 3

X

12

Prints new results

Lab tech 3

X

13

Provides new results to HOD for approval

Lab tech 3

X

14

HOD approves results

HOD

Four major changes were proposed after mapping the process.

  1. Compared to the current system in which samples were handed from Technician 1 to Technician 2 to Technician 3, now each technician was tasked with starting their own batches in the centrifuge (24 samples) and carrying them to completion in flow, ensuring a minimum of wait time.
  2. Paper notebooks were used to track samples in the current system time. Barcode scanners were implemented to automate many steps.
  3. Several interfaces between the test machines and the computer system were not working, which meant that manual entries and batched report preparations were the only options. These technical issues were fixed to enable more automation and improve flow with batching.
  4. Electronic signatures were recommended to make hard copies unnecessary for work to progress. Although electronic signatures were introduced, some hard copies were still maintained due to statutory requirements. They could be signed later by the HOD.

Step 5: Check the result

The first trials – even without all of the IT changes in place – demonstrated that one technician could process 24 samples, from requesting a test to issuing the report, within 60 minutes. Two technicians could start a batch of samples every 30 minutes and together complete 48 samples per hour.

Hourly load arrivals were measured; peak ICU lab loads were only 38 samples per hour. The lab could, therefore, handle the ICU loads with a maximum one 60-minute cycle. If all of those ICU samples were in the first batch of 24, TAT would be further reduced. The proposed process would be time-bound – reports would be issued every 30 minutes for all ICU samples received within the previous hour. After the biochemistry department went through the improvement process, the hematology department was similarly improved.

As the automation was implemented and employees got accustomed to the new process, TAT improvement continued as shown in Figures 3 and 4 below.

Figure 3: X-bar Control Chart for TAT Delay (After)
Figure 3: X-bar Control Chart for TAT Delay (After)
Figure 4: TAT Delay (Average)
Figure 4: TAT Delay (Average)

Before

After

Average delay

77 minutes

No delay – improved to one minute early

Average + 3σ

300 minutes

< 40 minutes

The ICU doctors proclaimed, “There is no delay from the laboratory.” More than 99.7 percent of reports were completed within 40 minutes, meeting the needs of the doctors.

At the start of this project, the doctors set a desired time of 20 minutes per test. When asked for targets for individual tests, the numbers for desired TAT ranged from 5 minutes to 120 minutes. The average delay resulted in 50 percent completed on time and a maximum delay of 40 minutes per test. The doctors proclamation of “no delay” was their impression of a far faster system. Although some tests continued to take longer than 40 minutes, they no longer took hours.

Step 6: Standardize the results

Team members were trained to draw the control charts, regularly review performance and eliminate the causes of any deviation.

Step 7: Quality improvement story

The quality improvement story was prepared and presented to management.

The healthcare system can be improved enormously by using TQM – first in the supporting processes and then in the medical processes themselves.

About the Author