SECTION 4

EXPERT DECISION
RULES
LEARNING OBJECTIVES

When you complete this section, you will be able to:

• Define rules

• Understand the anatomy of rules

• Understand the main types of rules used in the Hematology laboratory

• Understand the key benefits of implementing middleware decision rules in the Hematology laboratory

LEARNING GUIDE: MIDDLEWARE USE & OPERATION 16

SECTION 4 – EXPERT DECISION RULES
The Expert Decision Rules function is one of the most popular capabilities and benefits of
middleware. They help automate the manual tasks and processes in the laboratory environment,
enabling greater efficiency, productivity and error reduction along the testing continuum.

Rules are logic statements that define a set of conditions. These logic statements use IF, THEN and
ELSE to define conditions. Rules take the form: IF some condition is true, THEN perform a certain
action, or ELSE perform a different action. The IF, THEN and ELSE statements make up a rule.

• The IF portion of the rule defines a condition or series of conditions to determine if the follow
up actions (in the THEN or ELSE portion) should be taken.
• The THEN portion defines an action or series of actions to be taken if the IF condition(s)
is true.
• The ELSE portion defines an action or series of actions to be taken if the IF condition(s)
is false.
Rules can be applied to different data elements and processes in the sample testing process, such as
orders, results, and specimen processing status (e.g., specimen received).

The expert rules found in middleware essentially emulate advice that the Technologist would
provide, enabling automation of manual tasks and processes associated with the laboratory
workflow. In the Hematology laboratory environment, these processes include managing CBC
and differential test results, from reviewing results to determine if they can be released to the
LIS or further action is needed, to reviewing results against established laboratory guidelines to
determine if it’s clinically necessary to perform a slide review or manual differential.

Perhaps the most common uses of rules in middleware is for autoverification and autovalidation.
Autoverification is the process where test results are automatically reviewed and released
to the LIS or held for manual review based on a predetermined set of rules established by the
laboratory. Autoverification eliminates the need for human intervention.11 As results are filed into
the middleware from the analyzer, they are automatically reviewed against predefined conditions
(rules), and either automatically released to the LIS or held for manual review by the Technologists
if there are exceptions.

17 LEARNING GUIDE: MIDDLEWARE USE & OPERATION

 Result is transmitted
from instrument

Compare with
previous; release Is this
if within YES a rerun?
acceptable limits

NO

Hold for Are results

manual review NO numerical or equal
(possible rerun) to “<#.##” or
“>#.##”?

YES

Hold for
manual review; Does result
add additional have instrument flags
YES
autoverification (or those deemed Round result to
actions here relevant?) the correct
based on flag decimal place

NO

YES
Does the Does the Is the Are previous
result exceed result exceed result Is this the first result and current
NO NO analysis on this NO
high AMR low AMR numerical? result within
(“>#.##”)? (“<#.##”)? patient? acceptable
limits?

YES YES YES

Parameter- Parameter-
dependent dependent
options (hold for Is this result Hold for
options (hold for within the
manual review, NO manual review
review, rerun, verification
rerun, dilute, or release)
or release) range?

YES

Does the
result compare
well with other NO Hold all related YES
related assays for
parameters? manual review

YES

Result autoverified

Figure 4.1 An example of a sequential set of autoverification rules for Hematology results release.11 Although multiple
results are transmitted from the instrument as part of the CBC results for each sample, middleware evaluates each result
independently as shown here. Advanced middleware decision rules can manage this evaluation and complex combination
of steps.

LEARNING GUIDE: MIDDLEWARE USE & OPERATION 18

Autoverification of test results in middleware are evaluated against several different parameters.

• Patient demographic information – factors in associated patient demographics (e.g., age and
gender) when analyzing results to determine whether to auto release results, hold for manual
review, rerun or conduct additional testing.
• Critical values – ensure test results are not automatically released when clinical validity may
be suspect.
• Delta checks – compares current results for a given test against previous results that exist for
the same patient to determine if, based on established laboratory standards, the difference is
acceptable for release to the LIS.
• Quality control – blocks the automatic release of results if quality control issues are
identified.
• Instrument flags/exceptions – prevent the automatic release of results when an associated
instrument flag or exception is detected.

Condition 1 Condition 2 Condition 3

IF NEUTROPHIL absolute value IF NEUTROPHIL absolute value
is less than user defined value OR is greater than user defined value & IF this is the first time seen
(example <1.0) (example >20.0)

Action

Hold sample results (for manual review)

Figure 4.2 Example of an autoverification rule. There are 3 conditions within this rule. Condition 1 and condition 2 are
evaluated first; these conditions look at the absolute value for the Neutrophil result and whether the result is less than
1 or greater than 20. If condition 1 or condition 2 is true, then condition 3 is evaluated to determine if this is the first
occurrence of the Neutrophil result for the patient. The rule will perform both action (hold sample results) if condition 1
OR condition 2 is TRUE, and condition 3 is TRUE.

Autovalidation is the process whereby sample results are validated automatically based on
a predetermined set of conditions. These conditions may be established by a combination of
laboratory standard operating procedures, instrument vendor recommendations and/or published
guidelines, such as the 41 Consensus Rules developed by the International Society for Laboratory
Hematology (ISLH).12 These rules automatically review and validate CBC and differential results to
take actions that might include:

• Ordering a rerun
• Reflexing to a slide review
• Adding comments to notify the laboratory and/or clinical staff

19 LEARNING GUIDE: MIDDLEWARE USE & OPERATION

 Although middleware platforms offer different criteria that trigger actions to review CBC and
differential results from automated Hematology analyzers, one of the most widely used standards
is the ISLH Consensus Guidelines. Recognizing that there was little uniformity among laboratories
on what prompted the review of CBC and differential results, Dr. Berend Houwen invited 20
Hematology experts to determine criteria and develop rules in 2002. These rules were tested on
over 13,000 blood samples in 15 laboratories. Following testing, the data was thoroughly analyzed,
leading to further refinement of the rules to a final set of 41 rules.13 The guidelines include rules
for CBC numeric results, differential numerical results and morphological flags. The 41 ISLH
Consensus Rules and modified versions are widely used in Hematology laboratories worldwide.

Middleware platforms with sophisticated rules-based solutions, offer the ISLH Consensus Rules as
a predefined rules package for Hematology laboratories. This rules package is typically preloaded
with the 41 Consensus Rules and are often configurable so that laboratories may customize
parameters and values to meet their operating procedures and patient population needs. Some of
the criteria evaluated in the rules set include:

• Analytical measurement ranges

• First time events
• Patient age
• Sample age
• Various instrument and morphological flags
• Delta checks

Based on these criteria, the middleware can trigger a variety of actions:

• Slide review
• Rerun
• Hold results for manual review
• Block partial or full results (e.g., block percentage and/or absolute differential results)
• Add comments

Condition 1 Condition 2 Condition 3

IF WBC is less than user IF WBC is greater than user
defined value (example <4.0) OR defined value (example >30) & IF this is the first time seen

Actions

Hold sample results + Order slide review

Figure 4.3 Example of an autovalidation rule. There are 3 conditions within this rule. Condition 1 and condition 2 are
evaluated first; these conditions look at the value for the WBC result and whether the result is less than 4 or greater than
30. If condition 1 or condition 2 is true, then condition 3 is evaluated to determine if this is the first occurrence of the
WBC result for the patient. The rule will perform both actions (hold sample results and order slide review) if condition 1
OR condition 2 is TRUE, and condition 3 is TRUE

LEARNING GUIDE: MIDDLEWARE USE & OPERATION 20

Key Benefits of Implementing Middleware Rules
in the Hematology Laboratory
Implementing Expert Decision Rules provided by advanced middleware solutions can automate
complex, manual processes and improve the efficiency of Hematology laboratories. Some of the key
benefits to the CBC results review process include:

1. Automation of the results review process can significantly improve results turnaround
time.

2. Elimination of the review of routine, manual results can improve productivity by allowing
the laboratory staff to focus on actual sample exceptions.

3. Minimize tech-to-tech variation during results review through standardization.

Advanced rules solutions are typically customizable and can be adapted and revised as the
workflow of the laboratory evolves.

Now complete the section Quiz.

21 LEARNING GUIDE: MIDDLEWARE USE & OPERATION