Quality

Quality, in pathology, has got a lot of attention lately because there have been high profile screw-ups that lead to significant harm.

General

The keys to quality are understanding the:

1. Needs of the stakeholders (surgeons, oncologists, patients, other pathologists, the public at large).
2. Processes.
3. Developing measures of quality.
4. Tracking the measures of quality & assessing their validity.
5. Understanding the causes of failures/adverse events in the context of the processes.
6. Continually doing all of the above with the aim of improving outcomes.

Analysis

Overview

Quality issues can be examined in a number of different ways.

Finding a problem:

• Root cause analysis.

Anticipating problems:

• Failure mode and effects analysis (FMEA).

General error analysis

Pathology errors happen any time from when the lab gets the specimen until after the report is issued.

When errors happen:

• Work-up the problem.
  • Where did the error occur? Pathologist error?
• Talk to the clinician.
  • If it is a critical diagnosis contact the most-responsible physician immediately... if they are unreachable call the physician on-call for the most-responsible physician... if the patient is out-of-town you may have to coordinate with the local emergency department.
• Talk to the chief of pathology.
• Incident report.
• Reconstruct error.
  • Was it a specimen mix-up?
    • Is there another error?
• Amend the report(s).
• Remedy the source of error.

The classic structural break down

A classic structal break down for error analysis is:

Errors in pathology

Pre-analytical errors

Analytical errors

Post-analytical errors

Note:

• This break down is arbitrary and in of itself most useful for answering exam questions.
• In a practical context, it is a frame work for classifying errors. It is not useful for understanding the source of an error or addressing it.

Pre-analytic errors

• Container mix-up - pre-lab & in-lab.
• Block mix-up.
• Slide mix-up - labels wrong.
• Poor quality slides (fixation, processing, staining).
• Lost specimen - can be potentially anywhere in the process.

Analytic errors

• Interpretation wrong.
  • Factors:
    • Difficult case.
    • Technical factors (quality of slides).
    • Lack of clinical history.

Post-analytic errors

• Wrong case signed-out.
• Filing problem/lost report.
• Interpretation of report problem (poorly written report, misinterpretation).

Sources of error

• "Human error".
  • Training.
  • Work flow.
• Process gaps.
  • Process control.
  • Lack of redundancy.

Types of errors

Can be subdivided into the following groups:

• False-negative - missed diagnosis.
• False-positive - diagnosis made that on review considered not to be present.
• Threshold - difference of opinion regarding a diagnostic threshold.
• Type and grade.
• Missed margin.
• Other.

Grading of errors

May be subdivided by three groups:

• Grade 1: no consequence.
• Grade 2: possible consequence.
• Grade 3: definitely a consequency.

Error reduction

Various strategies can be employed:

• Training of staff - on error handling.
• Computer order entry.
  • Avoid duplication fatigue.
  • Quick correlation with several identifying features.
    • Full name, sex, date of birth -- these all appear when one opens a case.
• Barcode use.
  • Avoid transcription errors.
• Clinical information entry required.
  • Allow correlation with test.
    • The interpretation may differ if the history says "screening coloscopy" versus "large cecal mass, anemia and weight loss" versus "breast cancer".

Other strategies:

• Statistical process control.

Biopsy size

Very small tissue fragments are associated with a decreased diagnostic yield and an increased diagnostic uncertainty.

Measures of quality

Any number of parameters can be used to measure quality. The when, where and how-often something is measured depends on the value-added.

General measures of quality

There are really only two:

1. Timeliness, i.e. turn-around time (TAT).
2. Error rate.

Note:

• 1 and 2 can be examined/quantified in any number of ways.
• Error, in the context of a measurement, has to be defined.

Smaller categories

Smaller categories - errors:

• Analytic: specimen identification & transport.
• Preanalytic/analytic: tissue processing, e.g. fixation, blocking, embedding, sectioning, staining.
• Analytic: interpretation.
• Postanalytic: reporting/report integrity.

Individual measures

Specific measures:

• Preanalytic:
  • Identification - numbers match requisition.
  • Appropriate container.
• Analytic:
  • Mislabeling.
  • Interpretation errors - based on:
    • Internal review.
      • Cytology-histology correlation.
      • Biopsy-resection correlation.
      • Frozen section-permanent section correlation.
      • Internal comparisons, e.g. ASCUS/LSIL between pathologists.
    • External review.
      • External standards/expected rate.
  • Amended reports - captures several of the above.
• Postanalytic:
  • Completeness of report.
  • Critical diagnosis timely?
  • Report delivered to appropriate person?

Immunohistochemistry

Main article: Immunohistochemistry

Classification of IHC tests

IHC tests are classified in a paper by Torlakovic et al.:

• Class I:
  • Results used by pathologists.
  • Adjunct to histomorphology.
  • Examples: CD45, S-100.
• Class II:
  • Used by clinicans for treatment decisions.
  • Considered independent of the other information in the pathology report; thus, cannot be derived from other information in the report.
  • Examples: ER, PR, HER2, Ki-67, CD117, CD20.

The implication of irregularies in the different classes are different. Problems in Class II tests are potentially more severe, as there is no internal control.

Work-up of suspected IHC problems

• Review controls (internal and external).
  • Isolated to case vs. larger problem?
    • Discuss with lab/make other pathologists aware of the issue.
• Repeat test - to identify the cause.

IHC process:

1. Ischemia time - warm ischemia, preparation of specimen.
2. Fixation - under, over, defective fixative, not enough fixative.
3. Processing prior to antibody binding, usu. heating (antigen retrieval).
4. Antibody-antigen binding.
5. Reporter molecule binding.
6. Counterstaining.
7. Interpretation problem.
   • Known/expected epitope cross-reactions, e.g. CMV & HSV.
   • Unknown/unexpected epitope cross-reactions.

Notes:

• Problems can arise at any step.

Other

Failure-potential analysis

Adapted from Ullman:

1. Identify potential individual failures.
2. Identify the consequences of those failures.
3. Identify how the individual failures can arise.
4. Identify the corrective action.

See also

• Critical values.
• CAP checklists.