Defining Target Populations for Clinical Pathways


Seattle Cancer Care Alliance, Seattle, WA


The author reports no relevant financial relationships. 

A critical consideration in pathway development is identification of the target population for the pathway. A pathway is often developed for a specific population of patients with a given clinical diagnosis. However, variations in electronic medical record documentation create a challenge to the process of determining which patients should be eligible for a particular pathway. Greater refinement of clinical criteria for determining the target population for a pathway can help to overcome this challenge.
Related Content
Clinical pathways and personalized medicine
Addressing cost in clinical pathways

Clinical pathways are established for specific diseases to standardize care and reduce clinical variability. Within pathways for oncology, stages of the disease as well as phases of care—such as initial evaluation and surveillance—are identified to address the unique considerations of the disease state or phase of treatment. As pathways continue to evolve and become more integrated into different facets of the patient experience, they may also include other dimensions of care, such as supportive care, palliative care, and survivorship. Together, these dimensions can help produce comprehensive population health strategies.

Defining a pathway’s target population is an evolving process. One way of defining target populations for pathways is by their clinical diagnosis. For example, patients with non-small cell lung cancer (NSCLC) would be a group eligible for an NSCLC Clinical Pathway. Using electronic medical records (EMRs), clinicians are able to document the diagnosis of NSCLC using this acronym or by the specific histologic subtype a patient has, such as adenocarcinoma, large cell, or squamous cell. However, variation in EMR documentation—whether in a clinic note or a pathology report—represents a challenge to the process of designating a particular patient as being included in the target population or being eligible for a pathway.

At Seattle Cancer Care Alliance (Seattle, WA), we are experimenting with an approach that supports our goal of developing a clinical intelligence reporting platform. Our development approach to pathways focuses on defining the standard of care and establishing optimizations for superior outcomes, patient experience, and the mitigation of variation. To achieve this goal, we have worked with our pathways development team to develop a process for identifying target populations for clinical pathways, which we are currently evaluating.

During the development of a pathway, along with our providers, pathologist, pathway associates, data abstractors, and allied health care team, we define three data elements that help to determine a target population: (1) primary site; (2) histology code with description; and (3) clinical diagnosis. A clinical diagnosis is one that is abstracted from a primary oncologist’s clinic notes, in combination with a pathology report, and is used to confirm the histology and a patient’s eligibility for a pathway.

The three elements address gaps that can appear in our clinical data systems and clinical notes. For example, we found that histology site and histology code (we use the International Classification of Diseases for Oncology [ICD-O] scale) are not sufficient to clearly define patient populations in all cases. This means that pathology reports may not stage a patient’s disease, which makes our analysis unable to place a patient on a pathway without considering a diagnosis with stage from a clinical note.

Consider a pathway that is developed for NSCLC. The histology site and code will indicate that the patient has NSCLC but not which stage. Some pathways, depending on their intent, scope, or development status, may not address all stages or metastasis status. We discovered that ICD-O codes do not include stage of disease in their definition. Further, the codes, in our implementation, do not indicate in the case of solid tumors whether the tumor is primary or a metastatic secondary mass. ICD-O–3 codes1 contain morphology modifiers /6 and /9, which do indicate metastatic state, but our implementation of ICD-O-3 follows Facility Oncology Registry Data Standards page 124 guidance2 not to use code modifiers beyond /3. In order to designate a patient as eligible for an NSCLC pathway and a specific stage—stage IV for example—the ICD-O codes for primary histology site and tissue type may not be sufficient due to lack of stage information and metastatic status in some pathology reports.

There are likely opportunities to single source the information needed to determine patient eligibility to a pathway. In fact, some institutions are implementing templated notes or discrete information entry via an EMR that captures necessary histology, stage, and clinical diagnosis with stage. Alternatively, similar approaches to discrete data entry may require a provider to designate which pathway a patient should be considered for and, ultimately, whether they meet all inclusion criteria.

Our approach works within the constraints of how our data is captured and abstracted currently. However, our process only determines patient target population. We have another process that considers our exclusion and inclusion criteria for a pathway and whether eligible patients are ultimately placed on a pathway. That process is the topic for a future column.

Institutions should share information about how they establish pathway target populations and how these approaches address the nature of clinical data availability. 



1.    International Agency for Research on Cancer, World Health Organization. International Classification of Diseases for Oncology (ICD-O-3). Accessed October 24, 2016.

2.    Commission on Cancer, American College of Surgeons. Facility Oncology Registry Data Standards, Revised for 2016. Accessed October 24, 2016.