Decoding the Phenotype of Sickle Cell Disease with OMOP

The phenotype of sickle cell disease is the clinical story that unfolds from a single, tiny error in a person's genetic code. It’s the full range of observable traits and health problems-from the misshapen red blood cells that give the disease its name to the cascade of severe, often life-threatening complications. This phenotype is the direct, physical manifestation of a specific defect in the hemoglobin gene.

Defining the Sickle Cell Disease Phenotype

It’s tempting to think of sickle cell disease (SCD) as a single condition, but that's not quite right. A better way to picture it is as a wide spectrum of health outcomes, all originating from a single point mutation. This change creates an abnormal type of hemoglobin called hemoglobin S (HbS). If you're new to this concept, our guide on what "phenotype" means in a data context is a great place to start.

When a person has HbS, their normally pliable, disc-shaped red blood cells can deform under stress, becoming rigid, sticky, and twisted into a crescent or "sickle" shape. This fundamental change in cell structure is the primary event that triggers the entire clinical phenotype of SCD.

Genotype vs. Phenotype in SCD

In my experience, one of the most critical distinctions to make is between genotype-the genetic instructions-and phenotype, which is how those instructions play out in the real world. A patient's specific genetic combination is the key to anticipating the severity of their disease.

This table gives a high-level overview of how different genotypes often translate into clinical outcomes.

Core Sickle Cell Genotypes and Phenotypic Outcomes

Genotype	Common Name	Typical Phenotype Severity
HbSS	Sickle Cell Anemia	Severe
HbSC	Hemoglobin SC Disease	Moderate
HbSβ⁰-thalassemia	Sickle Beta Zero Thalassemia	Severe
HbSβ⁺-thalassemia	Sickle Beta Plus Thalassemia	Mild to Moderate
HbAS	Sickle Cell Trait	Generally Asymptomatic

As you can see, the most common and severe form is homozygous sickle cell anemia (genotype HbSS), which occurs when a person inherits a sickle cell gene from both parents. However, there are many other possibilities that produce a wide range of clinical pictures.

Two important distinctions are:

• Sickle Cell Trait (HbAS): These individuals inherit one sickle cell gene and one normal one. They are usually asymptomatic carriers, but their red blood cells can still sickle under extreme physiological stress, like severe dehydration or low oxygen. This is not considered sickle cell disease.
• Other SCD Genotypes: Compound heterozygous forms, like HbSC disease or sickle beta-thalassemia, happen when someone inherits the sickle cell gene plus another abnormal hemoglobin gene. This results in different, and often less severe, phenotypes compared to HbSS.

This distinction is crucial for understanding the true scale of the disease. Globally, the number of people living with sickle cell disease jumped by 41.4% between 2000 and 2021, climbing from an estimated 5.46 million to 7.74 million. You can dig into the data and learn more about global sickle cell prevalence to see the growing impact.

The genotype provides the genetic blueprint for making abnormal hemoglobin. The phenotype is how that blueprint actually gets built in a person's body, ranging from no noticeable issues to a lifelong battle with debilitating health crises.

Practical Tips for Data Representation

For anyone working with clinical data, capturing this phenotypic variability is absolutely essential for meaningful analysis. When you first start building a cohort for SCD, resist the urge to lump all related diagnoses together.

Here’s my advice:

• Be Specific: Your first step should be to create separate concept sets for symptomatic disease (like HbSS and HbSC) versus the asymptomatic trait (HbAS). Mixing them will completely distort your results.
• Use the Right Tools: A terminology browser like the OMOPHub Concept Lookup on our website can be invaluable. Use it to find the precise concept IDs for each genotype. This level of precision is the only way to ensure your data accurately reflects the clinical reality of this complex disease.

From Clinical Reality to Actionable Data

To build a meaningful phenotype of sickle cell disease, we have to bridge the gap between a patient's lived experience and the structured data in our systems. The key is realizing that nearly every clinical event in SCD traces back to one of two fundamental problems: vaso-occlusion (blocked blood vessels) and chronic hemolysis (the rapid breakdown of red blood cells).

For anyone working with the data, this two-pathway framework is a powerful simplification. It acts as a decoder ring, allowing you to connect the dots between a specific complication-from excruciating pain to organ failure-and its biological origin. This approach makes the immense clinical complexity of SCD far more manageable, even without a background in hematology.

The Signature Events of Vaso-Occlusion

When you think of sickle cell disease, you're likely thinking of vaso-occlusion. It's the direct result of rigid, sickle-shaped cells piling up and blocking tiny blood vessels. This obstruction starves tissues of oxygen, leading to intense pain and organ damage. In electronic health records (EHRs) and claims data, this process leaves behind a clear trail of major clinical events.

• Vaso-Occlusive Crises (VOCs): These are the most common and defining feature of the SCD phenotype. You'll spot them in the data as emergency room visits or hospitalizations where the primary diagnosis is "SCD with crisis." These are episodes of severe, acute pain and are a major focus for nearly all clinical research.

• Acute Chest Syndrome (ACS): This is a life-threatening form of VOC that unfolds in the lungs and stands as a leading cause of mortality in SCD patients. In the data, ACS typically appears as a hospitalization that includes diagnoses for both SCD and pneumonia or acute respiratory failure.

• Stroke: Vaso-occlusion in the brain can trigger devastating strokes, with some studies showing that up to 24% of patients experience an overt stroke by age 45. Researchers can pinpoint these events by querying for cerebrovascular accident codes within the records of patients with an established SCD diagnosis.

Capturing the phenotype of sickle cell disease means moving beyond just a diagnosis code. It requires identifying the specific, measurable clinical events-like VOCs or ACS-that define the patient's lived experience with the disease.

Tracing the Impact of Chronic Hemolysis

While vaso-occlusion drives the acute crises, chronic hemolysis is the engine of long-term, systemic damage. This constant, premature destruction of red blood cells puts the body under immense stress. To fully capture this aspect of the SCD phenotype, it's crucial to pull in laboratory results, starting with basics like understanding the blood count test.

The fallout from rapid cell breakdown shows up in the data in a few key ways:

• Chronic Anemia: This is a universal feature of SCD. In lab data (which populates the Measurement domain in OMOP), you'll see this as persistently low hemoglobin and hematocrit levels.
• Jaundice and Gallstones: The breakdown of red cells releases a yellow pigment called bilirubin. High levels cause jaundice (icterus) and can lead to the formation of gallstones (cholelithiasis), both of which are documented with specific diagnosis codes.

If you're looking to dive deeper into how these clinical ideas are technically represented in a database, our post on semantic mapping is a great resource. By systematically connecting these biological drivers to their data footprints, you can assemble a truly comprehensive, data-driven picture of the SCD phenotype.

Unlocking Phenotypic Variability with OMOP

To really dig into the phenotype of sickle cell disease across large populations, we have to move beyond narrative chart notes and translate the disease's complex clinical story into a standardized, computable format. This is precisely the job of the Observational Medical Outcomes Partnership (OMOP) Common Data Model. By mapping all that messy, diverse clinical data into OMOP's standard vocabularies, we can create a consistent and reliable picture of the disease.

Think of it as the crucial step that turns raw, unstructured data from countless sources into a structured asset, ready for real, sophisticated analysis. It’s the bridge between a single patient's chart and a research database powerful enough to study thousands.

Translating Clinical Concepts into OMOP IDs

So, where do we start? It all begins with identifying the specific codes-the concept IDs-that represent the different facets of the SCD phenotype. These IDs are the building blocks, drawn from standard terminologies like SNOMED CT and ICD-10-CM that have been harmonized within the OMOP framework.

For instance, a general diagnosis of "Sickle-cell anemia" is a world away from "Sickle-cell/Hb-C disease" or the asymptomatic "Sickle cell trait." Nailing the right concept ID is absolutely critical. This is how we solve common data quality headaches, like distinguishing between a patient who has the symptomatic disease versus someone who is simply a carrier. The same goes for key clinical events like vaso-occlusive crises (VOCs) and acute chest syndrome, which all have their own unique concept IDs.

Here are a few foundational OMOP concept IDs for any SCD phenotype work:

• Sickle cell anemia (HbSS): SNOMED CT Concept ID 443011
• Sickle cell trait (HbAS): SNOMED CT Concept ID 83833003
• Sickle-cell/Hb-C disease (HbSC): SNOMED CT Concept ID 414879002
• Vaso-occlusive crisis (VOC): SNOMED CT Concept ID 23633003

Pro-Tip: I always recommend starting by building separate concept sets for the different genotypes and, crucially, for the asymptomatic trait. A great tool for this is the OMOPHub Concept Lookup. It lets you explore the vocabulary hierarchies, so you can find not just the main concepts but also all their descendants (more specific child terms) to build a truly comprehensive list.

The screenshot below shows what you see when you search for "sickle cell" in the OMOPHub Concept Lookup tool. It immediately highlights the variety of related terms available.

This view really drives home just how many specific terms fall under the broad "sickle-cell disorder" umbrella. It's a clear reminder that precision in concept selection is non-negotiable.

Finding Concepts Programmatically

When you're building complex phenotype algorithms, manually searching for concepts one-by-one just isn't practical. For more advanced workflows, you can automate this discovery process using the OMOPHub APIs. This is a game-changer when you need to assemble large, intricate concept sets.

The omophub-python SDK, for example, gives you a straightforward way to query the entire vocabulary and pull the exact IDs you need.

Here’s a quick look at how you could search for concepts related to "sickle cell crisis" using the Python SDK. This code is verified against our documentation examples.

from omophub import OMOPHub
# Authenticate with your API key
hub = OMOPHub(api_key="YOUR_API_KEY")
# Search for concepts containing "sickle cell crisis"
# We can also filter by vocabulary, domain, etc.
concepts = hub.concepts.search(
    query="sickle cell crisis",
    vocabulary_id=["SNOMED"],
    page_size=5
)
# Print the results
for concept in concepts.items:
    print(f"Name: {concept.concept_name}, ID: {concept.concept_id}")

For a deeper dive, you can find more detailed examples in the official omophub-python SDK on GitHub or by checking out our API documentation. If you're an R user, there's also an equivalent omophub-R SDK available. Getting comfortable with these tools is how you move from theory to practice, efficiently translating any clinical phenotype into a precise, analyzable definition inside the OMOP CDM.

Building a High-Fidelity SCD Cohort in OMOP

Taking the rich, complex picture of sickle cell disease and translating it into a computable cohort is a genuine challenge. You can't just cast a wide net for any diagnosis code mentioning "sickle cell." That approach almost always pulls in a lot of noise, especially from asymptomatic carriers of the sickle cell trait. To build a truly high-fidelity cohort, you need a precise, logical algorithm that separates the signal from the noise in your structured data.

This really comes down to defining sharp rules for who gets in and who stays out. It means hand-picking concept codes for symptomatic disease while actively filtering out any codes that only point to the trait. We also have to think about how to define key clinical events, like a vaso-occlusive crisis (VOC). A reliable method is to combine specific diagnosis codes with visit information, such as an emergency room admission, to confirm a clinically significant event. This is where longitudinal data becomes invaluable, allowing us to see these patterns unfold over a patient's journey.

Defining Inclusion and Exclusion Criteria

The foundation of a strong cohort is a set of watertight entry criteria. For anyone with symptomatic SCD, this involves building a concept set with specific OMOP concept IDs for diagnoses like homozygous sickle cell anemia (HbSS) and compound heterozygous forms like HbSC disease. These codes are our confirmation of the disease itself.

Just as crucial, though, is your exclusion list. This concept set needs to capture all known OMOP concepts for the sickle cell trait (HbAS). By making a rule to exclude any patient whose only relevant diagnosis is the trait, we can clean up our cohort significantly. It’s a common but critical step for working with messy EHR or claims data where the distinction isn't always clear.

The decision tree below provides a simple visual for how this phenotyping logic works in OMOP. You can see how patients are triaged into either a "disease" or "trait" category based on the hierarchy of their diagnosis codes.

This flowchart shows the critical fork in the road: a patient with any qualifying SCD diagnosis code is funneled into the disease cohort. On the other hand, if a patient only has codes for the trait, they're correctly filtered out. For a deeper dive into these kinds of strategies, our guide on advanced cohort study design offers some great insights.

The table below outlines what the logic for this kind of cohort definition might look like in practice.

Sample OMOP Cohort Definition Logic for Symptomatic SCD

Criteria Type	OMOP Domain	Example Logic and Concept Sets
Initial Population	Person	All persons in the database.
Inclusion Criteria	Condition	Must have at least 1 occurrence of a diagnosis from the SCD Symptomatic Disease concept set.
Exclusion Criteria	Condition	Exclude persons who have 0 occurrences of diagnoses from the SCD Symptomatic Disease concept set AND at least 1 occurrence of a diagnosis from the SCT Carrier concept set.
Outcome: VOC	Condition & Visit	A record of a "Vaso-occlusive crisis" diagnosis that occurs during an "Emergency Room Visit" or "Inpatient Hospitalization" visit.

This structured approach is what makes robust, reproducible research possible. It ensures that when we define a cohort of patients with symptomatic SCD, we are all working from the same playbook.

Expert Tip: When defining a key outcome like a VOC, always try to anchor it to a clinical encounter. Combining a diagnosis concept for "Vaso-occlusive crisis" with a visit concept for an "Emergency Room Visit" or "Inpatient Hospitalization" is a great way to confirm you’re capturing an acute, meaningful event, not just a passing note in the patient’s chart.

This level of precision is vital for tackling global health challenges. The worldwide prevalence of sickle cell disease phenotypes shows dramatic regional differences, with sub-Saharan Africa carrying the heaviest burden-nearly 80% of all global cases. In 2021, an estimated 7.74 million people were living with SCD across the globe.

These staggering numbers, which you can read more about from the World Health Organization, underscore why we need cohort definitions that are not only accurate but can also work consistently across diverse populations, especially in federated data networks. This ensures that the insights we generate are both reliable and globally relevant.

Using Data to Address Health Disparities in SCD

The technical work of defining the phenotype of sickle cell disease is more than just an academic exercise. It's a direct line to addressing the urgent, real-world need for health equity among patients around the globe.

Sickle cell disease is a condition that overwhelmingly affects people of African, Mediterranean, and South Asian ancestry. To understand-and ultimately correct-the deep-rooted health inequities they face, we first have to see them clearly in the data. This is where standardizing the SCD phenotype within a framework like the OMOP Common Data Model becomes absolutely essential.

For anyone working with clinical data, this underscores a powerful ethical responsibility. Building accurate, reliable cohorts isn't merely a technical task; it's the foundational work required to conduct research that can drive justice and fairness in healthcare.

Once we can define patient groups and outcomes with high fidelity, we can finally start asking the right questions and begin to measure the true impact of care access, socioeconomic status, and other social determinants on people's lives.

Quantifying Global Disparities

The mortality data for sickle cell disease reveals a staggering global health burden, with outcomes that vary dramatically based on where a person is born and the care they can access. The numbers themselves tell a grim story.

Even in a high-resource country like the United States, the estimated life expectancy for an individual with SCD is still more than 20 years shorter than for the general population. The outlook is far worse in lower-resource settings. In India, for instance, a heartbreaking 20% of children born with SCD die before their second birthday. These statistics highlight just how critical it is to get the SCD phenotype right. You can find more data on SCD from the CDC to see the full scale of these disparities.

This shows that an accurate phenotype is the first step toward identifying at-risk populations and channeling resources where they can save lives.

Actionable Tips for Health Equity Research

Building precise cohorts is the bedrock of any research aiming to influence policy or improve how care is delivered. If you're using OMOP data to study health disparities, here are a few practical places to start:

• Integrate Social Determinants of Health (SDOH): Don't stop at clinical data. Whenever you can, link your SCD cohort to SDOH information. This often includes area-level data on poverty, education, and housing, which can frequently be mapped to the OMOP OBSERVATION domain.
• Analyze Care Gaps: Use your cohort as a lens to examine patterns in healthcare use. You can investigate disparities in emergency room visits, the time it takes to receive treatment, or access to specialists like hematologists across different demographic groups.
• Build Better Concept Sets: Coding practices for SCD can vary significantly between institutions and populations. Use tools like the OMOPHub Concept Lookup to make sure your phenotype algorithm is robust enough to capture these nuances, ensuring you don't miss patients.

Common Questions When Phenotyping SCD

When you start digging into large-scale clinical data to phenotype sickle cell disease, a few common roadblocks always seem to pop up. Let's walk through some of the most frequent questions we see from researchers working in an OMOP environment and how to tackle them effectively.

How Do I Separate Sickle Cell Disease from Sickle Cell Trait?

This is probably the first and most critical challenge you'll face. The last thing you want is to accidentally include individuals with the benign trait in your disease cohort. The key is to be very deliberate with your concept sets.

Your strategy should be twofold. First, create a concept set specifically for Sickle Cell Disease, pulling in codes for homozygous forms (like SNOMED CT ID 443011 for Sickle cell anemia) and the various compound heterozygous forms. Second, create a separate concept set for Sickle Cell Trait using only concepts that explicitly mention "trait" (like SNOMED CT ID 83833003).

From there, building a clean cohort definition means defining your initial SCD group and then applying an exclusion criterion that removes any patient whose only relevant diagnosis is from your trait concept set. For a deeper dive into this logic, you can check out the examples in our API documentation.

Which OMOP Domains Are Most Important for an SCD Phenotype?

Don't just stop at the Condition Occurrence domain. A truly comprehensive SCD phenotype pulls from several corners of the OMOP CDM to build a complete clinical picture.

Beyond the initial diagnosis codes, you'll find critical information in these domains:

• Measurement: This is where you'll find essential lab results, like the percentage of Hemoglobin S.
• Procedure Occurrence: Look here for evidence of interventions like blood transfusions.
• Drug Exposure: This domain tracks key medications, most notably hydroxyurea.
• Visit Occurrence: Crucial for identifying hospitalizations linked to complications like vaso-occlusive crises.

Can I Phenotype a Specific Complication Like Acute Chest Syndrome?

Absolutely. This is where multi-domain phenotyping really shines. Building a reliable phenotype for a complex event like Acute Chest Syndrome (ACS) requires combining several pieces of evidence.

A strong ACS phenotype algorithm would look for a patient who has a hospitalization visit, a diagnosis code for ACS, and evidence of a new pulmonary infiltrate found on a chest X-ray. That X-ray finding might live in the Procedure domain (as a radiological procedure) or the Observation domain (as the textual report finding).

Expert Tip: You can use our SDKs, like the omophub-python SDK on GitHub, to programmatically fetch all the necessary concept IDs from SNOMED and LOINC. This makes assembling the building blocks for a complex phenotype like ACS much faster. We also offer an omophub-R SDK for R users.

Piecing these elements together ensures you're capturing a true clinical event, not just a rule-out diagnosis. If you need to find the right codes to start with, the Concept Lookup tool on our website is a great place to begin your search.

---

At OMOPHub, we handle the heavy lifting of managing and providing access to clinical vocabularies. Our developer-first platform gives your team immediate API access to OHDSI ATHENA, so you can build cohorts, run analyses, and get to your results faster. See how we can accelerate your research at https://omophub.com.