ODS-C common mistakes
ODS-C · Understand the traps and how to avoid them.
Applying the current-year staging manual and current data-item rules to a case diagnosed years earlier, instead of using the edition and coding rules in effect at the time of diagnosis.
Candidates study from the newest reference materials and habitually apply them to every abstract, forgetting that staging criteria, site-specific data items, and coding instructions are time-dependent and that a case must be abstracted per the rules in force on its diagnosis date.
What to do instead
Before answering, identify the diagnosis date and explicitly name which staging edition and which data-item version applies. Practice with mixed-era cases and write the applicable edition at the top of your scratch notes for each question.
Confusing clinical and pathologic staging elements, or choosing the overall stage group from the wrong element set, when the case has both clinical and pathologic information.
The abstract narrative contains imaging, biopsy, and resection findings interleaved, and candidates grab the most complete or most recent information rather than the information belonging to the specific staging classification being asked about.
What to do instead
Sort every finding into clinical versus pathologic buckets before assigning any T, N, M, or group. Recall that pathologic staging requires resection of the primary (or equivalent criteria) and that clinical staging is not overwritten by later pathologic data when the question asks for clinical stage.
Missing or misdating the reportability decision, especially for cases that become reportable only through a later event, or treating a non-reportable diagnosis as reportable because a malignancy word appears in the pathology text.
Candidates pattern-match on words like 'carcinoma' or 'malignant' without checking the reportable-list logic, ambiguous terminology rules, and the diagnosis-date rules that determine whether and when a case enters the registry.
What to do instead
Build a repeatable decision tree: is the term on the reportable list or an ambiguous term requiring a definitive diagnosis, and what date does the case become reportable? Drill ambiguous terminology pairs and borderline histologies until the decision is automatic.
Recording the wrong date as the date of diagnosis, such as using the pathology report date, the consultation date, or the date of first treatment rather than the earliest date of definitive diagnostic confirmation.
The abstract presents many dates and the pathology date is visually prominent, so candidates default to it without applying the priority rules for establishing date of diagnosis.
What to do instead
Memorize the date-of-diagnosis priority sequence and apply it explicitly to every case. When several dates appear, list them in order and justify your selection in one sentence before committing to an answer.
Selecting the wrong primary site code for overlapping or ill-defined lesions, or defaulting to an unspecified site code when the workup supports a more specific site.
Candidates are uncomfortable with the coding rules for overlapping lesions and 'unknown primary' scenarios and take the path of least resistance rather than applying the site-coding conventions and the rules for when an unspecified code is acceptable.
What to do instead
Review the coding rules for overlapping lesions and for unknown primary site, and practice with cases where imaging and surgical findings point to a site not named in the pathology report. Treat 'unspecified' as a last resort that must be justified.
Misapplying multiple-primary rules, either collapsing two distinct primaries into one record or splitting a single tumor with extension or recurrence into two primaries.
Candidates rely on intuition about whether tumors 'look related' instead of applying the formal rules on timing, histology, site, and whether a lesion is an extension, recurrence, or metastasis.
What to do instead
Work through the multiple-primary rules by histology group and time interval, and explicitly classify each additional tumor as same primary (extension/recurrence) or new primary using those criteria. Practice paired scenarios that differ only in timing or histology.
Treating treatment data collection as a simple yes/no for surgery, radiation, or systemic therapy, missing the required detail such as modality, sequence relative to other treatment, and whether the intent was curative or palliative.
Candidates focus on whether treatment happened and overlook the granular data items the registry requires, and they conflate neoadjuvant or adjuvant timing with the primary treatment field.
What to do instead
For each treatment mention, capture modality, date, sequence, and intent as separate fields. Practice abstracts where surgery follows neoadjuvant therapy, or where palliative radiation is given for a different site, to lock in the distinctions.
Answering quality and compliance questions from general principle rather than from the specific standard, such as confusing completeness, timeliness, and accuracy measures or misidentifying which edit check or standard applies.
Compliance content feels intuitive, so candidates reason about what 'should' be required instead of recalling the actual standard, its threshold, and the mechanism used to monitor it.
What to do instead
Build a one-page comparison table of the major quality standards with their measure definitions, thresholds, and the type of edit or audit used to detect failures. Test yourself by naming the standard from a described failure scenario.
Mishandling confidentiality and data-release questions by releasing or discussing identifiable data when only aggregate data is permitted, or by failing to apply the minimum-necessary and suppression concepts.
Candidates conflate internal registry use with external release and underestimate how small cell counts, dates, and geographic detail can re-identify patients.
What to do instead
Separate internal operational use from external release in every scenario, then apply the release rules and small-cell suppression logic. Practice spotting re-identification risk in tables that look de-identified but contain dates or rare combinations.
Misreading analytic outputs, such as interpreting age-adjusted rates without noting the standard population, treating crude and adjusted rates as interchangeable, or drawing conclusions from counts without denominators.
Candidates are trained on abstraction and coding and have less practice with epidemiologic interpretation, so they accept a rate or trend at face value without checking the denominator, adjustment, or comparison group.
What to do instead
For every analytic question, state the measure type, the denominator, and any adjustment before interpreting. Practice distinguishing crude versus age-adjusted rates, incidence versus prevalence, and relative versus absolute comparisons using small worked examples.