What is the STUMBL / Battle score?

The STUMBL score takes its name from the STUdy of the Management of BLunt chest wall trauma programme. In the literature it is also called the Battle score after the group that derived and validated the model. It is a clinical prediction rule designed to stratify risk of adverse outcomes after blunt chest wall trauma using a small set of variables that are usually available during an emergency department (ED) assessment.

Unlike some thoracic trauma scores built mainly from anatomy and age in multi-injury populations, STUMBL explicitly incorporates chronic lung disease and pre-injury anticoagulant use together with rib fracture burden and peripheral oxygen saturation (SpO₂). The intent is to give clinicians a structured way to think about which patients may develop delayed respiratory and related complications, so that observation, admission, and escalation decisions can be aligned with estimated risk.

Clinical problem: blunt chest trauma in the ED

Blunt chest trauma is a common presentation to emergency services worldwide. Many patients have injuries that are not immediately life-threatening on arrival yet still carry meaningful risk of delayed complications such as pneumonia and other lower respiratory tract infections, atelectasis or consolidation on imaging, pleural collections, and clinical deterioration over hours to days. Rib fractures—when present—mark increased morbidity, especially in older adults and in those with limited pulmonary reserve.

Guidelines and pathways for major trauma exist, but day-to-day decisions for the large group of patients with non-catastrophic blunt chest wall trauma can be less standardized. Symptoms and initial examination findings may underestimate later risk. Prognostic models such as STUMBL were developed to complement—not replace—bedside judgement by making risk factors explicit and combining them into a single numeric score.

What outcome does the score target?

In the primary derivation work, investigators defined a composite complication endpoint that included clinically important thoracic and selected associated events (for example lower respiratory tract infection, pulmonary consolidation, empyema, pneumothorax, haemothorax, specified solid-organ injuries when applicable, and short-term mortality, depending on the publication’s exact endpoint definition). When applying the score in practice or in research, it is essential to read the endpoint definition used in each cohort, because minor differences in what counts as a “complication” change reported sensitivity, specificity, and calibration.

Score components (how each domain is defined)

Age (points per decade)

Age contributes one point for each full decade of life. Conventionally this is implemented as the integer part of (age in years ÷ 10). For example, patients aged 10–19 years receive one point from this item, those aged 20–29 receive two points, and so on. The component captures the well-established association between increasing age and poorer outcomes after rib fractures and chest wall trauma, including tolerance of pain, cough effectiveness, and physiological reserve.

Number of rib fractures (three points per fracture)

Each identified rib fracture contributes three points. The count should reflect the best available imaging interpretation—typically a formal radiology report. Plain radiography underestimates rib fractures compared with computed tomography; cohorts that rely predominantly on chest X-ray may therefore under-ascertain fracture counts, which can lower calculated scores relative to CT-based cohorts. When no fracture is seen but suspicion remains high, clinical pathways may still warrant caution even if the fracture subscore is zero.

Pre-injury anticoagulant therapy (four points if present)

Patients taking anticoagulant medication before injury receive four additional points. This variable proxies bleeding risk and complexity of management around haemothorax, contusion progression, and need for interventions. Exact drug classes (vitamin K antagonists, direct oral anticoagulants, antiplatelet agents, etc.) and timing of last dose are not distinguished in the published point system; local protocols still govern reversal, imaging, and disposition.

Chronic lung disease (five points if present)

Chronic lung disease—commonly operationalized as chronic obstructive pulmonary disease and similar chronic pulmonary conditions—adds five points. Original data-abstraction notes in related STUMBL work have specified that asthma alone was not counted in the same way as chronic obstructive pathology; when reproducing the score for audit or research, teams should document their own inclusion rules for transparency. Reduced baseline lung function increases the impact of splinting, atelectasis, and respiratory infection after chest wall injury.

Oxygen saturation in the ED (categorical subscore)

The SpO₂ subscore uses five-percent bands downward from the high nineties, with increasing points as saturation falls. The widely reproduced summary assigns no points for saturations in the 95–100% range, then steps through higher point values through the 90s and 80s, up to ten points for the lowest tabulated band (approximately 70–74% in common tables). Saturation should ideally be the first measurement on room air in the ED; readings taken while the patient receives supplemental oxygen can mask hypoxemia and artificially lower the score. Work of breathing, perfusion, carbon monoxide exposure, and probe placement also influence SpO₂ interpretation.

Calculating the total

The total STUMBL score is the sum of the age-decade points, rib-fracture points, anticoagulant points, chronic lung disease points, and the SpO₂ subscore. There is no upper bound in principle: very elderly patients with multiple fractures, anticoagulation, lung disease, and hypoxemia can accumulate large totals. The calculator on this site walks through each contribution so users can see how individual factors drive the aggregate risk estimate.

Interpreting the total: risk strata from the derivation cohort

The original publication presented approximate probabilities of the composite complication endpoint across score ranges. Although exact percentages vary by table version, the pattern is monotonic: higher scores align with higher complication risk. For example, low totals (roughly in the single digits to ten) correspond to substantially lower reported event rates than totals in the mid-teens and twenties. These probabilities describe the derivation and validation populations; they will shift when transported to new settings with different injury mix, imaging practice, and outcome definitions.

Disposition cut-points in common use

Two thresholds are frequently quoted in summaries of the Battle model:

• Score ≥ 11: Often described as a level at which hospital admission should be considered because complication risk in the original work rose meaningfully above the lowest band. Subsequent external evaluations have asked whether adhering strictly to this threshold changes admission rates and whether it adds information beyond unstructured clinician judgement; results are mixed and context-dependent.
• Score ≥ 26: Proposed in the derivation literature as a stratum where risk is high enough that critical care or higher-acuity monitoring might be appropriate, subject to injury pattern, physiology, institutional resources, and regional trauma systems. This threshold identifies a smaller, sicker subgroup; sensitivity for complications may be limited if used alone.

Neither threshold is a mandatory rule; both should be integrated with repeat examination, analgesia and respiratory care, social factors, and available outpatient safety nets.

Strengths of the STUMBL approach

• Feasible inputs: Variables are routinely documented in many EDs.
• Transparent arithmetic: The score is easy to audit and teach.
• Face validity: Domains align with mechanisms clinicians already worry about after chest wall trauma.
• Discrimination in development cohorts: The model showed useful separation between patients who did and did not experience the composite endpoint in published ROC analyses.

Limitations and cautions

• Transportability: Calibration changes when fracture detection, oxygen measurement conditions, and comorbidity prevalence differ from the original cohorts.
• Composite endpoints: A single score predicting several distinct events can be difficult to interpret at the bedside for one specific outcome (for example isolated pneumonia versus haemothorax).
• Clinician judgement: Some prospective and retrospective studies report that fixed cut-points admit more patients than usual care without clearly superior discrimination, underscoring that scores are adjuncts.
• Exclusions: Derivation work focused on defined patient groups; patients with immediate life threats, other major injuries, or incomplete data may not match the model’s assumptions.
• SpO₂ context: Supplemental oxygen, poor signal quality, and patient agitation can mislead if not charted carefully.

Using this calculator responsibly

This tool is provided for education and clinical decision support. It does not establish a diagnosis, dictate legal standards of care, or replace serial reassessment, imaging pathways, analgesia and pulmonary toilet, or institutional trauma guidelines. Always document reasoning independent of the numeric score, and reassess if symptoms, vitals, or imaging evolve.