AI-Driven Clinical Deterioration Index Flags High-Risk Geriatric Hip Fracture Patients After Surgery: 93.3% Complication Rate and 20.0% One-Year Mortality at CDI ≥65

In geriatric hip fracture patients after surgery, an elevated AI-driven CDI was associated with worse outcomes: CDI ≥65 versus <65 showed complication rates of 93.3% vs 26.4% (P < 0.001), length of stay 10.1 vs 5.2 days (P = 0.047), discharge ambulation 8.8 vs 42.3 feet (P < 0.001), and one-year mortality 20.0% vs 3.8% (P = 0.02), with CDI ≥65 associated with postoperative complications at an odds ratio of 37.9. The institutional cutoff of 65 had been validated previously but not for an orthopaedic population; this study tests the tool in the specific setting of postoperative geriatric hip fracture and reports differences in complications, length of stay, ambulation, and one-year mortality. Retrospective cohort at a single Level I trauma center, 197 patients (age 65–98, 66% female), therapeutic Level III evidence; only 15 patients exceeded the threshold, a small subgroup, with effects expressed as odds ratios and between-group rate differences.

The institutional CDI cutoff of 65 performed with high specificity and low sensitivity in this cohort: specificity 99.3% and sensitivity only 22.6%. This indicates that a threshold validated in other settings would miss many patients who ultimately had complications in the orthopaedic postoperative population; the source frames this as a reason to recalibrate cutoffs by specialty rather than as a failure of the tool. Threshold performance computed within the same 197-patient cohort, a retrospective diagnostic-accuracy assessment without prospective validation.

Optimizing the threshold to 47.7 raised sensitivity from 22.6% to 77.4% while accuracy remained comparable at 75.0%. This provides concrete numbers for a specialty-specific cutoff, showing that the proportion of high-risk patients detected can be substantially increased without sacrificing overall accuracy. Threshold re-analysis within the same cohort, a data-driven cutpoint optimization; the source does not state whether the new threshold was validated in an independent cohort.

The results apply to patients aged 65 and older who underwent fixation of OTA/AO 31-A/B/C hip fractures at a single Level I trauma center and had 30-day follow-up, excluding pathologic fractures, revision surgery, and missing follow-up. For such a setting, the work supports a rapid stratification using peak CDI within the first 48 postoperative hours and suggests that a specialty-specific cutoff (such as 47.7) may identify more high-risk patients than the inherited 65, informing closer monitoring, early rehabilitation, or more intensive follow-up.

Only 15 patients exceeded the threshold, so the stability of between-group comparisons awaits larger samples; the 47.7 cutoff came from data-driven optimization in the same cohort, and whether it retains 77.4% sensitivity and 75.0% accuracy in independent populations remains open; the association between CDI ≥65 and complications is observational and does not establish causation; the single-center design and the institutionally validated 65 cutoff mean external use requires recalibration. The source is full text, but supplemental digital content is not expanded in the body, so readers needing specific measure weights or the full complication list would still consult the supplement.