Measuring the Quality of Neurosurgical Care Through Intracranial Suppurations: A Twenty-Indicator Framework Tested on 150 Cases at Bouaké University Hospital

Background. Global surgery indicators measure a country’s capacity, not what happens to a patient between the first symptoms and discharge. Intracranial suppurations (ICS) test every link in the chain of care. We used them as a tracer condition to build and test a framework for measuring quality in neurosurgery. Methods. Single-center cohort study, retrospective then prospective, of 150 observations (148 patients) of ICS managed at Bouaké University Hospital (Côte d’Ivoire) from 2016 to 2026. This is a secondary analysis of a cohort whose main study is the subject of a separate manuscript. The ISQF framework comprises twenty indicators distributed according to the Donabedian model: six of structure, eight of process, and six of outcomes. A composite compliance score (ISI-QI) was calculated for each patient. Two Firth penalized logistic regression models (in-hospital mortality; unfavorable outcome, GOS 1–3) combined the ISI-QI, a Glasgow Coma Scale (GCS) score below 9, and an age of at least 65 years. Internal validation relied on 2,000 bootstrap samples with optimism correction. No protocol was registered. The report follows STROBE, SQUIRE 2.0, and TRIPOD+AI; no machine learning method was used. Results. Overall process compliance was 89.0% (878/987 care opportunities; 95% CI: 86.9–90.8). Clinical indicators met or approached their targets. Timeliness indicators fell short: 67.0% of patients with a surgical indication were operated on within 48 hours, and 2.0% were admitted within seven days of symptom onset. The median delay to surgery was 20 days, of which about 95% elapsed before reaching the hospital. Only 34 of 150 patients (22.7%) received antibiotic therapy guided by the antibiogram. The ISI-QI was not associated with mortality (aOR 1.18 per 10-point increment; 95% CI: 0.72–2.06) but was associated with unfavorable outcome (aOR 1.52; 1.04–2.29), solely because of the early-surgery indicator (aOR 3.72; 1.31–11.41): patients operated on early were more severe (GCS < 9 in 18.2% of cases versus 2.6%). Without this indicator, the association disappeared (aOR 1.15; 0.80–1.69). After optimism correction, the ISI-QI added nothing to initial severity for mortality (C statistic 0.646 versus 0.651), and for functional outcome the apparent gain (0.676 to 0.739) disappeared without the early-surgery indicator (0.685). Conclusion. ICS lend themselves to quality measurement from ordinary records. The failure lies before the hospital and in the microbiological chain, not in the operating room. The ISI-QI describes the compliance of an organization; it does not predict the outcome of a patient. From this we derive a ten-indicator dashboard, the ISI-QI-10. Level of evidence. III. Protocol registration. None (secondary analysis of an observational cohort).