Clinical utility of the HALP Index in the classification of schizophrenia: Logistic regression and ROC analysis

Aim: This study compared the Hemoglobin–Albumin–Lymphocyte–Platelet (HALP) index between individuals with schizophrenia and healthy controls and evaluated its discriminative performance in distinguishing schizophrenia. 

Materials and Methods: Demographic and laboratory data from 69 patients with schizophrenia and 62 healthy controls were analyzed. A multivariable binary logistic regression model predicting schizophrenia diagnosis included logHALP, age, sex, smoking status, and C-reactive protein (CRP). Model discrimination was evaluated using receiver operating characteristic (ROC) analysis and the corresponding area under the curve (AUC). In the schizophrenia group, associations between the HALP index and cognitive performance or functional outcomes were examined using Spearman’s correlation.

Results: Among 131 participants (schizophrenia, n=69; control, n=62), the median HALP index was significantly lower in the schizophrenia group (51.88 [18.89–116.69]) compared with controls (63.47 [31.57–167.54]; p<0.001). In multivariable logistic regression, logHALP demonstrated an independent protective association (B= −3.181; odds ratio [OR]=0.042; 95% confidence interval [CI]: 0.002–0.886; p=0.042), whereas CRP emerged as an independent risk marker (B=1.077; OR=2.936; 95% CI: 1.742–4.947; p<0.001). ROC analysis based on predicted probabilities yielded an AUC of 0.832 (95% CI: 0.763–0.900); as this value is derived from a case–control sample without prevalence information, it reflects discriminative rather than diagnostic performance. Using the Youden index, an optimal cutoff of approximately 0.45 provided 78.3% sensitivity and 74.2% specificity.

Conclusion: Lower HALP index values and elevated CRP levels are independently associated with schizophrenia and may serve as supportive biomarkers to aid in clinical classification and risk stratification.