| dc.description.abstract | Abstract:
Introduction: Neurotic disorders often remain unrecognized by general practitioners in primary care. By creating a two-stepped screening tool non-recognition is counteracted. The first step consists of a prediction rule derived from a logistic regression model. Whereas the second step of screening uses the variable perceived support and focuses on the high risk group, screened in step one.
Method: A total of 3791 participants, aged 19-89, were included from general practitioners centers in Leidsche Rijn. Neurotic disorders were assessed in five year follow up with the ICPC-P coding system. Predictor variables for the first step of screening are gender, age, educational level, alcohol problems, works status, ethnicity, amount of somatic complains, and BMI. The second step of screening assesses the perceived support of patients.
Results: The first step of the screening tool was a multivariable logistic regression model including only easily obtainable characteristics had a fair discriminative value. The Area Under the Receiver Operating Characteristic Curve (C-statistic) was found to be 0.650 (95%CI: .630-.670). The prediction rule derived from the model had a sensitivity of 50.1%, specificity of 69.8%., positive predictive value (PPV) 38.3% and negative predictive value (NPV) of 78.9%. The second step of screening, the assessment of perceived support in the high risk group, improved the discriminative value of the screening tool to a sensitivity of 17.4%, a specificity of 95.9%, a PPV of 58.8% and a NPV 77.5%.
Discussion: The two-stepped screening model identifies patients as having a predispositioned high risk for obtaining a neurotic disorders, this group can be monitored by the GP to improve detection. | |
