From May 2011 to January 2017, patients diagnosed with CuTS were recruited from the outpatient service of the hospital. The number of variable was used to calculate the optimal number of participants for this study. In this study, the sample size was typically expressed in terms of events per variable. An events per variable of 10 was widely used as the low limit for developing the models that predicted a binary outcome . The sampling method was convenient sampling. The diagnosis of CuTS was made using a combination of clinical evaluation and nerve conduction study. Clinical assessments included the history of initial presence with intermittent paresthesias, numbness, and tingling in the small finger and ulnar half of the ring finger . According to the guideline of the American Association of Electrodiagnostic Medicine , all patients underwent the nerve conduction study. Confirmatory criteria included: (1) motor nerve conduction velocity (MNCV) across the elbow less than 50 m/s; (2) an MNCV difference of greater than 10 m/s between the elbow segment and forearm segment; (3) a conduction block with compound muscle action potential decreased more than 20% (amplitude measured from the elbow to upper arm). Electrodiagnostic studies were performed by a specialist technician using a Dantec Keypoint Portable Nerve Conduction/ electromyography machine (Dantec Dynamics, Bristol, Bristol, UK) and reported by a consultant neurophysiologist. The specific process is shown in Figure 1.
Patients were selected based on one of the following criteria: (1) patients with subjective symptoms, no matter the presence of intrinsic muscle atrophy or not; (2) electrodiagnostic evidence of CuTS; and (3) age > 18 years because the study just focused on adult patients. Patients with one of the following criteria were excluded: (1) age ≤ 18 years; (2) patients with other neuropathy confirmed electrophysiologically; (3) patients refused to attend the study; (4) patients who had undergone previous treatments, such as splinting, steroid injection, or cubital tunnel release; (4) a previous diagnosis of anxiety, depression, and other psychiatric disorder; (5) pregnant and lactating women because they were at high risks of depression and anxiety that distorted the assessments .
Demographics and Clinical Evaluation
We used the self-administered questionnaire to assess the patients, which consisted of two parts. The first part included patients' demographic data (age, gender, educational level, marital status, job status, and socioeconomic status). The second part was patients' clinical data (hypertension, diabetes mellitus, tobacco use, alcohol use, history of cancer, and duration of symptoms). Patients were determined as older (> 50 years) and younger (≤ 50 years) adults. Educational level was registered as university degree, primary and middle degree, and illiterate degree. Marital status was coded as married, single (with or without cohabiting), widow, and divorced. Job status was registered as employed (in vacation or not) and unemployed (including students). Based on the classifying occupation, employed patients were classed into three groups: unskilled group, semi-skilled / skilled group and semi- professional/professional group. The criteria of classifying occupation were shown as follows: unskilled occupation: work not requiring education/training - eg: peon, watchman, domestic servant, laborer; semiskilled occupation: need some training to do the routine job efficiently - eg: laboratory attendant, library attendant; skilled occupation: long training in complicated work - eg: carpenter, mechanic, driver, telephone operator; semi-professional - high school teacher, lecturers in college, junior administrators, research assistants; professional- those involved in decision making, laying down policies and executing them - eg: doctors, senior administrative officers, senior lecturers, readers, professors, principals of colleges, advocates, engineers. The socioeconomic status was recorded based on the total amount of family income (US Dollar/per year) (high > 12, 735; medium = 4,126 to 12, 735; low < 4,126). The duration of symptoms was determined as long (> 2 years) and short (≤ 2 years) terms. Based on the standard of World Health Organization, alcohol use was defined as ≥ 60g on one occasion in the past 30 days . Patients were asked to complete the Quick Disabilities of the Arm, Shoulder, and Hand (QuickDASH) questionnaire to assess the hand function (0= no disability, 100= total disability) [16, 17]. It is a more quickly administered version of DASH, developed using Rasch analysis . It has adequate convergent and discriminant validity, excellent internal consistency reliability with a Cronbach α of 0.89 in the primary care setting . Based on the modified McGowan grade , the patients were classified into four groups (grade I: subjective symptoms, no abnormal objective findings; grade IIa: good intrinsic strength (4/5), no detectable muscle atrophy; grade IIb: fair intrinsic strength (3/5), detectable muscle atrophy; grade III: profound sensory and motor disturbances with marked intrinsic atrophy). We used the Chinese version Hospital Anxiety and Depression Scale (HADS) (≥11points, probable disorder; 8 to 10, possible cases; and ≤ 7, no case) to determine the anxiety and depression . The HADS (21-item) contains anxiety sub-scale (HADS‐A) and depression sub-scale (HADS‐D) with each including 7 questions. We classified the patients as being depressed or anxious (present case, ≥8 points) or nondepressed/nonanxious (absent case, ≤7 points). The cut-off scores of depression and anxiety were 8 points. In China, the validated HADS is commonly used for interviewing participants. The Chinese version of HADS demonstrated the similar satisfactory linguistic equivalence, conceptual equivalence, and scale equivalence (concordance rates at the cutoff of 8 for anxiety and depression sub-scales were 89% and 87%, respectively; and at the cutoffs of 11 were 87% and 91%, respectively) compared with the English version .
Data was collected by a clinical psychologist (XS) who filled out the pre-coded structured questionnaire that comprised demographic and clinical data collected from the patients.
The main outcomes were the proportions of anxiety and depression, as well as the associated factors. Associations between potential prognostic determinants and outcomes were examined using univariate logistic regression analysis. Predictors univariately associated with outcome (p < 0.05) were included in a multiple-predictor logistic regression model. Then multiple logistic regression (Backword-Wald) was carried out to identify the variables that were independently associated with anxiety and depression. Logistic regression was expressed as odds ratios with a 95% confidence interval (CI). Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS, version 25, Chicago, IL) for windows.