Study design and participants
A cross-sectional study was conducted at the Psychiatric Hospital of the Cross – Lebanon (HPC), between July 2019 and Mars 2020. The study enrolled 120 inpatients diagnosed with schizophrenia and schizoaffective disorders and 60 healthy controls, matched for age, education and sex. The inclusion criteria for patients were as follows: inpatients aged between 18-60 years; having an educational level over five years, meeting the DSM-5 criteria (Diagnostic and Statistical Manual of Mental Disorders, fifth edition) for schizophrenia and schizoaffective disorders; receiving antipsychotic medication and clinically stable. There were no specific medication criteria for inclusion in the patient group. The healthy individuals were recruited from the staff of the HPC hospital and meeting the following criterion: absence of a history of major psychiatric disorders. Exclusion criteria for all participants included brain trauma, neurological disorder or current substance use disorder that would influence cognitive performance.
Based on a list generated from the hospital's computer software, out of 180 patients selected according to the inclusion criteria, 120 patients (71 males and 49 females) were included. Sixty patients were excluded (40 males and 20 females) for the following reasons: 22 patients refused to participate, 21 left the hospital, 13 refused to continue the assessment and 4 have a difficulty of performing the cognitive tests (Figure 1). Those who agreed to participate in the study were requested to sign a written informed consent form without any monetary compensation.
Sample size calculation
Using the Gpower 3.1.9.2 software for the calculation of the minimal sample size needed for our study, with a power of 80% (1-β=0.8) and an error α of 0.05, an effect size of 0.47 was calculated based on the original study done by Keefe et al [10]. In this study the BACS battery composite score was highly correlated with the standard battery composite scores in patients with schizophrenia (r = 0.76) and healthy controls (r = 0.90). Considering a ratio of 2:1 in each group, the results showed that the minimal sample size needed was 164 (55 in the healthy control group and 109 in the patient group).
Adaptation and assessment procedure
The Arabic version of the BACS has already been translated from English into classical Arabic language by Ateeq and collaborators using the standard translation/back-translation procedure among 33 patients with schizophrenia from Riyadh city (unpublished study). This Arabic version was sent to us by Pr. Richard Keefe and was checked and approved by the Department of Psychiatry and Behavioral Sciences at the University of Duke Medical Center.
Demographic variables (age, gender, marital status, educational level and monthly income) and clinical information of the participants (diagnosis, duration of illness in years, duration of hospitalization in years, number of hospitalization, medication intakes, and family history of mental disorders) were collected from medical files. The socioeconomic status was divided into four levels (no income, low (< 1.000 USD), intermediate (1.000-2.000 USD), high (> 2.000USD)) and the education level was divided into three levels (complementary level (> 5 years), secondary level (> 9 years) and university level (> 12 years).
Participants were tested on two separate days by two independent psychologists (one trained in the administration of the BACS and the other one trained in use of the standard tests battery) with less than two weeks between the two assessments. There were two versions of the BACS (version A and version B). These versions have the same tests with alternate forms. In our study we used version A. For the first test session, subjects received version A of the BACS and for the second session, the standard battery of cognitive tests was administered.
The BACS, included the tests listed below in the administered order:
Verbal memory – list learning
Subjects were presented with 15 words and then asked to recall as many words as possible. This procedure was repeated five times. Performance was measured as number of words recalled per trial, in any order (with a range between 0 and 75).
Working memory – digit sequencing task
Subjects were presented with clusters of numbers of increasing length. They were asked to tell the experimenter the numbers in order, from lowest to highest. Performance was measured as number of correct responses (with a range between 0 and 28).
Motor speed – token motor task
Subjects were given 100 plastic tokens and asked to place them two at a time into a container as quickly as possible. A 60-s time limit was imposed. Performance was measured in terms of the number of tokens correctly placed into the container for the first half-minute, second half-minute and the 1-min total (with a range between 0 and 100 at the final outcome).
Verbal fluency
Category instances. Subjects were given 60 seconds to name as many words as possible within a given category (animals).
Controlled oral words association test. In two separate trials, subjects were given 60 seconds to generate as many words as possible that began with a given letter: T, R. The letters “م” (similar to M in English) and letter “ج” (similar to G in English) were used since these letters in Arabic had as much word redundancy as the letters T and R in English. The overall test score refers to the number of words generated correctly within 60 seconds. The total score for the verbal fluency test refers to the sum of the three trials. Higher scores reflect a better performance.
Attention and speed information processing – symbol coding
Subjects were required to write as quickly as possible the numerals 1–9 as matches to symbols on a response sheet for 90 seconds. The measure of performance consisted in the number of correct numerals (with a range between 0 and 110).
Executive functions – Tower of London
Subjects were shown two pictures simultaneously. Each picture showed three balls of different colors arranged on three pegs, with the balls in a unique arrangement in each picture. Subjects were asked to give the total number of times the balls in one picture needed to be moved in order to make the arrangement of balls identical to that of the other, opposing picture. There were 20 trials. The items were of variable difficulty, with a general tendency for later items to be more difficult. If patients responded correctly to all 20 trials, two additional trials of greater difficulty were administered. The measure of performance consisted in the number of correct responses (with a range between 0 and 22).
Standard battery
The standard battery consisted of tests designed to examine the same constructs as the BACS. In the order given, the tests and their respective constructs are listed as follow: 16 item Free and Cued Recall test (RL/RI-16) (verbal memory), Forward and Backward Digit Span Sequencing from the WAIS-IV (working memory) [25], Trail Making Test A (TMT-A) (motor speed), Controlled Oral Word Association Test (letter "ب" (similar to B in English) and "ف" (similar to F in English), Category Instances (Fruit category) (verbal fluency), Digit Symbol Coding from the WAIS-IV (attention and speed of information processing) [25] and Block Design Test from the WAIS-IV (reasoning and problem solving) [25].
Data analysis
Data analysis was done using the SPSS software version 25. To check the distribution of normality for the BACS scale, we used the Shapiro Wilk test and we found that the major dependent variable was normally distributed. A descriptive analysis was performed where categorical variables were expressed as absolute frequencies and percentages and quantitative variables as means and standard deviations. The Chi-square and Fisher exact tests were used to compare between categorical variables and patients and controls groups while the Student T-test was used to compare continuous variables between groups.
The Arabic-BACS composite scores and the standard neurocognitive battery were determined by averaging all the subscales of each instrument, after transforming them to z-scores. Concurrent validity of the Arabic-BACS subscales was tested by using Pearson correlations with the corresponding scales of the standard battery. Construct validity of the BACS scores was assessed using the principal component analysis. To ensure the model’s adequacy, Kaiser-Meyer-Olkin measure of sampling adequacy and Bartlett’s test of sphericity were calculated. Factors with eigenvalues values larger than one were retained and scree plot method was used for determining the number of components to extract [26]. Only items with factor loading larger than 0.4 were considered [27]. Moreover, internal consistency of the Arabic-BACS was assessed using the Cronbach's alpha. Face validity was examined by comparing subtest scores between patients and controls through Student T-test. Threshold for discrimination between schizophrenic cases and controls was determined, in addition to sensitivity and specificity, using receiver–operator characteristics (ROC) curves, where all schizophrenic patients were considered “cases” and all controls “non-cases”. P-values less than 0.05 were considered statistically significant.