The European Portuguese Version of the Brief Negative Symptom Scale

Negative symptoms reflect a currently much-untreated loss of normal functioning and are frequently found in psychotic disorders. We present the first translation of the Brief Negative Symptom Scale (BNSS) to European Portuguese and evaluate its validity in a sample of Portuguese male patients with a psychotic spectrum disorder. The Portuguese BNSS showed excellent internal consistency, high convergent validity (i.e., strong correlation with the PANSS negative factor), and high discriminant validity (i.e., a lack of association with the PANSS positive factor). In sum, the present European Portuguese BNSS has shown to be reliable, thus extending this instrument’s clinical availability worldwide.


Introduction
Negative symptoms relate to the inability to experience thoughts and behaviors that are usually present in most individuals.These were classically described in the Four A's mnemonic regarding schizophrenia, at the beginning of the last century: alogia, autism, ambivalence, and affective blunting [1].These symptoms are associated with poor functional outcomes at the work, social, self-care, or family levels [2] and still lack effective treatment solutions [3].In addition to being present in the general population with neurological disorders [4][5][6][7], negative symptoms are frequent in psychosis, particularly schizophrenia.The first step to effective treatment is to recognize these symptoms and objectively characterize their dimensions.To meet this demand, and following the Consensus Development Conference on Negative Symptoms [8], we have previously developed the Brief Negative Symptom Scale (BNSS) [9].As a 15-min interview, the BNSS can be delivered by trained mental health professionals in a clinical trial setting or a regular appointment; and consists of 13 items that evaluate consensual domains (anhedonia, asociality, avolition, blunted affect, and alogia) and an additional item that measures a reduction in normal distress.Each item is scored from 0 to 6, ascending with severity, and the score for each domain is calculated as the sum of the scores of its items.
Translated into multiple languages, psychometric evidence suggests that BNSS is reliable for evaluating the underlying mechanisms of negative symptoms [10].To our knowledge, the BNSS has not been assessed in a European Portuguese sample.We found two translation studies in Brazilian samples [11,12] and one in a Spanish [13] sample, which are the closest to European Portuguese, in terms of culture and language.Despite being the same language, the Brazilian Portuguese variety has substantial differences from European Portuguese, such as in: 1) grammar (for example, the frequent use of the gerund tense in Brazil) and 2) vocabulary (several words and expressions are not commonly used in one on the varieties or are so but have a different meaning (for example, in one question evaluating item 3 under anhedonia: "Existe alguma atividade que você fica na expectativa para fazer?" [Brazilian version] vs. "Haverá algo mais que o deixe entusiasmado por vir a realizar?"[Portuguese version]).Thus, in this study we aimed, for the first time, to translate BNSS to European Portuguese (from Portugal) and to validate it in a sample of Portuguese patients with a psychotic spectrum disorder.

Methods
We recruited patients from Hospital Júlio de Matos of Centro Hospitalar Psiquiátrico de Lisboa (CHPL).The study was approved by the ethical (Ref.CES005/2020) and scientific (Ref.CCP0031/2020) boards of CHPL, and data were collected from June 1, 2020, to January 31, 2021.All subjects provided informed consent.Fifty-one male patients were included.The inclusion criteria comprised being male, aged 20-55 years old, having Portuguese as a native language, at least 4 years of education, and a diagnosis of a psychotic disorder according to the International Classification of Diseases -10th version (ICD-10): of schizophrenia (F20), persistent delusional disorder (F22), or schizoaffective disorder (F25), confirmed with the Mini International Neuropsychiatric Interview (MINI; Psychosis section) [14] and by hospital chart-review, for at least 2 years and not before the age of 15; taking the same psychiatric medications for at least 6 weeks; and, as opportunistically part of an umbrella pharmaconeuroimaging study, accepting blood sampling, drug administration, and fMRI brain scanning.Exclusion criteria were hospital admission due to psychiatric-related symptoms during the previous 2 months; illegal drug consumption in the previous month; alcohol consumption greater than 28 units per week; previous diagnosis of neurological, hormonal, or serious hepatic, gastrointestinal, infectious, cardiovascular, renal or hematological conditions; a pervasive developmental disorder (ICD-10, F80-F89); color blindness, previous traumatic brain injury with loss of consciousness, history of seizures, and premature birth (i.e., before the 37th week of pregnancy gestation).
Sociodemographic and clinical data were collected by phone (Table 1) and the BNSS [9] and the Positive and Negative Syndrome Scale (PANSS) [15] were delivered either by video call or in person at CUF Infante Santo hospital.The PANSS includes three subscales (positive, negative, and general psychopathology subscales) with a total of 30 items rated from 1 ("Absent") to 7 ("Extreme").The positive subscale measures seven symptoms (P1-P7), likewise in the negative subscale (N1-N7), and the General subscale consists of 16 European Portuguese BNSS items (G1-G16).The subscales' scores are the result of the sum of its items.These assessments were conducted by three previously trained clinicians: two medical doctors (H.S., A.R.) and a clinical psychologist (S.F.).The same clinicians utilized the BNSS and the PANSS for rating every patient.We used existing and widely used, albeit not academically published, European Portuguese (i.e., from Portugal) versions of the PANSS and MINI.We translated the BNSS (version 2.0), which was then linguistically validated by the scale's original author (B.K.) on December 3, 2019, after back-and-forth translations and resolutions and a clinical review.The translated version is provided in the online supplementary Material (for all online suppl.material, see https://doi.org/10.1159/000530705.Convergent and discriminant validity were performed using the recent PANSS 5factor model consensus [16].According to this model, the PANSS negative factor features the items Blunted Affect (N1), Emotional Withdrawal (N2), Poor Rapport (N3), Passive/ Apathetic Withdrawal (N4), Lack of Spontaneity and Flow of Conversation (N6), Motor Retardation (G7).The description of the other factors can be found in the online supplementary Material S1.Data (N = 51) were analyzed using SPSS v27, IBM.For the reliability analysis of the BNSS validation, Spearman's correlations were conducted between the BNSS and PANSS scores, and Cronbach's alpha ≥0.7 and ≥0.8 were considered acceptable and excellent internal consistency values, respectively.Results were considered statistical significant when p < 0.05.Online supplementary Table S2 and Table 2 show the means and standard deviations of the PANSS and BNSS measurements, respectively.

Results
BNSS items showed excellent internal consistency, with a Cronbach's Alpha of 0.907 (Table 2).All BNSS items were significantly correlated with the BNSS total scale score (Table 2).Alpha if-item-deleted coefficients ranged from 0.90 to 0.91.Regarding convergent validity, BNSS  was strongly correlated with the PANSS negative factor (r = 0.926, p < 0.001) (Table 3).In terms of discriminant validity, we did not find statistically significant correlations between BNSS, PANSS positive factor (r = 0.041, p = 0.775), and PANSS depressed factor (r = 0.110, p = 0.443) (Table 3).Other exploratory correlations of BNSS and its domains, with PANSS scale and its factors, are provided in the supplementary material (online suppl.Table S3).

Discussion
In our sample, the BNSS's internal consistency and validity were congruous with a recent review [10].Convergent validity was confirmed by the strong and significant correlations between BNSS, its domains, and the PANSS negative factor.This suggests that the BNSS is able to accurately address negative symptoms in the Portuguese population, although our study is only generalizable to male patients with psychotic disorders.We further note that we found the scale to also be valid in a heterogeneous sample including several types of psychotic illnesses, e.g., delusional disorder, schizophrenia, and schizoaffective disorder, which indicates a high degree of applicability within psychosis.
The effect size (r = 0.93, p < 0.001) was en par, and slightly higher, than that of previous validation studies ((r = 0.89, p < 0.001) [17]; and (r = 0.866, p < 0.001)) [11].The effect size increase versus the later (Brazilian) study may be because, in ours and the German study, the rater of the PANSS and BNSS was the same person withinstudy; or it could be due to our study using an only-male sample, or a more accurate translation.Furthermore, in our study, BNSS did not significantly correlate with the PANSS positive factor and PANSS depressed factor, suggesting good discriminant validity since the mentioned PANSS factors measure other constructs.
When comparing our validation findings with the BNSS validation findings in samples closely related to European Portuguese, culturally and linguistically, such as the existing studies with a Spanish sample [13] and Brazilian patients [11,12], we found similar consistency metrics.While our Cronbach's alpha was 0.91, the Spanish study [13] found one of 0.98 and the Brazilian studies found 0.94 [11] and 0.88 [12].Our sample size was surpassed by only one of these three studies that collected 111 patients [11], while the Spanish including 20 [13] and another Brazilian including 30 patients [12].In sum, our findings confirm the cross-cultural properties of BNSS and extend previous validations in Brazilian Portuguese and Spanish cultures to European Portuguese.
As a limitation, our study was opportunistically part of a larger one-session male-only pharmaco-neuroimaging umbrella study, which reduced the number of patients enrolled and the representativeness of its clinical population to only those that were male, willing and able to participate in the umbrella study.However, our sample size is still in line with the original scale study [9] and several other validation studies [13,[18][19][20].Second, both BNSS and PANSS scales were rated by the same clinician, possibly leading to stronger correlations between the scores of both scales.Third, the SARS-CoV-2 pandemic and its mobility restrictions also contributed to selection bias and may have influenced some of our results like the assessment of gestural expressions for BNSS-blunted affect and self-reported participation in recent social activities for BNSS-avolition and BNSS-asociality due to a decreased opportunity to socialize.
In conclusion, BNSS, which has consistently demonstrated effectiveness in assessing negative symptoms, is herein shown reliability in male Portuguese patients with psychotic disorders.Given the large impact of negative symptoms on functional outcomes [21,22] and their currently difficult treatment, we hope this tool can help characterize this symptomatology in this population.

Table 1 .
Means Number of individuals; mg, milligrams, Antipsychotic doses were converted to daily ChlorPromaZine Equivalents (CPZE).**Antidepressant doses were converted into daily fluoxetine equivalents (FluE).Methods for calculating CPZE and FluE are detailed in the online supplementary Material S1.

Table 2 .
Means, standard deviations (SDs), and correlations of BNSS total scale and subscales