Validation of the lithuanian version of the self-evaluation of negative symptoms scale (SNS)

Abstract

Purpose

To validate the Lithuanian version of the Self-Evaluation Negative Symptoms Scale (Lith-SNS).

Materials and methods

A double translation from French to Lithuanian and back was performed. We included patients from in-patient and out-patient settings that had a diagnosis of paranoid schizophrenia according to ICD-10 criteria and were screened as free from acute psychotic symptoms using the Mini International Neuropsychiatric Interview (MINI). Participants were evaluated using the Brief Psychiatric Rating Scale (BRPS) and completed the Lith-SNS scale. We measured internal consistency, convergent validity, and discriminant validity of Lith – SNS comparing its scores with BPRS negative and positive symptom subscores.

Results

A total of 67 participants were evaluated. Cronbach’s alpha (α) for all 20 items of Lith-SNS (α = 0.82), and for the five subscores (α = 0.76) showed good internal consistency. Factor analysis showed a 2-factor solution which accounted for 70.12% of the variance with the first factor accounting for 53.3% and the second factor accounting for 16.8% of the variance. Lith-SNS total scores and all five subscores significantly correlated with BPRS negative symptoms subscores showing good convergent validity. There was a correlation between the Positive subscore of BPRS and the alogia subscore of Lith-SNS (r = 0.39, p = 0.001), but no correlations with other subscores or the total Lith-SNS score showing adequate discriminant validity.

Conclusions

Lithuanian version of SNS is a valuable tool to evaluate negative symptoms of schizophrenia with good internal consistency, convergent, and discriminant validity.

Keywords:

• Schizophrenia
• negative symptoms
• self-evaluation
• SNS

Introduction

Negative symptoms are regarded as one of the major and most common features of schizophrenia. Ever since 1974, these symptoms have been considered as an isolated syndrome [1]. It usually progress gradually, precede positive psychotic symptoms and have a great detrimental effect on the quality of life of the patients [2,3].

More specific diagnostic tools for negative symptoms are needed. During a Consensus Development Conference on Negative Symptoms, it was agreed upon that negative symptoms constitute a distinct therapeutic indication area and include blunted affect, alogia, asociality, anhedonia, and avolition. Negative symptoms and cognitive impairments represent separate domains. Development of new diagnostic tools that would evaluate all 5 agreed-upon symptoms and would distinguish them from depressive or cognitive impairment symptoms was set as a priority [4,5].

Items created before the Consensus Development Conference often have content validity flaws. Scale for the Assessment of Negative Symptoms (SANS), Positive and Negative Symptoms Scale (PANSS) do not evaluate all the agreed upon negative symptoms domains, they include questions evaluating other symptoms (for example cognitive impairment, depression, etc.). Moreover, SANS or PANSS have questions evaluating inadequate behavior, poor or disorganized speech, which are not considered as a part of the negative symptoms cluster and anhedonia is not assessed as one dimension and does not take into account its anticipatory and consummatory components. Brief Psychiatric Rating Scale (BPRS) is one of the oldest and most widely used psychiatric rating scales in the world. Only 3 questions out of 18 address negative symptoms, and rather vaguely evaluate blunted affect, emotional withdrawal, and motor retardation [4,6–9].

According to a recent systemic review of negative symptoms evaluation tools, very few of them have a self-evaluation format and most require an interpretation by the independent observer [6]. Therefore, results can be significantly affected by the interviewer’s experience and subjectivity. Moreover, they do not evaluate the burden of the negative symptoms experienced by patients. New generation observer rated tools like Brief Negative Symptom Scale (BNSS) and Clinical Assessment Interview for Negative Symptoms (CAINS) are of no exception in this regard and the risk that the patient might be misinterpreted, over- or under-diagnosed remains [6,7,10].

Recognizing that self-evaluation scales could improve observer rated evaluation by providing researchers and clinicians with important insight into the patient-perceived severity of negative symptoms, and adhering to the guidelines of the NIMH Consensus Statement, Dollfus et al. developed a Self-Evaluation Negative Symptoms Scale (SNS) [11]. This instrument is comprised of 20 statements about the experience of possible negative symptoms during the last week. Each statement is evaluated and scored by the patient from 0 to 2 depending whether he or she disagrees, neither agrees nor disagrees or strongly agrees with the statement. The total score ranges from 0 to 40, bigger score showing greater manifestation of negative symptoms. The 20 statements are divided into 5 sub-groups with 4 items each evaluating one of the five negative symptoms domains with a maximum score of 8 in each group. Strengths of this scale are its effectiveness evaluating each of the five negative symptoms domains and providing sub-scores for each of them; it is concise and comprehensible even for patients with thought disorders making it easily applicable in clinical and scientific environments. Unique feature of SNS is that most items are verbatim statements collected from the focus group patients with schizophrenia, therefore items are focused on the internal experience of the patients. According to Dollfus et al., SNS has good psychometric properties which were replicated in multiple SNS validation studies [11–16].

Good convergent and divergent validity and reliability are prerequisites of a good self-evaluation scale. SNS had statistically significant correlations with negative scores of PANSS [15], SANS, negative subscores of BPRS showing good convergent validity and insignificant correlations with positive subscores of BPRS and Clinical Global Impression (CGI) Parkinsonism scale illustrating adequate divergent validity [11,14]. As reported by Dollfus et al., SNS has excellent test-retest reliability with the intraclass correlation coefficients (ICCs) for the total scores of the 20 items between baseline and retest being high (ICC = 0.942, 95%, CI 0.883 − 0.971) [11]. Similar results had been shown in other SNS validation studies highlighting good reliability [12–15].

The main aim of this study was to validate the Lithuanian version of the SNS. Currently there are no self-evaluation tools for negative symptoms in Lithuania, only observer-rated tools such as BPRS or PANSS are available. Therefore, validation of SNS could significantly enrich the arsenal of both clinicians and researchers in Lithuania.

Methods

Linguistic validation

Lithuanian version of SNS has been validated in the Department of Psychiatry of the Lithuanian University of Health Sciences (LUHS). Consent and methodological support were provided by Prof. Sonia Dollfus, University of Caen, France. Original French version of the scale was translated into Lithuanian and back to French by two independent certified translators. The back-translation was reviewed by prof. Sonia Dollfus. The final reconciliation of the Lithuanian version of the scale and its clinical review was done in accordance with her remarks.

Participants

Participants of the study were recruited from the patients of the inpatient and outpatient settings of Psychiatry Department of the Hospital of LHUS. Inclusion criteria were age between 18 and 65 years, a diagnosis of paranoid schizophrenia according to ICD-10 criteria, no acute psychotic symptoms at the time of participation. Exclusion criteria were comorbid organic, including symptomatic, and mental disorders, mental retardation, mental and behavioral disorders due to psychoactive substance use, acute psychotic symptoms.

Bioethics

Bioethics approval Nr. BE-2-22 was received from the Kaunas Regional Biomedical Research Ethics Committee on the 1st of March 2019.

Procedure

Patients from inpatient and outpatient settings of Psychiatry department of the Hospital of LUHS were invited to join the study from the 1st of April 2019 until the 28th of February 2020. Each potential participant was informed about the purpose, methodology of this study and voluntary participation in it. After receiving an informed consent, participants were screened for psychotic symptoms using the Lithuanian version of Mini International Neuropsychiatric Interview (MINI) module L ‘Psychotic disorders’. Patients with no acute psychotic symptoms completed the Lithuanian version of Brief Psychiatric Rating Scale (BPRS – a short semi-structured interview aimed to evaluate general psychiatric condition, including depressive symptoms, positive and negative symptoms) and the Lithuanian version of the SNS scale (Lith-SNS).

Statistical analysis

Internal validity of Lith-SNS was calculated for the entire 20 items and 5 negative symptoms subscores using Cronbach’s alpha. Construct validity was evaluated using principal component factor analysis (PCA) with varimax rotation on the 5 subscores to examine the factor structure of the Lith-SNS. The factors retained had an eigenvalue > 0.8. Convergent and discriminant validity were assessed calculating Pearson’s correlations between Lith-SNS scores and BPRS positive and negative symptoms subscores. Statistical analysis was performed using IBM SPSS 22.0 software. The level of statistical significance was set at p < 0.05.

Results

Descriptive statistics

Study sample consisted of 67 participants with 35.8% (n = 24) male and 64.2% (n = 43) female. Age average 41.51 ± 13.76 years without significant difference between male and female.

Mean BPRS positive and negative subscores, Lith-SNS scores, are provided in Table 1. Positive BPRS subscore is a sum of BPRS items 4 (Conceptual Disorganization), 12 (Hallucinatory Behavior) and 15 (Unusual Thought Content). Negative BPRS subscore is a sum of BPRS items 3 (Emotional withdrawal), 13 (Motor retardation) and 16 (Blunted affect). These scores did not differ significantly regarding the gender.

Table 1. BPRS and Lith-SNS scores and subscores*.

	Mean score (± SD)	Differences between gender
		Male	Female	p
BPRS POZ	4.15 (± 1.46)	4.46 (± 0.36)	3.98 (± 0.19)	0.09
BPRS NEG	10.42 (± 3.95)	10.25 (± 3.96)	10.51 (± 3.98)	0.997
SNS SW	4.33 (± 2.64)	4.46 (± 2.77)	4.26 (± 2.6)	0.597
SNS DE	3.49 (± 1.69)	3.88 (± 1.96)	3.28 (± 1.5)	0.244
SNS A	4.03 (± 2.93)	4.5 (± 3.41)	3.77 (± 2.64)	0.998
SNS AV	4.03 (± 2.2)	3.83 (± 2.01)	4.14 (± 2.32)	0.29
SNS AN	3.27 (± 2.1)	3.17 (± 2.06)	3.33 (± 2.15)	0.486
SNS OV	18.61 (± 8.79)	18.83 (± 8.84)	18.49 (± 8.87)	0.705

*BPRS POZ = positive subscore of BPRS; BPRS NEG = negative subscore of BPRS, SNS SW = subscore of SNS social withdrawal questions; SNS DE = subscore of SNS diminished emotional range questions; SNS A = subscore of SNS alogia questions; SNS AV = subscore of SNS avolition question; SNS AN = subscore of SNS anhedonia questions; SNS OV = overall score of SNS.

Internal consistency

Cronbach’s alpha (α) for the 20 items of Lith-SNS (α = 0.82), and for the five subscores (α = 0.76) showed good internal consistency.

Construct validity

All five factors correlated together significantly; detriment was equal to 0.241. The Kaiser-Meyer-Olkin (KMO) index was calculated to be 0.76, and the Barlett test value was χ² = 90.486 (p < 0.001) showing that our data suited for factor analysis.

Factor analysis showed a 2-factor solution which accounted for 70.12% of variance with first factor accounting for 53.3% (eigenvalue = 2.664) and second factor accounting for 16.8% (eigenvalue = 0.841) of variance. Factor 1 consisted of social withdrawal, reduced emotional range and anhedonia and factor 2 grouped alogia and avolition subscores (see Table 2).

Table 2. Principal component analysis, varimax rotated component matrix.

SNS subscores*	Factors
	1	2
Lith-SNS SW	0.851**	0.196
Lith-SNS DE	0.786	0.116
Lith-SNS AN	0.753	0.403
Lith-SNS AV	0.240	0.779
Lith-SNS A	0.158	0.833

*Lith-SNS SW = subscore of Lith-SNS social withdrawal questions; Lith-SNS DE = subscore of Lith-SNS diminished emotional range questions; Lith-SNS A = subscore of Lith-SNS alogia questions; Lith-SNS AV = subscore of Lith-SNS avolition question; Lith-SNS AN = subscore of Lith-SNS anhedonia questions; **Correlations of each of the factors with the SNS subscores.

Convergent validity

Lith-SNS total scores and all five subscores significantly correlated with BPRS negative symptoms subscores showing good convergent validity (see Table 3). Only correlation between Lith-SNS avolition subscore and BPRS emotional withdrawal subscore was moderate (r = 0.411, p < 0.001), all other correlations between Lith-SNS subscores and BPRS blunted affect, emotional withdrawal, motor retardation, and total negative subscores were high (r > 0.5, p < 0.001).

Table 3. Correlation between BPRS* negative subscores and Lith-SNS scores**(Pearson r scores, all p < 0.001).

	Lith-SNS SW	Lith-SNS DE	Lith-SNS A	Lith-SNS AV	Lith-SNS AN	Lith-SNS OV
BPRS BA	0.624	0.553	0.591	0.512	0.667	0.749
BPRS EW	0.626	0.405	0.565	0.411	0.542	0.657
BPRS MR	0.683	0.536	0.732	0.64	0.722	0.85
BPRS NE	0.726	0.566	0.714	0.595	0.731	0.853

*BPRS BA = BPRS item 16, blunted affect score; BPRS EW = BPRS item 3, emotional withdrawal score; BPRS MR = BPRS item 13, motor retardation subscore, BPRS NE = sum of BPRS BA, EW and MR; ** Lith-SNS SW = subscore of Lith-SNS social withdrawal questions; Lith-SNS DE = subscore of Lith-SNS diminished emotional range questions; Lith-SNS A = subscore of Lith-SNS alogia questions; Lith-SNS AV = subscore of Lith-SNS avolition question; Lith-SNS AN = subscore of Lith-SNS anhedonia questions; Lith-SNS OV = overall score of Lith-SNS.

Discriminant validity

Discriminant validity was assessed calculating Pearson’s correlations between Lith-SNS scores and BPRS positive symptoms subscores. BPRS positive subscores significantly correlated with alogia subscores of Lith-SNS (r = 0.39, p = 0.001), but did not correlate to any other subscores or the total scores of Lith-SNS. Also, Lith-SNS negative symptoms subscores and total score did not correlate with the depression question of BPRS (item 9), showing good discriminant validity.

Discussion

SNS is the only self-evaluation scale that evaluates all five negative symptoms domains. Other scales such as Social Anhedonia Scale, The Subjective Experience of Deficits in Schizophrenia Scale or Community Assessment of Psychic Experience only evaluate a portion of negative symptoms and include questions evaluating cognitive decline, depressive symptoms, etc [17–19]. MAP-SR is a scale that was developed after the consensus conference and can be called a second-generation scale, but it only is adequate in evaluating motivation and anhedonia [20].

We found adequate psychometric properties of the Lithuanian SNS which are comparable to other validation studies. In the original validation study Dollfus et al. found good internal consistency for the 20 items of SNS (alfa = 0.867) and for the five subscores (alfa = 0.784), good convergent validity with significant correlation with BPRS negative subscores (r = 0.298, p = 0.037) and the SANS global evaluations (r = 0.628, p < 0.0001). Adequate discriminant validity was found because SNS scores did not correlate with the BPRS positive subscores (r = 0.253, p = 0.079) [11]. These results were replicated by Wojciak et al. in their Polish SNS validation study and Hervochon et al. in their French SNS validation study [12,14]. Some differences with other SNS validation studies could occur because different items could be included in the BPRS positive subscores in different studies, as there is no clear agreement on which items should be included in the positive subscore. We were not able to perform comparisons of Lith-SNS with SANS or BNSS for convergent and discriminant validity (as was done by authors of other validation studies) because these scales are not available in Lithuanian language [11–14,21].

When performing the principal component analysis we found that a two factors solution explained 70.12% of the variance with the KMO index of 0.76. This is comparable to the original validation study performed by Dollfus et al. in French and Spanish validation studies of SNS. However, Dollfus et al. found one factor consisting of anhedonia, avolition, asociality and alogia subscores (apathy factor) and second factor consisting of diminished emotional range (emotional factor) while in our study we found one factor consisting of social withdrawal, reduced emotional range and anhedonia, and factor 2 grouped alogia and avolition subscores [11,13,14]. Different makeup of factors could be due to smaller sample sizes compared to Spanish validation study or different sample compared to Dollfus et al. study.

The main drawbacks of this study were a relatively small sample size, use of only one scale (BPRS) as a reference for convergent and discriminant validity. Larger sample size and more scales for comparison would allow performance of a more decisive principal component analysis and discriminant and convergent validity analysis. Also, there was a significant inequality between the number of women and men surveyed. Considering possible gender differences in the clinical picture of schizophrenia, the significant difference of the study groups could influence the study outcomes.

Conclusion

Results of this study show that Lithuanian version of SNS is a valuable tool to evaluate negative symptoms of schizophrenia with good internal consistency, convergent and discriminant validity.

The main directions for future research of negative symptoms in Lithuania should be a validation of observer-rated scale for negative symptoms (e.g. BNSS). Having both a self-evaluation and observer rated evaluation tools for negative symptoms would enable more precise evaluation of these symptoms in clinical and research settings.

Disclosure statement

No financial interest of benefit has arisen from the direct applications of our research.

Data availability statement

Due to the nature of this research and the requirements of Bioethics committee, data is not available publicly but could be received from the corresponding author upon demand.

Additional information

Notes on contributors

Jonas Montvidas

Jonas Montvidas is a resident of psychiatry in the Psychiatry Department of the Hospital of the Lithuanian University of Health Sciences.