Review
On the use of the Positive and Negative Syndrome Scale in randomized clinical trials

https://doi.org/10.1016/j.schres.2015.04.006Get rights and content

Abstract

In the last 25 years, the Positive and Negative Syndrome Scale (PANSS) has been largely used to assess schizophrenia symptom intensity, but little information is available on how this scale was generally applied when evaluating the efficacy of schizophrenia therapies in randomized clinical trials. In the attempt to address this topic, a systematic PubMed Search was carried out using the keywords “PANSS” and “Randomized Clinical Trials”. The analysis of retrieved articles highlighted that PANSS has constituted a suitable psychometric instrument to investigate the efficacy of pharmacological and non-pharmacological therapies. However, the information potentially provided by this scale was only partially reported in research articles, when characterizing the symptomatic features of patients at baseline. Furthermore, a consensus is needed to identify methodological strategies that may properly adapt PANSS-subscale structure with the symptomatic profiles of individuals enrolled in randomized controlled trials. The possibility that PANSS interview procedures and enrollment eligibility criteria may influence the symptomatic composition of patients involved in these studies is also discussed.

Introduction

Twenty-five years have already passed since Kay and collaborators developed the Positive and Negative Syndrome Scale (PANSS) with the purpose to provide a comprehensive and reliable instrument for assessing schizophrenia symptoms (Kay et al., 1987). The scale comprised thirty items previously included in the Brief Psychiatric Rating Scale (18 items) and in the Psychopathology Rating Scale (12 items). PANSS items were divided in three subscales (Positive, Negative and General Psychopathology subscales), starting from the hypothesis that positive and negative symptoms may constitute the main dimensions of schizophrenia (Crow, 1980, Andreasen and Olsen, 1982), and in order to provide a contextual assessment of patient general psychopathological status (Kay et al., 1987). PANSS was also combined with a semi-structured interview and detailed definitions of seven levels of symptom intensity (scores from 1 to 7) were provided for each item (Kay et al., 1987).

Preliminary studies indicated that PANSS possessed an adequate construct validity, inter-rater reliability and external validity (Kay et al., 1987). Further analyses confirmed that PANSS may constitute a suitable instrument for the assessment of schizophrenia symptoms (Kay et al., 1988, Kay et al., 1989), however some PANSS characteristics became a matter of discussion. Kay and collaborators indicated that a four dimensional pyramidal model could better describe the heterogeneous symptomatic features of schizophrenia (Kay and Sevy, 1990). Furthermore, PANSS-subscale structures including four, five, or six subscales have been proposed starting from the results of principal component analyses (PCAs) and confirmatory factor analyses (CFAs) (Lindenmayer et al., 1994, Peralta and Cuesta, 1994, von Knorring and Lindström, 1995, Marder et al., 1997, Van den Oord et al., 2006, Villalta-Gil et al., 2006). Recently, studies based on item response theory indicated that PANSS should be modified taking account the necessity to reduce item redundancy and response options (Khan et al., 2011, Levine et al., 2011). Furthermore, it was suggested that PANSS items should be rescaled with the assignment of score ranging from 0 to 6, so to avoid incorrect calculations when analyzing data (Leucht et al., 2010, Obermeier et al., 2010). Finally, Kay and collaborators combined the structured clinical interview for the PANSS (SCI-PANSS) (Opler and Ramirez, 1992) with the DSM-III-R Structured Clinical Interview and Rating Criteria (SCID), with the purpose to integrate the functional-dimensional assessment of PANSS into the diagnostic system of the American Psychiatric Association (Kay et al., 1991).

Beside the attempts to improve the validity and manageability of PANSS, this scale remains one of the most widely used psychometric tools to assess schizophrenia symptoms. PANSS was translated and validated in several languages (Kay et al., 1990, von Knorring and Lindström, 1992, Igarashi et al., 1998, Lançon et al., 1999) so that clinicians of different countries could use this psychometric instrument for their evaluations of schizophrenia symptoms in first-admitted patients and in those receiving therapies. Moreover, a variety of clinical trials adopted PANSS to evaluate the efficacy of new therapeutic strategies in controlling schizophrenia symptoms. Particularly, PANSS has been extensively used in randomized clinical trials (RTs), whose experimental design is considered highly trustable by researchers, since the occurrence of experimental bias or data misinterpretations is reduced by the randomized assignment of subjects to different experimental groups. The possibility to conjoin the reliability of RT experimental design with the completeness of PANSS was indeed regarded as an advantageous approach for a proper assessment of the efficacy of schizophrenia therapies.

In spite of the large use of PANSS, there is still little information on how this scale was generally applied in RTs. It is unclear which of the proposed modifications of PANSS have been accepted by researchers, and whether the large amount of quantitative and qualitative information potentially provided by PANSS was generally reported in research articles. Finally, there is no global data depicting the symptomatic characteristics of patients involved in RTs using PANSS (PANSS-RTs) to assess schizophrenia symptom intensity.

In the attempt to collect information on the use of PANSS in RTs, the present analysis reviewed the available scientific literature, focusing on (i) how PANSS results are depicted within research articles; (ii) the PANSS subscales mostly used in RTs; and (iii) the demographic and symptomatic characteristics of individuals enrolled in PANSS-RTs.

Section snippets

Data source and inclusion criteria

A systematic PubMed advanced search of PANSS-RT articles was carried out using the keywords “PANSS” and “Randomized Clinical Trial”. All English written research articles reporting PANSS-RTs published between 1987 and 2012 were collected, but only PANSS-RT articles showing original data (o-PANSS-RT articles) were included in the analysis, so to avoid overestimation when evaluating the characteristics of patients enrolled in PANSS-RTs.

The therapeutic area of interest of all o-PANSS-RTs was

Retrieved PANSS-RT articles

The PubMed search retrieved 526 articles and, among them, 363 articles (69%) were selected as o-PANSS-RTs. The number of o-PANSS-RT articles strongly increased from 1994, while fluctuations could be observed starting from 2007 (Fig. 1). More than half of o-PANSS-RTs evaluated the clinical efficacy of antipsychotic drugs, but an increasing amount of research articles focused on other therapeutic approaches, such as psychological interventions, nutraceutical/plant-extract/hormone and Transcranial

Discussion

A large amount of PANSS-RT research articles have been published since 1987, confirming the suitability of this investigational approach when assessing the efficacy of schizophrenia therapies. Interestingly, the number of PANSS-RTs fluctuated but remained rather stable after 2007, possibly mirroring the occurrence of a sort of ‘re-thinking time’ in the development of new therapeutic strategies, as also suggested by an enhancement of pull- and post-hoc analyses. To this regard, the growing

Role of funding source

Fondazione Banco di Sardegna” gently supported this study with a research grant.

Contributors

GM designed the study and wrote the protocol. GC and SP managed the literature searches and analyses. EN and GM undertook the statistical modeling and analyses, and GM wrote the first draft of the manuscript. All authors contributed to and have approved the final manuscript.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgments

We thank “Fondazione Banco di Sardegna” (2013-1522) that partially funded this study and Dott. Mario Palomba for technical support.

References (38)

  • E.J. Van den Oord et al.

    Factor structure and external validity of the PANSS revisited

    Schizophr. Res.

    (2006)
  • V. Villalta-Gil et al.

    Four symptom dimensions in outpatients with schizophrenia

    Compr. Psychiatry

    (2006)
  • R.S. Wallwork et al.

    Searching for a consensus five-factor model of the Positive and Negative Syndrome Scale for schizophrenia

    Schizophr. Res.

    (2012)
  • N.C. Andreasen et al.

    Negative vs positive schizophrenia. Definition and validation

    Arch. Gen. Psychiatry

    (1982)
  • N.C. Andreasen et al.

    Remission in schizophrenia: proposed criteria and rationale for consensus

    Am. J. Psychiatry

    (2005)
  • T.J. Crow

    Positive and negative schizophrenic symptoms and the role of dopamine

    Br. J. Psychiatry

    (1980)
  • G. de Girolamo

    WHO Studies of Schizophrenia: An Overview of the Results and Their Implications for an Understanding of the Disorder. Psychosocial Approaches to Deeply Disturbed Patients

  • S. Dollfus et al.

    Identifying subtypes of schizophrenia by cluster analyses

    Schizophr. Bull.

    (1996)
  • M. Harrow et al.

    Factors involved in outcome and recovery in schizophrenia patients not on antipsychotic medications: a 15-year multifollow-up study

    J. Nerv. Ment. Dis.

    (2007)
  • Cited by (20)

    • A schizophrenia-like behavioral trait in the SHR model: Applying confirmatory factor analysis as a new statistical tool

      2018, Progress in Neuro-Psychopharmacology and Biological Psychiatry
      Citation Excerpt :

      A Scale for Assessment of Negative Symptoms (SANS) (Sajatovic and Ramirez, 2012) was developed to be used with the SAPS, assessing five subdomains of negative symptoms (attention, anhedonia-asociality, avolition-apathy, alogia and affective flattening). Together, the positive and negative symptoms act as general indicators of psychopathological status in the disorder (Nicotra et al., 2015). In rodents, several tasks are used to infer various aspects of human schizophrenia (Barnes et al., 2014; Jones et al., 2011; Lipska and Weinberger, 2000; Mattei et al., 2015; Miyamoto and Nitta, 2014; Ratajczak et al., 2013); for instance, locomotor activity to assess positive symptoms, social interaction task and sucrose preference test to investigate negative symptoms, novel object recognition and fear conditioning for cognitive symptoms, and prepulse inhibition of startle (PPI) to study pre-attentional information processing (Table 1).

    • Amendment of traditional assessment measures for the negative symptoms of schizophrenia

      2018, European Psychiatry
      Citation Excerpt :

      The characteristics of the Positive and Negative Syndrome Scale (PANSS) [1] and the Scale for the Assessment of Negative Symptoms (SANS) [2] have been a matter of discussion over the last 20 years. For example, further PANSS-subscales including four, five, or six factors have been proposed with several studies underlining that five-factor models show an adequate reliability when tested in different subgroups of individuals with schizophrenia, confirming the suitability of this approach [3,4]. On the other hand, cross-sectional studies of the SANS identify three, four and five different symptom factors; and longitudinal research has replicated three factors [5].

    • Oxytocin effects in schizophrenia: Reconciling mixed findings and moving forward

      2017, Neuroscience and Biobehavioral Reviews
      Citation Excerpt :

      Furthermore, clinical assessment tools themselves may limit our ability to adequately capture the effects of oxytocin in schizophrenia. The widely-used PANSS, for instance, has several drawbacks including a 7-point rating scale that may not reflect clinically meaningful change (e.g., examiners may not actually be able to distinguish between auditory hallucinations that are a “4 – moderate” versus a “5 – moderate severe” in severity) (Levine et al., 2011), item redundancy (Lehoux et al., 2009), incomplete measurement of positive symptoms (Aboraya and Nasrallah, 2016), and reliance on a structured clinical interview (a requirement that many studies do not meet) (Nicotra et al., 2015). In addition, the PANSS may not capture the most salient elements of psychotic illness.

    View all citing articles on Scopus
    View full text