Scale This study (ICC) Alphs et al. (ICC)

NSA-4 0.94 0.82

NSA-16 0.97 0.87

Scale This study (α) Alphs et al. (α)

NSA-4 0.65 0.64

NSA-16 0.87 0.85

Variables This study Alphs et al.

NSA-16 0.86 0.85

NSA global rating 0.67 0.68

PANSS negative subscale 0.73 0.52

PANSS negative symptoms Marder factor 0.73 0.57

Variables This study Alphs et al.

PANSS Marder factor anxiety/depression -0.11 -0.03

PANSS Marder factor disorganized thought 0.29 0.42

PANSS Marder factor hostility/excitement 0.03 0.06

PANSS positive symptoms 0.13 0.23

PANSS

Neg

5 15

20

25

30

35

40

5 10 15 20NSA4

•

•

•

• •

• •

• •

• •

•

••••

••••

••

••••

••••

•• •

•

••••

••••

•••••

••

••••

••••

•••••

• •

•

••••

••••

•••••

••••

•

•••

••••

•••••

••••

••

•

• •

••••

•••••

••••

••••

•••

•••••

••••

••••

•• •

•••

•••••

••••

••••

•••

•

•

••••

•••••

••••

••••

•••

••

••••

•••••

••••

••••

•• •

•

••

••••

•••••

••••

••••

•

••

••••

•••••

••••

••••

•

•••

•••••

••••

••••

•

•

•• •

•••••

••••

•••

••

••••

••••

•• •

• •

•

••

• ••

••

PANSS

Anx

iety

5

10

15

2

0

5 10 15 20NSA4

•

•

•

•

•••

•••

• ••

•••

•

••••

•••

• •

••

••••

••••

••••

••••

••••

••••

••••

••••

••••

••••

••••

••••

••••

••••

••••

•••

••••

••••

••••

• •

••••

••••

••••

••••

••••

••••

••••

•••

••••

••••

••••

••••

••••

••••

••••

•• •

••••

••••

••••

•• •

••••

••••

••••

•••

••••

••••

••••

••••

••••

••••

••••

•

Get the free mobile app athttp:/ /gettag.mobi

RESEARCH INSTITUTE©2014 MedAvante Inc.

How Does the NSA-4 Compare to the NSA-16?

Williams, JBW1,2, Garzio, LM1, Osman, D1, Popp, D1

1MedAvante Inc.; 2Columbia University Depts. of Psychiatry and Neurology

RESULTS

REFERENCES

Alphs L, Morlock R, Coon C, Cazorla P, Szegedi A, & Panagides J. (2011). Validation of a 4-item Negative Symptom Assessment (NSA-4): a short, practical clinical tool for the assessment of negative symptoms in schizophrenia. Intl J Methods Psychiatr Res, 20(2),e31-e37. Alphs LD, Summerfelt A, Lann H, & Muller RJ. (1989). The negative symptom assessment: a new instrument to assess negative symptoms of schizophrenia. Psychopharmacol Bull, 25, 159-163. Fleiss, JL (1981). Statistical Methods for Rates and Proportions. 2nd ed. New York, NY: Wiley. George, D & Mallery, P (2003). SPSS for Windows step by step: A simple guide and reference. 11.0 update (4th ed.). Boston: Allyn & Bacon, p. 231. Williams JBW, Popp D, Osman DA, Cohen EA, & Detke MJ. The evaluation of negative symptoms by videoconferencing in a clinical trial. Poster presented at International Society for CNS Clinical Trials and Methodology, Eighth Annual Scientific Meeting, Washington, DC, February 2012.

The NSA-4 had very good overall agreement with the NSA-16, in the hands of highly trained and calibrated Central Raters, and even higher convergent and divergent validity and interrater reliability than demonstrated by Alphs et al (2011). Overall, results were very similar to those obtained by Alphs et al (2011).

The PANSS and NSA-16 in this study were not administered independently of one another, so the usefulness of the NSA-4 alone can only be evaluated in the context of its pairing with the PANSS. Further, scores on the NSA-4 were derived using the four items that were administered as part of the overall NSA-16 scale. While we examined the 4 items separately, their scores may have been informed by responses on the remaining 12 items.

A briefer but valid version of the NSA-16 would be useful in clinical and research work. Future research comparing independent administration of the NSA-4 to the NSA-16 is needed.

CONCLUSIONS

To be useful, a scale must have convergent validity (high correlation with other scales that measure the same construct) and divergent validity (relatively low correlations with other scales that measure different constructs).

The NSA-4 in this study had correlation coefficients of 0.67 or greater with the full scale (NSA-16), the global rating, the PANSS negative subscale, and the PANSS negative symptoms Marder factor. (Table 1)

Figure 1. Correlation between NSA-4 and PANSS negative subscale

In this study, the NSA-4 scale was more highly correlated with PANSS negative subscales than was found by Alphs et al. (2011) (Table 1), indicating even better convergent validity than was found previously.

Table 2. Divergent Validity – Subscale Correlations with NSA-4

Figure 2. Correlation between NSA-4 and PANSS Marder factor anxiety/depression

The NSA-4 is negatively correlated with the PANSS Marder factor anxiety/depression (r = -0.11), and poorly correlated with the PANSS Marder factors disorganized thought and

hostility/excitement (r = 0.29 and 0.03,respectively). In addition, the NSA-4 is correlated poorly with the PANSS positive symptoms (r = 0.13). (Table 2) These findings show even better divergent validity than found by Alphs et al. (2011).

Table 1. Convergent Validity – Subscale Correlations with NSA-4

Internal Consistency Reliability Cronbach’s alpha is an indicator of scale reliability in that it examines the correlations among different items in a scale to see to what extent they measure the same construct. Cronbach’s alpha ranges from 0-1.0, with 0.7-0.8 indicating acceptable to good internal consistency (George & Mallery, 2003).

Table 3. Internal Consistency (measured by Cronbach’s Alpha, α) of the NSA-4 and NSA-16

Table 4. Intraclass Correlation Coefficients (ICC) of the NSA-4 and NSA-16

Cronbach’s alphas for the NSA-4 and the NSA-16 were remarkably similar for both studies. The NSA-16 showed excellent internal consistency in this study (α = 0.87) and in the previous study (α = 0.85). The NSA-4, as expected for a shorter scale, had lower internal consistency reliability in both data sets (α = 0.65 and 0.64). (Table 3)

Interrater Reliability One measure of the reliability of a scale is the degree to which different raters can assign similar scores to the same phenomena using the same scale, as measured by the Intraclass Correlation Coefficient (ICC). ICCs range from 0-1.0, with 0.75 and above being excellent (Fleiss, 1981).

In this study, the ICC = 0.94 for the NSA-4 and 0.97 for the NSA-16. In Alphs et al. (2011), the ICCs were slightly lower (0.82 and 0.87, respectively), but still good/excellent.

Negative Symptom Assessment (NSA-16) with NSA-4 items highlightedINTRODUCTION

• The Negative Symptom Assessment (NSA-16) is increasingly used as a validated measure to track response to treatment of negative symptoms in clinical trials of schizophrenia.

• NSA-16 takes up to a half hour to administer. As clinical trials have become more complex, a briefer assessment tool would be useful.

• The NSA-4 is proposed as a reliable and valid brief alternative to the NSA-16 (Alphs et al., 2011).

• Four NSA-16 items are included: restricted speech quantity, reduced emotion, reduced social drive, and reduced interests, as well as an overall global rating of negative symptoms.

• The goal of this study was to replicate the finding by Alphs et al. (2011) that the NSA-4 is an effective brief alternative to the longer NSA-16.

METHODS

• Data collected from two Phase 2 randomized double-blind studies of subjects with schizophrenia with prominent negative symptoms (n=483).

• All subjects were interviewed using live two-way videoconferencing at screen, baseline, and 11 more visits, including endpoint. This method has been demon- strated to be valid and reliable (Williams et al., 2012).

• Subjects were assessed by blinded independent Central Raters with ongoing calibration (n=29).

• At each visit, administration of the PANSS was immediately followed by the NSA-16 (n=2804 interviews).

• Correlation coefficients (r) were calculated for the NSA-4 and NSA-16 with each subscale of interest. (Tables 1 and 2)

• Internal consistency of the NSA-4 and NSA-16 were calculated using Cronbach’s alpha (α).

• Interrater reliability (determined by Intraclass Correlation Coefficient, ICC) was calculated between clinicians’ and observing trainers’ scores of the same subjects (n = 152).

1. Prolonged time to respond

2. Restricted speech quantity 1. Normal speech quantity 2. Minimal reduction in quantity; may be extreme side of normal 3. Speech quantity is reduced, but more obtained with minimal prodding 4. Flow of speech is maintained only by regularly prodding 5. Responses usually limited to a few words, and/or detail is only obtained by prodding or bribing 6. Responses usually nonverbal or limited to 1 or 2 words despite efforts to elicit more 9. Not ratable

3. Impoverished speech content

4. Inarticulate speech

5. Emotion: Reduced range (specify time frame for this assessment) 1. Normal range of emotion 2. Minimal reduction in range; may be extreme side of normal 3. Range seems restricted relative to a normal person, but during the specified time period subject convincingly reports at least four emotions 4. Subject convincingly identifies two or three emotional experiences 5. Subject can convincingly identify only one emotional experience 6. Subject reports little or no emotional range 9. Not ratable

6. Affect: Reduced modulation of intensity

7. Affect: Reduced display on demand

8. Reduced social drive 1. Normal social drive 2. Minimal reduction in social drive; may be extreme side of normal 3. Desire for social interactions seems somewhat reduced 4. Obvious reduction in desire to initiate social contacts, but a number of social contacts are initiated each week 5. Marked reduction in desire to initiate social contacts, but a few contacts are maintained at subject’s initiation (as with family) 6. No desire to initiate any social interactions 9. Not ratable

9. Poor rapport with interviewer

10. Sexual interest

11. Poor grooming and hygiene

12. Reduced sense of purpose

13. Reduced interests 1. Normal interests 2. Minimal reduction in interests; may be extreme side of normal 3. Range of interests and/or commitment to them seems diminished 4. Range of interests is clearly diminished and subject is not particularly committed to interests held 5. Only one or two interests reported, and these pursued superficially 6. Little or nothing stimulates interest 9. Not ratable

14. Reduced daily activity

15. Reduced expressive gestures

16. Slowed movements

GLOBAL NEGATIVE SYMPTOMS RATING

1. No evidence of this symptom 2. Minimal evidence of this symptom 3. Mild evidence of this symptom 4. Moderate evidence of this symptom; apparent to the casual observer 5. Marked evidence of this symptom; readily apparent to casual observer 6. Severe; not only obvious but has marked impact on functioning 7. Extremely severe symptom; it is incapacitating for subject