Association Between Social and Role Functioning and Transition to Psychosis in the North American Prodrome Longitudinal Study Consortium

Abstract

Background and Hypothesis

Functional deficits during adolescence have proved to be important risk factors for later psychosis at 2 to 3-year follow-up. The aim of the current study was to evaluate whether social and role functioning in clinical high-risk (CHR) youth is associated with transition to psychosis.

Study Design

Data from the North American Prodrome Longitudinal Studies (NAPLS 2, NAPLS 3) were combined with outcome data from the NAPLS Long-Term Follow-up study to assess the relationship between baseline social and role functioning and later transition to psychosis. Participants included 1474 CHR individuals from NAPLS 2 (n = 764) and NAPLS 3 (n = 710), of whom 198 (13%) are known to have transitioned to psychosis.

Study Results

Individuals who later transitioned to psychosis had more severe positive and negative symptoms and poorer social and role functioning at baseline compared to those who did not. Higher baseline social and role functioning was significantly associated with a reduced hazard of transition (HR = 0.864, P = .0013; HR = 0.898, P = .0004, respectively), whereas a decline in social and in role functioning was significantly associated with an increased hazard of transition (HR = 1.229, P = .0013; HR = 1.124, P = .002, respectively).

Conclusions

These results underscore the significance of social and role functioning in predicting subsequent transition to psychosis in CHR participants, suggesting an underlying vulnerability that is possibly independent of clinical symptoms. Early intervention that includes a focus on social and role functioning in high-risk youth may reduce the risk of future functional disability and perhaps a later psychotic episode.

Introduction

Poor social functioning is a key feature of schizophrenia and often proves more persistent and disabling than its positive symptoms.^1,2 Although functional disability is commonly linked with chronic schizophrenia, evidence has shown that these impairments emerge early in the disease.^3,4 In fact, functional disability, especially poor social functioning, is prevalent among individuals at clinical high-risk (CHR) for psychosis.⁵ To acknowledge the important work of Dr. Barbara Cornblatt in this domain of early functional disability, this paper aims to examine social and role functioning within the North American Prodrome Longitudinal Study (NAPLS) consortium.

Funded by the National Institute for Mental Health (NIMH), the NAPLS consortium began in 2004 under the leadership of Dr. Robert Heinssen. The original members Addington, Cadenhead, Cannon, Cornblatt, McGlashan, Perkins, Seidman, Tsuang, Walker, and Woods, representing eight sites, had independently conceived ongoing and/or completed studies with CHR youth. The objective was to retrospectively integrate these projects, and aggregate the descriptive, clinical, and functioning variables into a federated longitudinal database to address certain scientific questions that would later lead to new collaborative large-scale projects exploring the development of psychosis.⁶

NAPLS 2 was the first prospective longitudinal study collecting data from all eight sites with a 2-year follow up.⁷ NAPLS 3 followed as the second prospective longitudinal study building on NAPLS 2, adding investigators Bearden (University of California at Los Angeles [UCLA]) and Mathalon with UCSF as the 9th site.⁸ In 2021, the consortium was funded by NIMH for a longitudinal study, “Predictors and Moderators of Long-Term Outcome of Persons at Clinical High Risk for Psychosis (LTF).” The LTF project aims to locate participants from NAPLS 1-3 and affiliated studies from the NAPLS sites, 5-20 years after the initial involvement of CHR participants.⁹

Since NAPLS 1 data came from various projects, some variables were recoded to fit a standard measure.⁶ The assessment of functioning was inconsistent, ranging from lengthy interviews to self-reports to brief clinician-rated scales. Designed for individuals with schizophrenia, functioning measures then in use were not suitable for the CHR population, which includes younger individuals with varying levels of functioning. Thus, 2 new functional measures were created for NAPLS 1 to assess CHR individuals aged 12-25.¹⁰ These measures, Global Functioning: Social (GF:Social)¹¹ and Global Functioning: Role (GF:Role)¹² aimed to evaluate subtle levels of CHR functional difficulties with detailed anchors. Each scale is brief, easy to use, and provides a clinician rated single overall score similar to the Global Assessment of Functioning,^13,14 but focusing separately on social and role functioning without being confounded with psychiatric symptoms. Complete development and testing details are available in Cornblatt et al.,¹⁰ with a subsequent validation paper.¹⁵ Further details on the scales are presented in the Methods section below.

In addition to being the main functional measures in NAPLS 1-3, the GF:Social and Role are the most commonly used functioning measures in CHR research and are included in the NIMH-supported PhenX Early Psychosis Clinical Services Toolkit.¹⁶ They have been employed in multiple studies, demonstrating that CHR youth have significantly poorer functioning than age-matched healthy controls.^15,17-19 Notably, social functioning tends to remain poor over time with little improvement observed, even when there is an improvement in attenuated psychotic symptoms (APS).¹⁸ Social and role functioning has also been examined in relation to the transition to psychosis. In NAPLS 1, CHR individuals who developed psychosis exhibited significantly poorer social functioning than those who did not.²⁰ Furthermore, in NAPLS 1, it was shown that CHR participants who transitioned to psychosis were significantly more socially impaired compared to those who did not. This was not the case for role functioning, which, although impaired, was not associated with later transition to psychosis.⁵

A second NAPLS 2 paper by Carrión and colleagues²¹ reported that pre-baseline social decline was a predictor of poorer functional outcomes, as well as transition to psychosis. However, role functioning did not show the same predictor potential; instead it appeared to be more of an illness indicator that worsened over time.²¹ The next step in CHR research involved developing predictive models.²² The NAPLS group created an individualized risk calculator and pre-baseline decline in social functioning was a key variable in the model.²³ External validation of the NAPLS calculator²⁴ supported decline in social functioning as an important component of the model. In summary, the GF: Social and Role have been consistently utilized, highlighting the significance of social functioning in the outcome of CHR individuals.

The key aim of this article is to examine the association between functioning, as assessed by the GF: Social and Role, and transition to psychosis.

Methods

Participants

Participants for this paper were a pooled sample of 1474 CHR individuals from NAPLS 2 (n = 764) and NAPLS 3 (n = 710). NAPLS 2 and 3 were multisite longitudinal studies conducted at Emory University, Harvard University, University of North Carolina, Chapel Hill, Yale University, Zucker Hillside Hospital, University of Calgary, and University of California at Los Angeles (UCLA), at San Diego, and at San Francisco. Detailed information on recruitment and methods of NAPLS 2 and 3 is described in prior publications.^7,8

The inclusion criteria for NAPLS 2 and 3 required participants to be between 12 and 30 years old and to meet the Criteria of Psychosis-Risk Syndromes (COPS) based on the Structured Interview for Psychosis-Risk Syndromes (SIPS).²⁵ Exclusion criteria included meeting criteria for current or lifetime Axis I psychotic disorder, an IQ <70, or a history of a central nervous system disorder. Non-psychotic DSM-IV and -5 disorders were not exclusions provided they did not account for APS.

Centralized IRB approval was obtained to find, recontact and interview past CHR participants using a variety of methods for the NAPLS LTF study.⁹

Each site conducted extensive searches to find participants. To date, 452 participants from NAPLS 2 (n = 199) and NAPLS 3 (n = 253) have consented to follow-up and have been interviewed. In NAPLS 2 and 3 there were originally 166 CHR participants who transitioned to psychosis (95 from NAPLS 2, 71 from NAPLS 3). In the LTF study to date, 74 individuals who made the transition to psychosis have been interviewed. These include 42 of the original 166 individuals who transitioned in NAPLS 2 and 3 and another 32 individuals. Of those 32, 23 had transitioned after NAPLS 2 and 3 had ended, and 9 had transitioned to psychosis within the 2-year follow-up framework of NAPLS 2 and 3 but had dropped out prior to the time of transition. Therefore, the total transition rate of those we interviewed is 16% for the combined NAPLS 2 and 3 sample using LTF data.

Study Measures

The Structured Clinical Interview for DSM-IV and for DSM-5 was used to rule out psychotic diagnoses at entry to NAPLS 2 and 3, and to confirm psychotic diagnoses for those who made a transition to psychosis.^26,27 The SIPS was used to confirm NAPLS 2 and 3 participants met COPS criteria,²⁵ and in NAPLS 2, 3,f and LTF, the Scale of Psychosis-Risk Symptoms (SOPS) was used to assess APS and negative symptoms.²⁵ Details of the 5 positive and 6 negative symptoms of the SOPS are presented in Supplementary material.

Functioning was assessed using the GF:Social and Role scales.¹⁰ The GF:Social scale focuses on social functioning, including family relationships, with an emphasis on relationships and interactions with peers. The GF:Role scale evaluates school and work performance. Both are rated on a 10-point scale (10 = superior functioning, 1 = extreme dysfunction). Training and certification for these scales include a training video, manual, and reliability vignettes. Rater certification requires meeting reliability criteria, including alignment with gold standard scores and interclass correlations above 0.80 on vignette scores. The scales have been used in numerous CHR and early psychosis studies and have excellent interrater reliability and strong discriminant and convergent validity.^10,15 The GF scales have been translated into 8 languages demonstrating their international utility and applicability in CHR research—German, French, Italian, Spanish, Chinese, Korean, Japanese, and Danish—and have been included in the multisite international AMP-SCZ project.²⁸ The psychometric properties of the French and Italian versions have been published, demonstrating excellent interrater reliability and good to excellent concurrent validity.^29,30 In the United States, the GF scales are included in the Core Assessment Battery of the Early Psychosis Intervention Network (EPINET)³¹ and are used to monitor treatment outcomes of several thousand young people enrolled in Coordinated Specialty Care programs for first episode psychosis.³²

Procedures

Baseline measures for all NAPLS 2 and 3 participants were available from our databases. The data used in this paper included demographics, SOPS positive and negative symptoms, and GF: Social and Role. For participants in NAPLS 2 and 3 who transitioned to psychosis within 2 years of each study, the time from baseline to transition was documented.

Recruitment for participants for the LTF study utilized a methodical approach, targeting individuals who had previously consented to be contacted for future studies at the time of their initial participation. Multiple strategies were employed to establish contact with eligible participants. Research staff identified past participants who had explicitly indicated their preference not to be recontacted at the time of initial participation; these individuals were excluded from further outreach.⁹

Participants who had agreed to be re-contacted were approached via telephone and/or email, followed by letters or calls to other specified contacts. For all other past participants, that is those from whom permission to re-contact had not been sought at the end of NAPLS 2 and 3, letters were first sent. If there was no reply, research staff would call the participant, or anyone listed as an alternative contact. Upon establishing communication, the follow-up study would be explained. If appropriate, official sources such as hospital records, Department of Motor Vehicles (DMV), social security administration, and marriage and divorce records were consulted to obtain contact information for participants. As a final measure, social media and internet searches were conducted. Information regarding past participants was also gathered through relatives, as well as death and incarceration records. More specific details of the methods of the NAPLS LTF can be found in Cadenhead et al.⁹

LTF interviews were mostly conducted via secure teleconferencing due to the pandemic at the time the study started. Data collected at those interviews that were used for this article included the SIPS, to determine if the participant still met criteria for CHR or the criteria for conversion to psychosis had been met, the SOPS, to measure current positive and negative symptoms, and the GF: Social and GF: Role scales. All of these procedures and methods were approved by the Centralized IRB.

Statistical Analyses

Descriptive analyses to demonstrate differences between participants who transitioned to psychosis and those who did not, as well as baseline differences between participants who were interviewed and those who were not, were assessed using independent samples t-tests for continuous variables and chi-square tests for categorical variables.

Kaplan–Meier survival analysis was conducted to estimate the probability of remaining transition to psychosis-free over time (in days). Additionally, Cox proportional hazards models were used to examine the association between baseline Global Functioning: Social and Role scores and decline in Global: Social and Role with time to transition. These analyses were performed for the combined NAPLS 2 and 3 sample including all known transitions, that is including those observed in the LTF assessments. Due to the high likelihood of non-random missingness these analyses were then repeated with the transitions that were observed during the time frame of the original NAPLS 2 and 3 excluding the LTF participants.

All statistical analyses were conducted using the IBM Statistical Package for the Social Sciences (SPSS) version 25 and Statistical Analysis Software (SAS) version 9.4.

Results

Demographics

Table 1 presents demographics for the NAPLS 2 and 3 sample at baseline and for the LTF sample at the time of their interview.

Table 1.

Demographics of NAPLS 2 and 3 and Long-Term Follow-up Participants

	NAPLS 2 and 3    n = 1474    At baseline	LTF    n = 452    At interview
	Mean (SD)	Mean (SD)
Age	18.35 (4.14)	27.48 (5.37)
Years of education	11.36 (2.94)	14.80 (2.44)
	Frequency (%)	Frequency (%)
Sex
Male	821 (55.7)	213 (47.1)
Female	653 (44.3)	239 (52.9)
Racial a
Indigenous Groups	27 (1.8)	4 (0.9)
East Asian	61 (4.1)	16 (3.6)
Southeast Asian	26 (1.8)	5 (1.1)
South Asian	42 (2.8)	11 (2.4)
Black	200 (13.6)	62 (13.9)
Central/South American	75 (5.1)	28 (6.2)
West/Central Asian & Middle Eastern	14 (0.9)	2 (0.4)
White	829 (56.2)	263 (58.2)
Native Hawaiian or Pacific Islander	5 (0.3)	2 (0.4)
Multiracial	192 (13.0)	58 (12.9)
Maximum education achievedb
Some grade school	776 (52.6)	21 (4.7)
High school	549 (37.2)	195 (43.3)
Post-secondary education	133 (9.0)	181 (40.2)
Postgraduate	10 (0.7)	53 (11.8)

LTF—NAPLS 2 and 3 participants in the long-term follow-up study.

^aEast Asian includes: Chinese, Japanese, Korean; Southeast Asian includes: Cambodian, Indonesian, Vietnamese; South Asian includes: Indian, Pakistani, Sri Lankan; West/Central Asian and Middle Eastern includes: Egyptian, Lebanese, United Arab Emiratis, Afghan, Iranian.

^bGrade school includes: completion up to some grade school but not high school. Post-secondary education includes: College, Technical school, and University. Postgraduate includes: Graduate school and Professional school.

Differences Between Those in the LTF Study and Those Not Yet Found or Interviewed

Those who have been found to date and interviewed were compared to those who have not yet been found or interviewed across several key variables. A significantly higher proportion of females have been located and interviewed compared to males (χ = 18.44, P < .0001). Of the 821 males, 214 (26.1%) were interviewed, and of the 653 females, 238 (36.4%) were interviewed. There were no significant racial differences (χ = 9.76, P = .371) between those interviewed and those not. Comparisons on age, years of education, positive and negative symptoms, and social and role functioning at baseline are presented in Table 2. The only significant differences observed were that those who had been interviewed had significantly better ideational richness and less grandiosity. Those interviewed had higher ratings on current social and role functioning, and on the lowest social functioning in past year and highest role functioning in the past year. There was no difference in social decline.

Table 2.

Baseline Differences Between Participants Who Were Interviewed and Those Not Interviewed

	Interviewed    n = 452	Not interviewed    n = 1022	Test statistic	Significance value	Effect size
	Mean (SD)	Mean (SD)	t	P	Cohens d
Age	18.28 (4.09)	18.38 (4.16)	0.42	.674	0.024
Years of education	11.34 (2.93)	11.36 (2.94)	0.08	.937	0.005
Total SOPS positive	12.26 (3.74)	12.47 (3.62)	1.0	.319	0.056
Unusual Thought Content	3.55 (1.22)	3.53 (1.24)	−0.38	.702	−0.022
Suspiciousness	2.83 (1.49)	2.96 (1.47)	1.64	.100	0.093
Grandiose Ideas	0.86 (1.26)	1.01 (1.34)	2.01	.044	0.114
Perceptual Abnormalities	3.20 (1.40)	3.06 (1.47)	−1.74	.082	−0.098
Disorganized Communication	1.83 (1.52)	1.91 (1.46)	0.96	.339	0.054
Total SOPS negative	11.58 (6.15)	12.18 (6.22)	1.7	.089	0.097
Social Anhedonia	2.26 (1.67)	2.38 (1.73)	1.26	.209	0.071
Avolition	2.45 (1.61)	2.53 (1.63)	0.9	.367	0.051
Decreased Expression of Emotion	1.29 (1.41)	1.42 (1.49)	1.49	.135	0.085
Decreased Experience of Emotions and Self	1.87 (1.66)	1.87 (1.65)	−0.02	.981	−0.001
Decreased Ideational Richness	1.06 (1.24)	1.24 (1.33)	2.39	.017	0.136
Occupational Functioning	2.64 (2.03)	2.74 (1.97)	0.84	.401	0.048
GF: Social
Current	6.43 (1.54)	6.24 (1.55)	−2.24	.025	−0.127
Lowest in past year	6.11 (1.48)	5.91 (1.57)	−2.3	.022	−0.130
Highest in past year	7.07 (1.41)	6.95 (1.47)	−1.49	.135	−0.085
Decline in social functioning	0.64 (0.91)	0.72 (0.98)	1.35	.177	0.076
GF: Role
Current	6.27 (2.23)	5.99 (2.16)	−2.22	.027	−0.126
Lowest in past year	5.79 (2.26)	5.58 (2.18)	−1.7	.089	−0.096
Highest in past year	7.50 (1.54)	7.21 (1.74)	−3.25	.0012	−0.176
Decline in role functioning	1.24(1.71)	1.21(1.59)	−0.23	.819	−0.013

Abbreviations: GF: Role = Global Functioning: Role; GF: Social = Global Functioning: Social; SOPS = Scale of Psychosis-Risk Symptoms.

Differences Between Participants Who Transitioned to Psychosis and Those Who Did Not

The transition group consisted of any individual who had transitioned to psychosis either within the original NAPLS 2 and 3 projects or whose transition had been recorded through the LTF study. Differences between these 2 groups for age, education, race, SOPS symptoms, and social and role functioning are presented in Table 3. Significant differences, indicating more severe symptoms at baseline, for those who later transitioned to psychosis were observed for all positive symptoms except grandiose ideas and perceptual abnormalities, and for all negative symptoms except avolition and decreased experience of emotions and self. Those who transitioned had poorer social and role functioning, both currently and for the lowest score in the past year, and an increased decline in social functioning in the year prior to baseline.

Table 3.

Baseline Differences Between Those Who Transitioned to Psychosis and Those Who Did Not

	Transitioning group    n = 198	Non-transitioning    n = 1276	Test statistic	Significance value	Effect size
	Mean (SD)	Mean (SD)	t	P	Cohens d
Age	18.35 (3.88)	18.35 (4.19)	0.0	1.0	0.000
Years of education	11.26 (2.73)	11.37 (2.97)	0.47	.635	0.036
Total SOPS Positive	13.57 (3.84)	12.22 (3.60)	−4.83	<.0001	−0.37
Unusual Thought Content	3.98 (1.27)	3.46 (1.21)	−5.57	<.0001	−0.426
Suspiciousness	3.19 (1.51)	2.88 (1.47)	−2.76	.006	−0.211
Grandiose Ideas	1.04 (1.40)	0.95 (1.30)	−0.92	.356	−0.071
Perceptual Abnormalities	3.13 (1.59)	3.10 (1.43)	−0.23	.819	−0.019
Disorganized Communication	2.24 (1.64)	1.83 (1.45)	−3.28	.001	−0.274
Total SOPS Negative	13.30 (6.51)	11.78 (6.13)	−3.19	.002	−0.245
Social Anhedonia	2.66 (1.74)	2.29 (1.70)	−2.84	.005	−0.217
Avolition	2.70 (1.65)	2.48 (1.62)	−1.65	.099	−0.126
Decreased Expression of Emotion	1.62 (1.59)	1.34 (1.45)	−2.48	.013	−0.190
Decreased Experience of Emotions and Self	1.85 (1.69)	1.87 (1.65)	0.17	.866	0.013
Decreased Ideational Richness	1.48 (1.42)	1.14 (1.28)	−3.19	.002	−0.264
Occupational Functioning	3.03 (2.07)	2.66 (1.97)	−2.42	.016	−0.186
GF: Social
Current	5.98 (1.56)	6.35 (1.54)	3.05	.002	0.233
Lowest in past year	5.70 (1.54)	6.01 (1.54)	2.63	.009	0.201
Highest in past year	6.84 (1.45)	7.01 (1.45)	1.53	.126	0.117
Decline from highest to current	0.86 (1.04)	0.67 (0.95)	−2.6	.009	−0.199
GF: Role
Current	5.65 (2.17)	6.15 (2.18)	2.97	.003	0.227
Lowest in past year	5.27 (2.19)	5.70 (2.20)	2.58	.01	0.197
Highest in past year	7.14 (1.75)	7.32 (1.67)	1.46	.144	0.112

Abbreviations: GF: Role = Global Functioning: Role; GF: Social = Global Functioning: Social; SOPS = Scale of Psychosis-Risk Symptoms.

Survival Analysis

In Figure 1, for all known transitions, the Kaplan–Meier Survival Curve illustrates the cumulative probability of experiencing all known transitions over time, up to approximately 5000 days. The curve shows that the cumulative probability of transition increased rapidly during the early phase of follow-up and plateaued thereafter. By approximately 4000 days, the cumulative probability of transition approached 1.0, indicating that near all those who transitioned to psychosis had transitioned by that time.

Figure 1.

Kaplan–Meier Survival Curve (NAPLS 2 and 3, All Known Transitions Over Full Follow-up) for Transition Group.

In the combined NAPLS 2 and 3 sample using all known transitions, higher baseline GF: Social scores were significantly associated with a lower hazard of transition (HR = 0.864, 95% CI, 0.79-0.944, P = .0013), corresponding to a 13.6% lower hazard per one-unit increase in GF: Social. Conversely, decline in GF: Social was significantly associated with an increased hazard of transition (HR = 1.229, 95% CI, 1.084-1.393, P = .0013), indicating that each one-unit increase in decline was associated with a 22.9% higher hazard of transition.

Similarly, higher baseline GF: Role scores were significantly associated with a lower hazard of transition (HR = 0.898, 95% CI, 0.846-0.953, P = .0004), corresponding to an 18.6% lower hazard per one-unit increase. Decline in GF: Role was significantly associated with an increased hazard (HR = 1.124, 95% CI, 1.044-1.210, P = .002), with each one-unit increase in decline associated with a 12.4% higher hazard of transition.

To address the possibility of non-random missingness these analyses were re-run excluding the LTF participants. The results were as follows.

Higher baseline GF: Social scores were significantly associated with a lower hazard of transition (HR = 0.809, P = .0002), corresponding to a 19.1% lower hazard per one-unit increase in GF: Social. Decline in GF: Social was significantly associated with an increased hazard of transition (HR = 1.29, P = .0012); each one-unit increase in decline was associated with a 29% increased hazard of transition.

Similarly higher baseline GF: Role scores remained significantly associated with a lower hazard of transition (HR = 0.89, P = .003), corresponding to an 10.8% lower hazard per one-unit increase. GF: Role decline was significantly associated with an increased hazard of transition (HR = 1.14, P = .007), with each one-unit increase in decline associated with a 14.3% higher hazard of transition.

Discussion

This article examines the association of baseline social and role functioning and later transition to psychosis in 3 projects within the NAPLS consortium: NAPLS 2, NAPLS 3, and the LTF study. For the first time, social and role functioning has been assessed using the combined NAPLS 2 and 3 studies. Comparison between individuals who transitioned to psychosis with those who did not, revealed significant differences in both baseline positive and negative symptoms, as well as in current social and current role functioning.

Most transitions in NAPLS 2 and 3 occurred within the first year, with a steady decline over the 2-years of the study period. However, longer-term follow-up in the LTF study revealed that this decline continued to occur beyond 2 years, suggesting that even though transitions mostly occur earlier, there are individuals whose transition can occur many years after the initial assessment.

When considering transition to psychosis, regardless of whether they occurred within 2 years or later, social and role functioning as measured by the GF: Social and the GF: Role were both significantly associated with transition to psychosis. Moreover, a decline in both social and role functioning in the year prior to ascertainment were significantly related to later transition. The findings with respect to social functioning are consistent with previous research including the work of Dr. Cornblatt.^{5,20,21,23,33-36} For example, in a small sample in an early publication⁵ Cornblatt and colleagues first reported that even when controlling for neurocognition, those who develop psychosis had significantly poorer functioning than those who did not. This was later supported in the full NAPLS 2 sample.²¹ Furthermore, in this paper Carrion, Cornblatt and colleagues reported that amongst those who transitioned to psychosis, participants who had demonstrated a decline in functioning in the year prior to baseline had the poorest functioning at 12 months or post transition.²¹

In this current study, although social functioning has consistently been the stronger and more independent predictor when both NAPLS 2 and 3 samples were combined into a much larger sample, both role functioning and a decline in role functioning were significantly associated with later transition to psychosis. As noted earlier,¹⁷ although close, the GF: Role measure did not always demonstrate the same predictive value as social functioning. It is possible that role functioning may be more strongly influenced by environmental factors—such as changing schools, job loss, or low employment rates—than social functioning.¹ Cornblatt and colleagues suggested that role functioning is particularly responsive to these conditions and may represent a state factor linked to poor long-term functional outcome.³⁷ Thus, it is not surprising that role in addition to social functioning is associated with transition especially as they can interact, influencing each other over time.

Dr. Cornblatt introduced the neurodevelopmental RAP model, which links early decline in social functioning during adolescence to a biological vulnerability influenced by genetics and behavior, reflecting underlying brain pathology.³⁸ This model implies that abnormalities beyond APS may predict future psychosis.³⁷ Cornblatt also argued that social and role deficits are not merely consequences of increasing positive symptoms but represent enduring vulnerability traits evident in the early stages of illness.³⁸ Furthermore, there are several other possible explanations for why social dysfunction during adolescence and emerging adulthood may pose a heightened risk for adverse outcomes. This developmental period is marked by critical tasks such as forming and maintaining interpersonal relationships. Consequently, the onset of social dysfunction during this time can be particularly disruptive, potentially undermining the successful navigation of these essential developmental milestones. Importantly, early interventions targeting social and role functioning in high-risk youth could mitigate these vulnerabilities, reduce the likelihood of long-term disability and possibly prevent later psychotic episodes.³⁸

In summary, results of this current study underscore the continued significance of both social and role functioning, and a decline in social and role functioning, in predicting subsequent transition to psychosis in CHR participants in both short and long-term follow-up. Furthermore, these findings support the inclusion of the Global Functioning: Social and Role Scales in CHR assessment protocols and targeted psychosocial interventions.

Strengths and Limitations

Strengths of this study include examining the large sample of NAPLS 2 and NAPLS 3 combined as well as over 400 participants with longer-term follow up. There are limitations in this study. First, long-term outcomes could only be reported for 31% of the NAPLS 2 and 3 samples. To date, 452 individuals have been interviewed in the LTF study from NAPLS 2 and 3, with an additional 124 whose transition was confirmed in NAPLS 2 and 3 but not yet interviewed. Importantly, data from the 452 interviews indicated no differences in baseline age or education, or overall positive and negative symptoms. The interviewed group did, at baseline, have significantly better ratings on grandiosity and decreased ideational richness, however these differences were not clinically significant, and both were in the “questionably present” range. There were significant differences in baseline functioning with the interviewed group compared to those not interviewed having improved baseline social and role functioning. Again, these differences were not clinically significant in that differences were less than 1 point (i.e., differences of 0.19 and 0.27, respectively). It should be noted that the 2 groups did not differ in decline in social or role functioning. A notable observation was a higher number of females compared to males; possibly they were more willing to participate. Due to these differences in those interviewed versus those not interviewed and the number of participants who could not be interviewed, there is a high likelihood of non-random missingness.

However, to address the non-random missingness we repeated the Cox’s Hazard analysis with only the transitions that occurred during the original time frame of NAPLS 2 and 3 excluding the LTF participants. These findings were highly consistent with those from the full sample, suggesting that the original conclusions are robust.

Second, our transition rates are likely underestimated since we only found and interviewed 31% of the original sample.

Thirdly, since the purpose of the paper was to recognize the work of Dr. Cornblatt, we solely focused on the association of social and role functioning with transition to psychosis whereas other studies have explored additional predictive models.^22,39 Several studies improved prediction by incorporating cognition, negative symptoms as well as functioning.^37,40,41 For instance, the NAPLS calculator²³ included a range of variables in addition to a decline in functioning prior to baseline. A large meta-analysis⁴² classified both functioning and APS as highly suggestive as risk factors for transition. Overall, most prediction models identified social functioning as a key predictor.²³

Future Directions

Future research opportunities include investigation of the long-term trajectory (5+ years) of the functional deficits identified in youth at high-risk for psychosis available in the LTF study. Shorter term studies have confirmed the stability of functional deficits but the LTF longer-term follow-up will allow us to investigate subgroups with differing trajectories and outcomes as well as correlates of these outcomes.⁹ An examination of the correlates of functional outcome might shed some light onto factors that impact longer-term functional outcome that could possibly provide insight for interventions offered to high-risk individuals.

Conclusion

In conclusion, we want to further acknowledge the work of Dr. Barbara Cornblatt not only in her early genetic high-risk work but her insights in the emerging study of the psychosis prodrome. Her persistence in highlighting the core neurodevelopmental deficits manifesting as functional difficulties in adolescence, a key developmental period that shapes an individual’s life trajectory, will undoubtedly make a lasting impact on the early psychosis field.

Funding

NAPLS 2 was supported by the National Institute of Mental Health (grant U01 MH081984 to J.A.; grants U01 MH081928, P50 MH080272, and Commonwealth of Massachusetts SCDMH82101008006 to Dr. Seidman; grants R01MH60720, U01MH082022, and K24 MH76191 to K.S.C.; grant U01 MH081902 to T.D.C.; P50 MH066286 (Prodromal Core), the Staglin Family Music Festival for Mental Health, the Shear Family Foundation and the Don S. Levin Trust to C.E.B.; grant U01 MH082004 to D.P.; grant U01 MH081988 to E.F.W.; grant U01 MH082022 to S.W.W.; and U01 MH081857-05 grant to Dr. Cornblatt). NAPLS 3 was supported by the National Institute of Mental Health (grant U01MH081984 to J.A.; grant U01MH081928 to W.S.S.; grant U01MH081944 to Dr. Cadenhead; grant U01MH081902 to T.D.C. and Bearden; grant U01MH082004 to D.O.P.; grant U01MH081988 to E.F.W.; grant U01MH082022 to S.W.W.; grant U01MH076989 to D.H.M.; grant UO1MH081857 to Dr. Cornblatt). NAPLS LTF was supported by the National Institute of Mental Health (grants R01MH123641, and U01MH081944 to K.S.C.; grant U01MH081984 to J.A.; grant U01MH081928 to W.S.S. and M.K.; grant U01MH081902 to T.D.C. and C.E.B.; grant U01MH082004 to DO.P.; grant U01MH081988 to E.F.W.; grant U01MH082022 to S.W.W.; grant U01MH076989 to D.H.M.; grant UO1MH081857 to Dr. Cornblatt).

Conflicts of Interest

None declared.