Clinical Outcome Assessment Instruments in Schizophrenia: A Scoping Literature Review with a Focus on the Potential of Patient-reported Outcomes

Abstract

Objective

The complexity inherent in the treatment of schizophrenia results in a multitude of outcome assessments being employed when conducting clinical trials. Subjective outcome assessments and minimal clinically important differences (MCIDs) to evaluate clinical meaningfulness have gained traction; however, the extent of application in evaluation of treatments for schizophrenia is unknown. A scoping review was conducted to assess the availability of published psychometric evaluations, including MCIDs, for clinical outcome assessments used to evaluate treatments for schizophrenia.

Method of Research

Key databases (PubMed®, Embase®, APA PsycINFO®, International Society for Pharmacoeconomics and Outcomes Research) were searched for studies on schizophrenia published from 2010 to 2020. Secondary sources (ClinicalTrials.gov, PROLABELS™, FDA.gov) were also reviewed. Clinical outcome assessments were organized by type (patient-reported outcomes [PROs], clinician-reported outcomes [ClinROs], observer-reported outcomes [ObsROs]) and further classified by intended use (generic, mental health, schizophrenia). Reliability and internal consistency were evaluated using Cronbach’s α. External validity was evaluated by intraclass correlation coefficient (ICC).

Results

Across 140 studies, 66 clinical outcome assessments were identified. MCIDs were reported for eight of the 66 studies. Of these, two were PROs (generic) and six were ClinROs/ObsROs (three mental health-specific, three schizophrenia-specific). Reliability was good across generic, mental health-specific, and schizophrenia-specific categories, whereas external validity was strong mainly for schizophrenia-specific PROs. Overall, ClinROs/ObsROs that focused on mental health had good reliability and strong external validity.

Conclusion

This review provides a comprehensive overview of the clinical outcome assessments used in schizophrenia research during the past ten years. Results highlight the heterogeneity of existing outcomes and a growing interest in PROs for schizophrenia.

The current state of the management of schizophrenia typically involves a combination of pharmacologic and psychosocial interventions;¹ however, because of the complexity of the disease, establishing the effectiveness of any treatment is challenging.² Traditional approaches have focused on measuring psychopathology, including positive symptoms (e.g., delusions, hallucinations, and disorganized speech or behavior) and negative symptoms (e.g., affective flattening, alogia, and avolition), as well as quantifying cognitive impairment.^3,4 Other outcomes based on assessment of functioning are also of great interest. However, in contrast to studies on interventions in mood disorders, definitions of response and remission in studies on schizophrenia lack universally accepted or otherwise simple-to-understand definitions; this lack of consensus complicates the process of interpreting the clinical relevance of study results.^5,6 Thus, the implementation of standardized, validated clinical outcome assessments can allow researchers and clinicians to make sound inferences regarding the effectiveness of interventions,² provided that what is meaningful or not can be determined.

Clinical outcome assessments can be broadly divided into clinician- (ClinROs), observer- (ObsROs), or patient-reported outcomes (PROs). There is a rich tradition of utilizing clinician and observer outcomes; however, the self-assessment of patients/subjects can provide different and actionable information. A PRO is “any report of the status of a patient’s health condition that comes directly from the patient, without interpretation of the patient’s response by a clinician or anyone else.”^7,8 The potential value of PROs over ClinROs has become more evident relatively recently. The United States (US) Food and Drug Administration (FDA) recommends the use of PROs in pivotal clinical trials.⁹ Although patients with schizophrenia and their assessments of disease and treatment outcomes may be impacted by anosognosia,¹⁰ PROs are still seen as essential to understanding patient preferences for treatments and facilitating comparisons between different therapies.^7,8

The inclusion of PROs in clinical trials is now advised and supported by various guidelines, including some that focus on schizophrenia.^11–13 However, selecting a meaningful assessment that reflects the breadth of experiences and symptoms that patients with schizophrenia have and the impact of treatments may be challenging.^11–13 A potential solution to this obstacle would be the selection of a PRO with a reported minimal clinically important difference (MCID), which is defined as the minimal amount of change in an outcome assessment that would be important to a patient.¹⁴ MCIDs can help guide both clinicians in the assessment of treatment effectiveness and researchers in trial design by providing a rationale for the selection of the most suitable outcome assessment.

In this scoping review, a database and literature search was performed to identify existing outcome assessments used in clinical studies, assess the availability of MCIDs for these assessments used to evaluate patients with schizophrenia, and propose recommendations for furthering MCID research.

METHODS

Scoping review design. This scoping review was performed in accordance with the methodology outlined by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for scoping reviews (PRISMA ScR) checklist and the Joanna Briggs Institute guidelines for performing scoping reviews.^15,16 The key objective of this review was to identify the existing and relevant clinical outcomes for the assessment of schizophrenia, rather than to assess the quality of studies utilizing these outcomes.

Literature search strategy. Four key databases (PubMed®, Embase®, APA PsycINFO®, and International Society for Pharmacoeconomics and Outcomes Research [ISPOR; through Value in Health]) were used for this scoping review to assess a large variety of studies of interest. Searches were conducted between April 24, 2020, and April 27, 2020, to reduce the risk of missing newly published studies. EndNote™ X7 was used to import and screen the articles.

Initially, a search utilizing keywords and medical subject headings (MeSH) related to schizophrenia, clinical outcome assessments, and quality of life (QoL) was performed using the PubMed® database. Because Embase® and APA PsycINFO® do not use MeSH terms, keyword-only searches were performed for these databases; where appropriate, wildcard searches were performed to complement keyword-only searches. The ISPOR database does not permit wildcard searches; therefore, a single-word search (“schizophrenia”) was conducted on article titles, abstracts, and keywords. To verify the robustness of this search, an additional search was performed using alternative terms (“schizoaffective”), which yielded results already included in the first search. A summary of the search terms used can be found in Supplementary Table 1.

SUPPLEMENTARY TABLE 1.

Summary of search strategies for key databases

KEYWORD	FINAL SEARCH STRATEGYa
Schizophrenia	Schizophrenia[Title/Abstract] OR “Schizophrenia”[Majr]
Outcome assessment	“Outcome Assessment, Health care”[MeSH:NoExp]
Quality of life	“Quality of Life”[MeSH]
Patient-reported outcome measures	“Patient Reported Outcome Measures”[MeSH:NoExp] OR (“patient-reported outcomes”[Title/Abstract] OR “patient reported outcomes” [Title/Abstract])

Note: 4 key databases were selected: PubMed, Embase, PsycINFO, and International Society for Pharmacoeconomics and Outcomes Research (ISPOR; through Value in Health). The search strategy was first established and refined for PubMed, and then adapted to the other databases. Where possible, searches were limited to English language only studies published in or after 2010.

^aMeSH terms were only used in PubMed search.

NoExp: turn off explosion of MeSH headings; Majr: to search a MeSH heading that is a major topic of an article;

MeSH: medical search headings

Studies were included in this analysis if they met the following criteria: published during or after 2010; included patients with existing schizophrenia or schizoaffective disorder diagnosis; and reported on or used a patient-level outcome (patient-reported, observer/clinician-reported) for measuring disease progression, severity, and/or treatment effectiveness in schizophrenia. Any outcomes published only prior to 2010 were not considered for the purposes of this review.

An identified study was excluded from this analysis if the study included patients without schizophrenia or other relevant diagnosis; reported the outcomes of caregivers, partners, or relatives of patients with schizophrenia, but not patient outcomes; reported outcomes in “natural units” (e.g., biologic- or chemical-level, imaging results, life expectancy, hospitalization rate) rather than from an assessment tool or scale; measured the effects of secondary health conditions (e.g., smoking cessation, treatment-related diabetes, or high blood pressure); solely assessed medication adherence; reported utilities (e.g., for calculating quality adjusted life-years) from secondary data; evaluated costs or economic burden/cost minimization; was written as an editorial on clinical outcome assessments in schizophrenia; or not published in English.

Additional secondary sources were reviewed to support the primary literature search. ClinicalTrials.gov was examined for information concerning ongoing or completed schizophrenia-related, interventional, Phase III trials to identify clinical outcome assessments not already identified by the primary literature search. PROLABELS™ (through ePROVIDE™) was searched for the labels of medications approved within the last ten years (2010–2020) by the FDA and European Medicines Agency (EMA) for the treatment of schizophrenia. The identified labels were searched for additional clinical outcome assessments that could be included in this review. The FDA’s website (FDA.gov) was also reviewed for schizophrenia-related submissions to the FDA Clinical Outcome Assessment (COA) Qualification Program. Studies that were known to investigators prior to the literature search were included in this scoping review if they met the inclusion and exclusion criteria.

Data synthesis plan. Clinical outcome assessments that were identified in the scoping review were categorized according to three factors. Each category was divided into various subcategories to appropriately describe the nature of each outcome assessment. The three factors were: 1) who reported the outcome assessment (type), divided into ClinROs, ObsROs, and PROs; 2) focus or intended use of the outcome assessments (focus), divided into generic, mental health, and schizophrenia; and 3) domain of the outcome (domain), divided into QoL/health-related QoL (HRQoL), treatment-related, emotional/psychological wellbeing, symptomatic, cognition, and need for care (Figure 1).

FIGURE 1.

Categorization of clinical outcome assessments.

^aGeneric focus includes the application of assessments related to any condition (e.g., general wellbeing).

^bDomain is what the outcome aims to assess and includes the following: QoL (social and occupational functioning; e.g., Heinrichs–Carpenter Quality of Life Scale, Manchester Short Assessment of Quality of Life), HRQoL (not specific to any health condition; e.g., EuroQol-5 Dimensions, World Health Organization Disability Assessment Schedule), treatment-related (preference and/or perception of treatment; e.g., Clinical Global Impressions Scale-Improvement, Reasons for Antipsychotic Discontinuation and Continuation), emotional/psychological wellbeing (useful in measuring severity and treatment effectiveness; e.g., Brief Psychiatric Rating Scale, Montgomery–Åsberg Depression Rating Scale), symptomatic (evaluates specific symptoms of disease; e.g., Clinical Global Impressions Scale-Severity, Positive and Negative Syndrome Scale), and need for care (assesses costs and services; e.g., Client Sociodemographic and Service Receipt Inventory).

The psychometric properties for all identified clinical outcome assessments were identified to the extent that the literature allowed. The original development and original validation or schizophrenia-specific validation studies were located for each identified outcome assessment, if available. Where possible, the reliability and validity of each outcome assessment, as well as a brief interpretation of each scale or scoring system, were extracted. The performances of outcome assessments were evaluated according to the Consensus‐Based Standards for the Selection of Health Measurement Instruments (COSMIN) guidelines.¹⁷ Internal consistency (reliability) was evaluated by the Cronbach’s α method. Reliability was determined to be good if α was 0.8 or greater (+++), acceptable if α ranged from 0.6 to 0.79 (++), and poor or unacceptable if α was less than 0.6 (+/-). From each identified study, the reported intraclass correlation coefficients (ICCs) were assessed for each outcome assessment to determine the external validity. External validity was considered to be strong if ICC was 0.5 or greater (+++), moderate if ICC was 0.25 to 0.49 (++), and weak if ICC was less than 0.25 (+). Furthermore, information on MCIDs was extracted when possible, making note of the treatment modalities.

RESULTS

Literature search. Overall, 2,908 studies were identified by searching the key databases, and 17 studies of interest were identified through known sources prior to the primary literature search (Figure 2). After removing duplicate studies, the titles and abstracts of the remaining 2,329 studies were screened against the inclusion and exclusion criteria. The initial screening yielded 256 studies, which then underwent a full-text review for eligibility. Of those 256 studies, 140 studies were included in this scoping review.^11,18–156 Our search of ClinicalTrials.gov identified 294 ongoing or completed randomized controlled trials (RCTs) on schizophrenia with relevant information concerning clinical outcome assessments. A total of 15 EMA-/FDA-approved drugs and the related clinical outcome assessments were extracted from the PROLABELS™ database (Supplementary Table 2). Notably, none of the antipsychotic agents approved between 2010 and 2020 included a mention of PRO endpoints in their respective labels, and all mentioned were ClinROs/ObsROs. Additionally, two outcome assessments (Virtual Reality Functional Capacity Assessment Tool [VRFCAT] and Epidemiological Study of Cognitive Impairment in Schizophrenia [EPICOG]) were identified from schizophrenia-related submissions to the COA Qualification Submissions on the FDA’s website.

FIGURE 2.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram for the scoping review process

SUPPLEMENTARY TABLE 2.

Clinical outcome assessments in approved antipsychotic agents (2010–2020)

DRUG, GENERIC NAME (BRAND NAME)	APPROVAL DATE	SPONSOR	AGENCY	COA IN LABELa
Lumateperone (Caplyta)	20-Dec-19	Intra-Cellular Therapies, Inc. (NY, US)	FDA	PANSS
Asenapine (Secuado)	11-Oct-19	Hisamitsu Pharmaceutical Co., Inc. (FL, US)	FDA	PANSS, CGI-I
Brexpiprazole (Rexulti)	26-Jul-18	Otsuka Pharmaceutical Netherlands B.V.	EMA	PANSS, CGI-S
Aripiprazole (Abilify MyCite)	13-Nov-17	Otsuka Pharmaceutical Company, Ltd. (NJ, US)	FDA	PANSS, CGI-S
Cariprazine (Reagila)	13-Jul-17	Gedeon Richter, Ltd. (Hungary)	EMA	PANSS, CGI-S
Aripiprazole lauroxil (Aristada)	5-Oct-15	Alkermes, Inc. (MA, US)	FDA	PANSS, CGI-I
Cariprazine (Vraylar)	17-Sep-15	Allergan Sales, LLC (NJ, US)	FDA	PANSS, CGI-S
Brexpiprazole (Rexulti)	10-Jul-15	Otsuka Pharmaceutical Company, Ltd. (MD, US)	FDA	PANSS
Paliperidone (Invega Trinza)	18-May-15	Janssen Pharmaceuticals, Inc. (NJ, US)	FDA	PANSS
Paliperidone (Trevicta)	9-Dec-14	Janssen-Cilag International NV (Belgium)	EMA	PANSS, PSP
Lurasidone hydrochloride (Latuda)	21-Mar-14	Aziende Chimiche Riunite Angelini Francesco S.p.A. (Italy)	EMA	PANSS, CGI-S
Aripiprazole (Abilify Maintena)	15-Nov-13	Otsuka Pharmaceutical Netherlands B.V.	EMA	Relapse,b PANSS, PSP
Aripiprazole (Abilify Maintena)	28-Feb-13	Otsuka Pharmaceutical Company, Ltd. (MD, US)	FDA	PANSS, CGI-I, Relapse
Paliperidone (Xeplion)	4-Mar-11	Janssen-Cilag International NV (Belgium)	EMA	PANSS, PSP
Lurasidone hydrochloride (Latuda)	28-Oct-10	Sunovion Pharmaceuticals, Inc. (MA, US)	FDA	PANSS, BPRSd, CGI-S

^aBy necessity, COAs contained in labels are generally restricted to the primary and key secondary outcomes; additional secondary and exploratory outcomes are usually not mentioned in labels but can be found in regulatory documents (e.g., drug approval packages) and in the literature when eventually published.

^bCOA of interest not included in label.

BPRSd: Brief Psychiatric Rating Scale derived from the PANSS; CGI-I: Clinical Global Impressions Scale-Improvement; CGI-S: Clinical Global Impressions Scale-Severity; COA: clinical outcome assessment; EMA: European Medicines Agency; FDA: United States Food and Drug Administration; PANSS: Positive and Negative Syndrome Scale; PSP: Personal and Social Performance Scale

Summary of clinical outcome assessments. A total of 66 clinical outcome assessments were identified from the included studies and secondary sources (Table 1). Of these outcome assessments, 26 were ClinROs/ObsROs and 40 were PROs. Among the 26 ClinROs/ObsROs, there were 11 mental health-specific and 15 schizophrenia-specific outcome assessments; no generic ClinROs/ObsROs were reported in the included literature. Among the 40 PROs, there were 11 generic, 15 mental health-specific, and 14 schizophrenia-specific outcome assessments. Of these generic and schizophrenia-specific outcome assessments, select measures specific to cognition (n=5) were added to supplement and strengthen the cognition domain representation.

TABLE 1.

Classification of clinical outcome assessments identified in the scoping review

DOMAIN	CLINICAL OUTCOME ASSESSMENTS (N=66)
	ClinROs/ObsROs (n=26)			PROs (n=40)
	GENERAL (n=0)	MH-S (n=11)	SCH-S (n=15)	GENERAL (n=11)	MH-S (n=15)	SCH-S (n=14)
QoL/HRQoL	0	3	0	9	4	8
Treatment-related	0	3	0	0	4	1
Emotional/psychological wellbeing	0	2	2	0	6	1
Symptomatic	0	3	10	0	0	1
Cognition	0	0	3	2	0	3
Need for care	0	0	0	0	1	0

ClinRO: clinician-reported outcome; HRQoL: health-related quality of life; MH-S: mental health-specific; ObsRO: observer-reported outcome; PRO: patient-reported outcome; QoL: quality of life; SCH-S: schizophrenia-specific

ClinROs/ObsROs. Nearly half (n=11) of the ClinROs/ObsROs were mental health-specific (Table 2). These 11 outcome assessments were distributed between the QoL/HRQoL (n=3), treatment-related (n=3), symptomatic (n=3), and emotional/psychological wellbeing (n=2) domains. Overall, the mental health-specific outcome assessments identified here had relatively high reliability and strong validity, compared to other outcome assessments.^{56,57,102,157–168} The remaining 15 ClinROs/ObsROs were schizophrenia-specific (Table 3). Of these 15 outcome assessments, 10 focused on symptoms, while three focused on cognition, and two focused on emotional/psychological wellbeing. The reliability and external validity of the identified schizophrenia-specific assessments varied considerably.^{23,62,67,76,86,145,146,161,169–175} Certain outcome assessments, such as the Positive and Negative Syndrome Scale (PANSS), had good reliability and strong validity, while newly developed outcome assessments generally demonstrated lower reliability and validity. Reliability and external validity were not available for two outcome assessments.^27,36

TABLE 2.

Mental health-specific ClinROs/ObsROs

CLINICAL OUTCOME ASSESSMENT	CLINICAL OUTCOME ASSESSMENT REFERENCES	DOMAIN	INTERPRETATION	LANGUAGES	RELIABILITY	VALIDITY	ORIGINAL VALIDATION
Brief Psychiatric Rating Scale (BPRS)	Bech et al (2018),30 Fefeu et al (2018),55 Garcia-Portilla et al (2015),62 Masand et al (2011),99 Tachibana et al (2016),135 Yee et al (2017)144	Emotional/psychological wellbeing	18 items rated 1–7; score range: 18–126; higher score=greater severity	English, Malay, others	++ ICC=0.81 ++ Cronbach’s α: 0.64–0.76	++ on PANSS	Overall and Gorham (1962),166 Overall and Gorham (1972)230
Global Assessment of Functioning Scale (GAF)	Amri et al (2014),201 Boyer et al (2013),34 Hastrup et al (2011),72 Hosseini and Karkhaneh Yousefi (2011),77 Laengle et al (2010),88 Lin et al (2018),93 Masoomi et al (2018),100 Peuskens et al (2010),120 Schwartz (2007),231 Startup et al (2002),175 Tachibana et al (2016),135 Vaingankar et al (2017),139 Widschwendter et al (2018)143	Emotional/psychological wellbeing	3 domains; score range: 0–100; higher score=better function in every domain	English	+++ ICC=0.95	+on QLS (0.2)	Dufton and Siddique (1992)158
Montgomery–Åsberg Depression Rating Scale (MADRS)	Niitsu et al (2012),112 Pietrini et al (2015)121	Emotional/psychological wellbeing	10 items rated on 0–6 Likert scale; score range: 0–60; higher score=worse depressive symptoms	English	++ ICC=0.76	++ on Hamilton Depression scale	Davidson et al (1986),157 Montgomery and Asberg (1979)165
Social and Occupational Functioning (Assessment) Scale (SOFS)	Henry et al (2010),74 Higuchi et al (2017)75	Symptomatic	Scored from 0–100; higher score=better functioning	English, Japanese	+++ ICC=0.95	++ on PANSS	Saraswat et al (2006)168
Clinical Global Impression Scale-Severity (CGI-S)a	Durgam et al (2015),49 Falissard et al (2016),50 Fe Bravo-Ortiz et al (2011),54 Fefeu et al (2018),55 Gattaz et al (2014),63 Lee et al (2014),90 Masand et al (2011),99 Suzuki et al (2017),134 Thwin et al (2013)138	Symptomatic	Scored on 7-point scale; higher score=more severe	Multiple	++ ICC: 0.69–0.96	++ on PANSS, ++ PSP	Guy et al (1976)159
Brief Negative Symptom Scale (BNSS)	Kirkpatrick et al (2011)232	Symptomatic	15 items; 5 domains; score range: 0–78; higher score=more severe	Multiple	+++ ICC=0.96 Cronbach’s α=0.93	+++ on SANS, ++ on PANSS	Kirkpatrick et al (2011)232
Heinrichs–Carpenter Quality of Life Scale (QLS)	Barnes et al (2016),28 Chaves et al (2013),40 de Pinho et al (2018),46 Falissard et al (2016),50 Golubovic et al (2010),66 Lin et al (2018),93 Naber et al (2015),109 Niitsu et al (2012),112 Suzuki et al (2017)134	QoL	21 items rated on 7-point scale; higher score=better functioning	English, Portuguese, others	+++ ICC=0.942	+++ on GAF ++ on PANSS	Heinrichs et al (1984)160
Abbreviated QLS (4- and 7-item)	Fervaha et al (2014),56 Fervaha and Remington (2013)57	QoL	Abbreviated QLS (see QLS above)	English	N/A	+++ on full QLS (r=0.91) ++ on PANSS	Fervaha and Remington (2013),57 Fervaha et al (2014)56
The Manchester Short Assessment of Quality of Life (MANSA)	Hastrup et al (2011),72 Petkari et al (2020),118 Priebe et al (2012)123	QoL	12 items; higher score=better quality of life	English	++ Cronbach’s α=0.74	+++ on SQoL +++ on LQoLP	Priebe et al (1999)167
Udvalget for Kliniske Undersoegelser (UKU)	Baandrup et al (2017),26 Chaves et al (2013),40 Chen and Lung, (2017),41 Nielsen et al (2012),111 Widschwendter et al (2018)143	Treatment-related	Higher score=worse side effects	English	++ ICC: 0.6–0.71	+ on CGI-S + on PSP	Lingjaerde et al (1987)164
Reasons for Antipsychotic Discontinuation and Continuation (RAD-Q/RAD-I)	Matza et al (2011),101 Matza et al (2012)102	Treatment-related	Higher score=more reasons for discontinuation	English	++ Kappa ≥0.7	++ on PANSS ++ on CGI-S	Matza et al (2011)101
Clinical Global Impressions Scale- Improvement (CGI-I)	Masand et al (2011)99	Treatment-related	7-point scale; lower score=improvement; higher score=worsening	English, others	N/A	++ on PANSS	Leucht et al (2006),162 Leucht et al (2005)163

^aCGI-S was used in more than 50 of the included studies.

ClinRO: clinician-reported outcome; ICC: intraclass correlation coefficient; LQoLP: Lancashire Quality of Life Profile; N/A: not available; ObsRO: observer-reported outcome; PANSS: Positive and Negative Syndrome Scale; PSP: Personal and Social Performance scale; QLS: Quality of Life Scale; QoL: quality of life; SANS: Scale for the Assessment of Negative Symptoms; SQoL: Subjective Quality of Life Scale

TABLE 3.

Schizophrenia-specific ClinROs/ObsROs

CLINICAL OUTCOME ASSESSMENT	CLINICAL OUTCOME ASSESSMENT REFERENCES	DOMAIN	INTERPRETATION	LANGUAGES	RELIABILITY	VALIDITY	ORIGINAL VALIDATION
Positive and Negative Syndrome Scale (PANSS)a	Baandrup et al (2017),26 Birur et al (2016),32 Domenech et al (2019),48 Durgam et al (2015),49 Fe Bravo-Ortiz et al (2011)54 Gattaz et al (2014),63 Itokawa et al (2018),79 Jakubovski et al (2015),80 Lee et al (2014),90 Lin et al (2018),93 Lin et al (2018),94 Lloyd et al (2010),96 Niitsu et al (2012),112 Pagsberg et al (2014),116 Suzuki et al (2017),134 Thwin et al (2013)138	Symptomatic	30 items; total score (sum of positive, negative, and general psychopathology subscale scores) range: 30–210; higher score=more severe	Multiple	+++ Cronbach’s α ≥0.8	+++ on a number of COAs	Kay et al (1987)171
Scale for Positive Symptoms (SAPS) and Scale for Negative Symptoms (SANS)	Niolu et al (2015),113 Secher et al (2015)128	Symptomatic	SANS: 25 negative symptoms rated on 6-point scale; SAPS: 34 positive symptoms rated on 6-point scale; higher scores=greater severity	English	+++ ICC=0.97 (SAPS) ICC=0.96 (SANS)	++ on GAF	Startup et al (2002)175
Brief Assessment of Cognition in Schizophrenia (BACS)	Baandrup et al (2017),26 Ogino et al (2011),115 Takahashi et al (2017)136	Cognition	Composite score calculated as a mean z-score of the components of the score	English, Japanese	++ ICC ≥0.79	++ on GAF	Keefe et al (2004)172
Functional Recovery Scale in Schizophrenia (FROGS)	Capar and Kavak (2019)37	Symptomatic	19 items rated on 5-point Likert scale	English	+++ Cronbach’s α=0.909	N/A	Llorca et al (2009)233
Calgary Depression Scale for Schizophrenia (CDSS)	Addington et al (1990),169 Amri et al (2014),24 Farreny et al (2018),53 Hayhurst et al (2014),73 Maurino et al (2011),104 Yee et al (2017)144	Symptomatic	8 questions; score range: 0–3; higher score=more severe depression	English, Spanish, Korean, Thai	+++ Cronbach’s α ≥0.84	N/A	Addington et al (1990)169
Schedule for Deficit Syndrome (SDS)	Ahmed et al (2015)21	Symptomatic	Individuals meet criteria for the deficit syndrome if they have ≥2 negative symptoms that are clinically significant, and those symptoms are considered primary and stable	English	++ ICC=0.73	N/A	Kirkpatrick et al (1989)161
4-Item Negative Symptom Assessment (NSA-4)	Alphs et al (2011),23 Garcia-Portilla et al (2015)62	Symptomatic	Each of the 4 items and the overall global negative symptoms are rated on a 1- to 6-point scale, with higher scores indicating higher degree of impairment	English	++ Cronbach’s α=0.64 ICC=0.82	+++ on PANSS ++ QLESQ	Alphs et al (2011)23
Negative Symptom Assessment (NSA-16)	Alphs et al (2011)23	Symptomatic	Higher score=more severe	Multiple	+++ Cronbach’s α=0.92	+++ on PANSS	Axelrod et al (1993),170 Garcia-Portilla et al (2015),62 Kumari et al (2017)173
Psychosocial Remission in Schizophrenia Scale (PSRS)	Barak et al (2010)27	Symptomatic	Items scored from 1 (poor) to 4 (excellent); higher score=more severe psychosocial impairment	English	N/A	N/A	Barak et al (2010)27
Clinical Assessment Interview for Negative Symptoms (CAINS)	Blanchard et al (2017),33 Horan et al (2011)76	Symptomatic	2 scales (9-item Motivation and Pleasure scale, 4-item Expression scale), scored separately; higher score=greater severity	English	++ ICC ≥70 ++ Cronbach’s α=84	+ on CDSS	Horan et al (2011)76
Epidemiological Study of Cognitive Impairment in Schizophrenia (EPICOG-SCH)	Zaragoza Domingo et al (2017),145 Zaragoza Domingo et al (2015)146	Symptomatic	4 domains; impairment prevalence is percent of patients with scores <1.5 the standard deviation (or 10th centile) from the mean of standardized scores	English	++ Cronbach’s α=0.78	++ on WHO-DAS and CGI-S (≥0.4)	Zaragoza Domingo et al (2015),146 Zaragoza Domingo et al (2017)145
Virtual Reality Functional Capacity Assessment Tool (VRFCAT)	Keefe et al (2016)86	Cognition	4 scenarios simulating everyday activity; 12 objectives rated 1/0 for completion; 3 domains (errors, progression time, and completion time)	English	++ ICC (time)=0.81, ICC (errors)=0.65, ICC (progression)=0.61	+++ on Matrics/MCCB	Keefe et al (2016)86
The Measurement and Treatment Research to Improve Cognition in Schizophrenia (Matrics/MCCB)	Green et al (2011),67 Lystad et al (2014),97 Velligan et al (2014)140	Cognition	Score range: 1–100; higher score=better cognition	English, Norwegian	++ ICC ≥0.7	+ on CGI-S ++ on other COAs	Green et al (2011)67
Personal and Social Performance Scale (PSP)	Fleischhacker et al (2013),58 Gattaz et al (2014),63 Jelastopulu et al (2014),81 Lee et al (2016),91 Patrick et al (2009)174	Emotional/psychological wellbeing	4 subdomains rated 1–6; total score range: 1–100; higher score=better functioning	English	+++ ICC ≥ 0.87	++ on PANSS ++ on CGI-S	Patrick et al (2009)175
Clear Thinking Scale (CTS)	Bridges et al (2010)36	Emotional/psychological wellbeing	Complete scoring details unavailable at time of publication	English	N/A	N/A	Bridges et al (2010)36

^aPANSS was used in more than 50 of the included studies.

CDSS: Calgary Depression Scale for Schizophrenia; CGI-S: Clinical Global Impressions Scale-Severity; ClinRO: clinician-reported outcome; COA: clinical outcome assessment; ICC: intraclass correlation coefficient; GAF: Global Assessment of Functioning; MCCB: MATRICS Consensus Cognitive Battery; N/A: not available; ObsRO: observer-reported outcome; PANSS: Positive and Negative Syndrome Scale; QLESQ: Quality of Life Enjoyment and Satisfaction Questionnaire; WHO-DAS: World Health Organization Disability Schedule

PROs. A summary of the nine generic PROs is presented in Table 4. All generic PROs were focused on QoL/HRQoL or cognition. Most of the identified generic PROs displayed acceptable reliability, with Cronbach’s α of 0.8 or greater.^{117,122,176–183} External validity, however, was not particularly strong, whereas variation among the generic outcome assessments was quite high.^{117,122,176–178,180–183}

TABLE 4.

Generic PROs

CLINICAL OUTCOME ASSESSMENT	CLINICAL OUTCOME ASSESSMENT REFERENCES	DOMAIN	INTERPRETATION	LANGUAGES	RELIABILITY	VALIDITY	ORIGINAL VALIDATION
World Health Organization Quality of Life Scale Brief Version (WHOQOL-BREF)	Adewuya and Makanjuola (2010),20 Franz et al (2013),61 Sonntag et al (2015),131 Su et al (2014)132	QoL	26 items scored 0–100 through transformation; higher score=better health	Multiple	++ Cronbach’s α: 0.73–0.84	+++ on PANSS	The WHOQOL Group (1998)183
36-Item Short-Form Health Survey (SF-36)	Domenech et al (2019),48 Franklin et al (2019),59 Grunder et al (2016),69 Guo et al (2012),71 Papaioannou et al (2011),117 Su et al (2014)132	QoL	Lower score=greater disability	Multiple	+++ Cronbach’s α=0.9	+++ on PANSS (0.86) + on other COAs	Ware (2000),182 Papaioannou et al (2011)117
EuroQol 5 Dimensions (EQ-5D)	Adewuya and Makanjuola (2010),20 Papaioannou et al (2011),117 Pitkanen et al (2012),122 Sonntag et al (2015),131 Su et al (2014)132	HRQoL	0=death, 1=best possible health	Multiple	+ Cronbach’s α=0.63 ICC=0.62	++ on SQoL	Papaioannou et al (2011),117 Pitkanen et al (2012)122
Quality of Wellbeing Scale/Quality of Wellbeing Scale-Self Administered (QWB)	Thwin et al (2013)138	QoL	0=death, 1=best possible function	English	++ ICC=0.77 on test-retest	+++ to QWB interview	Kaplan et al (1997)178
Short-Form 6-Dimension (SF-6D)	Abdin et al (2019),18 Abdin et al (2019)19	QoL	Score range: 0.29–1; higher score=better QoL	Multiple	+++ ICC=0.92	+++ on PANSS (0.69)	Brazier et al (2002)177
Health Utility Index (HUI-3)	Abdin et al (2019),18 Seow et al (2019)130	QoL	Score range: -0.36–1; higher score=better health utility	Multiple	+++ ICC=0.83	+++ on PANSS (0.69)	Luo et al (2006)180
The Impact of Weight on Quality of Life-lite (IWQOL-lite)	Liu et al (2017)95	QoL	Items rated 1 (never true), 3 (sometimes true), or 5 (always true); higher score=worse QOL	English	+++ Cronbach’s α ≥0.92	N/A	Kolotkin and Crosby (2002)179
World Health Organization Disability Assessment Schedule (WHO-DAS II)	Guilera et al (2012),70 Mas-Exposito et al (2012)98	HRQoL	Score range: 0–100; higher score=greater disability	Multiple	+++ Cronbach’s α=0.94 ICC=0.92	+ on PANSS + on EQ-5D	World Health Organization (2000)181
Bernheim’s Anamnestic Comparative Self Assessment (ACSA)	O’Neill (2010)114	QoL	Various periods of life scored from -5 (worst) to +5 (best)	English	N/A	+ on subjective wellbeing	Bernheim et al (2006)176
Cognitive Assessment Interview (CAI)	Ventura et al (2010),154 Ventura et al (2013)155	Cognition	10 items; higher score=greater degree of impairment	English	++ Cronbach’s α: 0.77–0.78 ICC=0.92	+ on MCCB + on UPSA + on RFS	Ventura et al (2016)156
Brief Cognitive Assessment (BCA)	Velligan et al (2004)152	Cognition	3 domains; composite score is a mean z-score of the 3	English	++ Cronbach’s α: 0.77–0.78 ICC=0.821	+ on LWPRAS (0.24) +++ on SOFA (0.57) ++ on MCAS (0.42) +++ on FNA (0.66)	Velligan et al (2004)152

COA: clinical outcome assessment; FNA: Functional Needs Assessment; HRQoL: health-related quality of life; ICC: intraclass correlation coefficient; LWPRAS: Lehman Work and Productive Activity Scale; MCCB: MATRICS Consensus Cognitive Battery; MCAS: Multnomah Community Ability Scale; N/A: not available; PRO: patient-reported outcome; QoL: quality of life; PANSS: Positive and Negative Syndrome Scale; SQoL: Subjective Quality of Life Scale; RFS: Role Functioning Scale; SOFAS: Social and Occupational Functioning Scale; UPSA: Performance-based Assessment of Functioning

Most (n=15) of the identified PROs were mental health-specific (Table 5). Overall, six assessments focused on emotional/psychological wellbeing, four centered on QoL/HRQoL, four assessed treatment-related factors, and one evaluated the need for care. Similar to the generic PROs, the mental health-specific PROs had good reliability, whereas external validity varied.^{68,107,109,184–193} However, external validity was not available for several mental health-specific PROs.^{68,107,188,190} Reliability or external validity were not available for two of the treatment-related PROs and the need for care PRO.^45,194,195

TABLE 5.

Mental health-specific PROs

OUTCOME ASSESSMENT	OUTCOME ASSESSMENT REFERENCES	DOMAIN	INTERPRETATION	LANGUAGES	RELIABILITY	VALIDITY	ORIGINAL VALIDATION
Quality of Life Enjoyment and Satisfaction Questionnaire (QLESQ)	Abdin et al (2019),18 Bervoets et al (2012),31 Ngoc et al (2016),110 Petrović and Jankovic (2017),119 Pitkanen et al (2012)122	QoL	16 items; higher score=better subjective QoL	English, others	+++ Cronbach’s α=0.96 overall	+++ on LQoLP (0.91)	Endicott et al (1993),185 Ritsner et al (2005)191
Lancashire Quality of Life Profile (LQoLP)	Rossell et al (2016),127 Su et al (2017)133	QoL	Higher score=better QoL	English	+++ ICC=0.92 Cronbach’s α=0.93	+ on QLS (0.38)	Van Nieuwenhuizen et al (2001)193
Tolerability and Quality of Life Questionnaire (TOOL)	Falkai (2015),51 Montejo et al (2011)108	QoL	8 items scored on 4-point Likert scale; lower score=better functioning/QoL	English, Spanish	+++ Cronbach’s α=0.92 ICC=0.90	++ on EQ-5D (-0.72) + on PANSS (0.44)	Lindström et al (2009)186
Quality of Life Interview (QoLI)	Joseph et al (2015)82	QoL	8 domains	English	++ Cronbach’s α ≥0.8 + on various domains	++ on various measures of functionality	Russo et al (1997)192
Subjective Wellbeing Under Neuroleptic Scale-Short Version (SWN-K)	Aikawa et al (2018),22 Baandrup et al (2017),26 Dhanda et al (2019),47 Karow et al (2012),85 Maurino et al (2012),103 Maurino et al (2011),104 Naber et al (2015),109 Takeuchi et al (2016)137	Emotional/psychological wellbeing	20 items; score range: 20–120; higher score=better wellbeing	English, Japanese, others	+++ Cronbach’s α=0.92 on total score	+ to ++ on SF-36 ++ on CDSS	Naber (1995),189 Naber et al (2015)109
Subjective Wellbeing Under Neuroleptic Treatment Scale (SWN)	Karamatskos et al (2012),84 Vothknecht et al (2013),147 Vothknecht et al (2011)141	Emotional/psychological wellbeing	38 items, higher score=better wellbeing	English, German, others	+++ Cronbach’s α: 0.88–0.92	++ to +++ on PANSS, CGI-S	Naber (1995)189
Choice of Outcome in Cognitive Behavioral Therapy (CBT) for Psychoses (CHOICE)	Greenwood et al (2010)68	Emotional/psychological wellbeing	24 items are scored from 0 (worst)–10 (best). The mean score is used.234	English	++ ICC: 0.88–0.83	N/A	Greenwood et al (2010)68
Self-Appraisal of Illness Questionnaire (SAIQ)	Kao and Liu (2010)83	Emotional/psychological wellbeing	17 items scored on 4-point Likert scale; higher score=better awareness of illness	English, French, Chinese	++ Cronbach’s α=0.83	+ on PANSS	Marks et al (2000)189
Patient Assessment Questionnaire (PAQ)	Mojtabai et al (2012)107	Emotional/psychological wellbeing	53 items scored on 5-point Likert scale; higher score=better wellbeing	English	+++ Cronbach’s α=0.94	N/A	Mojtabai et al (2012)107
Recovery Assessment Scale (RAS)	Cavelti et al (2017),39 Girard et al (2016)64 Girard et al (2015)65	Emotional/psychological wellbeing	41 items scored on 5-point Likert scale; higher score=better wellbeing	English, German	++ Cronbach’s α: 0.74–0.87	++ on PANSS	Corrigan et al (2004)184
Preference of Medicine Questionnaire (POM)	Awad and Voruganti (2013),11 Isitt et al (2016),78 Mencacci and All (2012)105	Treatment-related	2 items scored 1–5; lower score=greater preference for current medications	English	N/A	N/A	Tandon et al (2006)195
New York Assessment for Adverse Cognitive Effects of Neuropsychiatric Treatment (NY-AACENT)	Correll et al (2019)5	Treatment-related	7 domains are scored on a 5-point Likert scale	English	N/A	N/A	Correll et al (2019)45
Questionnaire about the Process of Recovery (QPR)	Chien and Chan (2013),42 Law et al (2014)89	Treatment-related	22 items; higher score=recovery	English, Chinese	+++ Cronbach’s α ≥0.924	N/A	Neil et al (2009)190
Illness Management and Recovery Scale (IMRS)	Fardig et al (2011)52	Treatment-related	15 items scored on 5-point scale; higher score=better progress	English	++ Cronbach’s α (client version)=0.72 Cronbach’s α (clinician version)=0.80	N/A	Mueser et al (2006)188
Client Sociodemographic and Service Receipt Inventory (CSSRI)	King et al (2011)87	Need for care	Inventory for service use, no scoring applies	English	N/A	N/A	Chisholm et al (2000)194

CGI-S: Clinical Global Impressions Scale-Severity; EQ-5D: EuroQol-5 Dimensions; ICC: intraclass correlation coefficient; N/A: not available; QLS: Quality of Life Scale; QoL: quality of life; PANSS: Positive and Negative Syndrome Scale; PRO: patient-reported outcome

The remaining PROs were comprised of 14 schizophrenia-specific outcome assessments (Table 6). Among these 14 schizophrenia-specific PROs, more than half (n=8) evaluated QoL/HRQoL and three evaluated cognition. The remaining three schizophrenia-specific PROs were evenly distributed between the treatment-related, emotional/psychological wellbeing, and symptomatic domains. Generally, the QoL/HRQoL PROs were reliable (Cronbach’s α ≥0.8) and had strong validity.^{35,60,106,125,196–199} The treatment-related schizophrenia-specific PRO had good reliability and acceptable external validity.²⁰⁰ The emotional/psychological wellbeing schizophrenia-specific PRO had good reliability but lacked data on external validity.¹⁴² The symptomatic schizophrenia-specific PRO had variable external validity but lacked data on reliability.⁸²

TABLE 6.

Schizophrenia-specific PROs

OUTCOME ASSESSMENT	OUTCOME ASSESSMENT REFERENCES	DOMAIN	INTERPRETATION	LANGUAGES	RELIABILITY	VALIDITY	ORIGINAL VALIDATION
Schizophrenia Quality of Life Scale (SQLS)	Abdin et al (2019),18 Domenech et al (2019),48 Lee et al (2014),90 Su et al (2017)133	QoL	30 items, 3 subscores; 0=best QoL, 100=worst QoL	English, others	+++ Cronbach’s α (psychosocial)=0.93 Cronbach’s α (motivation and energy)=0.78; Cronbach’s α (symptoms and side effects)=0.8	+++ on HADS, GHQ-12	Wilkinson et al (2000)198
Schizophrenia Quality of Life Questionnaire Multidimensional Computerized Adaptive Test (SQoL-MCAT)	Michel et al (2018)106	QoL	41 items, 8 domains; score range: 0–100; higher score=better QoL	French, others	+++ ICC ≥0.9	+ on PANSS ++ on CDSS	Michel et al (2018)106
Schizophrenia Quality of Life Questionnaire (SQoL)	Baumstarck et al (2013),29 Chou et al (2011),44 Lee et al (2021),92 Niolu et al (2015)113	QoL	41 items; score range: 0–100; higher score=better QoL	French, English, Chinese	+++ Cronbach’s α=0.94	++ on EQ-5D ++ on PANSS	Auquier et al (2003)196
Quality of Life in Schizophrenia (QLiS) and Short Form (QiLS-SF)	Franz et al (2012),60 Franz et al (2013),61 Senin et al (2017)129	HRQoL	54 items; higher score=better QOL	English	++ ICC ≥0.7	++ on WHOQOL BREF ++ on SWN-K	Franz et al (2012)60
The Quality of Life for Mental Disorders (QoLMD)	Chiu and Lee (2018),43 Yee et al (2017)144	HRQoL	45 items, 8 domains (each domain scored separately); higher score=better HRQoL	Chinese	++ Cronbach’s α >0.7	N/A	Yu (1995)199
Schizophrenia Quality of Life Questionnaire-Short Form (SQoL18)	Baumstarck et al (2013),29 Boyer et al (2010),35 Caqueo-Urizar et al (2014),38 Girard et al (2016)64	QoL	18 items; score of 100=best possible QoL	French, English, Spanish	++ Cronbach’s α: 0.72–0.88	+++ on SQoL ++ on EQ-5D + SF-36	Boyer et al (2010)35
Riedel-Spellmann-Musil Scale (RSM)	Riedel et al (2009),232 Riedel et al (2011)125	QoL	36 items; complete scoring details unavailable at time of publication	German, English	+++ Cronbach’s α (self)=0.917 Cronbach’s α (observer)=0.915	+++ on QLS +++ on SWN-K	Riedel et al (2011)125
Satisfaction with Life Domains Scale (for Schizophrenia) (SLDS)	Azaiez et al (2018),25 Carlson et al (2009)235	QoL	7-point Likert scale; score range: 5–35; lower score=worse QoL (20=neutral point)	English	++ Cronbach’s α=0.84	N/A	Diener et al (1985)197
Assessment of Lifespan Functioning Attainment (ALFA)	Joseph et al (2015)82	Symptomatic	5 domains, combination of duration in state and binary score	English	N/A	+ to ++ on SANS for different domains +++ on QoLI	Joseph et al (2015)82
The Patient-Reported Experience of Cognitive Impairment in Schizophrenia (PRECIS)	Rosen et al (2018),126 Welch et al (2017)142	Emotional/psychological wellbeing	35 items, 8 domains; scored on a 5-point scale	English	+++ Cronbach’s α=0.942 ICC=0.78	N/A	Welch et al (2017)142
Personal Evaluation of Transitions in Treatment (PETiT)	Rajagopalan et al (2014)124	Treatment-related	30 items; higher score=better treatment outcome	English	+++ Cronbach’s α=0.92 ICC=0.97	+ on PANSS (0.18) + on QLS (0.21) + on GAF (0.24)	Voruganti and Awad (2002)200
Schizophrenia Cognition Rating Scale (SCoRS)	Keefe et al (2004)151	Cognition	18 items; higher score=greater degree of impairment	English	++ Cronbach’s α=0.79 ICC ≥0.81	+ on BACS + on ILSI + on UPSA	Keefe et al (2004)151
Clinical Global Impressions Scale-Cognition in Schizophrenia (CGI-CogS)	Ventura et al (2008)153	Cognition	38 items; higher score=greater degree of impairment	English	+++ Cronbach’s α (self)=0.93 Cronbach’s α (caregiver)=0.96 Cronbach’s α (rater)=0.95 ICC ≥0.92	++ on NCS	Ventura et al (2008)153
Brief Cognitive Assessment Tool for Schizophrenia (B-CATS)	Bell et al (2005),148 Hurford et al (2011),149 Hurford et al (2018)150	Cognition	3 domains; composite score is a mean z-score of the 3	English	++ Cronbach’s α=0.73 ICC=0.82	+ on FES study battery + on CHOR study battery	Hurford et al (2011)149

BACS: Brief Assessment of Cognition in Schizophrenia; CDSS: Calgary Depression Scale for Schizophrenia; CHOR: clozapine, haloperidol, olanzapine, and risperidone; EQ-5D: EuroQol-5 Dimensions; FES: first-episode; GAF: Global Assessment of Functioning; GHQ-12: 12-item General Health Questionnaire; HADS: Hospital Anxiety and Depression Scale; HRQoL: health-related quality of life; ICC: intraclass correlation coefficient; N/A: not available; NCS: Cogtest Neurocognitive Composite Score; PANSS: Positive and Negative Syndrome Scale; PRO: patient-reported outcome; QLS: Quality of Life Scale; QoL: quality of life; QoLI: Quality of Life Interview; SANS: Scale for the Assessment of Negative Symptoms; SF-36: 36-Item Short-Form Health Survey; SWN-K: Subjective Wellbeing Under Neuroleptic Treatment Scale-Short Version; UPSA: Performance-Based Assessment of Functioning; WHOQOL-BREF: World Health Organization Quality of Life Scale

MCIDs. MCIDs were reported for eight of the 66 identified outcome assessments (Table 7).^{24,50,69,91,138,201–203} Of these eight outcome assessments, six were ClinROs/ObsROs (3 mental health-specific and 3 schizophrenia-specific), while only two were PROs (both generic). One of the outcome assessments, the 36-item Short Form Health Survey (SF-36), reported predetermined MCIDs rather than derived values.⁶⁹

TABLE 7.

Summary of identified clinical outcome assessments with MCIDs

CLINICAL OUTCOME ASSESSMENT	COMPLETION/FOCUS	INTERPRETATION	MCID	REFERENCE
Positive and Negative Syndrome Scale (PANSS)	Clinician SCH-S	30 items; total score (sum of positive, negative, and general psychopathology subscale scores) range: 30–210; higher score=more severe	15.3 points (34.0%) change from baseline 4.25 for nonworking and 8.3 for working patients 20.2 points	Hermes et al (2012),202 Thwin et al (2013),138 Leddy-Stacy and Rosenheck (2016)203
Global Assessment of Functioning Scale (GAF)	Clinician MH-S	Score range: 0–100; higher score=better function in every domain	4	Amri et al (2014)201
Clinical Global Impressions Scale-Severity (CGI-S)	Clinician MH-S	Scored on 7-point scale; score range: 1–7; higher score=more severe	0.17	Thwin et al (2013)138
Heinrichs–Carpenter Quality of Life Scale (QLS)	Clinician MH-S	21‐items rated on a 7-point scale; score range: 0–126; higher score=better functioning	5.3 in Falissard et al (2016)50 1.13 in Thwin et al (2013)138	Thwin et al (2013),138 Falissard et al (2016)50
Calgary Depression Scale for Schizophrenia (CDSS)	Clinician SCH-S	8 questions scored 0–3, plus 1 observational item; score range: 0–27; higher score=more severe depression	1.3	Amri et al (2014)24
Personal and Social Performance Scale (PSP)	Clinician SCH-S	4 subdomains rated 1–6; total score range 1–100; higher score=better functioning	10.7 for single rater, 16.2 for multiple raters 7 for 1-category improvement on CGI-S	Lee et al (2016),91 Patrick et al (2009)177
Quality of Wellbeing Scale/Quality of Wellbeing Scale-Self Administered (QWB)	Self or observer Generic	0=death, 1=best possible function	0.17	Thwin et al (2013)138
36-Item Short-Form Health Survey (SF-36)	Self Generic	Lower score=less disability	5–10 (predefined before trial)	Grunder et al (2016)69

MCID: minimal clinically important difference; MH-S: mental health-specific; SCH-S: schizophrenia-specific

DISCUSSION

This analysis of schizophrenia-related literature identified 66 individual clinical outcome assessments, which represented patient- and nonpatient-reported outcome assessments, as well as various domains and focuses of outcome assessments for patients with schizophrenia. An advantage of a scoping review over a systematic review, in this context, was the ability to include a broader range of studies and review a larger share of the published literature. This benefit was reflected in the greater number of studies and outcome assessments that we evaluated in this review, as compared to previous efforts over the past decade.^25,204,205

The findings of this scoping review highlighted the heterogeneity of existing assessments in terms of scope, focus, and performance. While the reliability of most of the identified ClinROs/ObsROs was good, external validity varied considerably. Weak external validity was noted among newly developed schizophrenia-specific outcome assessments, suggesting a need for further validation through additional research. Overall, mental health- and schizophrenia-specific PROs demonstrated strong psychometric properties (reliability and external validity); these findings were consistent with previous studies.¹³⁰

The presence of published MCID values for only eight of the 66 identified outcome assessments indicates a substantial gap in knowledge. In particular, none of the schizophrenia-specific PROs had reported MCID values. Additionally, searches performed using the PROLABELS™ database confirmed that PRO endpoints were not mentioned in the labels of any antipsychotic agents approved over the last decade (2010–2020), further highlighting this gap.

While MCIDs are needed to help guide clinical practice and assess treatment effectiveness in research, the development of MCIDs requires substantial investments in research and trial resources. There is no consensus on a universal method for calculating MCIDs for PROs; anchor-based and distributive methods have been the most commonly used for the calculation of MCIDs, with distributive methods being the preferred method in schizophrenia.^206,207 By definition, PROs rely on self-report by patients, and the issue of reliability of PROs for MCID calculation has been raised in the literature.²⁰⁷ The main criticism has been that MCIDs based on PROs are unreliable in a population of patients with schizophrenia because of discrepancies between the perceptions of patients and those of other raters (i.e., clinician, observer, or proxy [e.g., family, caregiver]).²⁰⁸ Therefore, the selection of outcome measures should be considered carefully. In this review, criteria for the selection of an outcome measure for MCID evaluation that prioritized multiple aspects of an outcome measure (PROs, mental health- or schizophrenia-specific measures with strong psychometric properties, evidence of use in patients with schizophrenia, and established psychometric properties) were proposed. The application of this criteria to identified outcome measures resulted in a list of recommended outcome assessments for MCID research development. However, the list of suggested outcome measures for MCID evaluation is not exhaustive and only considers the outcome measures that met our recommendation criteria in the strictest sense.

Recommendations for MCID research. Based on the findings of this review, it is recommended that researchers consider the following priorities when selecting an outcome measure for MCID evaluation (Figure 3): type (prioritize PROs because of the lack of reported MCIDs among existing outcome measures), focus (prioritize mental health- or schizophrenia-specific measures based on the strong psychometric properties observed in this scoping review), evidence of use in patients with schizophrenia, and established and acceptable psychometric properties. By applying these criteria, three mental health-specific PROs were identified that should be evaluated for MCIDs: Quality of Life Enjoyment and Satisfaction Questionnaire (QLESQ), Subjective Wellbeing Under Neuroleptic Treatment Scale (SWN), and SWN–Short Version (SWN-K) (Table 8). Furthermore, three schizophrenia-specific PROs with potential for MCID evaluation were identified: Schizophrenia Quality of Life Scale (SQLS), Schizophrenia Quality of Life Questionnaire (SQoL), and SQoL–Short Form (SQoL18; Table 8). As of May 2020, the QLESQ, SWN, and SQLS were being utilized in ongoing Phase III trials indexed on ClinicalTrials.gov.^209–229

FIGURE 3.

Suggested factors of minimal clinically important difference (MCID) development

TABLE 8.

Recommended outcome assessments for MCID research development

OUTCOME ASSESSMENT	OUTCOME ASSESSMENT REFERENCES	DOMAIN	INTERPRETATION	LANGUAGES	RELIABILITY	VALIDITY	ORIGINAL VALIDATION
Quality of Life Enjoyment and Satisfaction Questionnaire (QLESQ)	Abdin et al (2019),18 Bervoets et al (2012),31 Ngoc et al (2016),110 Petrović and Jankovic (2017),119 Pitkanen et al (2012)122	QoL	16 items; higher scores indicate greater subjective QoL	English, others	+++ Cronbach’s α=0.96 overall	+++ 0.91 on LQoLP	Endicott et al (1993),185 Ritsner et al (2005)191
Subjective Wellbeing Under Neuroleptic Treatment Scale (SWN)	Karamatskos et al (2012),84 Vothknecht et al (2013),147 Vothknecht et al (2011)141	Emotional/psychological wellbeing	38 items; higher scores=better wellbeing	English, German, others	+++ Cronbach’s α: 0.88–0.92	++ to +++ on PANSS, CGI-S	Naber (1995)189
Subjective Wellbeing Under Neuroleptic Scale–Short Version (SWN-K)	Aikawa et al (2018),22 Baandrup et al (2017),26 Dhanda et al (2019),47 Karow et al (2012),85 Maurino et al (2012),103 Maurino et al (2011),104 Naber et al (2015),109 Takeuchi et al (2016)137	Emotional/psychological wellbeing	20 items; score range: 20–120; higher scores=better wellbeing	English, Japanese, others	+++ Cronbach’s α=0.92 on total score	+ to ++ on SF-36 ++ on CDSS	Naber (1995),189 Naber et al (2001)236
Schizophrenia Quality of Life Scale (SQLS)	Abdin et al (2019),18 Domenech et al (2019),48 Lee et al (2014),90 Su et al (2017)133	QoL	30 items; 3 subscores: 0 (best)–100 (worst) QoL	English, others	+++ Psychosocial α=0.93; motivation and energy=0.78; symptoms and side effects=0.8	+++ on HADS, GHQ-12	Wilkinson et al (2000)198
Schizophrenia Quality of Life Questionnaire (SQoL)	Baumstarck et al (2013),29 Chou et al (2011),44 Lee et al (2021),92 Niolu et al (2015)113	QoL	41 items; score range: 0–100; higher scores represent higher QoL	French, English, Chinese	+++ Cronbach’s α=0.94	++ on EQ-5D ++ on PANSS	Auquier et al (2003)196
Schizophrenia Quality of Life Questionnaire-Short Form (SQoL18)	Baumstarck et al (2013),29 Boyer et al (2010),35 Caqueo-Urizar et al (2014),38 Girard et al (2016)64	QoL	18 items; score range: 0–100; higher scores represent higher QoL	French, English, Spanish	++ Cronbach’s α: 0.72–0.88	+++ on SQoL ++ on EQ-5D + SF-36	Boyer et al (2010)35

CGI-S: Clinical Global Impressions Scale-Severity; CDSS: Calgary Depression Scale for Schizophrenia; EQ-5D: EuroQol-5 Dimensions; GHQ-12: 12-item General Health Questionnaire; HADS: Hospital Anxiety and Depression Scale; LQoLP: Lancashire Quality of Life Profile; MCID: minimal clinically important difference; PANSS: Positive and Negative Syndrome Scale; QoL: quality of life; SF-36: 36-Item Short-Form Health Survey

Limitations. The literature search was limited to studies published during or after 2010 and written in English. The review may have missed relevant outcome assessments used in earlier studies or published in other languages. However, if these missed outcomes were not used in the last decade or were not published in the de facto language of most regulatory agencies and medical research, the impact of their exclusion from our review can be considered minimal. In addition, some outcome assessments lacked reliability and/or validity, which makes comparisons with other assessments difficult. Please note that the categories of mental health- and schizophrenia-specific can appear arbitrary for those measures that can be employed transdiagnostically. Lastly, the strength of the reliability and validity of outcome assessments were only determined with Cronbach’s α method and ICC, respectively, which may have limited the breadth of this review.

CONCLUSION

This scoping review provided a comprehensive overview of the clinical outcome assessments utilized during the past ten years in the field of schizophrenia and related disorders. The 66 identified outcome assessments were found to be of varying focus, scope, and validity. A large portion of these outcome assessments were patient-reported, indicating the importance of a growing implementation of PROs in the field of schizophrenia. However, none of the schizophrenia-specific PROs had estimated MCIDs, suggesting significant gaps in knowledge and opportunities for future research. Mental health- and schizophrenia-specific PROs overall demonstrated strong psychometric properties (reliability and external validity), and mental health-specific ClinROs/ObsROs generally had good reliability. Additional research is warranted to interpret individual and group level changes in the most sensitive disease-specific PROs.

ACKNOWLEDGMENTS

Medical writing and editorial support for the development of this manuscript, under the direction of the authors, was provided by Adel Chowdhury, PharmD; Mark Skopin, PhD, CMPP; Jennifer C. Jaworski, MS, BCMAS, CMPP; and Frederique H. Evans, MBS, all of whom are with Ashfield MedComms, an Inizio company. Medical writing and editorial support were funded by Teva Branded Pharmaceutical Products R&D, Inc.

AUTHOR CONTRIBUTIONS

All authors contributed to the conceptual design and writing of the manuscript. All authors have approved the final manuscript.