Abstract

Objectives

Functional impairment is an important driver of disability in patients with bipolar disorder (BD) and can persist even when symptomatic remission has been achieved. The objectives of this systematic literature review were to identify studies that assessed functioning in patients with BD and describe the functional scales used and their implementation.

Methods

A systematic literature review of English‐language articles published between 2000 and 2017 reporting peer‐reviewed, original research related to functional assessment in patients with BD was conducted.

Results

A total of 40 articles met inclusion criteria. Twenty‐four different functional scales were identified, including 13 clinician‐rated scales, 7 self‐reported scales, and 4 indices based on residential and vocational data. The Global Assessment of Functioning (GAF) and the Functional Assessment Short Test (FAST) were the most commonly used global and domain‐specific scales, respectively. All other scales were used in ≤2 studies. Most studies used ≥1 domain‐specific scale. The most common applications of functional scales in these studies were evaluations of the relationships between global or domain‐specific psychosocial functioning and cognitive functioning (eg, executive function, attention, language, learning, memory) or clinical variables (eg, symptoms, duration of illness, number of hospitalizations, number of episodes).

Conclusions

The results of this review show growing interest in the assessment of functioning in patients with BD, with an emphasis on specific domains such as work/educational, social, family, and cognitive functioning and high utilization of the GAF and FAST scales in published literature.

1 INTRODUCTION

A meta‐analysis of 25 studies conducted in 15 countries estimated a pooled lifetime prevalence of 1.06% for bipolar I disorder (BP‐I) and 1.57% for bipolar II disorder (BP‐II).1, 2 However, US epidemiologic studies suggest that approximately 60% of patients with BP‐I receive mental health treatment, a statistic that is similar to what has been reported for other countries.3 This suggests that many individuals with bipolar disorder (BD) may experience continued disability, due in part to lack of treatment intervention, which could potentially limit their functioning and productivity and also decrease overall quality of life for them and their families.

The World Health Organization (WHO) has ranked BD as the 12th leading cause of disability worldwide, and the worldwide prevalence of moderate or severe disability for BD is estimated at approximately 22 million individuals.4 Poor functioning is considered a key driver of disability in patients with BD.5 The WHO International Classification of Functioning, Disability, and Health describes functioning and disability as multidimensional concepts involving the ability to control physical functions; perform activities related to domains such as self‐care and domestic, occupational, social, and civic life; engage in all aspects of life; and manage aspects of the environment that help or hinder these experiences.6, 7 Patients with BD report a variety problems in work functioning,8 with severe work impairment during a considerable portion of their long‐term course of illness,9 and high unemployment rates.10 They have decreased social engagement,11 weaker family relationships,12 and an increased likelihood of being separated, divorced, or widowed.13 For many patients with BD, these functional impairments persist into symptomatic remission, leading to difficulties in many aspects of their lives.14, 15 The negative effects of BD extend to caregivers, who report substantial burden and distress involving relationships and day‐to‐day activities.16 Thus, for many patients with BD and their families, functional outcome, measured as the ability to fulfill role expectations in all aspects of life and maintain interpersonal relationships, is at times more important than syndromal outcome.17

Although most interventional studies in patients with BD have focused on symptoms, recurrences, and mood states as the primary outcome variable, increasingly, studies are also assessing functioning as a key outcome.17 In addition, numerous studies have sought to identify determinants of functional outcome in patients with BD. A variety of correlates of poor functional outcome have been identified, including clinical factors such as lack of treatment adherence, comorbid substance abuse or anxiety disorder, and subsyndromal symptoms; demographic variables such as older age, male sex, and low socioeconomic status; and cognitive dysfunction, particularly verbal memory impairment and executive dysfunction.5 The growing interest in the relationship between cognitive and psychosocial function in BD is reflected in several recent reviews on this topic6, 17, 18, 19 and compels a better understanding of the tools used for functional assessment in this population. The purpose of this systematic literature review was to identify and describe scales that are used to assess functioning in studies of patients with BD and to gain an understanding of the domains of function that clinicians are measuring.

2 METHODS

2.1 Information sources and eligibility criteria

This systematic literature review was conducted according to the recommendations outlined in the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) statement.20 A search was performed in BIOSIS Previews, Embase, and MEDLINE for English‐language articles published in peer‐reviewed journals between 1 January 2000 and 6 November 2017. The following search string was applied using the abstract and title as search fields: (“functional impairment” OR “psychosocial outcome” OR “psychosocial functioning” OR “psychosocial treatment” OR “psychosocial impairment” OR “occupational function” OR “occupational impairment” OR “occupational” or “functional disability” OR “psychosocial disability” OR “disability or work” OR “cognition” OR “social cognition” OR “stress” OR “functional remediation” OR “cognitive remediation”) AND (“functional recovery” OR “functional outcome” OR “outcome n/1 [patient or recovery]” AND (“bipolar disorder” OR “bipolar I disorder” OR “bipolar II disorder” OR mania OR manic OR “bipolar mania” OR “bipolar depression” OR “bipolar I disorder” OR “manic psychosis” OR “bipolar disorder” OR “affective disorders, psychotic”) AND (scale* OR measure*). Conference publications (ie, posters, summaries, and abstracts), review articles, notes, letters, book chapters, interactive tutorials, or surveys; publications involving animal or in vitro studies; and publications on pediatric populations were excluded.

2.2 Article selection process

The title and abstract of each retrieved article were independently screened against eligibility criteria by one author. Selected full‐text articles were then divided between the authors for detailed review and inclusion assessment. Figure 1 shows the article selection process.

3 RESULTS

3.1 Article selection

The search retrieved 104 articles; 30 were determined ineligible based on screening of titles and abstracts, and 74 were selected for full‐text review. Forty articles met the eligibility criteria and were included in the qualitative data analysis (Figure 1).

3.2 Risk of publication bias

Several authors and research groups authored multiple articles identified in this review (Table 1 and Table 2), indicating the potential for publication bias.

3.3 Overview of selected studies

The eligible articles included 20 cross‐sectional studies (50%, 20/40; Table 1) and 20 longitudinal studies (50%, 20/40; Table 2). Most of the studies (78%, 31/40) were published after 2007. Twenty studies were conducted in Spain, eight in the United States, five in Argentina, two in Brazil, and one each in Denmark, France, Greece, Norway, and Turkey.

Across studies, a total of 24 different functional outcome assessments were identified, including 13 clinician‐rated scales, 7 self‐reported scales, and 4 indices based on residential and vocational data (Table 3). The most commonly used scales were the Global Assessment of Functioning21 (GAF), which was used in 17 studies (43%), and the Functioning Assessment Short Test22 (FAST), which was used in 15 studies (38%), both of which are clinician‐rated scales. Although the GAF, developed in 1976, dominated the early literature, the FAST, introduced in 2007, was used in nearly as many studies as the GAF in the last 10 years. Two clinician‐rated scales, the Levenstein Global Outcome Scale23 (LGOS) and the Strauss‐Carpenter Outcome Scale24 (SCOS) were used in two studies each. The remaining nine clinician‐rated scales were each used in only one study. All seven self‐reported functioning scales were used in only one study each. One study (3%) utilized both the Residential Status Index (RSI) and Vocational Status Index (VSI), while another utilized the Modified Location Code Index (MLCI) and Modified Vocational Status Index (MVSI).

Table 3 lists the 24 scales and functional domains addressed by each. Three scales, the GAF, the LGOS, and the Functional Recovery Self‐assessment25 (a 1‐question [yes/no] self‐assessment) only provide a global assessment of general functioning, whereas the remaining 21 scales provide assessments related to specific domains such as work/education, residential life, social relationships, family relationships, psychological health, autonomy and performance, and recreation. Thirteen instruments include an occupational functioning domain. Social life was the second most common functional domain included in the scales. Two scales, the FAST and the Structured Clinical Interview for Mood Spectrum26 (SCI‐MOODS), include a cognitive function domain. Most studies used ≥1 domain‐specific scale.

3.4 Cross‐sectional studies

Of the 20 cross‐sectional studies included in this review, 9 (all published between 2002 and 2016) used the GAF as the sole functioning scale.27, 28 Four of these studies compared overall functioning between patients with euthymic BP‐I or BP‐II vs healthy controls,27, 32, 33 and four compared functioning between BP‐I, BP‐II, and/or healthy controls or between patients with BP‐I and BP‐II with different mood states (ie, depression, major depression, mania, hypomania, and euthymia) or functioning (ie, low vs high).29, 30, 31 The remaining study28 used the GAF to compare functioning in patients with euthymic BD and patients with schizophrenia. Eight27, 28 of these nine studies that used only the GAF to measure psychosocial functioning evaluated the association between psychosocial functioning and ≥1 cognitive domain (eg, attention, learning and memory, executive function, language). Five of these studies also assessed correlations between the GAF and demographic27, 31 and/or clinical variables,27, 28 such as duration of illness, number of hospitalizations, number and type of episodes, symptom type and severity, and medications. One study also evaluated the relationship between psychosocial functioning and theory of mind, the ability to perceive other people's mental states.27 The ninth study35 used the GAF to assess functioning and symptoms in patients with BD.

In addition to these nine studies using only the GAF, one study36 used the GAF along with the Self‐reported Social Functioning Scale (SFS)37 to explore the relationships between executive functioning and both global functioning (GAF) and domain‐specific functioning (SFS; ie, withdrawal, interpersonal behavior, prosocial activities, recreation, independence‐performance, independence‐competence, and employment/occupation) in patients with euthymic BP‐I.

Six cross‐sectional studies38, 39 used the FAST to assess functioning; all were published between 2010 and 2017. Five of these studies assessed domain‐specific functioning (ie, autonomy, occupational functioning, cognitive functioning, financial issues, interpersonal relationships, and leisure time) in patients with BP‐I vs healthy controls,38, 39 in patients with BP‐I and BP‐II in remission,40, 43 or across different mood states in BP‐I and BP‐II.42 These studies used the FAST to determine correlations between psychosocial functioning and clinical variables (eg, symptoms, number of hospitalizations, number of episodes, duration of illness, trait‐impulsivity, neurologic soft signs),38, 39, 40 emotion processing,38 cognitive functioning (eg, verbal learning and memory, visual memory, executive function, perceived cognitive performance),38, 39, 43 social cognition,38 sleep disturbances,43 or demographic variables40 or to assess overall and domain‐specific functioning in different mood states.42 The sixth study that used only the FAST to assess psychosocial functioning41 compared overall functioning in healthy controls and patients with BP‐I or BP‐II and assessed the relationship to cognitive functioning and depression severity. In addition to the six studies using only the FAST, one study22 used the GAF to assess the validity of the Spanish‐language version of the FAST as a new functional scale in patients with BD.

The remaining three cross‐sectional studies44, 45 used functional scales other than the GAF and the FAST. One study46 used the RSI and VSI to assess the relationship between functional recovery and demographic, neurocognitive, and clinical factors in patients with BP‐I or BP‐II who were euthymic or had residual depression. Another study45 used the Social and Occupational Functioning Assessment Scale to assess the relationship between social life and occupational functioning and clinical, sociodemographic, and neurocognitive variables (ie, executive function, attention, verbal fluency, and verbal learning and memory) in patients with strictly defined euthymic BP‐II vs healthy controls. The third study44 used the Work and Social Adjustment Scale,47 which measures occupational, family, social life, and leisure domains, and the Young Schema Questionnaire Short Version48 to examine the relationships between domain‐specific functional impairment and early maladaptive schemas.

3.5 Longitudinal observational studies

Fifteen longitudinal observational studies met the inclusion criteria. Five studies (all published between 2011 and 2016) only used the FAST scale49, 50; two (published between 2010 and 2013) used both the GAF and the FAST54, 55; three (published between 2003 and 2017) used the GAF along with functional scales other than FAST25, 56, 57; and five (published between 2005 and 2010) used scales other than the GAF and the FAST to assess functioning.58, 59

Five studies assessed functioning using only the FAST scale. Two of the studies49, 53 assessed functioning in patients with euthymic BP‐I or BP‐II with a follow‐up of 1‐3 years. The first53 used the FAST to evaluate the relationships between overall psychosocial functioning and mood, symptoms, and antipsychotic exposure; the second49 used it to assess the relationships between psychosocial functioning and verbal memory and subthreshold depressive symptoms. One study50 compared 6‐year functional outcome in patients with euthymic BP‐I or BP‐II on lithium monotherapy vs healthy controls, assessing the relationship between changes in psychosocial functioning and cognitive performance (ie, executive function, inhibition, processing speed, verbal memory) over time in patients who responded to lithium. The two remaining studies, both published by the same research group, evaluated functional outcome in patients with BP‐I and BP‐II 6 months after an acute episode or subsyndromal state51 and in patients with BP‐I, BP‐II, and BD not otherwise specified after first or multiple affective episodes.52

The two observational studies that used both the GAF and the FAST54, 55 involved authors from the same institution and program. The studies included baseline assessment of psychosocial functioning using only the GAF, with addition of the FAST at endpoint to derive information related to specific functional domains. The first study54 sought to identify clinical and neurocognitive predictors of functional outcome in patients with euthymic BP‐I or BP‐II followed for 4 years. The second study55 assessed correlations between cognitive changes, including executive functioning, inhibition, attention, processing speed, verbal memory, and visual memory, and functioning in patients with euthymic BP‐I or BP‐II on lithium therapy at the time of enrollment vs healthy controls followed for 6 years.

One study57 used the GAF for severity assessment at baseline and the MLCI and the MVSI to evaluate functional recovery at baseline and follow‐up over 4 years in patients with BP‐I hospitalized for a first manic or mixed episode. A second study56 used the GAF along with the WHO Psychiatric Disability Assessment Schedule (WHODAS)63 to assess the relationship between global functioning (GAF); personal care, occupational, family, and social functioning (WHODAS); and neurocognition in patients with BP‐I, patients with schizophrenia, and healthy controls over 1 year. A third study25 assessed the relationship between functioning and number of previous episodes and the long‐term (4 years) functional outcome of patients with euthymic BP‐I or BP‐II under naturalistic treatment conditions, using the GAF to assess global functioning and a self‐reported assessment to measure functional recovery.

Five studies (published between 2005 and 2010)58, 59 used functional scales other than the GAF and the FAST to assess functioning in patients with various psychiatric disorders. One of these studies59 assessed the relationships between overall functioning using the LGOS, work disability/social adjustment using the SCOS, neurocognitive performance, and depressive symptoms in patients with BP‐I 15 years after an index manic episode. Similarly, the LGOS and SCOS were used in a second study60 to evaluate global functioning and work performance/social adjustment, respectively, in patients with bipolar mania, unipolar nonpsychotic depression, and unipolar psychotic depression followed for up to 8 years after an index episode, with the objective of examining longitudinal associations with life satisfaction. A third study61 used the Multidimensional Scale for Independent Functioning64 to compare social adjustment and independent living in inpatients and outpatients with BP‐I 1 year after hospital discharge and relationships with comorbid personality disorders and other clinical factors. The fourth study used the Streamlined Longitudinal Interview Clinical Evaluation from the Longitudinal Interval Follow‐up Evaluation65 to compare work functioning and life satisfaction in patients with manic or mixed bipolar episodes identified as responders or nonresponders 1 year after initiating treatment and determine predictors of treatment nonresponse.62 A fifth study58 used four questions from the SCI‐MOODS to measure self‐reported cognitive problems (ie, memory, decision‐making, concentration, and mental fitness) and the relationship to employment trajectory in patients with BP‐I treated with mood stabilizers, typical and atypical antipsychotics, antidepressants, benzodiazepines, or stimulants and followed for up to 43 months.

3.5.1 Longitudinal interventional studies

Five longitudinal interventional studies66, 67 (published between 2005 and 2015) met the inclusion criteria for this review. One study69 used the FAST to compare functional improvement in patients with euthymic BP‐I or BP‐II randomized to 21 weeks of a functional remediation program, psychoeducation, or treatment as usual. This study69 assessed the efficacy of the functional remediation program with respect to overall psychosocial functioning. A second study67 used the GAF, the Shortened version of the Quality of Life Scale of the WHO Quality of Life Assessment,71 and the Social Adjustment Scale Self‐Report72 to assess the effectiveness of psychoeducation compared with relaxation sessions in improving overall functioning, adjustment to the external environment, and social functioning, respectively, at 6‐ and 12‐month follow‐ups in patients with euthymic BP‐I or BP‐II receiving pharmacologic treatment.

The remaining three longitudinal interventional studies used functional scales other than the GAF or FAST. One study70 used the UCLA Social Attainment Survey,73 the Interpersonal Support Evaluation List, and the Family Assessment Device to measure functional impairment related to activities outside the home, social support, and family functioning, respectively, as a predictor of mood symptoms in patients with BP‐I 1 year after acute phase treatment consisting of pharmacotherapy alone or in conjunction with multifamily psychoeducation group therapy or individual family therapy. The second study66 used the Self‐reported Medical Outcomes Study SF‐36 Psychosocial Function and Health‐rating Questionnaire74 to assess improvements in physical, social, and major role activities due to poor health and the association with symptomatic remission and clinical recovery in acutely manic patients with BP‐I receiving daily olanzapine monotherapy in the 49‐week open‐label extension of a 3‐week randomized controlled trial. The third and most recent study68 used the Sheehan Disability Scale75 to assess functional recovery in family, social and work/school domains over 2 years in patients with BP‐I or BP‐II openly treated with aripiprazole and an adjunct mood stabilizer.

4 DISCUSSION

The objectives of this systematic literature review were to identify studies that assessed functioning in patients with BD and describe the functional scales used and their implementation. The results show an increasing number of such studies over the past several decades, suggestive of a growing appreciation for the importance of functional outcomes in this patient population. The 40 cross‐sectional and longitudinal studies identified utilized a total of 24 different functional scales, including global and domain‐specific instruments. However, one global assessment, the GAF, and one domain‐specific assessment, the FAST, have prevailed in the literature.

This is the first systematic literature review to examine the scales used in the assessment of functional outcomes in patients with BD. The focus of earlier systematic literature reviews has been on the relationship between cognitive impairment and functional outcomes in this patient population. A systematic literature review of 21 studies concluded that cognitive function is prognostic of general functioning in patients with BD and may be a better predictor than clinical severity measures.6 Similarly, a meta‐analysis of 22 studies reported a statistically significant correlation between neurocognitive and everyday functioning in patients with BD.19 Another systematic literature review of 12 studies observed an association between impaired cognition and impaired functioning in patients with BD but noted the need for additional studies using objective, performance‐based, functional measures.46 Nearly half of the studies included in the current systematic literature review sought to evaluate the relationship between psychosocial functioning and ≥1 domain of cognitive functioning. A similar number of studies examined the relationship between psychosocial functioning and clinical variables such as symptoms, duration of illness, number of episodes, and number of hospitalizations. A smaller number of studies assessed long‐term global or domain‐specific functional outcomes. Taken together with previous reports, the current systematic literature review confirms a growing interest in evaluating functional outcomes in patients with BD and an increasing demand for objective and domain‐specific scales that can address specific areas of interest such as cognition.39

The results of this systematic literature review show that the GAF was the most frequently used scale in cross‐sectional studies and the second most frequently used scale in longitudinal studies. After its introduction for use in the Diagnostic and Statistical Manual of Mental Disorders (DSM), Third Edition in 1980,76 the GAF gained widespread use and is now the most widely used clinician rating scale of disability.77 Although this literature review did not account for studies before 2000 and after the introduction of the GAF, there are several early studies that utilized the GAF, which likely set precedence for its use in later longitudinal and cross‐sectional studies. One key study used the GAF along with the MVSI and MLCI to measure global functioning, occupational status, and residential status, respectively, over 4 years78 in patients with BD; the aim of the study was to identify factors contributing to long‐term outcomes.

The reliability and validity of the GAF, however, depend on the rater's training and expertise, and GAF scores have been observed to correlate more with symptom severity than functional impairment.79 The FAST was introduced in 2007 as a simple interview to assess the following 6 domains of functioning considered to be the main problem areas for patients with mental illness, including BD: autonomy, occupational functioning, cognitive functioning, finances, interpersonal relationships, and leisure time.22 Results of this literature review show rapid and widespread adoption of the FAST in clinical studies, with use in 7 of 17 cross‐sectional studies, 7 of 12 longitudinal observational studies, and 1 of 4 longitudinal interventional studies published after 2007. The trend toward evaluating multiple domain‐specific aspects of functioning in patients with BD is expected to continue after the official deletion of the GAF from the DSM, Fifth Edition80 and the endorsement of the WHO Disability Assessment Schedule Version 2.0 (WHODAS 2.0),81 which measures cognition, mobility, self‐care, getting along, life activities, and participation, for clinical evaluation of disability.82 We see the adoption of the WHODAS as a valuable addition to the field as it provides more specific measures of functioning. On the other hand, the GAF provides an overall rating that includes functioning which in some cases may be useful because of its ease of use. As with any scale, investigators will need to decide which scale best meets their needs.

The remaining 22 scales of functioning identified in this systematic literature review were used in few studies. It is worth noting that all but 2 of these 22 scales are domain‐specific measures of functioning. The use of many different domain‐specific assessments during the early time frame of this systematic literature review may reflect the lack of availability during that period of a single functional scale capturing all functional domains of interest. Although there was considerable variability across these scales with respect to the nature of included functional domains, work/education, social, and family domains were most widely represented. The prevalence of scales incorporating these domains likely reflects an understanding that work, family, and social/interpersonal engagement are key areas of functional impairment in patients with BD that can persist into clinical remission5, 10 and affect the well‐being of patients and their families. However, the abundance of unique scales to assess the same functional domains presents a challenge when attempting to synthesize the literature. Although cognitive functioning was not broadly represented in these less widely utilized scales, it is included in the FAST and the WHODAS 2.0.

The results of this systematic literature review further the understanding of the scales used to assess functioning in patients with BD and the nature of the research questions to which they are applied. Such findings are important considering the recognition of different types of functional impairment in this patient population and the persistence of these impairments into symptomatic remission. However, results of this literature review show a relatively small number of interventional studies assessing functional outcomes in patients with BD. Thus, there is a continued need to ensure that functioning of patients with BD is included as an outcome in studies evaluating the effects of treatment interventions.

This systematic review has several limitations. Although the search parameters were designed to be comprehensive, it omitted non‐English articles and articles published before 2000, which could have limited its scope. In several areas of the literature, particularly studies using the GAF and FAST, multiple studies were published by the same research groups or authors at the same institutions. Although an effort was made to omit duplicate publications, it is important to recognize the potential for bias due to multiple publications by the same investigators.

In conclusion, this review shows high utilization of the GAF and FAST scales for assessment of functioning in cross‐sectional observational and interventional studies. Findings indicate interest in evaluating specific functional domains, with emphasis on work/educational, social, family, and cognitive functioning, and determining the relationships between psychosocial functioning, cognitive functioning, and clinical variables. More prospective studies are needed to assess the effect of interventional therapies on functional outcomes in patients with BD.

DISCLOSURES

Maxine Chen and Jessica J. Madera are employees of Otsuka Pharmaceutical Development & Commercialization, Inc. Heather M. Fitzgerald is an employee of Lundbeck LLC. Mauricio Tohen was an employee of Eli Lilly (1997‐2008) and has received honoraria from or consulted for Abbott, AstraZeneca, Alkermes, Allergan, Bristol Myers Squibb, GlaxoSmithKline, Eli Lilly, Johnson & Johnson, Otsuka, Merck, Sunovion, Forest, Roche, Elan, Lundbeck, Teva, Pamlab, Minerva, Pfizer, Wyeth, and Wiley Publishing; his spouse was a full‐time employee at Eli Lilly (1998‐2013).