INTRODUCTION
Instrumental activities of daily living (IADL) refer to the essential tasks required to maintain independence at home and in the community. Schizophrenia is a persistent condition characterized by recurrent episodes marked by symptoms, such as delusions and hallucinations, and cognitive impairments that affect daily functioning [1,2]. More than half of people with schizophrenia encounter difficulties with IADL; this subsequently impacts the lives of their family members and others and leads to considerable social costs [1,3]. Consequently, IADL training is a crucial component of psychiatric rehabilitation, which aims to reintegrate people with schizophrenia into the community. Assessing IADL in people with schizophrenia is of utmost importance for clinicians and researchers to understand their functional status and design appropriate treatment plans.
There are two distinct constructs of IADL measures: ability and actual performance [4,5]. The concept of the ability construct refers to a person’s underlying capacity to perform IADL. Evaluating the ability construct provides information on a person’s inability or capability to perform IADL. Cognitive impairment in people with schizophrenia influences their ability to perform IADL and learn how to do IADL, highlighting the critical importance of understanding their capabilities and implementing tailored interventions [2]. On the other hand, the concept of the actual performance construct refers to what a person actually does in his/her daily environment. Evaluating the actual performance construct can help practitioners determine the degree of IADL independence in practical situations. Avolition, characterized by a decreased level of goal-directed activity resulting from reduced motivation, is a negative symptom observed in people with schizophrenia that could affect them actually performing IADL [6]. Without performing IADL, people with schizophrenia become dependent on others to manage their daily tasks and maintain their daily routines. Hence, ability and actual performance constructs demonstrate unique conceptual characteristics and practical clinical value.
To the best of our knowledge, currently there are no IADL measures that simultaneously assess these two IADL constructs. In previous studies, some IADL measures have been applied to people with schizophrenia, such as the lawton instrumental activities of daily living scale (LIADL), University of California San Diego Performance-based Skills Assessment (UPSA), and independent living skills survey (ILSS) [7-9]. The LIADL and UPSA assess the ability construct, whereas the ILSS assesses the actual performance construct [4,10,11]. Among these measures, the LIADL has been widely used to identify IADL functions at home and in the community [3,7,12] and has been translated into numerous languages [12-14]. The LIADL is easy to administer and has sound psychometric properties [15-17]. Thus, the LIADL has been adopted to assess IADL in long-term care services [5,18]. However, the LIADL does not assess the actual performance construct. Therefore, in this study, we aimed 1) to develop a new IADL measure with two constructs based on the LIADL: the instrumental activities of daily living scale-ability and actual performance (IADL-AA) and 2) to examine construct validity, convergent validity, discriminative validity, and internal consistency reliability of the IADL-AA in people with schizophrenia.
METHODS
Item development
The IADL-AA items were developed according to the eight items of the LIADL: using the phone, taking transportation, handling finances, doing laundry, doing household chores, preparing food, shopping, and taking medications [19]. The eight items of the LIADL were expanded into 39 items in the IADL-AA. For example, the LIADL item “preparing food” was expanded in the IADL-AA to include more detailed tasks such as purchasing meals, cooking independently, and cleaning up after meals. In total, 39 items were designed and reviewed and modified by eight experts, including psychiatrists, occupational therapists, and researchers in the fields of IADL and psychometrics. The experts were asked to provide feedback on the appropriateness of the item, wording, relevance of the content, and scoring. After consultation with experts, three items related to personal grooming (i.e., regular personal hygiene, maintaining a neat appearance, and dressing appropriately for outings) were deleted. During preliminary item generation, we considered including these items to capture daily functions that reflected IADL [20]. However, personal grooming was classified as basic activities of daily living, rather than IADL. Thus, 36 items were retained after expert review. The ability domain of the IADL-AA was assessed based on the respondents’ perceived capability to perform each IADL task during the past month. The ability domain was rated on a 4-point scale: 0, unable to perform; 1, able to complete with assistance; 2, able to complete with reminders; and 3, able to complete. The actual performance domain of the IADL-AA was assessed through self-report by asking how frequently each IADL task had been carried out over the past month. The actual performance domain was scored on a 3-point scale: 0, rarely (≤3 times a week); 1, occasionally (4-6 times a week); and 2, often (≥7 times a week). Cognitive interviews were conducted with 15 people with schizophrenia in each round to collect comments on the clarity and comprehensibility of the 36 items. In the cognitive interviews, participants were asked three types of questions: 1) What did you think this item meant?; 2) Was this item easy to understand? Were there any words you did not understand?; 3) What alternative words would make this item easier to understand? Two rounds of cognitive interviews were administered for the 36 items. In the first round, 33 of the 36 items (91%) were judged by participants as understandable. In the second round, participants did not raise any further concerns or suggestions for any of the items. Face validity was confirmed by ensuring the items appeared relevant and appropriate to assess daily functioning, while content validity was supported by experts’ evaluations of the comprehensiveness and accuracy of the items. The face validity and content validity of the self-reported IADL-AA were established.
Validation of the IADL-AA
Participants
People with schizophrenia, living in the community, were recruited between October 2018 and September 2020 from a psychiatric center in northern Taiwan. The inclusion criteria were as follows: 1) schizophrenia diagnosis according to the Diagnostic and Statistical Manual of Mental Disorders, 5th edition; 2) aged >20 years; 3) stable mental condition, as determined by the clinical judgment of occupational therapists who had long-term experience working with people with schizophrenia who continuously monitored their mental status in the psychiatric center; and 4) ability to follow instructions (Mini-Mental State Examination [MMSE] score ≥22). People with a history of brain injury and a diagnosis of substance abuse or intellectual disability were excluded. To obtain stable item parameter estimates in the Rasch analysis, the sample size needed to be at least 250 [21,22]. Taipei City Hospital Research Ethics Committee provided ethical approval (TCHIRB-10412114) for the study, and written informed consent was obtained from all participants.
Procedures
People with schizophrenia who met the eligibility criteria were administered the IADL-AA and Clinical Global Impressions-Severity scale (CGI-S). The assessments were performed in a quiet environment to minimize any potential disruptions that could have affected the performance of the participants.
Measure
The MMSE was used to assess the overall cognitive functions (e.g., orientation, attention, and memory) of the participants [23]. The total MMSE score ranged 0-30, with higher scores reflecting greater overall cognitive function. Satisfactory test-retest reliability of the MMSE has been demonstrated in people with schizophrenia [24].
The CGI-S was used to describe the severity of psychiatric symptoms among participants in this study. The CGI-S scores were ranked on a 7-point scale: 1, no illness; 2, borderline illness; 3, mild illness; 4, moderate illness; 5, markedly ill; 6, severe illness; and 7, extremely severe illness [25]. The CGI-S has shown adequate convergent validity in people with schizophrenia [26].
Data analysis
For construct validity, the Rasch rating scale model was applied to examine the unidimensionality of each domain in the IADL-AA using the WINSTEPS® Rasch Measurement Computer Program (Winsteps® Software). Infit mean square (MnSq) and outfit MnSq were used to verify whether the item responses fit the unidimensional model. Items with infit or outfit MnSqs of <0.60 or >1.40 demonstrated a misfit [27]. The misfit item was removed from the two domains, and Rasch analysis was reapplied until the items conformed to the standards of infit and outfit MnSqs.
Correlations between the two domains were calculated using Pearson’s r to investigate the convergent validity. The standards of convergent validity were 0.40≤ r <0.75 for moderate correlation and r ≥0.75 for strong correlation [27]. We hypothesized that these two domains would moderately correlate with each other. Floor and ceiling effects were determined to examine discriminative validity and calculated as the percentages of participants who achieved minimum and maximum scores in the two domains, respectively. The domain showed floor or ceiling effects when the minimum and maximum scores reached ≥15% of all participants for one domain, respectively [28]. Internal consistency reliability was examined using Cronbach’s alpha (α). The standard of reliability was α ≥0.70 for group-level comparisons and α ≥0.90 for individual comparisons [27].
RESULTS
Table 1 presents the demographic and clinical characteristics of the participants. In this study, 266 participants were recruited to complete the IADL-AA to strengthen the robustness of the Rasch model estimates. Their mean age was 44.2 years, and 45.5 of them were males. The mean scores for the MMSE and CGI-S were 26.1 and 3.4, respectively.
To examine unidimensionality, 13 items were removed because of a misfit in one or two domains. The 23 remaining items in the ability and actual performance domains conformed to the Rasch model expectations. Table 2 presents the results of the infit and outfit statistics. The infit and outfit MnSqs of the ability domain ranged 0.76-1.17 and 0.81-1.40, respectively. The infit and outfit MnSqs of the actual performance domain were 0.77-1.21 and 0.68-1.39, respectively. The 23 items that fitted the unidimensionality were classified into 7 occupations according to the Occupational Therapy Practice Framework: Domain and Process, Fourth edition [29]: communication management (4 items: operate a phone, look up unfamiliar phone numbers, view messages, and send messages), driving and community mobility (2 items: follow traffic rules and transfer between public transportation), financial management (3 items: perform daily transactions, handle financial matters, and process bills), home establishment and management (3 items: do laundry, do simple household chores and clean [e.g., sweeping, wiping surfaces, folding clothes, or taking out the trash], and do heavy household chores and clean [e,g., mopping floors, washing windows, or moving furniture]), meal preparation and cleanup (6 items: buy a meal to go, use delivery services, reheat prepared meals, prepare and cook meals, put away and wash dishes, and dispose of garbage), shopping (3 items: make a shopping list, buy basic daily necessities, and buy non-essential items), and medication management (2 items: do timely hospital follow-up and make doctor’s appointments) [29]. The total scores for the ability and actual performance domains ranged 0-69 and 0-46, respectively.
For convergent validity, the actual performance and ability domains showed a moderate correlation (r=0.63). For discriminative validity, there were no floor effects (0.0% and 6.0% for the ability and actual performance domains, respectively) or ceiling effects (3.0% and 1.5% for the ability and actual performance domains, respectively). The internal consistency of the ability and actual performance domains was 0.83 and 0.90, respectively.
DISCUSSION
In this study, we aimed to develop the IADL-AA, a self-report measure used to assess two constructs of IADL (i.e., activity and actual performance) in the past month. The uniqueness of this tool is that it can be used to simultaneously assess multidimensional IADL in seven occupations. Using this measure, clinicians and researchers can assess the capability and degree of independence to perform IADL in people with schizophrenia.
We verified the construct validity using Rasch analysis. Item response patterns of 13 items displayed misfit, such as “drive a car” and “operate an ATM machine.” These items involve heterogeneous skills that may extend beyond the IADL constructs. Driving a car depends on specialized motor skills and decision-making skills [30] and operating an ATM machine demands technical and device-specific competencies [31]. The 23 remaining items of each domain in the IADL-AA exhibited a unidimensional fit. The unidimensional results show that each domain captured a single construct, and the scores of the items in each domain could be aggregated to represent specific IADL functions. A higher score in each domain showed a better domain-specific IADL function. Thus, the scores of the ability and actual performance domains can be used to comprehend specific IADL functions and monitor the progress of IADL function in people with schizophrenia in clinical and research settings.
Convergent validity refers to the extent to which theoretically related constructs correlate with each other in practice [27]. The ability and actual performance domains revealed moderate correlations, supporting the theoretical hypothesis that these two domains assess related aspects of IADL function. No strong correlation was observed between the two domains, as the ability and actual performance domains represent distinct IADL constructs, indicating their non-interchangeability [32]. The IADL-AA has adequate convergent validity in people with schizophrenia.
Floor or ceiling effects were not observed in the ability or actual performance domains, demonstrating that clinicians and researchers can use these two domains to distinguish minimum and maximum scores of IADL function in people with schizophrenia. Regarding internal consistency, the ability domain was adequate to compare the assessment results between groups. The actual performance domain exhibited high reliability (α=0.90), essential for making individual comparisons of assessment results in clinical application [27]. The lack of noticeable floor and ceiling effects and satisfactory internal consistency improved the psychometric properties of the IADL-AA in people with schizophrenia.
The IADL-AA offers three advantages for applications in clinical and research settings. First, it is easy to administer and efficient; the duration required to complete IADL-AA was approximately 10-15 minutes. Second, the IADL-AA comprehensively assessed IADL in the two domains at the same time. Third, the IADL-AA includes a diverse range of IADL items-seven occupations, allowing the assessment of IADL function in a broad spectrum of daily activities. Therefore, the IADL-AA appears to be useful for facilitating the thorough and efficient capture of IADL functions in people with schizophrenia.
Nevertheless, this study has four limitations that need to be addressed. First, this study used a convenience sample from one psychiatric center, which restricts the generalizability of our results. Second, the medication compliance item was not included in the IADL-AA. The medication compliance item in the ability domain demonstrated satisfactory fit statistics. However, in the actual performance domain, the same item did not meet the model expectations (infit MnSq=1.73; outfit MnSq=1.90). Actual performance on medication compliance could be influenced by external factors such as supervision or access to medication. In this study, all participants were required to have a stable mental condition as one of the inclusion criteria, which implied adherence to prescribed medication. Moreover, clinical staff at the psychiatric center routinely monitored participants’ medication-taking behavior. This high level of supervision and adherence may contribute to the elevated infit and outfit MnSq values for the medication compliance item. Future research is warranted to recruit participants with varying levels of medication adherence to cross-validate our results. Third, people with schizophrenia who had severe cognitive impairment were excluded, potentially limiting the applicability of the findings to all populations with schizophrenia. However, the cognitive criterion is essential due to the requirement for self-reported IADL in the IADL-AA. Future research should investigate the feasibility of using proxies to complete IADL-AA for people with schizophrenia and severe cognitive impairment. Fourth, the test-retest reliability of the IADL-AA was not examined, potentially limiting the interpretability of the test results. Future research is needed to examine the test-retest reliability of the IADL-AA.
In summary, the IADL-AA is a new measure designed to assess two IADL domains: ability and actual performance. Our results supported the psychometric properties of IADL-AA, including face validity, content validity, construct validity, convergent validity, discriminative validity, and internal consistency. The IADL-AA can be useful to clinicians and researchers in assessing IADL function in people with schizophrenia.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
