The ADL taxonomy for persons with mental disorders – adaptation and evaluation

Abstract

Background: There is a lack of occupation-focused instruments to assess Activities of Daily Living (ADL) that are intended for persons with mental disorders. The ADL Taxonomy is an instrument that is widely-used within clinical practice for persons with physical impairment. The aim of this study was to adapt the ADL Taxonomy for persons with mental disorders and evaluate its validity.

Methods: An expert group of Occupational Therapists (OTs) from psychiatric care adapted the ADL Taxonomy to fit the client group, including creating three new items. OTs in psychiatric care collected client data and evaluated the instrument for usability. Rasch analysis was used to evaluate the contruct validity of 16 activities separately.

Results: The OTs collected 123 assessments from clients with various mental disorders. Ten activities had excellent, and four had acceptable, psychometric properties with regard to item and person fit and unidimensionality. The activity managing the day/time gave complex results and would benefit from further development. The OTs found the test version intelligible, relevant and easy to use.

Conclusions: The ADL Taxonomy for persons with mental disorders has 16 activities with three to six actions each, and is now ready for clinical use.

Keywords:

• Activities of daily living
• assessment
• occupational therapy
• psychometric properties
• rasch analysis

Introduction

In psychiatric care, persons with mental disorders such as schizophrenia, or neurodevelopmental disorders such as ADHD, are common client groups seen by occupational therapists (OTs). These clients often have limitations that seriously affect performance of activities of daily living (ADL). Some common problems are due to limitations in cognitive functioning such as time management, initiation, planning and organization; other common problems are related to social interactions and emotions; physical limitations affecting ADL are not as frequent [1,2].

An important part of occupational therapy practice is to assess clients’ performance of ADL, both to identify occupational gaps (discrepancy between what a person wants or needs to do and what he/she does) requiring intervention and to evaluate the outcome of occupational therapy [3,4]. To accomplish this, occupational therapists must have valid and reliable instruments. There is a scarcity of descriptive ADL instruments developed for persons with mental disorders that cover a wide range of ADL activities. Instead OTs within psychiatric care use a variety of instruments to assess parts, or aspects, of ADL performance. Examples include instruments to prioritize activities and set goals, such as the Canadian Occupational Performance Measure (COPM) [5], or to observe selected activities, such as Assessment of Motor and Process Skills (AMPS) [6] and Perceive, Recall, Plan, Perform (PRPP) [7]. One limitation with many of the instruments used is that they assess a small selection of activities and do not give a broad picture of a person’s ADL-performance.

In Scandinavia, a widely-used descriptive ADL instrument that covers many activities is the ADL Taxonomy [3,4,8]. The original version of the ADL Taxonomy was developed during the late 1990s with the aim to describe a person’s performance of ADL within 12 defined activities which in turn are constituted by actions [9,10]. The ADL Taxonomy is restricted to describe performance within the areas of personal ADL and instrumental ADL and does not incorporate activities within work and leisure [3]. Different adaptations of the original version have been documented. A children’s version, and a version for persons with impaired vision was developed early after the original version [11]. Clinical use of the ADL Taxonomy has revealed that the instrument fails to capture difficulties in ADL performance due to cognitive limitations such as in executive functioning including initiative, time management and planning and organization. Within psychiatric care in Sweden this has led to the development and use of several modified versions, more or less comparable to the original version, but all communicated as the ADL Taxonomy. To our knowledge none of them have been evaluated.

The original ADL-taxonomy has been psychometrically evaluated regarding the hierarchical order of the actions within each activity in three different samples, mainly with physical limitations, in a total of 1,283 persons [10]. Content validity has been confirmed by an expert panel [3]. The ADL Taxonomy has also shown good discriminatory validity [4,10] A four-grade effort scale to be used together with the original ADL-taxonomy was developed and used to evaluate interventions in persons with Parkinson’s disease [12] and severe psychiatric disability [13]. Waerens and Fischer [4] used Rasch analysis to evaluate the unidimensionality of the 47 actions in the ADL-taxonomy in a sample of persons with moderate to severe brain injury, using a three-category rating scale. The results showed that the actions formed a unidimensional scale if deleting 10 actions. However, to keep them did not disrupt the measurement properties and the authors recommended to keep them as they were clinically valuable [4]. This new scale has later been used for research purposes in two different versions called the ADL Interview and the ADL Questionnaire, both employing a linear scale in order to get a global outcome measure of ADL ability [4,14–16]. To conclude, earlier research has foremost focused on development of a new instrument based on the actions within the ADL Taxonomy to obtain a global measure of ADL ability and not of new activities or actions needed to enhance description of occupational performance in different client groups. Therefore, there is a need in occupational therapy in psychiatric care to develop and psychometrically evaluate a new version of the ADL-Taxonomy, where both activities and actions are adapted and expanded to capture and describe difficulties in ADL performance among persons with mental disorders. The aim of this study was therefore to adapt the ADL Taxonomy for persons with mental disorders and evaluate its validity.

Materials and methods

This study was conducted in two steps: (1) Development of an adapted test version of the ADL Taxonomy for persons with mental disorders, and (2) Evaluation of construct validity, including the hierarchical order of the actions within activities, as well as the intelligibility and the usability of the adapted test version.

The original ADL taxonomy

The original version of the ADL Taxonomy was developed to describe a person’s performance in personal and instrumental ADL. It has 12 defined activities organized into a number of goal-directed, purposeful and intentional actions. The activities are eating and drinking, mobility, going to toilet, dressing, personal hygiene, grooming, communication, transportation, cooking, shopping, cleaning and washing [10]. The number of actions in each activity varies from three to six and they are ordered from the easiest to the most demanding. The ADL Taxonomy can be administered in three ways: interview, observation and questionnaire. Each action is assessed as ‘does’ or ‘does not’. In addition the options ‘wants to’/‘doesn’t want to’ and ‘can’/‘cannot’ can be added to discriminate between what a person actually does and what he wants to do or can do [11]. The result of an ADL Taxonomy assessment is recorded either in a table form with boxes to tick or in a spider diagram (The ADL Taxonomy Circle) with the easiest actions in the center.

Development of the adapted version of the ADL taxonomy for persons with mental disorders

An expert group led by the first author (KH) was formed. The group consisted of four clinical occupational therapists (OTs) from different areas of Sweden, representing both municipal (n = 2) and county-based (n = 2) psychiatric care. They had all been suggested by the psychiatric committee of the Swedish Association of Occupational Therapists based on their clinical expertise and interest in professional development. To ensure that the adaptations would be in line with the basic foundations of the original version of the ADL Taxonomy, i.e. which activities were to be counted as ADL activities, the two original authors (Kristina Törnquist and Ulla Sonn) participated in part. The group met on three occasions over a period of three months.

The expert group reviewed all known modified versions of the ADL Taxonomy used within psychiatric care to identify activities or actions that had been added or removed, and reviewed two of the published versions of the ADL Taxonomy [10,11]. These versions gave insight into the activities and actions that were relevant for clients with mental disorders, and which were not included in the original version. The discussions resulted in retention of all activities in the original version, with some changes. The activity mobility was expanded into orientation and mobility. The activity communication was divided into two activities, (reading and writing and managing communication aids). Three new activities were added; interacting with others, handling economy and managing the day/time (see Table 1). The appropriateness of each action within the activities was scrutinized. In the original activities, the wording and the tentative hierarchical order between the actions were modified for clients with mental disorders. For the new activities, actions were formulated and placed in a preliminary hierarchical order.

Table 1. Task measures and fit statistics from the rasch analysis. The easiest task in each activity first.

Activity/task	Task measure	SE	Infit statistics
			MnSq	z-value
Eating and drinking (n = 123)
Drinking alone	−1.84	0.76	0.94	0.1
Eating and drinking together with others	0.58	0.42	1.07	0.6
Eating alone	1.27	0.42	0.96	−0.4
Orientation and mobility (n = 123)
In known indoor environment*	−4.52	1.02	0.94	0.2
In known outdoor environment	−0.26	0.58	0.94	0.0
In unknown environment	4.79	0.73	0.98	0.2
Dressing (n = 119)
Undress	−3.67	1.07	1.3	0.6
Dress	−2.28	0.68	0.93	0.0
Choose appropriate clothing	0.53	0.44	0.66	−1.5
Change to clean clothing	1.90	0.40	0.72	−1.5
Changing clothes in public spaces	3.51	0.46	1.37	1.8
Going to the toilet (n = 99)
Cleaning oneself after elimination	−1.71	1.13	1.18	0.5
Perform urinary and bowel elimination	−0.76	0.85	0.81	−0.1
Arranging clothes	0.72	0.61	0.87	−0.5
Washing hands	1.76	0.59	1.16	1.2
Personal hygiene (n = 119)
Washing oneself	−1.14	0.57	1.28	1.4
Bathing/shower	0.39	0.58	1.10	0.5
Washing hair	0.75	0.62	0.49	−1.9
Grooming (n = 119)
Shaving/make-up	−0.45	0.38	1.14	0.8
Manicuring	−0.06	0.36	0.71	−1.9
Brushing teeth**	−0.01	0.36	1.43	2.4
Combing one’s hair/fix the hairstyle	0.20	0.37	0.91	−0.5
Pedicuring	0.33	0.37	0.77	−1.6
Reading and writing (n = 120)
Writing signature	−3.50	0.63	0.93	0.0
Writing messages	−0.75	0.34	0.83	−0.8
Reading mail/messages	0.26	0.30	1.02	0.2
Using own notes	0.55	0.30	0.83	−1.2
Reading papers/books	1.11	0.29	1.22	1.7
Filling in forms	2.34	0.31	1.06	0.5
Handling communication aids (n = 121)
Managing computer/tablet	−1.53	0.65	0.65	−0.7
Managing phone/cell phone	−0.64	0.46	0.82	−0.6
Managing TV/radio	−0.41	0.44	0.84	−0.5
Answering phone calls	1.22	0.34	1.28	2.5
Making phone calls	1.36	0.33	1.02	0.2
Interaction with others (n = 123)
Interacting with few people	−3.16	0.40	1.03	0.2
Taking formal contacts	−1.10	0.29	0.98	−0.1
Interacting with many people	0.71	0.29	0.80	−1.4
Ending contacts/relations	0.93	0.31	1.01	0.1
Creating and maintaining intimate relationships	1.62	0.33	1.18	1.1
Handling economy (n = 121)
Withdraw money	−2.91	0.48	1.16	0.6
Paying bills	0.83	0.34	0.68	−2.0
Having control over own economy	2.08	0.38	0.98	0.0
Shopping (n = 120)
Daily or small quantity shopping in neighborhood shop	−3.05	0.53	1.26	0.7
Making plans for shopping	−0.04	0.36	1.04	0.3
Weekly or large quantity shopping	1.44	0.31	0.84	−1.2
Shopping online	1.65	0.33	0.78	−1.6
Cooking (n = 122)
Heating up liquid or prepared food	−2.67	0.46	0.86	−0.5
Preparing a cold meal	−2.10	0.42	0.84	−0.6
Washing the dishes and pick up	2.27	0.34	1.07	0.5
Cooking a hot meal	2.50	0.35	0.89	−0.7
Cleaning (n = 119)
Identifying need for cleaning	−1.24	0.32	0.96	−0.3
Sorting waste	−0.21	0.31	1.09	0.7
Daily light cleaning	−0.11	0.30	0.85	−1.2
Weekly heavy cleaning	0.76	0.31	1.19	1.3
Cleaning out	0.79	0.33	0.85	−0.9
Washing (n = 118)
Washing	1.66	0.65	1.14	0.5
Planning for washing	0.24	0.48	0.79	−1.3
Taking care of clean laundry	1.42	0.47	1.04	0.3
Transportation (n = 122)
Going by car	−3.38	0.55	1.22	0.7
Going by bus/tram/tube	−0.38	0.31	0.82	−1.1
Going by train/boat/aeroplane	0.70	0.29	0.64	−3.3
Riding bicycle/moped	1.41	0.31	1.00	0.0
Driving car/motorcycle	1.64	0.34	1.47	2.9
Managing the day/time (n = 123)
Getting up in time	−0.27	0.30	0.79	−1.8
Doing things in time/being on time	−0.19	0.29	0.88	−1.1
Getting to bed in time	0.11	0.29	0.95	−0.5
Estimating time consumption**	0.35	0.29	1.38	3.3

SE: Standard error; MnSq: Mean square of the residuals.

* All patients scored “can” on the task “In known indoor environment” in the Mobility and orientation activity.

** Misfitting task in the Rasch analysis.

The test version

The adapted test version of the ADL Taxonomy consisted of 16 activities with three to six actions in each activity. The activities were eating and drinking (4 actions), orientation and mobility (3 actions), dressing (5 actions), going to toilet (4 actions), personal hygiene (3 actions), grooming (5 actions), reading and writing (6 actions), handling communication aids (5 actions), interaction with others (5 actions), handling economy (3 actions), shopping (4 actions), cooking (4 actions), cleaning (5 actions), washing (3 actions), transportation (5 actions) and managing the day/time (4 actions).

The report form was designed as the original version, but with a written instruction not to leave any actions unassessed. The form also included space for comments next to each action as well as at the end of the form to enable collection of data on the OTs’ opinions on intelligibility, usability and suggestions for changes. In addition, questions on age, sex and diagnosis of the client were included. The test version could be administered by interview or observation. In order to ensure that the assessments were as complete as possible, self-rating was not used.

Evaluation of construct validity, intelligibility and usability

Participants and procedure

Data for evaluation of the hierarchical order of the actions within each activity was collected from clients with a psychiatric diagnosis and for whom an ADL assessment was clinically needed. Clients were consecutively recruited at psychiatric units (n = 19), four within municipality care (three different municipalities) and 15 within county council care (five different county councils). Within municipality care, the clients lived in ordinary housing or sheltered housing. Within County council care, eight units had outpatient care only and seven units had both inpatient and outpatient care. Data was collected by thirty-two occupational therapists OTs over a data collection period of three months. The OTs collecting data were the clients treating OTs and all clients were assessed as a part of their ordinary OT intervention. The OTs were experienced users of the original ADL Taxonomy. Eighty-one percent of the assessments were conducted using interviews, the remainder were conducted through observations.

All data was deidentified by the OTs before distribution to the researchers, meaning that all data that could connect the assessment to a specific client was deleted, thus no sensitive personal client data was handled by the researchers. Because collected data was deidentified and collected as part of the ordinary intervention procedure, no consent was obtained from the clients. Prior to data collection, ethical approval for the study, including the procedure for collecting deidentified data, was obtained from the Regional Ethical Review Board in Uppsala, Sweden (Dnr 2014/359).

To collect data on the content, intelligibility and usability of activities and actions, all OTs had the opportunity to write comments on the adapted test version. In each score form, the OTs commented on the relevance of the activities and actions, the intelligibility and usability of the test version as well as suggested changes.

Data analysis

To evaluate the construct validity, including the hierarchical order, of the actions within each activity, the Rasch measurement model was applied [17–19]. A separate Rasch analysis was run on each of the 16 activities. This means that the ADL Taxonomy was not treated as a common, unidimensional scale, but rather each activity was treated as a scale with its own items (actions). In line with the evaluation of the original version, only the dichotomous alternative ‘does’/‘does not’ were used for each action in the analysis [10]. Consequently the dichotomous model within Rasch analysis was used [17]. For each activity, a number of Rasch parameters were evaluated. Firstly, the action and person fit, which indicates the extent to which the actions and persons adhere to the Rasch model assumptions, was evaluated by examining the mean square of the residuals (MnSq) and the standardized z-value. We chose to use infit statistics because these are more sensitive for detecting misfit close to the ability level of the action than outfit statistics, and were thus considered more relevant [17]. A MnSq of ≤1.3 in combination with a z-value of ≥−2 and ≤2 was considered as fit to the model [20]. Further, the hierarchical order of actions was noted because this plays a significant role in the assessment and interpretation of the ADL Taxonomy. For each activity, the unidimensionality was evaluated by a principal components analysis of the residuals, as well as inspection of the dimensionality map. The criteria for unidimensionality was an Eigen value of <2 for the unexplained variance in the first contrast. In addition, targeting was examined in each activity as proportion of participants who reached maximum and minimum score. The Rasch analysis was conducted using the Winsteps Rasch Measurement computer program (version 3.92.1, John M Linacre 2016).

To evaluate the intelligibility and usability of the test version, the OTs comments on the test version were analyzed (n = 126) using manifest content analysis [21]. Meaning units were identified, condensed and categorized from their content, resulting in five categories describing intelligibility, usability and suggestions for changes.

Results

The results were based on assessments of 123 clients. The clients’ mean age was 36 years (range 18–73); 55% were men. They were diagnosed with a variety of mental disorders; the largest groups were neurodevelopmental disorders (22%) followed by psychosis (18%) and depression (10%). Many clients had secondary diagnoses, such as depression and anxiety. The main diagnoses of the clients are presented in Table 2. Diagnoses represented by only one client were classified as ‘Others’.

Table 2. Client characteristics (n = 123).

Client characteristic
Age, years
Mean (SD)	36 (13,7)
Range	18–73
Sex	Frequency (%)
Men	68 (55)
Women	55 (45)
Diagnosis
Psychosis	22 (18)
ADHD, ADD, Attention deficit	16 (13)
Depression	13 (11)
ASD, Asperger’s	11 (9)
Anxiety disorder	6 (5)
Bipolar disorder	6 (5)
Eating disorder	6 (5)
Drug abuse with mental illness	6 (5)
Others	21 (16)
Missing/unclassified	16 (13)

SD: standard deviation; ADHD: Attention Deficit Hyperactivity Disorder; ADD: Attention Deficit Disorder; ASD: Autism Spectrum Disorder.

Validity including hierarchical order of actions

The Rasch analysis showed that 11of the 16 activities fulfilled all the criteria i.e. had good action fit, principal components analysis (PCA) results with an Eigen value <2 (Tables 1 and 3). These activities were eating and drinking, dressing, personal hygiene, handling communication aids, interaction with others, handling economy, shopping, cooking, cleaning and washing. The hierarchical order between actions is shown in Table 1. In six activities, the results did not fulfill the criteria (Table 3). Three of them had misfitting actions. In the activity grooming one action showed misfit, ‘brushing teeth’, but other results from this activity were good. In order to improve the scale of grooming, the action ‘brushing teeth’ was temporarily removed and placed in the activity personal hygiene, however this did not improve the properties of any of the activity scales; the action was kept within grooming as it was considered highly clinically relevant. In the activity transportation, the action ‘driving car/motorcycle’ showed misfit, which might be due to a lack of data from 38 clients for this action. In the activity managing the day/time, the action ‘estimating time’ showed misfit, and the PCA showed that the activity is possibly not unidimensional. There was, however, no ceiling effect and the person fit was good for managing the day/time. Two activities, going to the toilet and reading and writing, showed slightly high PCA results, indicating non-unidimensionality; however, there were no misfitting actions in these activities (Table 1). The activity orientation and mobility had one action, ‘known environment indoors’, that showed a ceiling effect because all participants scored ‘does’. This action could therefore be considered to be redundant.

Table 3. Summary of rasch analysis results.

Activities	n	New or modified activity compared to original	Action misfit	PersonMisfit(%)	Principal components analysis Eigen value	TargetingProportion of participants with
						Max score	Min score
Eating and drinking	123	Modified tasks	No	1.6	1.8	75%	3%
Orientation and mobility	123	Modified	No	1.6	0.0	63%	<1%
Dressing	119	Modified tasks	No	2.5	1.8	43%	<1%
Going to the toilet*	99	Modified tasks	No	0	2.1	74%	1%
Personal hygiene	119	No	No	0	1.9	71%	15%
Grooming*	119	No	1 action (brushing teeth)	0	1.7	48%	5%
Reading and writing*	120	Modified	No	2.5	2.0	33%	2%
Handling communication aids	121	Modified	No	0.8	1.6	61%	<1%
Interaction with others	123	New	No	2.4	1.6	20%	5%
Handling economy	121	New	No	4.1	1.5	34%	6%
Shopping	120	Modified tasks	No	3.3	1.7	18%	6%
Cooking	122	Modified tasks	No	4.1	1.5	41%	5%
Cleaning	119	Modified tasks	No	1.0	1.6	19%	22%
Washing	118	Modified tasks	No	1.7	1.7	49%	48%
Transportation*	122	No	1 action (driving car/motorcycle)	3.3	1.8	13%	<1%
Managing the day/time*	123	New	1 action(estimating time)	0	2.4	33%	19%

* The activity does not fulfill all criteria in the Rasch analysis.

The OTs opinions on intelligibility, usability and suggestions for changes

Five categories of comments were identified: (i) Intelligibility and usability, (ii) suggestions for new actions within the activities, (iii) adding new occupational forms, (iv) suggestions for the manual and (v) comments regarding the response alternatives.

Intelligibility and usability. The majority found the test version intelligible, relevant and easy to use. It was considered a good screening instrument of ADL ability. Usability could differ depending on the client group within the psychiatric domain, i.e. those with severe mental illness within hospital care did not perform some of the actions.

Suggestions for new actions within the activities were mostly within the activity managing the day/time. The suggestions were in regard to circadian rhythm and sleep, and consequences related to limited executive functioning such as: adapt to unexpected events during performance, adapt time to circumstances, establish and maintain routines, use of memory aids and ability to use/follow instructions. There was also a suggestion to divide the action ‘doing things in time/being on time’ into two.

Suggestions to add new occupational forms were submitted. Most common were work and leisure; for example, physical activity. Other suggestions were not regarded as ADL but rather limitations of the client which the OTs found important to assess, such as stress and hyperactivity.

Suggestions for the manual were mainly in regard to clarifications such as wording in some definitions e.g. to add ‘regularly’ to pinpoint that some actions need to be performed on a daily basis.

Comments regarding the response alternatives related to two factors, 1) the difficulty in deciding when a action is classified as ‘does’, and suggestions for more detailed response alternatives such as ‘does sometimes’ or ‘does with assistance’, and 2) general comments on the difficulties of making assessments of this client group with fluctuating abilities on a day-to-day basis; one suggestion was to expand the comments fields to make more room for descriptions in running text.

The results from the Rasch analysis and the content analysis of the comments resulted in a decision by the expert group that the ADL Taxonomy adapted for persons with mental disorders should have 16 activities with three to six actions each. In this adapted version, three activities are identical to the original version, three new activities have been added and 11 activities have been modified. In seven of the 11 modified activities, only the actions were modified. The hierarchical order of the actions within each activity obtained from the Rasch analysis was used in the manual (see Appendix A). Based on the comments from the OTs some definitions were clarified in the manual. Based on these results, the first version of the ADL Taxonomy for clients with mental disorder was set and added to the manual including the other versions of the ADL Taxonomy [11].

Discussion

The aim of this study was to adapt and evaluate the ADL Taxonomy for persons with mental disorders. This new version is expanded with three new activities compared with the original version, all developed to capture areas within ADL that are difficult to manage for persons with mental disorders. In addition, one activity in the original version was divided into two. Overall, the OTs collecting the data found this new version intelligible, useful and easy to administer.

The Rasch analysis showed that most of the activities had good validity. There were, however, some activities that had a misfitting action. The activity grooming had an unclear hierarchical action order, and the action ‘brushing teeth’ showed misfit. This is in line with earlier studies showing similar results [4,10]. One possible explanation is that all the actions within grooming are not relevant for all persons or not necessary to perform [10]. Additionally, the actions seem to have roughly the same level of difficulty. These factors put the hierarchical order out of play which led to misfit of one action. However, the PCA results showed satisfactory unidimensionality and, thus, the misfit was not considered a threat to validity. Further, the activity grooming and its actions was highly clinically relevant and was therefore retained. A clinical implication is that users should not attach too much importance to the hierarchy of the actions within grooming.

Within managing the day/time the action ‘estimating time consumption’ showed misfit, and the activity on the whole showed a less optimal PCA result. This indicates that the four actions in this activity might measure more than one dimension. Two of the actions are quite specific (‘getting up in time’ and ‘getting to bed in time’) while the other two are more general, ‘doing things in time/being on time’, and the misfitting ‘estimating time consumption’. The dimensionality map showed that these two perspectives might be the reason for the poor action fit and PCA results. The activity managing the day/time is new, and there were many discussions by the expert group on what actions to include. The intention was to capture time-related actions that are necessary to manage the daily activities that clients want to do or are expected to do. This resulted in both general and specific actions that are usually difficult for clients within psychiatric care. The many suggestions from the OTs collecting data highlight the importance of capturing the activity level of time management, as it is an extremely important area that affects the ability to manage daily activities successfully. It is therefore important to retain this activity and to develop it further in a future version of the ADL Taxonomy for persons with mental disorders.

Most of the OTs’ suggestions concerned including additional aspects of executive functioning with consequences spanning a wide range of activities such as problems related to adapting to unexpected events during performance, adapting time to circumstances, and establishing and maintaining routines. In this version of the ADL Taxonomy some of the new actions added within activities other than managing the day/time concern occupational problems related to executive functions such as ‘planning the washing’ or ‘having control over own economy’. In further development of this version of the ADL Taxonomy, the possibility to add actions related to executive functioning to more activities should be considered. However, as in all assessment procedures, it is crucial that OTs analyze and interpret the results to detect patterns in ADL performance between activities, and actions which need further assessment, with more specific instruments. For example, for clients with poor results on the managing the day/time activity, further assessment of executive functioning is recommended using instruments such as Assessment of Time Management Skills [22] or the Weekly Calendar Planning Activity [23].

The targeting analysis showed that in five activities more than half the sample had a maximum score and overall there was a tendency towards a ceiling effect. These results are not surprising because each activity have few actions and the rating scale of does/does not is dichotomous and not sensitive to different levels of ability. In clinical practice, we suggest to use the additional rating scales ‘wants to’/‘doesn’t want to’ and ‘can’/‘cannot’ to obtain a better description of the clients’ ADL performance. The sensitivity to change in the dichotomous scale of ‘does/does not’ is likely to be poor which makes the instrument less useful in research and for evaluation purposes. One way to remedy this could be to add response alternatives. Also, the OTs’ comments showed a need for more differentiated response alternatives in the ADL Taxonomy. In three earlier studies, the original version of the ADL taxonomy has been adapted with graded response alternatives. Common to these response alternatives is that they are ordered from less able to more able [4,13,14]. To rate the degree of effort needed to manage an action or activity is relevant also for persons with mental disorders, because initiation and mental effort are often a hindrance for these clients. Lindström and colleagues adapted the ADL Taxonomy with a five-point effort scale [13], which was originally developed for persons with Parkinson’s disease [12], and used it for persons with mental disorders. In future research, it would be interesting to evaluate the ADL Taxonomy for persons with mental disorders with ordered response alternatives, such as an effort scale.

An important feature of the ADL Taxonomy in all its versions is that each activity has its separate hierarchy with reporting in a spider diagram with the easiest action in the center of the diagram (3). This method of reporting is very useful in a clinical context, and may have contributed to the extensive clinical use of the ADL Taxonomy in Sweden. In research it would, however, be useful to be able to summarize the results in a global outcome measure. This is possible to evaluate using Rasch analysis if all actions are merged into one scale measuring ADL as a whole, and the activities are dropped. This would most likely also remedy the poor targeting in the present scales. This has been done for the original version of the ADL Taxonomy and formed the basis for creating a new ADL measure, where all actions are ordered from the easiest to the most difficult independent of activity (4). Creating a similar unified ADL measure based on the ADL Taxonomy for mental disorder was not within the scope of this study but is an interesting topic for future study. However, the ADL Taxonomy is primarily known as a clinically useful descriptive instrument and there is a risk that a global measure, and the omission of the activities, decreases the possibility to identify and describe which actions that constitutes an activity and what actions that makes the performance of an activity difficult. Therefore, it is clinically important to keep the descriptive focus of the ADL Taxonomy.

Methodological considerations

The mix of different competencies in the expert group was a strength in the development process. To have both clinical and research experience in the group meant that the test version of the ADL Taxonomy for persons with mental disorders was developed with clinically relevant activities and actions, and structured in a psychometrically sound way. The external validity was further strengthened by recruiting from both county and municipality care, with representation from both outpatient- and inpatient care.

The great majority of the client assessments (81%) were performed using interview. This might have affected the results. Earlier studies have shown that clients have a tendency to overestimate their problems when being assessed through interviews compared to self-ratings or observations [14,24]. On the other hand, clients with limitations in executive functioning, which is common in some mental disorders, might have limited awareness of their own difficulties leading to the opposite, an overestimation of their ability [25]. The aim of this study was not to compare results from different methods of administrating the ADL Taxonomy, but this could be a valuable aspect of further research.

A limitation of this study is the rather small sample size. However, the present sample size is sufficient to obtain stable item calibrations within ±0,5 logits with 99% confidence [26]. The mean age of the clients assessed in this study was 36 years. This is quite young and could pose a limitation in regard to generalization. Persons who have lived with their mental illness for a long time, especially schizophrenia, have shown more cognitive impairments than younger persons (1) and this affects their ADL performance. It is possible that the hierarchy would have been slightly different in an older group of clients. On the other hand, the range in age (55 years) and spread in diagnoses was good, which strengthened the validity.

Clinical implications and future research

This study resulted in the first version of the ADL Taxonomy for persons with mental disorders. OTs within psychiatric care now have a common, valid version of the ADL Taxonomy ready for use in clinical practice [11]. OTs using it should take into account the limitations with hierarchical orders in complex actions such as those performed in ADL, especially in relation to the activities grooming and managing the time/day. In future research, this first version of the ADL Taxonomy can be further developed and evaluated in regard to actions related to executive functioning as well as the possibility of using ordered response alternatives.

Acknowledgements

Thank you to the expert group: Desirée Abrahamsson, Agneta Strid Johansson, Kristina Rönnholm and Catharina Melin who have contributed with their knowledge, time and strong engagement in the development process. Thank you also to the OTs who collected data and for so generously sharing their opinions on the test version, and to Tina Sandulf at the Swedish Association of Occupational Therapists for coordination and administrative support.

Disclosure statement

The authors report no conflicts of interest.

Additional information

Funding

The study was funded by Örebro University and the Swedish Association of Occupational Therapists.

Appendix A

Operational definitions of activities and actions included in the ADL taxonomy for persons with mental disorders

Table

ACTIVITIES	ACTIONS*
Eating and drinking Defined as eating and drinking alone, or together with others, with acceptable table manners	(1) Drinking alone  (2) Eating and drinking together with others  (3) Eating alone Observe eventual over or under consumption
Orientation and mobility Defined as goal-directed orientation and mobility from one place to another	(1) In known indoor environment  (2) In known outdoor environment  (3) In unknown environment
Dressing Defined as dressing and undressing, regularly changing to clean clothing, choosing appropriate clothing for the weather and activity, and to manage changing clothes in public places (such as fitting room, gym, bathhouse)	(1) Undress  (2) Dress  (3) Choose appropriate clothing  (4) Change to clean clothing  (5) Changing clothes in public spaces Observe eventual hypersensitivity to materials
Going to the toilet Defined as get to the toilet room in time, performing the necessary elimination, arranging clothes, and performing hygiene afterwards	(1) Cleaning oneself after elimination  (2) Perform urinary and bowel elimination  (3) Arranging clothes  (4) Washing hands
Personal hygiene Defined as regularly washing body and hair, using appropriate amount of soap/shampoo and wiping oneself dry	(1) Washing oneself  (2) Bathing/shower  (3) Washing hair
Grooming Defined as regularly manage and take care of one’s outward appearance, and to use necessary hygiene products (i.e. tooth paste) Manicuring and pedicuring means cutting or filing one’s nails	(1) Shaving/make-up  (2) Manicuring  (3) Brushing teeth  (4) Combing one’s hair/fix a hairstyle  (5) Pedicuring
Reading and writing Defined as reading papers/books, own mail and notes, and writing and using own written material (incl. Internet)	(1) Writing signature  (2) Writing messages  (3) Reading mail/messages  (4) Using own notes  (5) Reading papers/books  (6) Filling in forms
Handling communication aids Defined as communicating through communication aids, absorbing information, managing communication equipment and its software (TV, radio, telephone, mobile phone, computer/tablet) and to make calls (incl. finding actual phone number)	(1) Managing computer/tablet  (2) Managing phone/cell phone  (3) Managing TV/radio  (4) Answering phone calls  (5) Making phone calls
Interaction with others Defined as following social norms with regard to their own and other people’s integrity, to interact with few or many people, to take formal contacts, to create and maintain intimate relationships and when needed end contacts/relationships	(1) Interacting with few people  (2) Taking formal contacts  (3) Interacting with many people  (4) Ending contacts/relationships  (5) Creating and maintaining intimate relationships
Handling economy Defined as planning own economy, withdrawing money, paying bills and having control over own economy (meaning to plan and monitor so that there is enough money)	(1) Withdraw money  (2) Paying bills  (3) Having control over own economy
Shopping Defined as making plans for shopping (i.e. making shopping list), getting to the store, getting the groceries, paying for them and bringing them home, alternatively shopping online	(1) Daily or small quantity shopping in neighborhood shop  (2) Making plans for shopping  (3) Weekly or large quantity shopping  (4) Shopping online
Cooking Defined as planning, getting together utensils, preparation, cooking, laying the table, washing the dishes and tidying after meals	(1) Heating up liquid or prepared food  (2) Preparing a cold meal  (3) Washing the dishes and tidying  (4) Cooking a hot meal
Cleaning Defined as light cleaning (i.e. making the bed, clearing away/tiding up, wiping off, dusting) and heavy cleaning (i.e. vacuuming/washing floors, washing toilet and bathroom, linen change). Cleaning also includes identifying the need for cleaning, clearing out and getting rid of things that are not used, and waste sorting	(1) Identifying need for cleaning  (2) Sorting waste  (3) Daily light cleaning  (4) Weekly heavy cleaning  (5) Clearing out
Washing Defined as making plans for washing (identifying need, booking laundry time, remembering the time), washing up, hanging up, folding, or ironing/mangling and sorting the laundry. Transporting oneself and the laundry to and from the washing place is also included.	(1) Washing  (2) Planning for washing  (3) Taking care of clean laundry
Transportation Defined as getting to and from, in and out, and traveling with public or private transportation.	(1) Going by car  (2) Going by bus/tram/tube  (3) Going by train/boat/aeroplane  (4) Riding bicycle/moped  (5) Driving car/motorcycle
Managing the day/time Defined as planning/structuring the day to be able to perform the daily activities that one wants or expects to do	(1) Getting up in time  (2) Doing things in time/being on time  (3) Getting to bed in time  (4) Estimating time consumption

* The actions are ordered by degree of difficulty with the easiest as number one.