Introduction
Mild cognitive impairment (MCI) is a heterogeneous clinical syndrome that lies between the cognitive functioning in normal aging and the early stage of dementia (Petersen et al., Reference Petersen, Caracciolo, Brayne, Gauthier, Jelic and Fratiglioni2014). Core clinical criteria in former and current major definitions of MCI include roughly preserved independence in functional abilities despite some objectively verified cognitive impairment and self- or informant-reported cognitive symptoms, thus not dementia (Petersen, Reference Petersen2004; Winblad et al., Reference Winblad2004; American Psychiatric Association, 2013). This description overlaps with criteria intended to describe preclinical Alzheimer's disease (AD) (Albert et al., Reference Albert2011). To date, no effective treatment of MCI is available (Kane et al., Reference Kane2017). Nevertheless, early detection of morbid cognitive and functional decline remains important; in the current situation in order to enable adequate and targeted support to handle the consequences of such declines in everyday life, and in the future to identify the optimal therapeutic window (Belleville et al., Reference Belleville, Fouquet, Duchesne, Collins and Hudin2014). In clinical practice, efficient and valid functional markers are needed that capture the subtle cognitive and functional decline in MCI. Impaired ability to manage finances and medications (Arrighi et al., Reference Arrighi, Gélinas, Mclaughlin, Buchanan and Gauthier2013) and withdrawal from participating in outings and other leisure activities (Arrighi et al., Reference Arrighi, Gélinas, Mclaughlin, Buchanan and Gauthier2013; Hedman et al., Reference Hedman, Nygård, Malinowsky, Almkvist and Kottorp2016) have been suggested as early functional signs of cognitive change. In addition, changes in the ability to use everyday technologies (ETs), i.e. the wide range of technical objects and services that are commonly present and used in our everyday lives at home and in the community, have proven to be one such potential marker (Malinowsky et al., Reference Malinowsky, Almkvist, Kottorp and Nygård2010; Nygård et al., Reference Nygård, Pantzar, Uppgard and Kottorp2012; Malinowsky et al., Reference Malinowsky2017). ET use can, for example, be studied by assessing the ability of individuals to use specific ETs or by focusing on the challenge level of specific ETs as perceived by users. This prospective five-year study explored whether the challenge level of certain ETs is more sensitive and well suited to detecting signs of functional change in MCI.
Cognitive ability has been shown to affect how many types of ETs people use (Czaja et al., Reference Czaja2006) and how efficient this use is (Slegers et al., Reference Slegers, Van Boxtel and Jolles2009). On a group level, both perceived (Nygård et al., Reference Nygård, Pantzar, Uppgard and Kottorp2012) and observed (Malinowsky et al., Reference Malinowsky, Almkvist, Kottorp and Nygård2010) ability to use ETs are significantly lower in persons with MCI compared to in persons with no known cognitive impairment, and even lower in persons with dementia. Ability in ET use as well as amount of ETs used have shown a potential to differentiate groups with different needs of assistance among persons with cognitive impairment (Ryd et al., Reference Ryd, Nygård, Malinowsky, Öhman and Kottorp2016). Furthermore, a declining pattern regarding the ability to use ETs and activity involvement during the years following the detection of MCI might be predictive of future dementia (Hedman et al., Reference Hedman, Kottorp and Nygård2017a).
The cognitive decline in MCI includes impaired memory related to recollective/associative abilities and short-term retention, impaired executive functions in new or demanding tasks, and reduced attentional control (Belleville et al., Reference Belleville, Fouquet, Duchesne, Collins and Hudin2014), which together with visuo-spatial function all exemplify cognitive abilities needed when using ETs such as coffee makers, cell phones, and cash machines (ATMs). Performance skills, i.e. observable actions when performing everyday activities, often involved when managing ETs and known to be, especially challenging are choosing the correct button or command, identifying the services or functions of the ET, and performing actions in a logical sequence (Malinowsky et al., Reference Malinowsky, Nygård and Kottorp2011). We also know that certain ET characteristics make their use more challenging for persons with and without cognitive impairments. ETs demanding a high frequency of performance skill actions, requiring use of more difficult performance skills, failing to provide feedback related to a variety of sensory functions (e.g. visual, auditory, and tactile), having complex design, and typically being infrequently used, pose higher levels of challenge on the users (Patomella et al., Reference Patomella, Kottorp, Malinowsky and Nygård2011, Reference Patomella, Kottorp and Nygård2013). Thus, complexity in process might be the common denominator challenging both cognition and ET use. Cross-sectional research has suggested that domestic ETs such as coffee makers, stoves, and microwave ovens are generally perceived and observed to be easier to use, while information and communication technologies (ICT) like cell phones and computers are more challenging (Patomella et al., Reference Patomella, Kottorp, Malinowsky and Nygård2011; Malinowsky et al., Reference Malinowsky, Kottorp and Nygård2013). However, longitudinal studies examining how the perceived challenge of specific ETs develops in persons with cognitive decline are lacking. Identifying ETs whose challenge level has the potential to detect functional change might be highly interesting for ensuring targeted support at an early stage. Therefore, we examined how older adults perceived the challenge levels of commonly used ETs at the time of MCI detection and five years later, with the aim to investigate whether the use of certain ETs is more sensitive to cognitive decline.
Methods
Sample
Participants were older adults with MCI diagnosed and recruited at a specialized outpatient memory clinic in Stockholm between April 2008 and May 2009 (Table 1). The inclusion criteria were: (a) fulfilled criteria for MCI as proposed by (Petersen, Reference Petersen2004), i.e. self-rated and clinically verified cognitive decline, no dementia, and essentially intact basic and instrumental activities of daily living; (b) at least 55 years old; (c) being a user of ETs; (d) the ability to take part in data collection in Swedish; (e) no cognitive comorbidities; and (f) no severe problems with hearing or vision that could not be compensated for. The Regional Ethics Committee in Stockholm approved the study, and informed consent was obtained from all participants. Figure 1 provides an overview of the sampling and of the data that were available and missing at the seven follow-up occasions over the five-year period. As the inclusion ended 2009, the final 5-year-follow-ups took place in 2014.
Participant characteristics at inclusion (n = 37)
Note: Mean and SD are presented for normally distributed data, while median and IQR are given for skewed data. ETUQ: Everyday Technology Use Questionnaire, where higher person measure indicates higher perceived ability in ET use (Nygård et al., Reference Nygård, Pantzar, Uppgard and Kottorp2012). MMSE: Mini-Mental State Examination (0–30), where higher score indicates better cognitive status (Folstein et al., Reference Folstein, Folstein and Mchugh1975).
Overview of sampling, data available and missing, and diagnostic information at the seven follow-up occasions.
Data collection
In the present study, we used baseline and five-year data because a previous study using the same sample (Hedman et al., Reference Hedman, Nygard and Kottorp2017b) had shown a significant decrease in perceived ability to use ETs only when comparing year five and baseline, and not between consecutive time points. The first author and five research assistants conducted the structured face-to-face interviews, most often in the homes of the participants. Those who so wished were accompanied by a significant other for support, but we based the scoring on the participants’ answers. To capture longitudinal changes in the perceived level of challenge of a range of ETs of varying difficulty, the Everyday Technology Use Questionnaire (ETUQ) was used (Rosenberg et al., Reference Rosenberg, Nygård and Kottorp2009). The ETUQ includes 92 ETs commonly used by older adults inside and outside the home. Each ET perceived as relevant by the person is scored on a 6-step scale, rating the person's perceived ability to use it (Table 2). In the ETUQ, a relevant ET is defined as an ET that the person has access to and has used in the past, currently is using, or intends to start using. In this study, both person measures and item measures were used (further explained in the data analysis section). The ETUQ has shown acceptable psychometric properties (unidimensionality, rating scale validity, and person response validity) in studies including older adults with cognitive impairments (Rosenberg et al., Reference Rosenberg, Nygård and Kottorp2009; Nygård et al., Reference Nygård, Pantzar, Uppgard and Kottorp2012). Additionally, cognitive function was assessed using the Mini-Mental State Examination (MMSE) (Folstein et al., Reference Folstein, Folstein and Mchugh1975), and we obtained diagnostic information from medical files at the memory clinic.
Description of the scale steps in the ETUQ
Data analysis
We used a Rasch model to transform the ordinal ETUQ scores into relative linear measures expressed in log odds probability units (logits) (Bond and Fox, Reference Bond and Fox2007). This is a suitable approach when analyzing ETUQ data because Rasch models can handle the typical scenario that not all ETUQ items are scored as relevant by all participants. That is, the response patterns of all the participants are used by the model to generate individual person measures, reflecting the perceived ability in ET use, as well as item measures, reflecting the relative challenge of each ET. These measures are displayed in two different hierarchy outputs, one placing the participants on a continuum from lower to higher perceived ability to use ETs, and another one placing the ETs on a continuum from being perceived as less to more challenging to use. A higher person measure indicates higher perceived ability in ET use for a person, while a higher item measure indicates a higher level of perceived challenge for an ET. In this study, special focus was placed on the item measures because our interest was to explore whether the challenge level of certain ETs was more sensitive to cognitive decline. However, we ensured stability and validity of the person measures by securing proper rating scale functioning according to standard procedures described earlier (Nygård et al., Reference Nygård, Pantzar, Uppgard and Kottorp2012).
For an ET to be included in the analysis, at least ten participants should have scored it as relevant both at baseline and at year five. This was true for 46 of the 92 items in the ETUQ (50%). In order to detect whether specific ETs presented more, or less, challenge over five-years’ time than expected based upon the Rasch model, we examined differential item functioning (DIF) in relation to time. To adjust for changes in the mean person measure of the sample, actual DIF (Petersson et al., Reference Petersson, Lilja, Hammel and Kottorp2008) between baseline and year five was calculated for each ET through the use of standardized z-comparisons calculated on the individual item measures from both of these time points and their individual standard errors. We used a z-score difference of ≥ ±1.96 as the criterion for significant actual DIF, and an item measure of 50.0 logits was used as a cut-off to classify the items as “relatively more easy” or “relatively more challenging” to use at baseline. An item measure of 50.0 logits represents an item where a person with a similar person measure (i.e. 50.0 logits) has an equal chance of scoring a 3 (The ET is sometimes/partly used together with another person) or a 4 (The ET is used with frequent/major difficulties), i.e. the cut-off between using the ET with or without another person. To explore potential patterns in the types of ETs that presented significant actual DIF, the analyzed ETs were also classified into the seven topic areas in which all ETs are organized in the latest ETUQ version (Nygård et al., Reference Nygård, Rosenberg and Kottorp2016) – home maintenance, information/communication, self-care, maintenance/repair, accessibility, economy/purchasing, and travel.
Results
At year five, 21 of the 37 included participants contributed with data, which gives a dropout rate of 43%. An independent-sample t-test revealed no significant baseline differences regarding age, MMSE, or ETUQ person measure between those who contributed with data at year five and those who had dropped out, and a χ 2 test indicated no significant relation between gender and attrition. The mean perceived ability to use ETs was 54.6 (SD 2.8) logits at baseline and 52.2 (SD 3.1) logits at year five. These person measures were used when adjusting the standardized z-score differences examining the actual DIF of the included ETs.
In Table 3, ET item measures of challenge at baseline and year five, standardized z-score differences, and item types of the 46 included ETs are presented. In total, seven (15%) of the analyzed ETs had positive z-score differences that were outside the ±1.96 interval, i.e. they showed significant actual DIF over time. The positive scores indicate that these ETs were perceived to become more challenging to use over time. For 39 (85%) of the ETs, no significant actual DIF was found, indicating overall significantly stable perceived challenge level over time for these ETs despite overall decreasing ETUQ person measures in the sample. A total of 30 of the 46 analyzed ETs (65%) were classified as “relatively more easy” to use at baseline, while 16 (35%) were classified as “relatively more challenging” to use. At year five, 33 ETs (72%) were classified as “relatively more easy”, while 13 (28%) were classified as “relatively more challenging.” Among the 30 ETs classified as “relatively more easy” to use at baseline, five (17%) – washing machine, TV with remote control, hand-held mixer, internet interaction, and cash machine (ATM) – showed significant actual DIF, indicating that these became more challenging to use over time. Among the 16 ETs classified as “relatively more challenging” to use at baseline, two (13%) – stereo/CD player and digital camera – showed significant actual DIF, which reflects that they became even more challenging to use over time.
Item measures, standardized z-score differences examining actual DIF, and ET types for the included ETs, here ordered according to their challenge level at baseline with a cut-off for “relatively more easy” versus “relatively more difficult” to use at 50.0 logits
Note: Positive z-score difference reflects increasing difficulty of the ET item over time. Negative z-score difference reflects decreasing difficulty of the ET item over time. ETs in bold print reach or exceed the ≥ ±1.96 cut-off for significant actual DIF.
As can be seen in Table 3, ET item types within the ETUQ topic areas of self-care, maintenance/repair, and travel are not represented in the analysis due to a small number of scorings.
Discussion
Our findings show that 15% of the analyzed ETs were significantly more challenging to use at year five compared to at baseline. Based on chance, we would have expected that just over two of the 46 ETs (5%) would demonstrate significant DIF. The findings exceeded this number. However, the majority of the ETs (85%) were perceived to be within a similar challenge level five years after inclusion, despite the fact that the participants’ perceived overall ability to use ETs had decreased and the fact that the diagnostic composition in the sample had changed from 100% MCI to 62% dementia. These findings deserve some reflection.
It is difficult to distinguish common characteristics among the seven ETs that showed significant DIF in this study, in relation to the ETs that did not demonstrate DIF over time. Some of the seven ETs that became significantly more challenging to use could be linked to activities known to reveal early functional signs of cognitive change. For example, perceiving the cash machine (ATM) as more challenging to use is in line with earlier findings of impaired ability to manage finances (Arrighi et al., Reference Arrighi, Gélinas, Mclaughlin, Buchanan and Gauthier2013), and the increased challenge level in using ETs like TV with remote control, hand-held mixer, internet interaction, stereo/CD player, and digital camera might be linked to the known early withdrawal from leisure activities in cognitive decline (Arrighi et al., Reference Arrighi, Gélinas, Mclaughlin, Buchanan and Gauthier2013; Hedman et al., Reference Hedman, Nygård, Malinowsky, Almkvist and Kottorp2016). However, several similar ETs typically used in similar activities were also found among the ETs with stable challenge level, for example, internet banking, receiver of digital TV, and internet information search. Similarly, explanations linked to the complexity of the activities where the ETs were used, or the complexity of the actions involved in the ET use (Malinowsky et al., Reference Malinowsky, Nygård and Kottorp2011), were not applicable.
The generated hierarchy displaying the ETs on a continuum from less to more challenging to use (Table 3) reaffirms earlier knowledge of domestic ETs as generally being easier to use (Patomella et al., Reference Patomella, Kottorp, Malinowsky and Nygård2011; Malinowsky et al., Reference Malinowsky, Kottorp and Nygård2013) because all but one ET within home maintenance were found among the “relatively more easy” ETs. Somewhat unexpectedly, a larger proportion of ETs classified as “relatively more easy” to use at baseline became significantly more challenging to use at year five (17%) compared to the proportion of ETs with significant DIF that at baseline were classified as “relatively more challenging” (13%). At baseline, the participants likely managed to use the washing machine, TV with remote control, hand-held mixer, internet interaction, and cash machine (ATM) without the support of another person, albeit with varying degrees of difficulty. By year five, however, several of these ETs had passed over to the “relatively more challenging” side of the 50.0 logit cut-off, thus becoming more likely for the persons to use with another person's support. This finding has at least two clinical implications. First, in order to detect possible functional decline it might be appropriate to ask persons with MCI whether there are ETs that they previously could manage by themselves that now require support from another person. Second, it is relevant to also attend to less challenging ETs within the area of home maintenance and not focus only on high-tech ETs within the areas of information/communication and economy/purchasing because home maintenance involved ETs with significant DIF.
Our finding of a predominantly stable challenge level of ETs over time corresponds in part with a previous study comparing the challenge level of ETs at two time periods three to five years apart in two different samples, including persons with MCI, persons with dementia, and controls (Malinowsky et al., Reference Malinowsky, Kottorp, Patomella, Rosenberg and Nygård2015). Also, in that study, stable challenge levels were most common with 85% stable ETs. However, contrary to the findings in our study, all but one of the ETs showing a significant change in the level of challenge were perceived as easier to use at the later time period in the previous study. This difference might be due to the fact that in the previous study there were different individuals included in the different time periods and the fact that the study also included older adults without cognitive impairments. The direction of the findings in the present study indicates that our underlying assumption might be correct; the challenge levels of certain ETs seem to be more sensitive to cognitive change and might thereby be a useful indicator of functional change in MCI. However, we are reluctant to highlight the specific ETs that showed significant DIF in this study as being, especially sensitive to cognitive decline because no clear patterns regarding their characteristics in comparison to the stable items could be found. We instead propose a comprehensive approach when evaluating the challenge levels of ETs over time in older adults with MCI and dementia, including both easier and more challenging ETs and including domestic ETs as well as ICT.
Given the lack of earlier longitudinal research that explores ET use on the item level in older adults with cognitive decline, this DIF study was undertaken despite the risk of being underpowered. The findings should therefore be interpreted with caution. As a consequence of the large attrition rate of 43%, the sets of persons generating the item measures differ between baseline and year five. However, the dropout analysis shows that the attrition did not differ at baseline from those who contributed with data at year five regarding MMSE, ability to use ETs, gender, or age, which suggests that the generated item calibrations at year five would have been similar without the attrition. Finally, it is important to remember that changes in challenge levels of specific ETs may occur for reasons other than functional change in the participants. For example, updating or replacement of technologies can also contribute to changing challenge levels. Although people with MCI generally are reluctant to replacing ETs with new ones (Nygård, Reference Nygård2008), it is possible that some ETs used by the participants were altered or replaced during the five years. For instance, new features could have made a cell phone easier or more challenging to use. Thus, ET use needs to be interpreted from several perspectives, not only based on the person's characteristics and diagnosis.
In conclusion, the findings of this study suggest that change of perceived challenge of ETs may indicate functional change in persons with cognitive decline, even if alternative explanations cannot be ruled out. Both easier and more challenging ETs typically used at home and in society need to be addressed to capture this functional change because significant changes may occur among ETs of all challenge levels and within all types of ETs in older people with MCI.
Conflict of interest
None.
Description of authors’ roles
All authors formulated the research question in consensus. Hedman took part in data collection, performed the statistical analysis, and wrote the paper. Kottorp was responsible for the statistical design of the study, gave advice during the statistical analysis, and provided input to the draft. Almkvist gave advice regarding the statistical analysis and provided input to the draft. Nygård planned the 5-year study, supervised the data collection, and assisted in writing the paper.
Acknowledgments
We thank the participants for sharing their experiences of ET use with us and the occupational therapists who helped us with recruitment and data collection. The following funders made the project possible: the Swedish Council for Health, Working Life and Welfare (FORTE); the Swedish Research Council (VR); the Strategic Research Programme in Care Sciences at the Karolinska Institutet; and the regional agreement on medical training and clinical research (ALF) between Stockholm County Council and the Karolinska Institutet.
Open access
