OBBLIGHI DI INFORMAZIONE
6. Le modifiche unilaterali al contratto quadro alla luce della PSD2
6.2. Le modifiche dei contratti stipulati con consumatori
The primary outcome measure of TCAS is quantitative: the number of falls using a falls calendar (see Appendix 8). Fall rates have not been widely used among Tai Chi studies because the primary outcome was not fall-related. Generally, however, when measuring falls, the calendar method is the most commonly used (Li et al., 2004; Logghe et al., 2009;
Taylor et al., 2012; Taylor-Piliae et al., 2014a; Tousignant et al., 2013), along with telephone calls to reduce self-reporting bias (Logghe et al., 2009; Taylor et al., 2012;
Tousignant et al., 2013). A systematic review of methods of measuring falls in randomised controlled fall prevention trials by Hauer, Lamb, Jorstad, Todd & Becker (2006) found eight primary collection methods: prospective registration, calendar, prospective registration, patient diary, prospective registration, postcard, recall, face to face interview, recall, postal questionnaire, recall, telephone, nursing home fall records and hospital fall or health
records. Routine health records have limited quality and availability in different settings. In the community, they have little significance because only less than 20 per cent of falls are reported by patients (Hauer et al. 2016). According to Hauer et al. (2016), prospective systems are superior, with falls diaries leading to an increased number of reported falls.
However, the authors found the three systems of prospective reporting to be difficult to determine, and may be under or over reported, often having a back-up retrospective recall, such as telephone calls. Thus, the authors acknowledge that retrospective recall may
introduce retrospective recall error into the data. Fall calendars with follow-up telephone calls for non-responders have been chosen as the method to record the number of falls in this study, thus minimising data error. Additionally, fall rates are the most direct measure to assess whether the number of falls is reducing, and obtaining fall rates from records is not suitable for TCAS because this study is exploring the number of falls following discharge from hospital during the intervention period.
Outcome measure for balance
To assess balance improvement as an outcome, numerous measures for balance have been developed (See Table 17). Because the TCAS study focuses on dynamic balance
(maintaining balance whilst in motion), balance assessment tools measuring this type of balance were evaluated. Dynamic balance assessment tools evaluated included the BBS, Dynamic Gait Index, EquiTest, Limit of Stability Test, Romberg Test of Balance,
Timed-Up-The SPPB is a brief performance battery designed for older adults to assess balance, gait speed and lower body strength (Guralnik et al., 1994). It involves aspects which are not being investigated by the TCAS study, e.g. gait and lower body strength.
The EquiTest and Limit of Stability Test [LoS] measure balance through computed dynamic posturography, which is expensive technical equipment. This equipment also requires special training, thus making these tests unsuitable for this study due to limited financial resources.
The Romberg Test and The Dynamic Gait Index focus on aspects other than balance caused by limb weakness. For example, the Romberg Test is used to investigate the cause of loss of motor coordination and is used more in conjunction with the vestibular system and vision rather than limb weakness. The Dynamic Gait Index assesses gait, balance and fall risk.
However, in stroke it has moderate responsiveness in depicting change at two months and five months after treatment. The Limits of Stability Test requires the use of force plates which are not available for the TCAS study.
The Timed-Up-And-Go [TUG] test, a tool used to assess a person’s mobility based on timing sitting to standing and walking, can be used as a simple measure of balance comparable to the BBS (Bennie et al., 2003). However, most studies in Table 15 have used TUG for general mobility. Although the TUG only takes five minutes, lessening the burden on
participants compared with the BBS taking fifteen minutes to perform, the BBS is already in use as part of a routine assessment in the local community where the TCAS study is
located. Harada, Chiu & Damron-Rodriguez, (1995) found that the BBS was a more valid measurement of balance than other clinical laboratory balance tests, and Stevenson and Garland (1996) found the BBS to have excellent reliability in elderly stroke survivors.
The TCAS study adopted the BBS (see Appendix 9), a 14-item/five-point ordinal scale which takes approximately 25 minutes and is easy to administer, requiring minimal equipment such as a step, stopwatch, ruler and a chair (Stevenson, 2001). The scale aims to
quantitatively assess balance and risk for falls in older people living in the community by performing tasks which are graded from zero to four. The total possible score adds up to 56 with higher scores reflecting better balance. The BBS can predict falls in the elderly by demonstrating scores less than 45 out of 56 (Bogle, Thorbahn & Newton, 1996). As well as assessing both dynamic and static aspects of balance (Blum & Korner-Bitensky, 2008), the BBS is also useful as a screening tool to identify if stroke survivors require physical therapy (Stevenson, 2001). Physiotherapists based at the study site use the BBS to assess stroke survivors who have been referred to community physiotherapy. Therefore, it would be easy to obtain a baseline BBS score for the Tai Chi study, but the purposes of a study, a measure
needs to be also adequate at examining the effectiveness of an intervention. Berg, Wood-Dauphinee, Williams, and Gayton (1989) designed the scale to provide a means to
determine change in balance ability over time, and, indeed the BBS has been administered before and after the intervention (Stevenson, 2001). It has been shown that the BBS has been responsive to change in stroke survivors at two, six and12 weeks post-onset (Wood-Dauphinee et al., 1999), with moderate to excellent sensitivity to change in the early post-stroke period (Blum & Korner-Bitensky, 2008). Establishing change and whether an
intervention may be effective or not is more difficult than predicting falls and assessing balance. It is not clear how much change in score is needed to claim that an intervention has been effective, and what constitutes an improvement in balance may vary between assessors. Therefore, some agreement is required. Stevenson (2001) suggests that the clinical interpretation of the before and after Berg Balance scores requires a change score of five BBS points or more to be 90 per cent confident that an individual’s BBS performance has changed when assessed by two different raters. For example, an initial Berg Balance score of 40 would have to increase to 46 to show an effect in the intervention. However, Blum and Korner-Bitensky (2008) suggest exercising caution when measuring change in stroke survivors who have mild balance impairment but did not state why. It is likely that stroke survivors with mild balance impairment may not improve balance because they may not show an improvement in balance because there is limited scope for improvement. A minimum change score of five was considered when analysing data from the BBS in this study to conclude that genuine change has occurred in participants.
One consideration with the BBS used with stroke survivors is the performance of the final two tasks where standing on one leg and standing in tandem are required. Stroke survivors with hemiparesis may find these tasks challenging. Kwong (2015) found that selecting the paretic leg as the primary weight-bearing leg in these two tasks resulted in significantly lower BBS scores in stroke survivors. The author suggests standardising the chosen leg for these two tasks. Therefore, in Kwong’s study, all participants were instructed to weight bear on the non-paretic leg to reduce the risk of falling. This means that the BBS score may potentially be even lower if the paretic leg was used.
Despite strong psychometric properties for assessing balance, previous studies with stroke survivors have demonstrated large floor and ceiling effects post-stroke (Blum & Korner-Bitensky, 2008). The authors thus recommended the use of other balance measures in conjunction with the BBS to address this, so possible additional measures were considered.
Only one stroke study discussed in Chapter 2 has used the BBS to measure dynamic balance; outcome measures varied greatly among studies, making it difficult to conclude which is the most appropriate.
Excellent between- and within-rater reliability for the BBS has been reported by Berg, Wood-Dauphinee and Williams (1995), as well as excellent test-retest and interrater reliability among chronic stroke (Blum & Korner-Bitensky, 2008). Interrater reliability was established for the TCAS, the results of which can be found in Appendix 10. For the TCAS study, baseline assessments were conducted by the community physiotherapists and 12-week BBS scores were conducted by the researcher. The researcher did not conduct the initial BBS because to do so would introduce researcher bias when the researcher conducts the 12-week scores. Additionally, both assessors need to be achieving the same results when calculating the scores. Therefore, it is important that both assessors at baseline and 12-weeks are assessing using the same parameters. In order to gain inter-rater reliability for the BBS, the researcher was taught how to perform the BBS by a senior physiotherapist based at the hospital. We performed three BBS tests on the same three patients without seeing what each other had scored. When the scores were compared, it was found that scores were similar. The setting and equipment used was the same equipment for the same 12-week assessments. The letter to confirm inter-rater reliability has been established is found in Appendix 11.
Table 17 Quantitative outcome measures related to balance used in previous Tai Chi stroke studies
Outcome measure Purpose Reference
Berg Balance Scale Dynamic balance Hart et al. (2004)
Gateview Static balance Kim et al. (2015)
Sensory
Organization Test Static balance Au-Yeung et al.
(2009)
Dynamic Gait Index Dynamic balance Kim et al. (2015)
EquiTest Dynamic balance Au-Yeung et al.
(2009) Limit of Stability
Test
Dynamic balance Au-Yeung et al.
(2009) Romberg Test of
Balance Dynamic balance Hart et al. (2004)
10min Walking Test Gait Kim et al. (2015)
2min Step Test Aerobic endurance Taylor-Piliae et al.
(2014) Timed-up-And-Go
Test [TUG] Gait Kim et al. (2015)
Mobility Au-Yeung et al.
(2009)
lower-body strength Taylor-Piliae et al.
(2014)
Outcome measure used for fear of falling
Few Tai Chi research studies involving stroke survivors have included self-efficacy as part of their outcome measures. Four popular outcome measures used by researchers for this are
the Activities-Specific Balance Confidence Scale [ABC], the FES, the Modified Falls Efficacy Scale [MFES] and the CONFbal Scale. The ABC and CONFbal Scale focus on confidence in performing an activity without losing one’s balance. The TCAS study’s primary outcome is falls. Therefore, confidence in performing an activity without falling would be more suitable to the study’s aims.
The FES is a reliable and validated ten-item questionnaire intended to be used for community and hospital patients with brain injury, multiple sclerosis, spinal cord injury, stroke and the general elderly (Dewan & MacDermid, 2014). The FES assesses fear of falling, recommended for older people living in the community and is a self-reported questionnaire which requires individuals to rate from one to ten how confident they are at performing certain activities without falling, regardless of whether they perform these activities (Dewan & MacDermid, 2014). Total scores range from ten (most confident) to 100 (least confident and greatest fear of falling). There is also a modified version of the FES with an additional four tasks. Due to the nature of a feasibility study, it was deemed unnecessary to expect participants to perform these four extra tasks. Therefore, the original 10-point FES was used in the TCAS study (see Appendix 12).
Outcome measure used for depression
There are many tests available to measure depression, such as the Beck Depression
Inventory, the Patient Health Questionnaire (PH2 and PH9), the GDS, the Zung Self-Rating Depression Scale, the Center for Epidemiological Studies Depression Scale [CES-D], the General Health Questionnaire, the Hospital Anxiety and Depression Scale [HADs] and the Hamilton Depression Rating Scale.
The 20-question CES-D was designed to measure the severity of depressive symptoms in the general population and is widely used in research as a screening instrument and has been used in the Tai Chi study by Taylor-Piliae et al. (2014a). However, it’s robustness and suitability of the commonly used four-factor 20-item CES-D model has been called into question. Carleton et al. (2013) investigated this and concluded that results should be interpreted with caution. For example, item 17, for example used crying as a severity of depression and this could not be accurately interpreted as a sign of depression.
The Hospital Anxiety and Depression Scale [HADS] was developed to detect states of depression who were treated for clinical problems (Zigmond & Snaith, 1983) but has been used among non-hospital-based patients with success (McDowell, 2006). It was not originally designed to be a clinical diagnostic tool (Whelan-Goodinson, Ponsford &
Schönberger, 2009) and has been found to perform as well as the Beck Depression Inventory and the General Health Questionnaire instruments (Mykletun & Stordal, 2001).
The Beck Depression Inventory is a 21-question multiple choice self-report tool designed to measure severity of depression, whereas the GDS is a 30-item self-report ‘yes’ or ‘no’ tool, aimed at older people to diagnose depression. Out of these two depression tools, the GDS was preferred because questions only required circling a ‘yes’ or ‘no’ to one response, as opposed to choosing one of four responses. The GDS scores result in a categorisation into normal, medium or severely depressed, and this should suffice for the TCAS study.
The Geriatric Depression Scale [GDS] has been shown to be useful in younger stroke survivors as well as older stroke survivors with minor depression and has demonstrated internal consistency and test-retest reliability (Sivrioglu et al., 2009). The GDS is a self-rating 30-item screening tool for depression developed for use in geriatric patients.
Questions refer to how one felt over the last week and can detect changes over time. There is also a shorter 15-item GDS but Chau and Mao (2006) found that although it was suitable to detect post-stroke depression, the 30-item version had stronger psychometric
characteristics in the stroke population. The GDS has been validated against the Hamilton Rating Scale for Depression and the Zung Self-Rating Depression Scale and was found to have a 92 per cent sensitivity and an 89 per cent specificity when evaluated against diagnostic criteria (Yesavage & Brink, 1983). The GDS was the chosen measurement tool for depression in the TCAS study because it has a ‘yes’ or ‘no’ answering system and can detect changes over time (see Appendix 13).
Outcome measure for quality of life [QoL]
The 11- item Patient Health Questionnaire asks questions relating to the last four weeks which stroke survivors may find difficult to recall. Similarly, the 12-item General Health Questionnaire requires recall over the last few weeks.
To measure QoL, there are numerous measuring tools, such as the SF 36, SF 12, Quality of Life Scale [Qols], WHOQoL-BREF, EuroQol, HRQoL, and the Duke Health Profile. The QoLs is designed for chronic conditions, intended for group assessment as opposed to individual patient assessment (Buckhardt & Anderson, 2003) and would be a good tool to use for the TCAS study. Another suitable tool is the HRQoL. The HRQoL is a four-item questionnaire with good retest reliability, validity and responsiveness (Yin, Njai, Barker, Siegel & Liao, 2016). An alternative tool to use is the Duke Health Profile, as used by Hart et al. (2004).
This tool is a 17-item questionnaire which assesses six health measures and four dysfunction measures. According to Vahedi (2010), the WHOQoL-BREF could be further improved with more research needed to increase measurement precision at the high-end of the scale. Though moderately reliable, Vahedi (2010) suggested this tool be used to assess moderate levels of quality of life.
The SF 12 questionnaire is a shorter version of the SF 36 questionnaire. The SF 36 questionnaire is a generic measure of health status that has been validated in stroke survivors and includes both physical and mental component summaries, known as PCS and MCS respectively (Pickard, Johnson, Penn, Lau & Noseworthy, 1999). The average time to fill in the SF 36 by the general population has been known to be 10 to 12 minutes but takes longer for stroke survivors; researchers using the SF 36 have found it to be burdensome for stroke survivors, resulting in missing data (Pickard et al., 1999). By using the SF 12,
missing data would be less problematic and thus increase the efficiency of the study. It is also quicker to complete, with the general population completing in two minutes. It is expected that stroke survivors would take longer but this is still far less time-consuming than the SF 36 (Pickard et al., 1999).
One disadvantage of using the SF 12 is the less-precise estimate of individual health and inability to calculate summary scores when one item has not been answered. Pickard et al.
(1999) suggest that the SF 12 may not be suited to evaluate changes over time because it only includes one third of the items contained in the SF 36. The authors analysed the SF 12 and SF 36 to determine the extent to which the summary scores of the SF 12 replicate the SF 36. Strong agreement was found between the two questionnaires, affirming that the SF 12 replicates the SF 36 summary scores without substantial loss of information. However, other measures in conjunction with the SF 12 were recommended by the authors.
No previous stroke studies as discussed in Chapter 2 have used the SF 12. The choice of outcome measure for QoL vary, making it difficult to conclude about the most appropriate one. The SF 12 was chosen for the TCAS study because of the little time it takes for stroke survivors to fill it in. It has been argued by some researchers that adopting a qualitative approach would explore QoL more accurately. Therefore, quality of life will also be
addressed by conducting an interview which will explore participants’ perceptions of QoL to enrich the data from the SF 12.
Initially, the EuroQOL questionnaire was considered, but after assessing its suitability compared with the SF 12, it was decided that this questionnaire is better utilised as part of the economic appraisal of health programmes and their incorporation into health technology assessments (see Appendix 14).
The Falls Efficacy Scale, Geriatric Depression Scale and the SF 12 were piloted using nine elderly Tai Chi practitioners who attended a local Tai Chi class. Nobody expressed any difficulties with reading or understanding the questionnaires.