Impact of Urinary Incontinence on Quality of Life
Impact of Urinary Incontinence on Quality of Life
Shared understanding between affected individuals and support persons dealing with UI is helpful for support persons to sensitively comprehend and address needs in continence care and achieve optimal QOL for affected individuals. Other re-searchers have described that when misunderstandings on UI experiences exist between affected individuals and support persons, support persons are challenged to provide optimal care – often a key factor in the affected individual's admission to longterm care (Garcia et al., 2005; Jansen et al., 2013). The main purpose of this study was to examine affected individual and support person agreement on the impact of UI on affected individuals' QOL. The secondary purpose was to conduct the first known psychometric testing of a newly developed tool (the 22-item SPIQOL) for support persons to infer the impact of UI on affected individuals' QOL.
When comparing participant characteristics in this study with those in Higashi et al.'s (2005) comparative study, we found consistency in the demographic profiles of affected individuals and support persons. No known population-based Canadian statistics are available on characteristics of support persons of affected individuals dealing with UI. However, the prevalence of UI increases with age, thus impacting mainly individuals older than 65 years of age and females (Canadian Continence Founda tion, 2007) as represented in our study sample. Affected individuals in both our study and the Higashi et al. (2005) study were 75 years and older (mean age) and mainly females, and support persons were 67 years and older (mean age), female, and either the spouse or daughter of the affected individual.
Our hypothesis aimed to test whether agreement between individuals affected by UI and their support persons existed on the I-QOL and the SPI-QOL tool. Our cursory analysis of proportional differences in problematic I-QOL items (items that had ratings of 1 to 3) indicated that more support persons than affected individuals reported 14 of 22 I-QOL items were "a problem." This finding is supported by consistent findings in several early reviews of response comparability studies, where support persons or proxy respondents (generally family or friends) tended to over-report deficits, depending on the HRQOL domain (Lobchuk & Degner, 2002; McPherson & Addington-Hall, 2003; Sprangers & Aaronson, 1992; Tang & McCorkle, 2002). While these authors noted less bias to over-estimate by proxies or support persons on more physical, factual, concrete, or observable QOL items (versus invisible psychological items), support persons in our study tended to over-report on concrete behavioral items (e.g., "He or she has to be careful standing up after he or she has been sitting down because of his or her urinary problems or incontinence") and less visible or more psychological items (e.g., "He or she feels depressed because of his or her urinary problems or incontinence"), which is difficult to interpret.
While concordant ratings may seem to be a reasonable goal in the provision of quality continence care, the value of discrepant reports between affected individuals and support persons is worthy of further consideration, as suggested by Pickard et al. (2004). Disagreement may actually be important information, especially when support persons offer additional information that the affected individual either lacks awareness of, denies, or is reticent to divulge to health care providers (Pickard et al., 2004). For instance, affected individuals dealing with UI may underreport issues in managing bladder control due to embarrassment, whereas the support person provides additional information that may be a more accurate reflection of the patient's management issues and contributes toward more optimal continence care.
Of clinical importance, top-ranked problematic I-QOL items (based on a items that had a similar rank order of moderate to high proportions of affected individuals and support persons who rated I-QOL items from 1 to 3) included the lack of personal control over bladder continence, worries about wetting oneself, having a hard time getting a good night of sleep, worries about not getting to the toilet in time, and having to make frequent trips to the bathroom. These problematic I-QOL items are corroborated in previous literature where, in particular, personal control over bladder continence is a consistently cited issue for affected individuals (Hayder & Schnepp, 2010; Nichols & Layton, 2012; Seshan & Muliira, 2014) and support persons (Cassells & Watt, 2003; Gallagher & Pierce, 2002; Hayder & Schnepp, 2008). Of note, spouses and adult children comprised the majority of this study's sample, and most reported they saw the affected individual on a daily basis. These types of close relationships appear to make spouses and adult children reasonable complementary re sources who can reliably report on highly problematic QOL items that are either observable behaviors they witness frequently on a daily basis (i.e., making frequent trips to the bathroom), or persistent and profound psychological experiences that affected individuals openly discuss with their support persons (i.e., worries about not getting to the toilet in time or wetting oneself). Clinicians need to pay close attention to highly problematic IQOL items similarly described by affected individuals and support persons that can guide their provision of person-centered continence care.
On the other hand, "worries about having sex" was the lowest ranked problem I-QOL item for affected individuals and support persons, which is difficult to interpret without further information from participants. Limited literature on the negative impact of UI on sexual intimacy describes that the loss of bladder control and associated smells can be influential factors (Nichols & Layton, 2012; Seshan & Muliira, 2014). Cassells and Watt (2003) described how older spousal support persons engaged in other forms of sexual behavior, such as kissing and hugging, to meet their intimacy needs. Further studies on the impact of UI on sexual behaviors and ability to meet intimacy needs are warranted, particularly for older individuals.
Although our study sample was too small to engage in a robust analysis of proportional differences on individual I-QOL items, support persons' mean ratings were not significantly different from affected individuals' mean ratings on total scale and sub-scale scores, and ICC correlations ranged from being modest to strong. Further, this study's differences of paired mean scores ranged from -6.7 to -0.1, which were less than 10% of the instrument's theoretical range of scores (0 to 100 units) or 10 points; we did not consider these to be clinically significant differences in participants' responses on the total scale and three subscales (Osoba et al., 1998; Ringash et al., 2007). Regardless, as reported in related studies on response comparability across varied non-urologic medical conditions (Dinglas, Gifford, Husain, Colantuoni, & Needham, 2013; Gundy & Aaronson, 2008; Jones et al., 2011; Oczkowski & O'Donnell, 2010), support persons in our study tended to have a non-significant response bias to over-estimate the impact of UI on avoidance and limiting behavior and psychosocial impact. Overall, a caveat is warranted due to this study's small sample size and wide 95% CIs that contribute toward less precision in reporting differences in paired mean ratings. Our findings of non-significant paired mean differences and moderate to strong ICC correlations, however, look promising if not conclusive for support persons' inferences to serve as reasonable "proxy" or "complementary" indicators for how UI impacts the QOL of affected individuals. Larger studies on levels of agreement between affected individuals and their support persons are warranted to provide more robust analyses on whether statistically and clinically significant differences exist between affected individual and support person I-QOL ratings.
Interestingly, some of our results corroborate those by Higashi et al. (2005) who also studied response comparability on UI-QOL. Higashi et al. (2005) conducted comparative analyses involving the distribution of differences on only 10 of the 22 IQOL items (i.e., they excluded 12 items that they felt represented "internal" experiences of the affected individual which could be difficult for the proxy to infer). Their modified tool was named the Proxy Urinary Incontinence survey that had a Cronbach's alpha of 0.86 (Higashi et al., 2005). Higashi et al. (2005) found in their sub-sample of 24 caregiving pairs that support persons significantly under-estimated the impact of UI on affected individuals' QOL as captured on their total 10-item scale. Similarly, we found that support persons tended to underestimate the impact of UI on the total 22-item SPI-QOL (albeit nonsignificant with a paired mean difference of -0.1), as well as on the social embarrassment sub-scale (also non-significant with a paired mean difference of -6.7). While Higashi et al. (2005) reported an ICC correlation of 0.52 on their total UI scale (10 items), we reported 0.74 (22 items) on the total SPIQOL scale. Higashi et al. (2005) speculated that their sample of proxies underestimated the impact of UI on affected individuals due to their reluctance to talk about the UI. However, more than 50% of affected individuals and support persons in our study reported that they engaged in open "talk" about UI that might have contributed in part toward a more strong ICC result on total I-QOL. In Shaw, McColl, and Bond's (2001) study, the concordance of patient and proxy responses on the use of incontinence pads revealed a substantial Kappa rating of 0.69, and the occurrence of UI had a low Kappa rating of 0.49. Moreover, the Kappa rating was slight to moderate on the frequency and timing of UI that they attributed to their small sample of 10; there was no systematic bias by proxies to under- or over-report on UI and a tendency for support persons who had closer contact with the affected individual to respond more similarly (Shaw et al, 2001).
Overall, it is difficult to compare findings from related response comparability studies that involved a range of urologic and non-urologic conditions, em ployed different scales, and conducted varied tests of inferential analysis. Across findings arising from this small study and related UI studies (that also comprised primarily small samples), inconsistencies remain on support person tendencies to over- or under-estimate and the magnitude of correlation between dyadic responses on UI QOL items. Nonetheless, our pilot findings are generally consistent with findings in related response comparability studies where better levels of agreement on HRQOL items occur when relatives serve as support persons and have daily contact with affected individuals. Given the growing expectations for informal supports and affected individuals to jointly manage chronic conditions like UI in the home, continued examination of perceptual agreement over time in larger studies are warranted. Such studies also need to compare the influence of relative versus nonrelative roles and relationship quality factors (e.g., degree of open, empathic communication) on levels of agreement to aid in designing interventions that boost the capacity of support persons to manage UI.
We engaged in the first psychometric evaluation of a new tool (the SPI-QOL), and therefore, could not conduct comparisons of the SPI-QOL with other studies. However, the SPI-QOL total scale and three sub-scales demonstrated promising results based on acceptable internal reliability consistency coefficients and strong item-to-scale correlations within this study's support person sub- sample. The SPI-QOL appeared to be acceptable in prompting support persons to provide comparable responses to affected individuals on the total scale and subscales, which supports the concurrent validity of the SPI-QOL. Overall, our analysis provides beginning evidence for the 22-item SPI-QOL as a valid, reliable, and feasible measure to capture either proxy or complementary reports by support persons who provide continence care; ongoing psychometric analyses in larger studies is recommended.
Discussion
Shared understanding between affected individuals and support persons dealing with UI is helpful for support persons to sensitively comprehend and address needs in continence care and achieve optimal QOL for affected individuals. Other re-searchers have described that when misunderstandings on UI experiences exist between affected individuals and support persons, support persons are challenged to provide optimal care – often a key factor in the affected individual's admission to longterm care (Garcia et al., 2005; Jansen et al., 2013). The main purpose of this study was to examine affected individual and support person agreement on the impact of UI on affected individuals' QOL. The secondary purpose was to conduct the first known psychometric testing of a newly developed tool (the 22-item SPIQOL) for support persons to infer the impact of UI on affected individuals' QOL.
When comparing participant characteristics in this study with those in Higashi et al.'s (2005) comparative study, we found consistency in the demographic profiles of affected individuals and support persons. No known population-based Canadian statistics are available on characteristics of support persons of affected individuals dealing with UI. However, the prevalence of UI increases with age, thus impacting mainly individuals older than 65 years of age and females (Canadian Continence Founda tion, 2007) as represented in our study sample. Affected individuals in both our study and the Higashi et al. (2005) study were 75 years and older (mean age) and mainly females, and support persons were 67 years and older (mean age), female, and either the spouse or daughter of the affected individual.
Our hypothesis aimed to test whether agreement between individuals affected by UI and their support persons existed on the I-QOL and the SPI-QOL tool. Our cursory analysis of proportional differences in problematic I-QOL items (items that had ratings of 1 to 3) indicated that more support persons than affected individuals reported 14 of 22 I-QOL items were "a problem." This finding is supported by consistent findings in several early reviews of response comparability studies, where support persons or proxy respondents (generally family or friends) tended to over-report deficits, depending on the HRQOL domain (Lobchuk & Degner, 2002; McPherson & Addington-Hall, 2003; Sprangers & Aaronson, 1992; Tang & McCorkle, 2002). While these authors noted less bias to over-estimate by proxies or support persons on more physical, factual, concrete, or observable QOL items (versus invisible psychological items), support persons in our study tended to over-report on concrete behavioral items (e.g., "He or she has to be careful standing up after he or she has been sitting down because of his or her urinary problems or incontinence") and less visible or more psychological items (e.g., "He or she feels depressed because of his or her urinary problems or incontinence"), which is difficult to interpret.
While concordant ratings may seem to be a reasonable goal in the provision of quality continence care, the value of discrepant reports between affected individuals and support persons is worthy of further consideration, as suggested by Pickard et al. (2004). Disagreement may actually be important information, especially when support persons offer additional information that the affected individual either lacks awareness of, denies, or is reticent to divulge to health care providers (Pickard et al., 2004). For instance, affected individuals dealing with UI may underreport issues in managing bladder control due to embarrassment, whereas the support person provides additional information that may be a more accurate reflection of the patient's management issues and contributes toward more optimal continence care.
Of clinical importance, top-ranked problematic I-QOL items (based on a items that had a similar rank order of moderate to high proportions of affected individuals and support persons who rated I-QOL items from 1 to 3) included the lack of personal control over bladder continence, worries about wetting oneself, having a hard time getting a good night of sleep, worries about not getting to the toilet in time, and having to make frequent trips to the bathroom. These problematic I-QOL items are corroborated in previous literature where, in particular, personal control over bladder continence is a consistently cited issue for affected individuals (Hayder & Schnepp, 2010; Nichols & Layton, 2012; Seshan & Muliira, 2014) and support persons (Cassells & Watt, 2003; Gallagher & Pierce, 2002; Hayder & Schnepp, 2008). Of note, spouses and adult children comprised the majority of this study's sample, and most reported they saw the affected individual on a daily basis. These types of close relationships appear to make spouses and adult children reasonable complementary re sources who can reliably report on highly problematic QOL items that are either observable behaviors they witness frequently on a daily basis (i.e., making frequent trips to the bathroom), or persistent and profound psychological experiences that affected individuals openly discuss with their support persons (i.e., worries about not getting to the toilet in time or wetting oneself). Clinicians need to pay close attention to highly problematic IQOL items similarly described by affected individuals and support persons that can guide their provision of person-centered continence care.
On the other hand, "worries about having sex" was the lowest ranked problem I-QOL item for affected individuals and support persons, which is difficult to interpret without further information from participants. Limited literature on the negative impact of UI on sexual intimacy describes that the loss of bladder control and associated smells can be influential factors (Nichols & Layton, 2012; Seshan & Muliira, 2014). Cassells and Watt (2003) described how older spousal support persons engaged in other forms of sexual behavior, such as kissing and hugging, to meet their intimacy needs. Further studies on the impact of UI on sexual behaviors and ability to meet intimacy needs are warranted, particularly for older individuals.
Although our study sample was too small to engage in a robust analysis of proportional differences on individual I-QOL items, support persons' mean ratings were not significantly different from affected individuals' mean ratings on total scale and sub-scale scores, and ICC correlations ranged from being modest to strong. Further, this study's differences of paired mean scores ranged from -6.7 to -0.1, which were less than 10% of the instrument's theoretical range of scores (0 to 100 units) or 10 points; we did not consider these to be clinically significant differences in participants' responses on the total scale and three subscales (Osoba et al., 1998; Ringash et al., 2007). Regardless, as reported in related studies on response comparability across varied non-urologic medical conditions (Dinglas, Gifford, Husain, Colantuoni, & Needham, 2013; Gundy & Aaronson, 2008; Jones et al., 2011; Oczkowski & O'Donnell, 2010), support persons in our study tended to have a non-significant response bias to over-estimate the impact of UI on avoidance and limiting behavior and psychosocial impact. Overall, a caveat is warranted due to this study's small sample size and wide 95% CIs that contribute toward less precision in reporting differences in paired mean ratings. Our findings of non-significant paired mean differences and moderate to strong ICC correlations, however, look promising if not conclusive for support persons' inferences to serve as reasonable "proxy" or "complementary" indicators for how UI impacts the QOL of affected individuals. Larger studies on levels of agreement between affected individuals and their support persons are warranted to provide more robust analyses on whether statistically and clinically significant differences exist between affected individual and support person I-QOL ratings.
Interestingly, some of our results corroborate those by Higashi et al. (2005) who also studied response comparability on UI-QOL. Higashi et al. (2005) conducted comparative analyses involving the distribution of differences on only 10 of the 22 IQOL items (i.e., they excluded 12 items that they felt represented "internal" experiences of the affected individual which could be difficult for the proxy to infer). Their modified tool was named the Proxy Urinary Incontinence survey that had a Cronbach's alpha of 0.86 (Higashi et al., 2005). Higashi et al. (2005) found in their sub-sample of 24 caregiving pairs that support persons significantly under-estimated the impact of UI on affected individuals' QOL as captured on their total 10-item scale. Similarly, we found that support persons tended to underestimate the impact of UI on the total 22-item SPI-QOL (albeit nonsignificant with a paired mean difference of -0.1), as well as on the social embarrassment sub-scale (also non-significant with a paired mean difference of -6.7). While Higashi et al. (2005) reported an ICC correlation of 0.52 on their total UI scale (10 items), we reported 0.74 (22 items) on the total SPIQOL scale. Higashi et al. (2005) speculated that their sample of proxies underestimated the impact of UI on affected individuals due to their reluctance to talk about the UI. However, more than 50% of affected individuals and support persons in our study reported that they engaged in open "talk" about UI that might have contributed in part toward a more strong ICC result on total I-QOL. In Shaw, McColl, and Bond's (2001) study, the concordance of patient and proxy responses on the use of incontinence pads revealed a substantial Kappa rating of 0.69, and the occurrence of UI had a low Kappa rating of 0.49. Moreover, the Kappa rating was slight to moderate on the frequency and timing of UI that they attributed to their small sample of 10; there was no systematic bias by proxies to under- or over-report on UI and a tendency for support persons who had closer contact with the affected individual to respond more similarly (Shaw et al, 2001).
Overall, it is difficult to compare findings from related response comparability studies that involved a range of urologic and non-urologic conditions, em ployed different scales, and conducted varied tests of inferential analysis. Across findings arising from this small study and related UI studies (that also comprised primarily small samples), inconsistencies remain on support person tendencies to over- or under-estimate and the magnitude of correlation between dyadic responses on UI QOL items. Nonetheless, our pilot findings are generally consistent with findings in related response comparability studies where better levels of agreement on HRQOL items occur when relatives serve as support persons and have daily contact with affected individuals. Given the growing expectations for informal supports and affected individuals to jointly manage chronic conditions like UI in the home, continued examination of perceptual agreement over time in larger studies are warranted. Such studies also need to compare the influence of relative versus nonrelative roles and relationship quality factors (e.g., degree of open, empathic communication) on levels of agreement to aid in designing interventions that boost the capacity of support persons to manage UI.
We engaged in the first psychometric evaluation of a new tool (the SPI-QOL), and therefore, could not conduct comparisons of the SPI-QOL with other studies. However, the SPI-QOL total scale and three sub-scales demonstrated promising results based on acceptable internal reliability consistency coefficients and strong item-to-scale correlations within this study's support person sub- sample. The SPI-QOL appeared to be acceptable in prompting support persons to provide comparable responses to affected individuals on the total scale and subscales, which supports the concurrent validity of the SPI-QOL. Overall, our analysis provides beginning evidence for the 22-item SPI-QOL as a valid, reliable, and feasible measure to capture either proxy or complementary reports by support persons who provide continence care; ongoing psychometric analyses in larger studies is recommended.
Source...