Abstract
Disabilities and handicaps affect health-related quality-of-life (HRQOL) and dysphoria symptoms (helpless, hopeless, worthless, dissatisfied with life, depressed, suicidal). Archived 2009 Canadian Community Health Survey (CCHS) data (N=124,188) replicated 49 positive correlations (p<.001) of seven kinds of disability measured by the Health Utilities Index (HUI) with seven measures of dysphoria. Cognition and pain disabilities appeared most dysphoric. Individuals (N=724) in the 2009 CCHS data with HUI total scores less than 0.00 are defined by HUI protocols as “worse than dead.” This HUI categorization has doubtful validity based on 17 measures of disability, suicidality, self-perceived health, and social engagement.
Experts in the field suggest that a suicidal person is feeling so much pain that they can see no other option. They feel that they are a burden to others, and in desperation see death as a way to escape their overwhelming pain and anguish. The suicidal state of mind has been described as constricted, filled with a sense of self-hatred, rejection, and hopelessness.
Physical disabilities have psychological concomitants. There are many kinds of disabilities, and combinations of disabilities, which can cause handicaps resulting in negative psychological outcomes. For example, motorically disabled people can become frustrated, angered, or depressed from the loss of mobility and thereby loss of activities. Social isolation can also arise from hearing and vision disabilities and from cognitive disabilities, again, with negative psychological outcomes (Morrison, 2008). Individuals with disabilities also experience increased barriers to obtaining health care as a result of accessibility concerns, such as transportation and problems with communication (Drainoni et al., 2006). Common effects of a disability may include mental health issues such as anxiety and depression, loss of freedom and independence, frustration and anger at having to rely on other people, loss of self-esteem and confidence, especially in social situations (London Health, 2014).
Effects of disabilities
Research on the effects of disabilities on psychological distress and suicidality has shown that several disabilities increase psychological distress and increase suicidality in patients. Pain seems to be the most studied variable of disability linked to increased psychological distress regarding suicidal thought and suicidal ideation, showing that chronic pain increases the risk of suicide attempts (Stenagera et al., 2014) and that illicit drug use as a form for pain relief also increases suicidality (Racine et al., 2014). There has also been research investigating the relationship between disability, impairment, and depression (Prince et al., 1997). The more general significance of poor physical health on suicidal ideation has also been established (Fairweather et al., 2006; Hawton and Fagg, 1988).
The health utilities index
Health economists’ health-related quality of life (HRQoL) is usually quantified by “preference-based” measures, on the presumption that less disability due to illness allows patients more choice in their life activities and thus greater ability to satisfy their preferences (Gamst-Klaussen and Rudmin, 2014).
One widely used measure of HRQoL is the Health Utilities Index (HUI). It is a multi-attribute health-status classification system, designed to provide a framework with which to describe consequences of health problems that involve impaired abilities. The HUI is a family of generic disability profiles for the purpose of measuring and for producing utility scores.
The HUI, mark 3 version (HUI3), utilizes eight sub-scales (vision, hearing, speech, ambulation, dexterity, emotion, cognition, and pain), which are identifiers of disability states, to compute a total HRQoL score using a proprietary mathematical formula (Horsman et al., 2003). Weighting of the sub-scales in the HUI is based on statistical modeling of responses from the general population who are asked to value health conditions using a 100-point visual analog scale and a standard gamble chance board. The HUI total score for each health state is calculated by means of a mathematical formula developed from the preference scores in line with the Von Neumann–Morgenstern utility theorem (Von Neumann and Morgenstern, 1944). Thus, the HUI total score represents mean community preferences on a 0.00–1.00 range, with 1.00 indicating optimal health utility, and 0.00 indicating “dead.” Scores can be lower than 0.00, indicating that the HRQoL is “worse than dead.” The HUI has been used in many studies since 1990, and has strong theoretical foundations, and has been well accepted by patients and professionals (Horsman et al., 2003: 2). There is a growing trend for the use of HUIs in HRQoL measures for describing treatment processes and outcomes in clinical studies, economic evaluations of health-care programs (cost-effectiveness and cost-utility analyses) and measurements and monitoring of population health.
Canadian Community Health Survey
The Health Utility Index is one of the many measures included in the biannual population health survey conducted by Statistics Canada (2010a). The Canadian Community Health Survey (CCHS) is a national cross-sectional survey of health determinants, health-status, and the utilization of the health-care system in Canada. The CCHS covers topics related to diseases and health conditions, general health, lifestyle, and social conditions. By collecting information about health at the community level, the CCHS aims to support provincial ministries of health, by providing them with the timely information they need to evaluate existing programs and to design new ones suited to their communities. The CCHS provides the current, detailed and uniform health information for every province and territory. The CCHS has been instrumental in drawing attention to emerging health issues (Statistics Canada, 2014).
The CCHS includes three types of content: core content, optional content, and rapid-response modules (Statistics Canada, 2010a). Core content is asked of all respondents, and has remained relatively stable for the 2001, 2003, 2005, 2007, and 2009 surveys. Optional content is chosen by health regions and is usually coordinated at the provincial level. Core variables have data on more than 100,000 respondents; whereas, optional variables have data on far fewer respondents, depending on how many provinces, and which provinces requested those variables to be measured. HUI measures are core variables.
The 2009 CCHS database was released for academic research in 2011 as part of Statistics Canada’s Data Liberation Initiative, the goal of which is to make data available to researchers at post-secondary institutions. Due to the depth and scope of the CCHS, several hundred articles have been written utilizing different cycles of the CCHS dataset.
General purpose
The purpose of this project was to use the variables available in the archived CCHS to examine the negative psychological concomitants of health-related disabilities, with particular focus on the extreme situation of suicidality.
Study 1
Purpose
The purpose of Study 1 was to examine correlations of HUI disability measures with available measures of suicidality in the 2009 CCHS.
Method
Measures
The Health Utility Index is based on eight sub-scales (Health Utilities Inc (HUInc), 2008), descriptions of which are shown in Table 1. Vision modality questions inquire about myopia, hypermetropia, and blindness. Hearing modality questions inquire about hard of hearing and deafness. Speech modality questions inquire about the ability to speak and to be understood. Dexterity mobility questions inquire about the ability to use and utilize hands and fingers. Mobility modality questions inquire about the ability to walk and the need for assistance from another person or from crutches or a wheelchair. Cognition modality questions focus on memory and thinking and the questions concern mostly problem-solving abilities and general memory. Pain mobility questions concentrate on pain level and how much pain level limits daily activities. Emotion modality questions inquire about the respondent’s state of happiness and despair. On the HUI sub-scales, higher values indicate more disability. As explained in the section “Introduction,” the sub-scales can be compounded by a proprietary weighting algorithm to produce a HUI Index score, for which a lower value indicates more disability.
Description of the Health Utility Index (HUI) sub-scales, showing score range and response options for the two lowest values and the worst (highest).
As shown in Table 2, conventional demographic measures were used as covariates, including gender, age, personal education level, and household income. The independent measures predicting suicidality were the seven HUI sub-scales that measure disability. The eighth HUI measure, the emotion modality, was considered for this study to be a measure of suicidality, since the highest value confirms the state of being “so unhappy that life is not worthwhile.” The HUI Index includes the HUI Emotion score; thus, for the purposes of this study, the HUI Index score was not used as either a measure of disability predicting suicidality nor as a measure of suicidality predicted by disability. Other dependent measures of suicidality available in the CCHS 2009 archived database with large numbers of respondents were diagnosed depression, dissatisfaction with life in general, feeling helpless dealing with problems in life, feeling hopeless during the past month, feeling worthless during the past month, and suicidal thoughts ever in life.
Descriptive demographic, disability, and suicidality measures for all respondents.
SD: standard deviation; HUI: Health Utilities Index.
N: number of respondents. For gender, 1: male; 0: female. Mean age was computed using category interval midpoints as the best estimate of age. For education 3 = post-secondary graduate. For income, 3: CAD40,000–CAD59,999. Lower Ns in the last five independent measures happen because these measures were optional depending on the interests of provincial ministries of health.
Data from the 2009 Canadian Community Health Survey (CCHS).
Sample
The sampling for the CCHS is based on the population in each Canadian province and the number of health regions within each province. A population sample of all regions and territories in Canada for people ages 12 years and older had been selected by Statistics Canada by first selecting households, and then selecting individuals within households. Households had been selected, half by random sampling of telephone listings and half by the existing sampling frames used by Statistics Canada’s Labor Force Survey. Within households, individuals were selected by random numbers picking age. For more exhaustive details of CCHS sample, see Statistics Canada (2010b). Thus, the CCHS data are considered to come from a representative sample of the non-institutional population above the age of 11 years. The CCHS 2009 questionnaire was administered by trained interviewers using computer-assisted interviewing (Statistics Canada, 2011). Participants were able to choose the language in which the interviews are conducted. It is important to note that CCHS sampling only included persons living in private homes. Persons residing in hospitals, retirement homes, dormitory residences, prisons, military barracks, and so on, were excluded from the sampling.
Procedures
As shown in Table 3, partial correlations were computed between each of the seven HUI sub-scale scores and the seven measures of suicidality. Four covariates were controlled in the partial correlation: age, gender, personal education, and household income. Five of the dependent measures had been coded in the CCHS database with high scores indicating positive well-being. Hence, diagnosed depression, dissatisfaction with life in general (labeled “satisfaction with life in general” in the CCHS database), feeling helpless dealing with problems in life, feeling hopeless during the past month, and feeling worthless during the past month were recoded to make a high value indicate suicidal tendencies.
Partial correlations of Health Utility Index (HUI) disability sub-scale scores with selected measures of suicidality, controlling for age, gender, personal education, and household income.
df: degrees of freedom.
All correlations are statistically significant (p < 0.001).
Data are from the 2009 Canadian Community Health Survey.
Results and discussion
Table 3 presents the results of the partial correlational analysis controlling for age, gender, education, and household income. All 49 correlations in Table 3 were positive; as a binomial probability, that could not have occurred as a random outcome. Because of the very large N in this study, all of the correlations were statistically significant (p < 0.001), indicating that any handicap, to some even small degree, is related to suicidality. The causal claim here is justified by the implausibility of suicidality causing physical disabilities. The HUI measures of hearing, vision, speech, and dexterity disabilities showed near negligible relationships with the suicidality measures, as shown by correlations ranging 0.02–0.06. The HUI measure of mobility disabilities showed stronger correlations with suicidality ranging 0.05–0.15. The HUI measure of cognition disabilities showed correlations with suicidality ranging 0.14–0.24. Similarly, the HUI measure of pain disabilities showed correlations with suicidality ranging 0.15–0.26.
These findings suggest that any disability has some statistically detectable effect in provoking suicidal tendencies. The disabilities of having trouble remembering or thinking and the disability of suffering chronic pain more strongly promote suicidality. But it is most noteworthy that the size of the impact of disability on thoughts to end life is relatively small. The r-square values arising from Table 3 ranged from less than 0.001 to a high of only 0.068, indicating the disabilities as measured by the HUI sub-scales are not indicators that the suffering individuals are at high risk of suicide.
Study 2
Purpose
Considering that the HUI measures of seven modality disabilities examined in Study 1 showed relatively low positive correlations with suicidality, then labeling patients as “worse than dead” as a result of very low HUI Index total scores may be unwarranted. The purpose of Study 2 was to investigate whether the effects of having multiple disabilities as well as negative emotional states, would justify a quality of life categorization “worse than dead.”
Method
Data from the 2009 CCHS database were selected for those individuals with HUI Index scores below 0.00. The HUI Index summation score is calculated from the eight sub-scales using a proprietary algorithm in order to quantify the respondents’ HRQoL. Because the HUI Index summation score considers 1.00 to have highest quality of life, and 0.00 to be equivalent to “dead,” then HUI Index scores below 0.00 are categorized as “worse than dead” (Feeny et al., 2002: 121). The 2009 CCHS survey data included 724 respondents scoring below 0.00 on their HUI Index measure. It is important to note that CCHS sampling only considers persons living in homes, excluding very ill people living in hospitals or in nursing homes.
Measures
As shown in Table 4, the 724 respondents were described by four demographic variables (gender, age, personal education level, and household income), by the seven HUI disabilities variables (HUI vision, HUI hearing, HUI speech, HUI mobility, HUI dexterity, HUI cognition, and HUI pain), by three health variables (self-perceived ill health, self-perceived ill health compared to 1 year ago, self-perceived mental health problems), by three social variables (employment, marital status, and sexual intercourse in the past year), and by four suicidality measures (HUI emotion, diagnosed depression, dissatisfaction with life in general, and suicidal thought ever in life). The three health variables and the satisfaction with life measure were answered in the CCHS protocol with a high value indicating negative conditions, even though the labels were expressed in positive terms. Therefore, in this report, these variables were relabeled with clear negative terms to indicate that high values mean worse life conditions. The three social variables, the depression measure, and the suicidal thought measure were all binary variables coded in the CCHS data as 1 = yes and 2 = no. These were recoded so that 0 = no, and 1 = yes, again so that a high value on these measures indicates worse life conditions.
Descriptive demographic, disability, health and social and suicidality measures for individuals with Health Utility Index (HUI) scores lower than 0.00 invoking the label “worse than dead.” Data are from the 2009 Canadian Community Health Survey (CCHS).
Mean age was computed using category interval midpoints as the best estimate of age.
Results and discussion
As shown in Table 4, descriptions of the 724 individuals in the 2009 CCHS data with HUI Index scores below 0.00 were not as catastrophically negative as the label “worse than dead” would imply. Demographically, this group was 58 percent women, with an average age of 61 (standard deviation (SD) = 17.89). They were generally low on education level (M = 2.00 SD = 1.34), living on an average household income of CAD20,000–CAD39,999 (SD = 1.30).
The disability data show that the group was not extremely disabled as measured by the seven HUI disability measures. The highest disability scores were for the HUI pain measure, showing mean score of 4.41 (SD = 1.27) in a scoring range 1–5, for the HUI cognitive measure, showing mean score of 4.04 (SD = 1.49) in a scoring range of 1–6, and for the HUI mobility measure showing mean score of 2.98 (SD = 1.12) in a scoring range of 1–4. Thus, this group labeled “worse than dead” displayed disabilities, but not to a maximal degree.
Their self-perceived ill health was very high, showing a mean of 4.41 (SD = 0.87) in a scoring range of 1–5. Self-perceived ill health compared to 1 year ago and self-perceived mental health problems both showed moderate mean scores, respectively 3.79 (SD = 0.96) and 3.66 (SD = 1.22) in a scoring range of 1–5. The social variables for this group were not as negative as the categorization of “worse than dead” would suggest. In this group, 14 percent were actively employed, 41 percent of were married, and of the 114 respondents who answered the sexual activity question, 69 percent had reported intercourse in the past year.
On the suicidality measures, this group had mean HUI emotion values and mean dissatisfaction with life in general values in the mid-range, respectively, 3.01 (SD = 1.38) and 3.42 (SD = 1.24) in a scoring range of 1–5. Only 51 percent reported diagnosed depression, and only 48 percent had reported suicidal thoughts ever in life.
In sum, this group’s reports did show them experiencing difficulties, but not at the extreme negative levels that the label “worse than dead” would suggest. On the HUI emotion sub-scale, only 18 percent selected the response that “life is not worth living.”
All respondents chosen for data collection in the CCHS were non-institutionalized individuals capable of living in their own homes despite their disabilities, and 14 percent of them were employed. In sum, “worse than dead” seems an invalid label for individuals with HUI summation scores lower than 0.00. Study 2 therefore recommends against continued use of that label, and also recommends that multiply disabled individuals with low HUI scores should be monitored for depression and suicidal thought.
General discussion
Study 1 used population data from the 2009 CCHS and found weak, but replicated positive relationships between disabled states and measures of life dissatisfaction and suicidality. The stronger relationships between disability, distress, and suicidality were with the disabilities concerning pain and cognitive functioning indicating that living with pain or having cognitive difficulties is perceived as more psychologically straining then having reduced sensory abilities or motor functions. The findings from Study 1 are consistent with earlier research concerning the effects of pain-related disability states and depression on suicidality (Edwards et al., 2006). Edwards et al. (2006) found that the magnitude of depressive symptom and the degree of pain-related catastrophizing were consistent predictors of the presence and degree of suicidal ideation. Similarly, Dour et al. (2011) found that emotion–cognition interaction in the statistical prediction of suicide attempts, linked problem-solving skills as a predictor for probability of a suicide attempts. Further support for more general findings on the psychological effect of disability is found in research done on the association between incident disability and depressive symptoms, showing that disability has a significant impact on the development of depressive symptoms (Chang et al., 2009).
Study 2 examined individuals with multiple disabilities and depressed emotions as measured by the HUI Index score (Gamst-Klaussen and Rudmin, 2014). Individuals scoring extremely low on the HUI summation score and thus categorized under the label “worse than dead” were descriptively investigated. This study used demographic, disability, health, social, and suicidality measures to describe the 724 individuals categorized by the HUI index as “worse than dead.” The study found that disabled states involving pain and cognitive functioning seem to be the highest represented disabilities reported for these individuals with extremely low HUI summation scores. However, there were no extremely high rates of diagnosed depression or suicidal ideation. A large portion of the respondents were married and most of those answering the sexual behavior question reported to have had sexual relations in the past year, thus showing social interaction and clear signs of life quality. Thus, Study 2 recommends against continued use of the label “worse-than-dead" when in fact mean scores of this group demonstrate that they do not consider themselves as better off dead.
Future research
The CCHS does not include the part of the population living in institutions. Therefore, the HUI does not account for older individuals with assumed poorer health residing in institutions. It seems likely that individuals with severe, chronic health problems would be institutionalized making it likely that the HUI scores in this study were biased toward better quality of life. There is also a lack of onset data for the disability sub-scales within the HUI. This lack of onset data eliminates the ability to differentiate congenital disabilities from acquired disabilities.
These studies exploited archived data that had been collected for purposes other than those of this study. Future studies designed to study the HUI and suicidality might employ psychometric measures created for the purpose of measuring suicidality. Some possibilities are the Beck Scale for Suicide Ideation (Posner, Brown, Stanley et al, 2011). Future research might also consider more refined differentiation of individuals with extremely low HUI summation scores, particularly in reference to depression and suicidality.
Conclusion
The results from these studies show that there are low positive correlations between disabilities and suicidality, based on a large, representative sample of a general population. This suggests that disabled individuals may be slightly more statistically prone to suicidal thought than the general population. Another implication is that medical staff, families, and the general public should be educated that multiple disabilities do not necessarily mean that the individuals feel despair or are suicidal since the study also establishes that suicidal thinking is not a universal outcome of living with various or multiple disabilities. These studies also imply that the HUI is a useful tool in identifying disabled states, but its labeling of individuals as “worse than dead” should no longer be usedsince this might mislabel individuals with a health state involving living with multiple disabilities catastrophizing their actual state of life quality.
Footnotes
Acknowledgements
The Canadian Community Health Survey data for this study were provided by Statistics Canada via their Data Liberation Initiative. The author would like to thank Floyd W Rudmin for his help and guidance in writing, editing, and methodology. Without his help, this article would never have been possible. Thank you for the help and mentoring. The author would also like to thank Olena Vasylenko and Fredrik Høie Jordet for taking the time to help with reviewing the revision.
Declaration of conflicting interests
The analysis and conclusions in this study are those of the author, and are not those of Statistics Canada.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.