A Global Family Quality of Life Scale: Preliminary psychometric evidence

The FQOL-2006 Survey [3] measures FQOL in families who have a member with intellectual or developmental disability. Mean scale scores are calculated for each domain. The survey contains two sections for each domain, (1) quantitative and qualitative items about characteristics of the domain, and (2) six items for each dimension and one open ended question for each domain, “What things do you think could improve your family quality of life?”. Additionally, two global items were included which are described below. Isaacs et al. tested the factor structure of the FQOL-2006 Survey in a large sample across four countries (Nigeria, Australia, Canada, US) in families with a member with an intellectual disability ( n = 402) and found the importance dimension ratings across the nine domains were skewed (all domains rated high) [17]. The stability dimension across the nine domains was not supported due to low factor loading leading to a final factor structure of 4 dimensions: opportunity, initiative, attainment, and satisfaction [17]. Confirmatory factor analysis supported a good fit for a model with nine domains (Health, Finances, Family, Others, Services, Values, Careers, Leisure, Community) and an overall latent FQOL factor. Although an overall latent FQOL factor was supported in this analysis, it has not been used in the literature by other investigators [17].

1.3 Global measures of overall FQOL

Single item individual QOL measures have been validated and supported [10, 18]. Some researchers feel that the overall global perception of QOL that reflects the individual’s emphasis on important domains may be useful as an outcome [10, 19, 20]. Similarly, a global concept of a family member’s perception of FQOL can include the domains important to the family.

Two global items at the end of the FQOL-2006 Survey address FQOL: “Overall, how would you describe your family’s quality of life?” and “Overall, how satisfied are you with your family’s quality of life?” [3]. Response pattern for the overall global item was 5 (excellent) to 1 (poor). Response pattern for the satisfaction question was 5 (very satisfied) to 1 (very dissatisfied). Only the global satisfaction item has been reported in the literature. In addition to the two global measures of overall FQOL described, a single item measure, “How would you rate your family’s quality of life?” was used by Sawin, Brei, Buran, and Fastenau to capture FQOL in families who had an adolescent with SB [10]. The response pattern was zero (poor) to 100 (excellent).

1.4 Summary of studies using FQOL measures

A small number of studies have reported measurement of FQOL in a variety of ways, mostly in correlational descriptive studies. While assessment of FQOL has been conducted since 2002, 12 studies have examined FQOL in families with children with either chronic health conditions or intellectual and developmental disabilities [11]. Only two studies examined families with children with and without chronic health conditions (CHC) [6] and without disabilities [15]. A majority of studies were conducted in families of children with intellectual and developmental disabilities. See Table 2 for a summary of studies using FQOL measures.

2. Methods

This psychometric analysis used data from a larger study on secondary conditions and adaptation in adolescents with SB. This study included 209 families who had an AYA with and without SB. Participants (parents and AYA) were asked to rate their own, their parent (or AYA) and their family’s QOL.

OPEN IN VIEWER

3. Results

Parents were primarily female (94%), Caucasian (86%), well-educated either attending or completed college/vocational training/graduate program (71%) and married (74%) (see Table 3 for sample characteristics). Combined family income was significantly lower for families with AYA with SB, 18% earned less than $20,000 and 50% earned greater than $50,000. In comparison group, 4% earned less than $20,000 and 73% earned greater than $50,000. About half of the AYA were female (54%) with a mean age of 15 years (range 12–25). Eighty-four percent of AYA with SB had a shunt and half used a wheelchair (50%). AYA with SB had a broad range of functional motor levels, high lumbar (19%), mid lumbar (33%), low lumbar (35%) and sacral (8%).

Principal component analysis of these parent and AYA data and subsamples (with or without SB) with Varimax rotation of the 3-item scale of FQOL items identified that the items load on one factor. In the total and subsamples with parent or AYA ratings (with and without SB), the Bartlett tests were significant and KMOs were 0.654–0.692 indicating factorability. In each analysis, the scree plot test suggested 1 factor and the loading on to a single factor ranged from 0.79–0.94. The amount of variance explained by a single factor ranged from 78–83% (see Table 4).

Parents reported high FQOL (parent report SB group m = 82.47, σ = 14.77; comparison group, m = 89.25, σ = 10.10). Parents with AYA with SB had slightly lower FQOL with greater variance. This difference was significant ( t = - 3.915; 0.001) but relatively small (e.g., less than 7 points on the 100-point scale). The AYA with and without SB also reported high FQOL and there was no significant difference between groups ( t = - 0.052; p = 0.959) (see Table 5 for descriptive statistics for QOL and overall FQOL). Internal consistency reliability for the scale in total sample and subsamples was strong ( α = 0.86–0.90) for parents and AYA.

Parent report of FQOL was moderately correlated with their report of family satisfaction and resources ( r = 0.51–0.57) and AYA report of FQOL was correlated with AYA report of family satisfaction ( r = 0.39). These moderate correlations support FQOL as a related but different construct than family functioning measured by family satisfaction and resources.

The relationship of FQOL in parent and AYA pairs was poor to moderate with wide confidence intervals indicating poor agreement and variability between parent and AYA report of FQOL scale scores for the total and sub-samples: ICC 0.41, 95% CI [0.22–0.56] for total sample; ICC 0.38, 95% CI [0.10–0.58], for SB group; ICC 0.48, 95% CI [0.22–0.65] for comparison group. Agreement between parents and AYA was highest in the group without SB.

Two approaches to FQOL measurement have been used in the literature: first, measures that tap domain-specific aspects of family life; and second measures that assess overall perspective of FQOL. The G-FQOLS, an example of the second, is a weighted perception of the domains important to the reporter about the family as a whole, which in this case is, an average of a family member’s perspectives of the individual, the child or parent, and their family’s quality of life. The nature of FQOL is a dynamic approach which facilitates parents’ and AYA’s report of important domains of life which can vary across time. A strength of the G-FQOLS is its focus on perception versus function.

In the current overall G-FQOLS instrument, perception of individual, other family members (either parent or child) and FQOL loaded onto one factor. Investigators might wonder about the close relationship of the three items. However, strong relationship between these items ( r = 0.60–0.76) has been reported previously [10] and there is evidence that family is the predominate concept when evaluating both AYA and parent individual quality of life. In this study, when queried about the meaning of individual quality of life, family was the concept most commonly nominated by all participant groups (AYA with and without SB and their parents) (unpublished data). The “core” contribution of family in individual QOL supports the factor loadings for the G-FQOLS three items. It is interesting to note that family is rarely a domain included in individual QOL measures, especially those based on function.

In one study that used the FQOL-2006 Survey with both overall and domain-specific measures, the global single FQOL item was related ( r = 0.62, 0.001) to the composite domain-specific score of FQOL [24]. The overall shorter measure is highly related to domain-specific total score. Therefore, using single items (or in this case 3 items) may be an effective screening tool. Conducting clinical research with instruments measuring predetermined domains is an efficient way to identify need for specific support services. Assessing overall FQOL may help us understand factors that contribute to FQOL. If family member reports a relatively poor FQOL, then further assessment of factors contributing to their perception would be warranted. It is important to distinguish between this measure of global FQOL and family functioning, the latter being a process measure of how families operate rather than a measure of a rating from the perspective of a family member of their overall FQOL. Factors related to FQOL may vary in different environments or conditions. The G-FQOLS, as a potential outcome measure, provides a way to go beyond what has been identified as relevant, since these domains may change because of varied needs and resources over the life course. This approach facilitates the identification of modifiable factors to inform interventions for individuals and families.

Poor agreement between parent and AYA report of FQOL indicates measure is sensitive to differences in perspectives by reporter. Findings support importance of obtaining AYA appraisal from AYA, rather than by parent proxy. This instrument offers the capacity to assess discrepancies between family members which may identify a meaningful target for intervention. Further assessment of the meaning of FQOL for both parents and AYA can determine if family members have the same meaning of FQOL or prioritize the domains similarly. If AYA had moderate to severe intellectual disabilities, they may not have the capacity to appraise their own FQOL. In this instance, ratings by parent proxy might be used to assess FQOL.

Differences between groups with and without SB were evident in parents, but not in AYA data. Although differences were detected in parent report, those differences were small. The lack of difference by group with and without SB in AYA may indicate that the meaning of FQOL for AYA differs from that of parents. Parents may place more weight on aspects of life related to their experiences as caregivers. In families of children with moderate to severe intellectual disabilities, parents’ appraisal of FQOL may be influenced by continued responsibility for care as their AYA transitions to adulthood.

Another explanation for the lack of difference between the report of FQOL in AYA’s with and without SB may be due to the “disability paradox” seen when adults with disabilities report a high QOL despite limitations in activities of daily living, difficulty in social roles, and experience with persistent discrimination [25]. The perception of AYA with SB may be based on meaning and values different from individuals without disability. While professionals might expect lower QOL in youth with SB, there is conflicting evidence about QOL in youth with SB in other than the physical domain [10, 26]. The universal finding of lower physical QOL is based on tools that use function such as items about walking or running to assess QOL/health-related QOL(HRQOL). Murray et al. acknowledge their findings of lower HRQOL could be a result of the measure they used where physical domain was measured by function [27]. Among QOL/HRQOL researchers studying individuals with disabilities, there is strong support for measurement of HRQOL based on perception (the WHO definition) rather than function.

Either a comprehensive FQOL instrument, which addresses domain-specific issues in a family who has a member with a disability, or a brief measure of overall perception can be useful in different research or clinical settings, depending on the purpose of the project. Future research may use rigorous mixed methods to develop/refine an item pool and test content validity for a domain-specific measure of FQOL in SB. This would help researchers/clinicians identify domain-specific targets for intervention. Additionally, item response theory may be applied to extend our measure of FQOL to predict expected ratings given specified parameters or our measure may be converted to a Patient-Reported Outcomes Measurement Information System (PROMIS ® ) metric to enable comparisons to other global health or well-being measures across studies. This study supported the use of a new brief summative instrument, the G-FQOLS. The analysis of total and subsamples provides evidence of construct validity and internal reliability of the G-FQOLS. This global appraisal of FQOL reflects parent’s and AYA’s personal weighting of domains important to them. Although domains or their weight may change over time, the overall rating is an indicator of the family’s overall quality of life.

5. Limitations

As a self-report measure, there is potential for social desirability response bias. Previous research using similar items in the SB population found scores of QOL/FQOL single item means between 72–77 [10], slightly lower than our current findings. The G-FQOLS items were administered in a set of 10 single-items including measures of stress, health and responsibility for care which might have facilitated a response bias. However, all items had a broad response range, supporting respondents’ ability to determine differences between items. For example, the mean stress of everyday life for parents of AYA with and without SB specifically was 53.33 (SD 26.32) [6]. However, in future testing, to protect against any response bias, FQOL items should be interspersed with items using a different response pattern. Test-retest reliability and sensitivity to change have not been determined and are needed prior to use in intervention studies. Another limitation is the lack of diversity in the sample tested. This sample was mostly Caucasian and resided primarily in Midwest (Wisconsin, Indiana, Ohio, and Virginia). Evaluation in a more diverse sample is indicated.

6. Implications for practice and research

This measure may be useful in primary or specialty care settings to identify families at-risk who need further assessment and intervention. Evaluation of families can begin with a preliminary screening of G-FQOLS. Although there is not enough evidence to determine cut points using the G-FQOLS, clinicians can either use the tool as a screening tool or as an entry to discussion. Further assessment and discussion might be indicated when the G-FQOLS score is 2.0 standard deviations below the mean of the evidence to date (e.g. values of 56.84–59.16 using data from this study). Alternatively, when used primarily as an entry to discussion, probes used could include: “What does FQOL mean to you? What are the barriers to achieving your FQOL goals? What would make your FQOL better?”. Additional evaluation could follow with a standardized assessment for area(s) of concern such as depressive symptoms, pain, or other issues.

Measuring FQOL over time would enable clinicians and researchers to understand developmental influences on FQOL. While most positive appraisals may occur in the younger years when discrepancies for development between themselves and peers are smallest, parents’ future expectations for their family members may influence their FQOL. Families with difficulty in typical tasks of family life in the context of chronic illness may be identified with a substantial decrease in score or low G-FQOLS score. This should lead to follow-up assessment. FQOL measurement may be particularly helpful during times of transition when additional independence is expected (e.g., transitioning to adolescence or emerging adulthood). Changes in G-FQOLS scores over time, when experiencing acute illnesses, secondary complications, or hospitalizations would be indications for specific assessment. Understanding perceptions of FQOL may help develop targeted family interventions over the life course.

Further research is needed to establish G-FQOL test-retest reliability and sensitivity to change. If established, the instrument can be used to monitor and evaluate future family-centered interventions in research and practice. Research to date has used the measure to quantify FQOL from a family member’s perspective. Future research should also address the use of the instrument with multiple family members.

7. Conclusion

This study provides preliminary evidence that supports use of a simple 3-item global measure of FQOL in research and practice. The G-FQOLS rating may be used for additional assessment of the components of FQOL in parents and AYA. If G-FQOLS is low, it may prompt referral to support services or screening for mental health concerns in both pediatric and specialty care settings.