Development and Psychometric Properties of the Self-Efficacy Scale for Transfers for Caregivers of Children Requiring Transfer Assistance

Article Outline

• Abstract
• Methods
  • Preliminary Phase: Development and Evaluation of the Initial CSEST
  • Preliminary Phase: Test of the CSEST
  • Preliminary Phase Data Analysis
  • Preliminary Phase Results
  • Phase II: Revision and Testing of the CSEST
  • Phase II Data Analysis
  • Phase II Results
• Discussion
  • Study Limitations
  • Future Research
• Conclusions
• APPENDIX 1. Examples of 2 items from the caregiver self-efficacy scale for transfers
• References
• Copyright

Abstract 

Thomas JJ, Shuford DM, Duke JC, Cipriani III DJ. Development and psychometric properties of the Self-Efficacy Scale for transfers for caregivers of children requiring transfer assistance.

Objective

To test the psychometric properties of a newly developed measure of caregiver self-efficacy, the Caregiver Self-Efficacy Scale for Transfers (CSEST), a 14-item questionnaire that measures the perceived self-efficacy of adult caregivers of dependent children and focuses on their perceived abilities and the challenges of assisting a child with restricted transfer ability.

Design

Reliability and validity study using survey data.

Setting

Community-based survey study.

Participants

Adult caregivers (eg, mother, father, grandparent) of dependent minors who require transfer assistance.

Interventions

Not applicable.

Main Outcome Measure

Using the Rasch measurement model, the CSEST was tested for reliability and separation, validity (hierarchical and content), and factor structure.

Results

The CSEST yields data with sufficient reliability (reliability indices of .96 and .94 for respondents and items, respectively) as well as sufficient separation of levels of self-efficacy (item separation index, 4.13; person separation index, 5.23). The CSEST has a stable hierarchical structure with adequate content validity (standardized χ² fit statistics <2.0) and point biserial correlations greater than .67. Factor analysis confirmed 2 insubstantial subscales; the 14-item CSEST accounted for greater than 85% of the total variance.

Conclusions

The CSEST yields reliable data and valid inferences of self-efficacy from adult caregivers of children requiring transfer assistance. The CSEST may identify areas of need and/or difficulty with transfers for parents caring for dependent children.

Key Words: Caregivers, Psychometrics, Rehabilitation, Reliability and validity, Self efficacy

STUDIES OF CAREGIVERS of children have reported a high rate of self-identified musculoskeletal problems including high incidence of neck and back pain.¹, ², ³, ⁴, ⁵, ⁶ Lifting children between lower- and higher-level surfaces and lifting children under stressful circumstances such as handling a wet child, coupled with use of poor body mechanics pose significant challenges for maintaining back and neck health.

Parents and caregivers of children who have disabilities are a special subset of the larger population of caregivers of children. Children with disabilities, especially those who are nonambulatory, present additional caregiver challenges.⁷, ⁸ The caregiver continues to provide transfer assistance on a daily basis throughout the child’s life. As their children grow, they frequently do not develop independence in mobility, resulting in long-term risk of stress and strain on caregivers. Nonambulatory children often use wheelchairs or other ambulatory devices that also require transferring into and out of vehicles and homes.

Occupational and physical therapists who treat children with disabilities are in a position to provide caregiver education regarding safe transfers, proper use of body mechanics, and transfer equipment. Parents and caregivers of children with disabilities vary in ability to accomplish needed transfers without injuring their backs and necks. Therapists need to be able to screen caregivers for their confidence in their ability, or self-efficacy,⁹ to meet the handling needs of their nonambulatory children without the caregivers injuring their backs. Assessing caregivers’ degree of confidence in physically handling their children is a first step in providing appropriate, individualized preventive strategies aimed at avoiding back strain and injury for caregivers.¹⁰ We found no assessments that focused on caregiver self-efficacy for transferring their children and children’s equipment without injuring their backs. The purpose of this study was to develop the Caregiver Self-Efficacy Scale for Transfers (CSEST) and test its psychometric properties. We report on the initial development and refinement of the CSEST, including Rasch analysis producing item and caregiver reliability indices, infit statistics, determination of order of perceived item difficulty, and item and caregiver separation indices. We used factor analysis to explore conceptual factors within the CSEST.

Back to Article Outline

Methods 

The initial development process and revision process of the CSEST took place in 2 phases. The preliminary phase involved development of the CSEST items, refinement of the items, and the first round of testing with caregivers. During the second phase, we revised the CSEST, based on the data analysis from the preliminary phase. The CSEST was again used by a sample of caregivers and we analyzed its psychometric properties using the Rasch measurement model. The Rasch measurement model has been used extensively as an alternative to classical test theory to analyze the psychometric properties (ie, reliability and validity) of measures in health care and rehabilitation.¹¹, ¹², ¹³, ¹⁴, ¹⁵, ¹⁶, ¹⁷, ¹⁸, ¹⁹, ²⁰, ²¹ The Rasch measurement model was first introduced by physicist George Rasch²² as a means to overcome the problems of measuring psychologic constructs on an ordinal, nonadditive scale. Rasch proposed a model that used the probability of responses to estimate a person’s ability or an item’s difficulty, thus eliminating dependence on raw, nonadditive scores. With the application of logarithmic functions to transform ordinal level data into equal-interval, additive data, Rasch provided a means to use a probabilistic approach to test the psychometric properties of measures. The resulting Rasch measurement model provides useful statistics and information regarding the reliability of data from measures and the validity of the inferences of these measures. For instance, the Rasch measurement model provides evidence of reliability of the data using a separation index (ie, how well the measure separates people into distinct levels of the latent trait) and a reliability index that is similar to the Cronbach α.²³, ²⁴ Validity evidence in the form of fit statistics and hierarchical structure provide additional information regarding the quality and utility of the measure.¹², ²⁵, ²⁶ This report summarizes both phases of the CSEST development.

Preliminary Phase: Development and Evaluation of the Initial CSEST 

We began the development of the CSEST with discussions with parents of nonambulatory children about what they found physically easy and demanding in caring for their children with special needs. From these discussions, we identified areas of common concerns and developed items around common areas of concern under specific conditions (see appendix 1). We asked caregivers to rate how confident they felt in their ability to do a specific type of transfer with their child under specific conditions. Conditions included adverse scenarios (ie, feeling hurried, running late, and experiencing inclement weather conditions) and more optimal conditions (ie, feeling calm, having plenty of time, and experiencing pleasant weather). Sixteen items consisted of 8 daily living transfers, including (1) storing the wheelchair in the car or van; (2) getting the wheelchair out of the car or van; (3) moving the child from the wheelchair into the car or van; (4) moving the child from the car or van into the wheelchair; (5) moving the child from the wheelchair to the toilet, a toileting device, or to where the child is diapered, and back to the wheelchair; (6) transferring the child from the wheelchair to the floor and back to the wheelchair; (7) moving the child from the wheelchair into bed and back to the wheelchair; and (8) moving the child from the wheelchair into the bathtub and from the bathtub back to the wheelchair. To assess a transfer of individually identified importance, we added an item that allowed the caregiver to select a public place of choice as the setting for a transfer.

For each of the 17 preliminary CSEST items, we developed a response scale based on Bandura’s⁹ recommendation of a 0- to 10-point scale anchored at one end with “not at all confident” and at the other end with “extremely confident.” After initial item development, 14 expert occupational and physical therapists evaluated the items on the scale for face validity including rating the items for clarity, conciseness, and importance. Twelve of the 14 experts returned the evaluations. The experts rated the items highly and provided important suggestions for wording changes. We revised items based on the evaluations from the experts.

Preliminary Phase: Test of the CSEST 

The University of Toledo Health Science Campus Institutional Review Board approved the plans for distribution of the scale and related demographic questions to caregivers. Parents and caregivers of children with disabilities were identified through the regional public school systems, outpatient pediatric clinics, and community support groups for caregivers. Contacts were also made via Eparent.com, the website for Exceptional Parent magazine, which elicited the cooperation of several groups throughout the United States including Family Voices, Support for Families of Children with Disabilities, Celebrating Families, and Vail Meadows Therapeutic Riding Academy.

Preliminary Phase Data Analysis 

We generated descriptive statistics for the respondent demographics using SPSS.²⁷,^a They included ages of children and caregivers, sex of caregivers, relationship of caregivers to children, caregiver health ratings, weight of children, diagnoses, transfer assistance level, and wheelchair type. The Rasch measurement model provided by Winsteps program²⁸,^b was used to test the psychometric properties of the preliminary version of the CSEST. The properties tested included person and item reliability, item infit statistics, and hierarchical order of the items.

Preliminary Phase Results 

Ninety-four questionnaire packets were collected in time for the data analysis. Twelve questionnaires were unusable for the following reasons: 7 were returned blank; 2 were inappropriate because the child did not need assistance with transfers; 1 was not used because a husband had filled out the scale for his wife who did not read English, and in 2 cases, 1 person had filled out 2 questionnaires, so only 1 questionnaire per person was used. A total of 82 questionnaires were included in the data analyses. Any incomplete data were entered as missing data. Table 1 summarizes the descriptive statistics for caregivers and their children. Total CSEST scores ranged from a low of 3 to a high of 170 (mean ± standard deviation [SD], 108.80±44.04). Possible scores range from 0 to 170. Twenty diagnoses were represented among the children. The most prevalent diagnosis identified was cerebral palsy, representing 46 children (57.5%).

Table 1. Descriptive Statistics Comparing Phase I and II Caregiver Samples

Characteristics	Phase I (N=82)	Phase II (N=71)
Age of child (y)
Range	3.5–22.0	3.0–22.0
Mean ± SD	12.0±5.5	10.7±5.1
Weight of child (kg)
Range	9.0–78.8	9.5–74.3
Mean ± SD	31.4±16.2	28.8±15.3
Transfer assistance, n (%)
Maximum	65(79.0)	57(80.3)⁎
Moderate	10(12.0)	9(12.7)
Minimum	5(6.0)	3(4.2)
None	2(2.0)	1(1.4)
Ambulation device, n(%)
Manual wheelchair	66(81.0)	59(83.1)
Power wheelchair	14(17.0)	9(12.7)
Stroller	2(2.0)	3(4.2)
Age of caregiver(y)
Range	25.0–65.0	24.0–69.0
Mean ± SD	42.2±8.9	41.1±8.5
Sex of caregiver, n(%)
Men	20(24.0)	13(18.3)⁎
Women	62(76.0)	57(80.3)
Relationship to child, n(%)
Mother	54(65.9)	52(73.2)
Father	16(19.5)	10(14.1)
Grandparent	3(3.6)	1(1.4)
Other	9(11.0)	8(11.3)
Caregiver health rating, n(%)
Excellent	22(27.0)⁎	18(25.4)
Good	40(49.0)	36(50.7)
Average	19(24.0)	17(23.9)
Poor	0(0.0)	0(0.0)

Abbreviation: SD, standard deviation.

The Rasch measurement model estimate of person reliability was .96 and for item reliability was .97. Several items had infit statistics in excess of 2 z scores (ie, standardized χ² fit values), which is considered to be evidence of poor fit to the model.²⁹ We evaluated the hierarchical structure of the preliminary CSEST and found that the sample of caregivers perceived items suggesting adverse conditions for transfers as more difficult compared with the items suggesting more optimal conditions.

Phase II: Revision and Testing of the CSEST 

Evaluation of the infit statistics from the preliminary CSEST revealed 3 items that did not fit the model sufficiently (z scores >2.0). We inspected these items (transferring the wheelchair in and out of the van under adverse conditions, transferring the child from the wheelchair to the floor under optimal condition, and transferring the child in a public place of choice) and eliminated them because they were either redundant with other items or did not yield a sufficient response pattern from the sample. Thus, a slightly shorter version of the CSEST was created with the remaining 14 items.

We sent the second version of the CSEST, which included demographic questions, to parents and caregivers of children with disabilities via the same mechanisms described for the initial testing of the CSEST. We added a demographic question that asked caregivers whether they had an ailment that affected their ability to perform transfers with their children. We gave caregivers the option of returning the paper copies of the CSEST by mail or requesting a telephone interview where a research assistant read the demographic questions and the CSEST items to the caregiver and recorded the responses.

Phase II Data Analysis 

The same descriptive statistics were collected in phase II. We used the Rasch measurement model to estimate the reliability of the data obtained from the CSEST based on the separation index and the reliability index. We estimated the separation index and reliability index for both the respondents as well as the individual items of the CSEST. In addition, we used the Rasch measurement model to estimate the validity of the inferences from the responses of the CSEST. Validity estimates included the individual infit statistics (ie, standardized χ² fit statistics) for items and respondents. The Rasch measurement model “generates goodness-of-fit statistics that can be used to perform confirmatory construct validity analyses of the constructed scale based on (a) formulating expectations about what should happen when a group of persons take a test, and then (b) confirming that the test items fit that model.”^14(p320) The infit statistic is an empirical method to discern the construct validity of an instrument. We examined the hierarchical structure of the items to test for order validity of the items. Finally, we conducted a factor analysis with Winsteps using raw scores to test for subscales of the CSEST. Factor analysis with Winsteps is an unrotated, orthogonal, principal components analysis of the standardized residuals of the observations.

Phase II Results 

We collected 74 questionnaires. Three could not be used because the children were noted to be completely ambulatory, leaving 71 questionnaires to be included in the data analyses. Table 1 summarizes the descriptive statistics for both phases. The majority (69%) of caregivers from phase II reported a physical ailment that limited their ability to care for their children. Thirty-eight (53.5%) of the children in phase II had a diagnosis of cerebral palsy.

The separation index for the respondents was 5.23, showing that the sample of subjects separated into more than 5 distinct levels of self-efficacy. The items had a separation index of 4.13, indicating that the items of the CSEST separated into more than 4 distinct levels of difficulty of endorsement for self-efficacy. These separation indexes translated into reliability indices of .96 and .94, respectively, for the respondents and the items.

The standardized infit statistics for the items and respondents did not exceed 2 absolute units from the expected chi-square values. Items and respondents interacted as expected, given a person’s level of self-efficacy and the perceived difficulty of a given transfer. The point biserial correlations, a further measure of item fit, exceeded .67 for each item. Table 2 provides each item’s individual infit statistics.

Table 2. CSEST Items in Descending Order of Perceived Difficulty

We examined the hierarchical order of the items by estimating the measures for each item. Table 2 provides the estimated item measures and shows the progression of item difficulty from least difficult to most difficult, in terms of self-efficacy endorsement. In table 2, bathtub transfer under adverse conditions was the most difficult item to endorse (measure ± standard error, .75±.10 logit) whereas transfer from wheelchair into a car or van under optimal conditions was perceived as the least difficult item to endorse (measure, −.84±.10 logit). The remaining items fall within an ascending order of difficulty from least to most difficult. A construct validity key map, displayed in figure 1, illustrates the order of perceived difficulty of items.

• 
  • View Large Image
  • Download to PowerPoint

• Fig 1. 

  Construct validity key map indicating the hierarchical order for indoor transfers (shaded bars) versus outdoor transfers (unshaded bars) noting adverse and optimal conditions. NOTE. N=71; items=14. Abbreviations: M, mean; S, 1 SD; T, 2 SDs.

Factor analysis suggested 2 main subscales for the CSEST. The first factor, which we defined as indoor transfers and outdoor transfers, explained 3.5% of the unexplained total variance. The second factor, which we defined as optimal transfer situations and adverse transfer situations, explained 2.5% of the total unexplained variance. Overall, the 14-item CSEST as a single measure explained 87% of the total variance.

Back to Article Outline

Discussion 

We tested the revised 14-item version of the CSEST for reliability and validity evidence. The Rasch measurement model reliability index indicated sufficiently high person reliability. The Rasch measurement model reliability index for items was also sufficiently high, well beyond the level Polit and Hungler³⁰ consider acceptable for the majority of instruments. The sufficiently high item reliability strengthens confidence in the stability of the CSEST for repeated administrations. There is a very high likelihood that the items will retain their relative positions within the 4 levels of self-efficacy that were identified with the separation index of 4.13.

The separation index indicated how well the CSEST separates into different levels of transfer self-efficacy (fig 2). A scale needs to produce at least 2 distinct levels of a construct to be considered sensitive. For example, the bathtub transfer under adverse conditions will likely remain the most difficult item for respondents to endorse and transferring with a wheelchair under optimal conditions will likely remain the easiest item to endorse. The remaining items will likely occupy 2 distinct levels of mid-range difficulty, and some changes in the hierarchical order within these 2 levels of difficulty can be expected on repeat administrations (ie, based on the standard errors as noted in table 2). In addition, the person separation index of 5.23 lends support to the sensitivity of the CSEST. It appears that the CSEST could be useful in clinical settings because of its ability to separate respondents into multiple levels of self-efficacy.

• 
  • View Large Image
  • Download to PowerPoint

• Fig 2. 

  Graphic representation of item difficulty compared with caregivers’ self-efficacy scores. NOTE. Numbers correspond to items on the CSEST in placement of perceived item difficulty. Abbreviations: a, adverse transfer condition in item; o, optimal transfer condition in item; X, caregivers’ scores (N=71).

Figure 2 also illustrates the relationship between the separation index and the overall reliability of the CSEST data. The graphic representation of item difficulty compared with caregivers’ self-efficacy scores was more condensed in the middle of the scale (see fig 2). Therefore, the caregivers within ±1 SD from the mean are measured the most accurately. The caregivers who are more than ±2 SDs from the mean are not measured as well because the items do not span the entire ruler. That is, fewer items measure the extreme range of self-efficacy in this sample. The graphic representation illustrated a wide variety of self-efficacy levels among the caregivers, which means that there are a variety of items that tap into a variety of levels of transfer self-efficacy. This variety further supports the reliability of the CSEST.³⁰

In terms of validity evidence, the Rasch measurement model provides several useful measures to support the construct validity of the CSEST. The standardized infit statistics indicated that the items were less than 2 standardized units from the expected values. This lends support to the content of the CSEST. Further, the point-biserial correlations all approached a value of 1.0. Fisher¹⁴ states that point-biserial correlations should be close to 1.0 for a good item fit. The point-biserial correlations were all greater than .67, which also contributes to the scales construct validity. Finally, looking at the construct validity key map (see fig 1), it is apparent that there is a logical hierarchical structure to the CSEST. The construct validity key map is a hierarchical order of item difficulty with the easier to endorse item (higher self-efficacy) at the bottom and the harder to endorse items (lower self-efficacy) at the top of the figure. The caregivers are displayed along the bottom of the graph, according to their individual perceived transfer self-efficacy. This indicates that all caregivers’ “confidence” levels are represented throughout the rating scale and items. For example, a self-efficacy rating of 10 on the easiest item involving transferring the wheelchair into the car or van under optimal conditions would be equivalent to a self-efficacy rating of an 8 for the hardest item involving bathtub transfers under adverse conditions. In other words, it was more difficult for the caregivers to report a higher level of transfer self-efficacy in adverse situations when compared with optimal situations. The highest self-efficacy reported for the hardest transfer was a score of 8 on a scale of 0 to 10. The construct key map also shows that persons of lesser perceived self-efficacy are less likely to endorse difficult items compared with persons of greater perceived self-efficacy. For instance, a person with a measure of confidence estimated at −2.0 logits will have at least a 50% probability of endorsing at least some level of confidence on most items with the exception of the most difficult item (bathtub transfer under adverse conditions). For this item, a person with an estimated measure of −2.0 will have almost no chance of endorsing any level of confidence.

The results point to a trend concerning perceived transfer self-efficacy of caregivers of children requiring transfer assistance. Caregivers perceive the adverse transfer conditions to be more difficult in which to feel efficacious than the optimal conditions. Under time constraints, caregivers may not lift the child in the safest manner. Griffin and Price stated that the lifting “methods chosen were not, however, always those that were ‘best’ for the mother’s back, just those that were time saving.”^2(p9) In other words, caregivers may report decreased self-efficacy for transferring their children without straining their back for adverse conditions because they may resort to unsafe lifting techniques in more adverse, time-constrained situations.

An additional source of construct validity evidence was found with the factor analysis. The CSEST appears to function well as a single, 14-item measure of self efficacy. We identified 2 factors within the CSEST (an indoor and outdoor factor and an adverse and optimal factor); however, neither factor contributed sufficiently to the overall unexplained variance relative to the total variance explained by the CSEST. These 2 factors certainly make conceptual sense in that indoor transfers are likely perceived to be different from outdoor transfers, and transfers under adverse conditions are likely to be perceived to be more difficult than transfers under optimal conditions. The weak contribution of these factors to the unknown variance, however, suggests that the 14-item version of the CSEST works sufficiently as a single measure of self-efficacy.

Even though there was insufficient factor analysis evidence of distinct subconstructs in the CSEST, caregivers seemed to perceive a hierarchy of difficulty with indoor transfers under adverse conditions perceived as more difficult to endorse than were outdoor transfers under optimal conditions (see fig 1). There are reasons why caregivers may be less confident during indoor transfers than during outdoor transfers. The conditions and types of transfers needed in the home are more varied, requiring more complex physical demands on the caregiver than the car or van transfers out-of-doors. Also, transfers performed inside the home, such as toilet transfers, occur in a greater frequency during the day than transfers performed outside the home. Griffin and Price state that car transfers were not often commented on because it was a “task that was not performed as frequently as other tasks.”^2(p10)

Caregivers perceived the bathtub transfer and floor transfer under adverse conditions to be the most difficult transfers to perform. The hierarchical ranking of items according to perceived item difficulty are split with the adverse situations being perceived as more difficult in which to express confidence in one’s skills, and the optimal transfer situations being less difficult except for 1 item, bathtub transfer under optimal conditions. In a study performed by Nicholson,⁷ 96% of parent caregivers assist their child with transferring in and out of the bathtub, which is the highest percentage of required parent assistance when compared with all the other transfers examined (get up from seat; get up from seat or floor; standing; walking; dressing; get in or out of bed; get on or off toilet). The bathtub transfer under both optimal and adverse conditions fell in the upper half of perceived item difficulty indicating that caregivers perceived the bathtub transfer to be particularly more difficult than other transfers. According to Griffin and Price, “for many of the mothers, bathing did not become a problem until they [the child] moved from a baby bath to a proper bath.”^2(p12) The CSEST would seem to identify and measure accurately this perception of caregivers.

Bandura stated “efficacy beliefs affect performance both directly and by influencing intentions.”^9(p43) Thus, therapists can assume that a decreased level of self-efficacy indicates a decreased ability to perform a specific transfer. Using the CSEST in the clinical setting can help occupational and physical therapists target their family education, addressing those areas of particular concern as indicated by a low self-efficacy score for particular items. By increasing a caregiver’s self-efficacy, therapists are changing the caregiver’s behavior.³¹ Bandura stated that “physical health is largely determined by lifestyle habits and environmental conditions. People suffer physical impairments and die prematurely mostly because of preventable detrimental habits.”^9(p258) Therapists are able to help caregivers change their behavior and increase their transfer self-efficacy, which may in turn have lasting impacts on a caregiver’s overall health and well-being.

Study Limitations 

There are several limitations to this study. The multiple methods used to reach caregivers made it difficult to track the exact number of surveys that reached caregivers’ hands. Several organizations received multiple surveys to pass out to eligible caregivers. We do not know the number of packets that reached eligible caregivers. Some survey packets were probably discarded and we do not know how many eligible caregivers who received surveys chose not to participate. This makes it impossible to calculate a response rate for the sample. The sample size for each phase was relatively small. This may limit the generalizability of the results. In fact, the sample consisted largely of female primary caregivers, which makes the results less applicable to male and secondary caregivers. A further limitation to this study is that we do not know if caregivers used any assistive devices or equipment to help them perform transfers, such as a wheelchair lift or bath hoist.

Future Research 

Future research needs to address the limitations of this study. To be confident in the reliability and validity of the CSEST as it applies to gender and caregiver role, the scale needs to be administered to a larger sample size of men and secondary caregivers. A larger sample size would increase the ability to generalize results more broadly and permit secondary comparisons such as whether mothers respond differently from fathers and whether primary caregivers can be differentiated from secondary caregivers. Rasch analysis of the 0- to 10-point rating scale used for the CSEST items would reveal if the rating scale is performing as expected. The clinical utility of the CSEST would be enhanced by studies examining whether different CSEST scores correspond to different functional levels among caregivers or differentiate caregivers with ailments from those without reported ailments or correspond to different sizes or weights of children. These studies would use differential item functioning analysis and require a larger sample size. Future studies could document the prevalence of equipment use and its effects on caregiver self-efficacy for performing transfers. Studies should explore how useful occupational and physical therapists find this scale in the clinical setting. Finally, research could examine CSEST responsiveness to change in transfer self-efficacy over time or in response to interventions.

Back to Article Outline

Conclusions 

This study describes the development, revision, and testing of the CSEST. Rasch measurement found that the person scores and the item measures of the CSEST have high reliability. The 14 items of the scale fit the unidimensional ruler created by Rasch measurement, which contributes to the scale’s construct validity. We found that the items of the CSEST follow a hierarchical order in which caregivers perceive the bathtub transfer under adverse conditions to be the hardest transfer and transferring from the wheelchair into the car or van during optimal conditions to be the easiest transfer. These findings are consistent with previous studies⁶, ⁷, ⁸ regarding the perceived difficulty of caring for a dependent child. The scale is sensitive enough to separate caregivers into different levels of self-efficacy and items into different levels of difficulty. Finally, factor analysis of the CSEST documented 2 subscales; however, these subscales did not contribute sufficiently to the overall measure of self-efficacy as individual measures. Therapists can use the CSEST in practice to assess caregivers’ level of self-efficacy to perform transfers without straining their backs and to target specific transfers during family education. This study shows that the CSEST yields reliable data, results in valid inferences regarding self-efficacy, and is a sensitive measure that can be used by occupational and physical therapists in clinical settings to help caregivers of nonambulatory children to safely transfer their children with less fear of straining their backs. Future research is needed to evaluate the performance of the rating scale and continue to explore further reliability and validity characteristics of the CSEST.

Suppliers

Back to Article Outline

APPENDIX 1. Examples of 2 items from the caregiver self-efficacy scale for transfers 

Item 2 You are home and feeling calm. You are going on an outing that you and your child have enjoyed before. You have plenty of time, and the weather is pleasant.

➢ How confident are you that you can move your child from the wheelchair or stroller into the car or van without straining your back?

Item 9 You are home. You are feeling hurried, and you are running late.

➢ How confident are you that you can move your child from the wheelchair or stroller into the bathtub and from the bathtub back to the wheelchair without straining your back?

NOTE: The complete CSEST is available from the author.

Back to Article Outline

References 

1. Calabro KS, Bright KA, Cole FL, Mackey T, Lindenberg J, Grimm A. Child care work: organizational culture and health and safety. AAOHN J. 2000;48:480–486
   • View In Article
   • MEDLINE
2. Griffin SD, Price VJ. Living with lifting: mother’s perceptions of lifting and back strain in childcare. Occup Ther Int. 2000;7:1–20
   • View In Article
   • CrossRef
3. King PM, Gratz R, Scheuer G, Claffey A. The ergonomics of child care: conducting worksite analyses. Work. 1996;6:25–32
   • View In Article
   • CrossRef
4. Grant KA, Habes DJ, Tepper AL. Work activities and musculoskeletal complaints among preschool workers. Appl Ergon. 1995;26:405–410
   • View In Article
   • MEDLINE
   • CrossRef
5. Brown MZ, Gerberich SG. Disabling injuries to childcare workers in Minnesota, 1985-1990. J Occup Med. 1993;35:1236–1243
   • View In Article
   • MEDLINE
6. Sanders MJ, Morse T. The ergonomics of caring for children: an exploratory study. Am J Occup Ther. 2005;59:285–295
   • View In Article
   • MEDLINE
   • CrossRef
7. Nicholson J. Management strategies for musculoskeletal stress in the parents of children with restricted mobility. Br J Occup Ther. 1999;62:206–212
   • View In Article
8. Washington K, Schwartz IS. Maternal perceptions of the effects of physical and occupational therapy services on caregiving competency. Phys Occup Ther Pediatr. 1996;16:33–54
   • View In Article
   • CrossRef
9. Bandura A. Self-efficacy: the exercise of control. New York: WH Freeman; 1997;
   • View In Article
10. Cornman-Levy D, Gitlin L, Corcoran MA, Schinfeld S. Caregiver aches and pains: the role of physical therapy in helping families provide daily care. Alzheimers Care Q. 2001;2:47–55
    • View In Article
11. Campbell SK, Kolobe TH, Osten ET, Lenke M, Girolami GL. Construct validity of the Test of Infant Motor Performance. Phys Ther. 1995;75:585–596
    • View In Article
    • MEDLINE
12. Chang WC, Chan C. Rasch analysis for outcome measures: some methodological considerations. Arch Phys Med Rehabil. 1995;76:934–939
    • View In Article
    • Abstract
    • Full-Text PDF (683 KB)
    • CrossRef
13. Creel GL, Light KE, Thigpen MT. Concurrent and construct validity of scores on the Timed Movement Battery. Phys Ther. 2001;81:789–798
    • View In Article
    • MEDLINE
14. Fisher AG. The assessment of IADL motor skills: an application of many-faceted Rasch analysis. Am J Occup Ther. 1993;47:319–329
    • View In Article
    • MEDLINE
15. Haley SM, McHorney CA, Ware JE. Evaluation of the MOS SF-36 physical functioning scale (PF-10): unidimensionality and reproducibility of the Rasch item scale. J Clin Epidemiol. 1994;47:671–684
    • View In Article
    • MEDLINE
    • CrossRef
16. Heinemann AW, Harvey RL, McGuire JR, et al. Measurement properties of the NIH Stroke Scale during acute rehabilitation. Stroke. 1997;28:1174–1180
    • View In Article
    • MEDLINE
    • CrossRef
17. Lai JS, Fisher AG, Magalhaes LC, Bundy AC. Construct validity of the Sensory Integration and Praxis tests. Occup Ther J Res. 1996;16:75–96
    • View In Article
18. MacKnight C, Rockwood K. Rasch analysis of the hierarchical assessment of balance and mobility. J Clin Epidemiol. 2000;53:1242–1247
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (138 KB)
    • CrossRef
19. Morris S, Morris ME, Iansek R. Reliability of measurements obtained with the Timed up and Go test in people with Parkinson disease. Phys Ther. 2001;81:810–818
    • View In Article
    • MEDLINE
20. Silverstein B, Fisher WP, Kilgore KM, Harley JP, Harvey RF. Applying psychometric criteria to functional assessment in medical rehabilitation: defining interval measures. Arch Phys Med Rehabil. 1992;73:507–518
    • View In Article
    • MEDLINE
21. Velozo CA, Kielhofner G, Lai JS. The use of Rasch analysis to produce scale-free measurement of functional ability. Am J Occup Ther. 1999;53:83–90
    • View In Article
    • MEDLINE
22. Rasch G. Probabilistic models for some intelligence and attainment tests. Chicago: Mesa; 1980;
    • View In Article
23. Wright BD, Mok M. Rasch models overview. J Appl Meas. 2000;1:83–106
    • View In Article
    • MEDLINE
24. Wright BD, Stone MH. Best test design. Chicago: Mesa; 1979;p 31
    • View In Article
25. Klauer KC. The assessment of person fit. In:  Fischer GH,  Molenaar IW editor. Rasch models: foundations, recent developments, and applications. New York: Springer-Verlag; 1995;p. 97–110
    • View In Article
26. McNamara T. Concepts and procedures in Rasch measurement. In: Measuring second language performance. Boston: Addison-Wesley; 1996;p. 149–181
    • View In Article
27. SPSS Inc. SPSS Base 10.0 for Windows user’s guide. Chicago: SPSS; 1999.
    • View In Article
28. Linacre JM. A user’s guide to Winsteps: Rasch model computer programs. Chicago: Mesa Pr; 2004;
    • View In Article
29. Bond TG, Fox CM. Applying the Rasch model: fundamental measurement in the human sciences. Mahwah: Lawrence Erlbaum; 2001;
    • View In Article
30. Polit D, Hungler B. Nursing research: principles and methods. 2nd ed.. Philadelphia: JB Lippincott; 1983;
    • View In Article
31. Maddux JE. Self-efficacy, adaptation, and adjustment: theory, research, and application. New York: Plenum Pr; 1995;
    • View In Article

• a Version 11.0; SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.
• b Version 3.56; Winsteps, PO Box 811322, Chicago, IL 60681-1322.