A scale for measuring health-related hope: its development and psychometric testing

ABSTRACT

BACKGROUND

Commonly used scales that measure hope in general may not capture how people experience their health and how healthcare affects hope as a patient-reported outcome. Therefore we tested question-items for measuring health-related hope (HR-Hope), and developed a scale for measuring HR-Hope in adults with chronic illnesses.

METHODS

Using results from a qualitative study, we wrote 45 question-items to measure 3 domains of HR-Hope: health & illness, role & social connectedness, and “something to live for.” We tested those items among 454 patients with chronic kidney disease (stage 3 to stage 5 requiring dialysis) in a cross-sectional survey. Using the results of factor analysis and of analyses based on item-response theory, we chose 18 of the 45 items, and proceeded to the next steps: reliability testing and criterion-based validation testing.

RESULTS

Exploratory factor analysis indicated unidimensionality. Confirmatory factor analysis showed that the data were consistent with a single second-order factor (health-related hope) and 3 first-order factors (the 3 domains). Coefficient α for total HR-Hope scores was 0.93. The HR-Hope scale was moderately correlated with both domains of Snyder’s hope scale. Compared with Snyder’s hope scale, the HR-Hope scale was more strongly correlated with the Center for Epidemiologic Studies Depression Scale and with the 8 domains of the Short Form-36-v2. Compared with Snyder’s hope scale, the HR-Hope scale was more sensitive to impaired performance status, depression, acceptance of illness, and the presence of family.

CONCLUSIONS

The 18-item HR-Hope scale can be used to measure health-related hope in people with chronic illness.

INTRODUCTION

The number of people who have chronic degenerative diseases is growing worldwide. Given that their medical conditions will not improve, what should be the goal of their care? Healthcare professionals might try to achieve good outcomes with respect to these patients’ inner lives: Are they in despair or are they hopeful? Assessing hope has been regarded as part of the “art” of medicine, and as depending on each healthcare professional’s interpersonal skills [1]. An instrument to measure health-related hope could help healthcare professionals understand the patients and families with whom they interact, and thus could motivate all concerned to communicate and perhaps even to increase hope.

Several scales have been used to measure hope. Among the most commonly used are Snyder’s Adult Hope Scale [2], the Herth Hope Index [3], the Miller Hope Scale [4], and the Beck Hoplessness Inventory [5]. These have been used in many healthcare-related contexts, but nonetheless they are general. That is, they measure hope (or lack of hope) in general, rather than hope with regard specifically to one’s health. The difference between hope in general and health-related hope may be understood by analogy to the difference between general quality of life (QOL) and health-related quality of life (HR-QOL, Supplementary Fig. 1 (Additional File 1)). Indices of general QOL are less useful than indices of HR-QOL when cross-sectional comparisons are used to answer questions about how health is associated with daily activities, role functioning, social functioning, etc. For example, happiness may be seen as one aspect of general QOL, and it may be more likely to be affected by factors other than health (religion, personality, etc.) than by health-related factors [6]. Similarly, in healthcare contexts when the goal is to make cross-sectional comparisons regarding hope, an index of hope in general may be less useful than an index of health-related hope. The analogy extends to longitudinal comparisons: when compared with indices of hope in general, an index of health-related hope can be expected to be more sensitive to the effects of healthcare interventions. We saw a need for an instrument that is responsive to changes in hope as they relate to interventions directly affecting health. We therefore developed a scale to measure not hope in general but rather health-related hope.

The sequence of steps in the development of this scale is illustrated in Fig. 1. The initial development and qualitative testing of items for this scale have been reported in Japanese [7], and so they are summarized briefly in Supplementary Item 1 (Additional File 2).

Fig. 1 The sequence of steps in the development of the Health-Related Hope scale.

The work began with a qualitative study [7] (the “Previous study” referred to in Fig. 1). Based on the findings of that study and on discussions among the research team, we conceived of health-related hope as having the three domains of health & illness, role & social connectedness, and “something to live for,” with each of those three domains having two aspects: prospects and self-efficacy. We then created 47 question-items asking about health-related hope. Below we report on the initial psychometric testing of those 47 items, and on the development of a shorter version of the Health-Related Hope (HR-Hope) scale (18 items).

METHODS

To begin testing the original 47 items, 28 chronic kidney disease (CKD) patients who were receiving dialysis were asked to respond to the items and to indicate whether or not they were easy to understand and respond to (the “Step 1” referred to in Fig. 1). We then removed or reworded items that the patients said were difficult. We also removed items with ceiling effects. The remaining 45 question-items are listed in Table 1. They were preceded by the following stem, which was used to establish the common context for all: “Please answer the questions below while keeping in mind how you feel about your future health.” Each item has four response choices: I don’t feel that way at all, I feel that way a little, I feel that way somewhat, and I feel that way strongly. The question-items were originally developed in Japanese.

Table 1 The 45 items developed to measure health-related hope

	Domain	Item content	% of missing	M	SD	Factor loading
a12	Something to live for	I think I’ll still be able to continue doing enjoyable things in the future.	0.9%	2.85	0.83	0.755
a18	Something to live for	I will probably be able to discover some meaning to my life.	0.9%	2.66	0.82	0.738
a31	Something to live for	I will probably be able to work, to the best of my ability, to find a sense of purpose in life.	0.9%	2.98	0.82	0.708
a21	Something to live for	I will probably be able to live each day to its fullest.	0.9%	2.97	0.80	0.683
a02	Something to live for	I feel that I can continue to experience a sense of fulfillment in my daily life.	0.9%	2.88	0.84	0.679
a07	Health & Illness	Even if I should feel down due to my illness, I could probably turn my feelings around.	0.9%	2.79	0.77	0.735
a28	Health & Illness	I feel I can adjust my health goals in a way that is consistent with my actual disease condition.	0.9%	2.80	0.80	0.733
a11	Health & Illness	I can probably alter my goals depending on changes in my illness or symptoms.	0.9%	2.70	0.82	0.711
a03	Health & Illness	Even if my health condition keeps me from achieving my present goals, I will probably be able to find a new goal.	0.9%	2.60	0.84	0.679
a23	Health & Illness	I can probably develop a personal lifestyle strategy for dealing with my disease condition.	0.9%	2.78	0.73	0.677
b12	Health & Illness	I will probably be able to find a way to keep my illness from worsening.	3.1%	2.56	0.82	0.618
a24	Role & Connectedness	I will probably be able to continue performing my role in society.	1.1%	2.45	0.87	0.646
e05	Role & Connectedness	I’ll probably be able to continue my usual role in support of my family.	1.1%	3.04	0.79	0.614
a05	Role & Connectedness	My disease experience will probably encourage those around me to be mindful of their own health.	0.9%	2.90	0.88	0.452
a14	Role & Connectedness	I feel I can deepen my relationships with my friends.	0.9%	2.84	0.84	0.662
a16	Role & Connectedness	Those around me will probably go along with any changes in my mood.	0.9%	2.63	0.82	0.644
e03	Role & Connectedness	I feel that I’ll continue to have a good relationship with my family.	1.1%	3.34	0.78	0.572
a04	Role & Connectedness	Those around me will probably continue to treat me the same way they always have.	0.9%	3.13	0.75	0.524
a06	Something to live for	I think I’ll still be able to live enjoyably in the future.	0.9%	2.61	0.95	0.765
a26	Something to live for	I think I can discover joy in my daily life.	0.9%	2.83	0.82	0.759
a29	Something to live for	I feel things will happen in the future that will bring me joy.	1.1%	2.91	0.87	0.712
a27	Something to live for	I would like to treasure my sense of fulfillment in life even if I’m sick.	0.9%	3.20	0.79	0.681
a30	Something to live for	I’d like to make an effort to experience events that will bring me joy.	0.9%	3.09	0.81	0.637
a17	Health & Illness	I can probably find a way to achieve my goals.	1.1%	2.63	0.79	0.719
b14	Health & Illness	I feel I will be able to keep my illness from worsening.	3.1%	2.55	0.85	0.658
b13	Health & Illness	I feel I will be able to maintain my current physical condition.	3.1%	2.62	0.80	0.637
a15	Health & Illness	My doctor will probably be able to provide me with a way to cope with my illness.	1.3%	3.04	0.78	0.554
c12	Health & Illness	I feel that my illness will not grow any worse than it is now.	0.0%	1.98	0.95	0.547
c11	Health & Illness	My symptoms will probably improve.	0.3%	2.05	0.94	0.518
b11	Health & Illness	I feel that my illness will improve over its present state.	3.1%	2.98	0.82	0.708
a13	Health & Illness	My doctor will probably consider my concerns attentively.	0.9%	3.31	0.73	0.509
a20	Health & Illness	No matter how my illness or symptoms may change, I can probably accept it.	0.9%	2.86	0.77	0.496
e01	Role & Connectedness	I’ll probably be able to fulfill any expected role in support of my family.	1.1%	2.93	0.81	0.533
e06	Role & Connectedness	My family will probably accept me when I’m ill.	1.3%	3.38	0.78	0.570
a08	Role & Connectedness	In the future, I hope to meet and interact with more new people.	1.1%	2.62	0.96	0.558
a22	Role & Connectedness	Even if my illness should progress, I feel my interactions with friends won’t decrease.	0.9%	2.51	0.92	0.550
a09	Role & Connectedness	The people around me will probably help me.	1.1%	2.72	0.86	0.535
e02	Role & Connectedness	I feel that my family will continue to support me.	1.1%	3.37	0.74	0.520
e04	Role & Connectedness	My family will probably discuss my health with me.	1.3%	3.28	0.82	0.519
a25	Role & Connectedness	I want to continue interacting with my friends.	0.9%	3.18	0.81	0.506
a01	Something to live for	No matter what happens with my health condition, I would like to continue living the way I do at present.	0.9%	2.89	0.89	0.275
d12	Role & Connectedness	I feel that the number of things I can do will increase in the future compared to now.	0.6%	1.77	0.90	0.351
d11	Role & Connectedness	I’ll probably be able to accomplish in the future those things I cannot accomplish in my daily life at present.	0.0%	1.80	0.88	0.320
a10	Role & Connectedness	I would like to try not to burden those around me with my illness.	0.9%	3.44	0.72	0.291
a19	other	I probably have nothing to worry about financially regarding continuing to receive medical treatment and nursing care.	0.9%	2.51	0.98	0.281

This shows the 45 items for measuring health-related hope constructed on the basis of the results of the qualitative study [7]. They are to be preceded by the following stem, which is used to establish the common context for all: “Please answer the questions below while keeping in mind how you feel about your future health.” Each item has four response choices: I don’t feel that way at all, I feel that way a little, I feel that way somewhat, and I feel that way strongly. The first 18 items constitute the final version of the Health-Related Hope (HR-Hope) scale.

PARTICIPANTS

Adults with CKD at stage 3 or higher (i.e. estimated glomerular filtration rate at or below 60 mL/min/1.73m²) who did not require dialysis, and those with CKD stage 5D who were receiving hemodialysis or peritoneal dialysis participated in this cross-sectional survey (the “Step 2” referred to in Fig. 1). All the procedures described after this section (i.e., the “Step 2” and “Step 3” referred to in Fig. 1) were done using the same participants’ data. In these patients the disease is usually incurable, the clinical course is indolent, deterioration can be delayed by dietary restriction and adherence to prescriptions for medications, and in Japan the crude mortality rate is lower than that due to advanced cancer (the annual death rate among dialysis patients in Japan was 9.8% in 2013) [8]. CKD patients with advanced cancer who were likely to die within one year and those with psychiatric disease were not included. A total of 454 patients from 5 institutions participated. The data were collected between February 2016 and September 2017 (Table 2).

Table 2 Baseline characteristics of the participants, n = 454

	Total* n = 454
Age, years	67.4 (13.8)
Women, %	143 [32%]
Performance Status
Requiring assistance at least occasionally	47 [10%]
Having family
No	47 [10%]
missing n = 5
Treatment status
Pre-dialysis	126 [28%]
Peritoneal dialysis	105 [23%]
Hemodialysis	223 [49%]

*Only patients who responded to the questionnaire were included.

ANALYSES USED IN ITEM SELECTION (THE “STEP 2” REFERRED TO IN FIG. 1)

For each of the 45 items asking about health-related hope (Table 1), the mean score, the standard deviation of the scores, and the percentage of missing responses were computed. The factor structure among the items was examined using exploratory factor analysis (maximum-likelihood method). As is also mentioned below, 5 items had factor loadings less than 0.4 and so those 5 were removed, which left 40 items. In addition, although the items were written with the intention that they would be useful for measuring 3 domains of health-related hope, the results of exploratory factor analysis suggested unidimensionality.

Therefore, unidimensional item-response theory (IRT) was applied to evaluate the 40 items that remained, to construct a shorter scale maintaining optimal precision and validity, and to evaluate the performance of the shorter scale [9]. As its name indicates, unidimensional IRT can be applied if the variable being measured is unidimensional [10, 11]. From among the many IRT models available, the generalized partial credit model (GPCM) was chosen because the responses to the health-related hope items are ordered [9]. Further details of the GPCM and its application are described in Supplementary Item 2 (Additional File 3).

CONSTRUCTION OF THE HEALTH-RELATED HOPE SCALE (THE “STEP 2” REFERRED TO IN FIG. 1)

To make an instrument that was shorter and more suitable for clinical use, we selected items on the basis of the results of the aforementioned item analysis, exploratory factor analysis, and IRT-based parameter estimates. Specifically, (1) each of the three domains should be represented, as should both prospects and self-efficacy for each domain (e.g., in the Role & Connectedness domain, the item “Those around me will probably continue to treat me the same way they always have.” measures prospects, whereas the item “I will probably be able to continue performing my role in society.” measures self-efficacy); (2) within each domain there should be items with widely different location parameters, i.e. a wide range of “difficulties” (Supplementary Item 2 (Additional File 3)); (3) items with higher values of the slope parameter were preferred; (4) items that were similar in their content to ones that had already been chosen were not included; (5) the number of items chosen should result in a scale that has adequate precision but still does not impose an unnecessary burden on the respondents. Application of those considerations resulted in an 18-item version of the HR-Hope scale (Table 3; the final Japanese version is provided in Supplementary Table 1 (Additional File 4)). Next, a test-information curve was constructed to facilitate examination of the scores’ precision across the entire range of health-related hope, which was expressed as θ values estimated using the GPCM.

Table 3 Slope and location parameters of items for the Health-Related Hope scale (both the 40-item beta version and the 18-item final version) estimated from a unidimensional item-response-theory model

	Domain	Item content	beta version 40 items		final version 18 items
			slope	location	slope	location
a12	Something to live for	I think I’ll still be able to continue doing enjoyable things in the future.	2.18	−0.36	1.27	−0.51
a18	Something to live for	I will probably be able to discover some meaning to my life.	2.00	−0.14	1.28	−0.20
a31	Something to live for	I will probably be able to work, to the best of my ability, to find a sense of purpose in life.	1.85	−0.45	0.94	−0.73
a21	Something to live for	I will probably be able to live each day to its fullest.	1.63	−0.53	1.06	−0.70
a02	Something to live for	I feel that I can continue to experience a sense of fulfillment in my daily life.	1.59	−0.41	0.97	−0.54
a07	Health & Illness	Even if I should feel down due to my illness, I could probably turn my feelings around.	2.28	−0.31	1.35	−0.35
a28	Health & Illness	I feel I can adjust my health goals in a way that is consistent with my actual disease condition.	2.10	−0.30	1.27	−0.40
a11	Health & Illness	I can probably alter my goals depending on changes in my illness or symptoms.	1.90	−0.20	1.19	−0.22
a03	Health & Illness	Even if my health condition keeps me from achieving my present goals, I will probably be able to find a new goal.	1.53	−0.15	1.00	−0.07
a23	Health & Illness	I can probably develop a personal lifestyle strategy for dealing with my disease condition.	2.13	−0.23	1.33	−0.34
b12	Health & Illness	I will probably be able to find a way to keep my illness from worsening.	1.54	−0.03	0.72	−0.07
a24	Role & Connectedness	I will probably be able to continue performing my role in society.	1.31	0.00	0.81	0.12
e05	Role & Connectedness	I’ll probably be able to continue my usual role in support of my family.	1.31	−0.46	0.57	−0.81
a05	Role & Connectedness	My disease experience will probably encourage those around me to be mindful of their own health.	0.73	−0.47	0.44	−0.70
a14	Role & Connectedness	I feel I can deepen my relationships with my friends.	1.41	−0.33	0.83	−0.47
a16	Role & Connectedness	Those around me will probably go along with any changes in my mood.	1.39	−0.11	0.84	−0.07
e03	Role & Connectedness	I feel that I’ll continue to have a good relationship with my family.	1.09	−0.93	0.63	−1.94
a04	Role & Connectedness	Those around me will probably continue to treat me the same way they always have.	1.01	−0.73	0.66	−1.13
a06	Something to live for	I think I’ll still be able to live enjoyably in the future.	1.52	−0.14
a26	Something to live for	I think I can discover joy in my daily life.	1.92	−0.30
a29	Something to live for	I feel things will happen in the future that will bring me joy.	1.61	−0.36
a27	Something to live for	I would like to treasure my sense of fulfillment in life even if I’m sick.	1.76	−0.73
a30	Something to live for	I’d like to make an effort to experience events that will bring me joy.	1.32	−0.72
a17	Health & Illness	I can probably find a way to achieve my goals.	2.19	−0.15
b14	Health & Illness	I feel I will be able to keep my illness from worsening.	1.55	−0.06
b13	Health & Illness	I feel I will be able to maintain my current physical condition.	1.55	−0.15
a15	Health & Illness	My doctor will probably be able to provide me with a way to cope with my illness.	1.12	−0.43
c12	Health & Illness	I feel that my illness will not grow any worse than it is now.	0.99	0.53
c11	Health & Illness	My symptoms will probably improve.	0.97	0.40
b11	Health & Illness	I feel that my illness will improve over its present state.	0.55	0.66
a13	Health & Illness	My doctor will probably consider my concerns attentively.	0.99	−1.20
a20	Health & Illness	No matter how my illness or symptoms may change, I can probably accept it.	1.00	−0.37
e01	Role & Connectedness	I’ll probably be able to fulfill any expected role in support of my family.	0.91	−0.55
e06	Role & Connectedness	My family will probably accept me when I’m ill.	0.89	−1.14
a08	Role & Connectedness	In the future, I hope to meet and interact with more new people.	0.81	−0.09
a22	Role & Connectedness	Even if my illness should progress, I feel my interactions with friends won’t decrease.	0.93	−0.08
a09	Role & Connectedness	The people around me will probably help me.	0.99	−0.25
e02	Role & Connectedness	I feel that my family will continue to support me.	0.87	−1.14
e04	Role & Connectedness	My family will probably discuss my health with me.	0.91	−1.35
a25	Role & Connectedness	I want to continue interacting with my friends.	0.91	−0.87

This instrument is protected by copyright: iHope International. All rights reserved.

When this instrument is used, please register through https://www.sf-36.jp/.

 

In addition, please cite this article as a reference:

Fukuhara S, Kurita N, Wakita T, Green J, Shibagaki Y.

A scale for measuring health-related hope: its development and psychometric testing.

Annals of Clinical Epidemiology 2019;1(3):102–119

CONFIRMATORY FACTOR ANALYSIS (THE “STEP 3” REFERRED TO IN FIG. 1)

Because the results of the qualitative study indicated that there were three domains but the results of the exploratory factor analysis indicated that there was only one, we used confirmatory factor analysis to evaluate the fit between the data and a second-order factor model. In that model, the first-order factors are the three domains identified in the qualitative study, and they are influenced by HR-Hope as the single second-order factor.

SCALE RELIABILITY AND SCALE SCORING (THE “STEP 3” REFERRED TO IN FIG. 1)

Internal consistency reliability was quantified as coefficient alpha. The scale score was computed as the average of the item scores. For patients who said that they had no family, the 2 items related to family were not applicable (Table 3, item e05 and e03), so those patients’ scores were computed as the average of the remaining 16 items.

MEASUREMENTS USED IN VALIDATION TESTING (THE “STEP 3” REFERRED TO IN FIG. 1)

The questionnaire included the Japanese version of Snyder’s hope scale [2, 12], the Center for Epidemiologic Studies Depression Scale (CES-D) [13, 14], and the SF-36v2 [15]. Presence of family was asked with the item “Do you have any family?” (responses: “yes” or “no”).

Snyder’s hope scale is intended to measure a “pathway” domain, which captures thoughts about one’s perceived ability to find a route to achieve one’s desired goal, and also an “agency” domain, which captures the perception that one will execute goal-oriented action [2, 12]. In this study, the internal consistency reliability (coefficient alpha) of Snyder’s hope scale was 0.76 for the pathway domain and 0.68 for the agency domain [12], and for the CES-D it was 0.84 [13]. Responses to the SF-36 were used to compute scores on 8 domains: physical function (PF), role physical (RP), bodily pain (BP), general health (GH), vitality (VT), social function (SF), role emotional (RE), and mental health (MH) [16].

Information collected from medical records included age, sex, and treatment (non-dialysis, hemodialysis, or peritoneal dialysis). Performance status was assessed by attending physicians, using the Zubrod Scale [17]. Scores on that scale range from 0 (normal activity) to 4 (bedridden). We defined impaired performance status as a Zubrod scale score of 2 (ambulatory >50% of waking hours; occasional assistance) or greater.

At one of the medical centers from which the participants were recruited, each patient’s physician or a clinical psychologist rated the “acceptance stage” of patients with CKD stage 5D who were receiving peritoneal dialysis. (Supplementary Table 2 (Additional File 5)) Through dialogue between the patients and a physician or a clinical psychologist, patients’ acceptance stage was routinely rated and recorded in the medical record. The acceptance stage was re-evaluated at the time of the questionnaire survey, with the evaluator unaware of the responses on the questionnaire. It was then defined as high-acceptance if the rating contained at least “reconciliation” or “acceptance” and low-acceptance if the rating contained neither of those two.

ANALYSES USED IN VALIDATION TESTING (THE “STEP 3” REFERRED TO IN FIG. 1)

For criterion-validation testing, we computed Pearson’s correlation coefficients for the associations of the HR-Hope scale scores with scores on other scales measuring related concepts: Snyder’s hope scale, the CES-D, and each of the 8 SF-36 domains. Similarly, we computed Pearson’s correlation coefficients for the associations of Snyder’s hope scale with scores on the CES-D and each of the 8 SF-36 domains. Next, we tested the null hypotheses that the correlations between HR-Hope scale scores and CES-D scores were the same as the correlations between Snyder’s hope scale scores and CES-D scores [18].

We estimated standardized differences between pairs of clinical groups for the HR-Hope scale’s total score and for each of its 3 domains, and Snyder’s pathway and agency domains, using unadjusted general linear models. The clinical groups used in these analyses were defined by Performance Status (impaired or not impaired), depressed state (a CES-D score of 16 or more), acceptance (high or low), and having family (yes or no). In addition, to compare the validity of the aforementioned scales, the Relative Validity (RV) was calculated using general linear models. The RV was defined as the ratio of the F statistic for the scales measuring each of the aforementioned domains divided by the F statistic for the HR-Hope total scores as the referent scale [19]. The RV indicates how much better or worse each scale is than the HR-Hope total score, for discriminating between patients who have different values on the 4 indices of clinical status. In this context, “better” is indicated by an RV greater than 1, while “worse” is indicated by an RV less than 1. The scales were the 3 domains of HR-Hope, and Snyder’s pathway and agency domains.

STATISTICAL SOFTWARE

Exploratory factor analysis, IRT-based analyses, and confirmatory factor analysis were done using R and PARSCALE 4.0 (Scientific Software International, Skokie, IL, USA). The other analyses were done using Stata/SE version 14 (Stata Corp., College Station, TX).

RESULTS

PARTICIPANTS

Characteristics of the 454 participants are shown in Table 2. The mean (SD) norm-based scores on the SF-36 domains were: PF 35.5 (19.0); RP 37.4 (16.3); BP 46.4 (11.4); GH 40.3 (10.6); VT 47.0 (10.6); SF 44.3 (13.6); RE 41.6 (15.2); and MH 49.3 (10.2). All scores but MH were lower than their respective national mean (by definition, for national norm scores the mean was 50.0 for all 8 domains) [16].

ITEM ANALYSIS, EXPLORATORY FACTOR ANALYSIS, AND IRT ANALYSIS

There were no ceiling or floor effects on any item. Missing responses ranged 0.0% to 3.1% (Table 1). Exploratory factor analysis indicated a strong unidimensionality: the first seven eigenvalues were 16.9, 2.8, 2.5, 1.8, 1.6, 1.4, and 1.3, and the first factor explained 36.4% of the variance in the data.

The factor loadings varied from 0.270 to 0.765. We decided to remove the 5 items with factor loadings less than 0.4. The remaining 40 items comprised the β version of the HR-Hope scale. There were no marked differences in the mean or in the standard deviation among those 40 items (Table 1).

Next, the IRT-based analysis was applied to the 40-item β version and then items for the final, shorter version were selected, resulting in the 18-item HR-Hope scale (Table 3, Supplementary Fig. 2 (Additional File 6)). With regard to the 18-item HR-Hope scale, the test information exceeded 9 from θ values of -2.4 to 1.87, and the peak of the test-information curve was 19.89 at a θ value of -0.8 (Supplementary Fig. 3 (Additional File 7)).

CONFIRMATORY FACTOR ANALYSIS

Confirmatory factor analysis of the second-order factor model resulted in a root mean square error of approximation of 0.051, a comparative fit of index of 0.947, and a standardized root mean square residual of 0.041, which indicate a good fit between the data and the model (Supplementary Fig. 4 (Additional File 8)).

RELIABILITY TESTS AND DISTRIBUTION OF SCORES

The distribution of the HR-Hope scores was approximately normal, and the mean and standard deviation were 2.80 and 0.56, respectively. The reliability of the 18-item HR-Hope scale was high: coefficient alpha = 0.93. The alpha coefficients of the 3-domains (health & illness, role & social connectedness, and “something to live for”) were 0.86, 0.74, and 0.86, respectively.

VALIDATION TESTING

The correlation of the HR-Hope scale with Snyder’s pathway domain was 0.56 and with Snyder’s agency domain it was 0.61. Correlations of scores on the HR-Hope scale with scores on the scales measuring other related concepts are shown in Table 4. The correlation with the CES-D scale was moderate: r = −0.50. Correlations with the SF-36 domain scores were weak to moderate (r = 0.23 to 0.49). The scales measuring hope-related concepts were correlated more strongly with HR-Hope scores than with Snyder’s pathway or agency scores. For example, the absolute value of the correlation coefficient of the HR-Hope score with the CES-D (r = −0.503) was larger than the correlation of Snyder’s pathway (r = −0.368) or agency (r = −0.366) with the CES-D scale (p < 0.001 for both). Similarly, for all 8 of the SF-36 domain scores the correlations with HR-Hope scores were stronger than with Snyder’s pathway or agency.

Table 4 Correlations (Pearson’s r) between the Health-Related Hope scale, Snyder’s hope scale, and scales measuring hope-related concepts

	HR-Hope scale	Snyder’s hope scale		p-value* HR-Hope vs Pathway	p-value** HR-Hope vs Agency	n
		Pathway	Agency
CES-D	−0.503	−0.368	−0.366	<0.001	<0.001	430
PF	0.228	0.108	0.129	0.006	0.018	439
RP	0.239	0.131	0.133	0.013	0.011	442
BP	0.239	0.135	0.136	0.018	0.014	439
GH	0.458	0.262	0.350	<0.001	0.004	439
VT	0.492	0.322	0.393	<0.001	0.007	439
SF	0.333	0.156	0.132	<0.001	<0.001	437
RE	0.311	0.226	0.225	0.046	0.034	442
MH	0.448	0.276	0.300	<0.001	<0.001	438

*Results of tests of the null hypotheses that the correlation between each hope-related scale and the HR-Hope scale was not different from the correlation between that hope-related scale the Snyder-pathway scale [18].

**Results of tests of the null hypotheses that the correlation between each hope-related scale and the HR-Hope scale was not different from the correlation between that hope-related scale the Snyder-agency scale [18].

CES-D: the Center for Epidemiologic Studies Depression Scale, PF: physical function, RP: role physical, BP: bodily pain, GH: general health, VT: vitality, SF: social function, RE: role emotional, and MH: mental health.

Results of validation tests using the four indices of clinical status are shown in Tables 5 through 8. To detect differences in performance status, the HR-Hope “something to live for” domain had the highest relative validity (1.201), followed by the HR-Hope health & illness domain (1.149) and the HR-Hope total score. The general hope measures (Snyder’s pathway and agency) had relative validities of 0.313 to 0.322. To detect differences in depressed state, the HR-Hope “something to live for” domain was most valid (1.044), followed by the HR-Hope total score (Table 6). The general hope measures (Snyder’s pathway and agency) had the lowest relative validities (0.420 to 0.449). To detect acceptance, the HR-Hope total score was most valid, followed by the HR-Hope health & illness domain (0.987) and the HR-Hope “something to live for” domain (0.959) (Table 7). The general hope measures had the lowest relative validities (0.258 to 0.261). To detect differences with regard to having family, the HR-Hope role & social connectedness domain was the most valid (1.777), followed by the HR-Hope total score (Table 8). The general hope (Snyder’s pathway) score had the lowest relative validity (0.132) while the HR-Hope health & illness score had the second-lowest (0.378).

Table 5 Mean scores and relative validity statistics by performance status group

Measure	Mean (SD) score		std. Δ	F Statistic	p-value	RV
	Not impaired n = 402	Impaired n = 46
HR-Hope scale
Total	2.85 (0.54)	2.63 (0.67)	−0.40	6.53	0.001	1.000
Health & Illness	2.73 (0.60)	2.47 (0.76)	−0.42	7.50	0.006	1.149
Role & Connectedness	2.91 (0.55)	2.77 (0.64)	−0.25	2.56	0.110	0.393
Something to live for	2.90 (0.64)	2.62 (0.77)	−0.43	7.84	0.005	1.201
Snyder’s hope scale
Pathway	2.68 (0.53)	2.55 (0.68)	−0.23	2.10	0.148	0.322
Agency	2.65 (0.57)	2.52 (0.65)	−0.22	2.04	0.154	0.313

SD: standard deviation, std. Δ: standardized difference, RV: relative validity.

The RVs were calculated by dividing the F statistics for each of the domains listed in the leftmost column by the F statistic for the HR-Hope total score. The resulting RV values indicate how much better or worse each scale is than the HR-Hope total score, with regard to discriminating between patients of different performance status. In this context, “better,” i.e. a higher relative validity, is indicated by an RV that is greater than 1, while “worse,” i.e. a lower relative validity, is indicated by an RV that is less than 1. The scales were the 3 domains of HR-Hope, and Snyder’s pathway and agency domains.

Values of std. Δ were calculated by dividing the mean differences (Impaired–Not impaired) of each domain score by the standard deviation of that domain score in this study’s participants.

Table 6 Mean scores and relative validity statistics by “depressed state” group

Measure	Mean (SD) score		std. Δ	F Statistic	p-value	RV
	Not depressed n = 285	Depressed n = 145
HR-Hope scale
Total	2.99 (0.50)	2.50 (0.53)	−0.88	88.87	<0.001	1.000
Health & Illness	2.88 (0.57)	2.38 (0.57)	−0.81	73.31	<0.001	0.825
Role & Connectedness	3.03 (0.51)	2.62 (0.56)	−0.73	58.18	<0.001	0.655
Something to live for	3.07 (0.58)	2.48 (0.63)	−0.89	92.76	<0.001	1.044
Snyder’s hope scale
Pathway	2.78 (0.51)	2.45 (0.55)	−0.60	37.33	<0.001	0.420
Agency	2.76 (0.52)	2.41 (0.59)	−0.62	39.88	<0.001	0.449

SD: standard deviation, std. Δ: standardized difference, RV: relative validity.

The RVs were calculated by dividing the F statistics for each of the domains listed in the leftmost column by the F statistic for the HR-Hope total score. The resulting RV values indicate how much better or worse each scale is than the HR-Hope total score, with regard to discriminating between patients of different “depressed state” status. In this context, “better,” i.e. a higher relative validity, is indicated by an RV that is greater than 1, while “worse,” i.e. a lower relative validity, is indicated by an RV that is less than 1. The scales were the 3 domains of HR-Hope, and Snyder’s pathway and agency domains.

Values of std. Δ were calculated by dividing the mean differences (Depressed – Not depressed) of each domain score by the standard deviation of that domain score in this study’s participants.

Table 7 Mean scores and relative validity statistics by acceptance group

Measure	Mean (SD) score		std. Δ	F Statistic	p-value	RV
	Low acceptance n = 35	High acceptance n = 56
HR-Hope scale
Total	2.69 (0.48)	3.01 (0.51)	0.62	9.07	0.003	1.000
Health & Illness	2.60 (0.51)	2.95 (0.55)	0.62	8.95	0.004	0.987
Role & Connectedness	2.74 (0.52)	3.00 (0.54)	0.49	5.34	0.023	0.589
Something to live for	2.73 (0.57)	3.09 (0.56)	0.61	8.70	0.004	0.959
Snyder’s hope scale
Pathway	2.69 (0.40)	2.86 (0.55)	0.33	2.37	0.128	0.261
Agency	2.63 (0.49)	2.80 (0.56)	0.33	2.34	0.130	0.258

SD: standard deviation, std. Δ: standardized difference, RV: relative validity.

The RVs were calculated by dividing the F statistics for each of the domains listed in the leftmost column by the F statistic for the HR-Hope total score. The resulting RV values indicate how much better or worse each scale is than the HR-Hope total score, with regard to discriminating between patients of different acceptance status. In this context, “better,” i.e. a higher relative validity, is indicated by an RV that is greater than 1, while “worse,” i.e. a lower relative validity, is indicated by an RV that is less than 1. The scales were the 3 domains of HR-Hope, and Snyder’s pathway and agency domains.

Values of std. Δ were calculated by dividing the mean differences (High acceptance–Low acceptance) of each domain score by the standard deviation of that domain score in this study’s participants.

Table 8 Mean scores and relative validity statistics by family-status group

Measure	Mean (SD) score		std. Δ	F Statistic	p-value	RV
	Not having family n = 47	Having family n = 400
HR-Hope scale
Total	2.63 (0.54)	2.85 (0.56)	0.38	6.26	0.013	1.000
Health & Illness	2.57 (0.56)	2.72 (0.62)	0.24	2.36	0.125	0.378
Role & Connectedness	2.64 (0.56)	2.92 (0.55)	0.51	11.12	0.001	1.777
Something to live for	2.70 (0.65)	2.89 (0.66)	0.30	3.72	0.054	0.594
Snyder’s hope scale
Pathway	2.60 (0.45)	2.67 (0.56)	0.14	0.83	0.363	0.132
Agency	2.46 (0.48)	2.66 (0.58)	0.34	4.80	0.029	0.766

SD: standard deviation, std. Δ: standardized difference, RV: relative validity.

The RVs were calculated by dividing the F statistics for each of the domains listed in the leftmost column by the F statistic for the HR-Hope total score. The resulting RV values indicate how much better or worse each scale is than the HR-Hope total score, with regard to discriminating between patients who reported having family and those who reported not having family. In this context, “better,” i.e. a higher relative validity, is indicated by an RV that is greater than 1, while “worse,” i.e. a lower relative validity, is indicated by an RV that is less than 1. The scales were the 3 domains of HR-Hope, and Snyder’s pathway and agency domains.

Values of std. Δ were calculated by dividing the mean differences (Having family–Not having family) of each domain score by the standard deviation of that domain score in this study’s participants.

DISCUSSION

We developed and tested a scale to measure HR-Hope in adults with chronic disease. The 18-item version covers 3 domains: effects of health on hope regarding (a) health & illness; (b) role & social connectedness; and (c) “something to live for.” The 18-item HR-Hope scale performed well on criterion-validation tests and it had good internal-consistency reliability.

Considering content validity, we note that the HR-Hope scale incorporates items asking about the content of hope and process of hope experienced by adults with chronic illness—items that are not included in scales intended to measure hope in general. First, the HR-Hope scale contains 6 items regarding health & illness. For example, the importance of measuring hope regarding health has been noted [1], and that is addressed in item a28 (Table 3), which asks about the patients’ self-efficacy for adjusting treatment goals based on their present and future health conditions. Second, the HR-Hope scale reflects processes of hope that are specific to people with chronic illness. For example, item a11 concerns modification or replacement of the object of hope when the prospects of a person’s health condition make what is hoped for at present less likely. Third, items in the HR-Hope scale ask about objects of hope that can be achieved even in a short period of time, thus avoiding asking about more distant goals that may be irrelevant in light of a patient’s medical condition. For example, item a21 asks about “each day” rather than about long time periods that might not be realistic for people with an incurable illness. Unlike this item, the Herth Hope Index asks about both short-range and long-range hope [3]. Fourth, the HR-Hope scale asks about some objects of hope that might be taken for granted by healthy adults but are uncertain for people with chronic illness. For example, rather than asking about the prospect of making new interpersonal connections, item a04 asks about the prospect of maintaining existing interpersonal connections.

Compared with Snyder’s hope scale, the HR-Hope scale performed better on criterion-based validation tests. Specifically, the CES-D and the 8 domains of SF-36 correlated more strongly with the HR-Hope scale than with Snyder’s pathway or agency domains (Table 4). It is important to note that the correlations of the 8 SF-36 domains with the HR-Hope scale were weak to moderate (r < 0.5 for each domain), and thus the HR-Hope scale is not measuring the same construct as the SF-36, that is, it is not measuring health-related QOL.

Similarly, for all 4 clinical indices the results of RV tests were better for the HR-Hope total score than for Snyder’s pathway and agency scores, which indicates that the HR-Hope total score is more sensitive to these clinical variables. First, the association of lower HR-Hope total scores with impaired performance status suggests that HR-Hope is partly determined by activities in daily life. This hypothesis is further supported by the findings that relative validities of the domain scores for “something to live for” and for health & illness were greater than 1 (i.e., greater than the reference value, which was the total score). Another explanation would be that maintaining a normal life is determined by functional independence [20]. Second, the finding of higher HR-Hope scores among people who reported having a family shows that family is a vital component of hope among people with chronic illness. People without a family may have had no role to fill in relation to someone important to them. Also, without assistance from their family, after they became sick with a chronic disease they may have been less likely to continue fulfilling their role in society or with friends, to deepen their connectedness with their friends, or to adjust their mood. This hypothesis is supported by the finding that to differentiate between those with and those without a family the RV of the HR-Hope role & connectedness score was the highest (1.777). Third, the finding that the HR-Hope score was positively associated with “acceptance stage” shows that those two phenomena—HR-Hope and acceptance of chronic illness—are linked in at least some people with chronic medical conditions. When acceptance is low (i.e. when there has been a loss of willpower, or a refusal, or a struggle to face one’s disease), then people might be unable to attain a new balance between life goals and disease experience [21]. As coping with disease experience is associated with better adherence to self-management in CKD patients [22], HR-Hope may also have a positive influence on adherence through better coping.

The results of the IRT-based analyses indicate that the 18-item version the HR-Hope scale will be able to measure health-related hope quite precisely, as the θ values at which test information exceeded 9 ranged from −2.4 to 1.87, and the estimated HR-Hope values of 96% of the study subjects were within that range.

Several limitations of this study warrant mention. First, the participants for psychometric testing of the HR-Hope scale were limited to patients with CKD. We tried to include both patients who were undergoing dialysis and those who were not undergoing dialysis, to reflect experiences of different symptoms, of different levels of symptom severity, and of different levels of functional status. Nonetheless, the psychometric properties of the HR-Hope scale should be further examined among patients with other chronic conditions. Second, whether our conceptual framework for HR-Hope and the question-items we used can be applied in other cultural contexts also requires further investigation, as cultural and religious differences may contribute to experiences of health-related hope. So that it may be tested in other countries, the HR-Hope scale has already been translated into English, and the conceptual equivalence of that English version has been checked via back-translation and consultation with the authors of the original Japanese version.

CONCLUSION

We developed the HR-Hope scale for adults with chronic illness and tested its psychometric properties. The results suggest that the HR-Hope scale can be applied to the intended population. This scale may be useful both as an outcome measure and as a stimulus to enhance self-management strategies, thereby helping to alleviate the suffering associated with disease complications.

CONFLICTS OF INTEREST

The authors declare no competing interests.

ACKNOWLEDGMENTS

This study was supported by JSPS KAKENHI (Grant Number: JP16H05216). The authors greatly thank the following researchers, research assistants, and medical staff members for their assistance in collecting the questionnaire-based and clinical information used in this study: Yoshitaka Ishibashi, MD, Mai Yanai, MD, Shino Fujimoto, MSc (Japan Red Cross Medical Center, Sibuya-ku, Tokyo), Masahiko Yazawa, MD, Kenichiro Koitabashi, MD, Tomo Suzuki, MD, Ms. Asako Tamura, Ms. Yuka Masuda, Ms. Takae Shimizu (St. Marriana University, Kawaaski-city, Kanagawa), Hiroo Kawarazaki, MD (Inagi Municipal Hospital, Inagi-city, Tokyo), Takayuki Nakamura, MD, Eiko Hashimoto, RN (JCHO Nihonmatsu Hospital, Nihonmatsu-city, Fukushima), Atsushi Kyan, MD, Masashi Saito, CE (Shirakawa Kosei General Hospital, Shirakawa-city, Fukushima), Ms. Lisa Shimokawa (Fukushima Medical University, Fukushima-city, Fukushima).

LIST OF ABBREVIATIONS

QOL: quality of life; HR-QOL: health-related quality of life; CKD: chronic kidney disease; CES-D: the Center for Epidemiologic Studies Depression Scale; PF: physical function; RP: role physical; BP: bodily pain; GH: general health; VT: vitality; SF: social function; RE: role emotional; MH: mental health; IRT: item-response theory; GPCM: generalized partial credit model; RV: Relative Validity.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE

Participants were included only if they gave written informed consent. This study’s protocols were approved by the ethics review boards of St. Marianna University (Number 3209) and Fukushima Medical University (Number 2417).

CONSENT FOR PUBLICATION

Not applicable.

AUTHORS’ CONTRIBUTIONS

S.Y. obtained funding; S.F., N.K., and T.W. participated in conception and design; N.K. and T.W. performed statistical analyses; S.F., N.K., T.W., J.G., and Y.S. analyzed and interpreted the data; N.K. and Y.S. collected data; S.F., N.K., T.W. and J.G. wrote the manuscript; S.F., N.K., T.W., J.G., and Y.S. reviewed the manuscript. S.F. and N.K. equally contributed to this study and share co-first authorship.

Supplementary Figure 1, Additional File 1 The difference between general quality of life and health-related quality of life.

The figure illustrates the difference between quality of life in general and health-related quality of life (Reprinted with permission) [1]. This figure shows an extension of the concept of health-related quality of life that was described by Ware [2]. In personal communications by e-mail on March 1, 2018, J. E. Ware stated that health-related quality of life consists of perceived functioning (what one thinks one is able to do), well-being (how one feels), and personal evaluation of health (what one says with regard to one’s health).

Measuring quality of life in general requires assessing self-perception of all aspects of a person’s life. Quality of life in general is affected not only by health but also by, for example, the safety of the environment, freedom, and income, which are distant from health per se. The difference between quality of life in general and health-related quality of life is analogous to the difference between hope in general and health-related hope.

 

References

1. Ikegami N, Fukuhara S, Shimozuma K, Ikeda S. In: The handbook of assessment of QOL for clinical practice. 2001, Igaku-shoin, Tokyo (in Japanese).

2. Ware JE. Measuring patient function and well-being: Some lessons from the Medical Outcomes Study. In: Lohr K, Rettig R, eds. Effectiveness and Outcomes in Health, Proceedings of an Invitational Conference by the Institute of Medicine. pp. 107–19, 1990, National Academy Press, Washington DC.

Supplementary Item 1, Additional File 2 Summary of the previous qualitative study referred to in Figure 1.

References

1. Wakita T, Kurita N, Tominaga N, Kato K, Konno S, Fukuhara S, et al. Hope in Adult Patients with Chronic Disease: Qualitative Study. Kansai Univ Psychol Res. 2016;17–33 (in Japanese except for English abstract).

Supplementary Item 2, Additional File 3 Details of the generalized partial credit model and its application.

Supplementary Table 1, Additional File 4 Japanese version of the 18-item Health-Related Hope scale.

Stem	これからのご自身の健康状態をふまえて、以下の質問にお答えください。
Question 1	これからも楽しみを持ち続けることができると思う。
Question 2	自分が生きる意味を見つけられるだろう。
Question 3	自分のできる範囲で、生きがいを見つけられるだろう。
Question 4	毎日を大切に過ごしていけるだろう。
Question 5	日々の暮らしの中で、生きがいを感じていけると思う。
Question 6	病気によって気分が落ち込むことがあっても、気持ちを切り替えられるだろう。
Question 7	病状に応じて、現実的な健康上の目標を決められると思う。
Question 8	私は、病気や症状の変化に応じて、目標を修正できるだろう。
Question 9	病気によって目標が達成できなくなっても、また新たな目標を見つけられるだろう。
Question 10	病状に併せて、自分なりに生活の工夫ができるだろう。
Question 11	病気を悪化させないための方法を探すことができるだろう。
Question 12	現在、社会の中で果たしている役割をこれからも続けられるだろう。
Question 13	私の病気の体験を知ることで、周囲の人も健康を気遣うようになるだろう。
Question 14	友人とより良い関係を作る事ができると思う。
Question 15	私の周囲の人は私の気分転換に付き合ってくれるだろう。
Question 16	私の周囲の人はこれからも今まで通り接してくれるだろう。
Response options for Questions 1 through 16	全くそう思わない 少しそう思う ある程度そう思う とてもそう思う
Question regarding family	ご家族はいらっしゃいますか？
Response options	はい いいえ
Question 17	現在、家族の中で果たしている役割をこれからも続けられるだろう。
Question 18	今後も家族との良い関わりが続くと思う。
Response options for Questions 17 and 18	全くそう思わない 少しそう思う ある程度そう思う とてもそう思う

This instrument is protected by copyright: iHope International. All rights reserved.

When this instrument is used, please register through https://www.sf-36.jp/.

 

In addition, please cite this article as a reference:

Fukuhara S, Kurita N, Wakita T, Green J, Shibagaki Y.

A scale for measuring health-related hope: its development and psychometric testing.

Annals of Clinical Epidemiology 2019;1(3):102–119

Supplementary Table 2, Additional File 5 The concept of acceptance stage regarding chronic disease.

Stage	Construct	Examples of patients’ behaviors and words
Loss	Psychological state in which one is losing “willpower” to face one’s disease and life	Speechless or expressionless during an office visit ‘Something to live for? ... Nothing for me’
Refusal	Psychological state in which one ignores or escapes from dealing with one’s illness, to avoid psychological distress associated with illness management.	‘So picky about my illness condition. Treating it is your business, isn’t it?’ ‘My dialysis therapy is left to my daughter’
Struggle	Psychological state in which one is too concentrated on coping with one’s disease to enjoy life. Sometimes the patient is obsessed with conquering the disease despite negligible curability and may alternate between joy and worry, for example when seeing the results of one’s medical tests.	‘Doctor, what is my renal function today?’ ‘I don’t want to be inferior to others because of my illness’
Reconciliation	Psychological state in which one is able to deal with one’s fear associated with disease, but on the other hand one may sometimes fail to care for oneself adequately (Note: This is “adequacy” as seen from a medical standpoint. It includes avoidance of medically undesirable habits, etc.).	‘I feel I choose treatment and self-management which is suitable to myself’ ‘I enjoy my hobby when I don’t need treatment’
Acceptance	Psychological state in which one has accepted one’s disease and fear, and enjoys something meaningful in life.	‘I’m grateful to everyone’ ‘I feel I received a new life’

The concept of acceptance stage was developed on the basis of Cohn’s theory of the 5 stages of adjustment to loss as applied to disabled persons (i.e. shock, expectancy of recovery, mourning, defense, and adjustment) [1] and the observed psychological trajectories of patients with CKD requiring dialysis therapy. From low to high levels of acceptance, the 5 acceptance stages are loss, refusal, struggle, reconciliation, and acceptance. As the acceptance stages are not necessarily mutually exclusive and could vary even within one day, overlapping descriptions such as “refusal—struggle stage” or “struggle—reconciliation stage” were allowed. A patient was considered to be at a high-acceptance stage if that patient’s rating contained at least “reconciliation” or “acceptance”. A patient was considered to be at a low-acceptance stage if that patient’s rating contained neither “reconciliation” nor “acceptance”. This cut-off was believed to be clinically meaningful for the physicians and clinical psychologists in the medical center who routinely evaluated acceptance stage in their daily practice. From their experience, they felt that patients at high acceptance stages tended to adapt to and continue cognitive-behavioral therapy for the dietary and fluid management prescribed in dialysis care [2].

 

References

1. Cohn N. Understanding the process of adjustment to disability. J Rehabil. 1961;27:16–8.

2. Fujimoto S, Muto T. Acceptance stage. In: Ishibashi Y, ed. Clinical practice that inevitably succeeds for chronic kidney disease and peritoneal dialysis: TRC. pp. 57–73, 2016, Chugai Medical Publishing Co, Tokyo (in Japanese).

Supplementary Figure 2, Additional File 6 Scatterplot of the slope and location parameters (based on item-response theory) of the 40 items in the β version of the HR-Hope scale.

Solid circles indicate the items chosen for the final 18-item version of the Health-Related Hope scale. Each upper-case letter indicates the domain that the item belongs to. H: health & illness, R: role & social connectedness, and I: Ikigai, “something to live for.” Each lower-case letter followed by a two-digit number identifies the item as listed in Table 3.

Supplementary Figure 3, Additional File 7 Test information curve for the 18-item HR-Hope scale.

The test information exceeded 9 from θ values of −2.4 to 1.87. The peak of the curve is at 19.89, at θ = −0.8.

Supplementary Figure 4, Additional File 8 Structural diagram of the model tested by confirmatory factor analysis.

This Figure shows the structure of the second-order factor model that was evaluated using confirmatory factor analysis. The first-order factors are the three domains that were identified in the qualitative work: Health & Illness, Role & Connectedness, and “Something to live for.” Those three are influenced by health-related hope (HR-Hope) as the single second-order factor. The Root Mean Square Error of Approximation was 0.051, the Comparative Fit of Index was 0.947, and the Standardized Root Mean square Residual was 0.041, which indicate a good fit between the data and the model. The estimated coefficient is shown for each path.

REFERENCES

• 1.   Bressan  RA,  Iacoponi  E,  Candido de Assis  J,  Shergill  SS. Hope Is a Therapeutic Tool. BMJ (Clinical Research Ed) 2017;359.
• 2.   Snyder  CR,  Harris  C,  Anderson  JR,  Holleran  SA,  Irving  LM,  Sigmon  ST, et al. The Will and the Ways: Development and Validation of an Individual-Differences Measure of Hope. J Pers Soc Psychol 1991;60:570–585.
• 3.   Herth  K. Abbreviated Instrument to Measure Hope: Development and Psychometric Evaluation. J Adv Nurs 1992;17:1251–1259.
• 4.   Miller  JF,  Powers  MJ. Development of an Instrument to Measure Hope. Nurs Res 1988;37:6–10.
• 5.   Beck  AT,  Weissman  A,  Lester  D,  Trexler  L. The Measurement of Pessimism: The Hopelessness Scale. J Consult Clin Psychol 1974;42:861–865.
• 6.   Shi  L,  Hasegawa  H. Bayesian Analysis of Happiness with Individual Heterogeneity. Behaviormetrika 2019;46:23–48.
• 7.  Wakita T, Kurita N, Tominaga N, Kato K, Konno S, Fukuhara S, et al. Hope in Adult Patients with Chronic Disease: Qualitative Study. Kansai Univ Psychol Res 2016;17–33 (in Japanese except for English abstract).
• 8.   Masakane  I,  Nakai  S,  Ogata  S,  Kimata  N,  Hanafusa  N,  Hamano  T, et al. An Overview of Regular Dialysis Treatment in Japan (as of 31 December 2013). Ther Apher Dial 2015;19:540–574.
• 9.   Edelen  MO,  Reeve  BB. Applying Item Response Theory (IRT) Modeling to Questionnaire Development, Evaluation, and Refinement. Qual Life Res 2007;16:5.
• 10.  Embretson SE, Reise SP. In: Item Response Theory for Psychologists. 2000, Erlbaum, Mahwah NJ.
• 11.   Muraki  E. A Generalized Partial Credit Model: Application of an EM Algorithm. Appl Psychol Meas 1992;16:159–176.
• 12.   Kato  T,  Snyder  CR. The Relationship between Hope and Subjective Well-Being: Reliability and Validity of the Dispositional Hope Scale, Japanese Version. Jpn J Psychol 2005;76:227–234 (in Japanese).
• 13.   Shima  S,  Shikano  T,  Kitamura  T,  Asai  M. New Self-Rating Scale for Depression. Clin Psychiatry 1985;27:717–723 (in Japanese).
• 14.   Radloff  LS. The CES-D Scale. Appl Psychol Meas 1977;1:385–401.
• 15.   Fukuhara  S,  Bito  S,  Green  J,  Hsiao  A,  Kurokawa  K. Translation, Adaptation, and Validation of the SF-36 Health Survey for Use in Japan. J Clin Epidemiol 1998;51:1037–1044.
• 16.  Fukuhara S, Suzukamo Y. In: Manual of SF-36v2 Japanese Version. 2015, iHope International Inc., Kyoto (in Japanese).
• 17.   Oken  MM,  Creech  RH,  Tormey  DC,  Horton  J,  Davis  TE,  McFadden  ET, et al. Toxicity and Response Criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol 1982;5:649–656.
• 18.   Diedenhofen  B,  Musch  J. Cocor: A Comprehensive Solution for the Statistical Comparison of Correlations. PLoS One 2015;10:e0121945.
• 19.   McHorney  CA,  Ware  JE, Jr,  Raczek  AE. The MOS 36-Item Short-Form Health Survey (SF-36): II. Psychometric and Clinical Tests of Validity in Measuring Physical and Mental Health Constructs. Med Care 1993;31:247–263.
• 20.   Mok  E,  Lam  WM,  Chan  LN,  Lau  KP,  Ng  JS,  Chan  KS. The Meaning of Hope from the Perspective of Chinese Advanced Cancer Patients in Hong Kong. Int J Palliat Nurs 2010;16:298–305.
• 21.   Poppe  C,  Crombez  G,  Hanoulle  I,  Vogelaers  D,  Petrovic  M. Improving Quality of Life in Patients with Chronic Kidney Disease: Influence of Acceptance and Personality. Nephrol Dial Transplant 2013;28:116–121.
• 22.   Palmer  SC,  Hanson  CS,  Craig  JC,  Strippoli  GF,  Ruospo  M,  Campbell  K, et al. Dietary and Fluid Restrictions in CKD: A Thematic Synthesis of Patient Views from Qualitative Studies. Am J Kidney Dis 2015;65:559–573.