Development and psychometric properties of the ‘Assessment of Burden of Chronic Conditions (ABCC) tool’ for people with chronic heart failure (CHF): a mixed methods approach

Abstract

Objective

To develop a chronic heart failure (CHF) module for the Assessment of Burden of Chronic Conditions (ABCC) tool and evaluate its psychometric properties.

Design

Mixed methods study.

Setting

The study was conducted in the Netherlands.

Participants

Healthcare providers (ie, general practitioners (GPs), practice nurses, cardiologists, heart failure nurse practitioners, heart failure specialist nurses) and patients with CHF were involved in the developmental phase. In the psychometric evaluation, patients with CHF were included.

Results

The ABCC-CHF module comprised 26 items across CHF-specific domains, accompanied by associated advice. It demonstrated high convergent validity with the Kansas City Cardiomyopathy Questionnaire (KCCQ) (97%) and successfully differentiated most known-groups. Internal consistency yielded a Cronbach’s α of 0.92 for the total scale, while test-retest reliability showed an intraclass correlation coefficient (ICC) of 0.97.

Conclusion

The CHF-specific module is developed and validated. It aims to integrate disease burden into routine healthcare, supporting person-centred care for people with CHF and multimorbidity.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Chronic heart failure (CHF) is highly prevalent and associated with substantial physical, emotional and social burden, contributing to reduced health-related quality of life (HRQoL).

• Existing patient-reported outcome measures (PROMs) mainly focus on specific aspects of disease impact and are not designed to support integrated, person-centred care.

WHAT THIS STUDY ADDS

• This study provides a valid and reliable tool to assess and visualise perceived disease burden in people with CHF, including those with multimorbidity.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE, OR POLICY

• The Assessment of Burden of Chronic Conditions (ABCC)-CHF module enables systematic integration of perceived disease burden into CHF care, supporting shared-decision making and personalised care planning.

Introduction

The prevalence of chronic diseases is rising due to an ageing population, placing a significant burden on individuals and negatively impacting their health-related quality of life (HRQoL), particularly among those with multimorbidity.¹ Chronic heart failure (CHF) stands out as one of the most prevalent and burdensome chronic conditions, particularly affecting older people. This complex syndrome, caused by functional and/or structural heart abnormalities, manifests in debilitating physical symptoms, as well as emotional and social burdens.² Consequently, patients with CHF experience a lower HRQoL compared with the general population and those with other chronic conditions.³ This contributes to frequent and expensive healthcare utilisation and poor prognosis.²

Recently, a shift is being advocated from disease-focused to person-centred care, emphasising self-management and shared decision-making.⁴ As the importance of capturing patients’ perceptions of the impact of their chronic conditions is increasingly recognised, efforts are being made to develop patient-reported outcome measures (PROMs). However, these PROMs often focus on specific aspects of burden, such as treatment burden.⁵ The Assessment of Burden of Chronic Conditions (ABCC) tool was developed to build on this approach by assessing the broader perceived burden of chronic conditions and facilitating communication between patients and healthcare providers (HCPs). As it aims to support person-centred care, it should cover the most common chronic conditions. Currently, modules exist for chronic obstructive pulmonary disease (COPD), asthma and type 2 diabetes mellitus (T2DM).⁶ The tool includes several steps and components, following a cyclical course (figure 1).

Figure 1. Steps and components of the Assessment of Burden of Chronic Conditions (ABCC) tool. First, the patient completes a self-reported questionnaire (the ABCC scale) at home assessing perceived burden of disease. Second, the outcomes of the ABCC scale are digitally visualised into a balloon diagram during the consultation with the healthcare provider (HCP). Each balloon symbolises a domain of perceived disease burden or lifestyle; height and colour are based on predefined cut-off points. Third, the patient and healthcare provider together decide about which domain(s) they want to talk. By clicking on a balloon, advice per domain can be requested as an in-screen pop-up. Fourth, the patient and healthcare provider formulate a personalised treatment goal to include in the individual care plan. This gives patients direction to work on their health at home until the next consultation. In preparation for the next consultation, the cycle begins again with the completion of the ABCC scale. During this consultation, current outcomes on the questionnaire appear in coloured balloons, outcomes of the previous consultation appear in grey. Visualising the differences in this way creates a monitoring function for both healthcare provider and patient.

The ABCC tool includes a generic section, addressing overall disease burden, applicable to all users, and one or more disease-specific section(s) that focus(es) on the burden of a particular condition.^{6 7} This creates a single questionnaire and balloon diagram adapted to the individual’s specific condition(s).

Since CHF is a prevalent and burdensome condition, a CHF-specific module within the ABCC tool would be valuable for improving its management. Therefore, this study aims to develop a disease-specific module for patients with CHF (ABCC-CHF module) that fits within the existing ABCC tool. The questionnaire (ABCC-CHF scale), a PROM, forms the basis of the balloon figure and facilitates shared decision-making. Therefore, its psychometric properties will also be evaluated. The components of the ABCC tool are briefly explained in table 1.

Table 1. ABCC tool terms explained.

ABCC tool	The instrument itself has a modular structure, consisting of a generic section supplemented with disease-specific sections for—currently—COPD, asthma, T2DM and CHF. It comprises a questionnaire measuring perceived burden of disease, visualised burden-of-disease domain scores and associated advice.
ABCC scale	The underlying questionnaire of the ABCC tool consists of generic items and is supplemented with disease-specific items on the perceived burden of disease.
ABCC-CHF module	The disease-specific section for people with CHF can be added to the existing ABCC tool. It consists of a questionnaire (the ABCC-CHF scale), burden-of-CHF domains and associated advice.
ABCC-CHF scale	The underlying questionnaire of the ABCC-CHF module consists of disease-specific items on the perceived burden of disease in people with CHF.

ABCC, Assessment of Burden of Chronic Conditions ; CHF, chronic heart failure; COPD, chronic obstructive pulmonary disease; T2DM, type 2 diabetes mellitus.

Methods

This study was conducted in the Netherlands. As presented in box 1, it employed a mixed methods approach, combining qualitative (ie, developing the ABCC-CHF module, steps 1–5) and quantitative methods (ie, psychometric evaluation of the ABCC-CHF scale, step 6). The study adhered to established steps and requirements for the ABCC tool.⁶,⁹ An iterative process was employed, allowing for continuous adaptation and refinement of the ABCC-CHF scale and domains.¹⁰

Box 1. Study overview.

1. Establishing a first draft of the Assessment of Burden of Chronic Conditions (ABCC)-chronic heart failure (CHF) scale and domains

Based on clinical guidelines, literature and input from consultation of experts.

2. Investigating content validity of the ABCC-CHF scale and domains

Based on input from

1. healthcare providers involved in primary and specialised CHF care;

2. people with CHF.

3. Investigating face validity of the ABCC-CHF scale and domains

Based on input from people with CHF.

4. Establishing the preliminary version of the ABCC-CHF scale and domains

Based on input from experts, for example, healthcare providers involved in primary and specialised CHF care, representatives of the Dutch patient organisation of cardiovascular disease (Harteraad), the Dutch Heart Foundation and the Dutch Heart Registration, as well as one person with CHF.

5. Establishing cut-off points and advice per domain of the ABCC-CHF module

Based on national guidelines and input from a subgroup of the expert groups in step 4, namely healthcare providers involved in primary and specialised CHF care, and one person with CHF.

6. Evaluating psychometric properties of the preliminary version of the ABCC-CHF scale and domains

Based on a questionnaire study in people with CHF evaluating

1. construct validity by testing convergent validity and discriminative properties (known-groups);

2. reliability by testing internal consistency and reproducibility (test-retest analysis).

Development (steps 1–5)

After creating a first draft of the ABCC-CHF scale and domains (step 1), interviews were held with HCPs and patients with CHF (step 2). Step 2a evaluated whether the ABCC-CHF scale was relevant and comprehensive for assessing and displaying disease burden, while step 2b focused on patients’ perceived burden in daily life, as well as on assessing the relevance, comprehensiveness and comprehensibility of the questionnaire. HCPs were recruited via email for face-to-face interviews, conducted either individually or in pairs. Patients were recruited by HCPs and were interviewed face-to-face, either individually or in focus groups. In step 3, patients were interviewed by phone, using the think-aloud method,¹¹ to evaluate whether the ABCC-CHF scale and domains were logical and understandable. All interviews were conducted by one researcher (the first author) from April to August 2019, until data saturation was reached. This was done from a constructivist perspective, and experiences were assimilated by means of phenomenology.¹² Interviews were audio-recorded and transcribed verbat, after which analysis was performed manually. Transcripts were inductively coded, and emerging themes were identified through thematic analysis. These themes were reviewed and discussed within the research group to generate items and domains for the ABCC-CHF scale. Steps 4 and 5 involved expert meetings to reach consensus on the accuracy and feasibility (ie, content and length) of the ABCC-CHF module. Items, domains and associated advice for a preliminary version of the ABCC-CHF module were selected based on expert knowledge, clinical expertise and guidelines and emerging themes from the interviews.

Evaluation of psychometric properties (step 6)

To assess the psychometric properties of the preliminary version of the ABCC-CHF scale, construct validity (step 6a) and reliability (step 6b) were tested through a questionnaire study, conducted from October 2019 to April 2020. Participants were recruited by the Dutch Heart Foundation and Harteraad, a Dutch patient organisation. Inclusion criteria were a CHF diagnosis, age >18 years and the ability to read and understand the Dutch language.

Data collection

At baseline (T0), participants completed a self-administered paper questionnaire at home, including questions regarding baseline characteristics and the New York Heart Association (NYHA) classification,¹³ as well as the preliminary version of the ABCC scale, the Kansas City Cardiomyopathy Questionnaire (KCCQ)¹⁴ and the Hospital Anxiety and Depression Scale (HADS).¹⁵ At T1 (2 weeks after T0), participants completed the ABCC scale again and gave an estimation of their health status (ie, much worse, about the same, much better) to evaluate reproducibility, while minimising the risk of both recall effects and health changes.¹⁶

The ABCC scale measures perceived disease burden, with higher scores representing greater burden (ie, lower health status).⁶ One missing value per domain was tolerated, except for single-item domains.^{7 9} Missing values for item 15 were not tolerated, as this item evaluates CHF-specific concerns. Details regarding the domains and items can be found in table 2.

Table 2. Final version of the ABCC-CHF scale.

Dear Sir/Madam, With this questionnaire, we would like to get an impression of how you are doing. During the consultation with your healthcare provider, you can talk about the topics that are important to you. These questions are related to the chronic condition(s) for which you visit your healthcare provider.
	In the past week, how often…	0 Never	1 Hardly ever	2 A few times	3 Several times	4 Many times	5 A great many times	6 Almost all the time
1	did you suffer from fatigue?	□	□	□	□	□	□	□
2	did you have a poor night’s rest?	□	□	□	□	□	□	□
3	did you suffer from sadness, fear, frustration, shame or other unpleasant feelings?	□	□	□	□	□	□	□
4	did you experience taking medication (eg, tablets, puffs, insulin) as a burden?	□	□	□	□	□	□	□
	In the past week, to what extent …	0 Not at all	1 Very slightly	2 Slightly	3 Moderately	4 Very	5 Extremely	6 Totally
5	were you limited in strenuous physical activities (such as climbing stairs, hurrying, doing sports)?	□	□	□	□	□	□	□
6	were you limited in moderate physical activities (such as walking, housework, carrying things)?	□	□	□	□	□	□	□
7	were you limited in daily activities at home (such as dressing, washing yourself)?	□	□	□	□	□	□	□
8	were you limited in your work or social activities (short trip, visiting friends and family)?	□	□	□	□	□	□	□
9	did your condition negatively influence your relationship with others?	□	□	□	□	□	□	□
10	did you have any difficulty with intimacy or sexuality?	□	□	□	□	□	□	□
11	did you worry about your future?	□	□	□	□	□	□	□
12	did you feel short of breath while lying (flat) or did this wake you up?	□	□	□	□	□	□	□
13	did you feel short of breath while sitting or standing still?	□	□	□	□	□	□	□
14	did you experience any of the following complaints: dizziness, stomach complaints, pain on the chest and loss of appetite or coughing?	□	□	□	□	□	□	□
15	did you feel anxious or concerned that something would happen to your heart?	□	□	□	□	□	□	□
16	did you feel aggravated to adjust your life (eg, planning activities and medication intake, weighing daily and paying attention to salt and fluid intake)?	□	□	□	□	□	□	□
17	did you have a bloated stomach, legs, feet or ankles?	□	□	□	□	□	□	□
	In the past 3 months, how often …
18	have you had an increase or decrease in weight (more than 2 kilograms in 3 days)?	□ □ □ □ □	Never Once Twice 3 times 4 times
19	have your medicines been adjusted due to an increase in complaints?	□ □ □ □ □	Never Once Twice 3 times 4 times
20	have you been hospitalised due to complaints?	□ □ □ □ □	Never Once Twice 3 times 4 times
The following questions are related to your lifestyle:
21	In the past week, how many days did you perform moderately intense physical exercise for 30 min or more? For example, walking or cycling at a fast pace. It may also be a minimum of 3×10 min.	□ □ □ □	0 days 1–2 days 3–4 days 5 days or more
22	How many glasses of alcohol did you drink in the past week?	___	Glasses per week
23	Do you smoke or have you smoked?	□ □ □	Yes. In the last week, how many (shag) cigarettes have you smoked on average per day? ___________ Previously. I stopped smoking since: ___ (month) / ___(year) No
24	What is your weight?	___	………… kg
25	What is your height?	___	………… cm
26	Is there anything else you would like to talk about or would like to receive more information about?
	_______________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

The final version of the ABCC-CHF scale consists of seven generic domains, four disease-specific domains and four lifestyle domains. The generic domains are fatigue (item 1), night’s rest (item 2), feelings and emotions (items 3, 11 and 15), medicines (item 4), physical limitations (items 5, 6 and 7), relations and work (items 8 and 9) and sexuality (item 10). The disease-specific domains for CHF are heart failure complaints (items 12, 13 and 14), adjusting your life (item 16), fluid retention (items 17 and 18) and stable heart failure (items 10 and 20). The lifestyle domains are physical activity (item 21), alcohol (item 22), smoking (item 23) and weight (BMI) (items 24 and 25).

ABCC, Assessment of Burden of Chronic Conditions; BMI, body mass index; CHF, chronic heart failure.

HRQoL is considered an integrated part of disease burden and was therefore used as a comparative measure to evaluate CHF burden.^{6 7} Since no gold standard exists for measuring perceived burden in CHF, the ABCC scale was compared with the Dutch version of the KCCQ¹⁴ to assess convergent validity. This is a self-administered, 23-item questionnaire that measures disease-specific HRQoL in patients with CHF, focusing on subscales like physical limitation, symptom burden, social limitation and quality of life. Scores range from 0 to 100, with higher scores indicating better health. Scoring and handling of missing data were performed as described in the KCCQ scoring instructions.¹⁷

The ABCC scale’s discriminative ability was assessed by means of known-groups analysis, based on expert opinion and literature.¹⁸,²⁶ Groups were classified based on severity of CHF, depression and anxiety. CHF severity was derived from the NYHA classification.¹³ The Dutch version of the HADS was used to evaluate known-groups validity discriminating for depression and anxiety.¹⁵ This self-administered, 14-item screening scale consists of 2 subscales: anxiety and depression (7 items each). Subscale scores range from 0 to 21. Exact scoring was performed according to the HADS scoring instructions.²⁷ As these instructions contain no guidelines for handling missing data, missing items were imputed using the mean for that domain.⁷ One missing item per domain was tolerated.

Sample size

According to recommendations, the participant-to-item ratio should range from 2 to 20 participants per item.⁷ The preliminary version of the ABCC-CHF scale in this study consisted of 20 items. Generic questions about lifestyle were excluded due to their non-subjective measurement, as well as the open question. For pragmatic reasons and based on previous ABCC tool studies, a participant-to-item ratio of about thr was deemed appropriate, resulting in an estimated sample size of 60 participants. Based on this sample size and assuming effect sizes that could detect 80% power and a two-sided significance level of α=0.05, prerequisites for data analysis were established. Power Analysis and Sample Size (PASS) software (2019; V.19.0.9) was used for all calculations. As for construct validity, a correlation as small as 0.345 would already yield the required power to detect a significant association.⁷ Regarding differentiation by known-groups, the same ratio was assumed for participants with and without depression based on HADS scores, as seen in previous comparable research.⁹ Although COPD was examined in that study, its impact on mental health is comparable to that of CHF.^{28 29} Assuming the same group sizes (N1/N2 is about 4.25), a sample size of 60 could demonstrate statistical significance with an effect size of 0.9. Internal consistency was evaluated using Cronbach’s α for the total score and its multi-item domains. K was set to 3, as fewer items require larger sample sizes. Under these conditions, a Cronbach’s α of ≥0.466 could be detected. Test-retest reliability was estimated using the intraclass coefficient (ICC), with ICCs of ≥0.31 providing at least 80% power.

Data analysis

To assess the psychometric properties of the ABCC-CHF scale, prerequisites for validity and reliability were established based on sample size calculations, expert opinions and literature (box 2).¹⁸,²⁶³⁰ The KCCQ was used as a comparative questionnaire to evaluate convergent validity,¹⁴ requiring at least 75% of the predefined correlations to match the hypotheses.³⁴ Since both the ABCC-CHF scale and the KCCQ include single-item domains, these domains were hypothesised to significantly correlate on a moderate level between −0.3 and −0.7.³⁰

Box 2. Prerequisites to assess psychometric properties of the Assessment of Burden of Chronic Conditions (ABCC)-chronic heart failure (CHF) scale.

1. Convergent validity

Hypothesis for total ABCC-CHF scale: correlation lower than −0.7.

Hypothesis for ABCC-CHF single-item subscales: correlation in between −0.3 and −0.7 for:

• fatigue versus Kansas City Cardiomyopathy Questionnaire (KCCQ) overall summary score, overall clinical score, total symptom score, symptom frequency and symptom burden;

• night’s rest versus KCCQ overall summary score, overall clinical score, total symptom score and symptom frequency;

• feelings and emotions versus KCCQ overall summary score and quality of life;

• physical limitations versus KCCQ overall summary score, overall clinical score and physical limitation;

• relations and work versus KCCQ overall summary score and social limitation;

• sexuality versus KCCQ overall summary score and social limitation;

• adjusting your life versus KCCQ overall summary score;

• fluid retention versus overall summary score, overall clinical score, total symptom score, symptom frequency and symptom burden;

• heart failure complaints versus KCCQ overall summary score, overall clinical score, total symptom score, symptom frequency and symptom burden.

2. Known-groups validity

Hypothesis for mild versus severe CHF assessed by the New York Heart Association (NYHA) classification: significantly higher scores on ABCC-CHF subscales:

 Fatigue, feelings and emotions, physical limitations, stable heart failure, adjusting your life, fluid retention and heart failure complaints.

Hypothesis for depression versus no depression assessed by Hospital Anxiety and Depression Scale (HADS): higher scores on ABCC-CHF subscales:

 Fatigue, night’s rest, feelings and emotions and relations and work.

Hypothesis for anxiety versus no anxiety assessed by HADS: higher scores on ABCC-CHF subscales:

 Fatigue, night’s rest, feelings and emotions and relations and work.

3. Reliability

Internal consistency

Acceptable Cronbach’s α:

• ≥0.90 for the total ABCC-CHF scale;

• ≥0.70 for subscales with two or more items.

Reproducibility

Acceptable intraclass correlation coefficient (ICC) of ≥0.90.

To evaluate the discriminative properties, three pairs of groups were created, based on (1) NYHA classification¹³ (mild vs severe CHF), (2) depression as measured by the HADS depression subscale¹⁵ (depression vs no depression) and (3) anxiety as measured by the HADS anxiety subscale (anxiety vs no anxiety). Subscale scores of ≥8 discriminated between the known-groups of (borderline) depression and anxiety.³⁵ Statistical significance was determined at values of p≤0.05.

Cronbach’s α was used to assess internal consistency. For test-retest analysis, only participants whose self-administered health status remained unchanged between T0 and T1 were included. Reproducibility was expressed using the ICC. Depending on the normality of the data (assessed via histogram and Q-Q plot), either t-tests or Mann-Whitney U tests were conducted. Statistical analyses were performed using IBM SPSS Statistics V.25.0.

Patient and public involvement

Patients with CHF were involved in multiple stages of this study. Patient involvement began early in the research process, during the development of the ABCC-CHF scale and domains (steps 2 and 3). Patients also contributed to the evaluation of the psychometric properties in this study (step 6).

Results

Development

The first draft of the ABCC-CHF scale and domains was based on (international) guidelines,^{18 19 36} literature³⁷,⁴² and experts’ consultations. For step 2a, 19 HCPs were interviewed, including 4 general practitioners (of which 2 specialised in cardiovascular disease and 1 in palliative care), 6 practice nurses, 4 cardiologists, 2 heart failure nurse practitioners and 3 heart failure specialist nurses. In step 2b, 15 patients were interviewed. To evaluate face validity, five additional participants were interviewed by phone.

The results from steps 1, 2 and 3 were combined and discussed during 3 expert meetings, to develop a preliminary version of the ABCC-CHF scale and domains. Following consensus among the expert group and developers of the ABCC tool,⁶ an additional item was incorporated into the generic domain of relations and work (item 9 in final version, table 2). For the CHF-specific module, nine items across four domains were identified: heart failure complaints, adjusting your life, fluid retention and stable heart failure. A balloon diagram visualising the ABCC-CHF scale scores is presented in figure 2.

Figure 2. Visualisation of the perceived burden of disease of a person with CHF. On the far left, four disease-specific domains for CHF are shown, followed by seven generic domains, ranging from physical limitations to medicines. Four lifestyle domains are on the far right. All balloons represent domain scores derived from the ABCC scale: green for a low score, orange for a moderate score and red for a high score. To follow and visualise changes over time, grey balloons illustrate the domain scores from the previous visit to the healthcare provider. In cases of comorbidities, the associated disease-specific domains will be shown to the left of the CHF domains. *Translated from the original Dutch version.

In two expert group meetings, cut-off points for the ABCC-CHF scale scores, as well as heights and colours of the balloons in the diagram, were determined based on current guidelines. Advice per domain was based on average scores or specific score combinations. An example of such advice is provided in online supplemental material 1.

Evaluation of psychometric properties

In total, 68 participants were included in the questionnaire study (step 6), with 3 lost to follow-up between T0 and T1 due to personal circumstances. Baseline characteristics are presented in table 3.

Table 3. Baseline characteristics of the evaluation study population (n=68).

Female sex, n (%)	35 (51.5)
Mean age, years (SD)	61.5 (13.0)
Level of education*, n (%)
Low†	15 (22.1)
Middle‡	22 (32.4)
High§	29 (42.6)
Other	2 (2.9)
NYHA classification, n (%)
Class I	15 (22.1)
Class II	34 (50.0)
Class III	18 (26.5)
Class IV	1 (1.5)
Diagnosed with heart failure since, n (%)
<1 year	8 (11.8)
1–5 years	29 (42.6)
>5 years	30 (44.1)
Treated in, n (%)
General practice	2 (2.9)
Hospital	63 (92.6)
Unknown	3 (4.4)
Medicines, n (%)
None	3 (4.4)
At least one	65 (95.6)
ACE inhibitors	29 (42.6)
Angiotensin II receptor inhibitors	27 (39.7)
Beta-blockers	57 (83.8)
Digoxin	6 (8.8)
Diuretics	48 (70.6)
Ivabradine	3 (4.4)
Statins	30 (44.1)
Comorbidity, n (%)
None	4 (5.9)
At least one	64 (94.1)
Anaemia	5 (7.4)
Anxiety disorder	3 (4.4)
Asthma	7 (10.3)
Cardiac arrhythmias¶	36 (52.9)
Cognitive impairment**	1 (1.5)
COPD	8 (11.8)
Coronary artery disease††	12 (17.6)
Depression	1 (1.5)
High blood pressure	5 (7.4)
Renal impairment	13 (19.1)
Sleep apnoea	18 (26.5)
Stroke‡‡	5 (7.4)
T2DM	12 (17.6)

^*Based on the Dutch Standard Education Format.⁵⁰

^†Classified as preparatory or lower vocational education, general secondary education or secondary vocational education.

^‡Classified as higher general secondary education or secondary university education.

^§Classified as higher professional education or university.

^¶Including atrial fibrillation (AF), implantable cardioverter defibrillator (ICD) or pacemaker support.

^**For example, dementia.

^††Including heart attack or chest pain due to heart problems.

^‡‡Including transient ischaemic attack (TIA), cerebral infarction or cerebral haemorrhage.

COPD, chronic obstructive pulmonary disease; NYHA, New York Heart Association; T2DM, type 2 diabetes mellitus.

Due to missing data for 1 participant, analyses for convergent validity included 67 participants. Eleven participants reported a change in health status between T0 and T1 and were excluded from the reproducibility assessment, leaving 57 participants for test-retest analyses. Additional participant characteristics can be found in online supplemental material 2, while outcomes on the ABCC-CHF scale, KCCQ and HADS can be found in online supplemental material 3. Psychometric properties of the ABCC-CHF scale are summarised in table 4, with complete outcomes available in online supplemental material 4.

Table 4. Psychometric properties of the ABCC-CHF scale.

Convergent validity (n=67)
	KCCQ
	OSS	OCS	TSS	PL	SF	SB	QoL	SL
ABCC-CHF scale, r
Total	−0.865*
Fatigue	−0.777*	−0.744*	−0.584*		−0.251	−0.739*
Night’s rest	−0.543*	−0.526*	−0.576*		−0.357*
Feelings and emotions	−0.503*						−0.478*	−0.497*
Physical limitations	−0.881*	−0.906*		−0.850*
Relations and work	−0.755*							−0.779*
Sexuality	−0.512*							−0.504*
Adjusting your life	−0.613*
Fluid retention	−0.482*	−0.597*	−0.691*		−0.564*	−0.417*
Heart failure complaints	−0.598*	−0.639*	−0.616*		−0.463*	−0.591*
Known-group validity (n=67)
	NYHA †			HADS-D ‡			HADS-A ‡
	I–II (n=48)		III–IV (n=19)	<8 (n=51)	≥8 (n=12)		<8 (n=51)	≥8 (n=16)
ABCC-CHF scale, r
Total	1.5 (0.9 to 1.9)		2.8 (2.2 to 3.2)§	1.6 (0.9 to 2.3)	2.9 (2.2 to 3.2)§		1.6 (0.9 to 2.6)	2.2 (1.7 to 3.0)§
Fatigue	3.0 (2.0 to 4.0)		5.0 (3.0 to 6.0)§	3.0 (2.0 to 5.0)	5.0 (4.0 to 6.0)§		3.0 (2.0 to 5.0)	4.0 (2.0 to 5.0)
Night’s rest				2.0 (1.0 to 3.0)	3.5 (2.0 to 4.8)§		2.0 (1.0 to 3.0)	3.0 (2.0 to 4.0)§
Feelings and emotions	1.3 (0.7 to 2.0)		2.3 (1.5 to 3.7)§	1.3 (0.7 to 2.0)	3.0 (2.1 to 3.9)§		1.0 (0.7 to 1.7)	2.7 (2.0 to 3.6)§
Physical limitations	1.3 (0.7 to 2.5)		4.0 (3.0 to 4.7)§
Relations and work				1.5 (0.5 to 2.5)	4.5 (2.4 to 5.0)§		1.5 (0.5 to 3.0)	2.0 (0.0 to 4.0)
Sexuality	0.0 (0.0 to 0.5)		1.0 (0.5 to 2.0)§
Adjusting your life	1.0 (0.0 to 3.0)		4.0 (2.0 to 5.0)§
Fluid retention	0.0 (0.0 to 1.5)		1.5 (0.0 to 2.0)
Heart failure complaints	0.3 (0.0 to 1.0)		1.3 (0.7 to 2.0)§
Reliability
Internal consistency	Cronbach’s α (95% CI)
Total scale	0.92 (0.88 to 0.95)¶
Feelings/Emotions	0.75 (0.62 to 0.84)¶
Physical limitations	0.90 (0.84 to 0.93)¶
Relations and work	0.86 (0.84 to 0.93)¶
Stable heart failure	0.38 (0.00 to 0.60)
Stable heart failure**	0.65 (0.40 to 0.80)
Heart failure complaints	0.79 (0.69 to 0.87)¶
Fluid retention**	0.66 (0.42 to 0.80)
Test-retest reliability	ICC (95% CI)
	0.97 (0.94 to 0.98)††

Due to non-normal distribution of data regarding the subscales of the ABCC-CHF scale, Pearson’s rank correlation coefficient (r) with 95% CI was calculated.

^*r<−0.7 for total ABCC-CHF scale or −0.7<r<−0.3 for single-item correlations.

^†NYHA I–I being mild CHF versus NYHA III–IV being severe CHF.

^‡No depression versus depression and no anxiety versus anxiety were discriminated by scores of ≥8 on HADS-D and HADS-A, respectively.

^§P≤0.05.

^¶α≥0.90 for total ABCC-CHF scale or α≥0.70 for subscales with two or more items.

^**After reallocation of item 18.

^††ICC ≥0.90.

ABCC, Assessment of Burden of Chronic Conditions; CHF, chronic heart failure; HADS, Hospital Anxiety and Depression Scale; ICC, intraclass correlation coefficient; KCCQ, Kansas City Cardiomyopathy Questionnaire; NYHA, New York Heart Association; OCS, overall clinical scale; OSS, overall summary scale; PL, physical limitation; QoL, quality of life; SB, symptom burden; SF, symptom frequency; SL, social limitation; TSS, total symptom scale.

The ABCC-CHF total score correlated significantly with the KCCQ overall summary score (r=−0.865). Of the 30 correlations investigated (box 2), 29 met the defined criteria, indicating almost 97% of the correlations aligned with the hypotheses. Only the correlation between fatigue on the ABCC-CHF scale and symptom frequency of the KCCQ (r=−0.251) did not meet the criteria. Participants with a higher NYHA class scored significantly higher on the hypothesised domains of the ABCC-CHF scale, except for fluid retention. Similarly, those reporting depression scored significantly higher on the hypothesised domains of the ABCC-CHF scale. Participants with anxiety showed significantly higher scores on the hypothesised domains of night’s rest and feelings and emotions, but not on fatigue or relations and work.

The ICC for the ABCC-CHF scale was 0.97 (95% CI 0.94 to 0.98). The total ABCC-CHF scale demonstrated strong internal consistency (α=0.92, 95% CI 0.88 to 0.95). However, internal consistency of stable heart failure was insufficient (α=0.38, 95% CI 0.00 to 0.60). Analysis of internal consistency revealed that the highest α was achieved by reallocating item 18 to the domain fluid retention. This change increased the α for stable heart failure to 0.65 (95% CI 0.40 to 0.80) and for fluid retention to 0.66 (95% CI 0.42 to 0.80), after which the expert group decided to reallocate the item permanently. This reallocation did not change the sequence or content of the questionnaire. All analyses were then repeated with the updated domains (table 4).

Table 2 presents the final version of the ABCC-CHF scale. Outcomes are scored using a 7-point Likert scale, except for items 18–20, which are scored from 0 to 4. High scores indicate a greater perceived disease burden. Please note that while an English translation of the ABCC-CHF scale is provided for transparency, the scale was developed, administered and evaluated in Dutch. The English version has not (yet) been formally validated and therefore cannot be used as such.

Discussion

This study describes the development and psychometric evaluation of a CHF module to be included into the ABCC tool. Development was based on (inter)national guidelines, literature and input from patients, HCPs and experts, similar to the development of the other modules.^{6 8} The findings confirm that the ABCC-CHF scale is a valid and reliable measure of disease burden in patients with CHF. Convergent validity was demonstrated, with nearly 97% of the predefined correlations aligning with hypotheses. The ABCC-CHF scale showed strong known-groups validity, distinguishing between varying levels of CHF severity, depression and anxiety in most cases. Finally, the ABCC-CHF scale exhibited sufficient reliability, with acceptable internal consistency for the total score and most multi-item domains, and exceptionally high reproducibility.

Several exceptions observed in the results warrant further attention. First, the correlation between fatigue on the ABCC-CHF scale and symptom frequency of the KCCQ did not reach the predefined threshold. This is likely because the KCCQ symptom frequency captures a collection of symptoms (ie, shortness of breath, fatigue and oedema), whereas the ABCC-CHF domain fatigue focuses specifically on fatigue alone. Second, the fluid retention domain of the ABCC-CHF tool did not discriminate between mild and severe CHF based on NYHA classification. A possible explanation for this could be the reallocation of item 18, which resulted in averaged, relatively low domain scores. Moreover, the relatively high reported prevalence of diuretic use among respondents (70.6%), which helps control clinical signs of fluid retention, may have attenuated differences between NYHA classes. Third, anxiety did not significantly differentiate fatigue and relations and work. A possible explanation for this can be that anxiety is a collective term for many different types of complaints. Each form of anxiety has its own specific red flags.²² Fourth, the initial internal consistency of the stable heart failure domain was initially insufficient, which can possibly be attributed to the conceptual nature of this domain, which comprises a number of clinically distinct items reflecting different aspects of perceived stability. Unlike other domains, where items often overlap, these indicators represent complementary facets of a stable clinical state rather than a single underlying symptom.

The International Consortium for Health Outcomes Measurement (ICHOM) emphasises assessing patient-relevant outcomes in chronic disease management. However, using all suggested outcome measures from the ICHOM Heart Failure Standard Set, as well as those for multimorbidity, is often not feasible in routine care.⁴³ The ABCC-CHF scale offers a concise way to capture many of these outcomes in one questionnaire, which is visually displayed, facilitating tailor-made care plans and shared decision-making. Furthermore, the grey balloons allow for monitoring of disease burden over time.⁶

Blended care, combining remote monitoring and in-person consultations, is increasingly essential for managing chronic conditions like CHF, especially for homebound or frail patients.⁴⁴ The COVID-19 pandemic underscored the importance of self-management, as face-to-face consultations became limited.⁴⁵ The ABCC tool can facilitate blended care by integration into HCPs’ information systems, as well as digital applications with video calling capabilities, enabling seamless interaction between HCPs and patients, while also promoting self-management.

This study showed several strengths. Qualitative and quantitative methods were combined, involving various stakeholders, including patients, HCPs and experts. The development and psychometric evaluation of the ABCC-CHF module followed an iterative approach,¹⁰ allowing for continuous adjustments as new insights emerged. Additionally, the module was constructed through triangulation, incorporating literature,³⁷,⁴² stakeholder input and alignment with healthcare policies and expert guidelines.^{18 19 36} This also applied to the hypotheses for evaluating the psychometric properties.¹⁸,²⁶³⁰

The study also had some limitations. First, item selection and clustering were based primarily on expert opinion and usability, rather than traditional methods like item banks or factor analyses. Second, participants appeared to have a low disease burden, likely due to their relatively young age and fewer complaints, as they were recruited from a patient organisation.⁴⁶ Therefore, the extent to which the observed validity and reliability apply to populations with a higher perceived disease burden remains uncertain.

This study suggests opportunities for both future clinical practice and future research. Given that the ABCC tool has previously demonstrated validity in people with COPD, asthma and T2DM,⁷ and that the present study extends this evidence to CHF, the tool shows promise for broader implementation across both primary and secondary care.

From a research perspective, following completion of the present study, additional studies have been conducted evaluating the effectiveness, cost-effectiveness and patient experiences associated with the use of the ABCC tool in practice.⁴⁷,⁴⁹ Building on this growing evidence, and given that the conditions included in the ABCC tool frequently co-occur, future research is needed to assess the validity and reliability in populations with multimorbidity. In addition, further research is planned to explore its use in other healthcare settings, such as physiotherapy.

Conclusions

The ABCC-CHF module includes a 26-item self-administered questionnaire (ABCC-CHF scale) that assesses the perceived disease burden in patients with CHF, subsequently generating a balloon diagram that visualises the scores of perceived burden and provides tailored advice per domain. The study confirms the validity and reliability of the ABCC-CHF scale among Dutch people with CHF. By integrating disease burden into consultations, the ABCC tool enhances person-centred care for individuals with CHF and multimorbidity.

Data availability statement

Data are available on reasonable request.