Carers of patients receiving peritoneal dialysis (PD) or other chronic diseases can be at risk of work overload due to the burden of caring for and treating these dependent individuals.1–3

Overload is a key dimension in the evaluation of carers. Analysis of the variables affecting carers has led to the development of conceptual models that show overload to be caused by different factors and components.4–7 This is why it is so important to measure the degree of overload in this population.

Several studies conducted over the past 30 years have revealed significant degrees of carer overload in connection with different diseases, and have highlighted the need to design respite protocols for carers.8–12

The burden of carers of patients on renal replacement therapy with haemodialysis (HD), PD and kidney transplant has already been evaluated using questionnaires that address various different dimensions: health-related quality of life,1,2,13–15 emotional state,2,16,17 depression,14–18 vital satisfaction, family stress, marital relationship,19 psychosocial repercussion,13 exhaustion or burnout syndrome and, more specifically, subjective or objective caring-related personal overload.2,13,14,16,18,20

However, most studies have so far used general questionnaires to determine carer overload,1,2,14,16 even though generic measures may not be sensitive enough to identify differences, particularly in a research setting.21 Few studies have used questionnaires designed specifically for PD22,23 or HD plus PD.4

The characteristics of PD patients differ considerably from patients with dementia, cancer, or severe functional limitations, and, therefore, the stress factors affecting their carers also differ to a certain extent. Thus, some years ago our group developed the Overload Questionnaire for Carers of Patients on Peritoneal Dialysis (CSCDP, in its Spanish acronym). This instrument includes 3 dimensions: patient's degree of dependence, carer's subjective overload, and carer's objective overload, and has shown optimal reliability and validity for clinical use.24

Nevertheless, validation of health-related questionnaires or scales is an ongoing process, as the results and corresponding interpretation may vary depending on the specific characteristics of the population in which they are used.25

Questionnaires are validated on the basis of their reliability and validity. Reliability indicates the extent to which a measuring instrument produces accurate results on repeated tests. It is evaluated on the basis of internal consistency, stability over time, equivalence, and inter-observer agreement. Validity, meanwhile, shows the degree to which the instrument measures what it really intends to measure in terms of content, construct, and criteria.25–27

In this study, we have validated the CSCDP in a wider sample of carers using the 36-item short form (SF-36) health-related quality of life questionnaire28 and the Zarit Burden Interview (ZBI)29–32 to test the criterion validity. The ZBI has been translated into many languages, and is widely used and recommended.33–36 It has also been used for carers of patients with chronic kidney failure on replacement therapy with HD,13,17,37,38 PD1,2,17,22,23,38,39 and kidney transplant.16

The objectives of this study have been to validate the CSCDP by analysing the reliability (internal consistency) and validity (construct validity and criterion validity compared to the ZBI), and to construct a number of rating scales from the questionnaire to determine the degree of overload of the carers in our study sample.

All patients included in a PD programme from 8 hospitals were invited to take part in the study. To be included in the study, the patient must have been receiving PD for at least 3 months, and must be cared for by a primary carer. The primary carer was defined as the person close to the patient who, when asked, confirmed that he/she is responsible for the care of the patient on PD. Both participants, patient and carer, were asked to sign an informed consent form before inclusion in the study. The study was approved by the ethics committee of the reference hospital (H. Universitari Germans Trias i Pujol, Badalona) and complies with the principles of good clinical practice.

Data collected from the patients included age, gender, education, work status, concurrent diseases, Charlson index, choice of treatment modality, and degree of independence when performing PD. The tests administered were: the Mini-Mental State Examination,40 the Karnofsky Performance Status scale,41 which assesses the patient's physical capacity and degree of independence, and the Lawton–Brody Instrumental Activities of Daily Living Scale.42 In the latter, we used the original 0–8-point score that measures the patient's degree of independence, and an alternative score developed by us that shows the degree of limitation resulting from not being able to perform the different activities, in order to correlate this with the degree of dependence. Health-related quality of life was assessed using the SF-36 questionnaire validated for the Spanish population.28,43

In the case of the carers, we recorded their personal characteristics and work status and the degree of kinship or relationship with the patient, and their score on the Mini-Mental State Examination, the SF-3628 and the 22-item ZBI29,30 adapted and revalidated for the Spanish population.31,32 In this questionnaire, answers were scored from 0 to 4 (never: 0; rarely: 1; sometimes 2; quite frequently; 3; nearly always: 4). Following the system used by many other authors,17,34,37 overload was classified into 4 different levels, according to the following scores: (1) from 0 to 20, slight overload or no overload; (2) from 21 to 40, mild to moderate overload; (3) from 41 to 60, moderate to severe overload, and (4) 61 or more, severe overload.

The CSCDP comprises 30 questions relating to 3 dimensions: patient dependence (8 items), subjective overload (12 items) and objective overload (10 items) of the PD carer.24 Unlike the previous study published by our group, each item is scored using a scale from 0 to 4, with 4 being the worst situation. These 2 tests were administered with an interval of between 15 and 30 days. The order was changed to avoid bias.

One hundred and sixteen patients and carers were included initially; of whom 9 were excluded because they did not complete the protocol, leaving 107 carers and patients from 8 hospitals for the final analysis.

The average age of patients was 63.49±13.29 years, and carers was 57.50±14.69 years; this difference was significant (p=0.002). The patients had been receiving PD for 24.07±19.15 months, with a Charlson comorbidity index of 7.05±2.71; 34.6% were diabetic and 49.5% were treated with a cycler. There were no significant differences between centres (ANOVA).

The majority of carers were women, while the patients were mostly men (p<0.001). Both carers and patients had a similar level of education. The work status of most patients was retired, and most carers were housewives.

PD had been prescribed by the patient's doctor in 35.5% of cases, and had been chosen by the patient or his/her family in 64.5% of cases; in 31.8% of these, it was chosen because it enabled them to pursue their leisure and free time activities (Table 1). Forty-three (40.2%) patients always needed help to perform PD, 14 (13.1%) needed help sometimes, and 50 (46.7%) never or hardly ever needed help.

The relationship between carers and patients was: spouses 83 (77.6%), children 17 (15.9%), father or mother 5 (4.7%) and other 2 (1.8%). The carer had been caring for the patient for: less than 6 months – 11 (10.3%); between 6 months and 1 year – 27 (25.2%); between 1 and 2 years – 27 (25.2%); between 2 and 3 years – 18 (16.8%) and more than 3 years – 24 (22.4%). Twenty-five carers had a full-time job (23.4%), 14 (13.1%) worked part-time, and 68 (63.6%) did not work outside the home. Nearly all (92.5%) carers lived in the same house as the patient.

Results of the questionnaires administered

The results of the descriptive statistics are shown in Table 2. In the Mini-Mental State Examination, patient scores were significantly lower than those of the carers (p<0.001). Eleven patients and 2 carers scored less than 23 points (cognitive impairment). The analysis by centres (ANOVA) found that the carers from centre 3 scored significantly lower on the Mini-Mental State Examination than those from centres 5 and 8 (Bonferroni: p=0.011).

Table 2.

Results of the questionnaires administered to patients and carers.

	Median	Mean	Standard deviation	Lower limit of 95% CI	Upper limit of 95% CI	Min.	Max.
Patients
Mini-Mental	31	29.29	6.07	28.12	30.46	6	35
Karnofsky	60	63.68	18.53	60.11	67.25	30	100
Lawton–Brody	4	4.28	2.18	3.86	4.7	0	8
Lawton–Brody (alternative)	10	10.79	6.19	9.61	11.98	0	23
Carers
Mini-Mental	33	31.83	3.39	31.18	32.48	21	35
Zarit overload	22	25.51	16.08	22.43	28.59	0	71
CSCDP
Dependency	13	14.43	7.98	12.9	15.96	1	31
Subjective overload	9	9.57	7.29	8.17	10.98	0	35
Objective overload	11	12.75	8.65	11.09	14.41	0	35
Subjective+objective overload	21	22.32	14.39	19.56	25.08	0	68
Total questionnaire	34	36.75	18.03	33.29	40.2	7	89

95% CI: 95% confidence interval; CSCDP: Overload Questionnaire for Carers of Patients on Peritoneal Dialysis.

On the Karnofsky scale, the median score was 60 and the average score was 63.67%, corresponding to the need for occasional help. The ANOVA by centres showed significant differences between centre 3 and centre 8 (post hoc, Bonferroni: p=0.002).

On the Lawton–Brody scale, the average score using the instrument's own system showed moderate dependence, and light-to-moderate dependence using our alternative system. The reliability of this scale was good (Cronbach's α 0.760) with the instrument's own system, and somewhat better (Cronbach's α 0.835) with our alternative system.

The health status of carers (SF-36) was better than that of patients in terms of physical function (p<0.001), physical role (p<0.001), general health (p<0.001) and vitality (p=0.02); there were no differences in bodily pain, social functioning, emotional role and mental health (Fig. 1).

Fig. 1.

SF-36 Health Questionnaire. The statistical evaluation is given in the text. BP: bodily pain; ER: emotional role; GH: general health; M. 55–64: general population of men aged between 55 and 64 years; MH: mental health; PF: physical functioning; PR: physical role; SF: social functioning; VT: vitality; W. 55–64: general population of women aged between 55 and 64 years.

(0.17MB).

The health of the patients (mostly men) was significantly worse than that of men in the general population aged between 55 and 6443 in all dimensions (p<0.001). The carers (mostly women) showed worse health status in terms of their physical role (p=0.023), bodily pain (p=0.036), social functioning (p<0.001) and emotional role (p=0.011) than women in the general population aged between 55 and 6443 (Fig. 1).

Overload Questionnaire for Carers of Patients on Peritoneal Dialysis

The results of the CSCDP, and the wide range between minimum and maximum scores, are shown in Table 2. The reliability of the questionnaire was excellent, with high internal consistency (Cronbach's α) in each dimension: dependence (α 0.808), subjective overload (α 0.812), objective overload (α 0.868) and total score (α 0.887).

Factor analysis initially showed 8 factors with an eigenvalue greater than 1, which explained the 67.6% of total variance. On the basis of the Scree plot (Fig. 2) we chose the first 3 factors which explained the 45.7% of variance (Table 3). Component 1 showed high factorial loading on objective overload items (items 21–30; r=0.417–0.811) and on 8 of the 12 items of the subjective overload dimension of the CSCDP (r=0.264–0.644). Component 2 only showed high factorial loading on items of the dependency dimension (items 1–8) of the CSCDP (r=0.350–0.799). Component 3 showed high factorial loading on 5 items of the subjective overload dimension (r=0.497–0.745).

Figure 2.

Scree plot of the main components of the Overload Questionnaire for Carers of Patients on Peritoneal Dialysis.

(0.06MB).

Table 3.

Rotated component matrix of the Overload Questionnaire for Carers of Patients on Peritoneal Dialysis.

Itemsa	Component
	1	2	3
OLdep1		0.522
OLdep2		0.790
OLdep3		0.610
OLdep4		0.799
OLdep5		0.671
OLdep6		0.749
OLdep7		0.711
OLdep8		0.350
OLsubj9	0.549
OLsubj10	0.428
OLsubj11	0.615
OLsubj12			0.596
OLsubj13			0.731
OLsubj14	0.399
OLsubj15	0.644
OLsubj16	0.509		0.497
OLsubj17			0.745
OLsubj18	0.264
OLsubj19			0.537
OLsubj20	0.598
OLobj21	0.576
OLobj22	0.544
OLobj23	0.729
OLobj24	0.803
OLobj25	0.811
OLobj26	0.521
OLobj27	0.672
OLobj28	0.417
OLobj29	0.798
OLobj30	0.651

a

CSCDP items: OLdep: dependency overload; OLsubj: subjective overload; OLobj: objective overload.

This study, carried out on 107 PD patients and their carers, confirms the results obtained in our previous paper in which the CSCDP was developed.24 It shows that the scale has excellent reliability and good validity, and can determine the rating scales and cut-off points for different degrees of carer overload. The sample of patients obtained from 8 hospitals with ongoing PD programmes supports the value of our results.

Carmona Moriel et al.22 applied our questionnaire to 18 carers of PD patients using a 1–5 rating scale. They reported an average overload score of 41.89±14.93; if we had scored the items from 1 to 5, we would have obtained a similar average of 44.27±14.35 points for overload. These results indicate that the CSCDP obtains similar results in an independent sample.

The reliability of the questionnaire is confirmed, insofar as in the high internal consistency of the 3 definitive dimensions used in the first validation study24 (Cronbach's α 0.86–0.89) has been reproduced in this new sample (Cronbach's α 0.887). Similarly, the acceptable results of the limited test-retest reported in our previous study shows the stability of the questionnaire over time.

Content validity was already evaluated by our group during development of the questionnaire by eliminating items that were not consistent with the purpose of the questionnaire.24

The exploratory factor analysis of the main components of the CSCDP reveals some interesting information with regard to construct validity. Component 2 correlates solely with patient dependency items, while component 1 correlates with all the objective or repercussion overload items, which is consistent with the design of the questionnaire. In contrast, subjective overload correlates with components 1 and 3, suggesting that some items in this dimension are perceived by the carers as an objective overload or as something with repercussions on their life, while the items that correlate with component 3 should relate to the carer's personal perception of caring for or helping the patient. Either way, the 2 overload dimensions measure the carer's perception of their relationship with the patient in respect of aspects that impact on the life and on the most personal feelings of the carer, which we conceive as overload. Moreover, the dependency dimension is well delimited and differentiated from the other 2 dimensions, all of which confirms the construct validity of the questionnaire.

To test the criterion validity, we compared the CSCDP with two similar questionnaires: the ZBI29,30 and the SF-36.28

The ZBI, widely used internationally33–37 and translated and validated in Spanish,31,32 consists of 22 items to evaluate the carer's overload. Items are rated from 0 to 4 in the English versions and in some Spanish versions13,22,32 and from 1 to 5 in other Spanish versions, including the validated version.31,32 The results of this scale vary considerably, depending on the rating system used and the population studied.

The score awarded to the items (0–4 or 1–5) will determine the cut-off points for different degrees of overload. The version in English and languages other than Spanish, with 22 items and a rating scale of 0–4, considers 4 levels of overload (see the methods section), but is only an approximation, since the results vary greatly depending on the population studied or the patients’ diseases.17,34,37 In the validated Spanish version with a 1 to 5 rating scale, 3 degrees of overload were determined with 2 cut-off points: no overload-slight overload 46/47, mild overload-severe overload 55/56.31 In a subsequent study, the cut-off points were re-set to 56/57 and 59/60, respectively, relative to a general health status questionnaire.32 Most Spanish studies, even those using the 0–4 rating scale22,38 or the 21-item ZBI44 use this classification; the resulting confusion has been described by Álvarez et al.45

It is interesting to note that studies published in English have reported the wide range of mean scores, from 22.40 to 39.10, obtained when applying the ZBI in diseases such as cancer, dementia and acquired brain injury.34,46 Likewise, in studies on carers of patients with PD, or PD vs. HD, the average scores range from 20.4 to 43.75, with no significant differences between the 2 treatments,2,13,14,37 although a Japanese study reported surprisingly low scores of between 12.5 and 14.1 points in PD.39

In our subjects, the ZBI and CSCDP showed good correlation (r=0.683). This was confirmed in the overload dimensions, and is further evidence of criterion validity. Carmona Moriel et al. confirmed the very good correlation (r=0.875; p=0.000) between both questionnaires administered in the same visit.22 In our study, however, they were separated by at least 15 days, and were not administered in the same order, which may explain in part the different degree of correlation. This good correlation confirms the criterion validity of the CSCDP.

The analysis of correspondence between the 3 (grouped) levels of overload of the ZBI and the tertiles of the CSCDP shows acceptable, though not excellent, levels of correspondence (Kendall's τ-b 0.570; p<0.001). This could be due to several factors: (1) changes in the carer's perception of overload in the time elapsed between the 2 questionnaires; however, the conditions that can lead to overload in the PD carer do not change in that 15–30 day period, so any perceived change would be minor; (2) the established degrees of overload do not correspond entirely; (3) the 2 instruments do not measure the same type of overload; for example, the dependency dimension of the CSCDP has very little correlation with the ZBI, and the exploratory factor analysis of the main components of the ZBI shows that 1 of the first 3 has no correlation at all with the CSCDP. When different scales measure different types of overload, this can affect the correspondence between different degrees used in the scales, and highlights the need to use questionnaires designed specifically for PD.

The authors declare that they have no conflicts of interest concerning the preparation and publication of this article.