Symptom clusters in chronic kidney disease and their association with people's ability to perform usual activities.

BACKGROUND: People living with a long-term condition, such as chronic kidney disease (CKD), often suffer from multiple symptoms simultaneously, making symptom management challenging. This study aimed to identify symptom clusters in adults with CKD across treatment groups and investigate their association with people's ability to perform their usual activities.
METHODS: We conducted a secondary analysis of both cross-sectional and longitudinal data collected as part of a national service improvement programme in 14 kidney centres in England, UK. This data included symptom severity (17 items, POS-S Renal) and the extent to which people had problems performing their usual activities (single item, EQ-5D-5L). We categorised data by treatment group: haemodialysis (n = 1,462), transplantation (n = 866), peritoneal dialysis (n = 127), or CKD without kidney replacement therapy (CKD non-KRT; n = 684). We used principal component analysis to identify symptom clusters per treatment group, and proportional odds models to assess the association between clusters and usual activities.
RESULTS: Overall, clusters related to: lack of energy and mobility; gastrointestinal; skin; and mental health. Across groups, the 'lack of energy and mobility' clusters were associated with having problems with usual activities, with odds ratios (OR) ranging between 1.24 (95% confidence interval [CI], 1.21-1.57) for haemodialysis and 1.56 for peritoneal dialysis (95% CI, 1.28-1.90). This association was confirmed longitudinally in haemodialysis (n = 399) and transplant (n = 249) subgroups. IMPLICATIONS: Our findings suggest that healthcare professionals should consider routinely assessing symptoms in the 'lack of energy & mobility' cluster in all people with CKD, regardless of whether they volunteer this information; not addressing these symptoms is likely to be related to them having problems with performing usual activities. Future studies should explore why symptoms within clusters commonly co-occur and how they interrelate. This will inform the development of cluster-level symptom management interventions with enhanced potential to improve outcomes for people with CKD.

Introduction

Symptom burden is high in people with long-term conditions [1-4]. This is also true for people with chronic kidney disease (CKD) across disease stages and treatments [5, 6]. Those with CKD may suffer from 6 to 20 symptoms simultaneously [7], which negatively affects their quality of life and increases the risk of treatment non-adherence, health care utilization, and mortality [8-10]. It is therefore not surprising that people with CKD have identified improving symptom management as a research priority [11], and that clinical practice guidelines include it as a key element of CKD care [12, 13]. However, considering the number of concurrent symptoms, it is challenging for healthcare professionals and patients to assess each symptom separately and develop treatment plans accordingly. This may explain why CKD symptom management is often suboptimal [14-16].

One way to address this problem is to develop management strategies that target multiple, potentially related, symptoms at once. This aligns with the suggestion that within and across long-term conditions symptoms often co-occur, i.e. cluster together [17-19]. In CKD, previous research proposed clusters related to fatigue, pain, gastrointestinal, and skin symptoms [20-28]. They also reported negative associations between these clusters and outcomes linked to quality of life, such as health functioning [21, 25–27, 29] and depression [21]. This implies that managing clusters, rather than individual symptoms, may be an effective way to improve outcomes related to quality of life for people with CKD [24] and to reduce the overall challenge faced by both healthcare professionals and people with CKD in considering the number of concurrent symptoms.

CKD symptom clusters have been mainly investigated in people receiving dialysis [28]. Studies considering other kidney replacement therapies (KRTs) or CKD stages often identified clusters across rather than stratified by treatment or stage [20, 21, 23, 26, 27]; one reason for this lack of stratification might have been the modest sample sizes of <450 people [20, 21, 23, 26, 27]. Only few studies looked at symptom clusters specifically for people with a kidney transplant [22] or with CKD but not on KRT (CKD non-KRT) [25], despite these people reporting a similar symptom burden [5, 6, 30]. This leaves it largely unknown to what extent current clusters generalise across or differ between treatment groups.

This study, therefore, aimed to (1) explore symptom clustering in a large data set, stratified by CKD treatment group, (2) assess the relevance of the identified clusters by investigating their association with people’s ability to perform their usual activities, and (3) determine if these associations were stable over time. We anticipate the findings to contribute to developing cluster-level symptom management strategies with potential to improve outcomes for people with CKD.

Methods

We conducted a secondary analyses of cross-sectional and longitudinal data collected in the context of a national service improvement programme in 14 kidney centres across England (UK) called Transforming Participation in Chronic Kidney Disease (TP-CKD) [31]. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [32] (see S1 Table) and the nomenclature for kidney function and disease proposed by Levey and colleagues [33] for reporting our findings.

Transforming Participation in Chronic Kidney Disease (TP-CKD)

The TP-CKD programme aimed to support people with CKD to better manage and make decisions about their own care and treatment. This included introducing collection of patient-reported outcome data in 14 English kidney centres [34]. Between December 2015 and December 2017, members of local kidney care teams approached eligible patients in their centre. People were eligible if they were aged 18 years or over and receiving care for any stage of CKD or on any form of KRT. The number and sociodemographic characteristics of people on KRT treated in each centre are provided in S2 Table; this does not include people with CKD not on KRT because this data was not available. People who were interested received a paper copy of the questionnaire, which they returned to their unit upon completion. Centres did not record information on whom they screened for eligibility, who had been confirmed eligible but declined participation, or on reasons for declining participation. In twelve centres, people who had previously taken part were invited to complete the questionnaire again at a later time.

Measures of exposure and outcome

In the current secondary analysis, we used symptom burden as the measure of exposure. To assess exposure, we analysed data collected through the Palliative care Outcome Scale-Symptom (POS-S) Renal [35], which was one of the patient-reported outcome measures included in the paper-based TP-CKD questionnaire. The POS-S Renal consists of 17 symptoms that are common for people with CKD [35], such as pain, weakness or lack of energy, and itching. For each symptom, respondents are asked to indicate to what extent they have been bothered by it over the last week on a scale from 0 (not at all) to 4 (overwhelmingly).

As the outcome measure we used an item from the EuroQol 5 dimensions—5 level (EQ-5D-5L) version [36], which was also included in the TP-CKD questionnaire. The EQ-5D-5L assesses general health status and asks respondents to rate whether they had problems doing their usual activities (e.g. work, study, housework, leisure activities) on a scale from 1 (no problem) to 5 (unable). People with CKD previously reported this item as being important to ‘living well’ (29). We did not consider the remaining EQ-5D-5L items as part of the outcome measure because: some fully overlapped with symptoms in the POS-S Renal (pain/discomfort; anxiety/depression; mobility); one was not reported by people with CKD as a priority outcome in research (self-care: wash and dress) [37, 38]; and one because its responses could not be scanned into electronic format (self-rated health status).

Data processing, linkage and ethics

The completed TP-CKD paper-based questionnaires were sent to the UK Renal Registry (UKRR) and scanned into electronic format. The UKKR used people’s unique National Health Service (NHS) number to link questionnaires to data on: date of birth; gender; ethnicity; index of multiple deprivation area quintile (proxy of socio-economic status [39] derived from postcode, with higher quintiles representing more social deprivation); treatment group; primary kidney disease diagnosis; and time on KRT. The UKRR annual report describes the definitions and measurement methods for these data items [40].

Because the primary purpose of TP-CKD was service evaluation and improvement rather than research, no formal ethical approval was required. People taking part in the TP-CKD programme implicitly consented to their data being processed and linked by returning a completed questionnaire(s) to their kidney centre. The UKRR holds permissions under s251 of the NHS Act 2006, to gather, process, and share confidential patient information for the purposes of audit and research. These permissions are renewed annually by the UK’s Health Research Authority’s Confidentiality Advisory Group. The collection and secondary analysis of the data for this study were approved by the UKRR (ref: UKRR ILD32) and carried out under the ethical permissions granted to the UKRR by the Research Ethics Committee.

Secondary data analysis

We included data collected during the TP-CKD programme in the secondary analysis if questionnaires had fewer than 4 missing symptom scores for the POS-S Renal and the ‘usual activities’ item on EQ-5D-5L had been completed. For people who completed a questionnaire on more than one occasion, we used the first questionnaire for the cross-sectional analyses, and the first and last questionnaires for the longitudinal analyses. For the latter, we only included people who were in the same treatment group when completing the first and last questionnaire. In the absence of information on non-responders, we assessed potential selection bias by comparing characteristics of people included in the analysis to those of the overall CKD and KRT population in the UK in 2016, which was the most recent information available at the time.

To identify symptom clusters, we conducted principal component analyses (PCA). We assumed polychoric correlations between symptoms due to the ordinal nature of the scores and stratified by treatment group (CKD non-KRT; peritoneal dialysis; haemodialysis; transplant). We applied oblique promax rotation to account for potential correlation between clusters. For determining the optimal number of clusters we considered: eigenvalues (>1); scree plots; clarity of clustering patterns (i.e. variables loading strongly onto one cluster only; no clusters of single variables); and the variance explained by each cluster [41]. Where there were no clear clustering patterns, we excluded symptoms that were inconsistent with the other symptoms (as indicated by Cronbach’s α), and re-ran the PCA [41]. We assigned symptoms to the cluster for which they had the highest factor loading, but only if this loading was >0.5 [27] and was >0.2 higher than loadings for other clusters [20]. We assessed each cluster’s internal consistency by calculating Cronbach’s α.

For the cross-sectional analyses investigating the association between symptom clusters and people’s ability to perform their usual activities, we developed proportional odds models per cluster-treatment group combination. The usual activities score was included as the ordinal outcome, and the total symptom cluster score (i.e. sum of individual symptom scores within the cluster) as a continuous predictor. In addition to unadjusted models (i.e. only including the cluster score), we also developed partially adjusted (i.e. also including age, gender, ethnicity, socio-economic status, and time on KRT) and fully adjusted models (i.e. additionally including scores of the other clusters and of symptoms not assigned to any cluster). We checked for violations of the proportional odds assumption and for non-linearity of covariate effects.

To handle missing data on symptoms, ethnicity, gender and socio-economic status, we used multiple imputation with fully conditional specification, assuming the data were missing at random. We carried out 20 imputations using all available symptom scores, usual activity scores, clinical and socio-demographic data as predictor variables. We checked distributions and correlations between variables, comparing imputed and observed data. Since the UKRR only has full coverage of people currently on KRT, we classified questionnaires as pertaining to the ‘CKD non-KRT’ group if they could not be linked to a UKRR record and the type of KRT was missing in the questionnaire.

To confirm the findings from the cross-sectional analyses, we conducted longitudinal analyses using the first and last questionnaire from the subsample of people who had been followed up over time. We developed the same set of proportional odds models using the clusters identified in the cross-sectional analyses, but now with (a) a within-person change in usual activities score as an ordinal outcome (decreased; stayed the same; increased) and a within-person change in total symptom cluster score as a continuous predictor; and (b) including baseline usual activities and cluster scores and time between surveys as additional covariates in the partially and fully adjusted models.

For all analyses, we used SAS version 9.4 and considered a p value of <0.5 significant. We did not perform any sensitivity analyses.

Results

Characteristics of people in the TP-CKD programme included in the analyses

The UKRR received questionnaires from 3,421 people, of whom 282 were excluded because they had ≥4 missing symptom scores or did not complete the ‘usual activities’ item on EQ-5D-5L (see S3 Table for their characteristics). The remaining 3,139 people who completed at least one TP-CKD questionnaire were included in the analysis. Table 1 shows their baseline characteristics. Overall, demographic characteristics of those included in our analyses were similar to characteristics of the overall CKD and KRT population in the UK (see S4 Table). Across the sample, people reported a median of five symptoms (interquartile range (IQR), 3 to 8). Weakness, poor mobility, and difficulty sleeping were the most prevalent symptoms (see Fig 1). Across the total sample, 22% of people reported having severe problems performing their usual activities.

Table 1

Baseline characteristics of people included in our analyses (values are numbers (% after excluding missing), unless indicated otherwise).

	All	CKD non-KRT a)	Peritoneal dialysis	Haemodialysis	Transplant
Total n (first survey)	3139 (100)	684 (21.8)	127 (4.1)	1462 (46.6)	866 (27.6)
Gender (male)	1636 (61.8)	147 (62.8)	73 (61.9)	881 (61.4)	535 (62.0)
Missing	490	450	9	28	3
Age (Mean, SD)	61.4, 16.0	63.3, 17.1	62.4, 16.0	64.5, 15.2	54.4, 14.1
Missing	2	2	0	0	0
Ethnicity
White	2170 (81.0)	239 (89.9)	96 (81.4)	1052 (73.5)	783 (90.8)
Asian	301 (11.2)	21 (7.9)	10 (8.5)	224 (15.6)	46 (5.3)
Black	159 (5.9)	4 (1.5)	9 (7.6)	128 (8.9)	18 (2.1)
Other	48 (1.8)	2 (0.8)	3 (2.5)	28 (2.0)	15 (1.7)
Missing	461	418	9	30	4
Social deprivation b)
IMD Quintile 1 (least deprived)	485 (15.6)	105 (15.8)	28 (22.1)	186 (12.7)	166 (19.2)
IMD Quintile 2	529 (17.0)	137 (20.6)	29 (22.8)	191 (13.1)	172 (19.9)
IMD Quintile 3	514 (16.5)	113 (17.0)	22 (17.3)	236 (16.2)	143 (16.5)
IMD Quintile 4	656 (21.0)	127 (19.1)	20 (15.8)	303 (20.8)	206 (23.8)
IMD Quintile 5 (most deprived)	934 (30.0)	183 (27.5)	28 (22.1)	544 (37.3)	179 (20.7)
Missing	21	19	0	2	0
Time on KRT in years (Mean, SD)	7.9, 8.6	N/A	2.3, 3.5	5.3, 6.9	12.9, 9.2
Primary kidney diagnosis
Diabetes	501 (19.7)	31 (23.0)	27 (23.1)	359 (25.1)	84 (9.7)
Glomerulonephritis	467 (18.3)	27 (20.0)	19 (16.2)	215 (15.0)	206 (23.9)
Hypertension	190 (7.5)	6 (4.4)	14 (12.0)	131 (9.2)	39 (4.5)
Polycystic kidney disease	257 (10.1)	14 (10.4)	10 (8.5)	83 (5.8)	150 (17.4)
Pyelonephritis	246 (9.7)	6 (4.4)	11 (9.4)	134 (9.4)	95 (11.0)
Kidney vascular disease	102 (4.0)	10 (7.4)	7 (6.0)	70 (4.9)	15 (1.7)
Other	384 (15.1)	17 (12.6)	14 (12.0)	195 (13.6)	158 (18.3)
Uncertain aetiology	398 (15.6)	24 (17.8)	15 (12.8)	243 (17.0)	116 (13.4)
Missing	594	549	10	32	3
Where questionnaire was completed
At home	688 (22.6)	51 (7.7)	53 (42.1)	454 (32.0)	130 (15.4)
In clinical settings (outpatient clinic, kidney unit, GP practice)	2358 (77.4)	609 (92.3)	73 (57.9)	963 (67.9)	713 (84.6)
Missing	93	24	1	45	23
How questionnaire was completed
Alone, without support	1910 (63.4)	432 (66.5)	81 (65.3)	708 (50.4)	689 (82.7)
With support from family member/friend	595 (19.8)	178 (27.4)	29 (23.4)	282 (20.1)	106 (12.7)
With support from clinical staff	506 (16.8)	40 (6.2)	14 (11.3)	414 (29.5)	38 (4.6)
Missing	128	34	3	58	33
Exposure & outcome c)
POS-S Renal
Number of symptoms (median, IQR) d)	5 (3,8)	5 (3,8)	5 (3,7)	6 (3,9)	4 (2,7)
EQ-5D-5L
Problems with usual activities e)
No problems	1034 (32.9)	236 (34.5)	44 (34.7)	325 (22.2)	429 (49.5)
Slight problems	695 (22.1)	160 (23.4)	37 (29.1)	322 (22.0)	176 (20.3)
Moderate problems	726 (23.1)	170 (24.9)	23 (18.1)	368 (25.2)	165 (19.1)
Severe problems	434 (13.8)	76 (11.1)	18 (14.2)	264 (18.1)	76 (8.8)
Unable to do usual activities	250 (8.0)	42 (6.1)	5 (3.9)	183 (12.5)	20 (2.3)
Any problems related to f)
Mobility	2130 (68.3)	436 (64.0)	89 (70.6)	1166 (80.3)	439 (51.0)
Self-care	1140 (36.5)	198 (29.0)	38 (30.2)	707 (48.6)	197 (22.9)
Pain/discomfort	1955 (62.6)	417 (61.6)	75 (59.1)	1003 (68.8)	460 (53.4)
Anxiety/depression	1583 (50.7)	354 (52.2)	64 (50.4)	795 (54.8)	370 (42.8)

Note. CKD non-KRT, people with chronic kidney disease not receiving kidney replacement therapy; EQ-5D-5L, EuroQOL Five Dimensions—5 levels version; GP, general practitioner; IMD, index of multiple deprivation; IQR, interquartile range; N/A, not applicable; POS-S Renal: Palliative care Outcome Scale-Symptom Renal; KRT: kidney replacement therapy; SD: standard deviation.

a) This included people with any stage of CKD not on KRT who were under the treatment of a kidney centre. The dataset did not contain information on the stage of CKD at enrollment.

b) Based on index of multiple deprivation quintiles (27)

c) For people who completed more than one questionnaire, we used the response from their first questionnaire for this table

d) Refers to number of symptoms with a score of >1 (i.e. reports of being at least slightly bothered by a symptom) from the POS-S Renal

e) Domain from the EQ-5D-5L; primary outcome measure of the current study

f) Refers to the remaining four dimensions within the EQ-5D-5L where people scored >1 (i.e. reports of having at least slight problems)

Fig 1

Prevalence of individual symptoms, presented by treatment group; people were considered to have the symptom if they had a score of > 1 on the POS-S Renal (i.e. reported being at least slightly bothered by a symptom).

A subsample of 699 people completed follow-up questionnaires at a median of 203 days (IQR, 133 to 301) after baseline, while remaining in the same treatment group. This subsample were mostly people on haemodialysis (n = 399) and with a transplant (n = 249). Their baseline characteristics were comparable to those of the overall sample (see S5 Table).

Symptom clusters by treatment group

Fig 2 shows the clusters we identified and the symptoms within them.

Fig 2

Symptom clusters from principal component analyses (PCA), stratified by treatment group; (a) CKD non-KRT, (b) peritoneal dialysis, (c) haemodialysis and (d) transplant. The size of the circles is proportional to the PCA loading which indicates how prominent the variable is within that cluster.

Table 2 displays in more detail the final assignment of symptoms to clusters, their factor loadings, and each clusters’ Cronbach’s α (see S6 Table for factor loadings prior to final cluster composition).

Table 2

Clusters, symptom loadings and internal consistency (Cronbach’s ɑ) resulting from principal component analyses.

	CKD non-KRT			Peritoneal dialysis				Haemodialysis			Transplant
Symptom clusters	Lack of energy & mobility	GI	Mental health	Lack of energy & mobility	GI	Skin	Mental health	Lack of energy & mobility	GI	Skin	Lack of energy & mobility	GI	Skin
Individual symptoms
Pain	0.52						0.62	0.66			0.72
Shortness of breath	0.82							0.61			0.75
Weakness	0.69			0.75				0.82			0.74
Nausea		0.80			0.78				0.85			0.88
Vomiting		0.95			0.83				0.94			0.98
Poor appetite		0.61						0.53				0.65
Constipation		0.58		-	-	-	-
Sore/dry mouth	0.51										0.62
Drowsiness	0.74			0.73				0.70			0.79
Poor mobility	0.84			0.76				0.86			0.96
Itching						0.83				0.83			0.84
Difficulty sleeping										0.59
Restless legs				-	-	-	-			0.63
Changes in skin						0.74				0.71			0.78
Diarrhoea		0.61					0.75		0.55			0.61
Feeling anxious			0.89				0.76	0.59			0.55
Feeling depressed			0.79				0.61	0.60			0.55
Cronbach’s α	0.82	0.72	0.76	0.74	0.85	0.56	0.74	0.85	0.73	0.71	0.88	0.72	0.54
Variance (%)	44.6	7.5	6.3	40.3	9.2	7.2	10.7	43.9	5.9	7.0	49.3	7.1	6.8

Note. GI, gastrointestinal; CKD non-KRT, people with chronic kidney disease not receiving kidney replacement therapy

Based on the key symptoms making up the clusters, we labelled clusters as being related to: lack of energy and mobility; gastrointestinal; mental health; and skin. Post-hoc analyses to explore why pain was in the ‘lack of energy and mobility’ (CKD non-KRT, haemodialysis, and transplant) and diarrhoea in the ‘mental health’ cluster (peritoneal dialysis) suggested that this was because they strongly correlated with the most prominent symptom in that cluster (i.e. poor mobility and feeling anxious, respectively).

Relationship between symptom clusters and people’s ability to do usual activities

Table 3 displays the unadjusted, partially, and fully adjusted proportional odds ratios for the cross-sectional associations between people’s total symptom cluster scores and their ability to do usual activities. There was an indication of violation of the proportional odds assumption in some models where comparing a score of 5 versus a score of 4 and lower, but because few people reported a score of 5 (i.e., unable to do my usual activities) this only affected a small proportion of the data.

Table 3

Association between total symptom cluster scores and ability to perform usual activities.

	CKD non-KRT			Peritoneal dialysis			Haemodialysis			Transplant
Symptom clusters	Unadjusted	Partially adjusted	Fully adjusted	Unadjusted	Partially adjusted	Fully adjusted	Unadjusted	Partially adjusted	Fully adjusted	Unadjusted	Partially adjusted	Fully adjusted
Lack of energy & mobility	1.40 (1.35,1.46)	1.40 (1.35,1.46)	1.32 (1.26,1.38)	1.61 (1.4,1.86)	1.76 (1.5,2.07)	1.56 (1.28,1.9)	1.24 (1.22,1.26)	1.25 (1.23,1.27)	1.24 (1.21,1.27)	1.33 (1.3,1.37)	1.35 (1.31,1.39)	1.37 (1.32,1.42)
GI	1.30 (1.24,1.37)	1.34 (1.27,1.41)	1.02 (0.96,1.08)	1.58 (1.3,1.93)	1.29 (1.08,1.55)	1.10 (0.84,1.43)	1.30 (1.25,1.36)	1.33 (1.27,1.39)	1.00 (0.95,1.05)	1.39 (1.31,1.47)	1.41 (1.33,1.5)	0.96 (0.9,1.04)
Skin				1.29 (1.08,1.53)	1.64 (1.32,2.04)	1.15 (0.92,1.45)	1.24 (1.21,1.28)	1.26 (1.23,1.3)	1.03 (1.00,1.07)	1.33 (1.23,1.44)	1.30 (1.2,1.41)	0.86 (0.77,0.95)
Mental health	1.56 (1.45,1.69)	1.64 (1.51,1.78)	1.21 (1.1,1.33)	1.30 (1.16,1.44)	1.37 (1.21,1.54)	1.10 (0.95,1.27)

Note. GI, gastrointestinal; CKD non-KRT, people with chronic kidney disease not receiving kidney replacement therapy; Values are odds ratios (95% confidence intervals), reflecting the odds of having morea problems with performing usual activities per unit increase in total symptom cluster score; Partially adjusted—adjusted for age, sex, ethnicity, Index of Multiple Deprivation (IMD) quintile and time on kidney replacement therapy (KRT, for haemodialysis, peritoneal dialysis and transplant groups only). Fully adjusted–adjusted for all variables in the partially adjusted model, and additionally for the total scores of all other symptom clusters and scores of any individual symptoms not appearing in any cluster. Different symptoms comprised each cluster per group: Lack of energy and mobility (CKD non-KRT– 6 symptoms—pain, shortness of breath, weakness, sore/dry mouth, drowsiness, poor mobility; Peritoneal dialysis– 3 symptoms–weakness, drowsiness, poor mobility; Haemodialysis– 8 symptoms—pain, shortness of breath, weakness, poor appetite, drowsiness, poor mobility, feeling anxious, feeling depressed; Transplant– 8 symptoms—pain, shortness of breath, weakness, sore/dry mouth, drowsiness, poor mobility, feeling anxious, feeling depressed); GI (CKD non-KRT– 5 symptoms–nausea, vomiting, poor appetite, constipation, diarrhoea; Peritoneal dialysis– 2 symptoms–nausea and vomiting; Haemodialysis– 3 symptoms–nausea, vomiting, diarrhoea; Transplant– 4 symptoms–nausea, vomiting, poor appetite, diarrhoea), Skin (CKD non-KRT–not applicable; Peritoneal dialysis– 2 symptoms–itching, changes in skin; Haemodialysis– 4 symptoms–itching, difficulty sleeping, restless legs, changes in skin; Transplant—2 symptoms–itching, changes in skin), Mental health (CKD non-KRT– 2 symptoms—feeling anxious, feeling depressed; Peritoneal dialysis– 4 symptoms–pain, diarrhoea, feeling anxious, feeling depressed).

a ‘More’ refers to any one step increase on the ordinal scale of the outcome measure (e.g. from no to slight problems, or from severe to extreme problems)

Across all clusters and treatment groups, the unadjusted and partially adjusted models showed a significant association between total cluster symptom scores and having problems with usual activities. However, after fully adjusting for other clusters and individual symptoms, only the association between the ‘lack of energy and mobility’ cluster and problems with usual activities remained significant across treatment groups (i.e. accounting for the ‘lack of energy and mobility’ cluster attenuated the ORs for other symptom clusters). The fully adjusted models showed that haemodialysis patients had 24% (95% confidence interval (CI), 21–27%) higher odds of having more problems with doing their usual activities for each unit increase in the total score for the ‘lack of energy and mobility’ cluster. This was 32% (95% CI, 26–38%), 56% (95% CI, 28–90%) and 37% (95% CI, 32–42%) for the CKD non-KRT, peritoneal dialysis and transplant groups, respectively. Other associations remaining significant after full adjustment were those for the ‘mental health’ cluster for the CKD non-KRT group (OR, 1.21; 95% CI, 1.1–1.33) and the ‘skin’ cluster for haemodialysis (OR, 1.03; 95% CI, 1.00–1.07) and transplant groups (OR, 0.86; 95% CI, 0.77–0.95).

Since only 6% (n = 43) and 1% (n = 8) of CKD non-KRT and peritoneal dialysis of people in the these groups completed follow-up questionnaires, we limited the longitudinal analyses to haemodialysis and transplant groups (see S7 Table for a summary of changes in total symptom cluster scores and changes in problems with usual activities per treatment group). We found that within-person changes in people’s ‘lack of energy and mobility’ cluster score were associated with higher odds of having increased problems with performing usual activities. This confirms the findings from the cross-sectional analyses, except for the ‘skin’ cluster (see Table 4).

Table 4

Association between within-person change in total symptom cluster scores and within-person change in ability to perform usual activities.

	Haemodialysis			Transplant
Symptom clusters	Unadjusted	Partially Adjusted	Fully adjusted	Unadjusted	Partially Adjusted	Fully adjusted
Lack of energy & mobility	1.13 (1.09,1.18)	1.22 (1.16,1.29)	1.23 (1.16,1.31)	1.29 (1.19,1.39)	1.53 (1.37,1.71)	1.50 (1.32,1.71)
Gastrointestinal	1.10 (1.00,1.21)	1.21 (1.07,1.37)	0.95 (0.82,1.1)	1.30 (1.12,1.52)	1.46 (1.22,1.76)	0.95 (0.75,1.2)
Skin	1.12 (1.05,1.19)	1.15 (1.06,1.24)	0.97 (0.89,1.07)	1.13 (0.95,1.35)	1.41 (1.11,1.8)	1.11 (0.81,1.53)

Note. Values are odds ratios (95% confidence intervals), reflecting the odds of having increased problems with performing usual activities per unit increase in the change in total symptom cluster score; Partially adjusted—adjusted for age, sex, ethnicity, Index of Multiple Deprivation (IMD) quintile and time on kidney replacement therapy. Fully adjusted–adjusted for all variables in the partially adjusted model, and additionally for the total scores of all other symptom clusters and scores of any individual symptoms not appearing in any cluster. Different symptoms comprised each cluster per group (underline denotes symptoms appearing in one patient group but not the other): Lack of energy and mobility (8 symptoms in each, Haemodialysis—pain, shortness of breath, weakness, poor appetite, drowsiness, poor mobility, feeling anxious, feeling depressed; Transplant—pain, shortness of breath, weakness, sore/dry mouth, drowsiness, poor mobility, feeling anxious, feeling depressed); Gastrointestinal (Haemodialysis– 3 symptoms–nausea, vomiting, diarrhoea; Transplant– 4 symptoms–nausea, vomiting, poor appetite, diarrhoea), Skin (Haemodialysis– 4 symptoms–itching, difficulty sleeping, restless legs, changes in skin; Transplant—2 symptoms–itching, changes in skin).

Discussion

Summary of the findings

This study found that, overall, CKD symptom clusters related to lack of energy and mobility, gastrointestinal, skin, and mental health. Although clusters varied between treatment groups, the ‘lack of energy and mobility’ clusters were consistently associated with having problems doing usual activities across groups, both cross-sectionally and longitudinally.

Relation to other studies

We used the largest data set to date to investigate symptom clusters for people with CKD. It provides a robust external validation of clusters across treatment groups identified by previous studies, including skin [21, 25–27], gastrointestinal [20, 23], and mental health [25, 26], which suggests that our findings are generalisable beyond the study context. The alignment with previous research on CKD symptom clusters is in contrast with other disease areas, such as cancer, where there seems to be less consistency in symptom clusters across studies [19].

In keeping with previous research [21, 23, 25–27], we found that lack of energy, problems with mobility, pain, mental health, and shortness of breath were the most frequently co-occurring symptoms. Whereas we labelled this cluster as ‘lack of energy and mobility’ across all treatment groups, other studies referred to it as ‘uraemic’. Another difference was that in other studies, it included dizziness [21, 27] and difficulty concentrating [20, 26], which were not listed in the POS-S Renal, the measure of symptom burden in our study. Furthermore, the ‘uraemic’ cluster in other studies included nausea and poor appetite [20, 21, 26, 27], whereas in our study, these symptoms, together with vomiting, formed a gastrointestinal cluster. This finding strongly aligns with wider symptom cluster research commonly identifying a separate gastrointestinal cluster consisting of nausea and vomiting [19]. Lastly, unlike other studies [20, 21, 27], we did not find neuro-muscular clusters because the type of symptoms within these (e.g. muscle soreness, numbness or tingling of feet) were not part of our symptom burden score.

The POS-S Renal also does not include fatigue, which is a common symptom in people with CKD that others found to pertain to several symptom clusters [20]. Lack of energy and weakness represent some but not all aspects of fatigue, which is a multi-faceted symptom related to sociodemographic, psychological, clinical and biochemical factors [42]. This implies that part of the people reporting lack of energy and weakness in our study might have actually suffered from fatigue, which may partly explain the strong associations we found between ‘lack of energy and weakness’ clusters and people’s ability to do usual activities.

In the CKD non-KRT treatment group, we found that frequently co-occurring mental health symptoms, i.e. feeling anxious and feeling depressed, formed a stand-alone cluster, which differs from what Lee and Jeon reported in their smaller, Korean study [25]. An explanation for mental health being a separate cluster may be that, as people approach kidney failure, they often experience more psychological symptoms [7]. In the peritoneal dialysis group in the current study, this cluster also included pain and diarrhoea, both of which may negatively affect mental health [43, 44].

Our analyses showed that the ‘lack of energy and mobility’ clusters were associated with problems performing usual activities in all groups and models, which echoes the findings of previous research in which uraemic clusters were associated with impaired physical and mental functioning [20, 26, 27, 29]. For the haemodialysis and transplant groups, this association remained significant over time, similar to what Ng et al. found [26]. Additionally, an increase in the CKD non-KRT group’s mental health cluster score was associated with increased odds of problems with usual activities, which complements the findings of Lee and Jeon [25]. Compared to our findings, some previous research in people with CKD found more clusters to be associated with patient-reported outcomes [28]. This may be partly explained by the fact that the exposure and outcome measures in these studies often overlapped, thereby introducing an element of circular reasoning in their design. For example, ‘pain’ was measured both as a symptom (exposure) and as an aspect of physical functioning (outcome) in several studies [21, 24–27]. We addressed this by using a single item to measure the outcome, which reflected the patient-important outcome of ‘living well’ [37], thus reducing overlap with symptoms in the POS-S Renal.

Limitations

One limitation of our study is the lack of information on which people were selected and invited to complete the TP-CKD questionnaire, and what the characteristics were of those who declined. Although people included in our analyses were representative for the overall CKD population in the UK, we cannot rule out the potential presence of selection bias in our data set. In addition, the data analysed in this study was collected in 2017, and we propose repeating the analysis in the future when updated registry data becomes available.

Furthermore, we did not have access to comorbidity data. Comorbidities are common in people with CKD [45] and may cause symptoms that were not part of our symptom burden score. Furthermore, comorbidities may have a negative impact on people’s ability to perform daily activities by increasing the chance of health-related problems, such as medication side effects and hospital admissions. Comorbidities are, therefore, an unmeasured confounder in our analyses that might have explained part of the variation in our outcome measure.

Lastly, due to the small number of people in the CKD non-KRT and peritoneal dialysis groups who completed the questionnaire on more than one occasion, we could not perform a longitudinal analysis to confirm the findings from the cross-sectional analysis in these groups. Furthermore, variation in the timing of follow-ups did not allow assessment of the stability of the symptom clusters over time. The stability of symptom clusters over time has direct implications on symptom management strategies and is an important next step in symptom cluster research [17, 19, 28].

How healthcare professionals could use the study findings

The study findings provide pointers for how healthcare professionals could improve CKD symptom assessments in clinic to reduce symptom underreporting and undertreatment. Although the need to assess and treat people’s symptom burden has been recognised [46], many symptoms remain unreported [47] and go untreated despite the availability of nonpharmacological and pharmacological interventions [16, 48]. Current guidelines recommend the assessment of symptoms by validated measures [13], such as the POS-S Renal [46]. However, these assessment tools are often long lists of individual symptoms, and time-constrained clinics are unlikely to be able to assess each symptom separately.

Based on the study findings, healthcare professionals could consider the following step-wise, more efficient cluster-based approach to symptom assessment:

Ask people with CKD an open-ended question about what symptoms are bothering them; even though our analyses indicate which co-occurring symptoms are related to problems with usual activities across treatment group, they do not necessarily reflect which symptoms matter at the individual level. Therefore, opening the discussion with this question remains the most straightforward and personalised way of assessing symptom burden.

If a person volunteers a symptom, pro-actively ask about the other symptoms in the cluster to which that symptom pertains because our analyses showed that they frequently co-occur. People do not always volunteer all symptoms at clinic appointments for a variety of reasons, such as not knowing that some symptoms may be related to their CKD, or assuming there is no treatment available [15, 49]. Encouraging people with CKD to report specific symptoms could address this, especially since some may withhold this information until prompted by a healthcare professional [49].

If a person does not volunteer any symptoms, ask specifically about symptoms in the ‘lack of energy and mobility’ cluster. In all treatment groups, these were the most commonly reported and consistently associated with problems with usual activities. It is important to note is that mental health symptoms are key components of this cluster for the haemodialysis and transplant groups and formed a unique cluster for the CKD non-KRT treatment group, also with significant associations with problems with usual activities.

In addition, the results from our longitudinal analyses in the haemodialysis and transplant groups suggested that healthcare professionals might consider monitoring within-patient increases in the ‘lack of energy and mobility’ cluster score because these could indicate a reduced ability to perform usual activities.

Future research

Our analyses indicated that symptoms cluster together in certain ways for certain treatment groups, but without providing an explanation for why symptoms within these clusters co-occurred. Some have suggested that certain symptoms may share an underlying mechanism [18, 20, 28] or one symptom that ‘drives’ them [19]. Future studies, building on and adding to the wider cluster research agenda [19] are thus warranted to enhance our understanding in this area, alongside qualitative studies to ascertain the clinical face validity of symptom clusters [17]. Ultimately, this should inform future studies into effective interventions that incorporate our suggested cluster-based approach to symptom assessments, complemented with clinical practice guidance on how to manage symptoms [12, 13, 46].

Conclusion

This study identified symptom clusters related to lack of energy and mobility, gastrointestinal, skin, and mental health across CKD treatment groups, including both people on KRT and those with CKD but not on KRT. The ‘lack of energy and mobility’ clusters were consistently associated with having problems with usual activities across treatment groups. The study findings provide pointers for a more efficient cluster-based approach to symptom assessments by healthcare professionals. Future studies should focus on developing cluster-level symptom management interventions with enhanced potential to improve the care and outcomes for people with CKD.

STROBE statement [32]—checklist of items and where in the manuscript they have been included.

(DOCX)

Click here for additional data file.

Sociodemographic characteristics of people on kidney replacement treatments in the UK Renal Registry at 31st December 2016 for the participating centres.

(DOCX)

Click here for additional data file.

Baseline characteristics for people with ≥4 missing symptom or without a score for ‘problems with usual activities’ whose questionnaires were excluded from all analyses.

(DOCX)

Click here for additional data file.

Baseline characteristics for all people with CKD on KRT in the UK Renal Registry at 31st December 2016.

(DOCX)

Click here for additional data file.

Baseline characteristics of people with more than one survey.

(DOCX)

Click here for additional data file.

Principal component loadings of all symptoms prior to final cluster composition.

(DOCX)

Click here for additional data file.

Summary of changes total symptom cluster scores and changes in usual activities score and per treatment group.

(DOCX)

Click here for additional data file.

10 Jun 2021

PONE-D-21-05457

Symptom clusters in chronic kidney disease and their association with patients’ ability to perform usual activities

PLOS ONE

Dear Dr. Moore,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Jul 24 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see:  http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at  https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Boris Bikbov

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. If you are reporting a retrospective study of medical records or archived samples, please ensure that you have discussed whether all data were fully anonymized before you accessed them and/or whether the IRB or ethics committee waived the requirement for informed consent. If patients provided informed written consent to have data from their medical records used in research, please include this information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

Reviewer #3: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: I Don't Know

Reviewer #2: Yes

Reviewer #3: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Dear autor,

Several studies point to the worsening of quality of life in patients with chronic kidney disease and

I highlight some points of improvement in the manuscript:

1. The abstract, the methodology is not clear and objective and compromises the analysis of the results.

2. The background the introduction could focus more on the effects of symptoms and improve the quality of life of these patients.

3. Methods: This Topic could be more objective with a description of the applied methodology.

4. The statistical analyses: there was no robust statistical analysis.

5. this study uses data from 2017 and may be more outdated.

6. The author could discuss in more detail the follow-up of these patients and opportunities to apply and approach these symptoms.

7. Is important describe the sociodemographic data of these centers for explain the symptoms is regulary for everybody.

Thank you!

Reviewer #2: Dear Authors,

Thank you for an interesting manuscript and the following comments are provided to improve the manuscript.

Title, Abstract, Introduction and Background

- Please do not label patients by a disease or treatment (e.g. CKD patient groups, haemodialysis patient, TP-CKD participants, etc) – amend in the abstract and throughout the manuscript

- Please use the current international nomenclature for kidney disease published in the journal Kidney International (Levey et al 2020) – such as kidney replacement therapy.

- No need to use “our”, “we use” “we found”, “our findings” and so on – simply rephrase these types of sentences throughout the manuscript

- Add design and data collection methods (symptom tool, etc) into abstract

- Please write the abstract and the entire manuscript according to STROBE guidelines.

- Amend sentence (lines 81-83) as there is at least one study which included patients not receiving KRT (ref 20 - Almutary et al included almost 25% non-KRT [both CKD G4 and G5]).

- Almutary et al 2017 also constructed structural equation modelling demonstrating the direct (casual) relationships between symptom clusters and health-related quality of life

- Sentence (lines 85-86), the phrase “modest sample sizes” is vague – especially when cluster analysis is only possible when there is a sufficient sample size to do this type of statistical analysis. Arguably more studies on CKD symptom clusters because symptoms are subjective and could reflect particular cultures, health experiences, etc, etc. There is no need to avoid needing replication studies in another patient population to add to evidence.

- Clarify the aims related to longitudinal data collection.

Methods

- Ensure all STROBE aspects are explicitly described.

- For the TP-CKD program, at grade of CKD are people enrolled? If CKD G4 were included, then further clarity in data analysis and results will be required; that is, how were the different groups handled (i.e. CKD G4, CKD G5, KRT groups)

- Clarify how PROM data was collected.

- Clarify ethical procedures for TP-CKD.

- Data analysis – at what level was significance set?

Results

- Very thorough and interesting

Discussion

- First paragraph – there is no need to restate design, sample and sites for the study, instead interpret the meaning of the overall results (without restating actual numbers).

- Add to several paragraphs in the discussion what explanation can be made for the differences found between this study and previous studies on symptom clusters (why isn’t there more consistency which is only partly explained by different symptom measures used). C. Miaskowski research in cancer symptom clusters and H.A. de Von in CVD symptom clusters is likely to be insightful for this manuscript.

- Lack of energy and its similarity (or difference) to fatigue also requires further interpretation.

- In the discussion on this study’s symptoms and activity, it would also be worth reviewing and discussing with regard to Almutary et al 2017 CKD symptom cluster modelling which demonstrated significant casual relationships with physical function (and also with mental health)

Clinical Implications

- Clarify why the IPOS-Renal is not recommend as this PROM is widely used (rather than POS-S renal).

Future Research

- Again C. Miaskowski and colleagues are exploring underlying mechanisms in cancer symptoms and symptom cluster. There is an opportunity for researchers in CKD to draw from this work and extend it.

Reviewer #3: This was a study that used secondary analysis of data gathered as part of a national level research in the UK to examine association between symptom clusters and usual activities in chronic kidney disease (CKD).

The manuscript has been written according to the author guidelines. Introduction section provides a background and good overview of contemporary literature relevant to the study. Statistical analysis was adequately and appropriately described. Results, discussion and conclusions sections were described in line with study aims. The manuscript was well presented.

Minor comments

• Currently, the terminology of CKD has been revised, enabling researchers to compare results across studies conveniently. For example, end-stage kidney failure or end-stage kidney disease, now termed as kidney failure. Please refer...

Levey et al. (2020). Nomenclature for kidney function and disease: Executive summary and glossary from a Kidney Disease: Improving Global Outcomes (KDIGO) consensus conference. Journal of Nephrology, 33(4), 639–648. doi:10.1007/s40620-020-00773-6

Threfore, it is recommend to reword CKD terminology throughout the manuscript to fit current guidelines.

• It is also suggested not to label people living with chronic diseases as patients (e.g., line 89: CKD patient groups) consistently throughout the manuscript.

• Methods: Line 119: Spell out EQ5D5L first, before using it in abbreviated form.

• Methods: Line 128: It is not clear that what do you mean by ‘see below’ as there is no further information given in relation to self-rated health.

• Line numbers are missing after the line 266. Therefore, it is difficult to point comments on exact places.

• Under discussion, relation to other studies, second paragraph; it is better to mention frequently co-occurring mental health symptoms. i.e., anxiety and depression

• According to current evidence, fatigue is the most frequent symptom experienced by those with CKD. See for example, Almutary et al. (2013) and (2016) in your reference list. However, POS-S-renal doesn’t capture fatigue. It is note that weakness/lack of energy is different to fatigue. Therefore, it is important to discuss this point in the discussion under the section ‘relation to other studies’.

• Use the RRT abbreviation consistently throughout the manuscript. Abbreviated form of renal replacement therapy is missing in the conclusions section.

• It is suggested to check the reference list to align journal requirements.

Thank you for the opportunity given to review this manuscript.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: Yes: Dr Harith Eranga Yapa

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

18 Sep 2021

Point-by-point response to reviewers’ comments

Manuscript: Symptom clusters in chronic kidney disease and their association with patients’ ability to perform usual activities (PONE-D-21-05457)

Page and line numbers refer to those in the ‘clean’ version of the revised manuscript.

Reviewer #1

1. The abstract, the methodology is not clear and objective and compromises the analysis of the results.

We have revised the abstract and made several edits with the aim of making our study and methodology clearer.

2. The background the introduction could focus more on the effects of symptoms and improve the quality of life of these patients.

In the Introduction, we have clarified that symptoms, and particularly symptom clusters, are associated with poorer quality of life and health functioning and higher levels of depression, and that managing symptoms clusters may improve outcomes related to quality of life. For example, by including sentences such as:

• “Those with CKD may suffer from 6 to 20 symptoms simultaneously, which negatively affects quality of life and increases the risk of treatment non-adherence, health care utilization, and mortality.” (page 4, lines 63-65)

• “… previous research proposed clusters related to fatigue, pain, gastrointestinal, and skin symptoms. They also reported negative associations between these clusters and outcomes linked to quality of life, such as health functioning and depression. This implies that managing clusters, rather than individual symptoms, may be an effective way to improve outcomes related to quality of life for people with CKD…” (page 4, lines 75-80)

3. Methods: This Topic could be more objective with a description of the applied methodology.

We have further clarified our methodology by making several changes to the Methods section, such as clarifying:

• In the first sentence of the Methods that: “We conducted secondary analyses of cross-sectional and longitudinal data” (pg 5, lines 100-101)

• How patient-reported outcome data was collected: (pg 6, lines 116-117; pg 7, lines 143-144) “People who were interested received a paper copy of the questionnaire, which they returned to their unit upon completion. […] The completed TP-CKD paper-based questionnaires were sent to the UK Renal Registry (UKRR) and scanned into electronic format.”

• The assessment methods for demographic and treatment data (pg 7, lines 148-150): “The UKRR annual report describes the definitions and measurement methods for these data items [ref].”

• Explaining why the self-reported health status question was not included in the analysis (pg 7, lines 136-141): “We did not consider the remaining EQ-5D-5L items as part of our the outcome measure […] because […] responses could not be scanned into electronic format (self-rated health status).”

• That we did not perform sensitivity analyses (pg 10, line 213): “We did not perform any sensitivity analyses.”

4. The statistical analyses: there was no robust statistical analysis.

We have described our approach to the statistical analyses in the Methods section under ‘Secondary data analysis’ (pg 8-10, lines 161-213). This approach is in keeping with previous studies that investigated symptom clusters in people with chronic kidney disease, such as Almutary et al 2016 and Thong et al 2009. We have referenced these studies in our manuscript. We have also clarified that “For all analyses, we used SAS version 9.4 for all analyses and considered a p value of <0.5 significant.” (pg 10, line 212-213).

5. This study uses data from 2017 and may be more outdated.

This is a valid point for which we have added the following sentence in the Limitations section of the Discussion to address this (page x, lines x-x):

“In addition, the data analysed in this study was collected in 2017, and we propose repeating the analysis in the future when updated registry data becomes available.”

6. The author could discuss in more detail the follow-up of these patients and opportunities to apply and approach these symptoms.

In the Discussion, we have suggested to combine our proposed cluster-based approach to symptom assessment with existing clinical practice guidance on symptom management by adding the following sentence (pg 24, lines 361-363):

“… effective interventions that incorporate our suggested cluster-based approach to symptom assessments, complemented with clinical practice guidance on how to manage symptoms [refs].”

7. Is important describe the sociodemographic data of these centers for explain the symptoms is regulary for everybody.

Thank you for this valuable suggestion. We have added a table presenting the sociodemographic characteristics for all people on a KRT treated by each of the participating centres (see Supplementary material, Table S1).

Reviewer #2

Title, Abstract, Introduction and Background

1. Please do not label patients by a disease or treatment (e.g. CKD patient groups, haemodialysis patient, TP-CKD participants, etc) – amend in the abstract and throughout the manuscript

We very much appreciate this advice and have addressed this throughout the manuscript as much as possible. For example, we changed ‘kidney patients’ to ‘people with CKD’, and we now state that we ‘categorised the data into treatment groups’, rather than ‘categorising participants into patient groups’.

2. Please use the current international nomenclature for kidney disease published in the journal Kidney International (Levey et al 2020) – such as kidney replacement therapy.

We very much appreciate the reviewer pointing us to Levey et al’s 2020 nomenclature. We have made considerable changes throughout the text and believe it is now in line with this nomenclature. We have also stated at the start of the Methods that we followed Levey et al for reporting our study findings (pg 5, lines 103-106).

3. No need to use “our”, “we use” “we found”, “our findings” and so on – simply rephrase these types of sentences throughout the manuscript

We reviewed the manuscript and identified all occasions where we could reasonably remove ‘our’, or replace ‘we’ by a noun without changing the meaning of sentences. Examples in the abstract include:

• “In our analysis, We categorised data by treatment group: …” (Methods section)

• Overall, we identified clusters related to: …” (Results section)

However, in most instances, we used ‘we’ to avoid passive voice, which is in line with all major style guides and advocated by most academic journals. We have therefore kept these phrases as they were.

4. Add design and data collection methods (symptom tool, etc) into abstract

Thanks for recommending this edit - we have now stated this in the abstract as follows:

“We conducted a secondary analysis of both cross-sectional and longitudinal data collected as part of a national service improvement programme in 14 renal centres in England, UK. This data included symptom severity (17 items, POS-Renal) and the extent to which people had problems performing their usual activities (single item, EQ-5D-5L).”

5. Please write the abstract and the entire manuscript according to STROBE guidelines.

We have used the STROBE guideline to review our manuscript and revised it accordingly. For example, we have added information on:

• assessment methods for demographic and treatment data (STROBE item 8):“The UKRR annual report describes the definitions and measurement methods for these data items [ref].”

• sensitivity analyses (STROBE item12c): “We did not perform any sensitivity analyses.”

• numbers of individuals at each stage of study (STROBE item 13). For example: “The UKRR received questionnaires from 3,421 people, of whom 282 were excluded because they had ≥4 missing symptom scores or did not complete the ‘usual activities’ item on EQ-5D-5L (see Supplementary material, Table S3 for their characteristics). The remaining 3,139 people who completed at least one TP-CKD questionnaire were included in the analysis.” (Results pg 10, lines 217-221). Where we did not have this information because of the nature of a secondary data analysis, we further clarified this in the Methods section. For example, “Centres did not record information on whom they screened for eligibility, who had been confirmed eligible but declined participation, or people’s reasons for declining participation.” (Methods pg 6, lines 117-119).

• the external validity of our findings (STROBE item 21) (Discussion pg 22, lines 306-309): “It provides a robust external validation of clusters identified by previous studies, including skin (21, 24, 26, 27), gastrointestinal (20, 22), and mental health (24, 26), which suggests that our findings are generalisable beyond the study context.”

Lastly, we have stated at the start of the Methods section that we used STROBE to guide the reporting of our study (pg 5, lines 103-106), and included a completed STROBE checklist in the Supplementary material (Table S1) to indicate on which page each STROBE item has been included.

6. Amend sentence (lines 81-83) as there is at least one study which included patients not receiving KRT (ref 20 - Almutary et al included almost 25% non-KRT [both CKD G4 and G5]).

We have added the 2016 Almutary et al reference to the Introduction and changed the sentence as follows to clarify that most previous studies identified symptom clusters across rather than stratified for treatments/stages (page 5, lines 83-91):

“CKD symptom clusters have been mainly investigated in people receiving dialysis [refs]. Studies considering other kidney replacement therapies (KRTs) or CKD stages often identified clusters across rather than stratified by treatment or stage [refs incl 2016 Almutary]. … Only few studies looked at symptom clusters specifically for people with a kidney transplant or with CKD but not on KRT (CKD non-KRT)], despite these people reporting a similar symptom burden. This leaves it largely unknown to what extent current clusters generalise across or differ between treatment groups.”

7. Almutary et al 2017 also constructed structural equation modelling demonstrating the direct (casual) relationships between symptom clusters and health-related quality of life

We have added the 2017 Almutary et al study as a reference to the following sentence (pg 4, line 75-78): “In CKD, previous research proposed clusters related to fatigue, pain, gastrointestinal, and skin symptoms. They also reported negative associations between these clusters and patient outcomes linked to quality of life, such as health functioning [refs incl 2017 Almutary] and depression.”

8. Sentence (lines 85-86), the phrase “modest sample sizes” is vague – especially when cluster analysis is only possible when there is a sufficient sample size to do this type of statistical analysis.

We agree that ‘modest sample sizes’ is vague and have further specified this by changing the sentence as follows (pg 5, lines 85-87): “…; one reason for this lack of stratification might have been the modest sample sizes of <450 people.”

9. Arguably more studies on CKD symptom clusters because symptoms are subjective and could reflect particular cultures, health experiences, etc, etc. There is no need to avoid needing replication studies in another patient population to add to evidence.

To better explain the rationale for our study, we have amended the following paragraph in the Introduction as follows (page 5, lines 83-91):

“CKD symptom clusters have been mainly investigated in people receiving dialysis [refs]. Studies considering other kidney replacement therapies (KRTs) or CKD stages often identified clusters across rather than stratified by treatment or stage [refs]; one reason for this lack of stratification might have been the modest sample sizes of <450 people [refs]. Only few studies looked at symptom clusters specifically for people with a kidney transplant [refs] or with CKD but not on KRT (CKD non-KRT) [ref], despite them reporting a similar symptom burden [refs]. This leaves it largely unknown to what extent current clusters generalise across or differ between treatment groups.”

10. Clarify the aims related to longitudinal data collection.

Thank you for this helpful suggestion. We have clarified the aim of the longitudinal analyses in the the introduction as follows (pg 5, lines 92-95):

“This study, therefore, aimed to (1) explore symptom clustering in a large data set, stratified by CKD treatment group,(2) assess the relevance of the identified clusters by investigating their association with people’s ability to perform their usual activities, and (3) determine if these associations were stable over time.”

Methods

11. Ensure all STROBE aspects are explicitly described.

See our response to reviewer 2 point 5

12. For the TP-CKD program, at grade of CKD are people enrolled? If CKD G4 were included, then further clarity in data analysis and results will be required; that is, how were the different groups handled (i.e. CKD G4, CKD G5, KRT groups)

The data set did not include information on CKD stage at enrollment in the non-KRT group. We have clarified this in footnote a) to Table 1 as follows:

“[the CKD non-KRT treatment group] included people with any stage of CKD not on KRT who were under the treatment of a kidney centre. The dataset did not contain information on the stage of CKD at enrollment”

13. Clarify how PROM data was collected.

We have clarified this by adding/changing the following sentences to the Methods (pg 6, lines 116-117; pg 7, lines 143-144): “People who were interested received a paper copy of the questionnaire, which they returned to their unit upon completion. [..] The completed TP-CKD paper-based questionnaires were sent to the UK Renal Registry (UKRR) and scanned into electronic format.”

14. Clarify ethical procedures for TP-CKD.

We have clarified that no formal ethical approval was required for TP-CKD (pg 7-8, lines 151-160): “Because the primary purpose of TP-CKD was service evaluation and improvement rather than research, no formal ethical approval was required. People taking part in the TP-CKD programme implicitly consented to their data being processed and linked by returning a completed questionnaire to their kidney centre. The UKRR holds permissions under s251 of the National Health Service (NHS) Act 2006, to gather, process, and share confidential patient information for the purposes of audit and research. These permissions are renewed annually by the UK’s Health Research Authority’s Confidentiality Advisory Group. The collection and secondary analysis of the data for this study were approved by the UKRR (ref: UKRR ILD32) and carried out under the ethical permissions granted to the UKRR by the Research Ethics Committee.”

15. Data analysis – at what level was significance set?

Thank you for noting this. We have added the following sentence to the end of the Methods section (pg 10, lines 212-213): “For all analyses, we used SAS version 9.4 for all analyses and considered a p value of <0.5 significant.”

Discussion

16. First paragraph – there is no need to restate design, sample and sites for the study, instead interpret the meaning of the overall results (without restating actual numbers).

We have removed the first sentence stating the design, sample and sites, and in line with STROBE, focused the first paragraph on a summary of findings for each of the study objectives (pg 22, lines 299-303): “This study found that, overall, CKD symptom clusters related to lack of energy and mobility, gastrointestinal, skin, and mental health. Although clusters varied between treatment groups, the ‘lack of energy and mobility’ clusters were consistently associated with having problems doing usual activities across groups, both cross-sectionally and longitudinally.”

17. Add to several paragraphs in the discussion what explanation can be made for the differences found between this study and previous studies on symptom clusters (why isn’t there more consistency which is only partly explained by different symptom measures used). C. Miaskowski research in cancer symptom clusters and H.A. de Von in CVD symptom clusters is likely to be insightful for this manuscript.

Thank you for pointing us to the highly relevant paper by Miaskowski et al. We have referenced this work in the Discussion’s ‘Relation to other studies’ section as follows:

• (pg 22, lines 305-311) “We used the largest data set to date to investigate symptom clusters across treatment groups in people with CKD. It provides a robust external validation of clusters across treatment groups identified by most previous studies, including skin, gastrointestinal, and mental health […]. This alignment with previous research on CKD symptom clusters is in contrast to other disease areas, such as cancer, where there seems to be less consistency in symptom clusters across studies [ref to Miaskowski].”

• (pg 22-23, lines 318-322): “Furthermore, the ‘uraemic’ cluster in other studies included nausea and poor appetite (20, 21, 26, 27), whereas in our study, these symptoms, together with vomiting, formed a gastrointestinal cluster. This finding strongly aligns with wider symptom cluster research commonly identifying a separate gastrointestinal cluster consisting of nausea and vomiting [ref to Miaskowski].”

We also provided an explanation for the differences between our findings and those of some others with regards to mental health as a separate cluster (pg 23, lines 333-339):

“In the CKD non-KRT treatment group, we found that frequently co-occurring mental health symptoms, i.e. feeling anxious and feeling depressed, formed a stand-alone cluster, which differs from what Lee and Jeon reported in their smaller, Korean study. An explanation for mental health being a separate cluster may be that, as people approach kidney failure, they often experience more psychological symptoms. In the peritoneal dialysis group in the current study, this cluster also included pain and diarrhoea, both of which may negatively affect mental health.”

18. Lack of energy and its similarity (or difference) to fatigue also requires further interpretation.

Thank you for this constructive suggestion. We have added the following section to the Discussion to address this (pg 23, lines 325-332):

“The POS-S Renal also does not include fatigue, which is a common symptom in people with CKD that others found to pertain to several symptom clusters. Lack of energy and weakness represent some but not all aspects of fatigue, which is a multi-faceted symptom related to sociodemographic, psychological, clinical and biochemical factors. This implies that part of the people reporting lack of energy and weakness in our study might have actually suffered from fatigue, which may partly explain the strong associations we found between ‘lack of energy and weakness’ clusters and people’s ability to do usual activities.”

19. In the discussion on this study’s symptoms and activity, it would also be worth reviewing and discussing with regard to Almutary et al 2017 CKD symptom cluster modelling which demonstrated significant casual relationships with physical function (and also with mental health)

We agree this is a relevant study to reference in this context. We have added it to the following statement in the Discussion (pg 23, line 340-343):

“Our analyses showed that the ‘lack of energy and mobility’ clusters were associated with problems performing usual activities in all groups and models, which echoes the findings of previous research in which uraemic clusters were associated with impaired physical and mental functioning [refs incl Almutary et al 2017].”

Clinical Implications

20. Clarify why the IPOS-Renal is not recommend as this PROM is widely used (rather than POS-S renal).

The two main reasons for using the POS-S Renal instead of the IPOS-Renal are:

1. The POS-S Renal contains 17 common CKD symptoms to use for composing clusters, while the IPOS-Renal only includes a subset of these symptoms.

2. The IPOS-Renal includes additional items on concerns beyond symptoms, such as information needs and practical issues. These items would not have been included in the symptom cluster analysis.

We propose not to include this explanation in the manuscript because there is a risk it may cause confusion rather than provide clarity.

Future Research

21. C. Miaskowski and colleagues are exploring underlying mechanisms in cancer symptoms and symptom cluster. There is an opportunity for researchers in CKD to draw from this work and extend it.

Thank you for this helpful suggestion. We have added a more explicit reference to Miaskowski’s paper in the Discussion as follows (pg 27, lines 419-422):

“Some have suggested that certain symptoms may share an underlying mechanism or one symptom that ‘drives’ them. Future studies, building on and adding to the wider cluster research agenda [ref to Miaskowski], are warranted to enhance our understanding in this area, alongside qualitative studies to ascertain the clinical face validity of symptom clusters.”

Reviewer #3

1. Currently, the terminology of CKD has been revised, enabling researchers to compare results across studies conveniently. For example, end-stage kidney failure or end-stage kidney disease, now termed as kidney failure. Please refer. Levey et al. (2020). Therefore, it is recommend to reword CKD terminology throughout the manuscript to fit current guidelines.

We very much appreciate this advice and pointing us to the Levey et al 2020 nomenclature. We have made considerable changes throughout the text and believe it is now in line with this nomenclature. We have also stated at the start of the Methods that we followed Levey et al for reporting our study findings (pg 5, lines 103-106).

2. It is also suggested not to label people living with chronic diseases as patients (e.g., line 89: CKD patient groups) consistently throughout the manuscript.

We very much appreciate this advice and have addressed this throughout the manuscript as much as possible. For example, we changed ‘kidney patients’ to ‘people with CKD’, and we now state that we ‘categorised the data into treatment groups’, rather than ‘categorising participants into patient groups’.

3. Methods: Line 119: Spell out EQ5D5L first, before using it in abbreviated form.

Thank you for spotting this – we have now spelled it out on first mention as follows (pg 7, line 131-132):

“As the outcome measure we used an item from the EuroQol 5 dimensions - 5 level (EQ-5D-5L) version, which was also included in the TP-CKD questionnaire.”

4. Methods: Line 128: It is not clear that what do you mean by ‘see below’ as there is no further information given in relation to self-rated health.

We agree that this was not very clear in the text. We have removed the ‘see below’ reference and instead explained why the self-reported health status questions was not included in the analysis (pg 7, lines 136-141):

“We did not consider the remaining EQ-5D-5L items as part of our the outcome measure [..] because [… responses could not be scanned into electronic format (self-rated health status).”

5. Under discussion, relation to other studies, second paragraph; it is better to mention frequently co-occurring mental health symptoms. i.e., anxiety and depression

We have taken your suggestion on board and reworded this sentence as follows (pg 23, lines333-334):

“In the CKD non-KRT treatment group, we found that frequently co-occurring mental health symptoms, i.e. feeling anxious and feeling depressed, formed a stand-alone cluster, …”

6. According to current evidence, fatigue is the most frequent symptom experienced by those with CKD. See for example, Almutary et al. (2013) and (2016) in your reference list. However, POS-S-renal doesn’t capture fatigue. It is note that weakness/lack of energy is different to fatigue. Therefore, it is important to discuss this point in the discussion under the section ‘relation to other studies’.

Thank you for this constructive suggestion. We have added the following sentences to the Discussion Section ‘Relation to Other Studies’ to address this (pg 23, lines 325-332):

“The POS-S Renal also does not include fatigue, which is a common symptom in people with CKD that others found to pertain to several symptom clusters. Lack of energy and weakness represent some but not all aspects of fatigue, which is a multi-faceted symptom related to sociodemographic, psychological, clinical and biochemical factors. This implies that part of the people reporting lack of energy and weakness in our study might have actually suffered from fatigue, which may partly explain the strong associations we found between ‘lack of energy and weakness’ clusters and people’s ability to do usual activities.”

7. Use the RRT abbreviation consistently throughout the manuscript. Abbreviated form of renal replacement therapy is missing in the conclusions section.

Thank you for spotting this. We have changed renal replacement therapy to KRT in line with rest of the manuscript.

8. It is suggested to check the reference list to align journal requirements.

We have corrected instances where references were not in line with journal requirements.

Submitted filename: Revision.Response to reviewers.docx

Click here for additional data file.

14 Oct 2021

22 Oct 2021

Point-by-point response to reviewers’ comments

Manuscript: Symptom clusters in chronic kidney disease and their association with patients’ ability to perform usual activities (PONE-D-21-05457)

Page and line numbers refer to those in the ‘clean’ version of the revised manuscript.

Reviewer #1:

Dear author, I recommend reviewing the abstract needs to include that I work with adults.

Thank you for your suggestion. We have made it clear in the abstract (Line 29) that the data presented in the manuscript is with adults. In the Methods (Line 118), it is stated that people were eligible if they were ‘18 years or over’.

For methods you need We have clarified that no formal ethical approval was required for TP-CKD and the follow-up period.

In our previous Response to Reviewers, we revised the ethics section of the Methods to directly address this (see below - Lines 151-160). No formal ethical approval is required in the UK for service evaluation and improvement, regardless if it carried out in a cross-sectionally or longitudinally (like the TP-CKD).

We have revised Line 154 to indicate that no formal ethical approval was required if people completed more than one questionnaire:

“Because the primary purpose of TP-CKD was service evaluation and improvement rather than research, no formal ethical approval was required. People taking part in the TP-CKD programme implicitly consented to their data being processed and linked by returning a completed questionnaire(s) to their kidney centre.” Lines 151-154

Do you need to include more information about the inclusion and exclusion criteria (age, therapy, literate, other condition).

The full inclusion and exclusion criteria are listed in Lines 112-113: People were eligible if they were aged 18 years or over and receiving care for any stage of CKD or on any form of KRT.

To confirm, there were no other inclusion and exclusion criteria for TP-CKD.

Describe in the methodology how the questionnaires were applied.

Key methods for the questionnaires are described in lines 100-141 of the manuscript:

We conducted a secondary analyses of cross-sectional and longitudinal data collected in the context of a national service improvement programme in 14 kidney centres across England (UK) called Transforming Participation in Chronic Kidney Disease (TP-CKD) [31]. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [32] (see Supplementary material, Table S1) and the nomenclature for kidney function and disease proposed by Levey and colleagues [33] for reporting our findings.

Transforming Participation in Chronic Kidney Disease (TP-CKD)

The TP-CKD programme aimed to support people with CKD to better manage and make decisions about their own care and treatment. This included introducing collection of patient-reported outcome data in 14 English kidney centres [34]. Between December 2015 and December 2017, members of local kidney care teams approached eligible patients in their centre. People were eligible if they were aged 18 years or over and receiving care for any stage of CKD or on any form of KRT. The number and sociodemographic characteristics of people on KRT treated in each centre are provided in Supplementary material, Table S2; this does not include people with CKD not on KRT because this data was not available. People who were interested received a paper copy of the questionnaire, which they returned to their unit upon completion. Centres did not record information on whom they screened for eligibility, who had been confirmed eligible but declined participation, or on reasons for declining participation. In twelve centres, people who had previously taken part were invited to complete the questionnaire again at a later time.

Measures of exposure and outcome

In the current secondary analysis, we used symptom burden as the measure of exposure. To assess exposure, we analysed data collected through the Palliative care Outcome Scale-Symptom (POS-S) Renal [35], which was one of the patient-reported outcome measures included in the paper-based TP-CKD questionnaire. The POS-S Renal consists of 17 symptoms that are common for people with CKD [35], such as pain, weakness or lack of energy, and itching. For each symptom, respondents are asked to indicate to what extent they have been bothered by it over the last week on a scale from 0 (not at all) to 4 (overwhelmingly).

As the outcome measure we used an item from the EuroQol 5 dimensions - 5 level (EQ-5D-5L) version [36], which was also included in the TP-CKD questionnaire. The EQ-5D-5L assesses general health status and asks respondents to rate whether they had problems doing their usual activities (e.g. work, study, housework, leisure activities) on a scale from 1 (no problem) to 5 (unable). People with CKD previously reported this item as being important to ‘living well’ (29). We did not consider the remaining EQ-5D-5L items as part of the outcome measure because: some fully overlapped with symptoms in the POS-S Renal (pain/discomfort; anxiety/depression; mobility); one was not reported by people with CKD as a priority outcome in research (self-care: wash and dress) [37,38]; and one because its responses could not be scanned into electronic format (self-rated health status).

13. Clarify how PROM data was collected, this topic needs to be clear.

this topic is still unclear:

PROM data were collected as part of the TP-CKD survey which is described in lines 100-141 of the manuscript:

We conducted a secondary analyses of cross-sectional and longitudinal data collected in the context of a national service improvement programme in 14 kidney centres across England (UK) called Transforming Participation in Chronic Kidney Disease (TP-CKD) [31]. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [32] (see Supplementary material, Table S1) and the nomenclature for kidney function and disease proposed by Levey and colleagues [33] for reporting our findings.

Transforming Participation in Chronic Kidney Disease (TP-CKD)

The TP-CKD programme aimed to support people with CKD to better manage and make decisions about their own care and treatment. This included introducing collection of patient-reported outcome data in 14 English kidney centres [34]. Between December 2015 and December 2017, members of local kidney care teams approached eligible patients in their centre. People were eligible if they were aged 18 years or over and receiving care for any stage of CKD or on any form of KRT. The number and sociodemographic characteristics of people on KRT treated in each centre are provided in Supplementary material, Table S2; this does not include people with CKD not on KRT because this data was not available. People who were interested received a paper copy of the questionnaire, which they returned to their unit upon completion. Centres did not record information on whom they screened for eligibility, who had been confirmed eligible but declined participation, or on reasons for declining participation. In twelve centres, people who had previously taken part were invited to complete the questionnaire again at a later time.

Measures of exposure and outcome

In the current secondary analysis, we used symptom burden as the measure of exposure. To assess exposure, we analysed data collected through the Palliative care Outcome Scale-Symptom (POS-S) Renal [35], which was one of the patient-reported outcome measures included in the paper-based TP-CKD questionnaire. The POS-S Renal consists of 17 symptoms that are common for people with CKD [35], such as pain, weakness or lack of energy, and itching. For each symptom, respondents are asked to indicate to what extent they have been bothered by it over the last week on a scale from 0 (not at all) to 4 (overwhelmingly).

As the outcome measure we used an item from the EuroQol 5 dimensions - 5 level (EQ-5D-5L) version [36], which was also included in the TP-CKD questionnaire. The EQ-5D-5L assesses general health status and asks respondents to rate whether they had problems doing their usual activities (e.g. work, study, housework, leisure activities) on a scale from 1 (no problem) to 5 (unable). People with CKD previously reported this item as being important to ‘living well’ (29). We did not consider the remaining EQ-5D-5L items as part of the outcome measure because: some fully overlapped with symptoms in the POS-S Renal (pain/discomfort; anxiety/depression; mobility); one was not reported by people with CKD as a priority outcome in research (self-care: wash and dress) [37,38]; and one because its responses could not be scanned into electronic format (self-rated health status).

10. Clarify the aims related to longitudinal data collection.

We addressed this by explicitly stating the aims of the longitudinal analysis in lines 92-97 of the Introduction:

“This study, therefore, aimed to (1) explore symptom clustering in a large data set,

stratified by CKD treatment group,(2) assess the relevance of the identified clusters by

investigating their association with people’s ability to perform their usual activities, and

(3) determine if these associations were stable over time.” Lines 92-97

In the discussion, check whether the treatment time also interferes with symptoms.

In this study we aimed to explore symptom clusters and the association between these clusters and the ability to perform usual activities. We described time on KRT for the participants, and adjusted for time on KRT when exploring the association between symptoms and ability to perform usual activities. The effect of time on KRT on symptom burden was not an aim of this work so we have not included it in the discussion.

Reviewer #2:

No further comments to address.

Thank you, Dr. Bonner, for your constructive feedback on the manuscript.

Reviewer #3:

No further comments to address.

We appreciate your constructive feedback and kind words on the manuscript, Dr. Yapa.

Submitted filename: Response to reviewers.22 10 21.docx

Click here for additional data file.

22 Nov 2021

24 Dec 2021

Point-by-point response to reviewers’ comments

Manuscript: Symptom clusters in chronic kidney disease and their association with patients’ ability to perform usual activities (PONE-D-21-05457R2)

Note: Page and line numbers refer to those in the ‘clean’ version of the manuscript.

Reviewer #1

1. This manuscript is a cross-sectional analysis of patients undergoing renal treatment. In the Transforming Participation in Chronic Kidney Disease (TP-CKD) session, better describe the program, how were the patients asked to fill out and is there an interval for re-application of the questionnaires? Such missing information compromises the analysis of table 3.

In response to reviewer comments on previous versions of the manuscript, we endeavoured to provide the necessary background information on the TP-CKD programme in a dedicated section in the Methods (Ln 107-121):

“The TP-CKD programme aimed to support people with CKD to better manage and make decisions about their own care and treatment. This included introducing collection of patient-reported outcome data in 14 English kidney centres [34]. Between December 2015 and December 2017, members of local kidney care teams approached eligible patients in their centre. People were eligible if they were aged 18 years or over and receiving care for any stage of CKD or on any form of KRT. The number and sociodemographic characteristics of people on KRT treated in each centre are provided in Supplementary material, Table S2; this does not include people with CKD not on KRT because this data was not available. People who were interested received a paper copy of the questionnaire, which they returned to their unit upon completion. Centres did not record information on whom they screened for eligibility, who had been confirmed eligible but declined participation, or on reasons for declining participation. In twelve centres, people who had previously taken part were invited to complete the questionnaire again at a later time.”

2. Another important data for table 3 is to include the time between analyses. Several studies point to the impact of treatment time and outcomes of interest.

Table 3 (p. 20) presents the results of the cross-sectional analysis of the data, and as such, there is no time element for this part of our analyses. For the longitudinal analysis (presented in Table 4, p. 21), we reported the time between measurements in the text (Ln 234-5): “A subsample of 699 people completed follow-up questionnaires at a median of 203 days (IQR, 133 to 301) after baseline, while remaining in the same treatment group.”

3. Data on frequency, initiation of this analysis and inclusion of this questionnaire in renal care services are not described. Is there a protocol for applying the questionnaire annually?

In the Methods section (Ln 107-121; also see response to point 1), we provided details of the TP-CKD programme with regard to when the data was collected, in whom and how. As described in the Methods, the programme ran between December 2015 and December 2017 (Ln 110-111). We this we implied there is no protocol for applying the questionnaire annually.

4. lines 117 to 121: “Interested people received a printed copy of the questionnaire, which they returned to the unit after completing it. Sites did not record information about who they screened for eligibility, who was confirmed eligible but refused participation, or about the reasons for denial. In twelve centers, people who had already participated were invited to answer the questionnaire later.” It is important to inform the interval between responses for the analysis of table 3. The symptoms improved, which may be a reflection of the treatment and others not, such finding may be due to the treatment process itself. Such observations should be placed as limitations, as well as the non-homogeneous adherence of the centers.

Table 3 (p. 20) presents the results of the cross-sectional analysis of the data, and as such, there is no time element for this part of our analyses. For the longitudinal analysis (presented in Table 4, p. 21), we reported the time between measurements in the text (Ln 234-5): “A subsample of 699 people completed follow-up questionnaires at a median of 203 days (IQR, 133 to 301) after baseline, while remaining in the same treatment group.”

In relation to the symptoms having improved, we would like to clarify that the aim of our study was to identify symptom clusters and their relationship to people’s performing daily activities, both cross-sectionally and longitudinally. The aim was not to describe symptom trajectories over time and as such, we do not report whether symptoms improved in the main text of the manuscript. As background information to the longitudinal analysis, we did include information in the supplement (Table S7) on how symptoms and people’s ability to perform daily activities changed over time. However, this information suggests that symptoms only improved over time for approximately half of patients, while they deteriorated for the other half. We, therefore, do not think these observations indicate clear limitations or ‘non-homogeneous adherence of centres’.

5. In the discussion, the main symptom identified 'lack of energy and mobility' may be related to the treatment time and its impact, which must be considered, long therapies, displacements...

To address the impact of treatment time, we have included ‘time on kidney replacement therapy’ in the partially and fully adjusted models of all analyses (see Table 3 (p. 20) and Table 4 (p. 21) to account for the influence this factor can have on symptoms.

Submitted filename: Response to reviewers 13 12 21.docx

Click here for additional data file.

9 Feb 2022

Symptom clusters in chronic kidney disease and their association with people’s ability to perform usual activities

PONE-D-21-05457R3

Dear Dr. Moore,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Boris Bikbov

Academic Editor

PLOS ONE

18 Feb 2022

PONE-D-21-05457R3

Symptom clusters in chronic kidney disease and their association with people’s ability to perform usual activities

Dear Dr. Moore:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Boris Bikbov

Academic Editor

PLOS ONE