Lifestyle behaviour change for preventing the progression of chronic kidney disease: a systematic review

Abstract

Objectives

Modifying lifestyle can prevent the progression of chronic kidney disease (CKD) but the specific elements which lead to favourable behaviour change are not well understood. We aimed to identify and evaluate behaviour change techniques and functions in lifestyle interventions for preventing the progression of CKD.

Design

Systematic review.

Data sources

MEDLINE, EMBASE, CINAHL and PsycINFO.

Eligibility criteria

Trials of lifestyle behaviour change interventions (including diet, physical activity, smoking and/or alcohol) published to September 2018 in adults with CKD stages 1–5.

Data extraction and synthesis

Trial characteristics including population, sample size, study setting, intervention, comparator, outcomes and study duration, were extracted. Study quality was independently assessed by two reviewers using the Cochrane risk of bias tool. The Behaviour Change Technique Taxonomy v1 was used to identify behaviour change techniques (eg, goal setting) and the Health Behaviour Change Wheel was used to identify intervention functions (eg, education). Both were independently assessed by three reviewers.

Results

In total, 26 studies involving 4263 participants were included. Risk of bias was high or unclear in most studies. Interventions involved diet (11), physical activity (8) or general lifestyle (7). Education was the most frequently used function (21 interventions), followed by enablement (18), training (12), persuasion (4), environmental restructuring (4), modelling (2) and incentivisation (2). The most common behaviour change techniques were behavioural instruction (23 interventions), social support (16), behavioural demonstration (13), feedback on behaviour (12) and behavioural practice/rehearsal (12). Eighteen studies (69%) showed a significant improvement in at least one primary outcome, all of which included education, persuasion, modelling and incentivisation.

Conclusion

Lifestyle behaviour change interventions for CKD patients frequently used education, goal setting, feedback, monitoring and social support. The most promising interventions included education and used a variety of intervention functions (persuasion, modelling and incentivisation).

PROSPERO registration number

CRD42019106053.

Strengths and limitations of this study.

• We used comprehensive, evidence-based frameworks to identify and describe behaviour change techniques and intervention functions in lifestyle behavioural interventions for patients with chronic kidney disease.

• Coding of behaviour change techniques and intervention functions was systematically and independently conducted by three researchers, and risk of bias was assessed.

• Summary estimates could not be ascertained due to the heterogeneity of interventions and outcome measures.

Introduction

Preventing the progression of chronic kidney disease (CKD) is a high priority for patients and clinicians, to reduce the requirement for dialysis.^1–3 Lifestyle interventions which modify behavioural risk factors such as poor diet and low physical activity can prevent progression of CKD and life-threatening complications and improve quality of life and survival.^4–6 Addressing behaviour change is particularly relevant in CKD as lifestyle modification can be challenging. Poor adherence to diet, medication and other treatments is common in CKD.⁷ Barriers to modifying lifestyle include low health literacy, conflicts with cultural norms, complicated nutritional requirements and safety concerns.^7–11

Guidelines recommend the explicit use of behaviour change for addressing lifestyle risk factors when designing and reporting interventions for patients with CKD.^{12 13} However, it is uncertain which aspects of lifestyle behaviour change interventions are the most effective, and reporting of behavioural components is often unclear, making implementation in practice problematic.

The Behaviour Change Technique Taxonomy v1 was developed to provide a comprehensive framework that integrates behaviour change techniques used in interventions.¹⁴ The Taxonomy was further synthesised into a framework, the Health Behaviour Change Wheel which describes the intervention functions necessary to change health behaviours.¹⁵ The Health Behaviour Change Wheel provides a broad, overarching framework in which to characterise behaviour change interventions while the Taxonomy identifies specific techniques related to individual behaviours. The intervention functions described in the Health Behaviour Change Wheel can be delivered by a variety of behaviour change techniques. For example, the intervention function, ‘education’, outlined in the Wheel, can include the behaviour change techniques ‘instruction on how to perform the behaviour’ and ‘information about antecedents’, detailed in the Taxonomy. Similarly, the intervention function ‘incentivisation’ can incorporate techniques such as ‘feedback on behaviour’ and ‘rewards’.

Behaviour change interventions using the Wheel and the Taxonomy can effectively change lifestyle behaviours. For example, a text-messaging and pedometer programme improved physical activity in people at high risk of type 2 diabetes,¹⁶ a digital healthy eating programme increased consumption of fruit and vegetables and sustained this over a 6-month period¹⁷ and a digital behaviour change programme achieved significant weight loss results in individuals at risk of type 2 diabetes.¹⁸ The Taxonomy and the Wheel are recommended approaches to modify lifestyle risk factors for chronic disease prevention.^{12 16 18} However, these frameworks have not been used in designing and reporting behaviour change strategies in lifestyle interventions for patients with CKD.

We aimed to identify and evaluate behaviour change techniques and intervention functions used in lifestyle interventions for preventing the progression of CKD. This may inform the development of effective and replicable behaviour change interventions for the prevention of CKD, leading to improvements in patient outcomes.

Methods

We used the Preferred Reporting Items for Systematic Reviews and Meta-analyses Statement¹⁹ and checklist to report this systematic review (online supplementary file S1).

Selection criteria

We included randomised trials of lifestyle behaviour change interventions (including, but not restricted to diet, physical activity, smoking and alcohol consumption) in adult patients (aged over 18 years) with CKD stages 1–5 and not requiring renal replacement therapy. We did not apply restrictions based on outcomes or language. Studies including a combination of pharmacological therapy and lifestyle were included but trials involving only pharmacological therapies were excluded.

Literature search

A comprehensive search was conducted in MEDLINE (1946 to 20 September 2018), EMBASE (1996 to 20 September 2018), CINAHL (1982 to 20 September 2018) and PsycINFO (1806 to 20 September 2018) using Medical Subject Heading (MeSH) terms relating to CKD, and lifestyle behaviour change interventions (online supplementary file S2), and reference lists of relevant articles and reviews. Author NE screened the studies by title and abstract and assessed full-text articles for eligibility. Those that did not meet the inclusion criteria were excluded.

Data extraction and critical appraisal

The trial characteristics relevant to the population, sample size and study setting as well as intervention (type, mode of delivery, use of theory, intervention functions (as described in the Health Behaviour Change Wheel¹⁵ and behaviour change techniques (as described in the Behaviour Change Technique Taxonomy v1¹⁴)), comparator, outcomes and study duration, were extracted and tabulated. We assessed the risk of bias using the Cochrane tool for randomised studies.²⁰ NE and KM assessed the risk of bias in each study independently and any differences were resolved by discussion.

We contacted the authors of the studies when it was necessary to gather additional information. Supplemental data was available in 12 of the 26 studies. In six studies with no supplemental data, sufficient information was available in the published article. Therefore, we contacted eight authors to request further information and received responses from two authors.

Analysis of intervention functions and behaviour change techniques

The Behaviour Change Technique Taxonomy v1 (the ‘Taxonomy’) and Health Behaviour Change Wheel (the ‘Wheel’) are comprehensive tools for identifying behavioural components in interventions and how frequently they occur.^{14 15} The two frameworks are complementary and in addition to designing interventions, they have been used as a method for identifying behavioural components in public health interventions and clinical trials.²¹ The tools have been used in previous systematic reviews to identify behaviour change techniques and functions in health interventions.^22–28

Behaviour change techniques

The Behaviour Change Technique Taxonomy consists of 93 behaviour change techniques, such as goal-setting, self-monitoring, social support and re-structuring the physical environment (see online supplementary table S1 for the full taxonomy). The techniques are grouped into 16 domains: goals and planning, feedback and monitoring, social support, shaping knowledge, natural consequences, comparison of behaviour, associations, repetition and substitution, comparison of outcomes, reward and threat, regulation, antecedents, identity, scheduled consequences, self-belief and covert learning.

Intervention functions

There are nine intervention functions in the Wheel: education, persuasion, incentivisation, coercion, training, enablement, modelling, environmental restructuring and restrictions.¹⁵ These are activities designed to change behaviours and include one or more behaviour change techniques. Definitions of each intervention function have been described by Michie et al and were used to inform decisions about what functions were present in each study.¹⁵

Authors NE and KM completed online training for interpreting the Wheel and the Taxonomy to ensure consistency and reliability of coding.²⁹ N.E, KM and VS independently read intervention descriptions line-by-line to locate text matching a definition of an intervention function¹⁵ and the description of behaviour change techniques from the BCTTv1 coding frame (online supplementary table S1). Each of the 93 behaviour change techniques were indicated as either present or absent in a standardised data extraction form. A behaviour change technique had to be explicitly described to be coded and included in the analysis. The authors compared the codes and discussed discrepancies to reach consensus.

Patient and public involvement

No patient involved.

Results

Literature search and study characteristics

The literature search yielded 10 043 citations from which 26 studies (n=4263 participants) were eligible and included in the review (figure 1). Study characteristics are shown in table 1. The studies were conducted in 15 countries.

Figure 1.

PRISMA flowchart of included/excluded studies. *A behavioural intervention explicitly describes a behaviour change technique which can be coded using the Behavior Change Technique Taxonomy v1. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-analyses.

Table 1.

Characteristics of included studies

Study	N	CKD Stage	Age (years)	Country	Intervention	Comparator	Primary Outcomes	Study duration (months)
Dietary interventions
Campbell et al 38	56	CKD4–5	>18	Australia	Individualised nutritional counselling and regular follow-up	Usual care	Body composition	3
Clark et al 47	590	CKD3	18–80	Canada	Coaching to increase water intake (drinking containers and water vouchers also provided)	Coaching to maintain usual fluid intake	Change in eGFR	12
De Brito-Ashurst et al 37	56	eGFR <60 mL and BP >130/80 or taking BP medication; Bangladeshi origin	18–74	United Kingdom	Community cooking education sessions facilitated by Bengali workers	Usual care	Reduction in systolic/diastolic BP	6
Dussol et al 61	63	Type I/II diabetic nephropathy, eGFR60-100 mL	40–72	France	Low-protein diet with telephone calls every 6 weeks to help change dietary habits	Usual-protein diet	Decline GFR and 24-hour albumin excretion rate	24
MDRD Study (1995)*	840	eGFR 13–55 mL	18–70	United States	Low protein diet with dietician support	Moderate, low and very low protein diets compared	Decline eGFR, dietary satisfaction	45
Mekki et al 62	40	eGFR 60–90 mL	47–75	Algeria	Nutritional advice based on Mediterranean diet	Usual care	Dyslipidaemia	3
Meuleman et al 32	138	eGFR ≥20 mL	≥18	The Netherlands	Sodium restricted diet with self-management, education, motivational interviewing and self-monitoring	Usual care	Sodium excretion & BP	3
Paes-Barreto et al 46	89	CKD3–5	≥18	Brazil	Intense counselling/education on low protein diet	Standard counselling	Change in protein intake	4
Pisani et al 42	57	CKD3b–5	>18	Italy	Low protein, phosphate and sodium diet, ‘6-tips diet’ checklist	Non-individualised, moderately low protein diet	Protein intake, metabolic parameters and adherence	6
Rosman e t al 63	247	CrCl 10–60 mL/min	15–73	The Netherlands	Dietary protein restriction and dietician visits every 3 months	Usual care	Adherence	24
Saran e t al 64	58	CKD3–4	>18	United States	Dietary sodium restriction (<2 g sodium per day)	Usual diet	Change in hydration status	1
Physical activity interventions
Aoike et al 59	29	CKD3–4	18–70	Brazil	Home-based moderate-intensity aerobic exercise programme	Usual care	Cardiopulmonary/functional, BP, CrCl, eGFR	3
Barcellos et al 65	150	CKD2–4	>18	Brazil	Aerobic and resistance training	Usual care	Change in eGFR	4
Greenwood et al 43	20	CKD3–4	18–80	United Kingdom	Resistance and aerobic training (3 days per week)	Usual care	Change in eGFR	12
Kao et al 30	94	eGFR ≥15 mL	≥39	Taiwan	Group education lecture; individual exercise programme Trans-Theoretical Model	Not specified	Exercise behaviour, depression, fatigue	3
Leehey et al 66	32	CKD2–4	49–81	United States	Aerobic & resistance training, home exercise (plus dietary management)	Dietary management	Urine protein to creatinine ratio	12
Rossi et al 45	107	CKD3–4	≥18	United States	Guided exercise twice a week plus usual care	Usual care	Physical function, quality of life	3
Tang et al 49	90	CKD1–3	18–70	China	Individualised exercise programme (education and home-based aerobic exercise)	Usual care	Physical function, self-efficacy, anxiety, depression, quality of life	3
Van Craenenbroeck et al 34	40	CKD3–4	≥18	Belgium	Home-based aerobic training programme (four daily cycling sessions, 10 min each)	Usual care	Peripheral endothelial function	3
Lifestyle interventions
Flesher et al 39	40	CKD3–4	18–80	Canada	Individual dietary counselling, group nutrition and cooking classes, exercise programme	Usual care	Composite eGFR, TC, urinary sodium, urinary protein and BP	12
Howden et al 40	83	CKD3–4	18–75	Australia	Multi-disciplinary care, lifestyle and aerobic/resistance training	Usual care	Change in CRF	12
Ishani et al 41	601	eGFR <60	>18	USA	Care by a multi-disciplinary team using a telehealth device	Usual care	Composite death, hospitalisation, emergency visits and admission to a nursing facility	20
Jiamjariyapon et al 67	442	CKD3–4	18–70	Thailand	Integrated care by multi-disciplinary team and community care workers. Group counselling, home visits	Usual care	Change in eGFR	24
Joboshi and Oka44	65	Overt proteinuria and clinically diagnosed CKD	38–86	Japan	Self-management programme	Standard education	Self-efficacy and self-management behaviour	3
Patil et al 68	76	Diabetic nephropathy	30–70	India	Low-calorie diet, physical activity and behaviour	ACE inhibitor therapy	24-hour urine protein BMI	6
Teng et al 31	160	eGFR ≥30 mL/min/1.73 m2	≥20	Taiwan	Lifestyle modification programme based on Trans-Theoretical Model	Standard education	Health behaviours, knowledge, physical function	12

*MDRD study described in two main articles: Gillis et al ³³ and Coyne et al ⁴⁸.

BMI, Body Mass Index; BP, blood pressure;CKD, chronic kidney disease; CrCl, creatinine clearance; CRF, cardiorespiratory fitness; eGFR, estimated glomerular filtration rate; MDRD, Modification of Diet in Renal Disease study; TC, total cholesterol.

Risk of bias assessment

Overall, the reporting of studies was relatively incomplete, particularly for the blinding of participants and personnel which was missing or unclear in every study (figure 2). Allocation concealment was unclear or at high risk of bias in 20 (77%) studies. Blinding of outcome assessment was also poorly reported with 19 studies showing high or unclear risk of bias for this domain. Domains that performed better were selective reporting with low risk of bias in 21 studies, random sequence generation with low risk of bias in 17 studies and incomplete outcome data showing low risk of bias in 13 studies.

Figure 2.

Risk of bias for individual studies (n=26). MDRD, Modification of Diet in Renal Disease study.

Characteristics of the interventions

Across the interventions assessed in the 26 studies included, 11 were dietary interventions, 8 involved physical activity and 7 used any combination of diet, physical activity, weight reduction and/or smoking cessation (lifestyle).

Five studies were informed by theory, three used the Trans-Theoretical Model,^{30 31} one used self-regulation theory³² and another was informed by contemporary behavioural theory, in particular the self-management approach.³³ Two studies used Motivational Interviewing,^{34 35} a counselling approach which involves behaviour change strategies.³⁶

Only three studies included family members, friends or partners in the intervention to facilitate participant’s behaviour change (online supplementary table S2).^{31 37}

Behaviour change techniques

Table 2 outlines the number of behaviour change techniques present in each lifestyle behaviour change intervention. The number of behaviour change techniques used across interventions ranged from two to 20.

Table 2.

Cross matrix of behaviour change techniques and lifestyle behaviour change trials

	Meuleman et al 32	MDRD Study (1995)*	De Brito-Ashurst et al 37	Paes-Barreto et al 46	Campbell et al 38	Rosman et al 63	Dussol et al 61	Pisani et al 42	Saran et al 64	Clark et al 47	Mekki et al 62	Tang et al 49	Kao et al 30	Greenwood et al 43	Rossi et al 45	Aoike et al 59	Barcellos e t al 65	Van Craenenbroeck et al 34	Leehey et al 66	Howden et al 40	Ishani et al 41	Joboshi and Oka44	Teng et al 31	Flesher et al 39	Jiamjariyaponet al 67	Patil et al 68
	Diet											Physical Activity								Lifestyle
1.Goals and planning
1.1. Goal setting (behaviour)
1.2. Problem solving
1.3. Goal setting (outcome)
1.4. Action planning
1.5. Review behaviour goal(s)
1.7. Review outcome goal(s)
1.8. Behavioural contract
1.9. Commitment
2.Feedback and monitoring
2.1. Monitoring of behaviour by others without feedback
2.2. Feedback on behaviour
2.3. Self-monitoring of behaviour
2.4. Self-monitoring of outcome(s) of behaviour
2.6. Biofeedback
2.7. Feedback on outcome(s) of behaviour
3. Social support
3.1. Social support (unspecified)
3.2. Social support (practical)
3.3. Social support (emotional)
4. Shaping knowledge
4.1. Instruction on behaviour
4.4. Behavioural experiments
5.Natural consequences
5.1. Information about health consequences
5.2. Salience of consequences
5.4. Monitoring of emotional consequences
6.Comparison of behaviour
6.1. Demonstration of the behaviour
6.2. Social comparison
7.Associations
7.1. Prompts/cues
8.Repetition and substitution
8.1. Behavioural practice/rehearsal
8.2. Behaviour substitution
8.4. Habit reversal
8.6. Generalisation of target behaviour
8.7. Graded tasks
9.Comparison of outcomes
9.2. Pros and cons
10.Reward and threat
10.3. Non-specific reward
10.4. Social reward
10.10. Reward (outcome)
11.Regulation
11.2. Reduce negative emotions
11.3. Conserving mental resources
12.Antecedents
12.5. Adding objects to the environment
15.Self-belief
15.1. Verbal persuasion capability
15.3. Focus on past success
Number of BCTs	20	18	12	9	7	6	4	4	4	2	2	14	11	9	7	6	6	4	2	9	7	7	7	6	4	4

*MDRD study described in two main articles: Gillis et al ³³ and Coyne et al ⁴⁸.

BCT, Behaviour Change Technique.

The top five most frequently observed behaviour change techniques were instruction on how to perform the behaviour (23 interventions, 88%), social support (16, 62%), demonstration of the behaviour (13, 50%), feedback on behaviour (12, 46%) and behavioural practice/rehearsal (12, 46%). Of the 93 possible behaviour change techniques that could have been used, 12 techniques were used in more than 20% of trials, 27 were used at least once and 54 were never used. The mean number of behaviour change techniques was 5, the median was four and the range 2–20.

The two studies with the highest number of behaviour change techniques (20 and 18 in each study) were both informed by theory, with a particular focus on self-regulation and self-management.^{32 33}

Intervention functions

Table 3 lists the intervention functions present in each study (education, enablement, training, persuasion, modelling, incentivisation, environmental restructuring, coercion and restrictions). The number of functions used across interventions ranged from one to seven.

Table 3.

Cross matrix of intervention functions and lifestyle behaviour change trials

		Intervention functions
Studies	Type of intervention	Education	Enablement	Training	Persuasion	Environmental restructuring	Modelling	Incentivisation
Campbell et al 38	Diet
Clark et al 47
De Brito-Ashurst et al 37
Dussol et al 61
MDRD Study (1995)*
Mekki et al 62
Meuleman et al 32
Paes-Barreto et al 46
Pisani et al 42
Rosman et al 63
Saran et al 64
Aoike et al 59	Physical Activity
Barcellos et al 65
Greenwood et al 43
Kao et al 30
Leehey et al 66
Rossi et al 45
Tang et al 49
Van Craenenbroeck et al 34
Flesher et al 39	Lifestyle
Howden et al 40
Ishani et al 41
Jiamjariyapon et al 67
Joboshi44
Patil et al 68
Teng et al 31
Total		21	18	12	4	4	2	2

*MDRD study described in two main articles: Gillis et al ³³ and Coyne et al.⁴⁸

Education

Education was used most frequently as an intervention function, present in 21 (81%) interventions (table 3). Examples of educational strategies were: nutritional label reading,^{38 39} a resistance training booklet for home-based exercise,⁴⁰ a lecture/workshop about exercise recommendations with demonstrations,³⁰ online education modules on lifestyle modification⁴¹ and a written ‘six-tip diet’ checklist.⁴²

Enablement

Eighteen (69%) interventions used enablement. Examples include Motivational Interviewing to improve self-management of diet, lifestyle and physical activity,^{32 43} supportive telephone calls matching stages of behaviour change,³⁰ self-management techniques to foster self-efficacy^{38 39 44} and arranging support from friends and family members and ‘buddy’ visits.^{31 33} Four interventions were specifically designed using a self-management approach and assessed self-efficacy as an outcome.32 33 39 44

Training

Twelve (46%) interventions included training as an intervention function. Training was used in every intervention targeting physical activity but only used in two dietary interventions and two lifestyle interventions. Examples of training include home-based exercise training, guided exercise training in a gym,⁴⁰ physical therapy or cardiac rehabilitation facility⁴⁵ or hospital³⁴ and interactive cooking classes.³⁹

Persuasion

Four (15%) interventions used persuasion as an intervention function. A dietary intervention aimed to persuade participants about dietary salt intake by displaying test tubes of salt content alongside a range of high-salt food items.⁴⁶ In another dietary intervention, positive thinking was applied to participant’s goals and dieticians praised progress and focused on positive results.³³ Similarly, a lifestyle intervention used positive reinforcement to increase confidence and celebrate successes related to behaviour change and also discussed lack of exercise, poor dietary habits, risks of not exercising and associated consequences.³¹ Only one physical activity intervention used persuasion in designing and displaying printed health messages to promote exercise.³⁰

Environmental re-structuring

Four (15%) interventions used environmental restructuring. Two involved placing exercise equipment in the home environment (exercise bicycle, Theraband, weights and Swiss ball)^{40 43} and two included adding food products and equipment into the home environment (low sodium/protein meals and water bottles).^{33 47}

Modelling

Two (8%) dietary interventions incorporated modelling as an intervention function. Educators used food models and household measuring utensils to model appropriate food portion sizes⁴⁶ and food tastings provided an example of low protein meals.³³

Incentivisation

Two (8%) studies used incentivisation, one in the form of ‘appreciation gifts’ including certificates and mugs³³ and another included ‘self-rewards’ chosen by participants.³²

Coercion and restrictions

These functions were not used in any of the interventions.

Outcomes

A description of primary outcomes and results reported in studies is included in table 4. Primary outcomes of studies in this review were diverse and were mainly physiological metrics (for example, eGFR, blood pressure, peak VO₂ and sodium or albumin excretion). Only six studies included patient-reported and/or behavioural primary outcomes such as quality of life, fatigue, knowledge, self-efficacy, self-management, exercise and health behaviours.30 31 44 45 48 49

Table 4.

Effects of the behaviour change interventions on the primary outcome(s)

Study	Primary outcome/s	Measures	Intervention (n)	Control (n)	Intervention*	Control*	Mean difference (95% CI)	P value
Dietary interventions
Campbell et al 38	Body composition	Body cell mass, %	29	27	2.0 (1.9 to 5.9)†	1.5 (5.5 to 2.5)†	3.5 (2.1 to 9.1)	0.2
		Body cell mass, kg			0.5 (1.8 to 0.8)†	0.5 (0.7 to 1.8)†	1.1 (0.7 to 2.9)	0.2
Clark et al 47	Change in eGFR	Change eGFR, mL/min/1.73 m2	311	308	−2.2 (−3.3 to −1.1)†	−1.9 (−2.9 to −0.9)†	−0.3 (−1.8 to 1.2)	0.74
De Brito-Ashurst et al 37	Change in BP	Reduction systolic/diastolic BP	25	23	–	–	−8 mm Hg (−11 to −5)/2 (−4 to −2)	<0.001
Dussol et al 61	Decrease in eGFR	Decrease eGFR, mL/min/1.73 m2	25	22	−7±11	−5±15	–	–
	24-hour albumin excretion rate	Microalbuminuria, mg/d			+114±364	+156±486	–	–
MDRD‡ Study 1 (1995)	Dietary satisfaction (Study A: GFR 25–55 mL/min. 1.73m2)	Dietary satisfaction score	220	221	3.6±1.0	3.8±1.0	–	<0.05
	Dietary satisfaction (Study B: GFR 13–24 mL/min. 1.73m2	Dietary satisfaction score	65	59	3.1±0.9	3.6±0.9	–	<0.01
MDRD‡ Study 2 (1996)	Decline eGFR (Study A: GFR 25–55 mL/min. 1.73m2)	Decline eGFR, baseline to 3 years	291	394	–	–	3.8 (4.2)§	–
	Decline eGFR (Study B: GFR 13–24 mL/min. 1.73m2)	Decline eGFR, baseline to 3 years	126	129	–	–	4.0 (3.1)§	–
Mekki et al 62	Total cholesterol (TC)	TC/mmol L-1	20	20	4.1±0.5	5.4±0.4	–	<0.05
	Triacylglycerols (TG)	TG/mmol L-1			2.9±0.1	3.9±0.1	–	<0.05
Meuleman et al 32	Blood pressure	Office systolic BP, mmHg	67	71	–	–	−7.3 (−12.7 to −1.9)¶	<0.01
		Office diastolic BP, mmHg			–	–	−3.8 (-6.9 to -0.6)¶	<0.05
	Sodium excretion	Sodium excretion rate, mmol/24 hours			–	–	2.9 (−21.6 to 27.3)¶
Paes-Barreto et al 46	Change in protein intake	Change protein intake, g/day	43	46	−20.7 (−30.9%)††	−10.5 (−15.1%)¶††	–	0.04
Pisani et al 42	Protein intake	Change protein intake, g/kg/day	27	27	−0.1 (−0.17 to −0.03)†	−0.2 (−0.28 to −0.13)†	–	0.04
	UUN excretion	Change UUN, g/day			−1.3 (−2.1 to −0.5)†	−2.8 (−3.6 to −2)†	–	0.008
	SUN	Change SUN, mg/dL			2.96 (−7.71 to 13.64)†	−16.63 (−27.3 to −5.96)†	–	0.012
	Urinary phosphate excretion	Change phosphate excretion, mg/day			−27.6 (−93.7 to 38.4)†	−165.3 (−231.3 to −99.2)†	–	0.005
	Serum phosphate concentration	Change serum phosphate, mg/dL			0.2 (0 to 0.4)†	−0.1 (−0.3 to 0.2)†	–	0.093
	Adherence	Met criteria, n, %			19 (70%)‡‡	11 (44%)‡‡	–	–
Rosman et al 63	Adherence (Group A1 & B: CrCl >30)	Median 24-hour urea excretion mmol/24 hours	45	47	–	–	–	<0.01
	Adherence (Group A2 & C: CrCl ≤30)	Median 24-hour urea excretion mmol/24 hours	23	17	–	–	–	<0.01
Saran et al64	Change hydration status	Extracellular Volume, L	29	29	–	–	−1.02 (−1.48 to 0.56)	<0.001
		Intracellular Volume, L			–	–	−0.06 (−0.12 to 0.01)	0.02
Physical activity interventions
Aoike et al 59	Cardiopulmonary	Maximal ventilation, L/min	14	15	90.7±28.1	76.6±23.3	–	0.003
	parameters	Ventilatory threshold, VO2 peak, ml/kg/min			26.1±7.0	24.2±7.1	–	0.302
		VO2 in respiratory compensation point, ml/kg/min			21.7±5.5	19.0±5.6	–	0.073
		Speed in respiratory compensation point, Km/h			6.8±1.1	5.8±1.0	–	<0.001
	Functional capacity	6MWT, minutes			583.1±85.2	561.2±91.2	–	0.028
		Time up/go test, seconds			5.82±1.39	6.42±1.11	–	0.001
		Arm curl test, repetitions			22.8±4.8	18.1±3.1	–	<0.001
		STST, repetitions			24.0±7.1	18.3±4.8	–	<0.001
		2-minute step test, steps			219.3±36.7	179.9±36.3	–	<0.001
		Back scratch test, cm			6.4±6.6	12.6±9.9	–	0.05
	Systolic and diastolic BP	Systolic BP, mm Hg			118.7±7.3	126.8±6.7	–	0.012
		Diastolic BP, mm HgP			76.1±4.4	81.0±3.7	–	0.038
	Renal function	Serum creatinine, mg/dL			2.6±1.1	3.2±1.4	–	0.215
		eGFR, mL/min/1.73 m2			31.9±13.7	23.9±12.2	–	0.046
Barcellos et al 65	Mean change in eGFR	Change eGFR, mL/min/1.73 m2	76	74	61.5 (57.0 to 66.1)†	59.0 (54.2 to 63.8)†	0.7 (−4.0 to 5.4)	–
Greenwood et al 43	Mean change in eGFR	Change eGFR, mL/min/1.73 m2	8	10	−3.8±2.8	−8.5±6.4	7.8±3.0 (1.1 to 13.5)	0.02
Kao et al 30	Depression	Change depression (Beck Depression Inventory-II scale)	45	49	−3.71§§	1.33§§	–	<0.01
	Fatigue	Change fatigue			−4.74§§	1.91§§	–	<0.001
	Exercise behaviour	Change weekly exercise			4.28§§	−1.24§§	–	<0.001
Leehey et al66	UPCR ratio	UPCR (mg/g) at 52 wks	14	18	405 (225 to 1038)†††	618 (323 to 1155)†††	–	0.39
Rossi et al 45	Physical function	6MWT, minutes	59	48	210.4±266	−10±219.9	–	<0.001
		STST, seconds			26.9%±27% age prediction***	0.7%±12.1% age prediction***	–	<0.001
		Gait speed, cm			9.5 (−36.4 to 34)†††	0 (−9 to 13)†††	–	0.76
	QoL (RAND SF-36),	Role functioning/physical			19.0±31.7	−8.9±38.4	–	<0.001
	mean change from	Physical functioning			11.1±19.3	−0.7±18.7	–	0.004
	baseline	Energy/fatigue			9.8±17.6	0.5±18.0	–	0.01
		General health			4.9±15.3	−1.2±11.5	–	0.03
		Pain			5.7±20.0	−3.8±24.4	–	0.04
		Emotional well-being			4.2±16.9	−0.4±17.1	–	0.2
		Social functioning			4.2±20.8	1.6±22.6	–	0.57
		Role functioning/emotional			6.9±24.5	1.9±29.2	–	0.38
Tang et al 49	Physical function	Change 6MWT, minutes	42	42	41.93±14.57	−5.05±14.81	–	<0.001
		Change STST, seconds			−2.68±1.95	0.49±2.07	–	<0.001
	Self-efficacy	Change self-efficacy score			6.64±6.92	−3.72±6.80	–	<0.001
	Anxiety	Change HAD-A score			−1.02±1.47	0.21±2.17	–	0.003
	Depression	Change HAD-D score			−0.76±1.32	0.31±1.84	–	0.003
	QoL (KDQOL-SF),	Symptom/problem list			2.49±4.81	0.38±6.97	–	0.007
	mean change from	Effects of kidney disease			1.90±5.22	−1.56±9.64	–	0.005
	baseline	Burden of kidney disease			−0.45±15.27	−15.3±18.11	–	<0.001
		SF-12 PCS			1.08±3.60	−0.74±4.55	–	0.045
		SF-12 MCS			1.87±5.69	−0.73±4.53	–	0.002
Van Craenenbroeck et al 34	Peripheral endothelial function	Flow mediated dilation of brachial artery	19	21	4.6±3.0	5.3±3.1	0.32 (−1.88 to 2.53)	0.9
Lifestyle interventions
Flesher et al 39	Composite of eGFR, TC, US, UP, BP	Number of improved endpoints	23	17	83	30		0.028
Howden et al 40	Change in CRF	VO2, ml/kg/min	36	36	2.8±0.7	0.3±0.9	–	0.004
Ishani et al 41	Composite death, hospitalisation, emergency visits, admission nursing facility	Occurrence of primary outcome/HR	451	150	208 (46.2%)	70 (46.7%)	–	0.9
Jiamjariyapon et al 67	Mean change in eGFR	Change eGFR, mL/min/1.73 m2	234	208	42.4±1.5	39.9±2.8	2.74 (0.60 to 4.50)	0.009
Joboshi and Oka44	Perceived behaviour	Self-efficacy	32	29	r=0.27, U=318.5**	–	–	0.035
		Self-management			r=0.29, U=310.0**	–	–	0.026
Patil et al 68	24-hour urine protein	24-hour urine protein, g/d	23 (B)	22 (A),31 (C)	1284.74±1079.94	A: 1079.27±1269.20; C: 1187.61±756.92	–	–
	BMI	Change in BMI (paired t-test)			−1.95±1.10	A: −0.15±0.38 (p=0.069); C: −2.56±0.68 (p=0.000)	–	0.000
Teng et al 31	Health-promotion lifestyle	Stress management	45	45	–	–	2.76	0.10
	behaviours (HPLP-IIC)	Interpersonal relations			–	–	3.88	0.05
		Health responsibility			–	–	13.63	0.001
		Physical activity			–	–	7.50	0.01
		Spiritual growth			–	–	2.79	0.10
		Nutrition			–	–	2.62	0.11
	Renal function protection knowledge	Knowledge renal function, Chinese herbs and CKD diet			–	–	No data	0.001
	Physical function	6MWT, minutes	45	45	420.4±81.2	368.5±99.7	–	0.04

*Unless otherwise indicated, values are shown as mean+/-SD.

†Mean change (95% CI).

‡Modification of Diet in Renal Disease (MDRD) study (Gillis et al ³³, Coyne et al ⁴⁸).

§Mean decline +/-SD.

¶Mean change from baseline after 6 months.

**Effect size (r) Median, Mann-Whitney's U Test.

††Mean change and % reduction from baseline values.

‡‡ Number of participants who met adherence criteria (n,%).

§§Paired T test.

¶¶ p-value calculated as p<0.05 x group interaction (Aoike 2015).

***STST results standardized as a percentage of age-predicted value using prediction formulas (Rossi 2014).

†††Median (IQR)

BMI, Body Mass Index; BP, blood pressure;CrCl, Creatinine clearance; CRF, Cardiorespiratory fitness; eGFR, estimated glomerular filtration rate; HAD-A/HAD-D, Hospital Anxiety & Depression Scale; HPLP-IIC, Health Promoting Lifestyle Profile-II Chinese version (questionnaire); KDQOL-SF, Kidney Disease & Quality of Life Short Form; 6MWT, 6 min Walk Test; SF-12 PCS/MCS, Physical and Mental Health Composite Scores; QoL, Quality of life; RAND SF-36, 36-Item Short Form Survey; STST, Sit to Stand Test; SUN, Serum urea nitrogen; UP, Urinary protein; UPCR, Urine protein to creatinine ratio; US, Urinary sodium; UUN, urinary urea nitrogen.

Eighteen studies (69%) showed a significant improvement in at least one primary outcome and all of these studies included education, persuasion, modelling and incentivisation as an intervention function (see online supplementary table S3). A meta-analysis of the data was not possible due to heterogeneity of outcome measures across the included studies. The heterogeneity of outcomes also meant we could not link outcomes with specific behaviour change techniques. Many studies are likely to be underpowered to detect modest effects, and so the absence of a statistically significant effect should not be regarded as evidence of no effect.

Discussion

Behaviour change interventions in trials in patients with CKD mostly focused on diet and physical activity. The primary outcomes of the trials were diverse and most were biochemical outcomes (eg, eGFR, blood pressure, peak VO₂ and sodium or albumin excretion), with few clinical or patient-reported and/or behavioural outcomes such as quality of life, fatigue, knowledge, self-efficacy and self-management.30 31 38 39 44 45 Only five interventions were underpinned by theory. The most frequently used intervention function was education, followed by enablement and training. Persuasion, environmental restructuring, modelling and incentivisation were used less frequently. Coercion and restrictions (which includes regulation) were not used in any of the studies. The top five most common behaviour change techniques were instruction on how to perform the behaviour, social support, demonstration of the behaviour, feedback on behaviour and behavioural practice/rehearsal. Identity, scheduled consequences and covert learning were not used in any of the studies. No association between frequency of functions or behaviour change techniques and the effect of interventions on outcomes could be identified.

The use of multiple behaviour change techniques does not necessarily lead to better outcomes and some evidence suggests that fewer techniques and the right combinations of techniques suited to the context are more effective.^50–52 Education was the most frequent intervention function used across the studies, which may be because it has been consistently shown that patients with CKD lack awareness about lifestyle risk factors and have low health literacy.^{10 11 53} Specifically, the behaviour change technique, ‘instruction on how to perform the behaviour’, was the most frequently reported technique, used in all interventions except two. We suggest this is highly applicable because dietary interventions can involve complex dietary restrictions of sodium, protein, potassium and phosphate. Patients have sought practical advice about how to implement these restrictions.⁵⁴ However, most educational strategies used a didactic approach, with health professionals verbally conveying information or providing written materials. Patients with CKD prefer multiple problem-solving and collaborative approaches, in partnership with health professionals.⁵⁴ Also, written materials for patients with CKD have a reading grade of 9 (age 14–15 years), which is higher than the recommended level (grade 5).¹⁰

The intervention function ‘training’ was used in every study targeting physical activity but was only used in two dietary interventions. Patients with CKD are overwhelmed by dietary information which can be complex, restrictive and insensitive to cultural norms.⁵⁴ A recent review of educational interventions for CKD patients found that including practical skills and workshops was associated with better outcomes.⁵⁵ For example, a low-salt programme for Bangladeshi patients with CKD in the United Kingdom included cooking and educational sessions facilitated by Bengali workers in a community kitchen. It targeted both patients and family members who cooked their own low-salt version of Bangladeshi recipes and led to a reduction in salt intake and reduced blood pressure for participants.³⁷ Approaches to enabling and training patients for behaviour change incorporating hands-on training may be more effective.

Our findings are similar to recent reviews of behavioural interventions for other conditions (cardiovascular disease, obesity, rheumatoid arthritis, prostate cancer and diabetes), which also found that behavioural interventions are not well-reported, not informed by theory and have diverse outcomes and modes of delivery.25–27 51 56 The behaviour change techniques associated with goals and planning, feedback and monitoring and social support have also been frequently used in behaviour changes interventions in patients with other chronic conditions. These techniques are proven strategies for behaviour change and in line with evidence-based recommendations for lifestyle modification.^{12 13 57}

We identified and described the behaviour change techniques and intervention functions in lifestyle behavioural interventions for patients with CKD with comprehensive evidence-based frameworks. Coding of behaviour change techniques and intervention functions was systematically and independently conducted by three researchers, and risk of bias was assessed. Potential limitations relate to poor reporting. Some interventions may have used behaviour change techniques or intervention functions in their study but did not report them, or details of techniques were unclear. We contacted authors and examined all associated supplementary materials and papers to collect more information.

Lifestyle behaviour change interventions for patients with CKD appear to integrate recommended and proven behaviour change techniques and intervention functions. These techniques such as goals and planning and self-monitoring are important but focus on individual agency rather than external factors. Interventions could be improved by considering the context of behaviour change and the social and physical environment of participants. For example, most of the interventions for physical activity focused on structured exercise programme and a reliance on equipment (eg, exercise bikes). Patients with CKD need to be able to integrate physical activity in to their daily lifestyle.⁵⁸ However, only one intervention for physical activity gave instructions on how to incorporate physical activity to fit in with daily activities and in environments easily accessible to patients, without the use of equipment.⁵⁹ This study reported improvements in cardiopulmonary and functional capacities of overweight patients with CKD.

Optimising the social environment and arranging support from friends, family and the community may also improve lifestyle behaviour change interventions for patients with CKD. Family support was used rarely in interventions in this review and only included in two studies.^{31 37} However, informal caregivers play an important role in the management of CKD and are often required to change their own lifestyle behaviours to support patients with CKD.⁶⁰ Characteristics of effective educational interventions for patients with CKD involved the patient’s family.⁵⁵

The quality of the design and reporting of lifestyle behaviour change interventions for patients with CKD requires explicit description of behavioural strategies to ensure interventions are generalisable and replicable. There are numerous evidence-based guidelines that recommend the explicit use of behaviour change techniques for addressing lifestyle risk factors in chronic disease prevention and these may be better used when designing and reporting interventions for patients with CKD. Recently the National Institute of Health and Care Excellence in the UK published comprehensive guidelines specific to behavioural interventions and lifestyle modification.¹² The WHO’s recommendations on behaviour change support this and further reinforce the need to consider the social and environmental determinants of health in changing lifestyle behaviours.⁵⁷

Conclusion

Lifestyle interventions in trials conducted in patients with CKD mostly focus on goals and planning, feedback and monitoring and education. However, we suggest that interventions may be improved by using interactive and tailored training, and strategies to help patients incorporate lifestyle modification in their daily activities, and physical and social environments. Explicit application of behaviour change taxonomies may help to increase the effect of lifestyle behaviour change interventions for improved health outcomes in patients with CKD.