Meta-Analysis: Caregiver and Youth Uncertainty in Pediatric Chronic Illness

Abstract

Objective To conduct a systematic review on the construct of illness uncertainty in caregivers and youth as related to the following: demographic and illness variables, psychological functioning, illness-related distress, and reaction/coping style. Methods A meta-analysis was conducted with articles assessing the associations between illness uncertainty and variables of interest that were published between November 1983 and June 2016 (n = 58). Results Psychological functioning and illness-related distress had primarily medium effect sizes. Demographic and illness variables had small effect sizes. More positive and fewer negative reaction/coping styles were associated with less illness uncertainty, with primarily small effects. Conclusions Illness uncertainty may be an important factor that influences psychological functioning and distress and coping in the context of pediatric chronic illness. However, additional research is needed to determine more precise mean effect sizes, as well as the potential efficacy of intervention to address uncertainty. Key words: adolescents, children, chronic illness, coping skills and adjustment, meta-analysis, parents, psychosocial functioning.

Introduction

Illness uncertainty is a cognitive appraisal (i.e., the interpretation of an event as it pertains to one’s well-being) that is elicited when an individual has difficulty in understanding the meaning of an illness-related event because health outcomes are unpredictable (e.g., what side effects will be experienced when starting a new medication) (Mishel, 1984), illness symptoms or events are ambiguous (e.g., is leg pain due to overexertion or relapse?), or there is a lack information about the illness or treatment (e.g., unknown outcomes of experimental treatments) (Mishel, 1990). Illness uncertainty has been associated with psychological functioning and coping in multiple studies across adult and pediatric chronic illness populations (Stewart & Mishel, 2000; Wright, Afari, & Zautra, 2009). Pediatric illness uncertainty (i.e., illness uncertainty experienced by a patient or a caregiver in the context of a pediatric illness), in particular, requires consideration of the developmental level of the child, the need to independently assess patient and caregiver uncertainty, and the subsequent consideration of potential transactional relationships between patient and caregiver illness uncertainty and potential associated outcomes (Lazarus, & Folkman, 1984; Wallander & Varni, 1998; Fedele et al., 2013). The purpose of this study is to synthesize literature and describe the nature of pediatric illness uncertainty for both pediatric patients and caregivers of children with a pediatric illness.

The development of the Child Uncertainty in Illness Scale, based on Mishel’s illness uncertainty model (Mullins & Hartman, 1995; Pai et al., 2007), facilitated research that examined illness uncertainty in the context of pediatric illness. A subsequent factor analysis of the Child Uncertainty in Illness Scale revealed factors (e.g., unpredictability and lack of understanding) consistent with Mishel’s illness uncertainty model and adult illness uncertainty measures (Pai et al., 2007). The pediatric illness uncertainty literature has spanned the developmental spectrum from young children to adolescents and young adults (Fedele et al., 2012, Mullins et al., 2007, Page et al., 2012). Mirroring the adult literature, many studies show that increased child illness uncertainty is associated with poorer overall quality of life (Fortier et al., 2013) and psychosocial functioning (Hoff, Mullins, Chaney, Hartman, & Domek, 2002; Mullins et al., 2007), including increased anxiety (Hommel et al, 2003) and depression (Carpentier, Mullins, Wagner, Wolfe-Christensen, & Chaney, 2007). However, some pediatric studies show no relationship between illness uncertainty and psychological functioning or coping (Carpentier, Mullins, & Van Pelt, 2007a; White et al., 2005). Overall, the relationship of demographic (e.g., age) and illness-specific factors (e.g., illness duration) in relation to patient reports of illness uncertainty has varied across studies (Carpentier, Mullins, Chaney, & Wagner, 2006; Hommel et al., 2003 Mullins et al., 2007; Sterken, 1996).

The examination of pediatric illness uncertainty would be incomplete without also considering the influence of caregivers. Several studies have indeed examined caregiver uncertainty (i.e., uncertainty a caregiver has about their child’s illness) and shown that higher caregiver-reported uncertainty is associated with higher levels of caregiver-reported psychological stress and more difficulty in coping (Chaney et al., 2016; Molzon et al., 2014). A few notable studies have assessed caregiver and patient reports of uncertainty about the child’s illness, (Fedele et al., 2012; Yarcheski, 1988; Santacroce, Asmus, Kadan-Lottick, & Grey, 2010), including two that examined potential transactional relationships (Page et al., 2012; Stewart, Lynn, & Mishel, 2010). These studies showed that caregiver uncertainty was not only associated with their own psychological functioning but their child’s as well.

Research examining illness uncertainty in pediatric chronic illness has grown considerably in the past three decades. In the absence of a systematic analysis, the variability in chronic illness populations and reporters, and mixed findings across studies, hampers the ability of clinicians and scientists to interpret and translate the collective findings. Therefore, the purpose of this study is to conduct a meta-analysis that will consolidate and summarize the current pediatric illness uncertainty literature. Specifically, we will describe factors associated with and how pediatric illness uncertainty relates to psychological functioning in pediatric chronic illness for both patients and caregivers. The results of these analyses are needed to inform the direction of future research and help determine the viability of pediatric illness uncertainty (either patient- or caregiver-report) as a target for intervention. Based on previous illness uncertainty literature, we hypothesize that demographic-specific (i.e., age and income) or illness-specific (i.e., duration and severity) factors will not be related to pediatric illness uncertainty. However, we hypothesize a number of relationships between uncertainty and psychosocial functioning and coping: (1) patient-reported illness uncertainty will be associated with patient psychological outcomes; (2) caregiver-reported illness uncertainty (i.e., caregiver’s uncertainty about their child’s illness) will be associated with caregiver psychological functioning; and (3) caregiver uncertainty will be associated with child psychological functioning, illness-related distress, and reaction/coping styles.

Method

The review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (Moher, Liberati, Tetzlaff, Altman, & PRISMA group, 2009).

Search Strategy

To the authors’ knowledge, this is the first meta-analysis to be conducted that focuses on pediatric illness uncertainty. A literature search was done to identify all potential articles on pediatric illness uncertainty published from November 1983 to June 2016. The starting date was chosen to correspond with Mishel and colleagues’ hallmark article on parental perceptions of uncertainty in childhood illness (Mishel, 1983). The keyword search strategy using PubMed, Cinahl Plus, and PsychInfo included: [infant OR child OR adolescent OR teen OR young adult OR parent OR mother OR father OR caregiver] AND [chronic illness OR chronic disease OR illness OR disease OR asthma OR cancer OR diabetes OR epilepsy OR genetic disorder OR HIV OR NICU OR neonatology OR prematurity OR intensive care OR sickle cell disease OR transplant] AND [uncertainty OR illness uncertainty OR Mishel’s theory of uncertainty].

The initial search yielded 4,107 results. To best represent the current literature in pediatric illness uncertainty, the following inclusion criteria were used: the study collected original data; a quantitative measure of uncertainty was used that included more than one item; a relationship to at least one other variable was reported; the study target population was exclusively children, adolescents, or young adults (defined as ≤30 years of age) with a chronic illness or their caregivers; and the article was published in English in a peer-reviewed journal. Additionally, articles were excluded if the focus of the study was specific to the construct of “diagnostic uncertainty” because it focuses on the period before diagnosis and whether a diagnosis will be made (Santacroce, 2001). References of identified studies, as well as relevant qualitative reviews and manuscripts, were then reviewed to ensure a thorough representation of all available literature. This process yielded three additional studies to include in the meta-analysis.

On initial review of abstractions by one author (L.S.) and removal of duplicate articles, 312 articles were identified for full-text review to determine eligibility. For a complete breakdown of article selection and exclusion, see Figure 1. A total of 58 articles were included in the meta-analysis representing 5,373 caregiver participants and 2,407 patient participants. See Table 1 for a list of included articles with applicable study variables.

Figure 1.

The PRISMA four-phase flow diagram detailing study selection.

Table I.

Articles Included in Meta-Analysis With Individual Variable Correlations

	N	Patient’s age (mean (SD), range)	Patient’s gender (%male)	Illness group(s)	Uncertainty measure (s)	Uncertainty reporter	Variable	Measure	Effect size (r)
Boman, Viksten, Kogner, and Samuelsson, 2004	467 (C)			Cancer; diabetes	PPD-C	Caregiver	Illness Specific
							Illness duration		.13
Bonner et al., 2006	149 (C)	9.4 (5.26), .83–18	52.3	Brain tumor	PECI, subscale	Caregiver	Caregiver psychological functioning
							Anxiety	BSI-Anx	.15
							Depression	BSI-Dep	.17
							Psychological distress	BSI-GSI	−.33**
							Traumatic stress	IES	−.27
							Family impact	IFS	.31*
							Guilt/worry	PECI	.63**
							Caregiver illness-related distress
							Externalized subjective burden	CGSQ	.19
							Internalized subjective burden	CGSQ	.46**
							Objective burden	CGSQ	.18
							Caregiver positive reaction/coping
							Caregiver emotional resources	PECI	−.37**
Carpentier et al., 2006	30 (C)			Diabetes	PPUS	Caregiver	Demographic
							Caregiver age		−.28
							Family income		−.10
							Illness specific
							Illness duration		.11
							Caregiver psychological functioning
							Psychological distress	BSI-GSI	.46*
							Caregiver negative reaction/coping
							Negative attributional style	ASQ-	.00
Carpentier, Mullins, and Van Pelt, 2007	121 (P)	19.7 (1.26), 18–22	37.19	Asthma	MUIS-C	Patient	Demographic
							Patient age		.03
							Illness specific
							Illness severity		−.06
							Patient psychological functioning
							Anxiety	BSI-Anx	−.01
							Depression	BSI-Dep	−.02*
							Psychological distress	BSI-GSI	−.04
							Patient illness-related distress
							Intrusiveness	IIRS	.53**
							Subjective severity		.01
Carpentier, Mullins, Wagner, Wolfe-Christensen, and Chaney, 2007	127 (P)	10.14 (1.44), 8–12	47.24	Diabetes; asthma; cystic fibrosis	CUIS	Patient	Demographic
							Patient age		−.26**
							Family income		−.22*
							Illness specific
							Illness duration		.10
							Patient psychological functioning
							Depression	CDI	.57**
							Patient negative reaction/coping style
							General negative attribution style	CASQ-R	.31**
							Internal negative attribution style	CASQ-R	.13
Chaney et al., 2016	82 (C); 82 (P)	13.17 (2.80), 7–18	32	Juvenile rheumatic disease	PPUS	Caregiver	Demographic
							Patient age		.05
							Family income		.01
							Illness specific
							Illness severity	PEDS	.40**
							Patient psychological functioning
							Depression	CDI	.27*
							Caregiver psychological functioning
							Psychological distress	BSI	.28**
							Caregiver illness-related distress
							Caregiver demands	CMCRDS	.47**
Decker, Haase, and Bell, 2007	193 (P)	15, 11–21; 16, 12–21; 16.2, 12–22	55	Cancer	MUIS	Patient	Illness specific
							Illness duration		.05
Fedele et al., 2009	65 (P)	20.22 (2.22), 18–30	30.8	Asthma	MUIS-C	Patient	Patient psychological functioning
							Mental quality of life	SF-36	.28**
							Patient illness-related distress
							Physical quality of life	SF-36	.29**
Fedele et al., 2011	51 (C); 51 (P)	14.09 (2.43), 9–17	37.25	Juvenile rheumatic disease	PPUS	Caregiver	Patient psychological functioning
							Depression	CDI	.39**
							Caregiver psychological functioning
							psychological distress	BSI-GSI	.46**
Fedele et al., 2012	122 (C); 122 (P)	13.56 (2.67), 7–18	32.8	Juvenile rheumatic disease	PPUS; CUIS	Caregiver; patient	Patient psychological functioning
							Depression	CDI	.29**b; .26**c
							Caregiver psychological functioning
							Psychological distress	BSI-GSI	.31**b; .20*c
							Patient illness-related distress
							Illness functioning, caregiver report	JAF-P	.23*b; .09c
							Illness functioning, patient report	JAF-C	.20*c
							Caregiver illness-related distress
							Intrusiveness, relationship factors	IIS-P	.38**b; .20*c
							Intrusiveness, instrumental factors	IIS-P	.50**b; .09c
Fortier et al., 2013	120 (C); 120 (P)	12.9 (3.0), 8–18	53.3	Cancer	CUIS	Patient	Demographic
							Patient age		−.22*
							Patient psychological functioning
							Anxiety	STAIC	.34**
							Quality of life, worry	PedsQL-C	−.35**
							Quality of life, emotional	PedsQL-C	−.48**
							Quality of life, cognitive	PedsQL-C	−.33**
							Quality of life, social	PedsQL-C	−.43**
							Patient illness-related distress
							Subjective pain severity		.28**
							Current pain		.39**
							Quality of life, physical	PedsQL-C	−.37**
							Quality of life, physical	PedsQL-P	−.21
							Quality of life, pain	PedsQL-C	−.36**
							Quality of life, pain	PedsQL-P	−.30**
							Quality of life, nausea	PedsQL-C	−.41**
							Quality of life, nausea	PedsQL-P	−.25**
							Quality of life, appearance	PedsQL-C	−.29**
							Quality of life, appearance	PedsQL-P	−.28**
							Quality of life, procedural anxiety	PedsQL-C	−.31**
							Quality of life, procedural anxiety	PedsQL-P	−.37**
							Quality of life, treatment anxiety	PedsQL-C	−.47**
							Quality of life, treatment anxiety	PedsQL-P	−.33**
Franck et al., 2014	107 (C)	8.3 (6.1), 0–18	56	General medical; surgical; cancer; neurological; cardiac	PPUS	Caregiver	Illness Specific
							Illness duration		.14
							Caregiver psychological functioning
							Anxiety	HADS	.27**
							Depression	HADS	.28**
							Traumatic stress	IES-R	.31**
							Patient illness-related distress
							Health at follow-up		−.31**
							Patient positive reaction/coping style
							Active coping style	COPE	.00
							Optimism coping style	COPE	−.27**
							Social support coping style	COPE	−.02
							Patient negative reaction/coping style
							Disengagement/substance use coping style	COPE	−.02
							Distraction/humor coping style	COPE	.04
							Negative coping style	COPE	.11
Fuemmeler, Mullins, and Marx, 2001	28 (C)	14, 9–24	42	Brain tumor	PPUS	Caregiver	Caregiver psychological functioning
							Psychological distress	GSI-BSI	.56**
							Traumatic stress	PDS	.39*
Grootenhuis and Last, 1997a	163 (C); 84 (P)	13.06 (3.50)	52	Cancer	SSERQ	Caregiver	Demographic
							Caregiver age, maternal		−.20
							Caregiver age, paternal		−.02
							Illness specific
							Illness duration		−.05d; −.13e
							Patient psychological functioning
							Depression	DQC	.19d; .23e
							Caregiver positive reaction/coping style
							Interpretive control, maternal	CSS	.09
							Interpretive control, paternal	CSS	−.04
							Predictive control, maternal	CSS	−.51**
							Predictive control, paternal	CSS	−.62**
							Vicarious control, maternal	CSS	.08
							Vicarious control, paternal	CSS	−.04
							Caregiver negative reaction/coping style
							Illusory control, maternal	CSS	.15
							Illusory control, paternal	CSS	.17
Grootenhuis and Last, 1997	163 (C); 84 (P)a	12.88 (3.22); 13.24 (3.81)	52	Cancer	SSERQ	Caregiver	Caregiver psychological functioning
							Anxiety	TRAIT	.77**
							Depression	BDI	.71**
							Helplessness	SSERQ	.67**
							Loneliness	SSERQ	.50**
							Positive feelings	SSERQ	.08
He, You, Zheng, and Bi, 2016	95 (C)	6 (3.8)	66.3	Cancer	PPUS	Caregiver	Caregiver positive reaction/coping style
							Coping	CHIP	−.24*
Hoff et al., 2002	68 (P)	14.8, 13–18	51	Diabetes	CUIS	Patient	Demographic
							Family income		−.04
							Illness Specific
							Illness duration		.16
							Patient psychological functioning
							Psychological distress	BSI-GSI	.44**
							Patient positive reaction/coping style
							Perceived control	PCS	−.42**
Hommel et al., 2003	56 (P)	19.11 (.97), 18–21	29	Asthma	MUIS-C	Patient	Demographic
							Patient age		−.13
							Illness Specific
							Illness duration		.01
							Illness severity		.23
							Patient psychological functioning
							Anxiety	BAI	.56**
							Depression	IDD	.48**
							Patient illness-related distress
							Subjective severity		.17
Hullmann, Eddington, Molzon, and Mullins, 2013	148 (P)	19.77 (1.6), 18–25	37.8	Asthma; allergies	MUIS	Patient	Demographic
							Patient age, asthma		.14
							Patient age, allergy		.04
							Patient psychological functioning
							Quality of life, mental (asthma)	SF-36	−.35**
							Quality of life, mental (allergy)	SF-36	−.32**
							Patient illness-related distress
							Intrusiveness (asthma)	IIS	.46**
							Intrusiveness (allergy)	IIS	.46**
							Quality of life, physical (asthma)	SF-36	−.53**
							Quality of life, physical (allergy)	SF-36	−.34**
							Subjective severity (asthma)		.48**
							Subjective severity (allergy)		.11
Knapp, Sberna-Hinojosa, Baron-Lee, Curtis, and Huang, 2014	266 (C)	11.5 (5.5)	55	Not defined, part of palliative care program	DCS	Caregiver	Demographic
							Caregiver age		−.01
							Caregiver psychological functioning
							Family impact	IOF	.01
							Caregiver illness- related distress
							Perceived health		.04
Lee, et al. 2007	51 (C)	<1, N = 16; 1–3, N = 17; 4–7, N = 13; >7, N = 6; 0–9	49	Congenital heart disease	PPUS	Caregiver	Caregiver psychological functioning
							Parenting stress	PSI-SF	.46**
							Caregiver positive reaction/coping style
							Social support	PRQ 85	−.48**
Lipinski et al., 2006	363 (C)	7.4 (5)		Rare chromosomal disorders	PPUS	Caregiver	Demographic
							Caregiver age		−.13*
							Caregiver illness- related distress
							Subjective severity		.17**
							Perceived benefit of diagnosis		−.25**
							Caregiver positive reaction/coping style
							Perceived control		−.32**
Maurice-Stam, Oort, Last, and Grootenhuis, 2008	231 (C)	8 (4.4), 1.1–18.2	58.1	Cancer	SSERQ	Caregiver	Demographic
							Caregiver age, maternal		−.11
							Caregiver age, paternal		−.00
							Illness Specific
							Illness duration		.14
							Patient illness-related distress
							Visible consequences, mother		.01
							Visible consequences, father		−.12
							Caregiver positive reaction/coping style
							Active problem focusing, mother	UCL	−.29**
							Active problem focusing, father	UCL	−.33**
							Adaptability, mother	FACES	.16
							Adaptability, father	FACES	.05
							Predictive control, mother	CCSS	−.52**
							Predictive control, father	CCSS	−.56**
							Caregiver negative reaction/coping style
							Illusory control, mother	CCSS	.17
							Illusory control, father	CCSS	.05
							Palliative reaction, mother	UCL	.01
							Palliative reaction, father	UCL	.28*
							Passive reaction, mother	UCL	.61**
							Passive reaction, father	UCL	.60**
Mishel, 1983	272 (C)	0–4 months, N = 48; 5–9 months, N = 35; 10–12 months, N = 10; 13 months–3 years, N = 53; 3.5–6 years, N = 48; 7–10 years, N = 38; 11–14 years, N = 28; 15–18 years, N = 12	50	General medical; surgical; diagnostic	PPUS	Caregiver	Caregiver illness-related distress
							Subjective severity		.16**
Molzon et al., 2004	173 (C)		54	Cancers	PPUS	Caregiver	Caregiver psychological functioning
							Anxiety	BSI-ANX	.73**
							Depression	BSI-DEP	.49**
							Psychological distress	BSI-GSI	.48**
							Traumatic stress	IES-R	.47**
Mu, Ma, Ku, Shu, Hwang, & Kuo, 2001	100 (C)	9.7 (5.26); 0.5–19	65	Cancer	PPUS	Caregiver	Caregiver psychological functioning
							Anxiety	STAI	.25*
							Caregiver illness-related distress
							Boundary ambiguity	BAS	.46**
							Caregiver positive reaction/coping style
							Sense of mastery	SMS	−.50**
Mu, Wong, Chang, and Kwan, 2001	324 (C)	9.4 (4.9), 6–19	57.4	Epilepsy	PPUS	Caregiver	Demographic
							Patient age		.02
							Caregiver age		−.06
							Caregiver psychological distress
							Depression	BDI	.36**
							Caregiver illness-related distress
							Boundary ambiguity	BAS	.52**
Mu, Ma, Hwang, and Chao, 2002	80 (C)	9.6 (5.5), 0.6–19	63.8	Cancer	PPUS	Caregiver	Caregiver psychological functioning
							Anxiety	STAI	.31**
							Caregiver positive reaction/coping style
							Sense of mastery	PMS	−.43**
Mu, 2005	210 (C)	9.23(4.81) 0.6–19.8 years	65.25	Epilepsy	PPUS	Caregiver	Demographic
							Caregiver age		−.09
							Caregiver psychological distress
							Depression	BDI	.26**
							Caregiver positive reaction/coping style
							Family coping	CHIP	−.23**
Mullins, Chaney, Pace, and Hartman, 1997	49 (P)	19.8(2.0), 17–26	42	Asthma	MUIS-C	Patient	Patient psychological functioning
							Psychological distress	SCL-90-GSI	.46**
							Patient negative reaction/coping style
							Global negative attribution	ASQ-GN	.05
							Internal negative attribution	ASQ-IN	.27*
							Stable negative attribution	ASQ-SN	.08
Mullins, 2000	40 (P)	19.76 (1.77), 18–25	45	Asthma	MUIS-C	Patient	Patient psychological functioning
							Depression	IDD	.51**
							Patient illness-related distress
							Intrusiveness	IIS	.36*
							Subjective severity		.38*
Mullins, 2007	164 (C)	10.13 (1.42), 8–12; 14.79 (1.70), 13–18	82	Asthma; diabetes	CUIS	Patient	Demographic
							Patient age		−.17*
							Family income		−.18*
							Illness Specific
							Illness duration		.03
							Caregiver psychological distress
							Parenting stress	PSI-SF	.29**
							Caregiver illness-related distress
							Perception of child vulnerability	CVS	.32**
							Overprotection behaviors	PPS	.17*
Mullins, 2012	52 (C)	7.32 (4.32), 2–17; 9.06 (4.85), 2–16	53.8	Cancer	PPUS	Caregiver	Caregiver psychological functioning
							Psychological distress	SCL-90-r	.23
							Traumatic stress	IES-R	.16
Neville, 1998	60 (P)	18.98 (2.17), 14.8–22.7	40	Cancer	MUIS	Patient	Patient psychological functioning
							Psychological distress	BSI-GSI	.55**
							Patient positive reaction/coping style
							Social support	PRQ-85-2	−.30**
Page et al., 2012	103 (C); 103 (P)	10.14 (1.4)8–12	43.7	Diabetes; asthma; cystic fibrosis	PUIS; CUIS	Caregiver; patient	Patient psychological functioning
							Depression	CDI	.23*b; .53**c
							Caregiver versus patient uncertainty		.38**
Pai et al., 2007	373 (P)	12.64 (2.73)	51	Diabetes; asthma; cystic fibrosis; cancer; sickle cell disease; juvenile rheumatic disease	CUIS	Patient	Psychological functioning
							Depression	CDI	.33*
Patterson and McDonald, 2015	76 (P)	18.5 (3.4)	33.9	Cancer	Author developed	Patient	Patient positive reaction/coping style
							Mindfulness	CAMM	.27**
Phillips-Salimi, Haase, Kintner, Monahan, and Azzouz, 2007	201 (P)	16.4 (2.9), 11–26; 14.8 (2.0), 10–21	55.2	Cancer	MUIS-R	Patient	Patient illness-related distress
							Symptom distress	MSDS	.34**
							Well-being	IWB	.42**
							Patient positive reaction/coping style
							Hope	HHI	.41**
Ryan et al., 2011	37 (C); 37 (P)	12.39 (2.87) 8–18	70	Diabetes; cystic fibrosis; asthma; cancer; sickle cell disease	CUIS	Patient	Caregiver psychological functioning
							Parenting stress	PSI-SF	−.46**
							Caregiver illness-related distress
							Perception of child vulnerability	CVS	−.33
							Overprotection behaviors	PPS	−.39*
Santacroce, 2002	15 (C)	6.6 (5.99)	25	Cancer	PPUS	Caregiver	Caregiver psychological functioning
							Anxiety	STAI	.11
							Traumatic stress	RI	.11
Santacroce et al., 2010	19 (C); 16 (P)	21 (3.7), 15–25	48	Cancer	MUIS-C	Patient; caregiver	Patient psychological functioning
							Anxiety	STAI-S	.36
							Traumatic stress	RI	.46*
							Caregiver psychological functioning
							Anxiety	STAI-S	.62**
							Traumatic stress	RI	.57*
							Patient illness-related distress
							Health promotion	HPLP-II	−.23
							Caregiver illness- related distress
							Health promotion	HPLP-II	−.34
Senger, Ward, Barbosa-Leiker, and Bindler, 2016	231 (C)	NR	49	Mitochondrial disease	PECI	Caregiver	Caregiver illness- related distress
							Illness-related distress (frequency)	PIP-F	.49**
							Illness-related distress (difficulty)	PIP-D	.53**
Stam, Grootenhuis, Brons, Caron, and Last, 2006	235 (C); 60 (P)	8.3 (4.6) 1.1–18.2	56.7	Cancer	SSERQ	Caregiver	Illness specific
							Illness duration		.52**d; .54**e
Sterken, 1996	31 (C)	1–5 years: 12 (39%); 6–12 years: 10 (32%); 13–18: 9 (29%)	55	Cancer	PPUS	Caregiver	Demographic
							Patient age		−.40*
							Caregiver age		−.48**
							Illness specific
							Illness duration		−.30
							Caregiver positive reaction/coping style
							Emotive coping style	JCS	.37*
							Confrontative coping style	JCS	−.34*
Stewart, Mishel, Lynn, and Terhorst, 2010a	68 (C); 68 (P)	13 (2.9), 8–18	58.3	Cancer	USK; PPUS	Patient; caregiver	Demographic
							Patient age		.19b; .21c
							Illness specific
							Illness knowledge	CKS	.09b;−.15c
							Patient psychological functioning
							Anxiety	RCMAS	.18b; .56**c
							Depression	CDI	.20b; .59**c
							Caregiver versus patient uncertainty		.26*
Stewart, Lynn, and Mishel, 2010	68 (C); 68 (P)a	13 (2.9) 8–18	58.3	Cancer	USK	Patient
									−.06
							Illness specific
							Illness duration		−.13b; −.23c
							Patient illness-related distress
							Subjective severity	PRISM	.36**
							Patient positive reaction/coping style
							Cohesion	FACES	−.39**
							Social support	NSSQ	−.17
Straveski, 2016	68 (C)	NR	NR	Cardiac surgery	PPUS	Caregiver	Caregiver positive reaction/coping style
							Readiness for discharge	RHDS	−.36*
							Sense of mastery	PMT	−.24
Suorsa et al., 2015	51 (C)	9.7 (6.45)	48.1	Disorders of sexual development	PPUS	Caregiver	Demographic
							Patient age		.17
							Caregiver age		.10
							Family income		−.10
Tackett et al., 2016	105 (C)	8.6 (5)	52.4	Cancer	PPUS	Caregiver	Caregiver psychological functioning
							Psychological distress	BSI-GSI	.41**
							Posttraumatic stress	IES-R	.33**
Tiwaree, Kantawang, Wonghongkul, and Lertwatthanawilat, 2016	96 (C); 96 (P)	56.25% between 12 and 15years	54.2	Cancer	CUIS; PPUS	Patient; caregiver	Illness specific
							Illness knowledge	IKS	.01b; −.31c
							Patient illness-related distress
							Symptom pattern	SPCC	.01b; .23*c
							Information support from providers	ISHP	−.03b;−.32c
							Information support from parents	ISPS	−.21b; −.15c
							Information support from peers	ISPeersS	.09b; .08c
							Caregiver versus patient uncertainty		.22*
Tomlinson, 1996	40 (C)	.5–1	21	PICU admission	MUIS-P	Caregiver	Demographic
							Family income		−.06
							Illness specific
							Illness severity	PRISM	.36**
							Caregiver positive reaction/coping style
							Cohesion	FACES-C	−.39**
							Social support	NSS	−.17
Tomlinson and Harbaugh, 2004	156 (C)	5.6, .004–18		Not defined, PICU admission	PPUS	Caregiver	Caregiver illness-related distress
							Role ambiguity	FNBAS	.54**
White et al., 2005	50 (C); 50 (P)	13.62 (2.42), 9–17	38	Juvenile rheumatoid arthritis	CUIS	Patient	Demographic
							Patient age		−.06
							Illness specific
							Illness duration		−.21
							Illness severity		−.08
							Patient psychological functioning
							Depression	CDI	−.03
							Caregiver psychological functioning
							Psychological distress	BSI-GSI	−.09
							Patient illness-related distress
							Subjective severity		−.15
Williams, Drew, Deluca, and McCarthy, 2013	35 (C)	11.45 (3.36), 5.3–18.6, 17.77 (.83), 16.44–19.32	47.6	Cancer	PECI	Caregiver	Psychological functioning
							Behavior and emotional functioning	SDQ	.36*
Wolfe-Christensen, Isenberg, Mullins, Carpentier, and Almstrom, 2008	102 (P)	19.7 (1.25), 18–22	36.2	Asthma	MUIS-C	Patient	Demographic
							Patient age		−.08
							Illness specific
							Illness severity		.07
							Patient psychological functioning
							Psychological distress	BSI-GSI	.53**
							Patient illness-related distress
							Subjective severity		.40**
Wray, Lee, Dearmun, and Franck, 2010	29 (C)	6.6 (5.29), .1–14	54	Not defined, hospitalized	PPUS	Caregiver	Caregiver psychological functioning
							Anxiety	HAD	.39*
Yarcheski, 1988	32 (P)	15.55 (1.1), 14–18	37.5	Cystic fibrosis	MUIS; PPUS	Patient; caregiver	Patient illness-related distress
							Future time perspective	FTP	.19
Ye, 2015	111 (C)	6.85 (4.37)	62.2	Cancer	PPUS	Caregiver	Caregiver positive reaction/coping style
							Resiliency	CDR	−.36**

^*p < .05,

^**p < .01;

^arepresents same sample, only unique variables used in second study;

^bwith caregiver uncertainty;

^cwith patient uncertainty;

^dwith maternal uncertainty;

^ewith paternal uncertainty P = patient, C = caregiver; CUIS = Children’s Uncertainty in Illness Scale; MUIS = Mishel’s Uncertainty in Illness Scale; MUIS-A = Mishel’s Uncertainty in Illness Scale—Adults; MUIS-C = Mishel Uncertainty in Illness Questionnaire—Community Form; MUIS-P = Mishel Uncertainty in Illness Questionnaire—Community Form—Parent Child Form; PECI = Parent Experience of Child Illness; PPUS = Parent’s Perception of Uncertainty in Illness Scale; PPD-C = Parental Psychological Distress in Childhood Cancer; SSERQ = Situation-specific emotional reaction questionnaire; USK = Uncertainty Scale for Kids; ASQ = Attributional Style Questionnaire (GN = Global Negative; SN-Stable Negative; IN-Internal Negative); BAI = Beck Anxiety Inventory; BAS = Boundary Ambiguity Scale; BDI = Beck Depression Inventory; BSI = Brief Symptom Inventory (GSI-Global Severity Index; Anx-Anxiety; Dep-Depression); CAMM = Child and Adolescent Mindfulness Measure; CASQ-R = Children’s Attributional Style Questionnaire—Revised; CCSS = Cognitive Control Strategies Scale; CDI = Children’s Depression Inventory; CDR = Connor-Davidson Resilience Scale; CHIP: Coping Health Inventory for Parents; CGSQ: Caregiver Strain Questionnaire; CMCRDS: Caring for my Child with a Rheumatic Disease Scale; COPE: Brief COPE; CSS: Control Strategy Scale; CVS: Child Vulnerability Scale; DCS: Decisional Conflict Scale; FACES: Family Adaptability and Cohesion Evaluation Scale (C: cohesion subscale); FNBAS: Family–Nurse Boundary Ambiguity Scale; FTP: Future Time Perspective Inventory; HAD: Hospital Anxiety and Depression Scale; HHI: Herth Hope Index; HPLP-II: The Health Promoting Lifestyle II; IDD: The Inventory to Diagnose Depression; IES: Impact of Event Scale; IES-R: Impact of Events Scale- Revised IKS: Illness Knowledge Scale; IFS: The Impact on Family Scale; IIS-P: Illness Intrusiveness Scale, Parent Report; IIRS: Illness Intrusiveness Scale; IOF: Impact on Family Scale; ISHP: Information Support from Health Care Provider Scale; ISPeerS: Information Support from Peer Scale; ISPS: Information Support from Parent Scale; IWB: Index of Well-being; JAF-P/C: Juvenile Arthritis Functional Assessment, Parent/Child Report; JCS: Jalowiec Coping Scale; MSDS: McCorckle Symptom Distress Scale; NSS: Norbeck’s Social Support Scale; PCS: Perceived Control Scale; PDS: Posttraumatic Stress Diagnostic; PEDS: Physician Estimate of Disease Status; PedsQL: Pediatric Quality of Life Inventory, Parent and Child Report; PedsQL-C: Pediatric Quality of Life Inventory Cancer Module, Parent and Child Report; PIP: Pediatric Inventory for Parents (F: Frequency, D: Difficulty); PMS: Pearlin Mastery Scale; PPS: Parent Protection Scale; PMT: Parent Mastery Test; PRISM: Pediatric Risk of Mortality Index; PRQ-85-2: Personal Resource Questionnaire Part 2; PSI-SF: Parenting Stress Index: Short Form; RCMAS: Revised Children’s Manifest Anxiety Scale; RHDS: Readiness for Hospital Discharge; RI: Reaction Index; SCL-90-r: Symptom Checklist 90- Revised; SDQ: Strengths and Difficulties Questionnaire; SF-36: SF-36 Health Survey Questionnaire; SMS: Sense of Mastery Scale; STAI: State- Trait Anxiety Inventory; STAIC: State- Trait Anxiety Inventory for Children; STAIS-S: State- Trait Anxiety Inventory- State Subscale; TRAIT: Trait Anxiety Inventory; SPCC: Symptom Pattern Scale of Children with Cancer; UCL: Utrecht Coping List.

Missing Data

Of the original 71 articles that were determined to meet inclusion criteria, 20 did not include sufficient information for every variable that they examined, so additional information was requested from the study authors. Reasons for needing additional information to calculate effect size included failure to report effect sizes or reporting results of a multivariate analysis but not the corresponding bivariate relationships. Of the 20 articles, 13 were excluded in full because of authors not responding or no longer having access to the relevant data. For eight articles, only select variables could be used, whereas others were excluded.

Data Extraction

All articles were extracted by two authors (L.S. and S.B. or A.E.) using a standardized data collection template. Data retrieved from each article included study design, measurement of uncertainty and study variables, relationship with study variables, and patient demographic characteristics. Owing to the large prevalence bias, a prevalence-adjusted bias-adjusted kappa was used (Byrt, Bishop, & Carolin, 1993). Agreement was excellent between all extractors (L.S.:S.B. = 0.92, L.S.:A.E. = 0.80). All discrepancies were discussed and resolved through consensus (L.S. and A.P.).

Quality of the Evidence

Study quality was assessed using a 20-item rating rubric that was developed for this study and was based on relevant items from validated rating tools used in observational research (DuRant, 1994; Fowkes & Fulton, 1991). Of these 20 items, six were examined separately for risk of bias that focuses specifically on the potential limitations on study design. Raters coded each item as present or not present for a total possible rating of 0 (poor quality) to 20 (excellent quality). Quality ratings were conducted by two coders (L.S. and S.B.), with 25% of the studies being coded by both raters for inter-rater reliability. Agreement was moderate between authors (κ = .80). All discrepancies were discussed and resolved through consensus (L.S. and A.P.). The rubric can be found in the supplemental materials (available online with this report). Additionally, a fail-safe N was calculated for each of the statistical analyses to examine the risk of publication bias. This value indicates the number of missing studies with nonsignificant results that would be needed for the p-value of the summary effect to no longer be significant. See Table 2 for results.

Table II.

Mean Effect Sizes for all Examined Study Variables

Study variablea	Patient reported uncertainty	N	I2	Nb	Caregiver reported uncertaintyc	N	I2	Nb
Demographic
Patient age	−.064, p = .195, 95% CI [−.160, .033]	9	53.91	1	.008, p = .131, 95% CI [−.118, .135]	6	49.49	0
Caregiver age	–	–	–	–	−.077, p = .040, 95% CI [−.149, −.004]	8	60.75	24
Family income	−.133, p = .004, 95% CI [−.220, −.044]	5	.000	4	−.077, p = .538, 95% CI [−.311, .166]	2	.000	NAd
Illness specific
Duration	.005, p = .923, 95% CI [−.093, .102]	7	34.86	0	.094, p = .202, 95%CI [−.051, .236]	8	79.17	15
Severity	.116, p = .222, 95% CI [−.071, .295]	5	71.13	2	–	–	–	–
Patient knowledge	−.060, p = .442, 95% CI [−.211, .093]	2	.000	NA	−.121, p = .557, 95%CI [−.484, .277]	2	85.36	NA
Patient psychological functioning	.370, p = .000, 95% CI [.265, .466]	17	81.85	779	.259, p = .000, 95%CI [.175, .340]	6	.000	48
					.358, p = .034, 95% CI [.028, .618]e	1	.000	NA
Anxiety	.369, p = .006, 95% CI [.113, .580]	5	84.13	52	–	–	–	–
Depression	.375, p = .000, 95% CI [.227, .507]	9	84.04	311	.261, p = .000, 95% CI [.176, .341]	6	.000	49
Psychological distress	.398, p = .003, 95% CI [.143, .603]	5	86.48	73	–	–	–	–
Caregiver psychological functioning	.259, p = .006, 95% CI [.078, .424]	5	62.41	26	.328, p = .000, 95% CI [.210, .437]	20	92.77	1048
Anxiety	–	–	–	–	.427, p = .000, 95% CI [.199, .611]	9	92.23	423
Depression	–	–	–	–	.408, p = .000, 95% CI [.250, .545]	7	88.81	357
Psychological distress	.078, p = .589, 95% CI [−.203, .347]	2	65.42	NA	.311, p = .018, 95% CI [.055, .529]	8	90.47	88
Traumatic stress	–	–	–	–	.263, p = .067, 95% CI [.−.018, .505]	7	90.20	49
Patient illness-related distress	.242, p = .000, 95% CI [.113, .364]	13	79.45	243	.179, p = .000, 95%CI [.089, .265]e	6	29.16	30
	.215, p = .007, 95%CI [.058, .362]e	3	36.34	7
Caregiver illness-related distress	.161, p = .013, 95% CI [.034, .283]	4	18.09	7	.431, p = .000, 95%CI [.329, .523]	8	80.05	578
Patient positive reaction/coping style	−.004, p = .985, 95% CI [−.418, .411]f	4	94.29	0	–	–	–	–
Caregiver positive reaction/coping style	–	–	–	–	−.299, p = .000, 95% CI [−.364, −.231]f	14	45.57	445
Patient negative reaction/coping style	.199, p = .009, 95% CI [.051, .338]g	2	.000	NA	–	–	–	–
Caregiver negative reaction/coping style	–	–	–	–	.155, p = .035, 95% CI [.011, .294]g	4	39.24	4
Caregiver vs. patient uncertainty	.294, p = .000, 95% CI [.179, .401]					3	.000	15

^aSelf-reported variable, unless otherwise noted.

^bFail-safe N: Number missing that would bring p-value > alpha of .05.

^cUncertainty was self-reported, meaning that the individual reporting was also the person experiencing the uncertainty.

^dFail-safe N cannot be calculated when there are fewer than 3 studies in the category

^eIndicates that the reporter of the study variable was different than the subject (i.e. caregiver reporting on variables specific to the patient)

^fDirectionality indicates that the more the individual used this coping style, the less uncertainty that was experienced

^gDirectionality indicates that the more the individual used this coping style, the more uncertainty that was experienced

Statistical Approach

Study analyses were conducted according to guidelines on meta-analysis with correlational data (Borenstein, Hedges, Higgins, & Rothstein, 2009), as well as modeled after other published studies that compared correlational data for effect sizes (Hood, Peterson, Rohan, & Drotar, 2009). Comprehensive Meta-Analysis (Version 2) software was used to organize data and calculate weighted random effect sizes for bivariate correlation using the following thresholds: .1 = small, .3 = medium, and .5 = large (Cohen, 1988). Effect sizes were calculated by categories established by the authors. When an individual study reported more than one variable in a particular category or two separate manuscripts were determined to be from the same sample, a mean of all variables was used to ensure that there was no overrepresentation of a single group of reporters. The random-effects model was used for all analyses because of the assumption that the true effects varied by study. Weights were assigned to studies based on sample size. Relative weights, along with forest plots for each analysis, are reported in the Supplemental Materials (available online with this report).

Variables were collapsed into the following categories for caregivers and patients independently: psychological functioning (e.g., anxiety and depression), illness-related distress (e.g., subjective pain and perception of physical functioning), and reaction/coping style (e.g., utilization of social support and mindfulness). Additionally, because of the number of studies examining four particular variables that were part of the psychological functioning category, additional analyses were conducted on the following subcategories: anxiety, depression, psychological distress, and traumatic stress (caregiver only). Two demographic variables were examined, including age and family income. Three illness-specific variables were also examined, including duration, severity, and knowledge. The demographic- and illness-specific variables were not collapsed into a single category for analyses, as it was felt that they did not represent a single domain. Finally, group differences between caregiver- and patient-report of uncertainty were examined. To best represent the data, analyses were conducted separately based on reporter of both uncertainty and the study variable. Other than when explicitly stated, the study variable was self-reported. Additionally, the uncertainty variable was always self-reported, meaning that the reporter was the individual who is experiencing the uncertainty. Results were reported if there were at least two studies with the same combination of reporters. See Table 1 for all variables used in the meta-analysis and corresponding categories.

Quantification of heterogeneity. The degree of inconsistency across studies was measured using I² because of the variability in study metrics and number of studies that were included in each independent analysis. A percentage is calculated that indicates what proportion of the observed variance is not considered to be because of error (Borenstein et al., 2009). A value near zero indicates that the variance is likely because of error, whereas a high I² indicates that the variance may be because of other factors (Higgins, Thompson, Deeks, & Altman, 2003). Although the importance of inconsistency depends on several factors, the following guide to interpretation has been established: 0–40% = might not be important, 30–60% = may represent moderate heterogeneity, 50–90% = may represent substantial heterogeneity, and 70–100% = considerable heterogeneity (Higgns & Green, 2008). For analyses with more than 10 studies and significant heterogeneity, moderator analyses were conducted to determine whether the heterogeneity was because of a measured factor (e.g., illness group).

Uncertainty Measurement

There were six unique measures of caregiver-reported uncertainty and four unique measures of patient-reported uncertainty. Caregiver uncertainty measures ranged from 2 to 34 items. There were 39 studies that measured caregiver uncertainty, and 76.9% used a scale based directly on Mishel’s conceptualization of uncertainty. Patient uncertainty measures ranged from 3 to 34 items. There were 23 studies that measured patient uncertainty, and 87.5% of these used a scale based on Mishel’s theory of uncertainty. Based on the authors’ report, 96.6% of the measures had established validity; however, only 19% reported a quantitative measure of validity. Despite the majority of studies using measurements of pediatric illness uncertainty based on Mishel’s theory, variability remained in terms of number of items even within the same measure based on adaptations by various authors and refinement of the measure over time.

Results

There were 58 studies included in this meta-analysis. The weighted mean age of the patients represented in the sample was 12.15 years with a SD of 3.29; 10 studies did not report mean age. Fifty-two studies reported gender characteristics. The percentage of male patients ranged from 21 to 82%, with a mean of 49.8%. The reporter of uncertainty or the study variable included caregiver only (56.9%), patient only (32.8%), or caregiver and patient (10.3%). Of the studies that reported caregiver relationship, 75.2% were female caregivers. The mean caregiver age was 38. years, with a SD of 6.63. Of the studies that included caregivers, 50% did not report caregiver age. Data were reported with regard to race or ethnicity in 70.7%, family income in 36.2%, and caregiver education in 50% of articles. Studies were conducted in North America (n = 40), Europe (n = 7), Asia (n = 9), and Australia (n = 2). Eleven studies included two or more disease groups. Of the studies that included one disease group, specific diseases were as follows: cancer (n = 26), asthma (n = 7), arthritis (n = 4), diabetes (n = 2), epilepsy (n = 2), cardiac disease (n = 2), cystic fibrosis (n = 1), sexual disorder (n = 1), mitochondrial disease (n = 1), and chromosomal disorder (n = 1). See Table 1 for additional demographic information. Full meta-analytic results are displayed in Table 2.

Study Quality

The mean quality rating across studies was 15.6 out of a possible 20 items (SD = 1.9; range = 12–19). Of the 58 studies, 48.3% stated clear hypotheses, 73.3% reported inclusion/exclusion criteria, and 88.3% used at least one validated measure for a study variable. Power calculations were only reported in 13.3% of the studies.

Risk of bias. Of the 58 total studies, results were as follows: 5 (8.6%) reported on all 6 items, 18 (31%) reported on 5, 24 (41.4%) reported on 4, 10 (17.2%) reported on 3, and 1 (1.7%) reported on 2 of the 6 items. The item that was most commonly missed was having a strong recruitment rate (i.e., ≥90%), which was only met by 10% of the studies. The two items that were most commonly met were having a study sample that was representative of the population (95%) and using a validated measure of uncertainty (93.3%). Fail-safe N’s were calculated for each analysis. For patient-reported analyses, fail-safe N’s ranged from 0 to 779, and for caregiver-reported analyses, they ranged from 0 to 1,048. As expected, the majority of the very low fail-safe N’s were associated with summary effect sizes that were not significant. Based on the small number of studies in the individual analyses, many of the fail-safe N’s indicated that results are likely meaningful. For example, if only five studies existed to be included in a summary analysis, it would be unlikely that there were 50 unpublished nonsignificant studies. This holds particularly true for the higher values (i.e., >100). See Table 2 for full results.

Heterogeneity

Substantial to considerable heterogeneity existed across the majority of the analyses run. Results for I² statistics are found on Table 2. Additional analyses to compare the within- and between-groups heterogeneity based on disease groups revealed no differences between cancer and other diseases. Results from these analyses can be found in Supplemental Table 1 (available online with this report).

Demographic

Findings were inconsistent with regard to age and income factors depending on the uncertainty reporter. Patient age was not related to either patient or caregiver uncertainty. However, younger caregiver age was related to higher levels of caregiver uncertainty but not patient uncertainty. In contrast, lower family income was significantly related to patient uncertainty; however, it was not related to caregiver uncertainty. Effect sizes were all small.

Illness Specific

Illness duration, severity, and knowledge were not significantly related to either patient- or caregiver-reported uncertainty. Although there were no sufficient data available to break this down by illness type, the following illnesses were represented in one of these variables: allergies, arthritis, asthma, cancer, cardiac disease, cystic fibrosis, diabetes, general medical/surgical, neurological illness, and rare chromosomal disorders.

Psychological Functioning

As predicted, patient psychological functioning was significantly related to both patient- and caregiver-reported uncertainty, indicating that increased psychological functioning was related to decreased uncertainty. When specific psychological functioning variables were examined further, the same pattern held true for anxiety, depression, and psychological distress.

Similarly, caregiver psychological functioning and all examined domains (i.e., anxiety, depression, psychological distress, and posttraumatic stress) were significantly related to caregiver uncertainty, with higher levels of psychological functioning being related to lower amounts of uncertainty. Overall, caregiver psychological functioning was also related to patient uncertainty; however, specific psychological functioning domains were either not present in the literature or were not significantly related to patient uncertainty. Effect sizes for psychological functioning variables for both patients and caregivers were primarily in the medium range.

Illness-Related Distress

Illness-related distress was significantly related to both patient and caregiver uncertainty regardless of reporter and subject with small to medium effect sizes. This relationship was in the expected direction, with higher levels of illness-related distress being related to higher levels of uncertainty.

Reaction/Coping Style

The method used by the patient or caregiver to cope with the various aspects of the child’s illness was categorized as either positive or negative, which was determined based on the conceptualization by the authors of each study. When it was not specified by the author to be a positive versus negative style, the broader literature was used to make a determination. Both patient and caregiver positive and negative reaction/coping styles were related to uncertainty in the expected directions, with one exception: patient positive reaction/coping style was not significantly related to uncertainty. Otherwise, a greater degree of positive reaction/coping style was related to less uncertainty, and a greater degree of negative reaction/coping style was related to more uncertainty for both patients and caregivers. Effect sizes were small to medium.

Caregiver Versus Patient Uncertainty

There was a significant relationship between patient and caregiver uncertainty with a medium effect size.

Discussion

This is the first meta-analysis to synthesize the current state of the quantitative literature on pediatric illness uncertainty. A total of 58 articles spanning the past three decades were included. Articles examined the associations between patient or caregiver uncertainty and a wide array of demographic, illness, and psychosocial functioning and coping variables. Consistent with hypotheses, uncertainty was significantly associated with self-reported psychological outcomes for both patients and caregivers. Specifically, patient-reported uncertainty was associated with patient-reported general psychological functioning, as well as anxiety, depression, and psychological distress independently. Caregiver’s uncertainty about their child’s illness was also associated with caregiver general psychological functioning and coping style. Especially, notable effect sizes were observed for relationships between caregiver’s uncertainty regarding their child’s illness and caregiver anxiety and depression. Finally, important transactional relationships were also demonstrated, as caregiver uncertainty about their child’s illness was associated with the child’s psychological functioning. Although it cannot be concluded from these findings that illness uncertainty leads directly to poorer psychological functioning, the literature certainly suggests that the illness uncertainty could impact on emotional well-being in the context of pediatric chronic illness (Mullins et al., 2012).

Illness-related distress (e.g., perceived levels of pain, perceptions of the illness interfering with life events, and perceived physical functioning) was also significantly related to illness uncertainty regardless of the subject of the study variable or uncertainty; however, effect sizes were smaller than the relationships between illness uncertainty and psychological functioning.

Finally, specific reactions and coping styles relative to illness uncertainty suggest that caregivers using positive reaction and coping styles experienced less uncertainty. Although this finding was not replicated with patient-reported uncertainty, both patients and caregivers did experience more uncertainty when they used more negative reaction and coping styles. These findings suggest the potential impact of learning and using more efficacious coping methods on illness uncertainty. However, it is important to consider that these groups were formed based on each study’s individual definition of positive and negative coping styles, and these are not absolute categories, as some coping strategies may be effective in certain circumstances and ineffective in others (Cheng, Lau, & Chan, 2014).

Despite the evidence suggesting that pediatric illness uncertainty is related to psychological outcomes, the intervention literature is nascent. Three pilot studies have targeted illness uncertainty (Burke et al., 2014; Hoff et al., 2005; Mullins et al., 2012), and findings support the notion that illness uncertainty is modifiable and has an impact on various aspects of psychological functioning. Additionally, there are specific factors that have been found to be associated with uncertainty that may suggest a role for intervention across family members. For example, increased feelings of control (Lipinski, Lipinski, Biesecker, & Biesecker, 2006) and a sense of mastery (Mu, Ma, Ku, Shu, Hwang, & Kuo, 2001) have been associated with decreased amounts of uncertainty in caregivers of children with illnesses. Other aspects such as caregiver utilization of social support have had mixed findings (Franck et al., 2014; Lee, Yoo, & Yoo, 2007). Although there is some intervention literature showing increases in patient’s use of effective coping skills, increased feelings of control (Hoff et al., 2002) and utilization of mindfulness strategies (Patterson & McDonald, 2015) are associated with decreased uncertainty. Finally, there has been some evidence in the adult literature to suggest that illness uncertainty can be modified, and intervention has a positive impact on patient’s cognitions, knowledge, and coping skills (Gil et al., 2006). This evidence from the pediatric and adult literature suggests that illness uncertainty may be an important cognitive process for both patients and their family members and related to various aspects of functioning and coping; however, the absence of longitudinal studies and sufficiently powered clinical trials prohibits definitive conclusions about the potential of illness uncertainty as an intervention target or conclusive recommendations for clinical practice.

Certainly, one challenge with this meta-analysis was the large number of variables of interest and outcomes assessed. These variables were kept independent to best represent the data. As would be expected, effect sizes were typically larger when the same individual was both the reporter of uncertainty and the related study variables. However, despite a fewer number of studies examining the relationship between patient-specific variables and caregiver uncertainty, there is initial evidence to suggest that caregiver uncertainty does have a relationship with patient outcomes. Few studies used caregiver proxy report of psychological functioning and coping outcomes. This may strengthen the evidence, given that most of the study variables are internal processes and not observable, which tend to be less reliable across informants (De Los Reyes & Kazdin, 2005), but may also result in inflated effect sizes because of shared variance (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003) that may be remedied by having multiple reporters.

Findings from this investigation should be considered in the context of a number of methodological limitations. First, several of the authors who were unable to provide the necessary data to be included in analyses had initially published findings on variables with nonsignificant relationships with uncertainty. Second, effect sizes were calculated using correlational data. Therefore, the effect sizes observed cannot be interpreted as causal relationships, instead must only be interpreted as associations. Additional longitudinal and intervention data are needed to identify potential causal relationships between illness uncertainty and psychological functioning and coping. Additionally, there was significant heterogeneity in study outcomes. A large number of analyses were conducted, and each individual analysis had a relatively few number of studies contributing. This limited the subgroup analyses that could be conducted to more fully consider the heterogeneity of the findings. For example, understanding the differences in measurement of uncertainty or between illness populations would add to the literature but could not be conducted. Finally, it is necessary to consider the definition and measurement of pediatric illness uncertainty. Given that the majority of the published literature is based on Mishel’s theory of uncertainty and corresponding measures, it is difficult to fully understand the ways in which those studies can be incorporated with the remaining literature to develop one conceptual definition.

Despite these limitations, this meta-analysis is the first to provide a consolidated summary of the uncertainty literature in pediatrics. The findings strongly indicate that additional research to understand the construct of illness uncertainty and the robust relationship between uncertainty and psychological functioning for both patients and caregivers is warranted. Although there are currently a limited number of intervention studies targeting uncertainty, the findings are promising and suggest that uncertainty can be managed with a likely impact on psychological functioning for both caregivers and patients. Further research is needed to better understand the causality of the relationships between illness uncertainty and the study variable of interest, as well as to develop interventions to guide providers in addressing this modifiable variable.

Supplementary Data

Supplementary data are available at Journal Of Pediatric Psychology online.

Funding

The conduct of the study; collection, management, analysis, and interpretation of the data; and preparation of the manuscript were supported in part by grant T32HD068223 for L.S.

Conflicts of interest: None declared.