Guideline on HiV disclosure counsellinG for cHildren up to 12 years of aGe - appendices

(1)

hiv/aids Programme

Guideline on HiV disclosure

counsellinG for cHildren up to

12 years of aGe - appendices

(2)

WHO/HIV/11.05

© World Health Organization 2011. All rights reserved.

(3)

Database Search Strategy

PubMed

#1: ((“HIV Infections”[MeSH] OR “HIV”[MeSH])

#2: (((“Disclosure”[MeSH] OR “Self Disclosure”[MeSH]) OR

“non-disclosure”[All Fields] OR “non disclosure”[All Fields]) OR

“Counseling”[MeSH] OR “counselling”[All Fields] OR “truth telling”[All Fields] OR “truth-telling”[All Fields] OR “bad news”[All Fields] OR

“diagnosis disclosure”[All Fields]))

#3: ((“Child”[MeSH] OR “Child, Preschool”[MeSH] OR “children”[tw] OR

“child”[All Fields]) OR “Pediatrics”[MeSH] OR “pediatr”[All Fields] OR

“paediatr”[All Fields] OR “enfant”[All Fields])

#4: #1 AND #2 AND #3

Limits: Published before 7/1/2010 Total Citations: 1081

Embase

#1: ‘human immunodeficiency virus infection’/exp/mj OR ‘human immunodeficiency virus infection’/exp

Results: 241,151

#2: ‘hiv’/exp OR ‘hiv’ OR ‘hiv infection’/exp OR ‘hiv infection’

Results: 308,736

#3: ‘acquired’ AND ‘immune’ AND ‘deficiency’ AND (‘syndrome’ OR

‘syndrome’/exp OR syndrome) Results: 116,063

#4: #1 OR #2 OR #2 Results: 310,559

#5: enfant* OR pediatr* OR paediatr*

Results: 952,181

#6: ‘children’ OR ‘children’/exp OR children OR ‘child, preschool’/exp OR ‘child, preschool’ OR ‘child’ OR ‘child’/exp OR child

Results: 1,993,247

#7: #5 OR #6 Results: 2,294,314

#8: ‘nondisclosure’ OR ‘non-disclosure’ OR ‘bad news’ OR ‘diagnosis disclosure’ OR ‘truth telling’

Results: 2,142

#9: ‘counseling’ OR ‘counseling’/exp OR counseling OR ‘disclosure’ OR

‘disclosure’/exp OR disclosure Results: 361,097

#10: #8 OR #9 Results: 362,019

#11: #4 AND #7 AND #10 Results: 2,491

#12: #4 AND #7 AND #10 AND [humans]/lim Results: 2,316

aPPENdiX a: sEaRCh sTRaTEGY

(4)

Guideline on HiV disclosure counsellinG for cHildren up to 12 years of aGe

Embase (continued)

#13: #4 AND #7 AND #10 AND [humans]/lim AND ([internal medicine]/lim OR [pediatrics]/lim OR [public health]/lim) AND ([preschool]/lim OR [school]/lim OR [child]/lim)

Total Citations: 436

PsychINFO

S1: HIV infections or HIV infect* or human immunodeficiency virus or human immunedeficiency virus or human immuno-deficiency virus or human immune-deficiency virus or acquired immunedeficiency syndrome or acquired immuno-deficiency syndrome or VIH or SIDA Results: 27,964

S2: children or child or childpediatric or paediatric or paediatrics or pediatrics or paed or ped*

Results: 502,945

S3: disclosure or non-disclosure or non disclosure or counseling or counselling or truth telling or truth-telling or bad news or diagnostic disclosure or self-disclosure or self disclosure or diagnosis disclosure Results: 149,138

S4: S1 and S2 and S3

Limits: Published before 7/1/2010 Total Citations: 476

Web of Science

Topic=(HIV infections OR HIV OR HIV infect* OR human

immunodeficiency virus OR human immune deficiency virus OR human immuno-deficiency virus OR human immune-deficiency virus OR acquired immune deficiency syndrome OR acquired immuno-deficiency syndrome OR VIH OR SIDA) AND Topic=(non-disclosure OR

nondisclosure OR bad news OR diagnosis disclosure OR truth telling OR counseling OR disclosure) AND Topic=(children OR child, preschool OR child OR enfant* OR pediatr* OR paediatr*)

Timespan=1981-2010. Databases=SCI-EXPANDED, SSCI, A&HCI.

Total Citations: 549

CINAHL

S1: MH “HIV Education” or MH “HIV Enteropathy” or MH “HIV Infections+” or MH “HIV Seronegativity” or MH “HIV Seropositivity” or MH “HIV Wasting Syndrome” or MH “HIV-AIDS Nursing” or MH “HIV- Infected Patients+”

Results: 41,803

S2: TX truth or TX disclosure or TX non-disclosure or TX bad news or TX counselling or TX counseling or MH “Truth Telling (Iowa NIC))” or MH

“Truth Disclosure”

Results: 37,444

S3: Limiters - Age Groups: Fetus, Conception to Birth, Infant, Newborn:

birth-1 month, Infant: 1-23 months, Child, Preschool: 2-5 years, Child:

6-12 years Results: 220,473

S4: S1 and S2 and S3, Limit 2009-2010

Total Citations: 161

(5)

Social Work Abstracts

S1: ab disclosure or de disclosure or ti disclosure or counseling or counselling

Results: 3,678

S2: DE “HIV-positive persons*” or ti hiv or ab hiv Results: 1,127

S3: child or children Results: 21,909 S4: S1 and S2 and S3 Total Citations: 165 African Index

Medicus

HIV AND disclosure Total Citations: 3 LILACS HIV AND disclosure

Total Citations: 15 WHOLIS HIV AND disclosure

Total Citations: 7

WPRIM HIV AND disclosure

Total Citations: 2

IMSEAR HIV AND disclosure

Total Citations: 7

Total Citations: 11

(6)

a PPEN di X B Do H IV positive c hildr en 12 years and under whose H IV st atus is disclosed to them display equal or gr eater well-being than those childr en who ar e not disclosed to? G RADE T AB LE S: Author(s): Date: 2010-11-21 Question: Do H IV positive c hildren 12 years and under whose H IV status is disclosed to them display equal or greater well-being than those c hildren who are not disclosed to? Settings: Children in H IV treatment settings throughout the world Bibliography: W hat are the health, treatment, psyc hological, and social outcomes for H IV positive c hildren ages up through 12 who do and do not have their H IV status disclosed to them. Coc hrane Database of Systematic Reviews [Y ear], Issue [Issue].

(7)

Quality assessment Summary of Findings Import ance No of patients Effect Quality No of studies Design Limit ations Inconsist - ency Indirect - ness Imprecision Other considera - tions

Disclosed to versus Not disclosed to Control Relative (95% C I) Absolute

Death (follow-up mean 3 years; Clinic report) 1observational studies

no serious limitations

no serious inconsistencyserious1

no serious imprecision

strong association2 dose response gradient311/225 (4.9%)

11/100 (11%)

RR 0.44 (0.2 to 0.99)

62 fewer per

1000 (from 1 fewer to 88 fewer)

㊉㊉㊉〇 MODERATECRITICAL CD4 decline (follow-up mean 3 years; CD4 decline of 50 or more) 1observational studies

dose response gradient4

38/220 (17.3%) 20/94 (21.3%)

RR 0.81 (0.5 to 1.32)

40 fewer per 1000 (from 106 fewer to 68 more)

㊉㊉〇〇 LOWCRITICAL Adherence to medication regimen (follow-up median 2.5 weeks; pill-count, self-and caregiver-report, refills) 5observational studies5

no serious inconsistency

no serious indirectness no serious imprecision

6none4,6

902/1311 (68.8%) 606/1040 (58.3%)

RR 1.18 (0.88 to 1.57)

105 more per 1000 (from 70 fewer to 332 more)

㊉㊉〇〇 LOWCRITICAL Psychological distress by self-report (Child self report in semi-structured interview and self-report survey) 2observational studies

7none14/54 (25.9%)

50/103 (48.5%)

RR 0.64 (0.41 to 1)

175 fewer per 1000 (from 286 f

ewer to 0 more)

㊉㊉〇〇 LOWIMPORTANT Depression/dysthymia (measured with: Standardized measures: Caregiver symptom interview tied to DSM-IV and Children’s Depression Inventory8; Better indicated by lower values) 2observational studies

reduced effect for RR >> 1 or RR << 18,963106-M

D 0.67 higher (0.15 lower to 1.48 higher)

10㊉㊉㊉〇 MODERATEIMPORTANT General anxiety (measured with: Standardized measure: Caregiver symptom interview tied to DSM-IV8; Better indicated by lower values) 1observational studies

reduced effect for RR >> 1 or RR << 18,94377-MD 0.7 higher

(0.32 lower to 1.72 higher)

11㊉㊉㊉〇 MODERATEIMPORTANT Separation anxiety (measured with: Standardized measure: Caregiver symptom interview tied to DSM-IV; Better indicated by lower values) 1observational studies

reduced effect for RR >> 1 or RR << 18,94377-MD 1 higher

11㊉㊉㊉〇 MODERATEIMPORTANT 1.16% of study population of youth is 12 and under 2.Relative risk is <.5 3.Children who were disclosed to received more CD4 tests, i.e. more monitoring 4.Age-adjusted effect in largest study (Williams et al., 2008) 5.Odds for all adherence studies calculated by general inverse procedures 6.In largest study, adherence report is validated by CD4 (Williams et al., 2008) 7.Child self-report about child distress in face-to-face or self-report interview in care setting. Thus social desirability

responding is likely, but equally so for both those disclosed to and those not 8.In Gadow et al. (2008) means for 6-11 year old children are adjusted for gender, caregiver education, caregiver- reported symptoms, life stressors, and caregiver relationship to youth, e.g. biologiical parent, since non-biological parents more likely to disclose to youth 9.Rates of psychological diagnoses and disclosure both increase with age 10.Not statistically significant in Mellins et al. (2002); statistically significant difference in Gadow et al. (2008) 11.Not statistically significant in Gadow et al. (2008)

(8)

S U M MAR Y OF F IN DI NG S T AB LE: Health, and mental health outcomes for H IV positive c hildren ages up through 12 who do and do not have their H IV status disclosed to them Patient or population: H IV positive c hildren ages up through 12 who do and do not have their H IV status disclosed to them Settings: Children in H IV treatment settings in multiple countries Intervention: Disclosed to versus Not disclosed to (Control) Outcomes Illustrative comparative risks* (95% C I) Relative effect (95% C I)

No of Partici - pants (studies)

Quality of the evidence (G RAD E) Comments As - sumed risk Corresponding risk

ControlDisclosed to versus Not disclosed to Death Clinic report Follow-up: mean 3 years

110 per 1000

48 per 1000 (22 to 109)RR 0.44 (0.2 to 0.99)

325 (1 study)㊉㊉㊉〇 moderate 1,2,3,4Romania (Ferris, 2007) CD4 decline CD4 decline of 50 or more Follow-up: mean 3 years

213 per 1000

173 per 1000 (106 to 281)RR 0.81 (0.5 to 1.32)

314 (1 study)㊉㊉〇〇 low1,3,5Romania (Ferris, 2007) Adherence to medication regimen pill-count, self-and caregiver- report, refills Follow-up: median 2.5 weeks

583 per 1000

688 per 1000 (513 to 915)RR 1.18

(0.88 to 1.57)

2351 (5 studies7)㊉㊉〇〇 low5,6

Italy, Togo, Uganda, United States (Giacomet, 2003; Polisset, 2009;Nabu-

keera-Barungi, 2007; Bikaako- Kajura, 2006; Williams, 2006) Psychological distress by self-report

Child self report in semi-structured interview and self-report survey

485 per 1000

310 per 1000 (199 to 485)RR 0.64 (0.41 to 1)

157 (2 studies)㊉㊉〇〇 low8

France, Zambia (Funck- Brentano, 1997; Menon, 2007) Depression/dysthymia

Standardized measures: Caregiver symptom interview tied to DSM-IV and Children's 9Depression Inventory

The mean Depression/ dysthymia in the intervention groups was 0.67 higher (0.15 lower to 1.48 higher)10

169 (2 studies)㊉㊉㊉〇 moderate9,11

United States (Mellins, 2002; Gadow

, 2010) General anxiety Standardized measure: Caregiver symptom interview tied to DSM-IV9

The mean General anxiety in the intervention groups was

0.7 higher (0.32 lower to 1.72 higher)12

120 (1 study)㊉㊉㊉〇 moderate9,11

United States (Gadow

, 2010) Separation anxiety Standardized measure: Caregiver symptom interview tied to D

SM-IV

The mean Separation anxiety in the intervention groups was 1 higher (0.08 lower to 2.08 higher)12

120 (1 study)㊉㊉㊉〇 moderate9,11

United States (Gadow

, 2010)

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio; GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate. 1.16% of study population of youth is 12 and under 2.Relative risk is <.5 3.Effect persisted in regression after age accounted for 4.Children who were disclosed to received more CD4 tests, i.e. more monitoring 5.Age-adjusted effect in largest study (Williams et al., 2008) 6.In largest study, adherence report is validated by CD4 (Williams et al., 2008) 7.Odds for all adherence studies calculated by general inverse procedures 8.Child self-report about child distress in face-to-face or self-report interview in care setting. Thus social desirability responding is likely, but equally so for both those disclosed to and those not 9.In Gadow et al. (2008) means for 6-11 year old children are adjusted for gender, caregiver education, caregiver-reported symptoms, life stressors, and caregiver relationship to youth, e.g. biologiical parent, since non-biological parents more likely to disclose to youth 10.Not statistically significant in Mellins et al. (2002); statistically significant difference in Gadow et al. (2008) 11.Rates of psychological diagnoses and disclosure both increase with age 12.Not statistically significant in Gadow et al. (2008)

(9)

a PPEN di X C D o H IV positive c hildren 12 years and under whose H IV status is disclosed to them display equal or greater well-being pre- or post- disclosure? G RADE T AB LE S: Author(s): Date: 2010-11-23 Question: Do H IV positive c hildren 12 years and under display benefits P re- to P ost-disclosure Settings: Clinic-attending H IV positive c hildren Bibliography: W hat are the health, treatment, psyc hological, and social outcomes for H IV positive c hildren ages up through 12 who do and do not have their H IV status disclosed to them. Coc hrane Database of Systematic Reviews [Y ear], Issue [Issue].

1 Data extends 3 years pre-disclosure and 3 years post-disclosure 2 Subscale of the General Health Assessment for Children 3 Quality of life subscales were based on caregiver report. Caregivers tend to give less reliable reports for older children as compared to younger children. 4 Unadjusted results reported here. Please see additional table for adjusted results. 5 Results plotted over time 3 years pre- and post-disclosure, adjusted effect for interaction of disclosure point and time since disclosure. 6 Longitudinal study of 395 youth in multi-site clinical trial

(10)

Guideline on HiV disclosure counsellinG for cHildren up to 12 years of aGe Quality assessmentSummary of Findings

Import ance No of patients Effect Quality No of stud - ies Design Limit ations Inconsist - ency Indirect - ness Imprecision Other considerations Disclosed to versus Not disclosed to Control Relative (95% C I) Absolute

General health perception (follow-up mean 3 years1; measured with: General health perception2; range of scores: 0-100; Better indicated by higher values) 1observational studies no serious limitations

reduced effect for RR >> 1 or RR << 14 dose response gradient53956395-MD 0.1 lower

4㊉㊉㊉〇 MODERATEIMPORTANT Symptom distress (follow-up mean 3 years1; measured with: Symptom distress2; range of scores: 0-100; Better indicated by higher values) 1observational studies

㊉㊉㊉〇 MODERATEIMPORTANT Physical functioning (follow-up mean 3 years1; measured with: Physical functioning2; range of scores: 0-100; Better indicated by higher values) 1observational studies

㊉㊉㊉〇 MODERATEIMPORTANT Health care utilization (follow-up mean 3 years1; measured with: Health care utilization2; range of scores: 0-100; Better indicated by higher values) 1observational studies

7reduced effect for RR >> 1 or RR << 14 dose response gradient53956395-MD 0.5 lower (2.42 lower to 1.42 higher)㊉㊉㊉㊉ HIGHIMPORTANT Psychological status (follow-up mean 3 years1; measured with: Psychological status2; range of scores: 0-100; Better indicated by higher values) 1observational studies

㊉㊉㊉〇 MODERATEIMPORTANT Social/role functioning (follow-up mean 3 years1; measured with: Social/role functioning2; range of scores: 0-100; Better indicated by higher values) 1observational studies

reduced effect for RR >> 1 or RR << 14 dose response gradient53956395-MD 0.2 higher

4㊉㊉㊉〇 MODERATEIMPORTANT Experience of social isolation/stigmatization pre- to post-disclosure 0no evidence availablenone0/0 (0%)0/0 (0%) RR 0 (0 to 0)

0 fewer per 1000 (from 0 fewer to 0 fewer) 0%0 fewer per 1000 (from 0 fewer to 0 fewer)

(11)

S U M MAR Y OF F IN DI NG S T AB LE: Pre-disclosure to Post-disclosure well-being for H IV positive c hildren ages up through 12 who have their H IV status disclosed to them Patient or population: H IV positive c hildren ages up through 12 who have their H IV status disclosed to them Settings: Clinic-attending H IV positive c hildren Intervention: P re-disclosure versus P ost-disclosure Outcomes Illustrative comparative risks* (95% C I) Rela - tive effect (95% C I)

No of Partici - pants (studies) Quality of the evidence (G RAD E) Comments

As- sumed risk

Corresponding risk

Con- trolDisclosed to versus Not disclosed to

General health per

ception

General health perception

1. Scale from: 0 to 100.Follow-up: mean 3 years2

The mean General health perception in the post-disclosure group was 0.1 lower (.61 lower to .41 higher)3

395 (1 study)㊉㊉㊉〇 moderate3,4,5

This set of studies represents the work of Butler et al. summarized in a 2009 publication. Sites in

the prospective longitudinal study represent multiple cities on the US mainland and in Puerto Rico. Symptom distress 1Symptom distress. Scale from: 0 to 100. 2Follow-up: mean 3 years

The mean Symptom distress in the post-disclosure group was 0.10 lower (.65 lower to 0.45 higher)

395 (1 study)㊉㊉㊉〇 moderate3,4,5 Physical functioning Physical functioning1.

Scale from: 0 to 100. Follow-up: mean 3 years

2

The mean Physical functioning in the intervention groups was 0.6 lower (9.11 lower to 7.91 higher)

395 (1 study)㊉㊉㊉〇 moderate3,4,5 Health care utilization Health care utilization1. Scale from: 0 to 100. Follow-up: mean 3 years2

The mean Health care utilization in the intervention groups was 0.5 lower (2.42 lower to 1.42 higher)

395 (1 study)㊉㊉㊉㊉ high3,4,5 Psychological status Psychological status1. Scale from: 0 to 00. Follow-up: mean 3 years2

The mean Psychological status in the post-disclosure group was 0.20 lower (0.79 lower to 0.39 higher)

395 (1 study)㊉㊉㊉〇 moderate3,4 Social/role functioning Social/role functioning1.

Scale from: 0 to 100. Follow-up: mean 3 years

2

The mean Social/role functioning in the post-disclosure group was 0.2 higher (1.39 lower to 1.79 higher)3

395 (1 study)㊉㊉㊉〇 moderate3,4,5

Experience of social isolation/stigmatization pr No studies availableSee comment e- to post-disclosure

While negative life events are often

used as an independent predictor of well-being, no study e

xamined

variation in their occurrence pre- to post-e

xposure for children with their HIV status disclosed to them

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio; GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate. 1.Subscale of the General Health Assessment for Children 2.Data extends 3 years pre-disclosure and 3 years post- disclosure 3.Unadjusted results reported here. Please see additional table for adjusted results. 4.Quality of life subscales were based on caregiver report. Caregivers tend to give less reliable reports for older children as compared to younger children. 5.Results plotted over time 3 years pre- and post-disclosure, adjusted effect for interaction of disclosure point and time since disclosure.

(12)

aPPENdiX d: ChiLd REaCTiONs TO disCLOsURE

Child-reported immediate reactions to disclosure

Summarized from Blasini et al. (2004), Mellins (2002), Oberdorfer(2006), and Wiener (1996) representing Puerto Rico, Thailand, and the United States

Child-reported reactions 6-months post disclosure intervention

Summarized from Blasini et al. (2004), Puerto Rico.

Reaction N

(156) %

Sad/worried/depressed 56 35.90%

Neutral/normal 44 28.21%

Shock/surprise 25 16.03%

Angry/rebellious 23 14.74%

Confused 13 8.33%

Positive, e.g. relief 10 6.41%

Reaction N

(40) %

Sad/worried/depressed 4 10.00%

Neutral/normal 28 70.00%

Shock/surprise 5 12.50%

Angry/rebellious 0 0.00%

Confused 0 0.00%

Positive, e.g. relief 19 47.50%

(13)

Effect sizes adjusted for covariates pre-disclosure to post-disclosure

1. Butler et al. (2009), N=395

2. Positive effect estimates indicate improvement over time.

3. Generally, estimates <.40 are considered small, .40-.70 moderate, and greater than .70 large.

4. Model is adjusted for age, gender, race/ethnicity, CD4%, viral load, CDC classification, primary caregiver, primary caregiver education, antiretroviral regimen, number of hospitalizations since last visit, negative life event score, time since disclosure and the interaction between disclosure and time since disclosure. Although a range of 72-94 is presented, the measuring instruments have a range of 0-100.

Measure of outcomes 3 years before and 3 years after disclosure (Butler et al, 2009)

⁴³

Effect of Disclosure Effect of Time by Disclosure

Domain Estimate p Estimate p

General health perception .410 .70 -.040 .91

Symptom distress .588 .31 .170 .41

Physical functioning .536 .79 -.333 .63

Health care utilization -.275 .61 .075 .64

Psychological status .005 .999 .448 .23

Social/role functioning .380 .69 .002 .999

men other than HAART with PIs had the lowest average scores. Subjects with more previous hospitalizations had significantly lower general health perception, and those with more negative life events had significantly lower symptom scores and physical functioning. Other co- variates were not associated with any of the QoL domains but were retained in the model to control for confounding.

DISCUSSION

The results of this investigation suggest that disclosure did not significantly affect QoL in a population of peri- natally, HIV-infected children and adolescents. Although

primary caregivers reported lower QoL scores after disclo- sure for all of the domains except social/role functioning, these differences were not significant, even after adjust- ment for demographic and clinical factors. Current recom- mendations regarding diagnostic disclosure to children with HIV/AIDS are based on lessons learned from pedia- tric oncology. However, the psychological benefits of disclosure apparent among youth with cancer may be off- set by HIV-related stigma and discrimination toward the infected child and other family members.

¹⁶

A more com- prehensive understanding of the effect of HIV disclosure on QoL is important for developing appropriate strategies for disclosure to children and adolescents with HIV.

FIGURE 3

Predicted scores for 6 QoL domains for 395 subjects with both predisclosure and postdisclosure QoL assessments. All of the models are adjusted for age (5–9, 10 –11, 12–13, and14 years), gender, race/ethnicity (white non-Hispanic/other, black non-Hispanic, or Hispanic), CD4% (14%, 15%–24%, or25%), viral load (400 or400 copies per mL), CDC class (N/A/B or C), primary caregiver (biological parent, other relative, or other/unknown), primary caregiver education level (grade 1–11 or high school graduate or higher), antiretroviral regimen (HAART with PI, other therapy, or not on ART), number of hospitalizations since last visit (0, 1, or2), negative life-event score (0, 1, or2), disclosure, time since disclosure, and the interaction between disclosure and time since disclosure. Time of disclosure is denoted by the line at month 0.

TABLE 3 Unadjusted and Adjusted Effects of Disclosure on QoL Scores

Domain Unadjusted for Covariates^a Adjusted for Covariates: Full Model^b Effect of

Disclosure

Effect of Time by Disclosure

Effect of Disclosure

Effect of Time by Disclosure

Estimate P Estimate P Estimate P Estimate P

General health perception 0.546 .61 0.076 .82 0.410 .70 0.040 .91

Symptom distress 0.588 .30 0.154 .45 0.588 .31 0.170 .41

Psychological status 0.164 .87 0.488 .20 0.005 0.999 0.448 .23

Physical functioning 0.686 .73 0.283 .68 0.536 .79 0.333 .63

Social/role functioning 0.423 .65 0.004 .99 0.380 .69 0.002 0.999

Health care utilization 0.261 .63 0.079 .63 0.275 .61 0.075 .64

Data are based on ﬁtting a mixed-effects model for each domain. The effects of disclosure represent the difference in QoL scores between predisclosure and postdisclosure visits. Negative values indicate poorer QoL scores after disclosure.

aAll of the models have effects of disclosure adjusted for time since disclosure (6-month intervals) and the interaction between disclosure and time since disclosure.

bFull model is adjusted for age (5–9, 10 –11, 12–13, and14 years), gender, race/ethnicity (white non-Hispanic/other, black non- Hispanic, or Hispanic), CD4% (14%, 15%–24%, or25%), viral load (400 or400 copies per mL), CDC class (N/A/B or C), primary caregiver (biological parent, other relative, or other/unknown), primary caregiver education level (grade 1–11 or high school graduate or higher), antiretroviral regimen (HAART with PI, other therapy, or not on ART), number of hospitalizations since last visit (0, 1, or2), negative life-event score (0, 1, or2), time since disclosure, and the interaction between disclosure and time since disclosure.

940 BUTLER et al

. Provided by Health Professions on April 28, 2010 www.pediatrics.org

Downloaded from

(14)

a PPEN di X E D o c hildren 12 years and under who get the H IV positive status of their parent/caregiver disclosed to them display equal or greater well-being than those c hildren who are not disclosed to? G RADE T AB LE S: Author(s): Date: 2010-12-16 Question: Do c hildren 12 years and under who get the H IV positive status of their parent/caregiver disclosed to them display equal or greater well-being than those c hildren who are not disclosed to Settings: Community and clinic settings Bibliography: Health, mental health and social eff ects of disclosure of caregiver H IV to c hildren 12 and under . Coc hrane Database of Systematic Reviews [Y ear], Issue [Issue]. Coc hrane Database of Systematic Reviews [Y ear], Issue [Issue].

1 Although interviews concerned over 50 children, reports were made by only 10 HIV positive parents/caregivers. 2 Assessment of testing was based on caregiver report rather than documentation of testing. 3 Large effect in total sample. Relative risk is 12.14 if sample is limited to children 12 and under. 4 No caregiver who had not disclosed had their child tested. 5 Adherence was measured by caregiver report, unverified by biological outcomes. 6 Cross-sectional retrospective interview of HIV positive caregivers. 7 Data is based on interview of caregiver rather than interview of both caregiver and child. 8 Relative risk is statistically significant, but not reported in article. 9 Widely used Child Behavior Checklist forms the basis for these 2 studies. 10 Assessment is based on caregiver report of child’s problems. Caregiver reports vary as children age. 11 Bauman study uses age as a covariate; problems do not differ then by disclosure group. 12 Widely used adult depression inventory, the Hamilton Depression Inventory

(15)

Quality assessmentSummary of Findings

Import ance No of patients Effect Quality No of stud - ies Design Limit a- tions Inconsist - ency Indirect - ness Imprecision Other consid - erations Disclosed to versus Not disclosed to Control Relative (95% C I) Absolute

Child Tested 1observational studiesserious1no serious inconsistency no serious indirectness

serious2strong association3 dose response gradient44/27 (14.8%)0/25 (0%)

RR 8.36 (0.47 to 147.77)

0 more per 1000 (from 0 fewer to 0 more)㊉㊉〇〇 LOWCRITICAL 0%0 more per 1000 (from 0 fewer to 0 more) Child Adherence 1observational studies

strong association22/31 (71%)

20/43 (46.5%)RR 1.53 (1.03 to 2.26)

247 more per 1000 (from 14 more to 586 more)㊉㊉〇〇 LOWCRITICAL 46.5%246 more per 1000 (from 14 more to 586 more) Plan for Future Care Communicated to Child 16observational studies

no serious indirectness

serious7strong association83/8 (37.5%)

4/39 (10.3%)RR 3.66 (1.01 to 13.27)

273 more per 1000 (from 1 more to 1258 more)㊉㊉〇〇 LOWIMPORTANT 10.3%274 more per 1000 (from 1 more to 1264 more) Caregiver Standardized Report of Child Problems (Better indicated by lower values) 2observational studies

serious9,10reduced effect for RR >> 1 or RR << 111197560-MD 0.4 higher (0.1 to 0.7 higher)㊉㊉〇〇 LOWIMPORTANT Child Report of Externalizing Behaviors (measured with: Subscales of the Child Behavior Checklist; range of scores: 0-64; Better indicated by lower values) 2observational studies

none60133-MD .43 lower (0.82 to 0.05 lower)㊉㊉〇〇 LOWIMPORTANT HIV positive Caregiver Depression (measured with: Hamilton Depression Inventory; Better indicated by lower values) 1observational studies

12none4166-MD 0.32 higher (3.58 lower to 4.22 higher)㊉㊉〇〇 LOW Perceived Social Support for HIV positive Caregiver’s Personal/Private Feelings (measured with: Arizona Social Support Subscale for Personal/Private Feelings; Better indicated by higher values) 1observational studies

none4166-MD 1.11 higher (0.22 to 2 higher)㊉㊉〇〇 LOWIMPORTANT

(16)

S U M MAR Y OF F IN DI NG S T AB LE: Disclosed to versus Not Disclosed to for Patient or population: Caregivers and c hildren in multiple countries attending clinics or services Settings: Clinic and service organizations Intervention: Disclosed to versus Not disclosed to Outcomes

Illustrative comparative risks* (95% CI) Relative effect (95% CI)

No of Partici- pants (studies) Qual- ity of the evidence (GRADE)Comments

As - sumed risk Corresponding risk

Con- trolDisclosed to versus Not Disclosed to Child TestedStudy populationRR 8.36 (0.47 to 147.77)

52 (1 study)

⊕⊕⊖⊖

low1,2,3,4Uganda (Rwemisisi, 2008)

0 per 1000

0 per 1000 (0 to 0) Child AdherenceStudy populationRR 1.53

(1.03 to 2.26)

74 (1 study)

⊕⊕⊖⊖

low5Togo (Polisset, 2009)465 per 1000711 per 1000 (479 to 1000) Plan for Future Care Communicated to Child

Study population RR 3.66

(1.01 to 13.27)

47 (1 study8)

⊕⊕⊖⊖

low6,7African immigrants in Sweden (Sander, 2009)

103 per 1000377 per 1000 (104 to 1000) Medium risk population 103 per 1000377 per 1000 (104 to 1000) Caregiver Standardized Report of Child Problems

The mean Caregiver Standardized Report of Child Problems in the intervention groups was 0.4 higher (0.1 to 0.7 higher)

757 (2 studies)

⊕⊕⊖⊖

low9,10,11

US and 15 Centers in 10 W

estern European Countries (Bauman, 2002; Nostlinger, 2006)

Child Report of Externalizing Behaviors Subscales of the Child Behavior Chec

klist. Scale from: 0 to 64.

The mean Child Report of Externalizing Behaviors in the intervention groups was

.43 lower (0.82 to 0.05 lower)

193 (2 studies)

⊕⊕⊖⊖

low

US, Adjustment further moderated by perception of parental warmth (Jones, 2007; Murphy, 2001) HIV positive Caregiver Depression Hamilton Depression Inventory

The mean HIV positive Caregiver Depression in the intervention groups was 0.32 higher (3.58 lower to 4.22 higher)

107 (1 study)

⊕⊕⊖⊖

low12US (Murphy, 2001) Perceived Social Support for HIV positive Caregiver's Personal/Private Feelings Arizona Social Support Subscale for Personal/Private Feelings

The mean Perceived Social Support for HIV positive Caregiver's Personal/Private Feelings in the intervention groups was 1.11 higher (0.22 to 2 higher)

107 (1 study)

⊕⊕⊖⊖

lowUS (Murphy, 2001)

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio; GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate. 1.Although interviews concerned over 50 children, reports were made by only 10 HIV positive parents/ caregivers. 2.Assessment of testing was based on caregiver report rather than documentation of testing. 3.Large effect in total sample. Relative risk is 12.14 if sample is limited to children 12 and under. 4.No caregiver who had not disclosed had their child tested. 5.Adherence was measured by caregiver report, unverified by biological outcomes. 6.Data is based on interview of caregiver rather than interview of both caregiver and child. 7.Relative risk is statistically significant, but not reported in article. 8.Cross-sectional retrospective interview of HIV positive caregivers. 9.Widely used Child Behavior Checklist forms the basis for these 2 studies. 10.Assessment is based on caregiver report of child’s problems. Caregiver reports vary as children age. 11.Bauman study uses age as a covariate; problems do not differ then by disclosure group. 12.Widely used adult depression inventory, the Hamilton Depression Inventory

(17)

a PPEN di X F: D o c hildren 12 years and under who get their parent/caregivers’ H IV positive status disclosed to them display greater or equal well-being pre- to post-disclosure? G RADE T AB LE S: Author(s): Date: 2010-12-20 Question: Do c hildren 12 years and under who get their parent/caregivers’ H IV positive status disclosed to them display greater or equal well-being pre- to post-discolsure? Settings: Community and clinic settings Bibliography: Health, mental health and social eff ects of disclosure of caregiver H IV to c hildren 12 and under . Coc hrane Database of Systematic Reviews [Y ear], Issue [Issue].

1.Prospective longitudinal study of naturally occurring disclosure 2.Only 15 youth were disclosed to during the course of the study, with 13 to 14 children or mothers completing various instruments 3.Mother reports on child symptomatology. Several studies indicate mothers perceive more symptomatology as child's age increases, perception of child's symptomatology is related to mother's own symptomatology, and that mothers tend to perceive more symptomatology than children. 4.Standard valid and reliable measures were used 5.Children were in the desired age range of 6-11 6.Both the probability of disclosure and of perceived psychological symptoms increase with age 7.Comparison was made from pre-disclosure to 12 to 20 months post-disclosure 8.Repeated measures, respondents are own control 9.Significant at p=.0126 in t-dependent test 10.Significant at.04 in t-dependent test 11.Significant at .047 in t-dependent test

(18)

Guideline on HiV disclosure counsellinG for cHildren up to 12 years of aGe Quality assessmentSummary of Findings

Import ance No of patients Effect Quality No of stud - ies Design Limit a- tions Inconsist - ency Indirect - ness Imprecision Other considerations

Dis - closed to versus Not dis - closed to Con - trol

Rela - tive (95% C I) Absolute

Mothers' reports of child internalizing symptoms (CBCL) (follow-up 12-20 months; measured with: Mothers' report of child symptoms on standardized Child Behavior Checklist (CBCL) Internalizing subscale; range of scores: 30-100; Better indicated by lower values) 1observational studies1very serious2no serious inconsistencyserious3no serious imprecision4,5reduced effect for RR >> 1 or RR << 16 dose response gradient7141,2148-MD 2.62 higher (1.55 lower to 6.79 higher)㊉〇〇〇 VERY LOWIMPORTANT Childrens' report of internalizing symptoms on the Children's Depression Inventory (CDI) (follow-up 12-20 months; measured with: Children's self-report of Children's Depression Inventory symptoms; range of scores: 0-54; Better indicated by lower values) 1observational studies1very serious2no serious inconsistency

no serious imprecision4,5reduced effect for RR >> 1 or RR << 16 dose response gradient7142148-MD 1.07 lower (3.76 lower to 1.62 higher)㊉㊉〇〇 LOWIMPORTANT Mothers' report of child externalizing symptoms (CBCL) (follow-up 12-20 months; measured with: Mothers' report of child symptoms on standardized CBCL Externalizing subscale; range of scores: 30-100; Better indicated by lower values) 1observational studies1very serious2no serious inconsistencyserious3no serious imprecision4,5

strong association9 reduced effect for RR >> 1 or RR << 16 dose response gradient714148-MD 4.86 higher (1.57 to 8.15 higher)㊉㊉〇〇 LOWIMPORTANT Childrens' report of CBCL externalizing symptoms (Youth Self Report-YSR) (follow-up 12-20 months; measured with: Children's self-report of CBCL-YSR externalizing symptoms; range of scores: 0-19; Better indicated by lower values) 1observational studies1very serious2no serious inconsistency

no serious imprecision4,5reduced effect for RR >> 1 or RR << 16 dose response gradient714148-MD .07 higher (4.5 lower to 4.64 higher)㊉㊉〇〇 LOWIMPORTANT Mother's report of relationship quality (follow-up 12-20 months; measured with: Mothers' report of relationship quality on the Interaction Behavior Questionnaire (IBQ); range of scores: 0-20; Better indicated by higher values) 1observational studies1very serious2no serious inconsistencyserious3no serious imprecision4,5strong association10 dose response gradient714148-MD 1.72 lower (3.19 to 0.25 lower)㊉〇〇〇 VERY LOWIMPORTANT Children's report of relationship quality (follow-up 12-20 months; measured with: Childrens' report of relationship quality on the Interaction Behavior Questionnaire (IBQ); range of scores: 0-10; Better indicated by higher values) 1observational studies1very serious2no serious inconsistency

no serious imprecision4,5dose response gradient714148-MD 0.79 lower (1.91 lower to 0.33 higher)㊉〇〇〇 VERY LOWIMPORTANT Children's understanding of HIV (follow-up 12-20 months; measured with: Child understanding of HIV; range of scores: 0-3; Better indicated by higher values) 1observational studies1very serious2no serious inconsistency

no serious indirectness no serious imprecision4,5strong association11 dose response gradient713138-MD 0.53 higher (0.06 to 1 higher)㊉㊉〇〇 LOWIMPORTANT

(19)

S U M MAR Y OF F IN DI NG S T AB LE: Mental Health and H IV K nowledge Outcomes P re-disclosure to P ost-disclosure for Disclosure of Caregivers’ H IV positive Status to Children 12 and Under Patient or population: H IV positive mothers attending an H IV/primary care clinic in the US and their 6-11 year old c hildren Settings: H IV/primary care clinic in an urban setting in the United States Intervention: Naturally-occurring disclosure by caregiver during a longitudinal study Outcomes Illustrative comparative risks* (95% C I) Rela - tive effect (95% C I)

No of Partici - pants (studies)

Quality of the evidence (G RAD E) Comments

As- sumed risk

Corresponding risk

Con- trolPre-disclosure versus Post- disclosure Mothers' reports of child internalizing symptoms (CBCL) Mothers' report of child symptoms on standardized Child Behavior Checklist (CBCL)

Internalizing subscale. Scale from: 30 to 100. Follow-up: 12-20 months

The mean Mothers' reports of child internalizing symptoms (CBCL) in the intervention groups was 2.62 higher (1.55 lower to 6.79 higher)

28 (1 study8)

⊕⊖⊖⊖

very low1,2,3,4,5,6,7

Comparisons in this table are based on one prospective longitudinal study by Shaffer (2001). Only three results were statistically significant in bivariate testing Childrens' report of internalizing symptoms on the Children's Depression Inventory (CDI) Children's self-report of Children's Depression Inventory symptoms. Scale from: 0 to 54. Follow-up: 12-20 months

The mean Childrens' report of internalizing symptoms on the Children's Depression Inventory

(CDI) in the intervention groups was 1.07 lower (3.76 lower to 1.62 higher)

28 (1 study8)

⊕⊕⊖⊖

low1,2,3,4,5,6,7 Mothers' report of child externalizing symptoms (CBCL) Mothers' report of child symptoms on standardized CBCL Externalizing subscale. Scale from: 30 to 100. Follow-up: 12-20 months

The mean Mothers' report of child externalizing symptoms (CBCL) in the intervention groups was 4.86 higher (1.57 to 8.15 higher)

28 (1 study8)

⊕⊕⊖⊖

low1,2,3,4,5,6,7,9Statistically significant in t-dependent Childrens' report of CBCL externalizing symptoms (Youth Self Report-YSR) Children's self-report of CBCL-YSR

externalizing symptoms. Scale from: 0 to 19. Follow-up: 12-20 months

The mean Childrens' report of CBCL externalizing symptoms (Youth Self Report-YSR) in the intervention groups was .07 higher (4.5 lower to 4.64 higher)

28 (1 study8)

⊕⊕⊖⊖

low1,2,3,4,5,6,7 Mother's report of relationship quality

Mothers' report of relationship quality on the Interaction Behavior Questionnaire (I

BQ). Scale from: 0 to 20. Follow-up: 12-20 months

The mean Mother's report of relationship quality in the intervention groups was 1.72 lower (3.19 to 0.25 lower)

28 (1 study8)

⊕⊖⊖⊖

very low1,32,3,4,5,7,10Statistically significant in t-dependent Children's report of relationship quality

Childrens' report of relationship quality on the Interaction Behavior Questionnaire (I

BQ). Scale from: 0 to 10. Follow-up: 12-20 months

The mean Children's report of relationship quality in the intervention groups was 0.79 lower (1.91 lower to 0.33 higher)

28 (1 study8)

⊕⊖⊖⊖

very low1,3,4,5,7 Children's understanding of HIV Child understanding of HIV. Scale from: 0 to 3. Follow-up: 12-20 months

The mean Children's understanding of HIV in the intervention groups was 0.53 higher (0.06 to 1 higher)

26 (1 study8)

⊕⊕⊖⊖

low1,3,4,5,7,11,12Statistically significant in t-dependent

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio; GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate. 1.Only 15 youth were disclosed to during the course of the study, with 13 to 14 children or mothers completing various instruments 2.Mother reports on child symptomatology. Several studies indicate mothers perceive more symptomatology as child’s age increases, perception of child’s symptomatology is related to mother’s own symptomatology, and that mothers tend to perceive more symptomatology than children. 3.Standard valid and reliable measures were used 4.Children were in the desired age range of 6-11 5.Non-significant as well as significant effects were reported 6.Both the probability of disclosure and of perceived psychological symptoms increase with age 7.Comparison was made from pre-disclosure to 12 to 20 months post-disclosure 8.Prospective longitudinal study of naturally occurring disclosure 9.Significant at p=.0126 in t-dependent test 10.Significant at.04 in t-dependent test 11.Significant at .047 in t-dependent test 12.Both understanding and probability of disclosure likely increase with age

(20)

aPPENdiX G:

Child reaction to parent/caregiver disclosure

Degree of child reaction, in the short and long term, to caregiver HIV disclosure as rated by parents on rating scales 1. Delaney (2008) and Tompkins (2007) both reported means and standard deviations for parent ratings of

child reactions. Five of the most prevalent reactions were the same for both authors, as was the underlying continuum of the rating, not at all to very much. Delaney (N=41, children 5-18, US), however, used a 5-point rating scale (1-5) and Tompkins (N=14, children 9-16, US) a 4-point rating scale (0-3). The most conservative statistical procedure was to calculate the unaltered combined mean and standard deviation. The next step was to determine the adjusted center of the new scale and report deviations from that center. Thus 0 or positive deviations indicate moderate to more than moderate concern and negative deviations indicate less than mo- derate concern.

2. Generally, estimates <.40 are considered small, .40-.70 moderate, and greater than .70 large.

3. “Long term” or “over time” reactions were not defined for the parents in the Delaney study. In the Tompkins study a mean 3.7 years (sd=2.39) had passed between disclosure and interview.

4. Delaney also had items about inquisitive (curious, questioning), avoidant, and comforting reactions that did not appear in other studies. Tompkins also assesses children’s self-reported initial and current reactions, which show a consistent pattern of decrease in rating severity.

Reaction

Immediate standardized effect size based on deviation from

center of scale (N=55)

sd

Long-term standardized effect size based on deviation from

center of scale (N=55)

sd

worried 0.36 1.79 0.15 1.90

sad/cries, emotional

distress 0.31 1.62 -0.35 1.56

shocked 0.08 1.77 -0.44 1.53

confused/not

understand -0.06 1.59 -0.26 1.85

angry -0.31 1.73 -0.84 1.36

(21)

Parent-reported immediate reactions of children to disclosure of caregiver HIV status –reaction prevalence

1. Based on studies by Murphy (2006), children ages 6-11, US; Nostlinger (2006), children ages 1-18, 10 West European countries; Palin (2009), children 11-16, South Africa; and Tompkins (2007), children 9-16, US, with parents reporting on immediate reactions of children to caregiver/parent’s disclosure of caregiver/parent HIV status.

2. Delaney (2008) and Murphy (2006) also indicate that children comfort and reassure parents/caregivers, with data provided by parental or child report.

Reaction N=269 %

Negative emotions

sad 112 41.64

worried 63 23.42

fearful 10 3.72

angry 5 1.86

Neutral reactions

no reaction 49 18.22

Positive reactions

mature reaction 45 16.73

hopeful 7 2.60

relieved 3 1.12

positive reaction (unspecified) 2 0.74

Cognitive/emotional reactions

confused/not understand 14 5.20

shock/disbelief 14 5.20

Other

other reaction (unspecified) 12 4.46

act out 7 2.60

ignore/deny 4 1.49

(22)

a PPEN di X h : D o c hildren whose parent/caregiver H IV status is publicly kno wn display greater to equal well-being to community control compared to children whose parent/caregiver H IV status is confidential? G RADE T AB LE S: Author(s): Date: 2011-01-04 Question: Do c hildren whose parent/caregiver H IV status is publicly known display greater or equal well-being to community controls? Settings: Community , clinic, hospital and sc hool settings Bibliography: Health, mental health and social eff ects of disclosure of caregiver H IV to c hildren 12 and under . Coc hrane Database of Systematic Reviews [Y ear], Issue [Issue].

1 Lifetime report in Li study and 1 year retrospective report in Ishikawa study 2 Retrospective self-report for lifetime or one year 3 No control for serious illness in family in community condition 4 Figures do not deviate greatly from report by Nostlinger on prevalence of bullying for HIV-affected youth in Western European countries in reports by parents in HIV-affected families 5 Child sellf-report 6 Ishikawa article focuses on children 8-12 (3rd through 6th grades in Thailand) while Li article focuses on children 6-18 7 Effect consistently >1 across studies 8 Lifetime report for 6-18 year olds 9 HIV-related events were removed from the final calculation so that the probability of events would be the same for HIV-affected youths and controls 10 Authors of Li study report results stratified by age and demonstrate consistent effects within age groups 11 Significant at p<.01 in between group analysis 12 Standardized measures translated into Chinese 13 Means adjusted for age, sex and family socioeconomic status 14 Significant at p<.01 in multivariate analyses

(23)

Quality assessmentSummary of Findings

Import ance No of patients Effect Quality No of stud - ies Design Limit a- tions Inconsist - ency Indirectness Imprecision Other considerations

Commu - nity A ware Child H IV- affected versus Community Control

Control Rela - tive (95% C I) Absolute

Child reports being bullied or threatened (follow-up 1-18 years1; Child self-reports on questionnaire) 2observational studies2serious3no serious inconsistency4no serious indirectness5,6

5strong association784/1259 (6.7%)

69/834 (8.3%)

RR 1.65 (1.09 to 2.5)

54 more per 1000 (from 7 more to 124 more)㊉㊉〇〇 LOWIMPORTANT 53 more per 1000 (from 7 more to 122 more)8.1% Number of traumatic life events (follow-up 6-18 years8; measured with: Endorsement of occurrence of 18 traumatic life events9; range of scores: 0-18; Better indicated by lower values) 1observational studiesserious3no serious inconsistency

5

10strong association111221404-MD 1.2 higher (0.96 to 1.44 higher)㊉㊉〇〇 LOWIMPORTANT Depression (CES-DC, Adjusted Mean Scale Score) (measured with: CES-DC for children adjusted mean scale score12,13; range of scores: 0-3; Better indicated by lower values) 1observational studiesserious3no serious inconsistency

no serious imprecision5,12strong association14 increased effect for RR ~1131221404-MD 0.17 higher (0.12 to 0.22 higher)㊉㊉㊉〇 MODERATEIMPORTANT Loneliness (Children's Loneliness Scale, Adjusted Mean Scale Score) (measured with: Adjusted mean scale score of the Children's Loneliness Scale (Asher)12,13; range of scores: 1-5; Better indicated by lower values) 1observational studiesserious3no serious inconsistency

no serious imprecision5,12strong association14 increased effect for RR ~1131221404-MD 0.37 higher (0.29 to 0.45 higher)㊉㊉㊉〇 MODERATEIMPORTANT Self-esteem (Rosenberg, Adjusted Mean Scale Score) (measured with: Adjusted Mean Scale Score for the Rosenberg Self-Esteem Scale13; range of scores: 0-3; Better indicated by higher values) 1observational studiesserious3no serious inconsistency

no serious imprecision5,12strong association14 increased effect for RR ~112,131221404-MD 0.13 lower (0.18 to 0.08 lower)㊉㊉㊉〇 MODERATEIMPORTANT Future Expectations (Adjusted Mean Scale Score) (measured with: Adjusted Mean Scale Score for the Children's Future Expectations Scale (Bryan)12,13; range of scores: 1-5; Better indicated by higher values) 1observational studiesserious3no serious inconsistency

no serious imprecision5,12strong association14 increased effect for RR ~1131221404-MD 0.12 lower (0.22 to 0.02 lower)㊉㊉㊉〇 MODERATEIMPORTANT Hope for Future (Adjusted Mean Scale Score) (measured with: Adjusted Mean Scale Score for Hopefulness about Future Scale (Whitaker)12,13; range of scores: 1-4; Better indicated by higher values) 1observational studiesserious3no serious inconsistency

no serious imprecision5,12strong association14 increased effect for RR ~1131221404-MD 0.21 lower (0.29 to 0.13 lower)㊉㊉㊉〇 MODERATEIMPORTANT

(24)

S U M MAR Y OF F IN DI NG S T AB LE: Mental Health Eff ects of Community A wareness that Child is H IV-aff ected versus Community Control Patient or population: Children in H IV-aff ected families Settings: High H IV seroprevalence communities in T hailand and China Intervention: Community A ware that Child is H IV-aff ected versus Community Control Outcomes Illustrative comparative risks* (95% C I) Rela - tive effect (95% C I)

No of Partici - pants (stud - ies)

Quality of the evidence (G RAD E) Comments

As- sumed risk

Corresponding risk

Con- trolCommunity Aware Child HIV-affected versus Community Control Child reports being bullied or threatened Child self-reports on questionnaire Retrospective: 1-18 years1

Study populationRR 1.65 (1.09 to 2.5)

2093 (2 studies7)

⊕⊕⊖⊖

low2,3,4,5,6,

Thailand study of school children (Ishikawa, 2010) and community study in rural China (Li, 2009). All remaining studies in China are from the same study group. In China studies HIV-affected families are identified by community leaders and living arrangements (e.g., group homes, foster homes, orphanages)

83 per 1000

137 per 1000 (90 to 208) Number of traumatic life events Endorsement of occurrence of 18 traumatic lif

e events8.

Scale from: 0 to 18. Follow-up: 6-18 years

9

The mean Number of traumatic life events in the intervention groups was 1.2 higher (0.96 to 1.44 higher)

1625 (1 study)

⊕⊕⊖⊖

low2,4,10,11Rural China (Li, 2009) Depression (CES-DC, Adjusted Mean Scale Score) CES-DC for children adjusted mean scale score12,13. Scale from: 0 to 3.

The mean Depression (CES-DC,

Adjusted Mean Scale Score) in the intervention groups was

0.17 higher (0.12 to 0.22 higher)

1625 (1 study)

⊕⊕⊕⊖

moderate 2,4,12,13,14Rural China (Li, 2009) Loneliness (Children's Loneliness Scale, Adjusted Mean Scale Score)

Adjusted mean scale score of the Children's L

oneliness Scale (Asher)12,13. Scale from: 1 to 5.

The mean Loneliness (Children's

Loneliness Scale, Adjusted Mean Scale Score) in the intervention groups was

0.37 higher (0.29 to 0.45 higher)

1625 (1 study)

⊕⊕⊕⊖

moderate 2,4,12,13,14Rural China (Li, 2009) Self-esteem (Rosenberg, Adjusted Mean Scale Score)

Adjusted Mean Scale Score for the Rosenberg Self-Esteem Scale

13. Scale from: 0 to 3.

The mean Self-esteem (Rosenberg, Adjusted Mean Scale Score) in the intervention groups was

0.13 lower (0.18 to 0.08 lower)

1625 (1 study)

⊕⊕⊕⊖

moderate 2,4,12,13,14Rural China (Li, 2009) Future Expectations (Adjusted Mean Scale Score)

Adjusted Mean Scale Score for the Children's F

uture Expectations Scale (Bryan)12,13. Scale from: 1 to 5.

The mean Future Expectations

(Adjusted Mean Scale Score) in the intervention groups was

0.12 lower (0.22 to 0.02 lower)

1625 (1 study)

⊕⊕⊕⊖

moderate 2,4,12,13,14

Rural China (Li, 2009) Scale measures generalized expectations, e.g., school, career Hope for Future (Adjusted Mean Scale Score)

Adjusted Mean Scale Score for Hopefulness ab

out Future Scale (Whitaker)12,13. Scale from: 1 to 4.

The mean Hope for Future (Adjusted Mean Scale Score) in the intervention groups was 0.21 lower (0.29 to 0.13 lower)

1625 (1 study)

⊕⊕⊕⊖

moderate 2,4,12,13,14

Rural China (Li, 2009) Scale measures concrete expectations, e.g., graduation from school

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio; GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate. 1.Lifetime report in Li study and 1 year retrospective report in Ishikawa study 2.No control for serious illness in family in community condition 3.Figures do not deviate greatly from report by Nostlinger on prevalence of bullying for HIV-affected youth in Western European countries in reports by parents in HIV-affected families 4.Child self-report 5.Ishikawa article focuses on children 8-12 (3rd through 6th grades in Thailand) while Li article focuses on children 6-18 6.Effect consistently >1 across studies 7.Retrospective self-report for lifetime or one year 8.HIV-related events were removed from the final calculation so that the probability of events would be the same for HIV-affected youths and controls 9.Lifetime report for 6-18 year olds 10.Authors of Li study report results stratified by age and demonstrate consistent effects within age groups 11.Significant at p<.01 in between group analysis 12.Standardized measures translated into Chinese 13.Means adjusted for age, sex and family socioeconomic status 14.Significant at p<.01 in multivariate analyses

(25)

aPPENdiX i:

Example disclosure models/ approaches presented by Guideline Group members at the consultation meeting, June 2009

Country / Program Summary of model

Democratic Republic of Congo (DRC)

Used by DRC/ Sustainable AntiRetroviral Access Program (SARA)/Kalembelembe Pediatric Hospital (KLL) Summary of disclosure process:

When to disclose: Begin disclosure from when child is 5 years old, full disclosure by 11 years. Earlier start of disclosure process if a child asks questions, has an aggressive attitude or decreased adherence to treatment and care.

Who discloses: Health-care workers (HCWs), parents/caregiver with HCW’s assistance if necessary How disclosure is done: Approach comprises four steps, step duration depends on a child’s level of psychological development.

• Step 1: Obtain good adherence to care and treatment and maintain confidentiality.

• Step 2: Child learns about the functioning of the human body, acute and chronic illness. The child should understand that he/she has a chronic illness and the disease can be controlled by treatment.

• Step 3: Disclosure (naming HIV)

• Step 4: Support following disclosure.

Tools used during the disclosure process, adapted to the child’s age: tales for younger children (5–8 years old); illustrated explanations (children 8–10 years old); comic strips for children older than 10 years (following full disclosure).

Key lessons learnt during implementation:

• Principal barriers to disclosure are parents’ underestimation of children’s knowledge of HIV and lack of initiative from HCWs.

India/ WAG CHELSEA (Women’s Action Group CHELSEA, where CHELSEAs tands for children, health, education, ladies, senior citizens, environment, awareness)

An approach used for provision of information to young children up to 9 years of age. This is done through:

• A series of informal group sessions targeted at children 5–9 years of age

• Objective of sessions: To provide relevant information in a graded, child-friendly manner, and to help find answers to common questions.

• Sessions are interactive and start with some simple activities like a poem/simple dance to relax the participants.

• Subjects covered: General health and hygiene, healthy lifestyle, care during sickness and nutritious foods. The topic of illness caused by HIV is introduced without disclosing the child/parent’s status.

• At the end of each session children are asked to draw what they have learnt.

• Follow-up review (with parent) of child’s behaviour. Parents are prepared to answer questions that children may ask after sessions.

Tools used: Stories, mostly using animal characters; flash cards; role-play by counsellors; puppet shows;

games/activities Advantages of sessions:

• Prepare the background for full disclosure

• Help explore children’s feelings (e.g. guilt, sadness, anger) and provide an opportunity for expressing and dealing with them

• Facilitate bonding and establishment of communication channels and support among the children and between children and counsellors.

Kenya/ Muangalizi

project, EDARP A community health worker (CHW) model When to disclose:

• Start process when caregiver and child are ready (when a child can process basic abstract information, preferably from 5 years of age).

• Full disclosure to children from 10 years of age, preferably done at the time of diagnosis of HIV.

Who discloses: Caregiver disclosure preferred (or), facilitated by the health-care provider.

How disclosure is done: Process and not event, partial then complete.

• Help the caregiver understand the why, when, how of HIV communication.

• Prepare the caregiver by supporting sustainable coping mechanisms. Anticipate challenges, model communication.

• Address myths that the child may have; begin with what the child knows.

• Disclose to significant adults to whom the child might disclose.

• Provide follow-up support for the child and caregiver. Look out for adverse reactions. Observe play to assess knowledge and coping mechanisms. CHW to provide routine feedback on progress of the child.

Results and key lessons learnt:

• Trend: increased rates of disclosure with increasing age.

• Challenges at operational level include HIV-related stigma, lack of an operational definition of disclosure, and lack of clear guidance on the disclosure process

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio; GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate. 1.Lifetime report in Li study and 1 year retrospective report in Ishikawa study 2.No control for serious illness in family in community condition 3.Figures do not deviate greatly from report by Nostlinger on prevalence of bullying for HIV-affected youth in Western European countries in reports by parents in HIV-affected families 4.Child self-report 5.Ishikawa article focuses on children 8-12 (3rd through 6th grades in Thailand) while Li article focuses on children 6-18 6.Effect consistently >1 across studies 7.Retrospective self-report for lifetime or one year 8.HIV-related events were removed from the final calculation so that the probability of events would be the same for HIV-affected youths and controls 9.Lifetime report for 6-18 year olds 10.Authors of Li study report results stratified by age and demonstrate consistent effects within age groups 11.Significant at p<.01 in between group analysis 12.Standardized measures translated into Chinese 13.Means adjusted for age, sex and family socioeconomic status 14.Significant at p<.01 in multivariate analyses

Guideline on HiV disclosure counsellinG for cHildren up to 12 years of aGe - appendices

hiv/aids Programme