Aller au contenu principal

Adherence to ready-to-use food and acceptability of outpatient nutritional therapy in HIV-infected undernourished Senegalese adolescents

Fatou Niasse , Marie Varloteaux , Karim Diop , Sidy Mokhtar Ndiaye , François Niokhor Diouf, Pape Birane Mbodj , Babacar Niang , Aminata Diack and Cecile Cames

Abstract

Open Access

Background: Ready-to-use food (RUF) is increasingly used for nutritional therapy in HIV-infected individuals. However, practical guidance advising nutrition care to HIV-infected adolescents is lacking, so that little is known about the acceptability of such therapy in this vulnerable population. This study assesses the overall acceptability and perception of a RUF-based therapy and risk factors associated with sub-optimal RUF intake in HIV-infected undernourished adolescents in Senegal.

Methods: Participants 5 to 18 years of age with acute malnutrition were enrolled in 12 HIV clinics in Senegal. Participants were provided with imported RUF, according to WHO prescription weight- and age-bands (2009), until recovery or for a maximum of 9–12 months. Malnutrition and recovery were defined according to WHO growth standards. Adherence was assessed fortnightly by self-reported RUF intake over the period. Sub-optimal RUF intake was defined as when consumption of the RUF provision was < 50%. RUF therapy acceptability and perceptions were assessed using a structured questionnaire at week 2 and focus group discussions (FGDs) at the end of the study. Factors associated with sub-optimal RUF intake at week 2 were identified using a stepwise logistic regression model.

* Correspondence: cecile.cames@ird.fr


Fatou Niasse and Marie Varloteaux are first co-authors.
Institut de recherche pour le développement (IRD), UMI233 TransVIHmi, INSERM U1175, Université de Montpellier, 911 Av Agropolis, 34394 Montpellier, France
Full list of author information is available at the end of the article

Introduction

Malnutrition is common among children living in low incomes settings and is even more critical in HIV- infected children, including those on antiretroviral treat- ment (ART) [1, 2], as HIV infection can induce growth failure and malnutrition in children, which in turn con- tribute to faster disease progression. Successful treat- ment of acute malnutrition in young children < 5 years [3], including HIV-infected children [4–6], has been achieved using Plumpy Nut® and Plumpy Sup® (Nutriset, Malaunay, France) food supplements, developed for pediatric use and generically referred to as imported ready-to-use food (RUF). During the last decade, most African countries adopted national guidelines for RUF- based treatment of acute malnutrition in children <5 years, including those with HIV.

The vulnerabilities of HIV-infected adolescents, espe- cially those with perinatal HIV infection, have been ex- tensively described [7]. They often have more advanced HIV disease with related co-morbidities associated with delayed HIV treatment and lifelong infection [8]. Peri- natally HIV-infected children ≥5 years and adolescents are also vulnerable to malnutrition and often face severe forms. Recent studies reported 16 to 33% of acute mal- nutrition in children and adolescents 5 to 19years-old on long term ART, of whom one-third was severely undernourished [9, 10]. Because National health systems have been oriented towards monitoring and care of chil- dren < 5 years and adult, adolescents health care services, including those living with HIV, suffer from multiple de- ficiencies, including the lack of nutritional care [11]. Despite World Health Organisation (WHO) recommen- dations for the integration of nutritional interventions into comprehensive HIV treatment programmes [12], nutritional therapy at HIV treatment sites in sub- Saharan Africa remains poorly available [13]. This is par- ticularly true with respect to the recommendations for adolescents, 2009 WHO Preliminary guidelines for an

integrated approach to the nutritional care of HIV- infected adolescents [14].

Consequently, acceptability data and practical guid- ance advising the implementation of nutrition care in these vulnerable patients are lacking. This is of concern, as the effectiveness of medical therapy is strongly associ- ated with its level of acceptance, adherence to prescrip- tion, timing of intake, and medical and psychosocial surroundings [15], all of which may differ among pa- tients of different age classes. Experimental data on the acceptability of RUF were obtained primarily from stud- ies conducted in adults since RUF provisioning has be- come a widely used outpatient strategy among vulnerable adults through food-by-prescription pro- grammes. Qualitative studies conducted in Southern Af- rica reported poor overall acceptability of RUF [16–18], whereas its organoleptic qualities were rated as good in studies from Vietnam [19] and Haiti [20]. Assessment of a pilot nutrition support providing RUF to 43 HIV- infected children and adolescents with acute malnutri- tion in two HIV clinics in Senegal shown that beneficiar- ies consumed 50% of their RUF on average and 20% defaulted from the programme [21]. Following this pilot study, the SNACS Study was implemented to assess the acceptability, effectiveness, and feasibility of outpatient RUF-based nutritional therapy among HIV-infected chil- dren and adolescents presenting with acute malnutrition across Senegal. The intervention protocol followed the 2009 WHO Guidelines [14]. The operational objective of this research, conducted in close partnership with the Senegalese Ministry of Health, was to provide the na- tional HIV programme with evidence-based recommen- dations for outpatient nutritional therapy in HIV- infected children and adolescents.

The objectives of the present report are first to describe the overall acceptability of the nutritional intervention in- cluding perceptions, behaviours, and constraints sur- rounding RUF use. Secondly, we assess adherence to RUF

Niasse et al. BMC Public Health (2020) 20:695

prescription and investigate risk factors associated with sub-optimal intake. To our knowledge, the present study is the first to conduct such assessments among HIV- infected undernourished adolescents.

Methods

Study setting

The SNACS study was carried out in Senegal from April 2015 to January 2017 in 12 public paediatric clinics: two in the Dakar district, the capital city, (Albert Royer Na- tional Paediatric Hospital and Roi Baudouin Hospital) and 10 in decentralized settings (Regional Hospitals of Saint Louis, Louga, Mbour, Kaoloack, Ziguinchor and Kolda; Health centres of Thiès, Nioro du Rip, Bignona and Kolda). Study implementation was gradual, starting with the central sites in April 2015 and ending with the Kolda sites in March 2016. SNACS, which refers to “snacks”, is the French acronym for nutritional support of growing adolescents in Senegal.

Study population and procedures

The analyses include participants aged 5 to 18years under active follow-up for their HIV infection and pre- senting with acute malnutrition. Children with acute malnutrition have low weight for their height due to a rapid weight loss, while those with chronic malnutrition have a low height for their age. Moderate acute malnu- trition (MAM) was defined as body mass index-for-age z-score (BMIZ) of < − 2 and ≥ − 3, and severe acute mal- nutrition (SAM) as BMIZ < − 3, according to the WHO growth standards [22]. Participants also were without medical complications and succeeded at an appetite test, which assesses their ability to consume a weight- appropriate portion of RUF [23].

Participants were seen every 2weeks for clinical as- sessment, provision of RUF for the next 2 weeks, and monitoring of adherence to RUF over the preceding 2 weeks and provided with RUF until they recovered from acute malnutrition (BMIZ ≥ − 2), discontinued the study (default, death), and at the latest, until the study was concluded, i.e. 9months in decentralized clinics or 12 months in Dakar. A social health worker counselled par- ticipants and caregivers to distribute RUF intake throughout the day, to eat the RUF in small bites to cope with nausea and lack of appetite, and to have sufficient drinking water at their disposal to prevent diarrhoea. Participants were also advised to avoid RUF intake dur- ing mealtimes, in order to preserve family meals and avoid sharing with other family members.

The study used imported RUF, Plumpy Nut® and Plumpy Sup®. Both are packaged in a 92 g individual sa- chet providing 500 kcal in the form of an energy-dense lipid paste made from peanut butter, oil, sugar and a high vitamin and mineral supplement (the first contains

milk powder, while the second contains soy proteins) [24]. Plumpy Sup® is recommended in the management of MAM, amongst other options, while Plumpy Nut® is dedicated to the treatment of SAM. WHO recommends a Plumpy Nut® prescription of 75 to 100kcal/kg/d in children aged 5 to 10 years and 60 to 90 kcal/kg/d above that age [14]. Across all ages, Plumpy Sup® was pre- scribed at the lowest recommended dose of 60 kcal/kg/d to participants with MAM. Prescription by weight may involve a high daily dose in older HIV-infected children and adolescents. The lowest value was used and max- imum energy intake provided by RUF was limited to 2000 kcal/d i.e. 4 sachets in all participants to preserve habitual diet and prevent appetite saturation.

Data collection and analysis

Participants’ characteristics at enrolment
Clinical and therapeutic characteristics of the participant were recorded from medical files while socioeconomic and food security data were collected with the caregiver. The household food insecurity access scale (HFIAS)—a list of nine specific questions about the availability and accessibility of foods and food-related worries for the household during the previous month—was used to as- sess food insecurity [25]. This access scale generates a 4- class variable of food insecurity—none, mild, moderate, and severe. In this analysis, we tested both the 4-class and the derived binary variable (food insecurity: yes vs. none). Minimum dietary diversity was defined as having consumed at least 5 food groups out of 10 the day before the study visit, using the Global Dietary Diversity Indica- tor for Women, as measured by participant 24-h recall of food consumption at enrolment and week 2 [26]. The recall was primarily conducted with the participant (with the help of the caregiver if need be).

At week 2, a structured questionnaire covering 5 topics, organoleptic appreciation of RUF, mode of intake, 24-h recall of RUF intake, self-stigma associated to RUF intake and RUF sharing, was administered primarily to participants ≥7 years or caregivers below that age. Ques- tions about RUF sharing were administered separately to participants ≥7 years and caregivers.

Adherence to RUF prescription

Based on participant and caregiver reports on RUF consumption, adherence was calculated as the percent- age of reported RUF sachets consumed out of the num- ber of sachets provided at each follow-up visit. Early sub-optimal RUF intake was defined when this ratio was <50%atweek2.

Risk factors of early sub-optimal RUF intake

Associations between early sub-optimal RUF intake with socioeconomic and demographic profiles, RUF acceptability

Niasse et al. BMC Public Health (2020) 20:695

Page 4 of 12

data, as well as ART drug history and severity of malnutri- tion, were investigated using univariable regression, after controlling for collinearity, as well as stepwise logistic re- gression models. Explanatory variables with P < 0.20 from the univariable analysis were included in the multivariable models and excluded by the stepwise procedure at P > 0.10. The final set of variables retained were tested for interac- tions. Differences were considered statistically significant at P <0.05. Medians are presented with their interquartile range (IQR). All statistical analyses were performed in SAS, version 9.3 (Cary, North Carolina, USA).

Expectations, perceptions and experiences of participants in the nutritional intervention
Focus group discussions (FGD) were conducted at the Dakar study clinics, after the end of the total study period (12months), with a convenience subsample of participants aged at least 7years, which aimed to be fairly representative of the study population with regards to age, sex and severity of malnutrition. Sessions were conducted in Wolof, the national language of Senegal, or French, by a social health worker and the doctoral stu- dent (MV) and focused on (i) participant expectations from the study, (ii) perceptions of RUF and (iii) experi- ences of the nutritional intervention, depending on their outcome in the study (recovery, failure, default). FGD were audiotaped and transcribed verbatim into French. The thematic analysis which consisted in coding, synthe- sizing and grouping sections of text from the FGDs that covered similar issues or experiences, was performed by the doctoral student (MV) in Dedoose, version 7.0.21 (Los Angeles, California, USA).

Ethical approval

Ethics clearance for the SNACS study protocol was given by the Ethics and Regulatory Committee and the Ministry of Health in Senegal. All parents or surrogate caregivers provided written informed consent. Partici- pants aged ≥7 years received extended information about the research and provided verbal assent [27].

Results

Study population

From April 2015 to September 2016, the SNACS study enrolled 184 HIV-infected participants of whom 173 were between the ages of 5 and 18years old (median age: 12.5years, IQR: 9.5–14.9), 104 presenting with MAM and 69 with SAM, were included in these ana- lyses (Table 1). All but one participant had been infected through mother-to-child transmission. The mother or father was the primary caregiver in 44 and 8% of partici- pants, respectively, while 48% lived with a surrogate caregiver who was an ascendant (grandparents, uncles, aunts: n = 24), collateral (siblings, cousins: n = 57), or a

neighbour (n = 2). Eighty-seven per cent were on ART at enrolment for a median duration of 48 months (15–75). Only 28% had had their HIV serologic status disclosed for whom the median time since the disclosure was 11 months (5–26). HIV status disclosure was associated with virologic suppression (< 50 copies/ml, P < 0.0001). Sixty-one per cent of participants recovered from wast- ing after a median follow-up duration of 29days (16– 65), 31% failed to reach BMIZ ≥-2 during the time of the study, 6% defaulted and 2% died. The overall median follow-up duration was 66days (21–224), which in- creased from 30 days (16–115) among participants pre- senting with MAM to 156days (85–275) in SAM participants (P < 0.0001).

Acceptability of RUF and feeding practices at week 2

Overall, 87, 79, 85, and 80% of participants initially rated the RUF appearance (colour and texture), taste, smell, and mouth feeling, respectively, as good. Those who stated they disliked RUF taste perceived it as too salty (38%), too sour (24%), too greasy (24%) or too sweet (14%). Participants who disliked the mouth feeling stated it was too pasty (80%) or sticky (20%). However, up to 30% reported feeling at some point disgusted by RUF (Table 2), and 19% stated there had been occasions when they refused to eat it. At least one episode of diar- rhoea and/or vomiting related to RUF intake was re- ported by 30% of participants at the first follow up visit, 2 weeks after initiating therapy. 24% of participants pre- sented 1 to 3 episodes of diarrhoea and 12% presented 1 to 3 episodes of vomiting. The primary mode for eating RUF was directly from the sachet (85%) (Table 2). Many participants (63%) consumed RUF in place of breakfast, while others reported eating it just before (10%), or just after (12%) meals. Only 15% reported consuming RUF as a snack. At enrolment and week 2, 56 and 39% of partic- ipants, respectively, reached the minimum dietary diver- sity the day before the visit. One-third of the participants stated they hid from others when eating RUF (Table 2). The reasons given were fear of envious reactions (42%), arousing pity (19%), or teasing (16%), while 23% chose not to respond to the question. Occa- sional RUF sharing within the household was reported by 38% of caregivers and participants (Table 2). Care- givers reported that they felt: they (82%), other adults (53%), or other children (67%) in the household needed the RUF. Acceptability data were also compared by the severity of malnutrition and no remarkable association was found.

Adherence to RUF

Most participants < 12 years were provided with 2 to 3 RUF sachets daily while most of their older counterparts received 3 to 4 sachets. If the maximum energy intake to

Niasse et al. BMC Public Health (2020) 20:695 Page 5 of 12

Table 1 Characteristics of HIV-infected participants and their caregivers at enrolment in the SNACS study, Senegal a-b

Characteristics

Girl
Decentralized study site Acute malnutrition

Moderate

Severe
HAZ, median IQR At school
Family status

Parents alive Mother orphan Father orphan Double orphan

HIV disclosed ARTc

No ART
ART – virologic suppressiond ART – no virologic suppression

CD4 cell/mm3, median IQR Caregiver school level

None
Primary
Secondary and more

Caregiver income None

Regular

Irregular
Food insecure household

< 12years N = 84

34 51

57 27 -1.7 58

31 20 16 17 5

17 26 40 603

36 25 23

40 9 35 65

≥ 12 years All N=89 N=

35 (39) 69 42 (47) 93

47 (53) 104 42 (47) 69 -2.0 (- 2.7 – - 1.2) -1.8 74 (83) 132

21 (24) 52 23 (26) 43 25 (28) 41 20 (22) 37 44 (49) 49

6 (7) 23 39 (44) 65 43 (49) 83 414 (209–707) 522

49 (55) 85 21 (24) 46 19 (21) 42

36 (40) 76 14 (16) 23 39 (44) 74 68 (76) 133

173

P value (40) 0.88

(40) (61)

(68)
(32)
(-2.4 – -0.8) (69)

(37) (24) (19) (20) (6)

(20)
(31)
(48) (343–861)

(43) (30) (27)

(48) (11) (42) (77)

(54)

(46)
(40)
(-2.6 – - 0.9) (76)

(30) (25) (24) (21) (28)

(13)
(38)
(49) (229–780)

0.07 0.04

0.09 0.03 0.24

< 0.0001 0.02

0.04 0.28

0.50

0.88

(49) (27) (24)

(44) (13) (43) (77)

aData are N, % unless otherwise indicated
bAbbreviations: HAZ height-for-age z-score, IQR interquartile range, ART antiretroviral treatment c2 missing values
dVirologic suppression is defined as viral load ≤50 copies/ml

be provided by RUF had not been limited to 2000 kcal/d, i.e. 4 sachets per protocol, 15 participants would have been prescribed 5 to 6 sachets according to weight and age prescription bands.

At week 2, 31% of participants were sub-optimal RUF consumers which means they had consumed less than 50% of their RUF provision. Only 11% of participants re- ported they had consumed all the doses prescribed.

Overall, participants reported having consumed 61% (45–81) of their RUF at week 2. This proportion was stable (varying between 64 and 57%) throughout the follow-up among the remaining participants. The energy provided by RUF per kg of body weight was significantly higher in MAM participants than in their SAM counter- parts in both age group. However, energy intake from

RUF per kg of body weight was far lower than expected based on the prescription weight bands for all partici- pants; while being significantly higher in the younger participants than in their older counterparts (Table 3).

In both age groups, there was no difference either in proportions of sub-optimal RUF consumers or propor- tions of RUF intake and energy intake per kg of body weight according to the severity of malnutrition.

Factors associated with sub-optimal RUF intake

To identify factors associated with early sub-optimal in- take, we ran a stepwise logistic regression model with sub-optimal RUF intake (yes/no) as the dependent vari- able. Dislike of the taste of RUF (aOR = 5.0, 95% CI: 2.0– 12.3), HIV-non disclosure (5.1, 1.9–13.9) and food

Niasse et al. BMC Public Health (2020) 20:695 Page 6 of 12

Table 2 Perceptions and behaviors related to the management of RUF therapy among undernourished HIV-infected participants and their caregiver 2 weeks after enrolment in the SNACS study, Senegala-b

Caregivers

Responsible for RUF management Caregiver

Participant
RUF sharing with caregiver
RUF sharing with other adults RUF sharing with other children Participantsc
RUF perceived as a drug

Yes
No
Don’t know

Reason for RUF therapy Sickness

Malnutrition

Don’t know
Participant needs encouragement for RUF feeding Participant is disgusted by RUF
Participant hides to take RUF
Single sachet taken over several intake
Main mode of intake

Direct feeding from the sachet Diluted in gruel
Spread on bread

< 12 years

67 16 25 6 13

63 8 12

20 48 15 39 27 21 44

67 14 2

≥ 12 years All

(81) 29 (33) 96 (56) (19) 60 (67) 76 (46) (30) 27 (30) 52 (30) (8) 6 (7) 12 (7) (17) 11 (13) 24 (15)

(76) 83 (93) 146 (85) (10) 3 (3) 11 (6) (14) 3 (3) 15 (9)

(24) 26 (30) 46 (27) (58) 62 (71) 110 (64) (18) – – 15 (9) (47) 33 (37) 72 (42) (33) 24 (27) 51 (30) (25) 29 (33) 50 (29) (53) 36 (40) 80 (47)

(81) 79 (89) 146 (85) (17) 10 (11) 24 (14) (2) – – 2 (1)

P value <.0001

0.98 0.83 0.42

0.006

0.0005

0.19 0.42 0.29 0.10 0.18

aData are N (%)
bRUF ready-to-use food
cQuestions were asked to the participant (and/or to the caregiver if need be when participant ≤7 years)

insecurity (2.8, 1.1–7.2) were the major risk factors asso- ciated with early sub-optimal RUF intake (Table 4). There was no association between food insecurity and RUF taste appreciation, however, there was a significant interaction between these two variables in the multivari- able model (p = 0.001). Stratified analysis indicated that in the absence of food insecurity, proportions of sub-optimal consumers were similar whatever the RUF taste appreci- ation (Mantel-Haenszel Chi2 test probability =0.36), while

in food-insecure households, dislike of the taste of RUF was strongly associated with a sub-optimal intake (p < 0.0001).

Expectations, perceptions and experiences of participants in the nutritional intervention
The qualitative analysis involved a convenience sub- sample of 24 participants in Dakar of whom 9 were girls and 8 were enrolled as SAM (Table 5). Four FGDs were

Table 3 Adherence to RUF among HIV-infected participants 2 weeks after enrolment in the SNACS study, Senegal a-b

Indicators

Sub-optimal RUF consumers, N (%) % of RUF intake/provided
Energy provided/Kg of BW, Kcal/kg/d Energy intake/Kg of BW, Kcal/kg/d

aData are median (IQR) unless otherwise indicated bAbbreviations: RUF ready-to-use food, BW body weight

< 12 years N=84

24 (29)
64 (46–90) 61 (55–68) 39 (29–50)

≥ 12 years N=89

30 (34)
58 (42–74) 55 (51–61) 31 (23–41)

P-Value 0.46

0.07
< 0.0001 0.001

Niasse et al. BMC Public Health (2020) 20:695 Page 7 of 12

Table 4 Risk factors associated with sub-optimal RUF intake in HIV-infected participants 2 weeks after enrolment in the SNACS

Study.

a–b–c

, Senegal

Effects
Girls vs. boys

< 12 years vs ≥12 years Decentralized setting vs Dakar School level

None vs. secondary

Primary vs. secondary
HIV status undisclosed: yes vs. no
Food insecurity: yes vs. no
Disliking RUF taste: yes vs. no
Disliking RUF taste * Food insecurity
Caregiver responsible for RUF management vs. participant Participant needs encouragement to eat the RUF: yes vs. no Participant eats RUF in several vs. single feeding
SAM vs. MAM
Virologic suppressiond: no vs. yes

Univariable analysis OR 95% CI

1.7 0.9–3.2 1.2 0.6–2.2 2.6 1.4–5.0

3.5 1.0–12.0 1.8 0.5–5.8 4.9 2.1–11.2 2.8 1.2–6.3 3.6 1.7–7.8 _ _

2.2 1.1–4.1 2.0 1.0–3.6 1.8 1.0–3.3 1.9 1.0–3.5 2.4 1.2–4.6

P value

0.09 0.60 0.003

0.05 0.36 0.0002 0.01 0.001 _

0.02 0.04 0.06 0.05 0.01

Multivariable analysis aOR 95% CI

0.7 0.3–1.5 

5.1 1.9–13.9 2.8 1.1–7.2 5.0 2.0–12.3 

2.0 0.9–4.4

P value _
0.38

0.002 0.03
< 0.001 0.001

0.07

aAbbreviations: aOR adjusted odds ratio, CI confidence interval, OR odds ratio, RUF ready-to-use food, MAM moderate acute malnutrition, SAM severe acute malnutrition
bSub-optimal RUF intake is defined as if < 50% of RUF provided
cExplanatory variables are included at P < 0.20 in multivariate analysis, exited at P ≥ 0.10

dVirologic suppression is defined as viral load ≤50 copies/ml

conducted: two FGDs with participants who recovered within the study (one group with HIV-disclosed and one group with HIV-undisclosed participants), one in each Dakar study site and two FGDs with those who failed or defaulted using the same format. Discussions last 1 h30 on average. FGDs revealed that all the participants had

Table 5 Characteristics of HIV-infected participants included in the focus group discussions in the SNACS study, Senegala

similar expectations of the therapy. They were very con- cerned about their physical appearance, associating their thinness with being short and weak. The SAM partici- pants reported they had been frequently left out of school sports programmes or street football games due to their poor fitness. Participants expressed their strong motivation to gain weight when they started the study. Those who recovered from wasting demonstrated satis- faction and pride in having gained weight. Since youths generally define themselves through the eyes of others, they also perceived the signs of their body shape’s normalization in friends’ comments: ‘People tell me I’ve got big now! Big as a baobab tree! That I’ve got stronger and have calf muscles! That I’ve got more muscle!’ Boy, 9 years, HIV-undisclosed, recovery.

Most participants reported that those around them were not told that they were participating in a nutrition intervention; a strategy recommended by health workers to avoid RUF sharing. Managing the daily intake of RUF out of sight proved complicated in homes - sometimes polygamous - where there is practically no space for young people; in particular for boys, to have some privacy: ‘I used to hide on the roof terrace to eat (RUF) because I’ve got a little brother who spends the day asking questions about everything he sees. If I don’t give him some he tells his mother and she tells me to give him some.’ Boy, 12 years, HIV-disclosed, recovery. This daily effort by

Characteristics

Girl
Median age, year (min – max) Acute malnutrition at enrollment

Moderate

Severe
At school Mother orphan HIV disclosed On ART Outcome

Recovery Failure Default

aData are N, unless otherwise indicated

Participants N = 24

9
13.7 (7.4–17)

16 8 20 13 15 20

12 8 4

Niasse et al. BMC Public Health (2020) 20:695

Page 8 of 12

participants and caregivers to conceal consumption of RUF, however, is not always enough to avoid the some- times necessary intra-familial sharing: ‘My grandfather takes some from me, and sometimes he takes a lot. One day 12 sachets went missing.’ Boy, 16 years, HIV- undisclosed, failure, as well as extra-familial sharing, which is sometimes deliberate: ‘My friends used to ask me to give them some. I used to steal them from my grand- mother’s bedroom and I hid them like that (he gestures) and I used to give some outside.’ Boy, 11years, HIV- disclosed, recovery.

This secrecy, also motivated by the desire to divert friends’ curiosity and comments; even from well- meaning adults shows the secrecy that surrounds HIV infection and taking ARV: ‘My friends used to ask me what I was taking. I didn’t tell them anything. You want to know my secret but I’m not going to say - because I can’t say!’ Boy, 14 years, HIV-disclosed, recovery. ‘Some- times I meet people who tell me I’ve got bigger but I don’t tell them anything. I don’t want them to ask me.’ Boy, 8 years, HIV-undisclosed, failure.

Close follow-up visits at the hospital affected partici- pants badly, worried that these repeated absences might arouse suspicion at school: ‘It used to bother me because I was absent too often. In the end, my classmates used to ask me what I had; sometimes it was my teachers. Only my teaching assistant knew that I was sick. It was my mother who went there. I don’t really know what she told them. My classmates used to ask me all the time why I was going to the hospital. I used to change the subject.’ Girl, 16 years, HIV-disclosed, recovery. Repeated ab- sences could sometimes disrupt following classes and learning: ‘I lost marks because of my absences. Our les- sons are long and catching up classes is complicated. It’s difficult to re-copy notes. It makes me tired.’ Boy, 12 years, HIV-disclosed, default.

The main divergence observed between participants who recovered and those who did not concerned the perception of RUF. The disgust of RUF, occurring early on or overtime, was a major constraint to RUF adher- ence: ‘We have an illness [HIV-infection] and on top of that we have malnutrition. My friends are well-built; more than me. I feel that I’m malnourished and I am until now. At the start, I wanted it [RUF]. I forced myself; I used to take four (sachets) per day. I even liked it but over time it was the smell. It disgusted me.’ Boy, 16 years, HIV-disclosed, default.

Nausea, diarrhoea and vomiting whilst taking RUF were most often reported by unsuccessful participants for whom this represented a huge obstacle: ‘The doctor asked me why I wasn’t putting on any weight. I told him every time I ate some I felt I was going to vomit. He said ‘Oh really?’ I replied yes. It was me that decided to give up.’ Girl, 8 years, HIV-undisclosed, failure.

Participants who recovered reported the various strat- egies they implemented to cope with the growing fatigue they felt in adhering to the RUF prescription, which con- sisted in alternating between modes of intake, or for some of the oldest, to allow themselves to have breaks of one to 2 days in RUF feeding. One participant, although affected by the study constraints, reorganized his daily routine around visits to the clinic and RUF feeding: ‘I had to take three and I set the times: at 7:30 am I used to get up and take it. At 12:30 pm I used to leave the room to go into the bedroom. At 7:30 pm I used to go into the bedroom, heat the water, cut [the RUF] in the cup with water. I’d stir it and drink it. I think about my health and growth above all. I’ve got used to it.’ Boy, 13 years, HIV-disclosed, recovery.

Discussion

Most participants initially reported a positive organolep- tic appreciation of RUF. However, from the earliest weeks of the study, many of them needed encourage- ment from the caregiver for continued RUF feeding, and feelings of disgust were common. On average, partici- pants reported having consumed two-third of their provision throughout the study, despite the adjustments made to reduce the recommended prescription. The ra- tionale of the WHO prescription weight bands refers to an increase in the recommended nutritional intake levels for asymptomatic and malnourished HIV-infected people, which could reach 50 to 100% of energy needs in children [28]. The RUF is thus expected to provide a complete diet to a child with SAM with, in replacement of the habitual household diet, and cover additional energy needs due to HIV-infection [14]. Our results suggest that age- and weight-based prescription bands might have overesti- mated adolescents’ ability to consume RUF in “real life” and that there might be a threshold of RUF intake that some individuals cannot overcome due to nutritional density and sticky composition. Indeed, except for a few studies which proposed 2 sachets daily over short dura- tions, in which HIV-adults stated that they appreciated RUF [19, 20] or adapted to the taste after the first weeks [29], a leading reason for non-adherence was disgust and growing tired of having to consume 3 to 4 sachets per day over a long time [16–18].

At week 2, 31% of participants reported having con- sumed less than half of the RUF provided. Not surpris- ing, disliking the taste of the RUF was associated with sub-optimal RUF intake. This observation is consistent with studies involving HIV-infected adults reporting that many disliked the RUF taste and smell and that RUF taste and consistency made it difficult to swallow [17, 18, 29, 30]. Food insecurity is known as a major factor of poor adherence to food-by-prescription programmes as individuals are unlikely to refuse food ration sharing.

Niasse et al. BMC Public Health (2020) 20:695

Page 9 of 12

Interestingly, stratified analyses showed that participants disliking RUF were more likely to present sub-optimal intake when they lived in food-insecure rather than in food-secure households. In case of participant reluc- tance, caregiver commitment to support RUF feeding might have been lower in the most vulnerable families, which thus encouraged sharing. The absence of associ- ation between sharing and sub-optimal intake could sug- gest that merely occasional sharing, perceived as acceptable, was reported. Group discussions with adoles- cents also suggest that some sharing escaped the atten- tion of the caregivers.

We found HIV-non disclosure to be associated with sub-optimal RUF intake. This finding echoes our recent analysis in the same cohort reporting that HIV-disclosed participants showed a better understanding of the re- search information at enrolment in the present study [27]. Disease disclosure proceeds together with a better understanding of therapeutic issues, such as the import- ance of adherence to medication [31]. Such empower- ment might have similarly benefited the early adherence to nutritional intervention.

The only other study which assessed similar support based on Plumpy Nut® among children and adolescents, conducted in an HIV clinic in Mali, reported greater ad- herence figures as 88% of participants were rated as “good” RUF consumers [32]. Comparisons between the two studies are however limited as they differ in pre- scription procedures. Also, the Malian study did not ex- plore RUF sharing practice and acceptability of nutrition therapy and enrolled younger participants (< 15 years) with less severe acute malnutrition than we did.

The high prevalence of undesirable effects (nausea, vomiting and diarrhoea) attributed to RUF in this study is of concern as it hindered RUF acceptability and adher- ence in some participants. Qualitative studies also report similar undesirable effects attributed to RUF in adults on ART or starting ART and pregnant women [16, 17, 30]. In some of these patients, the initial undesirable effects were transient and attributed to pregnancy or drugs ini- tiation. RUF feeding advice provided by the study staff, although partially followed, are questionable as they may not have been feasible in practice: for example to distrib- ute RUF intake throughout the day and avoid mealtime. Many participants used to hide from others when eating RUF, most of the time in the caregiver’ bedroom, pri- marily to avoid being asked to share. Limited time to eat the RUF due to school and transportation and space constraints in the household combine to hinder the integration of feeding therapy into the usual food practices. In addition, many participants substituted their habitual breakfast with RUF or ate it at meal- time, with possible negative impacts on the dietary di- versity provided by family meals. All in all, these

constraints on RUF intake contribute to appetite sat- uration and decreased adherence.

This study has several limitations. First, the validity of information about both adherence and sharing practices is generally difficult to ensure, as some degree of socially desirable response bias might be expected. Counting the RUF sachets would have been a reliable method to con- firm participants’ reports, but the pilot study showed lit- tle compliance with the instruction among participants and caregivers to return with unused sachets at each visit. Based on our mixed data analyses, we hypothesize that some participants’ reluctance to eat the RUF might have resulted in consistent family sharing, rather than regular sharing which would have reduced RUF feeding. Second, although RUF adherence data were computed at each follow-up point, we considered only the first 2 weeks of consumption to run the risk factors analyses. Early RUF adherence, while participants went through an adaptation phase, might not be representative of the overall RUF consumption. However, this methodological choice is justified for several reasons: (i) this avoided analytical bias due to cohort attrition (the first recoveries and defaults occurred at week 2), (ii) as in other studies [21, 32], adherence to RUF was rather stable over time among the remaining participants, (iii) adherence data coincide with acceptability data collected once at week 2. Third, we recruited participants in Dakar for the FGDs as these central sites have a longstanding and routine prac- tice of hosting discussion groups as part of follow-up for older children and adolescents. By contrast, the regional sites had no comparable experience or practices. Although FGDs results may not be generalizable to all children and adolescents in Senegal, they highlight perceptions, behav- iours, and constraints surrounding RUF use that are likely to be very similar in decentralized settings. Last, attend- ance at follow-up visits was demanding for our partici- pants. However, because of our providing remuneration for travel costs, we documented very few missed or de- layed visits. Attendance to follow-up visit and continuity of RUF provision might need specific attention in the rou- tine care.

Palatability of imported RUF constitutes a major con- straint for the acceptability of nutritional therapy. Stud- ies comparing the acceptability of a novel and local product versus imported RUF (Plumpy Nut®) in HIV- infected or uninfected populations could not demon- strate the superiority of this alternative product so far [19, 33, 34]. Imported RUF is expected to remain by far the most widely used and easily available products for outpatient therapy. Therefore, an alternative lever for improving the acceptability of RUF therapy in routine care is to improve follow-up and support procedures.

The socially and physiologically acceptable RUF intake could be assessed with the young patient, and possibly

Niasse et al. BMC Public Health (2020) 20:695

Page 10 of 12

adjusted during therapy according to their progression and tolerance. It would also facilitate increasing the time between follow-up visits to monthly or every other month since adolescents would not be given consider- able RUF provision.

Compared to other age groups, HIV-infected adoles- cents exhibit higher rates of loss to follow up [35, 36], poor ART adherence [37, 38], and increased needs for psychosocial support [39]. Nutritional rehabilitation is thus an additional challenge to routine HIV care that re- quires adequate counselling and continuous support to ensure adherence, including preventing or lessening un- desirable effects. Particular attention should be paid to adolescents with SAM - who were also older in this co- hort- as they are likely to experience longer follow-up duration. This might lead to saturation and discourage- ment and strongly affect adherence to RUF-based proto- cols. Strengthening HIV clinic capacity and adolescent empowerment, through peer and community support and early HIV-disclosure process, could benefit nutri- tional interventions and improve the acceptability of RUF therapy.

Food-by-prescription is mostly used to focus on HIV- related wasting. Our study did not distribute food for the entire family to complement individual-targeted RUF therapy. Strategies to reduce sharing such as the rationalization of RUF prescription suggested above, the enrolment of other undernourished family members in RUF therapy, together with food complements for the household need to be evaluated.

Conclusion

This study conducted at the national level in Senegal highlights several acceptability issues of RUF therapy based on WHO guidelines in HIV-infected children and adolescents, which may lead to decreasing adherence to therapy protocols. Beyond the initial barrier of RUF taste in some participants, the main factors affecting adher- ence were closely related to RUF feeding constraints and HIV non-disclosure. There is an urgent need for differ- entiated approaches to address specific health and nutri- tional needs of HIV-infected adolescents. Tailoring prescription guidance, empowering young patients in their care are and innovating in how to more effectively support them in their nutritional therapy are crucial le- vers for improving acceptability of RUF-based therapy in routine care. As such, our findings are valuable in informing improved management of nutritional care in this vulnerable population.

Abbreviations

ART: Antiretroviral treatment; BMIZ: Body mass index z-score (for age and sex); FGD: Focus group discussion; IQR: Interquartile range; MAM: Moderate acute malnutrition; RUF: Ready-to-use food; SAM: Severe acute malnutrition

Acknowledgements

The authors gratefully acknowledge all the participating children and their caregivers. They sincerely thank David Masson for his valuable advice and input during the conduct of the study. They thank all the clinics’ staff who participated in the study:

(1) Centre Hospitalier National d’Enfants Albert Royer de Dakar (Senegal): Aminata Diack (associated investigator), Aicha Dia, Babacar Niang (study clinicians); Ndeye Ngone Have, Astou Dieye, Khady Sidibe (social workers); Oumy Fall (nurse);

(2) Hospital Roi Baudouin de Guediawaye (Senegal): Baly Ouattara (study clinician); Abdou Kader Niang (nurse); Alhadji Bassine Diom, Adama Ndour, Lamine Mohamed Souane (social workers);
(3) Centre hospitalier régional de Saint-Louis (Senegal): Justine Ngom (study clinician); Ndéye Fatou Diédhiou (social worker), Ndéye Penda Yade (laboratory);

(4) Centre hospitalier régional de Louga (Senegal): Georges Antoine BAZOLO (study clinician), Georgette Badji, Fatou Tanon (social worker), Cheikh Mbacké Gueye (laboratory);
(5) Centre hospitalier régional de Mbour (Senegal): Abdoul Magib Cissé (study clinician), Ismaila Camara (social worker), Made Masséne Séne (laboratory);

(6) Centre hospitalier régional de Kaolack (Senegal): Papa Birane Mbodj (study clinician), Issakha Camara (social worker), Maimounatou Diallo (nurse), Sérigne Modou Séne (laboratory);
(7) Centre hospitalier régional de Ziguinchor (Senegal): François Niokhor Diouf (study clinician), Pascal Faye (social worker), Mame Thiaré Sarr (laboratory);
(8) Centre hospitalier régional de Kolda (Senegal): Daouda Djiba (study clinician), Gabriel Diombraise Diatta, Kadiatou Diallo (social workers), Jules Beckenbauer Diatta (laboratory);
(9) Centre de santé de Thiès (Senegal): Ousmane Junior Dieng (study clinician), Daya Diallo (social worker), Bassine Ka (laboratory);
(10) Centre de santé de Nioro du Rip (Senegal): Christophe Kanfom (study clinician), Aliou Diallo, Khadidiatou Ba (social workers), Cheikh Ndiaye (laboratory);
(11) Centre de santé de Bignona (Senegal): Ibrahima Tito Tamba (study clinician), Yaya Diédhiou, Béya Aïssatou Coly (social workers), Bintou Diédhiou (nurse), Néné Diatta (laboratory);
(12) Centre de santé de Kolda (Senegal): Thierno Cherif Sy (study clinician), Mahamady Souané, Kamissa Ba (social workers), Oumy Ndiaye (nurse), Antoine Ndiaye (laboratory);

Authors’ contributions
CC and FN formulated the research question and designed the study; KD, SMD, AD and MV implemented the study; AD, FND, PBM and SMD collected and monitored the data; CC and MV performed the analyses; CC, MV and FN wrote the first draft of the manuscript. All authors read and commented on drafts and approved the final version of the manuscript.

Funding

Financial support was provided by Expertise France – Initiative 5pc of the Global Fund and Sidaction. The United Nations Child’s Fund and the World Food Programme country offices in Senegal provided Plumpy Nut® and Plumpy Sup®, respectively, within the study. MV received a fellowship for her PhD thesis from Agence Nationale de Recherche sur le SIDA et les hépatites virales (ANRS). Supporting institutions had any part in the study design, analysis or writing of this manuscript.

Availability of data and materials

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate

Ethics clearance for the SNACS study protocol was given by the Ethics and Regulatory Committee and the Ministry of Health in Senegal. All parents or surrogate caregivers provided written informed consent. Participants aged ≥7 years received extended information about the research and provided verbal assent.

Consent for publication