COMMENTARIES
Evidence-based Medicine
Viewpoint
Relevance:
Childhood malnutrition is a clinically and socially
significant problem in many resource-constrained
settings in the world. Besides affecting individual
children and families, it has far reaching
consequences on society in general. Naturally, its
alleviation depends on many factors beyond
nutritional supplementation (of children and their
families). A recent collaborative study [1] by a
group of researchers from the United Kingdom,
non-governmental organizations and Patna Medical
College, explored whether financial empowerment of
women in disadvantaged rural communities (through
the Rojiroti scheme) could impact the nutritional
status of their children. In this scheme, women form
self-help groups voluntarily, attend meetings
regularly, contribute a nominal sum weekly, become
eligible for very small loans from the pool of
collected funds, and after six months can obtain
loans up to Rs 3000 based on credit-worthiness.
There is no restriction on what the loan amount can
be used for. The investigators chose a cluster
randomized trial design to compare rural units
(described as tolas) that implemented the
Rojiroti scheme with tolas that did not.
Anthropometric measurements of children younger than
five years old in both trial arms were done at
enrolment, repeated after 18 months, and compared
between the arms. Although a research question was
not articulated by the investigators [1], it can be
framed as: What is the effect of community-based
Rojiroti microfinance scheme (I = Intervention) on
the nutritional status of under-five children
(O=Outcome), in economically and socially
disadvantaged communities in rural Bihar
(P=Population) compared to no microfinance scheme
(C=Comparator) at the end of an 18 month period (T =
timeframe of outcome assessment)?
Critical appraisal:
Box I presents a summary of the
trial design and main results. The investigators
chose a cluster RCT design to address the research
question. Technically, this is the ideal design to
evaluate efficacy of potential interventions in
clusters of individual participants, wherein the
effects (of the intervention) are
expected/anticipated to spill over into/onto those
who are not directly receiving the intervention, but
are present in the same cluster. However, if the
effect of the intervention is expected to have
limited impact on non-participating individuals in
the cluster, then an individual RCT is more
appropriate. It is difficult to judge which of the
two designs is superior to compare community effects
through individual empowerment of some members, as
was done in this trial [1].
Box I Summary of the
Trial
|
Study design:
Cluster randomized trial with allocation of
rural community units called tolas
into the trial arms. The intervention was on
women in the tolas, and the outcomes
were measured in their children.
Study setting:
Four tehsils of Patna district comprising
about 60 tolas. A tola is described
as a rural community with a population of
approximately 500 people with similar social
and economic background [1]. In general, the
communities appear to be disadvantaged as
evidenced by absence of health-care centres,
lack of access to piped water, low level of
women’s education, social empowerment and
economic status. However, all tolas
had electricity supply and immunization
coverage was over 95%.
Study duration:
Tolas were recruited in three phases,
during 2 months in 2012, two months in 2013,
and 1 month in 2014. No other details were
mentioned.
Inclusion criteria:
Sixty tolas were selected for
implementation of the Rojiroti scheme;
however, the basis of selection and/or
eligibility criteria were not mentioned. Any
woman in the intervention tola could join
the Rojiroti microfinance scheme. Women in
the Comparison group (i.e control) tolas
could not join the Rojiroti scheme, but
could join other (unspecified) self-help
group (schemes). All children <5 years in
the tolas selected for Intervention
and Comparison groups were eligible for
outcome measurement, whether (or not) their
mothers availed the Rojiroti scheme.
Exclusion criteria:
None were described.
Enrolment process:
The basis for selection of tolas was
not specified. Tolas of similar size
(definition unspecified) but at least 15 km
apart, were paired, and randomly assigned to
either the Intervention or Comparator group.
Enrolment of tolas occurred in three
phases viz 2 months in 2012 (20 tolas
included), 2 months in 2013 (30 tolas
included) and 1 month in 2014 (6 tolas
included). Women in the Intervention
tolas were invited to join the Rojiroti
scheme through a “show of hands” and their
children were enrolled with verbal consent.
Intervention and Comparison
groups:
The Rojiroti scheme was implemented in the
intervention arm tolas. Nothing was
done in the comparison arm tolas.
Baseline demographic parameters of the
tolas, participating women and their
children were recorded in both groups.
Anthropometric measurements of all
under-five children were done using standard
tools and methods, at baseline and also
after 18 months; in both arms of the trial.
Outcomes:
All outcomes were measured 18 months after
enrolment, and compared between the two
trial arms (outcomes are listed in the
last row). Definition of two of the
secondary outcomes was not provided in the
article viz., proportion of women
with freedom to travel without permission of
a male relative, and forced asset sale.
Follow-up protocol:
Research staff conducted anthropometric
measurements in all children available 18
months after enrolment of tolas,
irrespective of whether the children and/or
their mothers participated in the trial.
Sample size:
A priori sample size calculation was
performed for a superiority trial, to detect
a 0.26 z score improvement in WHZ from an
estimated baseline of -0.96, with alpha
error 0.05 and beta error 0.20. Assuming 10%
attrition, the estimated sample size was
reported as 60 tolas. The
investigators observed approximately 40
under-five children per tola initially,
hence assumed that there would be
approximately 2400 children for
anthropometric measurements across the 60
tolas.
Data analysis:
Data of available children were analysed
between trial arms, calculating unadjusted
odds ratio. Subsequently odds ratio was
adjusted for baseline nutritional status,
age, gender and number of under-five
children per family. It was decided post hoc
to compare the outcomes in children in
Intervention tolas whose mothers did
(versus did not) join the Rojiroti scheme.
Comparison of groups at baseline:
•
The tolas in the two arms were
comparable for multiple parameters viz
connection to a paved road, distance from a
main road, presence of public distribution
scheme shop, presence of government primary
school, presence of other school,
availability of primary health centre,
access to ASHA worker and ANM, availability
of piped water supply, and electricity.
•
Participating mothers in the two arms
were comparable in terms of the number who
joined the Rojiroti scheme and age. However,
there were statistically significant
differences in terms of family land
ownership, freedom to travel without
permission, ability to read/write, and
school attendance- all in favour of those
who were in the Intervention arm.
•
Children in the participating
tolas were comparable in terms of median
number enrolled per tola, gender
distribution, mean age, proportion delivered
at home, immunization status, and proportion
having road-to-health cards. Most
anthropometric parameters were comparable
between arms, however HAZ and the proportion
of children with MUAC <12.5cm, were both
significantly better in the intervention
arm. However, proportion with wasting was
significant higher in the Intervention arm.
Summary of results: Intervention
versus Comparison arms
Primary outcome
•
Mean (SD) WHZ: -1.02 (1.11) vs
-1.37 (1.10), uOR* 0.16, 0.61; aOR** -0.03,
0.53
Secondary outcomes
•
Mean (SD) HAZ: -2.37 (1.29) vs
-2.53 (1.25), uOR -0.04, 0.37; aOR -0.24,
0.10
•
Mean (SD) WAZ: -2.13 (1.03) vs
-2.37 (1.05), uOR 0.11, 0.43; aOR 0.04, 0.49
•
Mean (SD) MUAC: 13.6 (1.1) vs
13.4 (1.1), uOR 0.03, 0.40; aOR -0.14, 0.38
•
Proportion with wasting: 18% vs
29%, uOR 0.28, 0.74; aOR 0.33, 1.14
•
Proportion with stunting: 63% vs
66%, uOR 0.60, 1.12; aOR 0.57, 1.64
•
Proportion with underweight: 53%
vs 63%, uOR 0.47, 0.84; aOR 0.29, 0.89
•
Proportion with MUAC <12.5 cm:13%
vs 18%, uOR 0.41, 1.05; aOR 0.27, 2.23
•
Proportion with MUAC <11.5 cm: 3%
vs 5%, uOR 0.36, 1.33; aOR 0.10, 6.14
•
Proportion of women with freedom to
travel‡: 5% vs 5%, uOR and aOR not specified
•
Forced asset sale: 2% vs 2%, uOR and
aOR not specified
|
*uOR is the 95% CI of the unadjusted odds ratio;
**aOR is the 95% CI of the adjusted odds
ratio;
‡travel without permission of a male
relative. |
The investigators used a computer program for
randomizing pairs of tolas, although since
only two tolas were randomized at a time,
simple coin tossing is sufficient. Paired
randomization obviated the scope for allocation
concealment. The outcome assessors were not blinded
to the intervention, but the reasons for this were
not specified.
Although, all 56 enrolled tolas were present
at the end of the study (i.e., zero
attrition), there was significant attrition amongst
individual participants (both mothers and children),
between the randomization (i.e., enrolment)
step, baseline variable measurement step and outcome
assessment step. For example, 2469 children were
eligible for anthropometric data assessment across
56 tolas at enrolment, but WHZ data could be
analysed in only 1718 (69.6%). Similarly, 2064
children were eligible for outcome measurement at
the end of the study, but WHZ (primary outcome)
could be analysed in only 1377 (66.7%). These
attrition rates are considerably high, although they
were comparable between the two groups. Further, it
is disconcerting that one-third of the potential
data was unavailable not because participants
dropped out, but because the anthropometric data
were not collected properly. This is unacceptable in
a well-funded RCT with appropriate training of
research staff.
It is also unclear what
proportion of the children whose baseline data were
collected, underwent data collection at the end of
the study. This has two entirely different
implications. First, if these proportions are
significantly different between the two trials arms,
a new confounding variable emerges. Unfortunately,
the authors did not show this data. Second, if the
intervention (i.e., implementation of
Rojiroti micro-finance scheme) is believed to impact
the whole community (and not just the participating
households), then we would expect to see the
benefits in children irrespective of whether or not
they were present when the intervention started or
whether their families availed the scheme. This
seems to have been the assumption of the
investigators in this study [1]. But if this is the
case, it can be argued that pre and post
intervention measurement of anthropometric
measurements would be more meaningful than
comparison between trial arms.
This raises another important issue. The
statistically significant ‘benefits’ in the
Intervention arm were not because children in this
arm showed improvement in anthropometric
measurements (as one would expect). In fact, 5 of 11
outcomes showed worsening over the 18-month
intervention period. These include mean HAZ
(declined from -2.00 to -2.37), mean WAZ (declined
from -1.89 to -2.13), proportion with stunting
(increased from 49% to 63%), proportion with
underweight (increased from 44% to 53%) and
proportion of mothers with freedom to travel without
permission (declined from 8% to 5%). Even the other
anthropometric measures showed no improvement, but
merely remained unchanged over 18 months. Thus, the
Intervention arm was proven superior [1] only
because the Comparison arm showed far greater
worsening of anthropometric parameters. The authors
interpreted this as empowerment of the community to
be resilient during food shortage, thus emphasizing
the benefit of the Intervention. However, this
explanation is unacceptable for three reasons.
First, it assumes that under natural circumstances,
children’s nutritional status declines over time.
However, the authors showed no data supporting this
presumption [1]. Second, the proportion of
households forced to sell assets was exactly 2% in
both arms, suggesting that apparent periods of food
shortage did not translate to loss of assets in
either arm. Third, analysis of the reasons for
taking loans in the Intervention arm shows that a
very small proportion was used for food and supplies
(in terms of percentage as well as absolute amount).
How to explain the differences in the two arms at
the end of the trial? One explanation could be that
mothers in the Intervention arm were more empowered
than mothers in the Comparison arm (literacy 21%
vs 16%, school attendance 19% vs 13%, and
freedom to travel without permission 8% vs
3%, and family land ownership 13% vs 8%).
Perhaps this could account for better child-care
practices even in the midst of acute shortages,
thereby preventing the pattern of decline seen in
the Comparison arm. However, these empowerment
indicators were present in less than 20% mothers in
the Intervention arm; hence, other unexplored
factors are likely. Had the authors re-collected
maternal baseline parameters at the end of the
study, a clearer picture of women empowerment (if
any) could be considered.
It should be remembered that children in the
Intervention arm had superior HAZ than those in the
Comparison arm. The impact of this on the final
outcome is unclear, although height is impacted much
later than weight and muscle mass, during food
deprivation.
A noteworthy point is that the authors [1] did not
report the number of deaths, or medical morbidities
amongst the children in either arm. Thus, the data
presented pertain only to survivors. It is
well-recognized that children with worse nutritional
state have greater likelihood of morbidity and
mortality. Thus, the available children no longer
represent all the eligible children. This
compromises internal validity. It can be further
argued that all-cause mortality data alone may be
insufficient, and malnutrition-related morbidity
should also have been measured.
The investigators reported that children in
Intervention tolas had similar anthropometric
outcomes, irrespective of whether or not their
mothers participated in the Rojiroti scheme. They
suggested that this indicated some kind of community
effect spilling over into non-participating
households. However, the proportion of participating
women in each tola were not described, hence this
assumption could be too simplistic.
It appears that 90% children in each arm of the
trial possessed road-to-health cards. These cards
provide valuable longitudinal anthropometric data.
This would have enriched the study by providing some
data for drop-out children, internal checks against
spurious data collected in the study, and also a
‘last recorded’ value for those older than 5 years
at follow-up. Most important, the inflection time
point(s) at which nutritional decline occurred could
have been calculated.
Another missed opportunity in this study is that
data were not analyzed in age bands, rather all
under-five children were clubbed together and
treated as single unit. This is important because
growth rates vary by age in under-five children.
As in many such studies, interesting data emerged
that were not the focus of the investigators. For
example, more than 95% children in both arms were
immunized [1]. This is somewhat surprising,
considering that the overall immunization coverage
(with BCG, 3 DPT, and measles vaccine among
12-23-month-old children) in Bihar during 2015-16
was 61.7%, coinciding with the national average of
62.0% [2]. How did the included children have such
excellent immunization coverage? This could be
because over 90% tolas had access to ASHA
workers as well as ANM in their community. Or
perhaps the reported immunization used some other
definition of immunization, or data were collected
unreliably. Since the baseline nutritional
indicators of children in terms of proportions with
stunting, wasting and underweight coincided with the
overall NFHS-4 data for Bihar [2], the latter
assumptions are more likely.
Each tola had only 500 people and around 40
under-five children. Although the age break-up of
the tolas is not known, India’s population
pyramid suggests just under 10% of the population is
in the age group 5-9 years [2]. This would translate
to about 50 primary school age children in each
tola. It is therefore impressive that all tolas
had a primary school and some had other schools as
well.
Conclusion:
This cluster RCT [1] suggested that participation of
disadvantaged rural women in a specific microfinance
scheme could prevent decline in the anthropometric
measurements of their under-five children over a
period of 18 months. However, the validity of the
trial is compromised by methodological issues and
compromised power due to significant attrition.
Hence it is difficult to draw firm conclusions from
this trial or recommend further similar studies.
Funding:
None; Competing interests: None stated.
Joseph
L Mathew
Department of Pediatrics,
PGIMER, Chandigarh, India.
Email:
[email protected]
References
1.
Ojha S, Szatkowski L, Sinha R, Yaron G,
Fogarty A, Allen SJ, et al. Rojiroti
microfinance and child nutrition: a cluster
randomized trial. Arch Dis Child. 2020; 105:229-35.
2.
Government of India, Ministry of Health and
Family Welfare. National Family Health Survey-4
2015-16. Available from:
http://rchiips.org/NFHS/pdf/NFHS4/India.pdf.
Accessed April 13, 2020.
Pediatrician’s Viewpoint
India is home to about one third of the stunted and
half of the wasted under-five children present
globally [1]. Malnutrition attributes to about 70%
of the under-five deaths in India during 2017 [2].
Apart from morbidities and mortality, malnutrition
is a key determinant for optimal cognitive growth
and development and overall health and productivity
in adulthood [3]. The UN Sustainable Development
Goal-2 targets elimination of child malnutrition by
2030 [4]. Child health and nutritional status is
reflecting a socioeconomic gradient [5]. The
economic growth in recent times has not optimally
transformed into reduction in childhood malnutrition
[6].
India has been making efforts towards reducing the
burden of malnutrition and the health adversities
through various programs including the nutrition
supplemen-tation and nutrition rehabilitation
centers. Recently National Nutrition Mission (NNM,
also called Poshan Abhiyan) has been initiated by
Government of India, which targets reducing
stunting, undernutrition, anemia and low birth
weight by 2%, 2%, 3% and 2% annually, respectively
by 2022 [7]. Globally, several efforts in past have
targeted the nutritional status of children and
women through various livelihood, agricultural and
conditional cash transfer systems with varied
results [8].
The current study documented the impact of the
Rojiroti microfinance effort through Self help
groups in Patna district, Bihar over 18 months
period [9].
Although this study was conducted in Bihar,
the context and underlying factors are applicable to
several parts of India. Malnutrition is a constant
challenge for the pediatrician. In clinical
practice, the pediatricians assess nutritional
status and give nutritional counselling including
breastfeeding, but the real change in family
practice and nutritional status dependents on the
food security, availability and home food
environment. Research from India revealed the roles
of social and economic competing forces for
persistence of undernutrition [10]. Although this
article does not include clinical dimension, but has
relevance for the pediatricians and child health and
nutrition functionaries.
The pediatricians have multiple opportunities and
roles to play in this context for all categories of
clients, especially those from the weaker social and
economic strata. Age-appropriate counselling and
empowerment of the parents and families for
preventive care including nutritional practices
(breastfeeding, weaning and complementary feeding,
especially targeting the locally available
nutritious foods ingredients), routine immunization,
vitamin A and deworming schedule and general hygiene
and sanitation at household level must be practiced
by all pediatricians. Rational medication and
supplementation prescription practice can be
critical in minimizing the out of pocket expenses
for the families. Apart from the prescription,
appropriate counselling for medicine and
supplementation adherence and continued feeding
during and after the illness are to be emphasized.
The pediatricians also have a stewardship role in
healthcare financing. While major share of the
curative healthcare services is provided by the
private sector, the preventive services are
delivered by public sector. Out of pocket
expenditure (OOPE) amounts to about 75% of
healthcare expenditure in India and the catastrophic
healthcare expense is an important cause of
impoverishment for the families [11]. The
catastrophic health related OOPE was also observed
in the current study [9]. Thus, the treating
pediatrician has a responsibility to understand the
financial implications of their clinical decisions.
The communication by the pediatricians to be
effective for parents, family and the community, it
must be clinically appropriate, transparent and
sociocultural context compatible.
Funding: None; Competing interests:
None stated
Manoja
Kumar Das
The INCLEN Trust International
New Delhi, India.
Email:
[email protected]
References
1.
Development Initiatives. 2018 Global
Nutrition Report: Shining a light to spur action on
nutrition [Internet]. Development Initiatives
Poverty Research Ltd, Bristol, UK; 2018. Available
from: https://globalnutrition
report.org/reports/global-nutrition-report-2018/.
Accessed April 10, 2020.
2.
Institute for Health Metrics and Evaluation.
GBD Compare data visualization [Internet]. Institute
for Health Metrics and Evaluation, University of
Washington, 2018. Available from:
http://www.healthdata.org/data-visualization/gbd-compare.
Accessed April 10, 2020.
3.
Victora CG, Adair L, Fall C, Hallal PC,
Martorell R, Richter L, et al. Maternal and
child undernutrition: Consequences for adult health
and human capital. Lancet. 2008;371:
340-57.
4. United Nations Department
of Economic and Social Affairs. Sustainable
Development Goal 2. [Internet]. United Nations;
2019. Available from:
https://sustainable-development.un. org/sdg2.
Accessed April 10, 2020.
5.
WHO Commission on Social Determinants of
Health, World Health Organization, editors. Closing
the gap in a generation: health equity through
action on the social determinants of health:
Commission on Social Determinants of Health final
report. Geneva, Switzerland: World Health
Organization, Commission on Social Determinants of
Health; 2008.p.246.
6. Subramanyam MA, Kawachi I,
Berkman LF, Subramanian SV. Is economic growth
associated with reduction in child undernutrition in
India?
PLoS Med. 2011;8: e1000424.
7. Ministry of Women and Child
Development, Government of India. Poshan Abhiyan.
PM’s overarching scheme for holistic nourishment
[Internet]. Available from:
https://icds-wcd.nic.in/nnm/home.htm#. Accessed
April 10, 2020.
8. Pandey VL, Mahendra Dev S,
Jayachandran U. Impact of agricultural interventions
on the nutritional status in South Asia: A review.
Food Policy. 2016; 62:28-40.
9. Ojha S, Szatkowski L, Sinha
R, Yaron G, Fogarty A, Allen SJ, et al.
Rojiroti microfinance and child nutrition: A cluster
randomised trial. Arch Dis Child. 2020;105:229–35.
10. Chaturvedi S, Ramji S,
Arora NK, Rewal S, Dasgupta R, V Deshmukh, et al.
Time-constrained mother and expanding market:
Emerging model of under-nutrition in India. BMC
Public Health. 2016;16:632.
11. Balarajan
Y, Selvaraj S, Subramanian S. Health care and equity
in India. Lancet. 2011; 377:505-15.
Nutritionist’s Viewpoint
Childhood malnutrition is increasingly recognized as
an important public health problem, for its adverse
effect on health and child survival, as well as for
long term growth and development. India is at the
epicentre of this global public health problem, with
22 million children wasted and over eight million
severely wasted at any one time [1]. Hence
implementation of evidence-based strategies for
prevention and management is topmost priority for
increasing child survival and productivity.
Malnutrition is a complex and multi-dimensional
issue, affected by poverty, inadequate food
consumption, inequitable food distribution,
suboptimal infant and child feeding and care
practices, equity and gender imbalances, poor
sanitary and environmental conditions and limited
access to quality health, education and social
services. Social protection involves policies and
programs that protect people against vulnerability,
mitigate the impacts of shocks, improve resilience
and support people whose livelihoods are at risk.
Social protection programs can improve food security
at household level, quality, and diversity; decrease
undernutrition; and help children reach their full
potential [2].
In the present study [3], authors assessed the
effect of microfinance initiative on the nutrition
status of children in a marginalized population. The
study suggests that, though micro-finance has been
able to reduce the deterioration in nutrition levels
in the children of extremely poor families, it has
not been able to actually improve or even maintain
the nutrition levels. This may be due to the
low-income gains from the scheme. From a policy
point of view, there are important conclusions which
may be drawn. One inference which may be drawn is
that schemes which directly tackle malnutrition and
help provide food to children must be continued and
encouraged in the poor states of the country.
Many nutrition-specific interventions to prevent
wasting and other forms of malnutrition are
delivered at community-level in India through
Anganwadi Services under the umbrella of the
Integrated Child Development Services (ICDS) scheme.
Nutrition-specific interventions are also delivered
during the VHSND or on separate days, including
growth monitoring, the promotion and support of
infant and young child feeding (IYCF), micronutrient
supplementation and supplementary feeding. While the
schemes, programs and delivery platforms are
nationwide in scale, the coverage, and quality of
interventions are insufficient to achieve the impact
required. It would be beneficial to channelize
efforts and funding to boost the efficacy of such
schemes.
To conclude, while microfinance schemes have their
own importance, they may not be the way to address
nutritional issues among children. A truly
multi-sectoral approach will achieve optimal
nutrition outcomes through greater coverage.
Funding:
None; Competing interests: None stated.
Praveen Kumar
Department of Pediatrics
LHMC & KSCH, New Delhi, India.
Email:
[email protected]
References
1.
UNICEF, WHO & World Bank Group 2018. Joint
Child Malnutrition Estimates 2019 Edition. Available
at http://data.unicef.org/resource/jme/.
Assessed on April 14, 2020.
2.
Ruel MT, Alderman H; Maternal and Child
Nutrition Study group. Nutrition-sensitive
interventions and programmes: How can they help to
accelerate progress in improving maternal and child
nutrition? Lancet. 2013; 382:536-51.
3.
Ojha S,
Szatkowski L, Sinha R, Yaron G, Fogarty A, Allen SJ,
et al. Rojiroti microfinance and child
nutrition: A cluster randomised trial Arch Dis
Child. 2020; 105:229-35.