review on child health priorities |
|
Indian Pediatr 2011;48:
183-189 |
|
UNICEF-PHFI Series on Newborn and Child
Health, India: Methodology for Systematic Reviews on Child
Health Priorities for Advocacy and Action |
Joseph L Mathew, *Dheeraj Shah, ¶Tarun
Gera, $Siddhartha Gogia, **Pavitra Mohan,
$$Rajmohan Panda, $$Subhadra Menon, and *Piyush
Gupta
From Advanced Pediatrics Center, PGIMER, Chandigarh;
*University College of Medical Sciences, New Delhi;¶Fortis
Hospital, New Delhi; $Max Hospital, Gurgaon, Haryana; **UNICEF,
India; and $$Public Health Foundation of India, New Delhi,
India.
Correspondence to: Joseph L Mathew, Advanced Pediatrics
Center, PGIMER, Chandigarh 160 012, India.
Email: [email protected]
|
India is committed to reducing childhood
mortality and morbidity. This requires evidence-based policy and practice
in the realm of public health. This in turn necessitates advocacy and
action (among all stakeholders), focused on locally relevant issues. A
collaboration to work towards this goal was forged between the Public
Health Foundation of India (PHFI), United Nations International Children’s
Emergency Fund (UNICEF), India; and a team of independent researchers. As
a first step, a systematic review of literature on four priority areas of
newborn care (community-based interventions) and child health (acute
respiratory infection, diarrheal disease, anemia), was undertaken to
address important issues including epidemiology, interventions for
management, and operational issues of planning, implementing, and
measuring actions at a programmatic level. This paper describes the
development of the methodology for undertaking these systematic reviews
including the process for framing of research questions, building a
research team, and executing the systematic review (literature search
strategy, data extraction, analysis, and reporting). The challenges
associated with ensuring robust methodology, are also described.
Keywords: Action, Advocacy, Anemia, ARI, Child health, Community
based newborn care, Diarrhea, India, Methods, Newborn, Systematic reviews.
|
Reducing childhood mortality and
forging improvement in Newborn and Child Health (NCH) are among India’s
most important public health goals [1]. The importance of generation of
evidence to inform policies and programs for achieving sustainable gains
in line with our country’s goals as well as the global Millennium
Development Goals [2] cannot be over-emphasized. In the era of
evidence-based health-care and informed health-care policy-making, all
initiatives should be developed on a solid foundation of evidence. The
Public Health Foundation of India (PHFI) collaborated with UNICEF to
generate an agenda for advocacy and action on key technical and
programmatic issues related to newborn and child health survival in India.
It was proposed to generate the evidence through systematic reviews,
discussions with experts and consultation with stakeholders. Based on
initial discussions in January 2010, four key areas were prioritized for
the review: community based newborn care, childhood anemia, acute
respiratory infections (ARI) and diarrheal diseases. This paper describes
the methodology employed for the identification, synthesis and collation
of evidence in these four key areas of newborn and child health in India.
The nature of the questions addressed (which included
technical as well as operational issues) and the intended purpose (to
pursue advocacy and action at a programmatic level) necessitated the
development of a separate methodology for this set of systematic reviews.
Standard systematic review methodology (based on conventional literature
searching) is designed to answer clinical questions (particularly
interventions) in health-care, hence was not expected to be sufficient for
this review.
The Review Process
Formulating Review Questions
A two-day consultative meeting of research experts was
organized on 9-10 February 2010 at New Delhi, under the aegis of UNICEF,
New Delhi, India and PHFI, to formulate questions for review from each of
the 4 thematic areas. Five to six experts per thematic group were invited
to formulate and refine the review questions. The approach was aimed to
systematically and comprehensively list the relevant questions, and give
equal treatment to questions in different domains: epidemiology,
preventive and therapeutic interventions, health systems and health
policy. Responses from these experts were discussed in a Delphi like
process to refine, sequence and prioritize the review questions [3,4]. The
final list of questions for each thematic area was generated summarizing
the input from the experts, following the discussions for refining each
research question, under a moderator. Within each question, one or more
sub-questions was/were also included. We grouped the final list of
questions under two domains: Technical issues; and Operational issues.
Since many of the operational issues were cross-cutting, these were
further refined to maintain uniformity across the thematic areas. The
final list of questions generated is presented in Tables I-IV.
Table I
Acute Respiratory Infections (ARI): Questions Reviewed
Technical issues |
Operational issues |
• What is the current status (and trends) of
morbidity and mortality from childhood ARI. Identify
risk/predisposing factors contributing to burden of ARI/mortality
due to pneumonia. What proportion of ARI cases have pneumonia/severe
pneumonia?
• What is the common etiology of community
acquired ARI, pneumonia and severe pneumonia? (India specific data
and trends)? What is the anti-microbial susceptibility pattern and
mechanism of surveillance for antibiotic resistance?
• What are the current National guidelines for
management of ARI (both at community and facility levels)? What is
the role of potential interventions (case finding and community
based management, antibiotics, zinc, vitamin A, measles, pertussis,
Hib and pneumococcal vaccine, supportive management and oxygen,
environment) that may result in reduction of mortality/prevention of
pneumonia?
• What proportion of ARI cases have wheeze
(audible/auscultation)? Is it predictive of severe
pneumonia/mortality/hypoxia? What proportion of cases with wheeze
respond to bronchodilator therapy? What is the impact of treating a
wheeze related ARI with/without bronchodilators?
• What are the client practices regarding
management of ARI (care seeking, home remedies)?
• What are the current prescription practices for
antibiotics, cough remedies, bronchodilators etc. in management of
ARI? |
• What are the existing
strategies/policies/programs at national/State levels? What are the
experiences? What are the indicators for monitoring/evaluation? And
what are the barriers (both from and within ‘System’ and ‘Demand’
side, and also technical/managerial) to implementation of standard
ARI case management? (these will include procurement, distribution,
advocacy, managerial spects, KAP, compliance etc).
• What are the experiences with integrated vs
vertical approach (in program mode)?
• What is the role of community health worker in
ARI/pneumonia control program? Are they equipped and empowered to
manage ARI? How do they perform after training? (Global
experiences)
• What is the feasibility of diagnosis and management of wheeze
in the community?
|
Table II
Diarrheal Disease: Questions Reviewed
Technical issues |
Operational issues |
• What is the magnitude/trend of diarrheal
morbidity (incidence) and mortality in India?
• What is the etiology (and trends therein) of
diarrhea, including neonates and young infants?
• What are the current National guidelines for
management of diarrhea including the role of low osmolarity ORS and
Zinc in management of diarrhea?
• What is the role of other interventions
(breastfeeding, immunization, vitamin A/zinc supplementation, hand
washing, environmental modification) in prevention of diarrhea?
• What are the client practices regarding
management of diarrhea (ORT, feeding, care seeking)?
• What are the current prescription practices for
ORS/ORT, Zinc, antibiotics, pre-probiotics, antisecretory agents,
antiemetics, other antidiarrheal drugs (metronidazole etc) in
management of diarrhea?
|
• What are the existing
strategies/policies/programs at national/State levels? What are the
past and present experiences? What are the indicators for
monitoring/evaluation? And What are the barriers (both from and
within ‘System’ and ‘Demand’ side, and also technical/managerial) to
implementation of the program? (these will include issues related to
procurement, distribution, advocacy, managerial aspects, KAP,
compliance etc.)
• What are the experiences with integrated vs.
vertical approach (in program mode)?
• What are the use rates of ORS and Zinc and
barriers for their use?
•
Formulation, branding, prescription, dispensing, quality,
manufacturing capacity, social marketing, and availability of
ORS/zinc
•
IEC, advocacy material, capacity building
•
Private sector/ AYUSH involvement
•
Current evidence/experience related to compliance of zinc use for
10-14 days.
• How to promote large scale use of ORS and Zinc
rapidly? (Global Experiences) |
Table III
Childhood Anemia: Questions Reviewed
Technical issues |
Operational issues |
• What is the prevalence of childhood anemia in
India and which are the groups at risk? Identify the target groups
for prevention and control of anemia.
• What are the major etiologic factors
contributing to anemia in children in reference to various regions
of India? (a) Nutritional iron, B12, FA, dietary habits,
bioavailability, PEM, others; (b) Maternal anemia (life cycle
approach); (c) Helminth infestation, H. pylori; (d)
Malaria; and (e) Hemoglobinopathies.
• What are the functional and economic
consequences of anemia in children, specially on (a) Mental and
motor development; (b) Physical growth; (c) Physical Capacity; (d)
Under-5 mortality; and (e) Infection?
• What are the available strategies/models for
anemia control and prevention, including (a) Maternal anemia control
programs (Life cycle approach); (b) Exclusive breastfeeding; (c)
Deworming; (d)Supplementation programs (Dose, Duration,
Formulations, Daily vs. intermittent, Iron vs. multimicronutrient);
(e) Food based approach; and (f) Fortification? What is the
effectiveness of different strategies to prevent and control anemia
among children, particularly with reference to impact on functional
consequences?
• What are the adverse effects and safety aspects
of supplementation programs? |
• What are the existing
strategies/policies/programs at national/State levels? What are the
experiences? Is National Anemia control program sufficient in
concept and implementation? Was it adequately managed? Whether the
program has been sufficiently evaluated and improvised/modified
based on the evaluation feedback? Has the cost effectiveness
ascertained? What are the indicators for monitoring/evaluation? And
What are the barriers (both from and within ‘System’ and ‘Demand’
side) to implementation? (these will include procurement,
distribution, advocacy, managerial aspects, KAP, compliance etc.)
• What are the experiences with integrated vs.
vertical approach (in program mode)?
• What can be done to scale up the effective anemia prevention
and control strategies?
|
Table IV
Community-based Newborn Care: Questions
Reviewed
|
|
• What is the magnitude of neonatal deaths in
India and what is the etiology. Are there any changing trends?
• What are the packages of interventions
(Typology? Timing?) that have been used for community newborn care?
• What has the effect of interventions been in
terms of outcome – mortality, neonatal care practices and health
seeking behavior?
• Are the studies comparable in terms of
interventions, baseline neonatal mortality, training of health
workers, type of workers, supervisory support, facility and referral
support and the population size of the area where intervention took
place?
• What are essential elements of home and
community based care that makes the most difference to neonatal
health outcomes? |
• What are the existing
strategies/policies/programs at national/State levels? What are the
experiences? What are the indicators for monitoring/evaluation? And
what are the barriers (both from and within ‘System’ and ‘Demand’
side) to implementation? (these will include procurement,
distribution, advocacy, managerial aspects, KAP, compliance etc.)
• What is the profile of the service provider in
different models? What training methods have been used for health
workers/supervisors for community newborn care interventions? Which
skills have poor retention amongst workers?
• Under what circumstances are these
interventions likely to have an impact? (Supervision, Logistics and
Support systems, Motivation of health providers)
• Have any of these interventions in studies been
taken to scale and do we have a measure of the effectiveness? What
are the difficulties in taking these interventions to scale? Which
of the interventions within the package have been difficult to
implement? Cost effectiveness? |
Research Team
To identify current evidence on the subject, a research
team comprising of one Coordinator, four Authors and four Reviewers (one
per topic area) was constituted. The team also included experts from
UNICEF, New Delhi and the Public Health Foundation of India.
Formulating Methodology
Following identification of questions, the Authors
developed a research protocol building up on the initial suggestions
provided by the experts. The initial protocol was prepared on 23 Feb 2010;
it was modified, refined, revised and finalized on 31 March 2010. Pilot
testing of this version was undertaken by each Author and the results
presented in a Consultative Meeting held on 6 April 2010. At this meeting,
the methodology for the systematic reviews was presented to all the
members of the research team including UNICEF and PHFI partners. Based on
feedback received, minor modifications were made which was then finalized.
The Methodology is briefly described below (also see
Web Table).
Each systematic review was undertaken by the respective Author based on
the Methodology agreed upon.
Sources of Literature
To address the technical issues (epidemiology, risk
factors, efficacy of interventions, client and professional practices,
etc), the primary databases employed were Medline through Pubmed and the
Cochrane Library (http://www.thecochranelibrary. com/view/0/index.html).
It has been shown that a brief RCT search strategy comprising these
databases is generally sufficient to locate trials for systematic reviews
in most cases and exhaustive searching is no longer regarded
cost-effective, if the most important databases have been covered [5].
However, it has also been recommended that search to other sources,
particularly regional databases is a preferred option [6]. Therefore,
IndMed (http://indmed.nic.in/) was also included as a primary
database. The Authors also accessed specific sources to address specific
questions within each thematic area. Some of these sources contained
generic information cutting across the four thematic areas; including the
National Family Health Survey (NFHS) website for the three NFHS reports (www.nfhsindia.org),
Sample Registration System (censusindia.gov.in/Vital.../SRS/Sample_
Registration_System.aspx) for data on (cause-specific) mortality, etc.
Some of the sources/databases were specific for a thematic area, for
example the National Nutrition Monitoring Bureau (www.nnmbindia.org)
for childhood anemia. The Authors were free to search any other additional
database with a specific justification for the same.
For the operational issues, it was recognized from a
pilot test-search that the literature sources listed above would be
inadequate; hence, the following additional sources were accessed:
Popline.org (www.popline.org), documents published by the World
Health Organization and available online (www.who.int), documents
of the UNICEF available online (www.unicef.org/india/), documents
of the Government of India available online, specific publications of the
Ministry of Health and Family Welfare (www.mohfw.nic.in),
publications of the Indian Council of Medical Research (www.icmr.nic.
in), and Central Bureau of Health Intelligence (www.cbhidghs.nic.in/)
reports. In addition, the Authors were encouraged to search other
literature sources including textbooks, abstract books, conference
proceedings, etc with justification for the same. Authors were free to
contact related subject experts. At the outset, it was realized that the
broad range of issues to be addressed and the type of questions raised
would necessitate a systematic review process, far beyond the usual
reviews undertaken to evaluate efficacy of interventions [7]. The usual
PICO (population, intervention, com-parator, outcome) framework [8,9]
would not be the best solution for many of the questions, especially the
Operational issues. In order to capture the vast amount of data likely to
be available, and at the same time retain (methodological) quality, it was
decided a priori to use the process described below.
Inclusion and Exclusion criteria
For the technical issues, the emphasis of literature
search was to collect and collate current, best evidence on the specific
question under review. The Technical questions themselves were broadly
categorized as Epidemiology questions (magnitude, time-trend, geographical
trends, social/cultural trend) and Intervention questions (prevention or
treatment). For the former, the following hierarchy was used: I:
Systematic review of interventional trials published within the last 10
years (in the Cochrane reviews, the date until which literature had been
searched is also available); II: Randomized controlled trial(s) (if
systematic review(s) within the past ten years, was/were unavailable);
III: Quasi-randomized trial (if no randomized trial(s) was/were
available); IV: Non-randomized trial (if neither a systematic
review nor randomized trial nor quasi-randomized trial were available);
V: Modelling/ secondary calculations based on studies within the past
ten years (if the preceding study types were not available); VI:
Modelling/ secondary calculations based on studies older than ten years
(if none of the above was available).
For the Intervention questions, the following hierarchy
was used: I: Systematic review of interventional trials within the
last 10 years; II: Randomized controlled trial(s) (if no SR
available within the past ten years); III: Quasi-randomized trial
(if no randomized trial was available); IV: Non-randomized trial
(if no SR or RCT was available); V: Modelling/ secondary
calculations based on studies within the past ten years (if the above
study types are not available); VI: Modelling/ secondary
calculations based on studies older than ten years (if none of the above
was available). Literature search using this hierarchy ensured that high
quality evidence was preferentially sought and included; at the same time,
it allowed different study designs to be included; without resorting to
data likely to be biased [5]. Table V summarizes the Methodology
process.
A similar hierarchy could not be utilized for
literature search on operational issues. Hence it was decided a priori
that no specific order of preference would be applied. Nevertheless, the
following were preferentially sought and included: (i) documents
originating from/ pertaining to India, (ii) documents originating
from developing countries, (iii) documents pertaining to developing
countries in general, (iv) documents originating from/pertaining to
developed countries or unspecified countries, that could have a bearing on
India and/or developing countries, based on the content presented. It was
also decided a priori that among the documents retrieved, those
which presented a Methodology section would be given higher preference, as
the risk of bias is lower; and among those where Methodology was
presented, those documents with robust Methodology were accorded greatest
preference. For both sets of questions, only English language publications
were sought and included.
Search Strategy and Selection of Publications
For the technical issues, each question necessitated a
separate search string. The question was converted to the PICO format and
key words selected for searching. The specific search terms used are
presented in the four individual reviews. Each search was designed as a
"broad sensitive" search, and if required tapered to a "narrow specific"
search. For the Operational issues, various search terms were tried singly
and in combination, to identify the string with the maximum output. After
several trials, the following string was selected and set for all the 4
reviews: * AND india AND (health policy OR health planning OR health
programs OR health services OR program evaluation OR operations research)
where the asterisk represents the topic under review (ARI, diarrhea,
anemia, or community-based newborn care).
The search date, search terms, search string and search
output were recorded and saved. The following format was used by each
Author to consider publications for inclusion/exclusion in the review: (i)
Examination of title; those titles obviously not relevant were excluded
and the rest processed further; (ii) Examination of Abstract or
Introduction (where abstract was not published) of the short-listed
titles; those which were not relevant were excluded; (iii)
Examination of full-text; those publications which did not match the
inclusion criteria were excluded; and the remainder processed to the next
filtration step; (iv) Amongst the selected publications, the
following were included: those pertaining to the pediatric age group (for
questions relating to epidemiology, interventions and programmes), for
interventions, standard dose/route/administration modalities were
preferentially included over experimental and/or special models in
research settings; and those with primary/hard outcomes were
preferentially included over secondary/surrogate/soft outcomes for
interventions (especially treatment and/or prevention modalities).
Data Collection and Analysis
The principal author for each topic studied each
included publication in detail and extracted data relevant to the review.
In case of anything lacking clarity, the Reviewer for the topic examined
the publication. If further clarification was required, the team of four
principal investigators deliberated on the issue and arrived at a
consensus.
Extracted data were synthesized in a descriptive
manner. It was decided a priori that no secondary data analysis
(meta-analysis or other statistical tests) would be performed, since the
objective of the systematic review was to identify issues for
advocacy/action; and the types of publications retrieved in each thematic
area were likely to preclude pooled analysis of aggregate data through
meta-analysis [10]. However, for the review on community-based newborn
care, a sub-group analysis was done.
There was considerable discussion on whether or not to
critically appraise each included publication for methodological quality.
Based on the facts that multiple study types would be eligible for
inclusion; high degree of variability in components, construction and
properties of critical appraisal tools for research reports and absence of
a "gold standard’ critical appraisal tool for any study design [11], it
was decided that no additional critical appraisal (except for Methodology
i.e validity) would be undertaken. Where possible, numerical data were
extracted and presented in a tabular format. Where essential, a footnote
for clarification was included.
Although there is a risk of publication bias when only
published literature is included; many systematic reviews themselves
largely do not recognize the impact of the problem [12]. Therefore, it was
decided that no statistical tests for publication bias would be
undertaken; and no additional searches for unpublished literature would be
undertaken to counter the potential of the same.
Quality Assurance
On completion of the four systematic reviews, each
underwent a first-level review by Reviewer with experience and expertise
in the topic under review as well as expertise in research methodology.
Based on feedback received, the reviews were modified, revised and
finalized. The detailed findings were presented at a third Consultative
Meeting held on 4-5 May 2010 at Manesar, Haryana. This was attended by the
Research Team, Senior Reviewers and partners from UNICEF and PHFI. Based
on feedback from all participants, the four reviews underwent further
revision and modification. A fourth Consultative Meeting was held on 21
July 2010 at New Delhi, which was attended by (amongst others) Dr H
Hombergh, Chief, Health, UNICEF India Country Office and Prof K Srinath
Reddy, President PHFI. The key findings of the four systematic reviews
were presented, additional feedback received and the reviews were
finalized.
The finalized systematic reviews were prepared for
publication in accordance with the standard guidelines for publication of
research papers. The first draft was prepared with a deadline of 15 August
2010, revisions were made and finalized by 15 October 2010.
Contributors: JLM devised the methodology, drafted
and finalized the manuscript. DS and PG provided important additional
inputs to Methodology and revised the manuscript. All authors contributed
to the process of finalizing the Methodology and manuscript.
Funding: UNICEF.
Competing interests: None stated.
Disclaimer: Opinions expressed by PM, RP and SM are
their individual views and do not in any way reflect upon the
institutions, they serve.
References
1. Ministry of Health and Family Welfare (Government of
India). Child Health Programme in India. Available from: http://www.mohfw.nic.in/dofw%20website/child%20healthrti.pdf.
Accessed November 11, 2010.
2. United Nations Millenium Development Goals.
Available from: http://www.un.org/millenniumgoals. Accessed November 11,
2010.
3. Byrne S, Wake M, Blumberg D, Dibley M. Identifying
priority areas for longitudinal research in childhood obesity: Delphi
technique survey. Int J Pediatr Obes. 2008;3:120-212.
4. Adler M, Ziglio E (eds). Gazing into the Oracle: The
Delphi Method and its Application to Social Policy and Public Health.
London: Jessica Kingsley Publishers; 1996.
5. Smith LA, Moore RA, McQuay HJ, Gavaghan D. Using
evidence from different sources: an example using paracetamol 1000 mg plus
codeine 60 mg. Available from: http://www.biomedcentral.com/1471-2288/1/1.
Accessed November 10, 2010.
6. Betrán AP, Say L, Gülmezoglu A, Allen T, Hampson L.
Effectiveness of different databases in identifying studies for systematic
reviews: experience from the WHO systematic review of maternal morbidity
and mortality. Available from: http://www.biomedcentral.com/1471-2288/5/6.
Accessed November 10, 2010.
7. Straus S, Haynes B, Glasziou P, Dickersin K, Guyatt
G. Misunderstandings, misperceptions, and mistakes. Evid Based Med.
2007;12:2-3.
8. Akobeng AK. Principles of evidence based medicine.
Arch Dis Child. 2005;90:837-40.
9. Mathew JL, Singh M. Evidence based child health: fly
but with feet on the ground! Indian Pediatr. 2008;45:95-8.
10. Lyman GH, Kuderer NM. The strengths and limitations
of meta-analyses based on aggregate data. Available from: http://www.biomedcentral.com/1471-2288/5/14.
Accessed November 10, 2010.
11. Katrak P, Bialocerkowski AE, Massy-Westropp N,
SaravanaKumar VS, Grimmer KA. A systematic review of the content of
critical appraisal tools. Available from: http://www.biomedcentral.com/1471-2288/4/22.
Accessed November 8, 2010.
12. Phillips CV. Publication bias in situ. Available
from: http://www.biomedcentral.com/1471-2288/4/20. Accessed November
13, 2010.
|
|
|
|