Diet quality and health in children and adolescents

Systematic review of diet quality indices and their associations with health-related outcomes in children and adolescents

Marshall et al. JHND Early View.


Diet quality indices add an important dimension to dietary assessment. The aim of this systematic review was to: (i) identify and describe the attributes and applications of diet quality indices developed for use or used in paediatric populations; (ii) describe associations between these diet quality indices and health-related variables in paediatric populations; and (iii) identify factors that are associated with diet quality in paediatric populations worldwide.


Studies were identified by searching electronic databases for relevant papers from 1980 to October 2013 using keywords. Inclusion criteria were original studies that utilised a quantitative measure of diet quality in children and adolescents aged 0–18 years.


One hundred and nineteen studies met the inclusion criteria, from which 80 different diet quality indices were identified. The majority of studies had >1000 participants and were of acceptable quality. Of the 56 studies that investigated health-related outcomes, weight status was the most researched. Europe produced the most number of diet quality indices (n = 27 indices). Of the 119 studies, seven intervention studies were identified. Paediatric diet quality indices were found to be associated with environmental, behavioural and maternal factors.


The use of diet quality indices in paediatric populations is a rapidly expanding area of research in diverse populations internationally. In economically disadvantaged countries, diet quality indices may be predictive of child growth. However, prospective cohort, intervention and validation studies are required to draw stronger conclusions concerning risk of future disease in paediatric populations in general.

How to write #3- The introduction

The nature of the problem

Almost everyone who works in science struggles with writing. The natural skills that earn us careers in science, such as observation or technical abilities tend to leave us disadvantaged relative to those with a flair for the arts or languages. There are also many aspects in the way we are trained that don’t equip us for something that (at least in academia) we end up doing on a daily basis. Most young scientists don’t perfect the art of writing until they have written a PhD thesis, but that particular writing exercise leaves them ill-equipped for writing papers. The thesis rewards excessive length and detail, whilst the scientific paper requires brevity. At the early stages of our writing careers we are not trained to deal with this format. Thesis writing, or producing undergraduate essays instills the instinct is to write a full review of the literature when we tackle the introduction to a paper. We forget that this element of the research article is not about showing how clever and knowledgeable we are. Before you start to write you need to accept that editors and reviewers don’t like long introductions and producing one could be a last straw prompting rejection, or at the very least will result in some annoying revisions.

What the introduction is for

Hopefully the paragraph above spells out what the introduction isn’t for. It is not a literature review. The introduction to a paper has one simple function, which is to presents the rationale for the work described in the paper. It should be short and provide a brief account of the landmark studies that got the field to where it is, what is currently known and what are the unanswered questions. The introduction gives an indication of what are you going to be presenting in the main body of the text. Above all this section of the paper should spell out why this work was important, why it is timely and why your readers should be interested in reading any further.


Where should we start?

It took me a while to work this out for myself, as it is counter-intuitive. Just as we don’t actually start writing the paper by doing the abstract, then the introduction, then the methods, etc (I actually suggest you start with the results section), we don’t start writing the introduction with the beginning. The best way of tackling this section of the paper is to start at the end, in fact with the very last lines. This will generally be the statement or statements of intent for the paper. You set out clearly what you were trying to achieve when you did your research. Not all research is hypothesis driven, but if yours is, then your hypothesis should be stated at the end of your introduction. Where there is no hypothesis you will still have had aims or objectives when you began your project, so end the introduction with those. At the end of your introduction the reader should be in no doubt about what you were trying to achieve. In some disciplines common to say how you went about it and so you might add a sentence or phrase to indicate that and link into the methods section that follows.


The hypothesis investigated in this study was that supplementation of the diet with iron would result in improvements in cognitive function in pre-school children This was addressed through a randomised controlled trial.


And now work we need to work backwards from that statement. Where did the hypothesis or aims come from? Is there some doubt or controversy that the study aimed to clarify? Is this work just the next logical step in a chain of research? Where did that controversy or chain of research come from? These are all questions that you should be able to answer as you prepare your introduction. You won’t have done your research on a whim, or conjured a hypothesis from nowhere. None of us work in a vacuum and so your introduction has to succinctly pull together the strands of other research that led to your particular piece of work.


Moving still further back from your final statements, you will need to open the introduction with something fairly general, but topical and up-to-date about your subject area. This will be the hook on which you hang the whole paper. Think in terms of an opening that says X is an emerging problem; X has been of concern for some time and is increasing; X may represent a novel strategy for dealing with Y. All of these statements are just ways of saying that the general field in which you are adding just one small piece of new knowledge is really, really interesting or topical. Hopefully you believe that is the case, since you have invested your precious time in researching it.


Length and structure

I would suggest that you aim for no more than 500 words and you could probably write your introduction around 3 themes. As I generally work backwards these would be:


  • What were we trying to do (hypothesis and aims)?
  • What is the current dogma/cutting edge/ controversy that needs to be challenged or resolved?
  • Define the general area of research

Flip them around and you have your discussion- these could even be 3 paragraphs of 150 or so words each.

  • Define the general area of research
  • What is the current dogma/cutting edge/ controversy that needs to be challenged or resolved?
  • What were we trying to do (hypothesis and aims)?


An example

Let’s just look at this in a different way by analysing one of my recent introductory sections to a manuscript. This is the complete introduction and the first thing to note is that it is just 333 words long, so barely longer than the abstract. This reinforces the fact that the introduction is there just to set up the reason for the study, to justify why this paper is important. It wasn’t written to educate the reader, in fact quite the opposite. I generally assume that 90% of the people who read my papers have some expertise in the subject and that in fact they may only scan the introduction to see if there are any new references that they’ve not seen before and to find out what the aim of the study was.


The worldwide increase in the prevalence of overweight and obesity is increasingly impacting across all age-groups in the population (Ogden et al., 2013; WHO 2013). As a result all developed countries are reporting high levels of obesity among women of childbearing age and this has important consequences for maternal and fetal health during pregnancy, and potentially for the longer-term health of the children of obese women (Normia et al., 2013; Langley-Evans 2014; Taylor et al., 2014). In the UK 13 % of 21- to 30-year-old and 22 % of 31- to 40-year-old women were estimated to be obese in 2007, and this was expected to rise to 30 and 47 % respectively by 2050 (Foresight, 2007). 20% of UK women aged 16-44 were obese in 2010 (National Obesity Observatory, 2014) and in the USA (Ogden et al., 2013) this figure was approximately 32% in the 20-39 year old population. A dramatic increase in the prevalence of severe or morbid obesity has occurred alongside the increasing prevalence of obesity in young women and in 2009 approximately 5% of all pregnancies in England were associated with maternal BMI of over 35 kg/m2, with approximately 2% of pregnant women having BMI in excess of 40 kg/m2 (National Obesity Observatory 2014). Pregnancy is recognised as a period during which women are vulnerable to excessive weight gain that they may find difficult to reverse, thereby increasing risk for subsequent pregnancies and their longer-term health (Groth et al., 2013; Von Rueslen et al., 2014).


So, here we have the opening paragraph. This is defining the area of study. It is simply stating that obesity is increasing, is common in women of child-bearing age and that being pregnant is in itself a common risk factor for women to become obese. Notice that there are references in the section, but that they are all to recent papers. The introduction is not the place for an in-depth review or a history lesson in your field of interest. Short, sharp, focused and up-to-date should be your watchwords.


Maternal obesity during pregnancy increases the risk of adverse pregnancy outcomes, including miscarriage, gestational diabetes and hypertensive disorders (Sebire et al., 2001; Wang et al., 2002; Jensen et al., 2003; Maconochie et al., 2007; Centre for Maternal and Child Enquiries 2010; Li et al., 2013; Sommer et al., 2014). Obesity is recognised as a significant risk factor for maternal and fetal death (Centre for Maternal and Child Enquiries 2010). The risks associated with maternal overweight and excessive weight gain are recognised by the US Institute of Medicine (2009), which has published guidance on optimal ranges of weight gain during pregnancy. These are based upon maternal weight prior to pregnancy, with obese mothers advised to gain 5-9 kg across pregnancy, compared to the 12.5-18 kg recommendation for women of normal weight. The UK does not have any formal, evidence-based recommendations for healthy weight gain in pregnancy, although a guidance range of 10-12.5 kg is included within Department of Health literature. However, National Institute of Health and Clinical Excellence (NICE) guideline of 2010 recommends that health professionals carefully manage maternal weight. The emphasis of these guidelines is on weight loss prior to, or after pregnancy (NICE 2010). Weight loss is not advised during pregnancy as it may pose a risk to fetal nutrition and development.


So here I spell out one of the important challenges that exist in this field. We know that obesity is linked to poor pregnancy outcomes but at the moment we have no clearly defined strategy for managing the problem in obese mothers-to-be. Again, the review element here is light touch and references kept current. This is the section of the introduction where we should be presenting the current state-of-the-art, current balance of opinion (or lack thereof).


The antenatal period puts women into greater contact with health professionals and is therefore an ideal time for health education. Mothers are generally open and more readily motivated to make lifestyle changes that could benefit the health of themselves and their baby (Ritchie et al., 2010; Wilkinson & McIntyre 2012; Wilkinson et al. 2014; May et al., 2014). A number of studies have evaluated the impact of antenatal diet, exercise or weight management programmes upon pregnancy outcomes. Thornton and colleagues (2009) found that monitoring the food intake of obese women was associated with lower gestational weight gain and lower prevalence of gestational hypertension. Shirazian et al., (2010) reported that a lifestyle modification in obese pregnant women reduced weight gain, but had no effect on adverse pregnancy outcomes such as pre-eclampsia. The meta-analysis of Thangaratinam et al., (2012) found that weight management interventions in pregnancy reduced the risk of pre-eclampsia, but had no impact upon other obstetric outcomes. There are also a number of ongoing studies evaluating intervention strategies, such as the LIMIT trial in Australia (Dodd et al., 2011) and the UK UPBEAT study (Poston et al., 2013). Alongside randomised controlled trials of interventions, there are many clinical interventions mounted on a local level that aim to reduce the impact of maternal obesity upon health in the community. In this paper we report the findings of a service evaluation of one such programme. One-to-one antenatal guidance from midwives and healthy lifestyle advisors resulted in lower gestational weight gain and a reduced prevalence of gestational hypertension.


The last section of this introduction is perhaps not my finest hour now that I read it again, as the final sentences should be a snappier statement of aims or hypothesis. However, you can see it there in the highlighted section and it is followed with a sentence that is a taster of the paper outcomes (some journal editors really like this) which gives an indication of how the study was performed. The start of the paragraph flows from the second section. Having set up the challenge (the fact that there is no defined strategy) there is an offering of examples that show the findings of the literature are inconclusive in evaluating what happens when trying to meet that challenge.


A potted introduction, or introductory tool kit

So the introduction boils down to this fairly generic formula:


My area of research is of importance/interest because of this ______________________________. This is a new/old/well-established/emerging area.

The current state of knowledge in this area is ______________________________. Researchers in this field disagree on several key areas/agree that it is now essential to explore. We know a certain amount, such as ______________________, but _____________________ and ___________________ are still unanswered questions/unknowns/areas of controversy. 

As a result of an inconsistent literature/ following on from previous work in this field, we set out to address the hypothesis that/aim of ______________________________________. We did this by conducting an experiment/systematic review/qualitative study/survey/randomised controlled trial/etc.

Try writing one for yourself using this kind of structure. Delete the alternatives that don’t apply to you and fill in the blanks. This will probably give you a core of 150 words without even having to think too hard and you can flesh out the rest around that skeleton.

Determinants of a regular intake of a nutritionally balanced school lunch

Determinants of a regular intake of a nutritionally balanced school lunch among 10–17-year-old schoolchildren with special reference to sense of coherence

Tilles-Tirkkonen et al JHND Early View


Free nutritionally balanced school lunches are offered to all schoolchildren in basic education in Finland in each school day. Having school lunch on a regular basis has been found to reflect overall eating patterns. However, skipping part of or even the entire lunch is common. The present study investigated the determinants of the regular consumption of a nutritionally balanced school lunch among schoolchildren, with special reference to the role of sense of coherence (SOC).


In total, 887 children (457 girls and 424 boys), aged 10–17 years from three municipalities in Eastern Finland, filled in a web-based questionnaire in class during a school day and reported eating patterns, body height and weight and perception of body image. SOC was measured by using the 13-item scale. The statistical analysis was carried out with logistic regression modelling and the chi-squared test.


In addition to female gender, frequent shared family meals, perception of body image as appropriate and younger age, SOC was a significant determinant of regularly eating a nutritionally balanced school lunch in the final multivariate modelling. Strong SOC was also associated with more regular meal frequency and health-promoting snack choices.


To promote healthy eating patterns among school-aged children, special attention should be paid to children with weak SOC because they may need specific support and encouragement. They might lack sufficient belief in their own capability and/or do not have adequate support from their family to influence their eating and other lifestyle patterns.

Liquid versus solid energy intake in relation to body composition among Australian children

Liquid versus solid energy intake in relation to body composition among Australian children

Zheng et al., JHND Early View



The debate about whether energy consumed in liquid form is more obesogenic than energy consumed in solid form remains equivocal. We aimed to evaluate the effects of liquid versus solid energy intake and different beverage types on changes in childhood adiposity.


Our analyses included 8-year-old Australian children (n = 158) participating in the Childhood Asthma Prevention Study. Dietary information was collected using three 24-h recalls at age 9 years. Multivariate linear regression was used to evaluate the effects of liquid versus solid energy intake and different beverage types on changes in body mass index (BMI) Z-score from ages 8 to 11.5 years (△BMIz8–11.5y) and percentage body fat (%BF) at age 11.5 years (%BF11.5y). Substitution models were used to evaluate the effects of substituting other beverage types for sugar-sweetened beverages (SSB).


Liquid energy intake (1 MJ day–1) was more closely associated with both △BMIz8–11.5y (β = 0.23, = 0.02) and %BF11.5y (β = 2.31%,= 0.01) than solid energy intake (△BMIz8–11.5y: β = 0.12, = 0.01 and %BF11.5y: β = 0.80%, = 0.07). SSB consumption (100 g day–1) was directly associated with △BMIz8–11.5y (β = 0.08, = 0.02) and %BF11.5y (β = 0.92%, = 0.004),whereas diet drinks (100 g day–1) were inversely associated with △BMIz8–11.5y (β = 0.18, = 0.02). Substitution of 100 g of SSB by 100 g of water or diet drink, but not other beverages, was inversely associated with both △BMIz8–11.5y and %BF11.5y (< 0.01).


Our findings indicate that liquid energy is more obesogenic than solid energy. In particular, SSB, but not other beverage types, are a significant predictor of childhood adiposity and replacing SSB with water can have long-term beneficial effects on childhood adiposity.

How to write #2: Planning your paper.

The rest of this series is going to focus on the mechanics of writing the various sections of a research paper, considering everything from abstract through to conclusions (though not necessarily in that order). I think, though, that before plunging into the maelstrom of producing actual text, it is important to think about what comes first. Planning and dealing with some of the perils and pitfalls of multiple authorship.

  1. Decide on what your story is going to be

Writing a research paper requires you to apply the simple rules of story writing that you learned in primary school. Your work needs a clear beginning, middle and end. You need to identify what all of those elements are going to be, before you start and then plan the flow of the story up front. I say ‘story’ because that is a great model for most scientific papers. We are essentially following the pattern of:

We are interested in this subject because previous work has shown it to be important. To advance the field we set out to assess ____________, which we did using these methods_____________________________. Our findings were ______________ and __________________ and _____________________. These were largely expected given previous work, but the novel finding from our research was _________________________. In conclusion, we have confirmed the importance of this area.

In developing the plan I always start by thinking about what the bit in the middle is. For all intents and purposes this is about looking at your data and deciding what is going to be in the results section. As I will describe in a separate post, there is a need to be reasonably selective. We don’t throw in every bit of material that was collected just because it is available. Your readers don’t want to wade through dozens of dense tables of numbers that don’t contribute to the ‘story’. The nature of the data that you do put in to the paper shapes the ending. Your discussion is going to reflect the outcomes of your work and how it fits with the literature. The beginning should be easy to frame. You are writing a paper because you have done some research. Your introductory text just has to spell out why you did it and why the research was necessary.

The best papers have the same flow as a work of fiction. Think about how you will link threads together, which order the data should be presented in and how best to portray something complex in a clear and concise manner. Look at some papers by other authors- we all have papers that we really like and value. How were they structured? Why do you think they were effective? What can you borrow for your own writing?

2. Decide on what format your report will be written for

I have to say that I am not really a fan of the short report and would much rather see papers written up in full so that the authors present a larger selection of data and properly discuss it. That should be your aim. There are situations however where the short report might be the way you want to go. Some experiments or studies are done on a shoestring budget and need to be published as a means of securing funding to do the work properly. Some undergraduate research may be noteworthy in that it has generated a small number of interesting results that would otherwise not get into the public domain.

So, my advice is, go for a full paper unless the study you have completed has yielded only enough data to generate a single graph or table. In those instances, short reports, letters to the editor and similar formats are the way to go

3. Consider your target journal

It helps a lot to decide on the target journal before you start to write. Each journal has a guide to formatting of papers, length of papers, referencing style, etc. Being aware of these from the start makes the writing process simpler and avoids last minute edits. There will be lots of factors that influence your choice of journal:

  • Impact factor
  • Speed of decision-making
  • Standing within your discipline
  • Time to actual publication
  • Availability of early online publication

These are all legitimate factors to shape your target, but it is also important to avoid wasting your time. Does the journal actually publish papers that are in your area? Do they have local rules about specific aspects of the work that you can’t meet (e.g. systematic reviews must meet PRISMA standard, weight studies need more than 12 month follow-up)? Does it meet the requirements of your university or funder for open access? Will it be the most suitable vehicle to showcase your work? Is there going to be a charge for publication? Check the guidance to authors very carefully before plunging in.

4. Shared authorship

Very few of us will work in isolation and with only occasional exceptions papers are published with multiple authors. If you are responsible for writing the paper you will need to agree first of all who is going to be an author and secondly what the order of authorship is going to be. The latter is subject to conventions depending on the discipline, but in my area the first author is going to be the person who contributed most to the collection of the data and the last author is the supervisor/principal investigator, who set up the study.  The people in between will generally be listed in order of contribution. Other disciplines vary and in some subjects the list is just alphabetical.

There are certain criteria to fulfil to qualify for authorship and the general expectation is that an author will have made a significant contribution to the design or execution of a research project, been involved in collection of or analysis of the data and to have played a role in the writing or the critique of the paper. People associated with the work who do not fulfil those criteria would be expected to be listed in the acknowledgements only.

An unfortunate outcome of collaboration and coauthorship is the direct correlation between the number of authors and the amount of hassle involved in producing a paper. Your coauthors may disagree on some of the fundamentals, so take care that when you decide what the story is going to be, that they all buy into it from the start, rather than arguing about it later. They will disagree on the format of the data, exact wording of points in the discussion and probably on the conclusion.  Again, try and resolve things at the early stage of planning, but do expect to spend some time sorting out these disagreements. Mostly (if you’ve chosen your team well) they will resolve amicably, but I have seen fur fly on occasions and the corresponding author needs to be confident about taking the lead and laying down the ground-rules. The most common problem is delays in the schedule. Some of your coauthors will respond to queries and comment on drafts in a timely manner. There will certainly be someone (as researchers tend to be academics) who has many other things to do and doesn’t get round to your paper as quickly as you would like. A week or two is never too bad, but delays of months are not good for harmonious writing.

My advice is to set your authors clear deadlines. ‘here is the first draft of the paper. If I don’t hear from you by the end of next week I will assume that you are happy with it and have no changes to make’ , is a device I tend to use and it usually keeps everyone on task and the writing process avoids undue delay.

5. Standing back and accepting peer review

I am an experienced writer and I can put together a manuscript pretty quickly. You may be less confident and might spend a few weeks putting your paper together and collating comments from coauthors. If you are the corresponding author who is coordinating the submission you will have a few challenges to get the final draft put together. Coauthors may not all agree, their comments may be contradictory, or with many cooks stirring the broth, your paper may not flow and speak with a consistent style. At some point you will need to take control and responsibility for a polished product.

My advice here is not to rush. Get the final draft together in your own time and then file it away somewhere for a week or so. Let it mature in the depths of your hard-drive and then bring it out again for a final check. You will almost certainly pick on something clunky or vague. Also don’t take sole responsibility for the paper- give it to someone else who is knowledgeable in your subject but isn’t a co-author. Their unbiased viewpoint and lack of familiarity with the work will pick up on obvious flaws, awkward expression and inconsistencies.

6. Submission

These days, journals run all of their manuscript submission and review services online and all of them will have their own specific rules and requirements for the submission. These will cover what sort of files can be uploaded, whether or not a cover letter is required, the length of the manuscript, format of figures and whether figures should be provided as separate files, or embedded in the main body of the paper. The frustration that online submission systems can cause you should never be under-estimated. You’ve got to the end of the writing process. You’re feeling excited and are desperate to push the submission button. But then… you get bogged down in it all and spend half a day changing the resolution settings on your graphs, producing a blinded copy without the authors names on, etc.

My simple advice here is do a dummy run as part of your preparatory work. It can save you a few points of blood pressure increase.