Evidence-Based Nutrition Practice

Applying the Concepts of EBP to Pediatric Nutrition Practice

(with thanks to Donna Johnson)

What evidence-based medicine is:

Evidence-based medicine is the conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients.

Sacket et al. BMJ 1996

What evidence-based medicine is:

The practice of EBM requires the integration of

individual clinical expertise

with the best available external clinical evidence

from systematic research.

Two Principles for Evidence Based Practice

Evidence alone is never enough to make a clinical decision (or a decision about guidelines or evidence)

Decisions & Recommendations should be guided by a hierarchy of strength of evidence

“If no randomized trials have been carried out for our patient’s predicament, we follow the trail to the next best external evidence and go from there.”

Sacket et al. BMJ 1996

Why Evidence-Base Practice?

Clinicians need information MDs said

» they need it twice a week,» they get it from our text books & journals.

RD’s said» They need it 5 times a week» They search the literature 5 times a month

Clinicians really need information!

If shadowed: they need it up to 60 times per week but only

actually find 30% of what they need and that comes from passers-by

» “my textbooks are out of date”» “my journals too disorganized”

Medical textbooks are out-of-date

Fail to recommend Rx up to ten years after it’s been shown to be efficacious.

Continue to recommend therapy up to ten years after it’s been shown to be useless.

Many Traditional Interventions are Out of Date: Ex. Infant GERD

Effective?

Placing in elevated prone position No

Placing in infant seat No

Thickening formula with infant cereal

No

Changing protein content of formula

No

Use of pacifier Unclear

5% vs. 10% dextrose water SlightlyCarroll et al. Arch Ped Adoles Med. 2002

Why don’t RDs seek out the best evidence?

Lack of time Lack of skills for critically appraising the

literature Discomfort with going against traditional

institutional practice

Thomas et al. J Hum Nutr Diet. 2003

Three solutions

1 Carefully research each clinical question that comes our way

2 Seek and apply evidence-based medical summaries generated by others

3 Accept evidence-based practice protocols developed by our colleagues.

Process of EB Practice

1. Define the question

2. Plan and carry out search of the literature

3. Critically appraise the literature

4. Apply the results to your practice

5. Evaluate your performance

Step 1: Define Question -- PICO

P - Patient and disease

I - Intervention

C - Comparative intervention (optional)

O - Outcome

Step 2: Search for Evidence

Translate PICO Question into a searchable question

Establish a search strategy» key concepts» boolean operators» synonyms» prioritize» limit

Step 2: Search for Evidence

Sources Tools

Articles & Reviews Medline

Meta-analysis Medline, Cochrane

Practice Guidelines Nat’l Clearinghouse

Sytematic Reviews Cochrane, EBM journals

Step 3: Critically Appraise

Basic Introduction and Tutorial at:

http://healthlinks.washington.edu/hsl/classes/evidence/

ADA Quality Indicators

Design appropriate to hypothesis Inclusion & exclusion criteria Sample size Key quality indicators for each study

design type (ex randomized trials or meta-analysis)

Step 4: Apply Results

Within context of individual patient preferences, values and rights (or population resources, values and culture)

Evidence, Values, and Resources

Values

Evidence Resources

The Strength of the Evidence Depends on the Rigor of the Studies

Randomized Controlled Clinical Trial Cohort Study Case-control Study Cross Sectional Study Meta-analysis

Randomized Controlled Clinical Trial

Involves one or more test treatments and a control treatment

Specified outcome measures for evaluating the intervention

Bias free method for assigning treatment

Randomized Controlled Clinical Trial

Advantages Disadvantages

Unbiaseddistribution ofconfounders

Expensive

Blinding more likely Volunteer bias

Randomizationfacilitates analysis

Ethically problematicat times

Confounding Variable

“An extrinsic factor that is associated with the predictor variable and a cause of the outcome variable.”

Hulley and Cummings, Designing Clinical Research

Cohort Study

Identification of two groups» one received exposure of interest» one did not receive exposure

Follow cohort through time to observe the outcome of interest

Cohort StudyAdvantages Disadvantages

Ethically safe Controls may be hard to ID

Subjects can be matched Exposure may be linked to aconfounder

Can established timing of events Blinding is difficult

Eligibility and outcomeassessment standardized

Ramdomization not present

Easier and cheaper than RCT Large sample or long FU may beneeded

Case-control Study

Identify patients who have the outcome of interest (cases)

Identify controls without the same outcome

Look back to see if they had the exposure of interest

Case-control Study

Advantages Disadvantages

Quick and cheap Reliance on recall to determineexposure

Good for raredisorders or longlag

Confounders

Selection of control groups isdifficult

Fewer subjects thanin cross-sectional

Potential selection bias

Cross Sectional Study

Observation of a defined population at a single point in time or time interval

Exposures and outcomes determined at same time

Cross Sectional Study

Advantages Disadvantages

Cheap and simple Can’t establish causality

Ethically safe Recall bias

Confounders may beunequally distributed

Group sizes may beunequal

Meta-analysis

Quantitative method of combining the results of independent studies

Larger sample size and stronger summaries and conclusions

The Five Strengths of Evidence1. Strong Evidence from at least one systematic review of

multiple well-designed RCT

2. Strong evidence of at least one well designed RCT of appropriate size

3. Evidence from well designed trials without randomization, single group pre-post, cohort, time series or matched case control

4. Evidence from well designed non-experimental studies from more than one research group

5. Opinions of respected authorities based on clinical evidence, descriptive studies or reports of expert committees

Systematic Reviews

Identify problem area or area of uncertainty Formulate question Search for evidence Select relevant evidence Abstract findings and evaluate reports Form recommendations Summarize the strength of the evidence

supporting the recommendation Disseminate the findings

EB Review: Example

Huang R-C, Forbes DA, Davies MW Feed thickener for newborn infants with gastro-oesophageal reflux (Cochrane Review). In: The Cochrane Library, Issue 2, 2004. Chichester, UK: John Wiley & Sons, Ltd.

Background

Gastro-oesophageal reflux (GOR) is common in newborn infants. A common first line management is the use of feed thickeners.

The prevalence of excessive GOR in children is approximately 8%, as diagnosed on 24 hour ambulatory pH manometry studies in an unselected healthy infant population. Symptomatic regurgitation alone is more common and has been found to occur in 18% of the general infant population

Step 1: Define Question

P - Patient and disease

I - Intervention

C - Comparative intervention (optional)

O - Outcome

P = Newborn infants with GOR & preterm infants up to 44 weeks

I = Thickeners of all types including rice, gum, or flour based, added to all types of milk including formula and human milk

C = Non-thickened feeds

O = signs and symptoms of reflux, 24 hour ambulatory pH monitoring and/or oesophagitis on biopsy

Step 2. Search for EvidenceWe searched MEDLINE from 1966 to December

2001, the Cochrane Controlled Trials Register, The Cochrane Library, Issue 1, 2002. CINAHL from 1982 to December 2001, and conference and symposia proceedings published in Pediatric Research 1990 to

1994. We also searched conference proceedings for the European Society for Paediatric Gastroenterology and Nutrition (ESPGAN) and the North American Society for Pediatric Gastroenterology and Nutrition (NASPGAN) from 1994 to December 2001. We did not restrict the searches to the English language.

Selection Criteria

All randomised controlled trials that examine the effects of thickening formulas on treating gastro-oesophageal reflux in neonates. The eligible studies were to compare thickened feeds to no intervention (unthickened feeds).

Step 3. Critically appraise the literatureKey Criteria: blindness of randomisation blindness of intervention completeness of follow up blinding of outcome measurement For individual trials, mean differences (and 95%

confidence intervals) were reported for continuous variables

For categorical outcomes the relative risk and risk difference (and 95% confidence intervals) were reported

Data collection and analysis

Two independent reviewers identified potential studies from the literature search. Quality was independently assessed by two independent reviewers.

Study Reason for exclusion

Bailey 1987  Patients' ages ranged from 4 days to 14 months. This was a cross over study in which each patient received both thickened and unthickened feeds, but it does not appear to be a randomised cross over trial. 

Carcassonne 1975 

No control group was used. The age group ranged from 40 days to five years. The patients used had anatomical abnormalites, severe burns or brain tumours. 

Gouyon 1989  This study evaluated smectite in newborn infants with gastroesophageal reflux. It was rejected because of the lack of randomisation and use of further "thickeners" in some patients in both intervention and placebo group on the basis of undefined symptoms. Gouyon 1988 is an abbreviated report of the same study. 

Greally 1992  The study population was aged between 2 and 18 months. Infants were randomised to receive either cisapride or gaviscon/carobel. There was no placebo group. 

Khoshoo 2000 

Age range was from 4 to 10 months. Not a RCT. 

Le Luyer 1992 

The study group ranged in age from 2 weeks to 57 months. Not a RCT (no control group). Patients were divided into two groups using different doses of a thickener (sodium alginate) without randomisation. 

Miller 1999  This study was rejected on the basis of the age group encompassing 0 to 12 months. The outcome data for newborn infants could not be separated from the data of older infants. 

Orenstein 1987 

Ages were between 4 to 34 weeks of age. This was a cross over study, not a RCT. Each patient received both thickened and unthickened feeds. 

Vandenplas 1994 

Ages ranged from 1 week to 4 months old. The outcome data for newborn infants could not be separated from the data of older infants. 

Weldon 1972  Not a RCT. No control group. 

Main Results

No studies fulfilled the requirements for inclusion in the systematic review.

Reviewers' conclusions

There is no evidence from randomised controlled trials to support or refute the efficacy of feed thickeners in newborn infants with GOR. Given the absence of evidence, we cannot recommend using thickening agents for management of GOR in newborn infants

Step 4. Apply resultsImplications for practice

At present, there is no evidence from randomised

controlled trials to support or refute the efficacy of feed thickeners in newborn infants with GOR. Although thickening feeds is a simple and cheap manoeuvre, there

are some theoretical side effects of this treatment such as delayed gastric emptying with increased caloric density of feed. Therefore, given the absence of evidence, we do not recommend using thickening agents for management of GOR in the neonatal population.

Medline

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed

Under “Limits” can select:» Review» Meta-analysis» Practice Guidelines» Randomized Controlled trial

Cochran Database of Systematic Reviews

http://www.update-software.com/clibng/cliblogon.htm

Can search and review abstracts for free Full text requires subscription

Cochran Database of Systematic Reviews (examples – 51 nutrition)

Dietary interventions for PKU Carnitine supplementation of parenterally fed

neonates Feed thickener for newborn infants with

GER Vitamin A supplementation for preventing

morbidity and mortality in very low birthweight infants

Formula milk vs. preterm human milk for feeding preterm or LBW infants

Cochran Database of Systematic Reviews (examples – 51 nutrition)

Growth monitoring in children Fat supplementation of human milk for

promoting growth in preterm infants Gastrostomy feeding versus oral feeding alone

for children with cerebral palsy Multicomponent fortified human milk for

promoting growth in preterm infants Enteral nutritional therapy for induction of

remission in Crohn's disease

National Guideline Clearinghousehttp://www.guideline.gov/

examples:

Early discharge of the term newborn (bottle and breast-feeding; National Assoc Neonatal Nurses) Nutrition practice guidelines for type 1 and type 2

diabetes mellitus (ADA) Guidelines for the evaluation of food allergies

(American Gastroenterological Association)

Bariatric surgery for severely overweight adolescents: concerns and recommendations. 

340 with “nutrition”; 60 with peds nutrition

National Guideline Clearinghousehttp://www.guideline.gov/

ADA currently has 12 guidelines here (ex.): chronic kidney disease gestational diabetes hyperlipidemia MNT type 1 and type 2 diabetes) hypertension – older adults

Most expected to be added in the future

Haynes & Haines, BMJ 1998