Reference

 

 

 

Compliant

 

 

 

 

 

Section Number

Section Topic

Sub-Section Number

Checklist Item

Yes

No

Page Number

Comments

1

Title

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2

Abstract

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3

Introduction

 

 

 

 

 

 

 

 

 

 

 

3.1

 Context

 

 

 

 

 

 

 

 

 

3.2

 Previous studies

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4

Study Design

 

 

 

 

 

 

 

 

 

 

 

4.1

Analyte

 

 

 

 

 

 

 

 

 

4.2

Subjects

 

 

 

 

 

 

 

 

 

4.3

Analytical Method

 

 

 

 

 

 

 

 

 

4.4

Length of Study

 

 

 

 

 

 

 

 

 

4.5

Samples

 

 

 

 

 

 

 

 

 

4.6

Analytical conditions

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5

Data Analysis

 

 

 

 

 

 

 

 

 

 

 

5.1

Outlier analysis undertaken

 

 

 

 

 

 

 

 

 

5.2

Heterogeneity of variance assessed

 

 

 

 

 

 

 

 

 

5.3

Statistical methods described

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6

Results

6.1

Terminology

 

 

 

 

 

 

 

 

 

6.2

Results clearly presented and managed

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7

Discussion

 

 

 

 

 

 

 

 

 

 

 

7.1

Limitations of the Study

 

 

 

 

 

 

 

 

 

7.2

Generalisability

 

 

 

 

 

 

 

 

 

7.3

Interpretation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


 

1.            Title

The title should indicate that the content relates to a study of biological variation, the   subject of the study, and the population studied.

Back to top page

2.            Abstract

Content will be limited by word count allowed by the journal in which the article is published. As a minimum it should indicate the headline biological variation data, the major characteristics of the population studied and clearly identify, the analyte and quantities studied, the duration of the study. and the sample matrix.

Back to top page

3.        Introduction

3.1 Context

 

3.2 Previous studies

 

Back to top page


 

4.            Study Design

4.1 Analyte

4.1.1        The study  should clearly identify the target analyte or measurand.  This is required to support the application of the derived data in the correct setting.  Specificity of analytical methods will impact on the utility of the data. Generational changes in methodologies that deliver  improved specificity of methods may exhibit  different between and within subject  variation. Ideally metrological  traceability will exist through to  a primary international standard. It is important that where there is complexity in terminology and loose  common usage of  names of analytes  that the  biological variation data are published  are identified in a way that enables future users to assess the applicability of the data set to their context.  A good example of where this might have been of importance in the past is in the context of  haemoglobin A1c measurement. Recent introduction of standards and reference methodologies enables a very clear definition of what is meant by the term HbA1c . Historicaly published biological variation data have been derived from studies using a variety of methodologies with varying degrees of specificity for HbA1c in many clinical settings.  It is possible to state that in such situations the data can only be validly when a method of the same specificity is applied to a similar population. In consequence biological variation data must carry sufficient metadata to enable their transportability and application.  The arguments are similar to those  used  to ensure transportability of reference data  

4.2 Subjects

The description of the subjects of any biological variation study is essential to enable  data to be critically assessed and applied. Reference data clearly require the population from which they were derived to be fully characterised. Biological variation data are being derived from a reference population and should therefore be described accordingly to enable transportability of the data set. Subjects should be objectively assessed against clear inclusion and exclusion criteria.  It is important that those wishing to apply published data to a target population are able to compare their characteristics of the population from which the data were derived. 

4.2.1        Number of subjects

4.2.2        Standard Demographics

4.2.2.1  Age

4.2.2.2  Sex

4.2.2.3  Ethnicity

4.2.3        State of well being. Define

4.2.3.1  Non-diseased

4.2.3.1.1      Evidence in the form of clinical and diagnostic assessment.

4.2.3.2  Diseased

4.2.3.2.1     Diagnostic criteria used

4.2.3.2.2     Staging of disease  

4.2.4        Physical characteristics

4.2.4.1  Height, weight, body mass index

4.2.4.2  Blood pressure, renal function etc

4.2.5        Nutritional status, unusual diets

4.2.5.1  Dietary assessment where relevant

4.2.5.2  Vegan, high protein, high fat

4.2.5.3  Food supplements

4.2.6        Drugs

4.2.6.1  Prescription and non-prescription

4.2.6.2  Contraceptive agents

4.2.6.3  Recreational drugs

4.2.7        Alcohol intake

4.2.8        Smoking status

4.2.9        Specific inclusion and exclusion criteria

4.3 Analytical Method

A clear description of the methodology used should be from part of the metadata.  This may be available via an appropriate reference or may need to be presented within the publication of the biological variation data. Known issues of specificity should be addressed where known of. Any  deviation  from standard operating procedures  indicated in the published method or from the manufacturers data sheets in the case of commercially available assays, should be clearly indicated.  These issues are again important in the context of transportability and generalisability of the  biological variation data. 

4.3.1        Clear description adequately referenced
End users of published biological variation  data must be able to access the fine detail of the analytical method used to assay the analyte of in question.    This will allow a local objective  assessment of comparability of methods in use with that used to generate the original data and thus on the transportability of the data. This will prove to be particularly important where complex molecules are the subject of measurement , where there is a lack of traceability to reference methodologies and international standards, where the methodologies are complex with many variables affecting the value obtained for an assay (e.g. temperature, pH, substrate concentrations, polyclonal versus monoclonal antibodies), when comparing data derived from new methods with those  older generation assay systems

4.3.2        Standardisation and traceability of methods should form part of the data set accompanying the biological variation data

4.3.3        Analytical conditions

4.3.3.1  Deviation from standard operating procedures, use of adaptations of published methods, and deviation from manufacturers recommended methods in the case of commercially available systems should be documented.

4.3.3.2  Optional assay conditions adopted should be clearly identified

4.3.4        Assay Specificity

4.3.4.1  The fact that specificity issues have been considered should be identified. Any known specificity issues should be identified or references given to them. These may have an impact on the applicability of the data as more specific methods become available.

4.3.4.2  Where physiological or pathological conditions (e.g. chronic kidney disease impacting on clearance of cross reactants (e.g. aldosterone) or uraemia leading to production of interfering moieties   (e.g. carbamylated haemoglobins)) are likely to be common and influence variability these should be  described or referenced.

 

4.4 Length of Study

Length of study may be of importance for many targets measurements. This should be clearly recorded and any heterogeneity in the sampling schedules that may confound the interpretation or application of the data by end users should be clearly recorded. There may be several factors that may need to be considered and accounted for ranging from the in vivo half life of the substance to be measured, to biological rhythms in measured indices ranging from hours, to days, to months, to seasonal. 

4.4.1        Biological rhythms operate across time periods ranging from minutes to months. The  level of variability will consequently  be a function of the sample timing and frequency relative to the periodicity of the target analyte  in vivo. This will impact on the utility of the data  for setting quality standards for assays used clinically across differing time periods in different contexts.

4.4.2        Turnover and in vivo half life of target analytes will be of relevance to the design of the study protocol.   

 

 

4.5 Samples

The number of samples required per individual will depend upon the periodicity of the target analyte, the proposed application of the assay, and practicability of the study. If for example  you are monitoring a disease state by  sampling once every two months, then you need to understand the magnitude of the variability over that period and whether it constant in those cases that do not progress or improve.  Smaller degrees of variability of the same analyte within a day may be measurable but less relevant in the context of the clinical application of the measurement.

4.5.1        Numbers of samples should be appropriate and based on a rationale that is explained

4.5.2        Timing of samples should be clearly described and form part of the metadata that accompanies the  biological variation data set

4.5.2.1  Time of day

4.5.2.2  Dates and location (northern/southern hemispheres have disparate seasons, latitude will effect hours of daylight at different times of year)

4.5.2.3  Duration of the study should be considered and discussed in the context of known biological rhythms and relevance in the context of the clinical applications of the target measurement.

4.5.3        Collection conditions should be adequately described

4.5.3.1  The study design should enable samples to be taken under conditions that are well characterised and  these should be described in detail in the publication of the data.  The following should be taken as a non exhaustive list of factors to be controlled:

4.5.3.1.1           Stress (e.g. venepunture with a needle versus in dwelling cannula  in situ).

4.5.3.1.2           Dietary restrictions (specific foodstuffs, calories, special diets versus fasting)

4.5.3.1.3           Posture: supine, erect, sitting (report should include time that the subject chosen posture prior to sampling. It  should be remembered that it takes a time equivalent  7 half lives  for a peak concentration to fall to <1.0% f there is no additional input into a system).

4.5.3.1.4           Duration of  an collection (e.g. urine (timed, random, mid flow, early morning), other fluid  or faeces)

4.5.3.1.5           Blood sampling should be characterised in terms of: -

·   Method of collection (e.g. syringe and needle, evacuated tubes, indwelling cannula/catheter, lancet(finger prick)

·   Volume of sample taken

·   Clear description of sites from which samples are taken (venous versus arterial, capillary) and anatomical site. The latter is important in that venous concentrations of an analyte in the ante-cubital vein may not reflect accurately concentrations at all other sites measured at that point in time.

·   Number of phlebotomist used

4.5.3.2  Anticoagulants, stabilisers and sample preservatives used should be described in detail or a suitable reference given to the technical specification for the system used.

4.5.3.2.1          Evidence should be presented to identify stability of the analyte under conditions collected

4.5.3.3  Stability of samples should be considered, as this should influence the analytical approach used to generate the required numerical data.

4.5.3.3.1          Pre-analytical processing conditions adequately should be described and assessed to be appropriate for the stability of the sample.

4.5.3.3.2          Storage conditions should guarantee stability of the analyte. Degradation during storage under certain conditions may affect older specimens more.

4.5.3.3.3          Evidence should be presented demonstrating stability under storage conditions used.

 

4.6 Conditions for analysis of samples

A description of the conditions  under which the samples were analysed is critical.

4.6.1        Model for analysis should  appropriate to the stability of the analyte

4.6.2        Analysis  should enable a robust measure of analytical imprecision and minimises the impact of analytical variability on the data set. (Optimal Conditions Precision preferred).

Back to top page

5.            Data Analysis

The usefulness of the reported indices of biological variation is dependent upon the fact that data set is truly representative of the population described, and that there is homogeneity the data, that is to say that the observations are representative of  a single “collective”. This means that outlying data  not representative of the population must be excluded. This further endorses the need for clear characterisation of the population studied to accompany the published data and to be accessible from   any data-based collection of biological variation data. Examination of the data and exclusion of outliers is a critical step to be undertaken prior to the final analysis of variance. If this is not undertaken the reported indices of biological variation   must be considered to be critically flawed.  Heterogeneity of the variances compromises the usefulness and generalisability of biological variation data.  It is highly recommended that the paper of Fraser and Harris1 is considered as a seminal paper in this context.

5.1 Outlier analysis

There a number of statistical techniques that can be used to exclude outliers These should be applied to the analytical data within  the following contexts with outliers identified excluded from the final data analysis: -

5.1.1        Outlying observations.

If the assays are done in duplicate and one result is excluded the remaining value should also be excluded.

If there is an outlying value observed within a subjects data set, that may be excluded and the remaining values used in the subsequent analysis of variance.

5.1.2        Outlying within subject variances (see 5.2)

5.1.3        Outliers among mean values of subjects:

A subject   within the group may have an outlying mean suggesting that the belong to another population. Again this can be tested using standard  statistical methods.  Graphical plots of each subjects mean value  with the range of surrounding values is a godd initial way of identifying the possibility of outliers in this context.

5.2 Heterogeneity of variance assessed

The question arise as to whether the within subject variances calculated fro each subject is an estimate of  a common true variance. This examination must be undertaken to enable delivery of data. This can be tested using the approach of Harris2 which enables an index of heterogeneity to be calculated. See also Fraser and Harris1.                                                                                                                               Back to top page

 

5.3 Statistical methods described and appropriate

Statistical methods used should be appropriately identified and referenced.  Methods for outlier identification should be robust and clearly identified as having been applied.

5.3.1        Methods described

Outlier identification.

Statistical packages used.

5.3.2        Data appropriately transformed

5.3.3        Groups appropriately stratified

Data should be assessed to identify whether there is a need to stratify  into subsets, based on characteristics such as sex, age etc.

 

 

Back to top page


 

6.            Results

 

6.1 Terminology

Unified terminology may aide understanding and  transportation of the data. Reference to “Group” and “Subject” using subscript “G” and “I” to refer to within group and within subject (used to be identified as between individual and intra-individual)   indices.

Therefore standard deviation group would be represented as SDG with coefficient of variation being identified as CVG. Within subject these would be SDI and CVI respectively. Analytical indices will carry the subscript “A”.  This is consistent with the terminology described by Fraser and adopted by Ricos et al.

6.2 Results clearly presented and managed

Format of results will need to meet the requirements of the publishing journal. Tabulation of appropriately stratified biological variation data with derived indices would be advantageous. The matrix studied and the duration of the study should be clearly identified.

Some work may need to be undertaken to define a data archetype model for biological variation data. This will include the relevant metadata and descriptors to accompany  the biological variation data to enable wider dissemination and appropriate usage. This may for instance include as a minimum description of the population from which they were derived, state of well being, specimen types, length of the study, the analytical method, used possibly coded in the form of a LOINC code (Logical Observation Identifiers Names and Codes (LOINC®)), and a reference to the published data.

 The results section should clearly identify the results of  outlier analysis undertaken and confirm homogeneity of the data sets

If data are stratified the parameters used to enable this should be clearly characterised.

 

 

Back to top page

7.            Discussion

The discussion of the data should clearly include a focus on the limitations; generalisability of the data should always be included in addition to any other focus

7.1  Limitations of Study

7.1.1        These may range from technical issues relating to the analytical system to issues relating to the data set identified during the statistical analysis.

7.2 Generalisability

7.2.1        Discussion  of Factors that may impact on  extrapolation  of the data  to similar or dissimilar populations at other locations

7.2.2        Potential difficulties with the analytical methods used that may affect extrapolation of the data to populations served by alternative techniques

7.3 Interpretation

 

Back to top page

 

 

 

 

 


References

  1. Fraser CG, Harris EK. Generation and application of  data on biological variation in Clinical Chemistry.  Crit Rev in Clin Lab Sci 1989:27(5);409-437.
  2. Harris, E.K. Distinguishing physiologic from analytic variation. J. Chronic Dis. 1970;23:469–480.

  3. Fraser CG. Biological Variation from Principles to Practice. Washington DC: AACC Press 2001
  4. Ricos C, Iglesias N, Garcia-Lario J, Simon M, Cava F et al. Ann Clin Biochem. 2007;44:343-352
  5. Logical Observation Identifiers Names and Codes, LOINC.  http://loinc.org/
  6.  
  7.  
  8.  
  9.  

Back to top page