A method for estimating age‐specific reference intervals (‘normal ranges’) based on fractional polynomials and exponential transformation

Summary. The age-specific reference interval is an important screening tool in medicine. Put crudely, an individual whose value of a variable of interest lies outside certain extreme centiles may be suspected of abnormality. We propose a parametric method for constructing such intervals. It provides smooth centile curves and explicit formulae for the centile estimates and for standard deviation (SD) scores (age-standardized values). Each parameter of an exponential-normal or modulus-exponential-normal density is modelled as a fractional polynomial function of age. Estimation is by maximum likelihood. These three- and four-parameter models involve transformations of the data towards normality which remove non-normal skewness and/or kurtosis. Fractional polynomials provide more flexible curve shapes than do conventional polynomials. The method easily accommodates binary covariates facilitating, for example, parsimonious modelling of ageand sex-specific centile curves. A method of calculating precision profiles for centile estimates is proposed. Goodness of fit is assessed by using Q-Q-plots and Shapiro-Wilk W-tests of the SD scores, and likelihood ratio tests of the parameters of an enlarged model. Four substantial real data sets are used to illustrate the method. Comparisons are made with the semiparametric LMS method of Cole and Green.

[1]  P Royston,et al.  Constructing time-specific reference ranges. , 1991, Statistics in medicine.

[2]  Bryan F. J. Manly,et al.  Exponential Data Transformations , 1976 .

[3]  S. Shapiro,et al.  An Analysis of Variance Test for Normality (Complete Samples) , 1965 .

[4]  D. Altman,et al.  Charts of fetal size: 1. Methodology , 1994, British journal of obstetrics and gynaecology.

[5]  T Tango,et al.  Estimation of normal ranges of clinical laboratory data. , 1986, Statistics in medicine.

[6]  P. Royston,et al.  Regression using fractional polynomials of continuous covariates: parsimonious parametric modelling. , 1994 .

[7]  R. Simon,et al.  Flexible regression models with cubic splines. , 1989, Statistics in medicine.

[8]  N. Draper,et al.  An Alternative Family of Transformations , 1980 .

[9]  H B Valman,et al.  Serum immunoglobulin concentrations in preschool children measured by laser nephelometry: reference ranges for IgG, IgA, IgM. , 1983, Journal of clinical pathology.

[10]  B. Lewis,et al.  Blood lipid concentrations and other cardiovascular risk factors: distribution, prevalence, and detection in Britain , 1988, British medical journal.

[11]  Patrick Royston,et al.  A Comparison of Statistical Methods for Age‐related Reference Intervals , 1997 .

[12]  D. Ruppert,et al.  Transformation and Weighting in Regression , 1988 .

[13]  T J Cole,et al.  Smoothing reference centile curves: the LMS method and penalized likelihood. , 1992, Statistics in medicine.

[14]  D. Altman.  Construction of age-related reference centiles using absolute residuals. , 1993, Statistics in medicine.

[15]  Interval Estimation for the Three‐Parameter Lognormal Distribution Via the Likelihood Function , 1980 .

[16]  A F Roche,et al.  NCHS growth curves for children birth-18 years. United States. , 1977, Vital and health statistics. Series 11, Data from the National Health Survey.

[17]  Tj Cole,et al.  Fitting smoothed centile curves to reference data (with discussion) , 1988 .

[18]  H E Solberg,et al.  The theory of reference values Part 5. Statistical treatment of collected reference values. Determination of reference limits. , 1983, Journal of clinical chemistry and clinical biochemistry. Zeitschrift fur klinische Chemie und klinische Biochemie.

[19]  Patrick Royston,et al.  Age-specific reference intervals (“normal ranges”) , 1997 .

[20]  M. Healy,et al.  Distribution-free estimation of age-related centiles. , 1988, Annals of human biology.

[21]  J. Wingerd.  The relation of growth from birth to 2 years to sex, parental size and other factors, using Rao's method of the transformed time scale. , 1970, Human biology.

[22]  P. Royston.  Approximating the Shapiro-Wilk W-test for non-normality , 1992 .

[23]  G. Box.  An analysis of transformations (with discussion) , 1964 .

[24]  A E Ades,et al.  Age-related reference ranges: significance tests for models and confidence intervals for centiles. , 1994, Statistics in medicine.

[25]  T. Cole.  The Use and Construction of Anthropometric Growth Reference Standards , 1993, Nutrition Research Reviews.