Abstract
Background
Many nursing and health related research studies have continuous outcome measures that are inherently nonnormal in distribution. The BoxCox transformation provides a powerful tool for developing a parsimonious model for data representation and interpretation when the distribution of the dependent variable, or outcome measure, of interest deviates from the normal distribution. The objectives of this study was to contrast the effect of obtaining the BoxCox power transformation parameter and subsequent analysis of variance with or without a priori knowledge of predictor variables under the classic linear or linear mixed model settings.
Methods
Simulation data from a 3 × 4 factorial treatments design, along with the Patient Falls and Patient Injury Falls from the National Database of Nursing Quality Indicators (NDNQI^{®}) for the 3^{rd }quarter of 2007 from a convenience sample of over one thousand US hospitals were analyzed. The effect of the nonlinear monotonic transformation was contrasted in two ways: a) estimating the transformation parameter along with factors with potential structural effects, and b) estimating the transformation parameter first and then conducting analysis of variance for the structural effect.
Results
Linear model ANOVA with Monte Carlo simulation and mixed models with correlated error terms with NDNQI examples showed no substantial differences on statistical tests for structural effects if the factors with structural effects were omitted during the estimation of the transformation parameter.
Conclusions
The BoxCox power transformation can still be an effective tool for validating statistical inferences with large observational, crosssectional, and hierarchical or repeated measure studies under the linear or the mixed model settings without prior knowledge of all the factors with potential structural effects.
Keywords:
Data transformation; NDNQI; Nursing quality indicator; ANOVA, Mixed modelBackground
Many health and nursing related studies focus on outcome measures that can be used
to identify superior treatments and/or to reveal deficiencies in practices [1]. While substantial effort has been made on research design and data collection, researchers
are more concerned with the validity of statistical conclusions should the reliability
of the measurement be compromised [2] or the basic statistical assumptions be violated because nonnormal data distributions
with these outcomes are common [3]. In the later case, data transformation is one of the powerful tools for developing
parsimonious models for detecting structural effects or predictive factors and for
better data representation and interpretation [46]. Ever since the pioneer works on the formal estimation of a suitable transformation
[3], the nonlinear monotonic power transformation family in the form of
Basic assumption for linear model methodology
Statistical analyses with the linear model methodology are based on the assumption that the population being investigated is normally distributed with a common variance and additive mean structure [10,11]. Let Y_{ijk }be the response for the k^{th }unit in the ij^{th }subclass for a twoway classification model; β is the vector of regression parameters, and X_{ij }is the design matrix for the ij^{th }subclass, the linear model (1) then assumes that the error is independent and identically distributed normal variable, ε_{ijk }~ N (0, σ ^{2 }), after removing the structural effect X_{ij}β.
When the theoretical assumption is not satisfied, data transformation can be applied so that inferences about unknown factors are still valid on the transformed scale [11]. Depending on the type of data and the form of their distribution, a number of different transformations were found so that the transformed data would meet the theoretical assumptions. These include: logit transformations for proportions; the square root transformation for count data; a logarithm or inverse transformation for continuous data skewed to either side with a heavy tail, etc. The family of power transformations is useful when the choice of transformation to improve the approximation of normality is not obvious [12]. The power transformation was first introduced by Tukey [13] and later modified by Box & Cox [3] to take account of the discontinuity at λ = 0. The BoxCox power transformation takes the following form (2) so that the transformed values are a monotonic function of the observations,
and for the unknown transformation parameter, λ,
where, Y_{ijk}, X_{ij}, β and ε_{ijk }are all defined as in equation (1). This transformation may allow the response variable to achieve simplicity and additivity in mean structure for the expected value of (y^{λ}) and make the variance more nearly constant among points in the factor space [14].
Substantial research has been conducted on the theoretical aspects of BoxCox modification [15], and a wide variety of applications used BoxCox transformation [1618]. It is reported that maximum likelihoodbased variance components analysis applied to nonnormal data had inflated type I errors, which were controlled best by BoxCox transformation [19]. BoxCox transformation can be used to improve signal/noise ratio, map families of distributions and result in more efficient and robust results [20]. Analysis of the diagnostic accuracy using the receiver operating characteristic curve methodology required a BoxCox transformation within each cluster to map the test outcomes to a common family of distributions [21]. Recently, median regression after applying the BoxCox transformation was reported as notably more efficient and robust than the standard least absolute deviations estimator [22]. Due to its highly structured nature, however, the BoxCox power transformation model is controversial, as some theoretical and Monte Carlo studies indicated that the data based estimate of λ is unstable and that, much like the case of multivariate collinearity, λ and β are highly correlated [79,16,17]. Other studies, however, downplayed the cost from databased BoxCox transformation, arguing the cost should be moderate on the whole and seldom large [23]. It has been suggested that we need to understand better the joint effects of variable selection and data transformation [7,8,23]. Under the BoxCox transformation (2), one can put the data on the correct scale for an ANOVA model when the predictor variables (X) are identified and included during the transformation process. Unfortunately, for many nonrandomized studies it is not clear what predictor variables should be included when the dependent variable deviates significantly from the normal distribution.
Under the linear mixed model setting, the error term of ε_{ijk }in model (3) is no longer independent and identically distributed (iid) normal, but rather correlated because sampling and experiment units may be hierarchical or each sampling unit may be repeatedly measured.
NDNQI database overview
In 1998, NDNQI^{® }was established by the American Nurses Association (ANA) to monitor nursingsensitive indicators that measure nursing quality and patient safety across all 50 states in the US [24]. Over the last decade, NDNQI has seen its participating hospitals grow from 35 in 1998 up to 1,450 by the end of 2009 [25]. With nursing data collected at the unit level within member institutions, NDNQI provides hospitals unitlevel performance reports with 8quarter trend data, along with national comparison data grouped by hospital staffed bed size, teaching status, Magnet status, various other hospital characteristics, and unit type [25].
Nursingsensitive indicators reflect the structure, process and outcomes of nursing care. Examples of nursing structure measures include the supply of nurses, skill level, RN education and certification [2426]. The Patient Falls indicator is an example of a nursing sensitive outcome and is defined as the rate per 1,000 patient days at which patients experience an unplanned descent to the floor during the course of their hospital stay.
Patient Injury Falls, as another example, is defined as:
Both Patient Falls and Patient Injury Falls have a common denominator of Total Number of Patient Days. Conceptually, a patient day is 24 hours, beginning with the hour of admission. The operational definition of patient days is the total number of inpatients present at the midnight census plus the total number of hours of short stay patients divided by 24. Short stay patients are patients on a unit for less than 24 hours either for observation or same day surgery.
Both Patient Falls and Patient Injury Falls are critical nursing quality indicators that may be associated with nursing workforce characteristics, as well as with unit type and some hospital characteristics such as teaching status and Magnet status. Other unknown factors might also affect the rates of Patient Falls and Patient Injury Falls in NDNQI hospitals across a wide spectrum of settings over the entire United States. Further, if such factors do exist, it would be of great interest to examine what administrative or nursing process adjustments a hospital might take to reduce these rates and thus improve the overall quality of service.
Methods
The BoxCox power transformation requires all predictor variables to be included in the model for estimating transformation parameter in order to put a skewed response onto the correct scale for the classic ANOVA model [27]. In this paper, a Monte Carlo simulation with a 3 × 4 factorial treatment design was used to contrast the properties of powertransformed response variables with and without the presence of the 3 × 4 factorial structural effects when the transformation parameter was estimated. The residual and the treatment main effects with the simulation were examined with twoway ANOVA model. NDNQI Patient Falls and Patient Injury Falls, collected on unit level, are correlated within hospitals and rightskewed in distribution. Statistical analysis without data transformation may violate the underlying assumption because of nonnormal error distributions, potentially also compounded with a correlated covariance structure. For illustration purpose, we first ignored the within hospital intra class correlation (ICC) and then extended the BoxCox power transformation into the linear mixed model framework [26] and analyzed NDNQI Patient Falls and Patient Injury Falls with mixed models assuming compound symmetric covariance structure [28] to contrast the effect of BoxCox transformations when predictor variable (Hospital Teaching and Magnet Status) were included in the transformation model with when they were ignored. Note, in NDNQI quarterly reports, ICC for all indicators were actually properly adjusted [29].
Patient Falls and Patient Injury Falls data from 6726 nursing units in 926 hospitals for the 3^{rd }quarter in 2007 were extracted from the NDNQI database maintained by NDNQI project at The Kansas University School of Nursing. The number of nursing units per hospital ranged from 1 to 36 with a median of 6 ± 5 (interquartile range). Along with the two indicators, hospital teaching status (Academic Medical Center; Other Teaching; NonTeaching) and Magnet status (Magnet vs. NonMagnet) were chosen from a variety of stratification variables for illustrative purposes. BoxCox transformation on Patient Falls and Patient Injury Falls were then applied both with and without inclusion of these predictors in the model with which the power transformation parameters were estimated.
Monte Carlo Simulations
All simulated data are based on a completely randomized block design with 3 × 4 factorial treatments, which can be expressed in the following model
where
The simulation study on a two factor, completely randomized, block design was aimed to answer the following two questions.
1. Will the goal of simplicity in structure and homogeneity in error for transformation be still achievable if predictor variables are omitted from the power transformation model?
2. What are the consequences of conducting the analysis of variance on the transformed response variable without including the predictor variables in estimating the transformation parameter (λ)?
Application to NDNQI Indicators
Suppose one is interested in investigating Patient Falls or Patient Injury Falls as a function of hospital teaching and/or Magnet status, then X_{ij }in (1) has 6 columns with the first being a column of 1's, the 2^{nd }and 3^{rd }representing the teaching status, the 4^{th }an indicator for Magnet status, and the 5^{th }and 6^{th }for the Teaching by Magnet status interaction. After exploratory data analysis using the ANOVA model with hospital teaching and Magnet status as having structural effects, Patient Falls and Patient Injury Falls were analyzed with the mixed model under a) without transformation, b) power transformed without teaching and Magnet effects during the parameter estimation for (λ_{0}), and c) power transformed with teaching and Magnet effects during the parameter estimation (λ_{1}). The power transformation parameter, λ_{0}, was obtained through a grid search by maximizing the log likelihood of the residual for the transformed response variable after removing the overall means. As Gurka et al. [7] proposed, we obtained λ_{1}through maximizing the residual maximum likelihood (REML) with the existing computational procedures (SAS PROC Mixed). Specifically, for each indicator, a scaled BoxCox transformation [3] for a wide range of the power parameter value, λ_{i }(i = 1 to 8 by 0.01) was first applied. Then, each transformed response was analyzed with the compound symmetry covariance structure to model the correlation among units within hospital. The λ_{i }that corresponds to the maximum REML was selected as λ_{0}.
Results
One of the main objectives for the BoxCox power transformation is to achieve normality
in random error distribution after removing the additive effects. With simulated data
under model (4), residuals from the 3 × 4 factorial ANOVA models with either
Figure 1. Test for normality by ShapiroWilk statistic for residual obtained from 3 × 4 ANOVA with response variable being nontransformed (red), power transformed without treatment effect (blue), and power transformed with treatment effect and their interaction (green). The horizontal dot line represents 5% significant level. All box plots are obtained from 30 datasets selected at random from 1000 simulation (for clarity) with whiskers representing the 10^{th }and 90^{th }percentiles.
Table 1. Statistics for power transformation parameter and statistical test for structural effects based on Monte Carlo simulations
The other objective with BoxCox power transformation is to achieve simplicity and
additivity by strengthening the main effects while reducing the effect of interaction
terms [3]. In regard to the two factor main effect, the same conclusion was reached with either
Figure 2. Fvalues from ANOVA table for the interaction effect by 3 × 4 factorial treatment design on response variable being nontransformed (red), power transformed without treatment effect (blue), and power transformed with treatment effect and their interaction (green). The green line represents 5% significant level. All box plots are obtained from 1000 simulated datasets with whiskers representing the 10^{th }and 90^{th }percentiles.
Table 2. Statistics for tests of structural effect with different transformation models based on Monte Carlo simulations
With the extracted NDNQI data, exploratory data analysis showed severely skewed distributions
for Patient Falls and Patient Injury Falls (Figure 3a, b). Without transformation, the residuals after removing the structural effects of
interest (teaching and Magnet status and their interaction) using mixed models differed
clearly from normal distribution (Figure 4a, b). Residual distributions from the mixed model analyses with the 3 × 2 structural
effects for hospital Teaching status and Magnet status for the transformed response
Figures 3. Distribution of Total Falls (a) and Total Injury Falls (b), for NDNQI hospitals reported for 3^{rd }quarter, 2007.
Figure 4. Residual distribution (without data transformation) of Total Falls (a) and Total Injury Falls (b), for NDNQI hospitals reported for 3^{rd }quarter, 2007.
Figure 5. Residual distribution of Total Injury Falls (a) and Total Falls (b), for NDNQI hospitals reported for 3rd quarter, 2007. Residuals were obtained after removing the structural effect on power transformed dependent variable by hospital teaching and Magnet status. Here, the BoxCox power transformation parameters were obtained with structural effects in the model.
Figure 6. Grid search for optimum BoxCox power transformation parameters. Residual Maximum Likelihood (REML) reached maxima at 3.34 and 4.82 for the BoxCox power transformation parameters for Total Falls (a) and Total Injury Falls (b) estimated from repeated measure analysis with the linear mixed models.
Figure 7. Residual distribution of Total Injury Falls (a) and Total Falls (b), for NDNQI hospitals reported for 3^{rd }quarter, 2007. Residual is obtained after removing the structural effect on power transformed dependent variable by hospital teaching and Magnet status. Here, the power transformation parameters were obtained without structural effect in the model.
Table 3. Repeated measure analysis with the linear mixed model for Patient Falls and Patient Injury Falls for 2007 NDNQI 3^{rd }quarter
Figure 8. Repeated measure analysis for structural effect by Magnet and teaching status for Patient Falls (a) and Patient Injury Falls (b) for NDNQI hospitals reported for 3^{rd }quarter, 2007.
Discussion
The BoxCox power transformation provides an effective tool to justify the use of
the linear model when the response variable is not normally distributed. It was originally
defined as highly structured and required all predictor variables to be included in
the power transformation model [3]. There is always a cost resulting from selection of the transformation expressed
as an inflated variance [7,16]. However, predictor variables may not always be clearly defined in practice. This
is especially true for exploratory data analysis, observational studies, or classification
and regression tree (CART) analysis aimed at finding potential relationships when
the distribution of the response variable deviates significantly from normality. In
such cases, applying the BoxCox power transformation to the response variable alone
and then searching for potential predictor variables was demonstrated to be effective
in terms of achieving constant error and simplicity of main effects in the simulations
and examples we examined. In our simulated data, the statistical tests for main effects
were slightly more conservative for
The real case examples with Patient Falls and Patient Injury Falls from the NDNQI database showed BoxCox power transformations both with and without structural effects for teaching and Magnet status included in the models for estimating the transformation parameters were equally effective in normalizing the residual distributions (Figures 5a, b, 7a, b). Table 3 shows the test statistics from hierarchical analysis allowing for correlation between error terms for structural effect by stratification variables (Teaching, Magnet, and their interactions).
With over 1800 hospitals (one in every thee general hospitals in the U.S.) contributing nursing indicator data to the NDNQI database today, it is as critical to provide users with valid national comparative data in nursingsensitive quality indicators. As hospitals are striving to improve the quality of their nursing service, they can turn to the NDNQI quarterly reports to identify potential problems. While most of the nursing quality indicators are skewed in distribution, the structural effects of hospital characteristics are not always clear. In such cases, the classic BoxCox power transformation can be applied to the nursing quality indicators, for a specific category of unit (such as pediatric or post surgical) with linear model analysis, or all units within hospital under the mixed model framework, prior to identifying the structural effects from a potentially large pool of variables.
Both the simulation study and real case analysis with NDNQI quarterly report data demonstrated that the consequence of omitting a structural effect from the BoxCox power transformation was limited. This is important given the fact that for many large healthrelated observational studies the number of potential structural effects may be quite large. As of 2008, NDNQI had over 20 potential structural effects for 34 nursing indicators. Participating hospitals benefit from meaningful, valid comparative information based on a number of demographic, social, administrative, and service related factors. Estimating BoxCox power transformation parameters on indicators without including the unknown, or sometimes unmeasured, structural effects can still provide participating hospitals with statistically valid comparisons.
A few limitations need to be noted. First, the BoxCox transformation works better only if the measure of interest relatively smoothly spread out. In other words, the method may fail if the data cluster on a few values. Secondly, it is necessary to conduct a grid search of the transformation in order to find the optimum parameter that maximizes the residual likelihood both under the linear and the mixed model settings. Otherwise, the subsequent analysis may differ depending on whether or not the structural effects were included in the estimating process for the transformation parameters. Our results suggested a fine grid search for the transformation parameter should be used regardless the inclusion of factors with potential structural effects and regardless of whether the analysis uses the linear or mixed model settings, because the agreement on test for the structural effects occurs only if both transformations are optimized. Lastly, potential interactions between parameter estimates for transformation and for linear and/or random effects remains unclear, and, interpretation for the transformed data analysis, as always, remains a challenge that warrants further research.
Conclusions
The validity of linear mixed modeling via maximum likelihood relies on the underlying assumption that the random effects and residuals of the dependent variable are normally distributed. Many health and nursing related outcome measures deviate from this assumption. While at the same time, factors with potential structural effects are of major interest and yet to be identified. Therefore, the BoxCox power transformation provides a powerful tool for developing parsimonious models (i.e. applying linear mixed modeling) for data representation and interpretation. By extending the power transformation into linear mixed model setting with NDNQI examples, we found limited difference from subsequent test of structural effects regardless of whether such structure is included or omitted during the parameter estimation for transformation. This allows analysts to transform variables earlier in the model building, making the process of applying BoxCox transformation much easier in practice.
Future work would be to employ some sort of a latent class analysis [30] on the NDNQI data and look for structural relationships within each class.
Abbreviations
NDNQI: National Database of Nursing Quality Indicators; ANOVA: Analysis of Variance; ICC: Intra Class Correlation; iid: independent and identically distributed; ANA: American Nurses Association.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
QH reviewed literatures, conducted statistical analysis, and drafted the manuscript; JM and BG advised on and supervised statistical analysis and provided critical input in drafting and revising the manuscript; ND supervised NDNQI data collection, evaluated unitspecific nursesensitive data, and provided overall guidance for the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This research was conducted under contract from the American Nurses Association (ANA). Dr. Nancy Dunton is the principal investigator.
References

Bonneterre V, Liaudy S, Chatellier G, Lang T, de Gaudemaris R: Reliability, validity, and health issues arising from questionnaires used to measure psychosocial and organizational work factors (POWFs) among hospital nurses: A critical review.
Journal of Nursing Measurement 2008, 16(3):207230. PubMed Abstract  Publisher Full Text

Strickland OL: Impact of Unreliability of Measurements on Statistical Conclusion Validity.
Journal of Nursing Measurement 2005, 13(2):8385. PubMed Abstract  Publisher Full Text

Ferketich S, Verran J: An overview of data transformation.
Research in Nursing & Health 1994, 17(5):393396. PubMed Abstract  Publisher Full Text

Leydesdorff L, Bensman S: Classification and powerlaws: the logarithmic transformation.
Journal of the American Society for Information Science & Technology 2006, 57(11):14701486. PubMed Abstract  Publisher Full Text  PubMed Central Full Text

Jaeger T: Categorical data analysis: away from ANOVAs (transformation or not) and towards logit mixed models.
Journal of Memory & Language [serial online] 2008, 59(4):434446. PubMed Abstract  Publisher Full Text

Gurka MJ, Edward LJ, Muller KE, Kupper LL: Extending the BoxCox transformation to the linear mixed model.

Lee JC, Lin TI, Lee KJ, Hus YL: Bayesian analysis of BoxCox transformed linear mixed models with ARMA(p,q) dependence.
J Statist Plan Infer 2005, 133:435451. Publisher Full Text

Spitzer JJ: A Monte Carlo investigation of the BoxCox transformation in small samples.
Journal of the American Statistical Association 1978, 73:488495. Publisher Full Text

Draper NR, Smith H: Applied Regression Analysis. Wiley Series in Probability and Statistics; 1998.

Johnson RA, Wichern DW: Applied Multivariate Statistical Analysis. PrinticeHall, Inc; 1998.

Turkey JW: The comparative anatomy of transformations.
Annals of Mathematical Statistics 1957, 28:602632. Publisher Full Text

Lindsey JK: The roles of transformation to normality.
Biometrics 1975, 31:247249. Publisher Full Text

Carroll RJ, Ruppert D: On prediction and the power transformation family.
Biometrika 1981, 68(3):609615. Publisher Full Text

Bickel PJ, Doksum KA: An analysis of transformation revisited.
Journal of the American Statistical Association 1981, 76:296311. Publisher Full Text

Oberg A, Davidian M: Estimating data transformations in nonlinear mixed effects models.
Biometrics 2000, 56:6572. PubMed Abstract  Publisher Full Text

Etzel CJ, Shete S, Beasley TM, Fernandez JR, Alliosn DB, Amos CI: Effect of BoxCox transformation on power of HasemanElson and maximumlikelihood variance components tests to detect quantitative trait loci.
Human Heredity 2003, 55(23):108116. PubMed Abstract  Publisher Full Text

Helene HT, Zwinderman AH: Comparing transformation methods for DNA microarray data.
BMC Bioinformatics 2004, 5:77. PubMed Abstract  BioMed Central Full Text  PubMed Central Full Text

O'Malley AJ, Zou KH: Bayesian multivariate hierarchical transformation models for ROC analysis.

Fitzmaurice GM, Lipsitz SR, Parzen M: Approximate median regression via the BoxCox transformation.

Carroll RJ, Ruppert D: The analysis of transformed data: comment.
Journal of the American Statistical Association 1984, 79:312313. Publisher Full Text

Dunton N, Gajewski BJ, Kluas S, Pierson B: The relationship of nursing workforce characteristics to patient outcomes.

Lake TL, Shang J, Klaus S, Dunton N: Patient falls: association with hospital Magnet status and nursing unit staffing.
Research in Nursing & Health 2010, 33:413425. PubMed Abstract  Publisher Full Text

Draper NR, Cox DR: On distributions and their transformation to normality.

Ramon CL, Milliken GA, Stroup WW, Wolfinger RD: SAS^{® }System for Mixed Models. Cary, NC: SAS Institute Inc; 1996.

Gajewski BJ, Mahnken JD, Dunton N: Improving quality indicator report cards through Bayesian modeling.
BMC Medical Research Methodology 2008, 8:77. PubMed Abstract  BioMed Central Full Text  PubMed Central Full Text

Di CZ, BandeenRoche K: Multilevel latent class models with Dirichlet mixing distribution.
Prepublication history
The prepublication history for this paper can be accessed here: