Box-Cox Transformaton

In statistical / biostatistical analysis, it is pretty common to apply the data transformation technique. The reason is to achieve the normality assumption. data transformation refers to the application of a deterministic mathematical function to each point in a data set — that is, each data point zi is replaced with the transformed value yi = f(zi), where f is a function.

The typical data transformations include logarithm, square root, Arcsine transformation. Log transformation is suitable for variables with log-normal distributions. The square-root transformation is commonly used when the variable is a count of something. For arcsin transformation, the numbers to be transformed must be in the range −1 to 1. This is commonly used for proportions, which range from 0 to 1.

Another popular data transformation technique Box-Cox transformation, which we may not use frequently in clinical trials. Box-Cox transformation belongs to the so-called 'power transform'. The Box-Cox family of transformations has two useful features: first, it includes linear and logarithmic transformations as special cases; and, second, it possesses strong scale equivariance properties, including the property that the transformation parameter is unaffected by the rescaling. Application of the Box-Cox transformation algorithm reduced the heterogeneity of error and permitted the assumption of equal variance to be met. Its main disadvantage is that both the domain and the range of the transformation are, in general, bounded.

Box-Cox transformation can be easily implemented with SAS Proc Transreg.

Further readings:

• Box-Cox Transformation: an Overview by Pengfei Li

• Basic Box-Cox Transformations using SAS Proc Transreg
• A Box-Cox Transformation using SAS Stat Studio
• What to do when data are non-normal?

Dealing with the paired data

Paired data contains values which fall normally into pairs and can therefore be expected to vary more between pairs than within pairs. The pairing is to reduce the variability. After the pairing, The between-subject variability will be eliminated. If pairing is effective it will reduce variability enough to justify the effort involved to obtain paired data.

There are many practical examples of paring. In clinical trial, crossover design is a special case of the pairing where the same subject receive more than one treatment. If all subjects receive treatment A, then treatment B, it can still be called crossover design (single sequence cross over design). In Epidemiology field, the case-control study is typically paring. There are terms 1:1 matched case-control, and 1:m matched case-control. In education, we can do the paring to compare the scores before and after the training;......

When outcome measures are continuous variable (such as drug concentration), without considering the covariates, analysis of paired data can be implemented by using paired t-test which can be easily performed using SAS PROC UNIVARIATE (calculate the difference for each pair, then run PROC UNIVARIATE) or SAS PROC TTEST (without calculating the difference first). Suppose x1 and x2 are paired variables,

proc ttest;
 paired x1*x2;
run;

If the normality assumption is questionable, the non-parametric tests (sign test and Wilcoxon signed rank sum test) can be used. UCLA's Statistical Consulting Services web site provided examples for these tests.

In more complicated situation (such as crossover design) or if we have to do the modeling to include the covariates, mixed model needs to be used. SAS PROC MIXED can implement the mixed model easily. See SAS/Stat User's Manual for PROC MIXED. In a research paper titled "Detection of emphysema progression in alpha 1-antitrypsin deficiency using CT densitometry; Methodological advances", I actually dealt with the paired data using so called 'random coefficient model'.


When outcome variable is discrete data, the easiest example is McNemar test. McNemar's test is performed if we are interested in the marginal frequencies of two binary outcomes. These binary outcomes may be the same outcome variable on matched pairs (like a case-control study) or two outcome variables from a single group.

In more complicated situation or if the covarites need to be included in the model, 'conditional logistic regression' needs to be employed. 'Conditional logistical regression' can be implemented using SAS Proc Logistic or SAS Proc PHREG. See following links for detail descriptions.

• A Tutorial on Logistic Regression by Ying So in SAS

• Condition Logistic Regression using SAS Proc PHREG procedure by David Brown

Pedistric use and geriatric use of drug and biological products

In the United States, every marketed drug or biological product needs to have its product label or package insert. The product label contains the use in special populations including pediatric and geriatric population. Here is a paragraph from FDA guidance on "Labeling for Human Prescription Drug and Biological Products — Implementing the New Content and Format Requirements"

Use in Specific Populations (§ 201.57(a)(13))

Information under the Use in Specific Populations heading includes a concise summary
of any clinically important differences in response or recommendations for use of the
drug in specific populations (e.g., differences between adult and pediatric responses, need
for specific monitoring in patients with hepatic impairment, need for dosing adjustments
in patients with renal impairment). Typically, information under this heading includes
limitations or precautions for specific populations or established differences in response.

Absence of the clinical study data in pediatric and geriatric population could sometimes cause problems in product label or in the drug approval process. During the drug development process, it is prudent to consider the inclusion/exclusion of patient population in terms of the age limit. In the study protocol, the inclusion criteria pertinent to the age limits (upper and lower limits) should be carefully considered. In the statistical analysis, when data for pediatric and/or geriatric population is available, subgroup analysis should always be performed.

In regulatory environment, the classification of the pediatric and geriatric population are defined as:

Pediatric population: according to ICH guidance E11 "Clinical Investigation of Medicinal Products in the Pediatric Population", the pediatric population contains several sub-cateogires:

• preterm newborn infants
• term newborn infants (0 to 27 days)
• infants and toddlers (28 days to 23 months)
• children (2 to 11 years)
• adolescents (12 to 16-18 years (dependent on region))

Notice that in FDA's guidance "General Considerations for Pediatric Pharmacokinetic Studies
for Drugs and Biological Products", the age classification is a little bit different. I am assuming that the ICH guidance E11 should be the correct reference.

Geriatric population:
Geriatric population is defined as persons 65 years of age and older. There is no upper limit of age defined. The Food and Drug Administration has regulations governing the content and format of labelling for human prescription drug products, including biological products, to include information pertinent to the appropriate use of drugs in the elderly and to facilitate access to this information by establishing a “Geriatric use” subsection in the labelling.

Further readings:

• Regulatory requirements for the development of medicinal products for pediatric use by Dr. Tom Sam
• FDA good review process "Labeling for Human Prescription Drug and Biological Products — Determining Established Pharmacologic Class for Use in the Highlights of Prescribing Information"
• FDA's Pediatric Drug Development

GxP: a collection of quality guidelines in clinical trial

GxP is now used to represent a collection of quality guidelines in clinical trial. The titles of these good practice guidelines usually begin with "Good" and end in "Practice", with the specific practice descriptor in between. A "c" or "C" (stands for 'current') is sometimes added to the front of the acroynm to form cGxP. For example, cGMP is an acronym for "current Good Manufacturing Practices."

Professionals who are working in the pharmaceutical or biotechnology industry should be very familiar with three common GxPs: GCP, GMP, and GLP.

GCP: Good Clinical Practices is an international ethical and scientific quality standard for designing, conducting, recording and reporting trials that involve the participation of human subjects. The GCP is governed by ICG guideline E6. To learn more about GCP, watch the GCP 101: An introduction at FDA website.

http://www.fda.gov/Training/CDRHLearn/ucm176411.htm

GLP: Good Laboratory Practice. Refer to Wikipedia for detail. GLP is the guidance for laboratory tests, pre-clinical tests, bioanalytical assays/measures, toxicology tests,...

cGMP: current Good Manufacturing Practice regulations for drugs contain minimum requirements for the methods, facilities, and controls used in manufacturing, processing, and packing of a drug product. FDA has many guidance on cGMP.

Recently, many other GxP terms have surfaced. It looks like that each functional area in clinical trial will have its own GxP. Below are some examples: GRP, GPP, GSP, and GCDMP.

GRP: Good Reprint Practices. In January 2009, FDA issued its final version of the guidance "Good Reprint Practices for the Distribution of Medical Journal Articles and Medical or Scientific Reference Publications on Unapproved New Uses of Approved Drugs and Approved or Cleared Medical Devices"

GPP: Good Pharmacovigilance Practices. In 2005, FDA issued its guidance on "Good Pharmacovigilance
Practices and Pharmacoepidemiologic Assessment" to provide guidance on (1) safety signal identification, (2) pharmacoepidemiologic assessment and safety signal interpretation, and (3) pharmacovigilance plan development.

While GRP and GPP are proposed by the regulatory agencies, there is no officially issued guidance on GSP (Good Statistical Practices) and GCDMP (Good Clinical Data Management Practices). However, the principles of these two good practices have been largely covered in ICH guidances, specifically, E9 (Statistical Principles for Clinical Trials) and E6 (Good Clinical Practice).

PSI Professional Standards Working Party developed a GUIDELINES FOR STANDARD OPERATING PROCEDURES for Good Statistical Practice in Clinical Research.

In several DIA presentations, Good Statistical Practices were said to include the following components:

Science:

• Protocol – Minimize bias – Maximize precision
• Analysis plan
• Presentation of results
• Leadership

Operational Processes

• Controlled statistical environment
  
• SOPs • Productivity tools
  
• Data standards
  
• Training
  

Credibility Results

• Reproducible research
  
• Transparent and efficient processes
  
• Validated analysis
  
• Data integrity assurance
  

The Good Clinical Data Management Practices (GCDMP) is developed by the SCDM (Society of Clinical Data Management). It provides assistance to clinical data managers in their implementation of high quality clincal data management processes and is used as a reference tool for clinical data managers when preparing for CDM training and education.

Biostatistics conferences/workshops

When I started my career in biostatistics, I joined the American Statistical Association (ASA) and attended its annual meeting (Joint Statistical Meeting) rotated in different large cities in North America (US and Canada). I have enjoyed the atmoshpere of the conference and networked with friends and professors in the statistical field.

For two consective years, I have skipped the meeting. Instead, I attended the FDA/Industry Statistical Workshop. JSM may be good to the students, but may not be good for professions (especially for statisticians who are working in drug development area). JSM has a lot of sessions/presentions that are unfiltered and too theoretic. A lot of stuff may never have the value in application. Even though it may be applicable one day, it may not be acceptable to the regulatory agencies.

The statistical conferences, symposiums, workshops with focus on clinical trial and drug development have thrived in recent years. Twice a year, FDA holds its workshops: one with Drug InformationAssociation "FDA/industry statistical forum" and one with ASA "FDA/industry statistical workhop". These conferences are more specific to the biopharmaceutical field and the topics are more relevant to the daily work of biostatisticians.

There are also several societies with focus on biostatics, for example, the International Society for Biopharmaceutical Statistics (ISBS) and the International Society for Clinical Statistics (ISCS). The International Chinese Statistical Association (ICSA) is also adjusting its focus to the biopharmaceutical field. Within ASA, biopharmaceutical network has been formed.

To get a flavor of the topics in these meetings, see the following links:

• 2009 FDA/Industry Statistical Workshop online program and presentations (The password to view presentations is Capitol
• Biopharmaceutical network's webinar topics and handouts

Positive Psychology - science to find happiness

When I see a news headline about "what is the most popular course in the Harvard University?", my curiosity drives me to find out what the course is. This leads me to the concept of "Positive Psychology". The most popular class is the Psychology 1504 (ie, positive psychology) taught by Dr. Ben-Shahar.

As mentioned in NPR news, "almost every semester for the past ten years, the most popular class at Harvard has been Intro to Economics, or as Tal Ben-Shahar likes to call it, how to get rich, but today there's an even bigger class on campus. It's Ben-Shahar's course on what he calls, how to get happy."

According to Wikipedia, Positive psychology is a recent branch of psychology that "studies the strengths and virtues that enable individuals and communities to thrive". Positive psychologists seek "to find and nurture genius and talent", and "to make normal life more fulfilling", not simply to treat mental illness. In other words, the positive psychology deals with love, happiness, job satisfaction, ...

In contrary to the Positive psychology, there should be a concept of negative psychology. However, even though the current psychology is so focused on the negative side (depression, fear, anxiety, mental illness,...), there is no formal definition of negative psychology.

Further readings about the negative psychology:

• The website of Tal Ben-Shahar
• Tal Ben-Shahar's video on Youtube
• Dr Seligman's vedio on Youtube
  
• NPR report on finding happiness
• Harvard psychology course information (including course slides)
• UPenn's Positive Psychology Center
• Good Life: Happiness on Youtube
  

Unlike the negative psychology which belongs to the medical science, the positive psychology has its applications in corporate business. It could be used to promote the positive culture, attitudes, employee's job satisfaction,...

However, there is also negative side about the positive psychology. See Dr. BARBARA S.HELD's argument.

In practice, Positive psychology encompass a variety of techniques that encourage people to identify and further develop their own positive emotions, experiences, and character traits. In many ways, positive psychology builds on key tenets of humanistic psychology. Whether or not the positive psychology techniques work will eventually rely on the evidence from the clinical trials. Since the psychology measures are typically intangible, how to design a trial or intervention, what to measure, how long to measure, what instrument to use,... could be challenging even more than the typically psychology measures (with the focus on disease or negative psychology). The following paper discussed this issue.

Overtreated, excess care

"Overtreated", "Overdiagnosed", and "Overdosed",... these are the terms I have used in one of seminars several years ago. By comparing the health care system between the United States and the China, you could easily think of these terms, especially when I heard the new medical conditions "ADHD - Attention-Deficit Hyperactivity Disorder", "M-IBS - Mixed Irritable Bowel Syndrome", "Chronic Fatigure Syndrome (CFS)", "fibromyhalia"; when I saw the images how many pills a patient took regularly.

Driven by the NPR interview with Shannon Brownlee (Are Today's Hospital Patients "Overtreated"?), I went to the local library to borrow her book "Overtreated: why to much medicine is making us sicker and poorer". I enjoyed very much in reading this book.

I intended to write a blog about this book, then found that many people had already expressed their opinion about this book. See Book Reviewer's comments from Amazon

Even though this book was written two years ago (in 2007), the arguments, the facts, the reasoning described in this book is very much relevant to the situation today (when the debate on the health care reform heats up). Below is a list of chapters:

• One: Too Much Medicine
• Two: The Most Dangerous Place
• Three: Your Local Hospital
• Four: Broken Hearts
• Five: The Desperate Cure
• Six: The Limits of Seeing
• Seven: The Persuaders
• Eight: Money, Drugs, and Lies (my favorite chapter)
  
• Nine: The Doctor Isn't In
• Ten: Less is More

Instead of going to detail, I would just cite some sentences from the book:

• "Doctors have a saying: Never get admitted to a teaching hospital in July, because that's when all the new interns arrive fresh from medical schools."
• "As research would show over the coming decades, stunningly little of what physicians do has ever been examined scientifically, and when many treatments and procedures have been put to the test, they have turned out to cause more harm than good."
• "Every patient admitted to a hospital risks being hurt or even killed by the very people who wish to help her."
• "Even as the number of [medical] imaging tests [X-ray, CT, MRI] is going up, numerous studies suggest that all those pictures are not nearly as effective at improving diagnosis as many doctors--and patients--tend to think."
• "The drug company representative, or drug rep, usually [is] a handsome young man or shapely young woman who has been recruited more for his or her good looks and outgoing personality than for his or her aptitude for science or medicine."
• "Among drug reps the unofficial name for thought leaders who work for multiple companies is 'drug whores'"
• "The more specialists involved in your health, the more likely it is that you will suffer from a medical error, that you will be given care you don't need and be harmed by it."
• "The Institute of Medicine estimates that only 4 percent of treatments and tests are backed up by strong scientific evidence; more than half have very weak evidence or none."
• "In the view of Richard Horton, a British physician and editor of the prestigious medical journal the Lancet, 'Journals have devolved into information-laundering operations for the pharmaceutical industry'"
• Says John abramson "The primary mission of medical research has been transformed. It used to be all about gathering information to improve health. Now clinical research is aimed at gathering information that will maximize return on investment"
  

Further readings:

• Are Americans Seeing Their Doctors Too Often? by Shannon Brownlee
• Overtreated blog by Shannon Brownlee
• Getting Well: It's About Time by Melinda Beck "Americans spend billions of dollars a year to treat things that would go away on their own."
• HARLOT plc: an amalgamation of the World's Two Oldest Professions by David Sackett and Andrew Oxman. BMJ
• Overdosed America: The Broken Promise of American Medicine by John Abramson