HKUMed Logo

HKU Li Ka Shing Faculty of Medicine
Biostatistics and Clinical Research Methodology Unit

Mission

Increasing importance is placed on evidence-based medicine, and the strength and quality of evidence used to guide clinical decisions. Clinical research studies require careful design and analysis in order to permit robust scientific inferences.

Recognising the need for core support in biostatistics and clinical research methodology, in January 2015 the Li Ka Shing Faculty of Medicine created the Biostatistics and Clinical Research Methodology Unit to provide support on quantitative research methods.

The mission of the Unit is to support research activities in the Faculty of Medicine through provision of expert advice on biostatistics and clinical research methodology. The objectives of the Unit are to provide expert advice on various aspects of biostatistics and clinical research methodology, including:

  • study design
  • sample size calculations
  • questionnaire design
  • planning and implementation of statistical analysis
  • analysis and interpretation of quantitative data

for grant applications, manuscripts for journal publication, and reports.

About Us

The Biostatistics and Clinical Research Methodology Unit is part of the Li Ka Shing Faculty of Medicine, and is hosted by the School of Public Health. The mission of the Unit is to support research activities in the Faculty of Medicine through provision of expert advice on biostatistics and clinical research methodology. The Unit also provides statistical consultancy to public bodies and private companies.

How much do we charge?

Advice on grant applications for Faculty members is free, while other services are charged based on hourly or daily rates. Please contact us to discuss the potential costs of ongoing support for data analysis and interpretation which will vary from project to project. Faculty members are encouraged to budget for any anticipated costs of data management and analysis in their grant applications, and to discuss this with us in advance of their grant application.

Contact Us

Please use the online form to make an initial request for support. This form includes basic details of your project that will help us to make efficient arrangements. We do not provide advice over the phone.

Our Team

Permanent Staff

Our Director is Dr. Helen Zhi, who graduated from Peking University in China before she went to the USA for graduate studies. She got PhD in statistics from Temple University, USA. Dr. Zhi has 15 years of work experience as Biostatistician/Associate Director in clinical trials from a top pharmaceutical company. She is an expert in all types of clinical studies and is familiar with various study designs. She was the lead statistician for FDA (US), EMA(EU), Health Canada and PMDA(Japan) submissions. She worked with many projects in different therapeutic areas, i.e., cardiovascular, metabolic, oncology, virology, neurology and musculoskeletal.

Senior Consultants

Ben Cowling is Professor and Head of the Division of Epidemiology and Biostatistics in the School of Public Health at HKU. He graduated with a PhD in medical statistics at the University of Warwick (UK) in 2003, and spent a year as a postdoc at Imperial College London (UK) before joining HKU in 2004. He has particular expertise in medical statistics including generalized linear models, survival analysis, study design, and meta-analysis.

Consultants

Location

BCRM Unit, School of Public Health,
Li Ka Shing Faculty of Medicine, The University of Hong Kong,
1/F, Patrick Manson Building (North Wing), 7 Sassoon Road,
Pokfulam, Hong Kong

Request Support

Useful Tools


  • To do a sample size calculation, you can use the online sample size calculator available at: http://www.math.uiowa.edu/~rlenth/Power/

    Screenshot of Sample size calculator, Scenario 1

    Worker example

    Scenario 1:
    "S-Syndrome (SS)" is characterized by profound irritability, disorientation and fatigue for those infected individuals. The efficacy of a vaccine (called "BG vaccine") in preventing adulthood SS remains uncertain, and a study is designed to compare the vaccination coverage rates in a group of MPH students infected with SS and a group of controls with equal sample size. Available information indicates that approximately 30% of the controls are vaccinated. The primary investigator plans to have an 80% chance of detecting an odds ratio significantly different from 1 at the 5% level of significance. If an odds ratio of 2 would be considered an important difference between the two groups, what should the sample size be included in each study group?

    Assumptions:
    Level of significance: 0.05,   Statistical power required: 0.8 

    Equation 1, for OR

    This can be rearranged as Equation 1, rearranged against p1

    Sample data for p2 and OR ?   Result p1

    Sample size calculations:
    Enter p1 = 0.462, p2 = 0.3, alpha = 0.05.
    Adjust sample size until reaching desired power.

    Sample size in each group: 152   Total sample size: 304 

    Sample size calculator results screenshot, Scenario 1

    Scenario 2
    If number of cases is limited to 100, untick "Equal ns", set n1 = 100, and increase n2 until the power reaches 80%.
    The required sample sizes are 100 cases and 293 controls to reach 80% power for OR of 2.

    Screenshot of Sample size calculator, Scenario 2

    If effect sizes smaller than OR = 2 are of interest, the sample size would be larger. Use the formula shown previously to calculate p1, based on particular values of p2 and OR.

    • Presents grouped data with rectangular bars with lengths proportional to the values that they represent
    • Can be plotted vertically or horizontally
    • Very useful for recording discrete data and show comparison
    Sample Bar chart of Population versus Weight Status Sample Bar chart of Population versus Major US cities
    • Represent the distribution of numerical data
    • Use for continuous data, where the bins represent ranges of data
    Sample Histogram of BMI
    • Display values for two variables for a set of data

      Sample Bar chart of Population versus Weight Status Sample Bar chart of Population versus Major US cities
    • Suggest various kinds of correlations between variables
    • Ability to show nonlinear relationship between variables

      Sample high positive correlated scatter plot
      High positive correlated
      Sample low positive correlated scatter plot
      Low positive correlated
      Sample negative correlated scatter plot
      Negative correlated
      Sample uncorrelated scatter plot
      Uncorrelated
      Sample non-linear correlated scatter plot
      Non-linear relationship
  • How to understand a Boxplot

    Sample Box Plot
    • represents the mean and variability of data
    • represents the overall distribution of the data
    Sample Means and Error Plot