Research has become an integral component of the academic missions of advanced and successful institutions of higher education in the world. University of Veterinary and Animal Sciences in Lahore, Pakistan (UVAS) with its new mandate as an independent institution of higher learning needs a strong research-infrastructure to meet the challenges of 21st century in veterinary and biomedical sciences. Applied statistics is the underpinning of such efforts through planned research, efficient experimental designs, and correct data analyses and validation of results. Descriptive and inferential statistical methodologies are employed in every phase of scientific endeavors to better understand the implications of the proposed research. Therefore, it is extremely critical that an experiment or survey is well thought out and designed in terms of its intended hypothesis. Sample size should be representative of the population in question and efficient in resource allocation using blocking or other techniques. Data should be analyzed correctly so that the results generated are truly reflective of the experimental treatments without systemic bias. Results should be interpreted and extrapolated carefully keeping the underlying assumptions and constraints in perspective. Collectively these measures will generate high quality research and lead to new discoveries. Therefore, biostatistical support is urgently needed in all aspects of scientific investigations undertaken by the veterinary and biomedical research faculty and students at UVAS and other collaborating institutes in Pakistan.
Such support will not only improve the quality of research and education in UVAS but also in the long run will produce a sustainable human resource in higher education and research in Pakistan.
The proposed biostatistical support will include:
• Planning and designing experiments in resourceful, efficient and statistically valid manners. Initial steps of research planning are critical as they lay foundation for research outcome. Oftentimes, investigators seek statistical help when the data have already been generated. Maneuvering data to fit into a design will generate results that may not necessarily be reflective of the underlying causes. It introduces researcher’s bias and violates several statistical assumptions leading to invalid conclusions. The proposed Biostatistical Consulting Center (BCC) will advise and train researchers during the initiation process of their research project regarding the crucial steps of experiments rather having to wait until the data is generated, which, most often, is too late to fix the anomalies.
• Formulating and stating the research hypotheses. Time and again in research it is not the wrong answer that is of concern, it is rather the wrong question that is being probed. Asking the right question(s) in experimental trials will ultimately lead to right answers. The proposed BCC will help research faculty and graduate students in formulating and stating right questions to ultimately generate right answers.
• Specifying a statistical model. There are various statistical models available for different experimental designs and data analysis. Choosing the right model is critical in addressing various hypotheses. For example, considering several variables at the same time and analyzing them using factorial analysis of variance can save resources, improve experiment efficiency and answer more questions at the same time.
• Selecting the efficient sample size to obtain the desired significance level. Sample size is often a contentious issue in empirical research and survey design. Small sample size may produce results with low significance level and may not represent the population in question. Large sample size, on the other hand, in human or animal research, may not be available or cost prohibitive. Finding an appropriate mix is the key to save time, efforts and resources. BCC will guide researchers in choosing the right mix of such samples.
• Helping in randomization process. Actual randomization process must not be overlooked. It should not be done in the statistician’s office but rather be actually carried out in the lab or field. The BCC will help train investigators in this process. • Helping in data management. Large experiments and surveys generate overwhelming amount of data sets. These data should be properly organized with right identifiers, portioned into manageable variables and may be divided into subsets for various analyses.
• Helping in statistical programming. Some statistical software, such as SAS/STAT requires programming to invoke various procedures from its library functions. Most often these commands are not intuitive. The BCC will organize workshops and seminars for the faculty and graduate students to understand such basic statistical programming.
• Helping in data analysis. Data analysis is the most postponed issue in many research projects as it often becomes a bottleneck. The BCC, through its early intervention in research projects, will help generate right data and subsequently guide investigators in data analyses by offering advice on the selection of specific methodologies and appropriate use of software for data analyses.
• Modifying statistical methodologies where assumptions are violated. Most often experiments and data analysis are conducted based upon certain underlying assumptions and sampling distributions. Care must be taken to validate these assumptions and if they are violated, alternative methodologies should be adopted.
• Providing assistance in interpretation of results. Once results are generated in a given research, it is not instinctive to understand their exact meanings and relation to stated hypothesis. Should the results be completely ignored if they do not reach to a p value of less than .05? Investigators grapple with these issues. Discoveries are often hidden in the results but are hardly recognized. The BCC will guide researchers to better understand statistical analysis results and their interpretations.
• Helping in presentations and publications. Presentations and publications are among the ultimate products of successful research projects. The BCC will steer investigators in such efforts to explain results carefully and succinctly to the scientific community. Publication in a peer reviewed scientific journals is a long hard battle that needs several iterations of revisions and editing before it is finally accepted. The BCC will assist researchers in this process to understand the reviewer’s critique on experimental design and data analysis and other related issues for successful publications.
Hence, the overarching goals and functions of the proposed BCC is to provide consulting and training in all aspects of biostatistics using variety of onsite and distance learning methodologies, to faculty, researchers and students in UVAS and other collaborating institutions in Pakistan.
Problem Statement
Applied statistical methodologies and data analysis techniques are critical under-girding factors in research and knowledge advancement. Appropriate experimental design lays foundation of sound research plans; sample size determines the power of significance and efficient utilization of scarce resources; and adequate data analysis generate results that are valid and true representative of the population in question. These are noble causes in scientific research, yet not intuitive. University of Veterinary and Animal Science, with its recent up-gradation and introduction of new disciplines involving graduate level research actively needs statistical infrastructure in order to conduct high impact research. Such statistical support, in the form of a consulting center will effectively provide the necessary know-how and expertise in applied experimental research and survey analyses, particularly in veterinary sciences and public health. The proposed BCC will be a significant step forward in achieving those objectives as it plans to develop human resources through variety of training and education methodologies, such as short courses, workshops, seminars, online courses and consulting on specific research issues.
Prior Experience / Capability
The principal investigator-US, Dr. Anwar Ahmad, is faculty manager of Jackson State University- Biostatistical Support Unit (BSU). The BSU was established in 2003 under NIH grant G12RR013459. Since its establishment, the BSU has been providing statistical services and collaborative research support to investigators, faculty, graduate students, and undergraduate students in the College of Science, Engineering and Technology (CSET) and Jackson State University (JSU).
The Unit has been organized to ensure better services to all investigators conducting research at JSU, particularly to those involved in biomedical and environmental sciences. Therefore, the BSU serves as a centralized resource for statistical expertise and data management services for current and future biomedical scientists at JSU.
The BSU has significantly contributed to the success of the Research Centers in Minority Institutions (RCMI) program at JSU, by providing RCMI investigators with capabilities in biostatistics, applied environmental statistics, statistical computing, and database and information management. Its impact in strengthening the research capabilities of CSET is evident from the significant number of research projects the Unit has advised, from inception to completion, and the number of research faculty and their graduate students it has mentored in their theses and dissertations during its present and previous funding cycles. Because of its critical nature, the BSU has been recently awarded a renewal grant of $ 808,539.00 for the period 2008-13 by the National Institutes of Health. It plans to strategically align itself with the other existing and new research groups and centers at JSU and other institutions, such as UVAS to help conduct high impact basic and applied research, particularly in the area of biomedical sciences and translational research.
Dr. Ahmad, PI-US, besides providing consulting services to faculty, RCMI investigators, researchers and graduate students, teach undergraduate and graduate courses in Biostatistics and is also actively involved in multidisciplinary and multifunctional collaborative research nationally and internationally with over 60 publications and abstracts to his credit. Some of his research areas include:
* Biostatistics Education and Consulting
* Neural Network Modeling of Physiological Variables
* Microbial Risk Assessmt
* Poultry Growth and Nutrition Modeling With this rich experience and infrastructure, the BSU in general and the Principal Investigator, Dr. Ahmad in particular, are ideally placed to advance knowledge and share expertise with UVAS in developing the proposed BCC and help building a strong research infrastructure in Pakistan.
|