D:/Biomedica Vol.26, Jan. – Jun. 2010/Bio-24.Doc P. 80 – 84 (WC)

A VIEW POINT

BIOSTATISTICS EDUCATION FOR UNDERGRADUATE MEDICAL STUDENTS

WAQAS SAMI Department of Biostatistics, University of Health Sciences, Lahore – Pakistan

Biostatistics is relevant to students and practitioners in medicine and health sciences and is taught as a part of the medical curriculum. The Biostatistical literacy of medical students is a problem all over the world including Pakistan which is actively discussed in different publications. Biostatistics is the subject in the medical curriculum that is not popular among the undergraduate students. The negativity of the medical students towards biostatistics is real and their perception can easily be changed through interactive teaching techniques by trained faculty. This paper describes the problems and challenges confront in teaching biostatistics to medical students. Some suggestions and guidelines are presented which may help to surmount these problems and enhance the teach- ing of biostatistics in health sciences. Keywords: Biostatistics, Medical Students, Curriculum, Computer – Based Learning.

BACKGROUND tical techniques have also become increasingly imp- Prior to the twentieth century, medical research was ortant in health sciences. primarily based on trial and empirical evidence. The medical practitioner in the 21st century will Diseases and the risk factors associated with a dise- need a far greater ability to evaluate new infor- ase were not well understood. Drugs and treatments mation than in the past. A good understanding of for diseases were generally untested. As medicine biostatistics can improve clinical thinking, decision has moved to become more evidence based, biosta- making, evaluations and medical research. tistics has become ever more important and rele- Undoubtedly, medical professionals are becom- vant to the practice of medicine and the education ing aware of the importance of leaning and applying of tomorrow’s doctors. It has also become increas- biostatistical methods in their research. This wish is ingly evident that the interpretation of much of the not constrained to medical researcher but medical research in health sciences depends to a large extent practitioners who read medical literature to keep on biostatistical methods. The applications of bio- them abreast also wish to gain a minimum know- statistics was started in the 17th century on causes of ledge in biostatistics. The role of biostatistics in death, births, marriages, construction of life tables medical education is now well recognised and the and estimating population sizes, that is now known curricula in almost all the medical institutions and as “Vital Statistics”. The field of genetics was the sec- universities across globe has provision of teaching ond epoch which was most benefited by biostatis- biostatistics to undergraduate and postgraduate tical ideas emerging in the works of Mendel, Bate- medical students / professionals. son, Darwin, Pearson and Fisher.2 In Pakistan, medical statistics is non-existent although medical research in Pakistan has gained INTRODUCTION momentum over the past several years. However, Biostatistics is a branch of applied statistics that is the logical conclusions based on information and concerned with the application of statistical metho- data are rarely witnessed. This problem is due to the ds to medicine, clinical trials, demography, popula- non-availability of Biostatistics faculty and practi- tion estimation, modelling, community diagnosis cally medical students, researchers and doctors are and surveys. In general, the purpose of using biosta- also unaware of its logic, uses and inferences to be tistics is to gather data that can be used to provide obtained. honest information about unanswered biomedical questions. Biostatistics is now considered as an ess- Teaching Biostatistics to Medical Students ential tool in planning and delivery of health care The role of biostatistics in medicine and health care systems. The knowledge and ability to use biostatis- is sometimes only fully understood and appreciated once the end – users are fully qualified in it. As alre-

Biomedica Vol. 26 (Jan. - Jun. 2010) 81 WAQAS SAMI ady stated, almost all health sciences disciplines from school to school reflecting the degree of faculty have the provision of teaching biostatistics. For a sophistication. In 1975, the American Statistical number of reasons biostatistics is one of those sub- Association in their meeting proposed the core cur- jects in the medical curriculum that is possibly un- riculum for Medical Statistics to be taught at under- popular among the medical students. graduate level which is being updated in light of the At undergraduate level there is a great diffe- usage of contemporary techniques. rence in terms of organization of teaching and time For teaching Biostatistics to medical students allotted for biostatistics instruction in medical the best time to start teaching is in 1st year of educa- schools / colleges globally. In United States, Cana- tion. A course in Biostatistics can only be helpful da, Australia and United Kingdom biostatistics is and valuable for the student if Biostatistical con- being taught in 1st and 2nd year of study with a very cepts and applications are reinforced throughout small percentage of schools spreading this instruc- his / her 5 years of education. tion over more than one term. While in Pakistan it There are numerous Biostatistical procedures is taught in 4th year of education as a component of and techniques which have proven useful, they are Community Medicine subject and the contents are widely used in biomedical sciences and majority of covered in just 3 – 5 lectures. Besides, for admis- these techniques are not a part of the Biostatistics sion in the medical schools of Saudi Arabia, apart curriculum in Pakistan. from passing the entrance test the students must In May 2002 the National Curriculum have also completed a pre-medical course in bio- Revision Committee on MBBS in its meeting held at statistics. Univer-sity Grant Commission, Islamabad revised In Pakistan at undergraduate stage, biostatistics the cur-riculum after due consideration of the courses are taught in a manner that are generally comments and suggestions from Universities and short and covers only introducing the concepts of Colleges across Pakistan. A comparison of Descriptive Statistics and a very little part of Infe- Biostatistics cur-riculum is given below: rential Statistics which does not help a medical stu- dent to groom himself for decision making. Globally the Biostatistics curriculum varies

Table 1: Curriculum of Biostatistics at Undergraduate level.

Curriculum of foreign medical Innovations in Pakistan’s Curriculum in Pakistan schools curriculum Definitions and Terms Concepts and Usage Experimental Design (ANOVA – Ways of Data collection Data and its types Descriptive Statistics Descriptive Statistics one and two way) • Central Tendency (Mean, • Central Tendency (Mean, Median, Mode) Median, Mode Regression Analysis • Measures of Dispersion • Measures of Dispersion Correlation Analysis • Graphical Presentation • Graphical Presentation Inferential Statistics Inferential Statistics Logistic Regression Analysis • Z – tests • t – tests t – tests Probability (Bayesian Statistics) • Anova Techniques Relevant Non-parametric tests • Multiple comparison tests Diagnostic Statistics • Chi – Square Chi – Square (Introduction) Discrete and Continuous (Sensitivity, Specificity, PPV. Distributions Probability NPV, ROC Curve etc) Non-parametric tests

Correlation Introduction to Survival Analysis Regression

Biomedica Vol. 26 (Jan. - Jun. 2010) BIOSTATISTICS EDUCATION FOR UNDERGRADUATE MEDICAL STUDENTS 82

Logistic Regression Analysis Sampling Survey Introduction to Sampling Survey

Vital Statistics Vital Statistics Introduction to Statistical

Packages

Biomedica Vol. 26 (Jan. - Jun. 2010) 83 WAQAS SAMI

Problems in Teaching Biostatistics • The teacher must convince the medical students The worldwide picture of biostatistical education in about the importance of learning the subject of medical schools is far better than that of what is biostatistics. The course should be aimed and being taught in Pakistan’s medical colleges. Now-a- taught in such a way that it increases student’s days medical students should be able to understand motivation towards it. One of the best ways of and interpret biostatistics so that they can use these motivating the students is to expose them to techniques both during training and most importa- medical literature with the examples of uses and ntly at postgraduate stage when they will be treating abuses of biostatistics. This could be accomp- patients. Some of the problems encountered in tea- lished most readily by preparing some interes- ching biostatistics to medical students are listed ting classroom sessions to provide students an below: opportunity to critique the reporting of biosta- • In Pakistan, a student is enrolled in a medical tistical procedures reported in the journals. college after completing 12 years of education. • The instruction method should be problem – At the pre-medical stage the curriculum has no oriented instead of technique – oriented. The provision of teaching statistical courses. When technique – oriented method is now an old style the students enter into medical colleges they are which contains hand calculations, formulas, introduced to the general concepts of medical drawing critical regions, traditional handouts statistics and even the introductory concepts fail and formal presentations which make the sub- to be understood because lack of prior know- ject learning boring for the students. They just ledge of statistics. This impacts negativity on take the subject as they have to sit for 40 minu- the importance of biostatistics. tes with eyes open in the class and mind resting. • Medical students are generally focused towards • Besides, problem – based approach should be the study of medicine and towards basic scien- introduced as this is a way of constructing and ces and have a very little desire of learning bio- teaching using problems as the stimulus and statistics. focus on student’s activity instead of presenting • The time allotted for Biostatistical teaching is the material through a traditional lecture for- not enough and thus prohibits the explanation mat. Problem based learning requires that stu- of most of the important techniques like Logis- dents take an active part in their own learning, tic regression analysis. Diagnostic Statistics and it is also best suited for motivated students who Survival Analysis etc. have the desire for own learning. In addition the • Instructors of biostatistics have varied backgro- students can emphasize their learning on the und and thus presenting the subject in their concepts and interpretations rather than the own understanding makes it difficult for the mathematical details according to their per- student to squeeze out the interpretation from sonal strength. the results. In Pakistan, main problem in tea- • Biostatistics is a branch of applied statistics and ching biostatistics is the unavailability of trai- it must be taught in terms of application in bio- ned biostatistics Faculty. Even with the promot- medical research. The students should be expo- ing research culture, the data generated is in a sed to real – life data instead of using the text- paralysed state with no one to analyse it. book examples which generally start from the • The formal teaching in biostatistics neither eng- statement “suppose”. ages the students nor meets their needs. A sur- • Medical Students come from different educa- vey conducted on biostatistics teaching20 sho- tional backgrounds; some have interest in num- wed that the students disliked the subject beca- erology which most of them dislike the biosta- use it was taught in a formal way (formulas and tistics as a subject. Thus any emphasis on the calculation). Lot of contents were unimportant. statistical proofs and probabilistic reasoning The lectures were poorly presented and it was should be discouraged. On the other hand if the difficult for them to know what they needed to instructor has no knowledge of teaching applied know. statistics then the whole course will be a roller- coaster ride for the students. A Pragmatic Approach for Teaching Biostatistics • Hand calculations should be avoided and for In this section, some guidelines and tips are given making the teaching interactive Computer – for biostatistics teaching which may help to over- based approach should be introduced. Nowa- come some of the problems encountered and enha- days a variety of Biostatistical software’s (SPSS, nce the learning in this subject. STATA, STATISTICA, NCSS and OPEN – EPI)

Biomedica Vol. 26 (Jan. - Jun. 2010) BIOSTATISTICS EDUCATION FOR UNDERGRADUATE MEDICAL STUDENTS 84

are available. Introducing CBL will also enhance is heavily dependent on the usage of accurate the student’s motivation for learning. Use of technique(s), for example comparing means computers should be encouraged in teaching to always address the differences, Chi-square tests allow the student to concentrate on the inter- addresses the associations. This core understan- pretation of the analysis rather than on hand ding is essential to be passed on to the students. calculations. On the other hand most major • The students should also be made aware of research projects involve a tremendous invest- using common words mistakes like association, ment in time and money. This results in a large correlation and differences; all are statistical data that needs to be analysed and the tasks can terms and should not be combined with english be easily achieved by using computers. language meanings. Moreover, the instructor • For designing of research reading medical lite- teaching biostatistics should strive to present a rature is a pre-requisite. A student cannot read well-balanced combination of lectures, tutorials and understand medical literature without hav- and practicals. ing knowledge of biostatistics and research In conclusion, there is a substantial disagre- methodology. A study performed by Altman ement in the course contents of the biostatistical 2000) reports that “statistical errors are so topics being offered in the medical curricula of Paki- common that 50% of the medical literature have stan as compared to other countries. The teaching statistical flaws. Another study carried out sessions should be enhanced to cover the major (McGuigan SM, 1995) reported that ‘serious sta- topics. The curriculum contents should also be up- tistical errors were found in 40% of 164 articles dated after every 3 years in the light of latest advan- published in a psychiatry journal and in 19% of cements both in the subject of biostatistics and 145 articles published in an obstetrics and gyne- Health Sciences. biostatistics and research metho- cology journal. dology should be taught as a continuum, with its • The instructor should educate the students abo- relevance to thinking about health and disease. In ut the methodology in designing the research all medical colleges, there should be a separate dedi- (research design, sampling technique / data col- cated Department of Biostatistics that can teach and lection and sample size etc) including the repor- guide the students in all phases of conducting rese- ting and reading of biostatistical language in a arch. Students should be taught from a common, research paper because nowadays, without pro- same standard, up to date, and self explanatory per reporting of biostatistics, the research arti- textbook in order to ensure consistency and avoid cle is rejected at the initial review. Furthermore, conflicting terminology. The negativity of our medi- the teaching of biostatistics should not be limi- cal students towards biostatistics is real and their ted to the use of data analysis techniques only. perception can easily be changed through interac- It is important that the students should be tive teaching techniques by trained faculty. One educated about the data collection techniques should remember that they want to become doctors because all the biostatistical techniques are ap- and not biostatisticians. It should be the primary plied on the data collected. goal of the biostatistics teacher to offset the fear of biostatistics among medical students. • The important steps in teaching biostatistics are analysis and interpretation of data. It is the ACKNOWLEDGEMENT primary responsibility of the instructor to make I am grateful to the Vice Chancellor of University of the students understand about using accurate Health Sciences for providing me the facility to data analysing technique(s) under given condi- complete this article. I am also obliged to the medi- tion(s). The inappropriate use of biostatistical cal students who gave me a lot of relevant informa- methods and techniques may mislead the stu- tion regarding the subject. dents understanding at a stage when their min- ds are fresh to gather knowledge. The instructor REFERENCES should address the questions of Why, Where 1. Altman DG, Bland JM. Improving doctors. and How while teaching the techniques. Inap- Under-standing of statistics. J R Stat Soc Ser A 1991; propriate use of biostatistical techniques can be 154 (2): 223-67. found in every stage of medical research related 2. A. G. Cock. William Bateson, Mendelism and to data analysis, design of the research, data bio-metry. Journal of the history of Biology. 1973: 6 collection and compilation, analysis, implemen- (1): 1-36. tation and interpretation. 3. McGuigan SM. The use of statistics in the British • The second area at which the instructor should Journal of Psychiatry. Br J Psychiatry 1995; 167: 683- focus on is the interpretation of the data which 8.

Biomedica Vol. 26 (Jan. - Jun. 2010) 85 WAQAS SAMI

4. Welch GE II, Gabbe SG. Review of statistics 16. Hillmer, S.C. “A Problem-Solving Approach to usage in the American Journal of Obstetrics and Tea-ching Business Statistics,” The American Gynecology. Am J Obstet Gynecol 1996; 175 (5): Statistician. 1996; 50: 249-256. 1138-41. 17. Report of a Committee of the American 5. Bland, J. M. Teaching statistics to medical Statistical Association Subsection on Teaching of students using problem-based learning: The Statistics in Health Sciences, Proposal for a core Australian expe-rience. BMC Medical Education, curriculum in medical biostatistics. Clinical 2004; 4 (31): 1472-69. Pharmacology and Therapeutics 1975; 18: 127-131. 6. Moussa, M. A. A. Developments in the 18. A.R. Feinstein, Clinical biostatistics, XXXIII. On instruction of Biostatistics at the Kuwait University tea-ching statistics to medical students. Clinical Health Science Centre in a decade. Teaching and Pharma-cology and Therapeutics 1975; 18: 121-126. Learning in Medi-cine, 2002; 14 (3): 194-198. 19. Chiang CL. What is biostatistics? (With 7. Sahai, H. and Ojeda, M. M. Teaching discussion by M. Zelen). Biometrics 1985; 41: 771– biostatistics to medical students and professionals: 775. Problems and solutions. International Journal of 20. Symposium on the teaching of statistics to Mathematical Education in Science and Technology. under-graduate medical students in Europe. 1999; 30 (2): 187-196. Copenhagen, Denmark: World Health Organization 8. Korea Biometric Society. Application of statistics regional office for Europe 1982. in medicine and public health workshop. Seoul: 21. Paul Mostert Stellenbosch University, South Korea Biometric Society. 1994. Africa, ICOTS-7, 2006. 9. General Medical Council: Tomorrow's Doctors: 22. Tae Rim Lee. Teaching Biostatistics to medical reco-mmendations on undergraduate medical per-sonnel with computer based supplement. education. 1993. Interna-tional Association for Statistical Education 10. Campbell MJ: Statistical training for doctors in and Inter-national Statistical Institute. 2001, Printed the U.K. In Sixth International Conference on in Grana-da, Spain. Teaching Statistics Cape Town, South Africa; 2002. 23. Thomas J, Melanie A. Using a problem-based 11. Garfield J: How students learn statistics. appro-ach to teach statistics to postgraduate science Interna-tional Statistical Review 1995; 63: 25-34. stu-dents: A case study. MSOR connections vol 9 no 12. Phillips, B., Francis, G., and Hutcheson, M. 2 May – July 2009. Statistics for workers in social and health sciences by 24. Jenny V Freeman, Steve Collier, David flexible learning modes 1998. Staniforth and Kevin J Smith. Innovations in 13. L. Pereira-Mendoza, L. Seu Kea, T. Wee Kee, & curriculum design: A multi-disciplinary approach to W. K. Wong (Eds.), Proceedings of the Fifth teaching statistics to undergraduate medical students International Conference on Teaching Statistics (pp. BMC Medical Education 2008, 8: 28 doi: 401-408). Sin-gapore: International Association for 10.1186/1472-6920-8-28. Statistical Edu-cation and International Statistics 25. Y.M. Lam. Teaching Statistics to Medical Institute. Students in the New Medical School. Journal of the 14. Marvin Zelen. Biostatisticians, Biostatistical Hong Kong Medical Association. 1986; Vol. 38: No. Science and the future. Statist. Med. 2006: 3409- 4. 3414. 26. Hans C. Van Houwelingen. The future of 15. Boyle, C.R. “A Problem-Based Learning biostatis-tics: expecting the unexpected. Statistics in Approach to Teaching Biostatistics,” Journal of medicine, vol. 16, 2773-2784, 1997. Statistics Educa-tion, (1999) 7, secure/v7n1/boyle.cfm.

Biomedica Vol. 26 (Jan. - Jun. 2010)