Selecting a research question
- 1. Abhaya Indrayan MSc,MS,PhD(OhioState),FAMS,FRSS,FASc Biostatistics Consultant, Max Healthcare Selecting a Research Topic
- 3. Fascinating World of Medical Research Despite enormous advances, the unknown segment is far more than known segment Do we have knowledge, skills, procedures, equipment for instant diagnosis of health problems? Do we have magic treatments that alleviate the sufferings immediately? Can we replace irreparable body parts that work as effectively as the originals? Can we accurately predict the outcome - when, how much residues will remain? Medical research very different from research in physics, chemistry, history, literature because health is vital to life
- 4. What is Research? RELENTLESS SEARCH FOR TRUTH (“Truth” changes - Ptolemy) Discovery of new facts (never thought of, in the domain of conjecture, or being looked at with suspicion) Enunciation of a new principle – wider ramifications (if two drugs work the same way, can we use one on the basis of established efficacy of the other) Fresh interpretation of the existing ideas (repurposing the existing drugs for COVID - plasma therapy) Research opens up new questions and it is endless search
- 5. Further on Research Why and how of a phenomenon (why Coronavirus does not affect children as much - how much is the difference in the incidence) Also, what, when, how much (what biochemical changes occur by eating citrus fruits, how much and for how long should be eaten for a significant change, how much change occurs) What the world will know that was not known earlier (key messages) – how does it add to the present knowledge For a PhD thesis, the topic can be confirmation (or denial) of the previous finding in a new setup (effect of celecoxib in osteoarthritis in old women in India) Detailed study of a phenomenon also sometimes comes in the ambit of research - low birth weight (why, how, where, how much, who, how to control, etc.)
- 6. Levels of Medical Research First level - PG thesis (proposing a hypothesis) - Use resources of the guides, generally two years, part-time - training in research methodology Second level - Doctoral dissertation (frank research as explained earlier), at least three years full time, must be publishable Third level - Institutional research, generally collaborative among the departments or among the faculty, may be externally funded - publishable in a good journal Fourth level - Huge multi-centric studies, requiring enormous resources, supported by scientific bodies such as ICMR, DBT in India
- 7. Scientific Methods of Reasoning (All sciences – not restricted to medical) 1. Induction – From a collection of specifics to a general statement – 95% of diabetics of age 60 seen in a clinic have survived for at least five years – thus P(5-year survival) = 0.95 – Used in most medical research by observing a sample of patients 2. Deduction – From general to specific – Positive predictivity of a PCR-CoViD test is 0.97 – thus, test +ve person has CoViD – Used in clinics for diagnosis 3. Abduction – Hypothetical best explanation of a phenomenon then test it – A patients gets repeated fracture – may have low BMD – By chance (doing something else but find another – – Alexander Flaming’s penicillin (Contaminated mold)
- 8. What is a Problem for Research? -1 GAP IN KNOWLEDGE OR DIFFICULT TO USE Unknown magnitude of the problem - unknown profile Perceived difficulty (premature delivery in anemic women) Feeling of discomfort about the way the things are (frequent monitoring of low birth weight babies) Presence of discrepancy (deterioration despite treatment) Two or more plausible answers (exercise or diet for weight control) Better (easy to administer, economical, less discomfort, etc.) way of managing patients – medical vs. surgery Propounding a new idea is also research provided it is rooted in science (the idea of positive health)
- 9. How to Find Problems for Research? Your own practice - encountering patients (management of a patient with long-standing pain in abdomen) Talk to your colleagues or guides Search literature - Choose reputed journals - Last paragraph may indicate what more is needed to be done - See if the results can be extended to other kind of patients or other areas - Whether new procedure can be used for the problem investigated earlier - Will the findings apply to your setup - what modifications can be considered Research into researches
- 10. CHARACERISTICS OF A GOOD RESEARCH QUESTION Timely – Cancer - Monkey pox Original – Not much discussed earlier – anti-ageing diet Novel – Biological markers of positive health Justified – People tried earlier but failed - medical therapy of breast cancer Anchored with prior evidence – sleep for health Focused – Which group (age, severity, hospitalized, low nutrition) Specific – Diet and cancer (contribution of diet to different types of cancer or its prevention)
- 11. What is a Good Problem for Research? Specify what the world does not know and you would like the world to know Merely filling up a knowledge gap is not enough - must have application to the real world now or in the future (speed of growth of nails) - DNA structure, genome sequence Must have a theoretical backup (biological plausibility - diet and heart diseases, mRNA-based vaccine) If you want to do research on age distribution of deaths by cancer, why do you want to do this research, how will it help the public or the science - what the world will gain The topic currently in discussion – positive health
- 12. Hot Topics for Medical Research BROAD AREAS – Fine tune as needed Prevention of chronic diseases (cancer, CVD, COPD, CKD) - diet, lifestyle, supplements, stress and anxiety Cancer cure (back to original health) Living healthy all through the life (without disease) Geriatric disorders - degenerative diseases Early signs of various diseases Diet and diseases Treatment of cognitive impairment
- 13. CHOICE DICTATED BY - 1 Researchability – Tractable – Vaccine for obesity Feasible – Sufficient resources available (expertise, time, equipment, enough cases, and facilities such as lab and radiological investigations) • How genes affect health • Phase I clinical trial – 24-hour monitoring, handling emergencies Capability – Interest, caliber, skill (Do you know enough about genomics)
- 14. CHOICE DICTATED BY – 2 Rational – Theoretical back-up Reasons available or can be speculated – biological plausibility (may emerge later) – impact of highly processed food in children Relevant – Useful – how will it contribute to improved health – applicability Ethical – No harm to anybody (no serious effects anticipated, patient’s time, money, convenience, pain, etc.)
- 15. TYPES OF RESEARCH QUESTIONS – 1 Questioning the existing practice – Dialysis in certain complicated cases Verification or refutation of a recent result - vitamin K2 linked to hyper-glycemia - epilepsy drugs causing birth defects Extension to a new group or emerging diseases - plasma therapy for CoViD Modification – chemotherapy for certain cancer
- 16. TYPES OF RESEARCH QUESTIONS – 2 Descriptive – Incidence, course and outcomes in septic shock in ICU cases Analytical – Cause-effect (childhood weight gain for adult disease) – Barker’s hypothesis – Efficacy of a treatment modality (clinical trials) Diagnostic – Biomarkers for diagnosis of liver diseases Prognostic – Prediction of outcome after cancer surgery Models – Development of scoring systems Statistical models for diagnosis and prognosis
- 17. R H C R A E S E Possible Research Topics (Universal) Knowledge limitation (Gaps in the knowledge) Unknown magnitude of the problem (patients requiring dialysis) Varying profile of cases (liver diseases) Difficulty in managing a patient (non-responders) Discomfort about the way the things are (lack of staff/facilities) Presence of discrepancy (high BP feeling OK) Two or more answers to the same question (medical/surgery) New (modified) method New drug/regimen/procedure/management New surgery New types of cases Existing method – validation in local conditions Doubt about its efficacy Different age-group Different population (obese, anaemic, diabetic, pregnant, etc.) Do not re-invent the wheel
- 18. Refine Improve the focus of your problem by defining the subjects, condition, population Decide depth (how much detail you would like) - correlates Prepare broad and specific objectives (measurable)