Research questions and design


Published on

Published in: Technology, Business
1 Like
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Research questions and design

  1. 1. Developing an Idea for a Study & Formulation of a Research Question By: Mahmoud Mahmoud, MD, PhD Prof. of Clinical Pharmacology, JFOM, KSA Prof. of Clinical Research, AUHS, CA
  2. 2. Research question is the uncertainty in the population that investigator wants to resolve How Why Where When How much Is it So What?
  3. 3. Anatomy and Physiology of Clinical Research • Anatomy: what research is made of? • Physiology: How it works ?
  4. 4. Anatomy of Research • Research questions (primary & secondary) • Significance (Rationale) • Design: a- Observational • 1- Cohort • 2- Cross sectional • 3- Case Control ( Descriptive & Analytical) (Prospective & Retrospective)
  5. 5. b- Experimental (Interventional) • 1- Randomization • 2- Blindness: Single blind, Double blind • Subjects: 1- Selections 2- Criteria • Variables Measurements: 1- Predictable 2- Confounding 3- Outcome
  6. 6. • Statistical issues: • 1- hypothesis • 2- sample size • 3- analytic approach
  7. 7. Physiology of Research • To study the universe through people and phenomena • Addressing the correct questions • Implementing the study (carry out) • Error of Research:
  8. 8. The scientific method The scientific method is the best way to solve truth from lies and delusion. 1. Observe some aspect of the universe. 2. Invent a tentative description, called a hypothesis, that is consistent with what you have observed. 3. Use the hypothesis to make predictions. 4. Test those predictions by experiments or further observations and modify the hypothesis in the light of your results. 5. Repeat steps 3 and 4 until there are no discrepancies between theory
  9. 9. Incredible Research Questions (Silly!!!) • Who has more bacterial flora in the colon, men or women? • What are the facial characteristics of faithful women ? • Optimizing the sensory characteristics and acceptance of canned cat food: use of a human taste panel. (Journal of Animal Physiology and Animal Nutrition) • Effects of cocaine on honeybee dance behavior. (Journal of Experimental Biology) • Swearing as a response to pain. (NeuroReport) • Pigeons can discriminate "good" and "bad" paintings by children. (Animal Cognition)
  10. 10. How to develop a RQ ? RQ develops from previous findings by investigator or others. – Scholarship: Mastering literatures can serve as source background – Mentors: are needed particularly for inexperience researcher.
  11. 11. Primary and Secondary questions • Many studies have more than one research question. • Primary Question: – Experiments often address the effect of intervention on several outcomes • Secondary Questions: – The advantages of having several questions: Several answers can emerge from single study – The disadvantages of having several questions: This will increase the complexity of the design and implementation – More sophisticated statistical techniques are needed
  12. 12. Being alert to new ideas and techniques • Attend conferences • A skeptical attitude – Not to believe what you see but try to find better criteria • New technology will be helpful for new insight and questions • Teaching is an excellent source of inspiration
  13. 13. Criteria of Good Research Questions • F… I…N…E….R • • • • • Feasible Interesting Novel Ethical Relevant
  14. 14. Feasible – To know the practical limits and problems of studying a questions • Adequate number of subjects, pilot survey is helpful • Adequate technical expertise for recruitment measuring variables, analysis, putting questionnaire.. Etc. Using consultants or coinvestigators can shape the research • Affordable in time and money, this has to be known and evaluated early • Manageable scope, big scope may not be feasible. Investigator need to narrow it
  15. 15. Interesting – To the investigator • Share this interest with consultant or outside expert before writing grant. – To the funding Agent – To the company (pharmaceutical co)
  16. 16. Novel – The aim is to find new information – Reiteration of previous research is not acceptable. – Novelty can be determined though literature review or funding agent records • Confirm or refutes previous finding • Extends previous findings • Provide new findings
  17. 17. Ethical – Good research must be ethically accepted. – Unacceptable risk or invasion of privacy is not accepted – Institution Review Board (IRB) is essential for consultation and approval
  18. 18. Relevant – Imagine the various outcome and the possibility that the outcome will add to: • To scientific knowledge • To clinical and health policy • To future research directions
  19. 19. Improving Research Question • An outline (1-2 pages) will be helpful for the investigator to clarify the ideas about the plan. • White paper or letter of understanding • The outline will provide basis for colleagues to react to with specific suggestions
  20. 20. Problems and Solutions Potential Problem Solutions A: The research question is not FINER 1. Not feasible: * Too broad - Specify a smaller set of variables - Narrow the question * Not enough subjects available - Expand the inclusive criteria - Eliminate or modify exclusion criteria - Add other sources of subjects - Lengthen the time frame for entry into study - Use strategies to decrease sample size * Methods beyond the skills of the investigator -Collaborate with colleagues who have the skills - consult experts and review the literature for alternative methods -Learn the skills
  21. 21. * Too expensive -Consider less cost study with * fewer subjects and measurements * less expensive measurements * Fewer follow up visits 2- Not interesting, novel or relevant -Consult a mentor - Modify research question 3- Uncertain ethical suitability -Consult with IRB - Modify the research question B- The study plan is vague Ambiguity -Write the research question at an early stage - Get specific in the one to two pages study plan . How subjects will be sampled . How variables will be measured
  22. 22. Translational Research • Translation of findings from Ivory tower to reality – T1: from lab to clinical studies (from Bench to bedside) – T2: from clinical studies to health practice (need process and collaborators) Lab Research T1 Clinical Studies Population research T2
  23. 23. Hypothesis • Definition: – It is a precise testable statement of what the researchers predict, it will be the outcome of the study. • It is transformation of research questions into final and most specific version to establish the basis for tests of statistical significance. • It summarizes the sample, design, the predictor and outcome variables • This usually involves proposing a possible relationship between two variables: the independent variable (what the researcher changes) and the dependant variable (what the research measures).
  24. 24. Characteristics of a good hypothesis • Simple versus complex: – One predictor and one outcome variable • Ex: Sedentary life style is associated with diabetes – Combined predictor or outcome variable can be useful • Specific versus Vague – To clarify subjects and the variables • In advance versus after–the- fact – Hypothesis needs to be state before the study not after reviewing the data. – The after the fact leads to over interpretation
  25. 25. Types of Hypothesis • Null and alternative hypotheses Null H (H0) There is no association between predictors and outcome Alternative H (Ha): There is association between predictors and outcome • One and two sided alternative hypotheses i.e. One direction or 2 directions
  26. 26. • If your prediction was correct, then you would (usually) reject the null hypothesis and accept the alternative. •If your original prediction was not supported in the data, then you will accept the null hypothesis and reject the alternative. •The logic of hypothesis testing is based on these two basic principles:
  27. 27. • In research, there is a convention that the hypothesis is written in two forms: – the null hypothesis, or Hn, H0 and – the alternative hypothesis (called the experimental hypothesis when the method of investigation is an experiment), or Ha or He (either shorthand version is acceptable). • Briefly, the hypotheses can be expressed in the following ways: – The Hn states that there is no relationship between the two variables being studied (one variable does not affect the other). – The Ha states that there is a relationship between the two variables being studied (one variable has an effect on the other).
  28. 28. • A research hypothesis (Ha) prediction is one that states that results are not due to chance and that they are significant in terms of supporting the idea being investigated. • A 'null hypothesis' (Hn) prediction is one that states results are due to chance and are not significant in terms of supporting the idea being investigated.
  29. 29. Types of Results • The problem if results in the samples do not reflect the population – Positive – False positive (Type I)…… Need to be zero • Alpha (reject the null hypothesis when it is actually true, level of statistical significance) • Alpha of 0.05 or less – False negative (Type II)….Need to be zero • Beta (failing to reject the null hypothesis, when it is actually false) • Power (1-beta) – Negative
  30. 30. Type 1 Error: False Positive • Claiming that two means are not equal when in fact they are equal. In other words, you • reject a null hypothesis when it is TRUE. Type 2 Error: False Negative • Not finding a difference between two means when in fact there is a difference. In other words, you reject a null hypothesis when it is FALSE.
  31. 31. • Avoiding of errors entirely is difficult but can be minimized by: – Increase sample size – Design – Different measurements
  32. 32. • Strategy: • 1- Put a scale • 2- Precision (reproducible) of observer, instruments and subject………. Repetition • 3- Accuracy (confidence to measure reality)……. Using gold standard
  33. 33. Designing Research
  34. 34. I- Designing an Observational Study 1- Cohort Study • Powerful study • Types: Descriptive, analytic, prospective and retrospective • Prospective Study: structure, strength, weakness • Retrospective study: structure, strength, weakness
  35. 35. • Nested Case-control and case-cohort studies: structure, strength, weakness • Multiple-cohort studies and external controls: structure, strength, weakness • PLANNING A COHORT STUDY
  36. 36. 1-Cohort Study Determine Incidence of Disease Prospective Retrospective Nested Case-control & Case cohort Multiple-cohort & External control Structure Define samples And follow variable over Time for future outcome Define samples and follow over time for outcome in the PAST Nested case Control: is cohort study (P or R) then evaluate outcome occurrence Vs the control Case-Cohort: Similar to NCC except no control cases are selected but random samples regardless of the outcome. Measurements can be used for the whole study samples Different samples with different risk exposure. In occupational and environmental study. Registry and census are useful Strength -Powerful -accurate measurements of variables in known outcome -Powerful - Unbiased - Less costly -less time -Useful for costly measurements -Unbiased -For rare exposure and potential hazards Weakness -Expensive -Inefficient in rare outcome -can’t evaluate pre-post disease -Limited control -May not have information to answer the questions - As cohort study -Baseline is affected by silent preclinical diseases -Storage of samples -Retrospective study lacks the recorded information
  37. 37. 2- Cross Sectional Study 3- Case Control Study Structure - Select sample of population - Measure Predictor and outcome variable All measurements are made at once. - Select sample of population with disease - Select sample at risk free of disease - Measure predictor variables Generally Retrospective Strength - Major source of health and habits - Epidemiologically identify risk information factors - Hypothesis determine cause and effect - Find strength between associated factors - Determine prevalence of - Efficient for rare disease disease. - Useful to generate hypothesis - No waiting for outcome to occur (fast, inexpensive) - Examine network of causal links - Serial surveys used to determine the changes observed at several times Weakness - Difficult to establish causal relation - Does not fit for rare disease - Can not define the time duration of disease - Increase risk of Bias - Can not measure prevalence or incidence - Limited information
  38. 38. II- Designing Experimental Clinical Trial A- Randomized Blind Clinical Trial • Powerful design for causality • Overcome: Bias, confounders • Blinding is effective
  39. 39. Steps • • • • • • • 1- Selection of subjects 2- Measuring baseline variables 3- Randomize participants 4- Apply interventions (Placebo included) 5- Follow up Cohort 6- Measure outcome variables 7- Analyses of results
  40. 40. Selection of Subjects • Inclusive Criteria: • Optimize the rate of primary outcome • Generalization with minimal high risk for feasibility • Exclusive Criteria • Avoid unnecessary exclusions • Exclude non-contributors to primary outcome. And who will be difficult to follow e.g. mental status, harmful, treatment is ineffective • Reasonable guided by weighing • Restrict recruitment • Overcome the confounding variables
  41. 41. • Sample size and recruitment strategy • Limited # of subjects is wasteful, unethical and misleading conclusions. • Design a trial with: • • • • Large # Accessible population Enough time Enough money
  42. 42. Baseline variables • • • • • Collect tracking information (name, SS..) Demographic data Measure predictable variables (smoking) Measure outcome variable (ECG) Establish banks of materials
  43. 43. Randomization • In 2 or more groups (placebo included) • Avoid Bias by correct randomization for treatment and groups • Use Algorithm, blocked, stratified blocked to ensure no influence of the investigator. • Separate team my be used • IVRS (interactive voice response system)
  44. 44. Applying Interventions 1- Placebo • Blinding is recommended to avoid: 1- Co-intervention 2- Biased assessment of outcome • Blinding is difficult than randomization • Unbinding and breaking the code must be available 24 hour (pharmacist is helpful) • At end of the study ask subject and investigator for the guess
  45. 45. 2- Active Drug • It must be: Effective, safe, highest tolerable dose • Use a range of doses • Choice of Control • Consider Co-intervention for ethical reasons 1- when patient is under regular medications 2- use statistical adjustment for difference between groups. • Equivalence trial: under standard treatment.
  46. 46. Follow up and Adherence to the Protocol • Strategy for maximization: 1- Subject compliance 2- Easy intervention administration 3- Convenient visits 4- Painless study measurements 5- Encourage subject to be on trial 6- Find subject who lost follow up
  47. 47. Measure outcome • • • • • Clinical outcome Vs Surrogate outcome Statistical characteristics Number of outcome variables Valid outcome Adverse effect
  48. 48. Staging of Testing New Therapy by FDA • Pre-clinical • Phase I: Unblind, uncontrolled, few volunteers to test safety • Phase II: small, randomized, controlled, blinded to test tolerability, intensity, dose, pharmacokinetics • Phase III: large, controlled, blinded to test effect and outcome • Phase IV: large, observational (post marketing surveillance)
  49. 49. TITLE: A prospective, randomized, openlabel, multi-center, pharmacoeconomic evaluation, comparing XXX to epoetin alfa in patients with chronic kidney disease (CKD) stage V on dialysis. INDICATION Anemia of chronic kidney disease OBJECTIVES Primary: The primary objective of the study is to demonstrate significant provider time saving using XX monthly intravenous injections compared to up to three times weekly epoetin alfa injections for anemia management in in-center hemodialysis patients, while maintaining adequate Hb control Secondary: Demonstrate safety of XX TRIAL DESIGN Prospective, randomized, open-label, comparative, active-control multi-center study
  50. 50. TARGET POPULATION Patients with chronic kidney disease (stage V) on dialysis treated with epoetin alfa for the treatment of XX IV monthly CKD-associated anemia LENGTH OF STUDY INVESTIGATIONAL MEDICINAL PRODUCT(S) DOSE/ ROUTE/ REGIMEN 9 months Titration period: 7 months Evaluation period: 2 months Epoetin alfa weekly dose XX Monthly dose <8000 IU 120 µg/mL 8000-16000 IU 200 µg/mL >16000 IU COMPARATOR “DRUG” (or STANDARD OF CARE) DOSE/ ROUTE/ REGIMEN 360 µg/mL Epoetin alfa 1-3/week IV per dialysis center protocol
  51. 51. ASSESSMENTS OF: - EFFICACY Co-Primary endpoints: Time and Motion and Hb change from baseline Time and motion will be average time spent on anemia treatment over months 7-9 after the first study dose.. The co-primary endpoint is defined as the change in Hb concentration from baseline to the average in months 7-9. - SAFETY Safety endpoints: oVital signs (Pre- and post dialysis: BP, heart rate, body weight) oECG oAdverse events oLaboratory parameters oIron parameters: serum ferritin, serum iron, serum transferrin or total iron-binding capacity (TIBC); calculated transferrin saturation (TSAT) or percentage of hypochromic RBCs oHematology: hemoglobin (safety and efficacy), hematocrit, white blood cell count, platelet count oBlood chemistry: aspartate aminotransferase (AST), alanine aminotransferase (ALT), creatine kinase (CK), alkaline phosphatase (ALP), albumin, C-reactive protein (CRP), glucose in non-diabetics, potassium, phosphorus oAnti-erythropoietin antibody determination oDialysis adequacy: Kt/V or urea reduction ratio (URR)
  52. 52. INCLUSION CRITERIA Age ≥ 18 Able to understand, provide and sign written informed consent CKD stage V on outpatient in-center hemodialysis therapy Unchanged mode of therapy for the last 3 months prior to entry into study Anemia of CKD treated with epoetin alfa iv 3x/W Average Hb 10-13 g/dL over last 3 months prior to randomization. No Hb values should be <10 or >13 g/dl during this period. Kt/V ≥ 1.2 or URR ≥ 65% Adequate iron status: Serum ferritin ≥ 100 ng/ml OR TSAT ≥ 20% (or percentage of hypochromic red cells < 10%). Iron supplement therapy, either iv or po is allowed.
  53. 53. EXCLUSION CRITERIA Poor compliance with dialysis treatment, evidenced by >2 missed treatment monthly over the previous 3 months. Patients hospitalized during the previous 3 months prior to randomization for any cause except minor surgery such as vascular graft revision. Renal transplant Epoetin alfa dose >300 U/kg/week. Patients expected to change dialysis modality over the course of the study Immunosuppressive therapy (other than corticosteroids for a chronic condition, cyclosporine and monoclonal/polyclonal antibodies) in the 6 months prior to enrollment Overt gastrointestinal bleeding or any other bleeding episode necessitating transfusion within 3 months prior to enrollment Hemoglobinopathies (e.g., homozygous sickle cell disease, thalassemias of all types Hemolysis Active malignant disease (except non-melanoma skin cancer)
  54. 54. - Chronic, uncontrolled or symptomatic inflammatory disease (e.g., rheumatoid arthritis, systemic lupus erythematosus) - Acute infection - C-reactive protein (CRP) > 30 mg/L - Temporary (untunneled) dialysis access catheter - Vitamin B12 deficiency - Folic acid deficiency - Uncontrolled or symptomatic secondary hyperparathyroidism - Poorly controlled hypertension (sitting pre-dialysis systolic blood pressure ≥ 170 mmHg – average of 2 screening values with at least one day between measurements) - History of seizures in previous 6 months prior to enrollment - Current (or history of) Pure Red Cell Aplasia - Platelets > 500x109/L - Chronic Congestive Heart Failure (NY Heart Association Class IV) - Myocardial infarction, severe or unstable angina, coronary artery disease, stroke, severe liver disease within the 12 weeks prior to enrollment - High likelihood of early withdrawal or interruption of the study for any reason
  55. 55. - Surgery during the course of the study or within 30 days prerandomization, unless simple surgery, i.e., simple dialysis access revision, laser photocoagulation, etc., - Pregnancy or breast feeding - Women of childbearing potential without effective contraceptive methods - Previous treatment with XX - Administration of another investigational drug within 6 months prior to randomization - Administration of erythropoietin-related compounds other than those specified by this protocol within 6 months prior to randomization - Known hypersensitivity to any constituent of the study or reference medication - RBC transfusions within 3 months prior to randomization - Life expectancy <12 months from randomization. - Participation in studies testing investigational devices or dialysis solutions, unless reported to the sponsor in advance and approved by the sponsor. - Erythropoietin-related compounds other than those specified by this protocol within 6 months prior to randomization, during the screening/baseline or the treatment periods