Research Methodology and study design.pptx

Research methodology, study design, level of evidence,

1. Research Methodology (D e sign of study, Sampling, data collection and Analysis, Level of Evidence and grades of Recommendation) Dr Uttam Nepal 2nd yr Resident MS ENT-HNS KIST MCTH

2. Roadmaps: R esearch: Introduction and its Types S tudy design: Different types of study design S ampling : Differnet techniques and its importance M ethods of data collection, analysis and interpretation E vidence and recommendations.

3. If we knew what it was we were doing, it would not be called research, would it? Albert Einstein

4. R esearch

5. research French   " recherche “ - to go about seeking Research is a quest for knowledge through diligent search or investigation or experimentation aimed at the discovery and interpretation of new knowledge (WHO). Redman and Mory defined research as a “ systematized effort to gain new knowledge .” Clifford Woody : Research comprises of defining and redefining problems , formulating hypothesis, collecting, organising and evaluating data, making deduction and reaching conclusions and at last carefully testing the conclusions to determine whether they fit the formulating hypothesis

6. R esearch: Need and purpose Improve existing knowledge – generate new technologies Identify priority problems – design and evaluate policies to deliver greatest health benefit with available resources Open the new horizon for research

7. R esearch: significance Evidence based practice Treatment protocol Treatment of sudden sensorineural hearing loss with IV steroid Policy making Iodine supplementation in salt to prevent goiter in adolescence Problem identification and prioritization Quality of health service Driving force – new research

8. R esearch: importance As a postgraduate student - To critically analyse the information - Thesis preparation As future practitioner - Evidence based practice - Professional status - Participate in research projects As an educated citizen - To understand difference between scientifically acquired information and others

9. R esearch: fields of application Useful in policy making Treatment protocol Investigation Sequel/Complication Quality of service Actual size of problem Avenue for further research

10. T ypes of research Basic research To generate new knowledge and technologies In vitro / in vivo experiments (animal experiments) eg. p53 gene mutation in HNSCC Applied/Action research Involving practical application eg. Monoclonal antibody therapy in HNSCC Discriptive research Eg. Prevalence of ca larynx in adult population in Kathmandu Analytical research Eg. Role of environmental pollution in allergic rhinitis

11. T ypes of research Quantitative research Based on the measurement of quantity or amount eg. Recurrence rate of sinonasal polyp after FESS Qualitative research Relating to quality or kind eg. Quality of life after surgery for head and neck cancer Outcome research Impact of intervention on the patient in terms of health status and health related quality of life eg : Hearing outcome after butterfly cartilage tympanoplasty Experimental research Comparing two groups on one outcome measure to test some hypothesis regarding causation eg. Comparison of dexamethasone versus bupivacaine in reducing early postoperative pain after tonsillectomy in children

12. T ypes of research Conceptual / Empirical Field-setting research or laboratory research or simulation research Clinical or diagnostic research Health systems research Exploratory Historical research

13. H ealth research: types

14. R esearch method and methodology

15. Components of R esearch Conceptualizing the problem Statement of problem Research questions Rationale Literature review Formulating objectives Developing testable hypothesis Ethical considerations Workplan

16. S election of topic Interest Relevance Magnitude of problem Severity of problem Population affected Avoidance of duplication Feasibility Political acceptability Applicability Cost effectiveness Ethical consideration

17. R esearch question “Uncertainty” about something in the population that the investigator wants to resolve by making measurements in the study population . It is best to frame the research question using the PICO format P - Population/patient/person with(or at risk of) a health problem on which research is based. I - Interventions C -Comparisons O -Outcome Often ( T ) – Time frame is also added to the PICO format

18. R esearch(contd.) Components of a Good Research F easible- Adequate number of participants , technical expertise and resources . I nteresting N ovel – Provides new information E thical- Amenable to a study that ethics committee will approve R elevant - To the scientific world of knowledge / clinical practice/ health policy.

19. Study Design

20. S tudy design

21. S tudy design: Types Non intervention studies Exploratory studies Descriptive studies Comparative or analytical studies Cross sectional Case - control Cohort Intervention studies Experimental studies Quasi – experimental studies Parallel design Cross over design Factorial design Meta Analysis

22. Non intervention studies Exploratory Study To gather preliminary information that will help define problems and suggest hypotheses Helps in selection of: Best research design Data collection method Selection of subjects Audits E.g. Quality of discharge Summary in different department of KISTMCTH D escriptive study In-depth description of characteristic of one or limited number of cases. Case studies: case report of uncommon diseases e.g.. B/L facial nerve palsy Case series

23. C ross sectional study Quantifying distribution of variable in population at a point of time Quantify disease burden Useful for hypothesis generation Covers a sample of population Total population covered:- census E.g.. Prevalence of COM, children with OME OUTCOME EXPOSURE STUDY

24. Cross sectional study Advantages: Cheap and simple; Ethically safe. Disadvantages: Establishes association at most, not causality; Confounders may be unequally distributed; Group sizes may be unequal.

25. CASE CONTROL STUDY Compares one group with problem to another group without problem(control group) Retrospective from outcome to exposure Confounding factors EXPOSURE STUDY DISEASE

26. CASE CONTROL STUDY Odds Ratio: Chances that a case was exposed to risk factor to the odds that a control was exposed to risk factor Estimation of risk & not a true risk value Case Ca Larynx Control Without Ca Smoking 33(a) 55(b) No smoking 2(c) 27(d) Odd Ratio: a x d/ c x d – 8.1

27. CASE CONTROL STUDY Advantages: quick and inexpensive; Good for rare disorders or those with long lag between exposure and outcome; fewer subjects needed than cross-sectional studies. Possibility of exploring multiple exposures Disadvantages: reliance on recall or records to determine exposure status; confounders; selection of control groups is difficult; potential bias: recall, selection.

28. COHORT STUDY Group exposed to risk factor (study group)  compared with group not exposed (control group) Prospective study, Incidence Longer, labour intensive, expensive & loss to follow up EXPOSURE STUDY OUTCOME

29. COHORT STUDY Relative Risk (RR): Incidence among exposed/ Incidence among non exposed = a/( a+b )/c/( c+d ) True estimate of risk,  value- strength Interpretation of Relative risk: RR =1: no association <1: protective effect <2 : weak association >30: almost aetiological association Ca Larynx No Ca Larynx Total Smoker 60(a) 40(b) 100 ( a+b ) Non smoker 30(c) 70(d) 100 ( c+d ) RR:Incidence in Exposed/Incidence of non exposed-60/30-2

30. COHORT STUDY Advantages: ethically safe; subjects can be matched; can establish timing and directionality of events; eligibility criteria and outcome assessments can be standardized; Disadvantages: controls may be difficult to identify; exposure may be linked to a hidden confounder; blinding is difficult; randomization not present; for rare disease, large sample sizes or long follow-up necessary.

31. INTERVENTION STUDIES Experimental study Randomized Control Trial( RCT) Randomization, Control, Blinding Quasi- Experimental Studies Parallel Design Cross-over design Factorial design Meta analysis

32. RANDOMIZED CONTROLLED TRAILS(RCT) Can actually prove causation At least 2 groups (experiment or intervention & non-intervention) Three characteristics: Randomization : to allocate subjects to control & experimental group Control: to compare with experimental group Manipulation: some intervention to one group

33. QUASI- EXPERIMENTAL STUDY Study  little or no control over allocation of the treatments or other factors being studied The key difference : lack of random assignment  impractical or unethical Easier to set up, minimizes external validity Threat to internal validity, value of result less than RCT

34. PARALLEL DESIGN Parallel Groups Multiple concurrent experimental arms Different treatments Different doses Control arm(s) Placebo, active control Balance/imbalanced randomization e.g. ISSNHL :- Placebo, i.v steroid, IT steroid

35. CROSS OVER DESIGN

36. CROSS OVER DESIGN

37. FACTORIAL DESIGN Evaluates two interventions simultaneously Four possible treatment combinations Efficient approach in some circumstances Potentially more informative approach Major concern: interaction of interventions

38. Clinical trails phases Phase I: Normal volunteers To find Maximum tolerated dose Phase II: Patients with Diseases To establish effects & side effects Phase III: Clinical trial Phase IV: Long term surveillance Phase I Phase II Phase III Phase IV

39. Sampling

40. SAMPLING Population or Universe : a set of all individuals or objects having common characteristics Sample : subset or part of the population Sampling : Process or technique of selecting a sample of appropriate & manageable size for study Sampling unit: breakdown parts of population which are distinct, unambiguous & non-overlapping e.g.. Hospitals, 15-25 yrs

41. SAMPLING Advantages Reduces cost of study considerably Greater speed (less time) Greater accuracy Increase scope of study: several aspects Some tests not possible in population can be applied Study can be in depth & more subtle

42. RELIABLE SAMPLE Efficient Representative Measurable Sizeable Coverage Goal oriented Representative of underlying population Feasible Economic & cost efficient

43. errors Sampling errors Repeated sample from same population, result different Collection, processing, analysis Decrease by increasing sample size Non sampling errors Observational & defective measurement technique Error in editing, coding & tabulating results Increased by increasing sample size

44. TYPES OF SAMPLING A. Probability Each element has equal chance of being included Simple random Systematic random Cluster Stratified B. Non-probability Each element may not have same chance Convenience Purposive Quota sampling

45. SIMPLE RANDOM SAMPLING Equal chance Decided by law of chance Provides greatest number of possible sample Small sample-Lottery Large sample-Table

46. SYSTEMIC RANDOM SAMPLING Pre determined system is followed Number of possible sample is greatly reduced Every 3 rd or 10 th First sample is selected randomly

47. CLUSTER SAMPLING Selection made of cluster E.g.. VDC, tole/ward; military camp; School Prevalence of OME among children in day care

48. STRATIFIED SAMPLING Population divided into different strata E.g.. Religion; socioeconomic status Random sampling done in each strata proportionate sampling less likely to miss smaller group e.g.. If out of 1000 in a community, planned for 100 Hindu = 850; sample = 85 Muslim = 100; sample = 10 Christian = 50, sample = 5

49. NON PROBABILITY SAMPLING Convenience: Haphazard, non representative Purpose usually for exploratory Patient visiting ENT OPD Purposive: Predetermined idea Results can not be generalized Natural history of AOM in immunodeficiency Quota sampling: Population divided into mutually exclusive groups and subjects or unit selected by judgment and not randomly

50. PROBABILITY VS NON PROBABILITY SAMPLING

51. D ata collection and interpretation

52. variables Variable Any observation that can take on different values Attribute A specific value on a variable Variable: Age, Gender Attribute: 18/19/20, Male/Female Types of Variables: Independent Variables vs. Dependent Variables Qualitative Variables vs. Quantitative Variables

53. variables Qualitative / categorical variable N on-numerical values Dichotomus : Antibiotic use –yes/no, Gender-Male/Female P olychotomus : N ationality Ordinal: Disease severity –mild/moderate/severe Quantitative variable Intrinsically numerical Continuous – Hearing threshold Discrete – Number of attacks of tonsillitis

54. variables Independent variable Factor that is assumed to cause / influence the problem e .g. Smoking in the study of relationship between smoking and laryngeal cancer Dependent variable Factor that gets modified under influence of independent variable e .g. Suffering from laryngeal cancer

55. S cales of measurement Nominal Qualitatative data (Male, Female) O rdinal Categories are ranked eg . Pain - Mild, Moderate, Severe Stage of cancer – I,II, III ,IV I nterval Intervals between Classes equal, zero point arbitrary eg. Pain scale (1-10) IQ score (90,95,100) R atio

56. SAMPLE SIZE ESTIMATION Sample size calculation depends on Type of study Type of statistical analysis required Formula depends on: The prevalence of the condition/attribute of interest Margin of error Confidence limits Difference in proportions required Expected RR required

57. S tudy instruments Tools by which data are collected Questionnaire and interview schedules Other methods: Medical examination Laboratory tests Screening procedures Previous hospital records Survey Focus group discussion Designing of recording forms Mailing/telephone/e-mail/face to face interview

58. P urpose of data collection Get information Monitor progress Evaluate performance & use of resources e.g. audit Identify gaps between knowledge & practice Evaluate impact of programs Make decision Utilize in planning

59. D ata collection: Steps Identification of objective Listing target population Identification of time frame Selection of data collection instrument Development of data collection method Selection of proper sampling design Organization Precoding Pretesting Field survey launching

60. D ata analysis and interpretation Plan for analysis beforehand Description should include Design of analysis form Plan for processing & coding data Choice of statistical method

61. C riteria in selecting appropriate analytical method Objectives of the study Design of the study Scale by which the variables are measured Sample size Number of variables to be analyzed

62. M ethods of data interpretation presentation T able: Simple Should be numbered Self explanatory title Heading of rows or column clear & concise Foot note may be given Presented according to size or importance

63. M ethods of data interpretation presentation G RAPHS: Methods of showing quantitative data using a coordinated system Arithmetic scale line Semi logarithm scale line Histogram Frequency polygon Scatter diagram

64. M ethods of data interpretation presentation Charts : Bar Pictogram Pie Geographic co-ordinate Flow Organization

65. M ethods of data presentation Percentage no of units with certain characteristics x100 no of units in sample n = 50 ; M = 20 : F = 30 ; % of M = 40% Proportion Compares relation in magnitude one part of study unit to whole M = 2/5 : F = 3/5 Ratio Relation in size between 2 or more parts M:F=2:3 Rate Amount or degree of something measured in specific period of time

66. M easurements Measures of central tendency Mean Median Mode Hb of patients admitted with epistaxis (in Gm%) 13.2, 9.5, 11.7, 13.6, 10.4, 11.7, 9.9, 14.0, 12.4, 11.9, 10.3 Mean = 11.7 Median = 9.5, 9.9, 10.3, 10.4, 11.7, 11.7 , 11.9, 12.4, 13.2, 13.6, 14.0 Mode = 11.7

67. M easurements Measurement of dispersion Range: Difference between largest and smallest value Percentiles: Parts that divide measurements into 100 equal parts e.g. 3 rd percentile, 10 th percentile Standard deviation: Measure which describes how much individual measurement differs on an average from mean ± 1 SD = 68.2% ± 2 SD = 95.4% ± 3 SD = 99.6%

68. D ata analysis Manually Using computer programs Excel Data base SPSS Epi-info SAS: Statistical Analysis Software

69. D eterming difference between groups

70. S ignificance test application Difference between groups X 2 test Sign test t-test Association between groups X 2 test Odds Ratio Relative Risk Pearson’s Spearman's

71. Significance test application Chi Square test Criterion for using  2- test Data should be categorical or qualitative one  2- test is used to: Find out the significance of difference between two proportion Find out the association or relation between two variables Students’ T test Criterion for using t-test distribution should be normal data set may be quantitative sample size  30 Z test Criterion for using Z-test Distribution should be normal Data set may be qualitative or quantitative Sample size  30 Spearmans’s rank correlation coefficient(  )  = 1 - 6  d 2 / n ( n 2 – 1) Where, d - difference of paired ordered observations & n - number of paired observations

72. E vidence and recommendation

73. evidence Evidence a thing or things helpful in forming a conclusion or judgment E videnced – B ased M edicine “ the integration of the best research evidence with clinical expertise and patient value ” “ conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients ” sackett et al 1996 BMJ

74. L evel of evidence

75. L evel of evidence Level A: Consistent Randomized Controlled Clinical Trial, cohort study, clinical decision rule validated in different populations. Level B: Consistent Retrospective Cohort, Exploratory Cohort, Ecological Study, Outcomes Research, case-control study; or extrapolations from level A studies. Level C: Case-series study or extrapolations from level B studies. Level D: Expert opinion without explicit critical appraisal, or based on physiology, bench research or first principles.

76. Q uantification of evidence

77. G rades of recommendation Grade A: Good scientific evidence suggests that the benefits of the clinical service substantially outweighs the potential risks. Clinicians should discuss the service with eligible patients. Garde B: At least fair scientific evidence suggests that the benefits of the clinical service outweighs the potential risks. Clinicians should discuss the service with eligible patients. Garde C: At least fair scientific evidence suggests that there are benefits provided by the clinical service, but the balance between benefits and risks are too close for making general recommendations. Clinicians need not offer it unless there are individual considerations.

78. G rades of recommendation Grade D: At least fair scientific evidence suggests that the risks of the clinical service outweighs potential benefits. Clinicians should not routinely offer the service to asymptomatic patients. Garde N : Scientific evidence is lacking, of poor quality, or conflicting, such that the risk versus benefit balance cannot be assessed. Clinicians should help patients understand the uncertainty surrounding the clinical service.

79. L evel of evidence and grade of recommendaton

80. summary Research is a quest for knowledge through diligent search or investigation or experimentation aimed at the discovery and interpretation of new knowledge R esearch is very important aspect in medical f ie ld. RCTs are regularly done trails in medical field S ampling and its different methods are used for minimizing error Mean, mode and median are used for measuring central tendency ; Range, percentile and SD are used for measuring dispersion X 2 is used for measuring association and difference between groups L evel of evidence and Grade of Recommendation are associated with each other

81. References: Scott-Brown’s Otorhinolaryngology, Head and Neck Surgery, 8 th edition Park’s Text book of social and preventive medicine ,21 st edition