Research One-Month Program Handouts (PDF)

**Research One-Month Program** **Including all the materials, links, assignments, pre-test and post-test.** **\*ALL rights reserved.** **Do not distribute it anywhere** **It's only for you.** **Please, write all what you will learn in this rotation in a special notebook or File titled "Research...

**Research One-Month Program** **Including all the materials, links, assignments, pre-test and post-test.** **\*ALL rights reserved.** **Do not distribute it anywhere** **It's only for you.** **Please, write all what you will learn in this rotation in a special notebook or File titled "Research Rotation" and keep it with you. Keep add to it all what you will learn after this rotation about research and EBM.** **The handout is two parts:** **1^st^ part- Basics** **Common terms you need to know;** - PI; primary investigator, who is full responsible person - Study coordinator; is the backup and facilitator for the PI - The author corresponding: he/she is the primary one to communicate with the journal - Research pharmacist; is important in any-study including medications - Recruiter; The one who find the study subjects - RA; Research Assistance; and volunteer assistance. They are usually an interns or medical students - Statistician are in different levels and certificates, like, Bachelor, Master, and PhD, etc.. - Cost and budgets; put the details and justifications, it may include the compensation for the side effects and any consequences of the trial, transportation, time out of works, and this calculation is per patient or per investigator. - Always have a time horizon or time-line for your study. It is the main indicator for your success. - The study protocols, data, usually preserved in your institution all the time and kept for 2 years usually, but there is no standards, so it differ between institutions. - Consent; voluntary, capacity, free of coercion, comprehensive, free to withdraw at any time with no consequences....easy contact for concern or even to know the results Always indicate alternate to participate, he may have a copy if he wants. - Ethics in clinical research; Link[, ] - IRB; institutional review board; https://www.youtube.com/watch?v=U8fme1boEbE&ab\_channel=U.S.DepartmentofHealthandHumanServices - For any new drug to get FDA approval and to be accepted for selling it has to go through 4 phases and minimum three. It is a very long process; it may take even 10 years. IN RARE SCENARIOS THEY MAY ALLOW FOR FASTER PROCESS SUCH AS IN A DISASTER PANDEMIC FOR VACCINES. - The 4 phases for the new drug to be marketed - There is different levels of evidence and class of recommendation. Link, https://www.ahajournals.org/doi/full/10.1161/cir.0000000000000312 - Incentive to the subjects can be, but it should be justified with aiming not to bias the study. - When we should stop for benefits (truncated trials) محاكمة مبتورة and harms? Is there any caveats? Link https://www.distillersr.com/glossary/truncated-trials - Always check whether your patient has been in other study. The patient should not be in two simultaneous interventional studies. - Different survey programs; Monkey, GCRC assistance, Adope live cycle designer, google forms - Difference between effectiveness, and efficiency? [[https://www.researchgate.net/post/What\_is\_the\_difference\_between\_efficacy\_and\_effectiveness\_with\_respect\_to\_clinical\_trials]](https://www.researchgate.net/post/What_is_the_difference_between_efficacy_and_effectiveness_with_respect_to_clinical_trials) - Clinical Practice Guideline; How it's maid? Standard of care Vs EBM, It should be clearly stated who did these guidelines, link; you'll find the most important guidelines - benefits and CAUTIONS of guidelines In general, the guideline is a process done by a team. They shouldn't have any conflict of interest, and they rely on EBM as much as possible, but may go with personal opinion. - How to cite and save the articles? you can use Refworks, Endnotes (paid software), or Mendely. - How to communicate with the statistician- You need to have such research rotation, so you can understand him/her and they do the same with you. **EARLY COMMUNICATION WITH THE STATICIAN MAKES DIFFERENT** - ICD- 10 codes is a code system for all diagnosis made. For instance, you can follow the previous charts for all PE patients, because you can just simply enter the code of PE and track all the previous cases. - Types of studies; either 1. Observational A. Analytics such as case-control or cohort B. Descriptive such as case report, case series 2. Experiemental where we do intervention such as RCT. - We have different types of control and each one has its indication and problems Placebo Standard No treatment Self Sequential control Cross-over control Historical control. [[https://www.scientific-european-federation-osteopaths.org/different-types-of-clinical-trials/]](https://www.scientific-european-federation-osteopaths.org/different-types-of-clinical-trials/) [https://www.youtube.com/watch?v=Opf9QbosgHM&ab\_channel=GCP-Mindset-AllAboutClinicalResearch] - **Data analysis programs and software like SPSS, SAS, and R program and STATA.** - Important book; Designing Clinical Research by Stephen B. Hulley. - **Important engines; data-base** Scopus, Science direct, Scirus, Pubmed, Google scholar, Medline, Orchid, TNT, Best-evidence, Chochrane, Ovid medline, web of science, DBLP, JSTOR, ISI web of knowledge, Embase, Sinomed, CINAHL. - Terms; Fraud تزوير Plagiarism سرقة أدبية.... - We have Central tendency measures like Mean, Mode, and Median - There is another term for variations of the variables as Range, Interquartile, and Standard deviations link [ ] [[https://www.statisticshowto.com/probability-and-statistics/interquartile-range/]](https://www.statisticshowto.com/probability-and-statistics/interquartile-range/) **Part 2- Advance** - **Cost-effective;** **These used mainly for the higher level, such as the ministry of health and CEO directors** you aim or target the less cost and at least equal effective, or most cost but most effective and reduce other cost like absence from the work. Cost-effectiveness ratio= cost per year of life gain. There is an important term in cost effective analysis studies which is QALYS= Quality Adjusted Life Years and it ranges from 0-1. We are not looking only for money, we look for feasibility, ethical, social etc,,,. There is another term; DALYS= Disability Adjusted Life Years. It represent combined mortality and morbidity of a disease We always; refuse any intervention or drug cost more than standard with less benefits and always do less cost and more or equal benefits. In case of more cost and more benefits or less cost and less benefits we will look for CEA (COST-EFFECTIVE ANALYSIS). [[https://www.youtube.com/watch?v=pv8M95aII2M]](https://www.youtube.com/watch?v=pv8M95aII2M) [[https://www.youtube.com/watch?v=TN1aZ9J-2zA]](https://www.youtube.com/watch?v=TN1aZ9J-2zA) There is another term called ICER= Delta Cost/ Delta Qalys, How we measure the quality of life; There is a standardized questionnaire called EQ or 5D. Link for critical appraisal for cost effective studies [ ] - **Scientific writing;** it's an important skill you have to master, and these a few links, website, book may help. The saurus.com where you will have different synonyms. There are two important sites you can check your words, sentences, and the whole paragraph if it has been written correctly and in a scientific way or not 1-Corpus of contemporary American English --COCA [https://corpus.byu.edu/coca/ ] 2-Word and phrase- [ ] There is a good book for scientific writing called- **Scientific Research Writing for non-native speakers of English-** **by Hilary Glasman-Deal.** - Understanding Confidence Interval link - What is the null hypothesis and alternative one. Link - What is the different between parametric tests and non-parametric tests. Link - Different statistical tests made for different types of data- normally distributed or abnormally- dependent and independent variables, how many groups, and is it correlation, association or causation study. You will see different tables for how to choose the statistical tests depends on the above-mentioned points. - **We have different terms we need to know in a diagnostic study and therapeutic studies, prognostic, correlation or degree of agreements studies and prevalence or incidence, or survival.** - **In therapeutic studies**; Different definitions; RR= 0.5 means , 1 means no different , less than one is favor intervention if you aim to reduce outcome and more than one if you want to increase the outcome = 1- 0.5= 50% reduction If the result is 0.25= 1- 0.25= 75% reduction So, the far from one is more significant On the other hand , if your target to increase the outcome, so it should be more than one..... Example 1.76 = 1.76 times or 1.67- 1= 67= 67% better **Excellent reference -- youtube- NCCMT** RRR= 40% means, REDUCE BY 40%\....IT MAY AUGMENT THE RESULT MAINLY IF THE RESULT IS NOT SIGNIFICANT ARR= 12% means, 12% MEANS REDUCTION BY 12% NNT NNH Table: 4x4 Outcome yes no Intervention A B ---------------- --- Placebo C D 1. **ARR= (A/A+B ) - ( C/C+D) = RISK DIFFERENCE= HOW MUCH PERCENTAGE = طرح** 2. **RR= (A/A+B)/ (C/C+D) = RISK RATIO** **تقسيم** **FROM THE ABOVE CALCULATION WE CAN CALCULATE THE NNT AND RRR** **NNT= 1/ ARR** **RRR= 1- RR X100 = wRong, wRong, wRong and misleading** **I remender it as 3R= 3 WRONGS** 3. **Odds ratio= (A/C)/ (B/D) = HOW MANY TIMES** Ratio Vs Odds example; you make a campaign and reach a 100 people then they call them and other 100 non involved in the campaign Received the shot Did not receive the shot -------------- ------------------- -------------------------- -- intervention 80 20 control 55 45 Odds of who received the shot = 80/20= 4 Odds in control = 55/45= 1.22 Overall = 4/1.22= 3.3 Odds= 1 = no effects Our campaign makes them 3.3 times than those didn't receive - Correlation and regression it ranges from -1 to +1 , where 0 is no correlation and +1 maximum positive correlation, and -1 is the negative correlation. Link, https://www.youtube.com/watch?v=xTpHD5WLuoA&ab\_channel=LEARN%26APPLY%3ALeanandSixSigma - Survival analysis and Kaplan-meier plots, IT IS ABOUT TIME TO EVENT OCCURS [] (see the three parts of it). **In diagnostic studies** - In diagnostic tests; Sensitivity, and specificity, which is patient oriented tests and not affected by the prevalence like the predictive value does. - Positive predictive and negative predictive value which affected by the prevalence of the disease and it is a test oriented. - Other important tests; Positive and Negative likelihood ratio, which both SENSITIVITY AND SPECIFITY above tests. [ ] [ ] excellent 2X2 TABLE ; patient positive patient negative Test positive TP A FP B -------------------- ------ Test negative FN C TN D **Sens= TP/ (TP+FN)= A/ (A+C)** **Spec= TN/(TN+FP) = D/ (D+B)** **PPV= TP/(TP+ FP) = A/ (A+B)** **NPV= TN/ (TN+ FN) = D/( D+C)** PPV INCREASE WITH INCREASE PREVALANCE AND DECREASE WITH DECREASE PREVALANCE. NPV IS INVERSELY TO PREVALANCE **Likelihood ratio:** It incorporates two tests together sensitivity and specificity. \+ LR= Sensitivity / 1- specificity -LR= 1- Sensitivity/ specificity https://www.youtube.com/watch?v=TzPvCSFZUSQ&ab\_channel=TerryShaneyfelt What's the different between prevalence and incidence? In incidence (overtime how many) you look for possible causation,,, you search over time like RCT or Cohort In prevalence (now how many) you just look now only,,,like cross sectional study... difficult to assess causation or association. - Area under the curve- roc curve? What does it mean? [] (he has excellent lectures) We are trying to avoid type 1 and type 2 in conducting our study, so, what are they? [(he has excellent lectures)] - What is the power of study means? - There is different between causation and association studies. We will discuss it in the meeting. - Different between Literature review and Systematic review? [ ] Writing literature review; [ ] Writing Systematic review; - Stages and steps for conducting the research; **Clear and excellent question,** **It should fill a gap in the knowledge and it should have a benefit for audiences** Document and save your work, Council colleagues you thrust, Writing proposal and grant, **it should have a solid methodology** Frequent meeting, Data collection, **Set a time- horizon** Entering and analysis, Writing a manuscript and application. Search journals. - Always increase and insure **validity= accuracy** and **reliability= precision**. - **Interim analysis**, we may need it to look for missing data or early stopping the study for clear benefits or harms. - Always have the approval of the IRB before you start your project. Don't ever start any project without the IRB approval first. The exception is the literature review and systematic review with or without meta-analysis. - Always have data secured, and HIPPA regulations. HIPAA stands for Health Insurance Portability and Accountability Act. Passed in 1996 HIPAA is a federal law that sets a national standard to protect medical records and other personal health information - **Types of data;** quantitative and qualitative, categorical=nominal, ordinal= ranks, interval has zero and ratio has no zero,.... [https://www.intellspot.com/data-types/] - What we usually look in our studies; it is also the common types of studies Prevalence, Incidence, Prognostic, Association and correlations Causation, Therapeutics, Diagnostic. Review such as literature review or systematic review - Types of statistical tests; 1. **Descriptive statistic**s one which look for the study data itself. 2. **Inference statistiscs** to infer the results to the general populations. - Different between inferential and descriptive statistic tests [ AND ] [[https://www.youtube.com/watch?v=VHYOuWu9jQI]](https://www.youtube.com/watch?v=VHYOuWu9jQI) - Always we aim to avoid the bias and to be generalized results. - We have to know the independent and dependent variables. [https://www.youtube.com/watch?v=utNpSEEyMIU&ab\_channel=MathHelp.com] - Pilot study?? [ ] - What is the power of study means? - Missing data, How we manage it? Very advance terminology and statistical issues- just to have an idea and not to deep understand [[https://www.youtube.com/watch?v=asyJCVLV4LI]](https://www.youtube.com/watch?v=asyJCVLV4LI) - **General rule in statistics;** **All tests assume that all data are normally distributed and all relations are linear correlation.** - Outliers; 5 ways to deal with outlier https://cxl.com/blog/outliers/ - Correlation and regression. What are they and what is the different? [[https://www.youtube.com/watch?v=4EXNedimDMs]](https://www.youtube.com/watch?v=4EXNedimDMs) ( best ever statistician speaker. I like him the most). - In multiple regression; When there is multi-variables independent and one dependent or outcome variables. The independent variables should not relate or correlate to each other otherwise we will have multi collinearity risk and redundant. - Logistic regression; we do it for binary outcome. - What is ANOVA? ANCOVA? [[https://www.youtube.com/watch?v=0Vj2V2qRU10]](https://www.youtube.com/watch?v=0Vj2V2qRU10) - General rules; any statistical test has its (assumptions) to be applied first. - One way and two ways ANOVA - How can we choose the statistical tests [https://twitter.com/research\_con/status/1695105608230359377/photo/1] - What types of data [ ] - What is normally distributed data - How to calculate the sample size - **Critique an articles;** **It depends mainly on your background knowledge and using pre-printed Checklist Questions for each type of studies, you can use it.** - Survival analysis/ Hazard ratio [https://www.youtube.com/watch?v=z1b2hFzXsrU] - Interpret Forest Plot - Types of statistical graphs - Intension to treat Vs per-protocol analysis ### - **Conduct a new study;** **We will have a one-day discussion (open discussion).** I will have your feedback at the end of this rotation. It is important to make the necessary changes for future rotations. Wish you all the best, DR. Yahia Akeely ED consultant Clinical Research Fellowship