1 / 25

ASPIRE CLASS 7: Interpreting Results and Writing an Abstract

ASPIRE CLASS 7: Interpreting Results and Writing an Abstract. Jordan King, PharmD, MS Clinical Pharmacy Specialist, Research/Ambulatory Care Investigator, Clinical Pharmacy Research Team Kaiser Permanente Colorado. Agenda for the presentation. Interpreting Results.

redell
Télécharger la présentation

ASPIRE CLASS 7: Interpreting Results and Writing an Abstract

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. ASPIRE CLASS 7: Interpreting Results and Writing an Abstract Jordan King, PharmD, MS Clinical Pharmacy Specialist, Research/Ambulatory Care Investigator, Clinical Pharmacy Research Team Kaiser Permanente Colorado

  2. Agenda for the presentation Interpreting Results Explain the difference between cumulative incidence, incidence rate, and prevalence Explain the difference between relative risk, odds ratio, and hazard ratio Calculate the number needed to treat and the number needed to harm Writing an Abstract List the essential elements of an abstract Recognize and apply succinct writing to an abstract

  3. A key feature of all clinical research is measurement. This is typically done in two or three steps: how frequently are events occurring in a group or population (measures of frequency) what is the effect of an exposure on the frequency of events in two or more groups (measures of association) how would changing the exposure impact the frequency of outcomes (measures of potential impact) Measures of frequency Measures of association Measures of Potential Impact How do we describe results and make comparisons? Prevalence Relative Impact on population Incidence Absolute Impact on exposed

  4. Measures of Frequency – Quantifying the occurrence of outcomes

  5. Example from VIGOR: What is risk? The probability that an event will occur in a given period Expressed as a % Calculated as: Important characteristics of risk: Easily interpreted Requires all patients to be observed/at-risk for same duration Must know duration of time to interpret Incidence proportion (risk) GI events: MI events: Rofecoxib = Rofecoxib = Naproxen = Naproxen =

  6. Example from VIGOR: What is a rate? The frequency at which an event occurs Expressed as events per person-time Calculated as: Important characteristics of rate: Assumes the likelihood of an event is constant during at-risk period Allows for variable follow-up between patients Must consider scale (per 10, 100, 1000 person-years) Incidence rate GI events: MI events: Rofecoxib = Rofecoxib = Naproxen = Naproxen = per 1000 person-years per 100 person-years

  7. Is this a rate or a risk?

  8. What is prevalence? The existing cases of disease at a given time Expressed as a proportion Calculated as: Important characteristics of prevalence: Easily interpreted Used to describe a point or period in time Affected by both incidence and duration of disease Prevalence

  9. Measures of Association – making comparisons between groups

  10. Risk difference, incidence rate difference, absolute risk difference, etc. The difference in risk or rate between two groups Measures absolute effect of an exposure Calculated as: Interpretation: Difference <0, lower in the exposed Difference >0, higher in the exposed Difference =0, no change in the exposed Absolute Measures Example from VIGOR: GI events: MI events: RD = RD = IRD = IRD =

  11. Risk ratio, relative risk, incidence rate ratio The ratio of risk or rates between two groups Calculated as: Interpretation: Difference <1, lower in the exposed Difference >1, higher in the exposed Difference =1, no change in the exposed Relative Measures Example from VIGOR: GI events: MI events: RR = RR = IRR = IRR =

  12. Absolute effect “When measuring the effect of an exposure on the health of a population, an absolute effect measure is needed.” Relative effect “These measures indicate the extent to which the exposure in question accounts for the occurrence of disease among the exposed people who get disease.” I often report both When to use absolute vs relative effect measures?

  13. VIGOR Trial: NNT risk = 1 / (0.014 – 0.030) = 62.5 NNT rate = 100 / (2.1 – 4.5) = 41.7 On average, 63 patients would need to receive rofecoxib, as opposed to naproxen, to prevent one additional GI event NNH risk = 1 / (0.004 – 0.001) = 333.3 NNH rate = 1000 / (7.3 – 1.7) = 178.6 On average, 333 patients would need to receive rofecoxib, as opposed to naproxen, to cause one additional MI The average number of patients who need to be treated to prevent (or cause) one additional event Provides an easily interpretable estimate of the magnitude of an effect Calculated from absolute differences • 1/ARD • people/IRD Must consider TIME! Number Needed to Treat and Number Needed to Harm

  14. Odds ratio (OR) Used for case-control studies or to determine likelihood of dichotomous outcome in cohort studies Calculated from logistic regression model or 2x2 table Odds = Hazard ratio (HR) Used for time to event analysis Calculated from a Cox Proportional Hazards model to compare relative event rates Commonly combined with Kaplan-Meier curves and statistics Other Common Measures

  15. Used to express the statistical variation (random error) that underlies our estimate Interpretation: “If the level of confidence is set to 95%, it means that if the data collection and analysis could be replicated many times and the study were free of bias, the confidence interval would include within it the correct value of the measure 95% of the time.” Measures of variability - 95% Confidence Intervals

  16. Case Study – Interpreting results of the VIGOR trial

  17. “During a median follow-up of 9.0 months, 2.1 confirmed gastrointestinal events per 100 patient-years occurred with rofecoxib, as compared with 4.5 per 100 patient-years with naproxen (relative risk, 0.5; 95 percent confidence interval, 0.3 to 0.6; P<0.001).” What was the measure of frequency? What was the measure of association? What was the reference group? How do we interpret these values? From the VIGOR abstract:

  18. “The incidence of myocardial infarction was lower among patients in the naproxen group than among those in the rofecoxib group (0.1 percent vs. 0.4 percent; relative risk, 0.2; 95 percent confidence interval, 0.1 to 0.7)” What was the measure of frequency? What was the measure of association? What was the reference group? How do we interpret these values? From the VIGOR abstract:

  19. When using updated MI data: Incidence Rate 7.4 events per 1,000 person-years (rofecoxib) 1.5 events per 1,000 person-years (naproxen) Incidence Rate Difference 5.9 MI events caused by rofecoxib for every 1,000 patients treated Number Needed to Harm with rofecoxib 1,000/5.9 = 170 We will harm 1 patient for every 170 patients that are treated with rofecoxib Public Health Impact of VIGOR

  20. Writing an Abstract

  21. An abbreviated and accurate representation of the paper contents The purpose of the abstract is to give the reader the essence of the research done • We are also trying to entice the reader to want to know more (i.e., read the rest of the manuscript, accept the manuscript for presentation at a meeting) It is often the only part of the article that is read by editors and readers On average you get about 250 words What is an abstract and why do we write them?

  22. The format and length required for abstracts differ from journal to journal (and conference to conference) Aim (or objectives, background, etc): State the rationale and/or hypothesis being tested or the procedure being evaluated in one to two sentences. (Why was the study done?) Methods: Briefly state what was done, including study design, population, and main exposure and outcome(s). The sample size should be included. State main analytic method(s). (What was done and how was it done?) Results: Provide the findings of the study, including indicators of statistical significance. The data should include actual numbers as well as percentages. (What was found?) Conclusions: Summarize in one or two sentences the conclusion made on the basis of the findings. (What was concluded?) Essential elements of an abstract

  23. Start with the conference/manuscript writing guide! Provide only main objectives, methods, results, and conclusions • Space is very limited (you can use abbreviations to save characters, but limit number) • Do not list results that are not stated in the methods • Statistics used is optional - I typically describe statistic used to assess primary outcome (e.g., Cox proportional hazards models were used to ….) Give just enough information in the abstract to make the reviewers interested in hearing more • A reviewer will be unable to evaluate the quality of the research from an abstract alone • Flow is very important Abstract must be stand alone Very rare to submit an abstract with results pending now that you are not a student (most conferences will not accept and I don’t recommend) Manuscript – not conference abstract – should be your goal Tips for writing a abstract

  24. QUESTIONS?

More Related