Society of Critical Care Medicine
Section of Clinical Pharmacy and Pharmacology

How to Write an Abstract Suitable for Publication
Hosp Pharm 2004:39:92-95
Keith M. Olsen, Pharm.D., FCCP
Sandy Kane, Pharm.D.

Abstracts submitted to a medical or pharmacy meeting or as part of a scientific article, have often been terse statements describing the research question. The outcome is often a poorly written, rambling abstract that may result in rejection or impair the evaluation of the scientific reviewer. Because many of the same mistakes in abstract writing are repeated from year to year, the Clinical Pharmacy and Pharmacology Section of the Society of Critical Care Medicine (SCCM) has charged the research committee develop a "How-to-Guide" to writing better abstracts for submission to the annual SCCM Symposium.

As authors of over 110 abstracts, we suggest following the steps below to increase the chance for abstract acceptance. We make the assumption in writing your abstract; the data is sound and with appropriate supporting evidence in study design. Thus, writing a clear concise abstract significantly enhances the chances for acceptance. Since abstracts are brief in nature, but highly informative, we have chosen to use brevity as well. However, this should not diminish from the content describe herein and following these simple rules increases the acceptance rate of an abstract.

What is and abstract?
An abstract is a brief, but comprehensive summary of the contents of a scientific investigation or scientific presentation. The latter could be an invited presentation at an SCCM symposium that requires submission of an abstract. Remember the word brief; you generally cannot present the entire research project within the word allotment.

How do I get started?
This a common question that often perplexes even the experience writer, unless a specific plan is implemented. To overcome the writer's block or in this case the starting (stumbling block) block mentality, we strongly encourage the use of a structured abstract. Although most abstracts follow the heading outline in the structured abstract, many authors do not use the structured titles. The structured abstract was created not only make the information more informative, but to also facilitate the writing. First review the structured abstract below, then some specific suggestions in writing a clear concise document is provided.

Objective/Purpose/Introduction (10% of abstract): 
The preference is to write one sentence containing a specific purpose statement. Too often writers make the mistake of writing several introductory sentences before stating a purpose. Avoid this unnecessary material. At most, the introduction may include one background statement (one sentence) and then the purpose or specific objective (second sentence).

Design or Methods (30-40% of abstract): 
Describe the study design employed, the study population, the dose or dosage of the drug if used, length of treatment, other collected data like lab tests (if presented in the results), statistics employed. Do not put results in methods or a method in the results. Be clear and concise.

Results (30-40% of abstract): 
What are the most important findings of the research, any specific adverse events or safety results, statistical data (p values)?

Conclusions (10% of abstract): 
List the most important points that the investigators have learned; what are the recommendations. This should be accomplished in one or two sentences. Remember the purpose when writing the conclusion; do not just restate the results.
_____________________________________________________________

Now, before writing, ask yourself the four WHAT'S of writing an abstract....

• What is the purpose of this study?
• What was done?
• What was found, identified or learned?
• What is the importance or implications of these findings?

What makes a good abstract?
Accurate - reread the abstract. Does the abstract reflect the purpose stated in the introduction.

Self-contained - define all abbreviations. Do not include methods that you do not provide results for. Make sure every result has a corresponding method. Use generic names for drugs. You may use common abbreviations, but try to define as many as possible.

Concise and Specific - Be as brief as possible, but still deliver your message. Remember most abstracts range from 150 to 250 words.

Coherent writing - write in a clear manor. Authorities suggest using verbs rather than noun equivalent and the active rather than the passive. Use the present tense to describe results and past tense to describe specific variables manipulated or tests applied.

Tables or Figures - with computer technology and the online submission of abstracts, the inclusion of tables and figures has become more common. If you have very detailed data, a table may be the best way to express them. Be sure to check abstract guidelines to confirm size and format.

A Check List of Do's and Don't

• Background no more than two sentences including the purpose
• Abbreviations defined on their first use
• Every result has a corresponding method
• Dose, dosage, population studied, duration or time period, other measurements to be included in the results, and setting of the research are included
• Dose/dosage contains units and time duration
• Conclusion extracts the one or two most important point
• If table included, clear and readable
• Only one table included
• Don't present work you are going to do or that should be done
• Don't overuse familiar terms or acronyms
• Don't add a statement in conclusion that is a stretch of the results or not supported by the data
• Don't use a lot of space reporting negative findings (negative abstracts are always harder to get accepted)
• Don't repeat or paraphrase title in purpose of abstract (common mistake)
• Purpose just doesn't repeat title but indicates the scope of the research (see both examples below of how to avoid this trap)
• Have used present tense throughout abstract, especially conclusion
• No jargon is included that may confuse the reader
• References - you may include a reference (abbreviated) if highly relevant to the current research. This is not the norm, however, and if in doubt whether to use, don't.
• Avoid abbreviations in title if possible
• Authors included had a specific contribution to the study
• Have an independent researcher review the abstract; let the abstract sit for a day or two, then edit again.

*Below are two example abstracts that were accepted to scientific     meetings. Notice the second abstract with only one sentence in the introduction. A very specific statement of the purpose was all that was needed. Also, notice in both abstracts the purpose is a specific statement of the research rather than a repeat of the title.

Notice the abbreviations are all defined early, there is a method for every result, and finally clear concise conclusions.

The Pharmacodynamic Activity and Efficacy of Linezolid in a Rat Model of Pneumococcal Pneumonia. K.M. OLSEN1, L.C. PREHEIM1,2,3, M.J. GENTRY-NIELSEN1,2,3. Univ. of Nebraska Medical Center1, Creighton Univ.2 and Veterans Affairs Medical Center3, Omaha, NE

Background: Linezolid is a new oxazolidinone with potent activity against gram-positive cocci, including Streptococcus pneumoniae. We determined the in vivo efficacy and pharmacodynamic (PD) activity of two doses of linezolid in comparison to ceftriaxone or a PBS placebo.

Methods: Rats infected by intrapulmonary instillation of 8 x 107 cfu of penicillin-sensitive S. pneumoniae (ATCC 6303) were treated for 5-days beginning 18 h post-infection. Groups of 12 rats received oral PBS, oral liquid linezolid (MIC 0.75 mg/L) at 25 or 50 mg/kg q12h, or subcutaneous (sub-Q), ceftriaxone (MIC 0.016 mg/L) at 100 mg/kg q24h. Mortality was followed for 10 days post-infection, with blood cultures performed on day 1 (pretreatment) and days 3, 5, and 10 post-infection. Serum was collected for pharmacokinetic (PK)/PD parameters at 1, 3, 5, and 12 h post-dose.

Results: The cumulative mortality for the PBS group was 100%, ceftriaxone 0%, linezolid (50 mg/kg) 8.3%, and linezolid (25 mg/kg) 58.3% (p < .05 for low-dose linezolid vs. other groups). Bacteremia mirrored mortality, with negative cultures obtained in 0%, 100%, 83% and 50% of rats, respectively, by day 5 post-infection. All PK/PD parameters reflected efficacy, but the strongest predictor of outcome was time drug concentration exceeded MIC (T > MIC) > 45% of the experimental dosing interval.

Conclusions: Oral Linezolid 50 mg/kg q12h was as effective as sub-Q ceftriaxone 100 mg/kg q24h in experimental pneumococcal pneumonia. T > MIC ratio was the best predictor of outcome when the serum concentration exceeded the MIC for at least 45% of the dosing interval.

A Randomized, Cross-Over Study of Duodenal or Jejunal Compared to Nasogastric Administration of Omeprazole Suspension in Critically Ill Patients
Jeffrey O. Phillips1, Keith M. Olsen2, Jill A. Rebuck2, Michael H. Metzler1
1Department of Surgery, School of Medicine, University of Missouri-Columbia
2Department of Pharmacy, College of Pharmacy, University of Nebraska Medical Center

Purpose: To characterize absorption and pH control of simplified omeprazole suspension (SOS) 2mg/mL in 8.4% sodium bicarbonate administered via the nasogastric versus jejunal or duodenal route.

Methods: Nine critically ill surgical patients, NPO, mechanically ventilated were enrolled in this randomized, cross-over study. Patients received a single dose 40mg SOS by nasogastric and either the jejunal or duodenal route. Twenty-four hour continuous intragastric pH monitoring was performed during the study period. Sequential blood samples were collected over 24-hours to characterize SOS absorption and the pharmacokinetic parameters.

Results: Nasogastric administration of SOS resulted in lower maximum mean ± SD serum concentrations compared to jejunal/duodenal dosing (0.970 ± 0.436 vs. 1.833 ± 0.416 :g/mL, p = 0.006). SOS absorption was significantly slower when administered via nasogastric tube (108.3 ± 42.0 vs. 12.1 ± 7.9 minutes, p < 0.001). However, all routes of administration resulted in similar SOS area under the serum concentration-time curves (AUC0-¥; 415.1 ± 291.8 vs. 396.7 ± 388.1 :g·hr/mL, p = 0.91). Mean intragastric pH values remained above 4 one-hour post SOS administration and remained greater than 4 for the entire 24-hour study (6.32 ± 1.04, 5.57 ± 1.15, nasogastric vs. jejunal/duodenal, p = 0.015), regardless of administration route.

Conclusions: In critically ill surgical patients, pharmacokinetic parameters and subsequent pH control following the administration of SOS are similar by the jejunal, nasogastric, or duodenal route. SOS suspension offers an alternative acid control measure when patients are unable to take oral medications, yet have an enteral tube in place.

References:

1. Fitzgerald JT, Smith HM. Word for Word, Ann Arbor, Michigan. 1991.
2. Instructions for preparing structured abstracts. JAMA 1991;266: 42-44.
3. Rennie D, Glass RM. Structuring abstracts to make them more    informative. JAMA 1991;266:116-117.

Below is a summary of the scoring sheet utilized by reviewers for the 2002 SCCM Symposium in San Diego. This only an example to guide reviewers in the decision making process. This scoring system is subject to year-to-year change.

Instructions for Grading

Submit a numerical grade for each abstract listed on either the On-line or Off-line grading sheet. As a separate attachment, a guide to numerical scoring of specific sections of the abstract is provided. You are not required to submit a score sheet of each graded abstract; it is intended to serve as a guide. If you do not consider yourself qualified to judge an abstract, please do not ask a colleague to grade it for you.

If an abstract would be more appropriate in another category or section, please indicate where it should be placed in COMMENT. For abstracts graded below 6 (six), please indicate in the COMMENT box a reason (s) for rejection. This information will be provided to the authors as an educational tool.

Note: We have provided a check box titled "newsworthy"; if you find an abstract that you believe is worthy of media attention, please indicate that by checking this box.

Each abstract must be given a NUMERICAL grade based on the following criteria:

Note: If an abstract is not scored, the scores assigned by the remaining graders will be averaged.

For questions concerning grader instructions, contact MSS Customer Service at ( ).

General guide for abstract scoring. Please note that some abstracts may be important and excellent but due to subject matter may not be structured in the following manner.

The following general guidelines will be used as the schema for evaluating each abstract on a 0-10 scale, with the best abstract having a 10 value.


Introduction/background                                   1 point

Hypothesis                                                     2 points total
   Stated/implied                                                1 point
   Quality assessment Novel                               0.5 point
   Important                                                       0.5 point

Methods                                                        3 points Total
   What was done/how appropriate?                    2 points
   Statistical methods/analysis                           1 point

Results                                                          2 points total
   Sufficient data presented
   Adequate data
   Relationship to hypothesis

Conclusions                                                   2 points
   Summary/Stem from data/nor overstated
   Future analysis of data unacceptable if not
   substantiated by results
                                                                     10 Total for "ideal"
                                                                     abstract

For abstracts with a score of below six, a reason for rejection should be included under COMMENTS.
Reason for rejection: indicate as many as appropriate
1. No hypothesis
2. Methods not stated/unclear
3. No summary of essential results
4. No conclusion
5. No new information
6. Lack of data
7. Numbers in study too small
8. No controls
9. Results don't support data
10. Case report with no unique character
11. Promotional/too commercial in nature
12. Methods are not clear or are inappropriate
13. Study in progress/not completed
14. No abstract text received