How to write an abstract

The importance of a well-written abstract can hardly be overstated. Evaluators, editors or judges invariably turn first to the abstract to grasp the intent and content of a presentation or paper. The abstract should be comprehensible, apart from the presentation or text, and should be able to stand alone.

An abstract needs to include: 1) title, 2) authors’ name(s), 3) the objectives of the investigation, 4) the experimental methods used, 5) the essential results, including data and statistics, and 6) conclusions.

The specific format for an abstract is often specified by the requesting organization, but some typical examples are provided as guides:

Example 1

Dentist Behaviors Which Reduce Patient Anxiety – Belief and Reality.
N.L. CORAH* and R.M.O’SHEA State University of New York at Buffalo, Buffalo, NY

Previous research on patient anxiety has shown a significant relationship between certain dentist behaviors and patient anxiety reduction. This study assessed dentist’s beliefs about the relative efficacy of their own behavior. A total of 605 dentists in general practice were asked to respond to a list of 25 dentist behaviors by rating the importance of each behavior for its effectiveness in dealing with patient anxiety. They were also asked to select the three most effective behaviors. Two indices of dentist perceived importance of their behaviors were derived and each compared with the actual effectiveness as determined from a sample of 250 adult dental patients. Although the majority of dentists rated 18 of the 25 behaviors as being very effective, there was no significant association between dentist perceived effectiveness and actual effectiveness. The implications for dentist training are discussed.

This study supported by NIDR Grant No. DE 04494.


Example 2

Tuberculosis Risk Factors in Adults in King County, Washington, 1988 through 1990.
Susan E. Buskin, PhD, MPH, James L. Gale, MD, MS, Noel S. Weiss, MD, DrPH and Charles A. Nolan, MD

Objectives. Tuberculosis has become a resurgent public health problem in the United States. Because resources are limited, control programs frequently must target populations at greatest risk. The purpose of the study was to examine risk factors for tuberculosis in adults.
Methods. In King County, Washington State, from 1988 through 1990, the characteristics of patients with tuberculosis were compared with census data, and a case-control study was conducted. Self-administered questionnaires were completed by 151 patients with active tuberculosis and 545 control subjects.
Results. Infection with the human immuno-deficiency virus, non-White race/ethnicity, and foreign birthplace were each associated with a sixfold or greater increase in risk. Each of the following was associated with at least a doubled risk: history of selected underlying medical-conditions; low weight for height; low socioeconomic status; and age 70 years or older. Men had 1.9 times the risk of women, smokers of 20 years or more duration had 2.6 times the risk of nonsmokers, and heavy alcohol consumers had 2.0 times the risk of nondrinkers.
Conclusions. Intervention targeting easily identified groups may be an effective way to reduce the incidence of tuberculosis. (Am J Public Health. 1994;84:1750-1756)


Example 3

Factors Associated with Changes in Untreated Bounded Edentulous Spaces Over 15 Years.
DM Singer, J Farhat, EA Krall, DA Sugars, RI Garcia
(Boston VA Outpatient Clinic, Tufts Univ., Univ. North Carolina, Boston Univ. School of Dental Medicine)
The assumption that teeth adjacent to untreated bounded edentulous spaces (BES) will shift is the primary reason for recommending replacement of a missing tooth. However, there is a paucity of data on the longer-term outcomes of untreated BES. We assessed changes over time (up to 15 years) in the distance between teeth adjacent to untreated posterior BES, using intraoral radiographs previously obtained from subjects (N=145) in the VA Dental Longitudinal Study, a close-panel observational study begun in 1968. Subjects were examined triennially. They are not VA patients and receive all dental care in the private sector. For this analysis, subjects were selected based on having an untreated posterior BES, at baseline or at a subsequent exam, and with readable radiographs available for a minimum of 3 consecutive exams. Measurements on radiographs were made using an electronic digital caliper to determine untreated BES distance between mesial and distal abutments, degree of abutment tipping, and status of the opposing arch. We found that teeth adjacent to untreated BES present in subjects at study baseline experienced a mean loss of 0.002 mm (SD=0.95), over a mean follow-up time of 15 years. Subjects who developed BES after entry into the study experienced a mean loss 0f 1.36 mm (SD=1.21) in abutment space over 9.5 years. A significant correlation was found between the change in abutment distance and mean bone loss, whether the tooth was present to begin with, the initial distance between the abutments, and the number of years between first and last measurements (p=0.0001). Using a regression model we can predict that 42.6% of the variability in BES change in width distance is explained by distal abutment tipping status, whether the tooth was present to begin with, and smoking status. In conclusion, subjects who lost a tooth during the study period demonstrated significantly more (p<0.0001) space loss over follow-up time compared to subjects who entered the study with pre-existing BES. However, even in the incident BES cases the majority of subjects experienced <1.4 mm change over time. That baseline BES experienced less change over time than incident BES cases suggests a ‘survivor’ effect or plateau effect may be at play, explaining in part the relative stability of baseline BES over the follow-up period. These results show that certain factors may be responsible for a slight shift in teeth adjacent to untreated BES as well as raise questions about the validity of the assumption that teeth adjacent to untreated BES will invariably and significantly shift.
Supported by VA HSR&D Services and NIDR (DE11878, DE07220, DE07268)