Quality of Abstracts of Papers Reporting Original Cost-Effectiveness Analyses

• Published in: Medical decision making 2005-07, Vol.25 (4), p.424-428
• Main Authors: Rosen, Allison B., Greenberg, Dan, Stone, Patricia W., Olchanski, Natalia V., Neumann, Peter J.
• Format: Article
• Language: English
• Subjects: Abstracting; Cost-Benefit Analysis; Guidelines as Topic; Peer Review, Research
• ISSN: 0272-989X; 1552-681X
• DOI: 10.1177/0272989X05278932

Background . Although many peer-reviewed journals have adopted standards for reporting cost-effectiveness analyses (CEAs), guidelines do not exist for the accompanying abstracts. Abstracts are the most easily accessed portion of journal articles, yet little is known about their quality. The authors examined the extent to which abstracts of published CEAs include key data elements (intervention, comparator, target population, study perspective) and assessed the effect of journal characteristics on reporting quality. Methods .Systematic review of the English-language medical literature from 1998 through 2001. The authors searched MEDLINE for original CEAs reported in costs per quality-adjusted life years(i.e., cost-utility analyses). Two independent readers abstracted data elements and met to resolve discrepancies. Results . Among the 303 abstracts reviewed, a clear description of the intervention was present in 94%, comparator in 71%, target population in 85%, and study perspective in 28%. All 4 data elements were reported in 20% of abstracts, 3 elements in 49%, 2 in 22%, and 0 or 1 in 9%. In journals with CEA-specific abstract reporting requirements, structured abstract requirements, or impact factors ≥ 10, significantly more data were included in abstracts than in journals without these features (P < 0.01 for all comparisons). Conclusions . Abstracts of published CEAs frequently omit data elements critical to proper study interpretation. An explicit core set of reporting standards is needed, based on the standards by the US Public Health Service’s Panel on Cost-Effectiveness for reporting of CEAs, but specific to the accompanying abstracts.