The use of "spin" in laparoscopic lower GI surgical trials with nonsignificant results: an assessment of reporting and interpretation of the primary outcomes.

BACKGROUND

Spin has been defined as "specific reporting that could distort the interpretation of results and mislead readers."

OBJECTIVE

The purpose of this study was to identify how frequently, and to what extent, "spin" occurs in laparoscopic lower GI surgical trials with nonsignificant results.

DATA SOURCES

Publications were referenced in MEDLINE and EMBASE (1992-2012).

STUDY SELECTION

Randomized controlled trials comparing laparoscopic with open surgical technique in lower GI surgery were sought. Trials were included if a nonsignificant (p > 0.05) result of the primary outcome(s) occurred.

INTERVENTION

The laparoscopic versus open technique in lower GI surgery was studied.

MAIN OUTCOME MEASURES

Trials were assessed for frequency, strategy, and extent of "spin," as previously defined.

RESULTS

Fifty-eight trials met the inclusion criteria. Sixty-six percent of these trials had evidence of "spin." In general, authors used significant results only (one of multiple primary outcomes, secondary outcomes, or subgroup analyses) (43%) or interpreted nonsignificance as equivalence (43%). Trials with spin were more likely to recommend the laparoscopic approach over the open technique (p < 0.001), were less likely to call for further trials (p = 0.003), and were less likely to acknowledge the nonsignificant differences (p < 0.001). Inadequate randomization was associated with decreased odds of spin (p = 0.03), as was an intent-to-treat analysis (p < 0.0001), whereas inadequate allocation concealment (p = 0.06) was weakly associated with a decrease in spin. No other a priori candidate risk factors were associated with the presence of spin.

LIMITATIONS

Funding source was rarely described, so the association between industry funding and spin could not be assessed.

CONCLUSION

The distortion of nonsignificant results in laparoscopic trials was highly prevalent in this review. Readers of trials with nonsignificant results should be cautious of the authors' interpretations. Editors, reviewers, and publishers should ensure that author's conclusions correspond to the study's results and design.