Crossing the zero line: A scientometric analysis of surgery's transition from disruptive to consolidating science

Scientific research is becoming less disruptive across disciplines. We characterized temporal trends in surgical innovation metrics over 5 decades.

We analyzed 168 million (M) publications (1970-2020) from the SciSciNet database across 5 fields: surgery (2.4M), medicine excluding surgery (41M), biology (34M), mathematics (23M), and computer science (66M). For each field, we computed decade means of 3 validated indicators: the disruption index (CD5), which distinguishes work that opens new directions (positive) from work that reinforces existing knowledge (negative); novelty (N; Uzzi atypicality), measuring the extent to which a publication cites unusual, rarely co-occurring combinations of journals; and the Sleeping Beauty Index, identifying publications with delayed recognition.

In the 1970s, surgery had the highest CD5 (+0.0049) of any field; by the 2010s, it had the lowest (-0.0017) and was the only field in net-negative territory, having first crossed zero in the 2000s, a 135% decline. All fields declined significantly (P < .05), but surgery's slope was steepest (-0.00154/decade, R² = 0.900; P = .014); computer science retained the highest disruption (+0.0019). Paradoxically, surgery showed the largest novelty increase (+63%) despite the steepest disruption decline. The Sleeping Beauty Index fell across all fields (-46% to -78%), consistent with accelerated information dissemination.

Surgery has transitioned from the most disruptive to the only net-consolidating field examined, becoming more creative in combining ideas, yet less likely to produce work that fundamentally redirects the field. Whether structural features of the surgical research enterprise drive this trajectory is an important question for consideration.