Pickleball Injury Counts Climbed Fast, But Per-Player Risk Did Not Move in Step
Emergency department-treated pickleball injuries rose sharply across the United States between 2014 and 2023, but the pattern did not prove that the sport became more dangerous for any individual player. The most striking finding from this ecological time series is the gap between absolute burden and rate-based interpretation: the injury count surged, while the injury-to-participation measure stayed statistically flat over the period with available denominator data.
The analysis drew on National Electronic Injury Surveillance System microdata and weighted 1,738 adjudicated cases to an estimated 101,450 injuries nationally. By 2023, estimated emergency department-treated injuries had reached 24,294, compared with 1,391 in 2014, although the earliest estimate was described as highly unstable. A large share of adjudicated cases involved older adults, with 70 percent occurring in people aged 60 to 79 years. That age concentration matters clinically because falls, sprains, fractures, and other musculoskeletal injuries tend to carry greater consequence in older patients, particularly when bone density, balance, or recovery capacity is limited.
Exposure Grew Almost as Fast as Injuries
The denominator changed quickly. From 2017 through 2023, estimated injuries increased 4.0-fold while estimated participation rose about 4.4-fold. On that narrower window, no temporal trend appeared in the injury-to-participation measure, with P=0.59. In practical terms, the available surveillance data cannot separate a larger player base from a higher injury risk per player.
One sentence captures the core problem: more players can produce more injuries even when the game itself is not becoming more hazardous. The study design is useful for describing that tension, but it cannot resolve it. An ecological time series can show the movement of populations, not the individual mechanisms of injury within those populations.
This distinction is especially important for a sport such as pickleball, where rapid adoption can concentrate attention, news coverage, and emergency department visits at the same time. The rise in cases may reflect greater exposure, older beginners entering the sport, more frequent play, or some combination of those forces. It may also reflect changes in help-seeking or coding practices. None of those explanations can be disentangled from the available data alone.
Search Interest Tracked Injury Counts Poorly Once the Time Pattern Was Adjusted
The investigators also asked whether general internet attention could stand in for injury burden. In annual levels, Google search interest for “pickleball” correlated tightly with total pickleball-related injuries, with R=0.96. Wikipedia pageviews tracked even more closely, with R=0.98. At first glance, those correlations look impressive.
They were not specific. Search interest for “sourdough” also correlated with injury counts at R=0.76. That negative control weakens any claim that pickleball search traffic is uniquely tied to injuries. Shared upward drift across years can make unrelated series look tightly linked even when no causal connection exists.
Once trend and seasonality were removed, the association largely dissolved. Monthly residual correlation fell to R=0.084, and year-on-year correlation over 96 months from 2015 through 2023 was only R=0.28. The seasonal patterns also ran in opposite directions. Injuries peaked in February, while searches peaked in July. On 12-month mean profiles, the sourdough negative control fit the injury season at R=0.96, while pickleball ran against it at R=-0.40.
That is an unusual result. A signal that looks persuasive in raw annual totals can collapse when the calendar is taken seriously.
News Coverage Moved With Search, Not With Injuries
General pickleball search interest and pickleball news coverage stayed linked in year-on-year changes during 2017 to 2023, with R=0.64. Their year-on-year correlations with total pickleball-related injuries were much smaller, at 0.25 and 0.20. News and search behavior therefore appear to move together more reliably than either one tracks injury burden.
For clinicians and public health observers, the implication is narrow but practical. Search data may capture attention, enthusiasm, or media exposure, but it is not a dependable substitute for injury surveillance. A surge in online interest does not automatically mean a surge in harm, and a surge in harm does not necessarily follow the same temporal rhythm as public curiosity.
The authors explicitly frame their findings as descriptive, not causal. The observed rise in emergency department-treated injuries is consistent with growth in exposure alone. The available data do not determine whether the sport has become more dangerous per participant.
What the Data Can and Cannot Support
This study adds three clinically relevant points. First, emergency department-treated pickleball injuries rose substantially over the decade, and older adults accounted for a large proportion of adjudicated cases. Second, participation appears to have expanded at a similar pace, which makes the flat injury-to-participation measure plausible. Third, broad internet attention correlated with injury counts in raw annual data but lost much of that relationship after adjustment for trend and seasonality.
The limitations are substantial and should stay in view. The analysis is ecological, so it cannot link an individual player’s exposure to an individual injury. The emergency department data capture only treated injuries, not the full spectrum of minor or self-managed events. Participation estimates come from survey-based industry data, and the injury denominator depends on those estimates. Seasonal behavior, media attention, and changing awareness of the sport may all distort apparent associations.
The preprint therefore supports a restrained conclusion: pickleball injuries are rising in absolute terms, but the available surveillance framework cannot say whether risk per participant has changed. Search volume is not a reliable injury barometer on its own, and parallel upward trends should not be mistaken for proof of a shared cause.