Study analysis · BMJ Open Sport & Exercise Medicine · 2023

A 7-year nutrition program for college runners failed its main goal - but may have cut a specific type of bone injury by 44% relative. The catch? It only worked at one school.

Teaching female college runners to eat enough for their training did not clearly reduce overall bone injuries, but it may have cut spongy-bone injuries by 44% relative, and it seemed to work at one school but not the other.

Reading level
Low certainty
Level 2b · Individual cohort studyAssociation, not causationNo causal claims

Overview

What the study found

The study in plain English — the bottom line, every takeaway we extracted, and what to do with them.

In simple terms

This study followed female college runners over several years to see if nutrition education changed how often they got bone injuries. Because the runners were not randomly assigned to get the education or not, and the comparison was to past years instead of a control group, we can't say the education caused the change. We can only say it might be linked to fewer certain types of bone injuries.

What’s the bottom line?

Researchers studied female college distance runners for 7 years. They taught teams about eating enough energy and bone-building nutrients. They compared injury rates before and during the program. Overall, bone stress injuries did not significantly drop, but some types of injuries and one school did show drops. Note: this study has corrections/errata; check notices.

How strong is this study?

The study is decent because it lasted many years and used medical records to track injuries, but it has some big weaknesses. There was no control group, and things like team culture and coaching changes could have affected the results. Also, the number of runners was smaller than planned, so the results are not very certain.

Reporting

40 / 100

  • COI disclosure+40/40
  • Data availabilitydata not shared
  • Code availabilitycode not shared
Methodology

25 / 100

  • Randomizationnot randomized
  • Blindingblinding unclear
  • Control groupno control group
  • Sample size (n=134)+9.8/20
  • Follow-up+10/10
Publication

100 / 100

Statistical

77 / 100

  • P-values+15/15
  • Effect size+20/20
  • Confidence intervals+15/15
  • Pre-registrationnot pre-registered

Each component is scored out of 100 and then capped by the study design — a case series cannot reach the ceiling a randomised trial can, however well it is reported.

Where it sits

RCT reviews

Max 100

Randomized Trials

Max 90

Reviews of Cohort Studies

Max 85

Cohort Studies

Max 72

Reviews of Case-Control Studies

Max 63

Case-Control Studies

Max 58

Cross-Sectional & Case Series

Max 50

Expert Opinion

Max 5
StrongerWeaker
Cohort Studies
Level 2b
56

56 / 100

Probability of being correct

Groups of people are followed over time to see who develops an outcome. Strong for identifying risk factors and associations, but cannot prove causation as firmly as RCTs.

This design cannot establish causation — the findings describe an association, not a cause. Observational design with no randomization, no concurrent control group, and historical comparison. Potential confounding by team culture, coaching changes, secular trends, and other unmeasured factors. Cannot rule out chance, especially given post hoc analyses and multiple comparisons.

No Conflicts

No conflicts of interest identified

No conflicts of interest or industry funding were identified in the provided text. A grant is mentioned but no funder name or role is disclosed.

The provided text appears truncated before any declarations/COI section. No author affiliations or industry ties are listed. The smartphone app 'Run Fueled' was developed by the study team, not a commercial entity.

Key takeaways

  1. 01

    Overall bone stress injuries: 0.52 per person-year before vs 0.43 during (p=0.34; absolute difference 0.09 fewer per person-year).

  2. 02

    Trabecular-rich injuries: 0.18 to 0.10 per person-year (absolute reduction 0.08; 44% relative decrease; p=0.047).

  3. 03

    Cortical-rich injuries: 0.31 to 0.35 (p=0.87).

  4. 04

    School 1 overall: 0.63 to 0.27 per person-year (absolute reduction 0.36; p=0.041).

  5. 05

    School 2 overall: 0.33 to 0.56 (p=0.20).

  6. 06

    In plain terms, overall there were about 52 injuries per 100 runners per year before and 43 per 100 during the program — about 9 fewer per 100 runners per year, but this was not statistically significant.

  7. 07

    Trabecular-rich injuries fell from about 18 to 10 per 100 runners per year (about 8 fewer per 100 per year).

  8. 08

    At School 1, overall injuries fell from about 63 to 27 per 100 runners per year (about 36 fewer per 100 per year), but School 2 did not improve.

  9. 09

    The absolute risk reduction was not reported as a percentage of people; these are event rates per person-year.

Surprising findings

  • Overall bone stress injury rates did not significantly drop, even though the intervention focused on a well-established risk factor: low energy availability.Many people assume that fixing underfueling will quickly and clearly reduce injuries. This study suggests the relationship is more complicated and may take years or additional changes to show up.
  • At one institution, BSI rates actually went up during the intervention (0.33 to 0.56 events per person-year), though not significantly.An intervention designed to help athletes would not be expected to coincide with higher injury rates. This highlights how team environment and implementation can overwhelm the intended effect.
  • A post hoc subgroup showed a 44% relative reduction in trabecular-rich BSI, while the primary outcome was null.It is unusual for a subgroup to show a clear signal when the overall trial misses its primary endpoint. This can be a true biological effect or a chance finding from multiple comparisons.

Practical takeaways

For coaches and athletes, prioritize adequate energy availability and bone-building nutrients, especially for runners at high risk of the Female Athlete Triad.

This study does not prove that nutrition education reduces overall BSI rates. The overall result was null, and the significant findings were post hoc and underpowered.

medium-low confidence

Pay attention to team culture and staff stability when implementing nutrition programs. Stable coaching and dietitian staff may make a difference.

This is an observational inference from a two-site study with no control group. The institution difference could be due to chance or unmeasured factors.

low-medium confidence

Don't expect quick fixes. Bone changes take time, and the study at Institution 1 showed a linear decline over the intervention years.

The study was cut short by COVID-19 and may not have been long enough to see full effects. The sample was underpowered.

medium confidence

Check for published corrections or errata before citing this study. The summary notes that corrections exist.

Always verify the latest version of the paper and any correction notices to ensure you are using accurate information.

high confidence

Why this study matters

The primary outcome flopped

Overall bone stress injury rates were 0.52 events per person-year before the program and 0.43 events per person-year during it. That is an absolute difference of 0.09 fewer events per person-year, or about a 17% relative reduction, but it was not statistically significant (p=0.34). The study was underpowered, with 78 runners instead of the target 96.

It challenges the simple idea that teaching athletes to eat more automatically prevents injuries. Sometimes the headline is 'no clear effect' even when there is a promising signal.

Spongy-bone injuries dropped significantly

In post hoc analyses, trabecular-rich BSI rates fell from 0.18 to 0.10 events per person-year. That is an absolute reduction of 0.08 events per person-year, or a 44% relative decrease, and it was statistically significant (p=0.047). These injuries occur at sites like the calcaneus, femoral neck, sacrum, and pelvis.

Trabecular bone is more metabolically active and sensitive to low energy availability, so nutrition may preferentially protect these high-risk sites.

Location, location, location: one school won, one didn't

At Institution 1, overall BSI rates dropped from 0.63 to 0.27 events per person-year (absolute reduction 0.36; about a 57% relative reduction; p=0.041). At Institution 2, rates went from 0.33 to 0.56 events per person-year (absolute increase 0.23; p=0.20). The interaction between institution and phase was significant (p=0.009).

The same intervention produced opposite patterns at two schools, suggesting team culture, staff stability, and existing resources may be as important as the nutrition advice itself.

Cortical-rich injuries didn't budge

Cortical-rich BSI rates were 0.31 events per person-year before the program and 0.35 during it (p=0.87). That is an absolute increase of 0.04 events per person-year, which is small and not statistically significant.

It suggests that nutrition education alone may not be enough for injuries in the hard outer shell of bone. Different bone types may need different prevention strategies.

Underpowered and cut short by COVID

The study enrolled 78 runners in the intervention phase instead of the planned 96, and it ended early because of the COVID-19 pandemic. The authors note this may have limited the ability to detect a true effect.

A null result doesn't always mean 'no effect' - it can mean 'not enough data to see the effect.' This is a key nuance for interpreting health research.

Want the whole report?

Detailed mode opens the full scientific breakdown — every score component, the methodology, conflicts of interest, the evidence analysis behind each claim, and the raw study data.

Standing

The people behind it

The researchers who wrote the study this analysis is built on.

Authored by

15 researchers

If this is your work, this is how we attribute it on Fit Body Science. Michael Fredericson is listed as the lead author.