REF-1311 · Fokkema et al., 2019 · Journal of Science and Medicine in Sport · Paper · PubMed

Why do new runners quit?

Of 774 beginner-programme runners, 29% had stopped within 6 months, half because of a running injury. Women and those who had never run (vs run in the past year) were likelier to stop. B.

Main caveat: Loss to follow-up. 56% of the cohort did not answer at 26 weeks.
limited evidenceMeasured on non-athletes and recreational people

Fokkema and colleagues followed novice runners from the Dutch Start to Run programme — a 6-week programme run twice a year by the Dutch Athletics Federation, with one group session and one or two individual sessions a week, aiming at 20 min of continuous running.

Of 7660 people who signed up in March or September 2013, 1936 were included in the NLstart2run study; 774 (43.7%) filled in the questionnaire sent 26 weeks after the start (20 weeks after the programme ended). At 26 weeks 546 (70.5%) were still running, on average 98.9 min a week; 225 had stopped (29.1% in the Results, 29.5% in the abstract).

A running-related injury was the most common reason to stop (108, 48.0%), ahead of running not being the preferred sport (13.8%), health issues (12.9%), an injury not related to running (11.6%) and no time (11.6%); 72% of those who stopped intended to start again.

In the multivariable model, being female, being unsure about continuing after the programme and (almost) no alcohol use went with a higher chance of stopping, and running experience in the past year and a higher RAND-36 physical-functioning score with a lower chance.

Responders were older, more often male and more experienced than non-responders, and the authors expect the true discontinuation to be higher — "at least one-third". Our reading: for a beginner, injury is not only lost weeks but the main route out of the sport; the predictors are modest associations from a cohort that lost most of its members.

Protocol

Setting: Start to Run, organised by the Dutch Athletics Federation (organiser, not a funder): one supervised group session plus one or two individual sessions a week for 6 weeks; goal 20 min of running without breaks.
Who: novice runners aged 18-65 who signed up for the March or September 2013 programme and gave digital informed consent; part of the NLstart2run study (ethics: University Medical Center Groningen, no. 2012/350).
Baseline questionnaire (one week before the start): sex, age, height, weight (BMI), smoking, alcohol, physical activity (SQUASH), running experience (none / less or more than a year ago), other sports, earlier running injuries, other musculoskeletal complaints, intention to continue (yes / maybe / no), motivation (BREQ-2, relative autonomy index) and perceived health (RAND-36 subscales).
Follow-up at 26 weeks: still running structurally (yes/no, no definition of distance or frequency), reasons to continue or stop, minutes per week, intention to restart.
Sample: n = 774 returned the follow-up (43.7%); 3 did not say whether they were still running and were excluded.
Analysis: univariable then multivariable logistic regression (enter method); p ≤ 0.05; SPSS 21.
Funding and interests: NLstart2run was funded by the Netherlands Organisation for Health Research and Development (ZonMw), grant 50-50305-98-12001; this study was partly funded by a programme grant of the Dutch Arthritis Foundation. No conflict-of-interest statement appears in the accepted manuscript read. The Dutch Athletics Federation ran the programme; it is not named as a funder.

Key Results

Who stopped, and whyResult
Still running at 26 weeks546 (70.5%); 98.9 (SD 89.7) min/week; 55.7% ran in a group
Stopped running within 26 weeks225 — 29.1% in the Results, 29.5% in the abstract
Reason: running-related injury108 (48.0%)
Reason: running not the preferred sport / health issues / other injury / no time / other31 (13.8%) / 29 (12.9%) / 26 (11.6%) / 26 (11.6%) / 5 (2.2%)
Of those who stopped, intend to restart72%
Predictor of stoppingUnivariable OR (95% CI)Multivariable OR, Results textMultivariable OR, abstract
Female (vs male)1.72 (1.17-2.53)1.68 (1.09-2.59)1.74 (1.13-2.68)
Unsure about continuing after the programme2.11 (1.40-3.20)2.06 (1.31-3.24)2.06 (1.31-3.24)
(Almost) no alcohol vs up to three times a week1.76 (1.23-2.51)1.61 (1.10-2.36)1.62 (1.11-2.37)
Running experience less than a year ago (vs none)0.55 (0.34-0.90)0.54 (0.30-0.98)0.46 (0.26-0.83)
RAND-36 physical functioning, per point0.97 (0.96-0.99), labelled "physical fitness"0.98 (0.96-1.00)0.98 (0.96-0.99)
Who answered: compared with non-responders, responders were more often male (24.9% vs 19.3%), older (44.6 vs 42.1 years), had a lower BMI, more running experience (43.2% vs 36.8%) and more other sports (46.9% vs 39.9%), and more often a past running injury (20.3% vs 15.1%).
Numbers that do not add up:
225 of 774 is 29.1% (and 225 of 771 is 29.2%; our arithmetic); the abstract and Discussion give 29.5%, which is 100% minus the 70.5% still running.
The multivariable odds ratios differ between the abstract and the Results text of the accepted manuscript (female 1.74 vs 1.68; recent running experience 0.46 vs 0.54); the published tables were not available to settle which is right.
774 is 43.7% of 1772, the number who completed a baseline questionnaire in the same cohort (REF-1312); it is 40.0% of the 1936 "included" in this paper's Results (our arithmetic).
Injury and non-injury quitters: a model restricted to those who stopped for reasons other than a running injury gave results similar to the full model (authors).

Mechanisms

Measured: self-reported reasons for stopping and baseline characteristics associated with stopping.
Authors' interpretation: injuries are a considerable problem among novice runners and make restarting hard; running courses should teach injury prevention and how to restart after an injury; a longer programme might lower discontinuation; novices with low perceived physical functioning might first improve fitness, for example by walking (their suggestion, untested).
Authors' interpretation: the factors associated with stopping were not those expected from injury: a model restricted to those who stopped for reasons other than a running injury gave results similar to the full model. Inference: who quits and why are partly separate questions; injury was the stated reason for half of those who stopped.
Inference: the time of stopping was not recorded, so injury and other reasons act as competing risks the study could not separate (the authors' limitation).

Practical Application

Why do new runners quit? Of 774 beginner-programme runners, 29% had stopped within 6 months, half because of a running injury. Women and those who had never run (vs run in the past year) were likelier to stop. B.
For a beginner, preventing injury is how the habit survives (B, novice cohort): the base's measured lever is the size of each step — no single run far beyond the past month's longest (REF-1102) — and the beginner build in SYNTH-75 (frequency, then duration, then intensity; practice). Tolerance-first preparation phases have not been shown to cut injuries (REF-1307, REF-925).
Plan the restart when an injury comes: 72% of those who stopped meant to start again. Not from this paper; the base's own practice: the return is gated by symptoms, not dates (SYNTH-73, SYNTH-36).
Ask one question at the start: "Do you plan to keep running after this programme?" Being unsure roughly doubled the odds of stopping (OR about 2.1; association). Not from this paper; the base's own practice: agree the next goal before the beginner programme ends.
Women and those who had never run stopped more often than those who had run in the past year (association, not cause): a reason to check in more, not to plan differently.

Limitations

Loss to follow-up. 56% of the cohort did not answer at 26 weeks; responders were older, more often male, fitter by self-report and more experienced, and those still running were probably keener to answer, so discontinuation is probably underestimated (authors).
Self-report. Continuation had no definition of distance or frequency; injury and reasons were self-reported and could be recalled differently; running and injury history are open to recall bias.
No time of stopping. A competing-risks analysis was not possible.
Reporting. 29.5% vs 29.1%, and two sets of multivariable odds ratios (abstract vs Results text); the tables were not in the manuscript file read.
Population. Participants of one supervised Dutch programme, three in four women; not people who start running alone.
Funding and interests: ZonMw grant 50-50305-98-12001 and a Dutch Arthritis Foundation programme grant; no conflict statement in the manuscript read.
How this entry was checked and changed
Why this level

a large prospective cohort sits in the rubric's B+ band; held at B because 56% of the cohort did not return the 26-week questionnaire and responders differed from non-responders, the reasons for stopping were self-reported without the time of stopping, and the multivariable odds ratios in the abstract differ from those in the Results text of the accepted manuscript read here (its tables are not in that file). Read from the full text (accepted author manuscript in the Erasmus University repository)

Full text (checked 2026-10-27)
Of the 7660 novice runners that signed up for Start to Run in March and September 2013, 1936 runners were included; the 26-week follow-up questionnaire was filled in by 774 participants (43.7%) (full text)
A total of 70.5% (n=546) continued running at 26 weeks, on average 98.9 (SD 89.7) min/week (full text)
In total 225 participants (29.1%) stopped running within 26 weeks; a running-related injury was the most frequently reported reason (n=108; 48.0%) (full text)
Abstract: 29.5% of the novice runners (n=225) had stopped running; being female (OR 1.74; 95% CI 1.13-2.68); previous running experience less than one year previously (OR 0.46; 95% CI 0.26-0.83) (full text)
Results text, multivariable: female OR 1.68 (95%CI 1.09-2.59); previous running experience less than one year ago OR 0.54 (95%CI 0.30-0.98); unsure about continuation OR 2.06 (95%CI 1.31-3.24) (full text)
Of the runners that stopped running, 72% indicated that they intended to start running again in the future (full text)
Both this study and the earlier Start to Run study had a high loss to follow-up (56% and 43%, respectively) (full text)
The NLstart2run study was funded by ZonMW, grant number 50-50305-98-12001, and partly by a program grant of the Dutch Arthritis Foundation (full text)