Using Machine Learning to Predict Nonachievement of Clinically Significant Outcomes After Rotator Cuff Repair.

Background: Although some evidence suggests that machine learning algorithms may outperform classical statistical methods in prognosis prediction for several orthopaedic surgeries, to our knowledge, no study has yet used machine learning to predict patient-reported outcome measures after rotator cuff repair. Purpose: To determine whether machine learning algorithms using preoperative data can predict the nonachievement of the minimal clinically important difference (MCID) of disability at 2 years after rotator cuff surgical repair with a similar performance to that of other machine learning studies in the orthopaedic surgery literature. Study Design: Case-control study; Level of evidence, 3. Methods: We evaluated 474 patients (n = 500 shoulders) with rotator cuff tears who underwent arthroscopic rotator cuff repair between January 2013 and April 2019. The study outcome was the difference between the preoperative and 24-month postoperative American Shoulder and Elbow Surgeons (ASES) score. A cutoff score was calculated based on the established MCID of 15.2 points to separate success (higher than the cutoff) from failure (lower than the cutoff). Routinely collected imaging, clinical, and demographic data were used to train 8 machine learning algorithms (random forest classifier; light gradient boosting machine [LightGBM]; decision tree classifier; extra trees classifier; logistic regression; extreme gradient boosting [XGBoost]; k-nearest neighbors [KNN] classifier; and CatBoost classifier). We used a random sample of 70% of patients to train the algorithms, and 30% were left for performance assessment, simulating new data. The performance of the models was evaluated with the area under the receiver operating characteristic curve (AUC). Results: The AUCs for all algorithms ranged from 0.58 to 0.68. The random forest classifier and LightGBM presented the highest AUC values (0.68 [95% CI, 0.48-0.79] and 0.67 [95% CI, 0.43-0.75], respectively) of the 8 machine learning algorithms. Most of the machine learning algorithms outperformed logistic regression (AUC, 0.59 [95% CI, 0.48-0.81]); nonetheless, their performance was lower than that of other machine learning studies in the orthopaedic surgery literature. Conclusion: Machine learning algorithms demonstrated some ability to predict the nonachievement of the MCID on the ASES 2 years after rotator cuff repair surgery.

RunIn3: the development process of a running-related injury prevention programme.

BACKGROUND: Running is an important type of exercise to keep people physically active. However, running also carries a risk of developing running-related injuries (RRI). Therefore, effective and evidence-based RRI prevention programmes are desirable, but are scarce in practice. An approach to face this problem might be the application of methods to develop RRI prevention programmes based on theories of behaviour change. OBJECTIVE: The purpose of the study was to develop an RRI prevention programme based on perspectives of behavioural and social science theories, as well as taking a framework development approach. METHODS: This was a qualitative study using the Intervention Mapping (IM) framework held between February and March 2018 in São Paulo, Brazil. The participants were involved in running practice. The data collection was conducted during focus group meetings. The data analysis was based on semantic thematic approach using a content analysis orientation based on inductive reasoning. RESULTS: The target population of the RRI prevention programme identified was 'adult recreational runners'. The objectives of the RRI prevention programme were established in two broad actions: (1) to provide feedback on individual training characteristics and RRI risk; and (2) provide/enhance knowledge, skills and self-efficacy on RRI preventive behaviours. The programme is aimed to be delivered through an online system. CONCLUSION: An RRI prevention programme was developed using the IM framework and a participatory approach. The programme was named 'RunIn3', and it is based on providing feedback on running volume and RRI risk, as well as providing knowledge, skills and self-efficacy on RRI preventive behaviours.

Can we explain running-related injury preventive behavior? A path analysis.

BACKGROUND: Behavioral and social science theories/models have been gaining attention in sports injury prevention. OBJECTIVE: To investigate the potential of the Theory of Planned Behavior in explaining running-related injury preventive behavior. METHODS: Six-month prospective cohort study based on data gathered from a randomized controlled trial. From a total of 1512 invited trail runners, 232 were included in this study. Preventive behaviors and their determinants were assessed at baseline and two and six months after baseline. Five-point Likert scales were used to assess the determinants of preventive behavior. A Bayesian path analysis was conducted applying mixed models and mediation analysis. RESULTS: A 1-point increase in intention, attitude, subjective norm, and perceived behavioral control predicted an increase of 54% (95% Bayesian credible interval [BCI]: 38, 71) in the rate of performing running-related injury preventive behavior, explaining 49% (R2 0.49; 95% BCI: 0.41, 0.56) of the variance around preventive behavior. Intention and perceived behavioral control predicted running-related injury preventive behavior directly, while 40% (95% BCI: 21, 61) and 44% (95% BCI: 20, 69) of the total effect of attitude was mediated by intention and perceived behavioral control, respectively. Attitude, subjective norm, and perceived behavioral control predicted intention. CONCLUSIONS: The Theory of Planned Behavior may have the potential to explain half of the variance around running-related injury preventive behavior and intention. Therefore, such theory may be considered a relevant and useful tool in developing, investigating, and/or implementing programs aimed at preventing running-related injuries.

Cost-effectiveness and implementation process of a running-related injury prevention program (RunIn3): Protocol of a randomized controlled trial.

BACKGROUND: Running is one of the most popular and accessible physical activities in the world. However, running-related injuries are unfortunately very common. Scientific evidence is limited and scarce regarding (cost-)effectiveness and implementation process of interventions for running-related injuries prevention. Thus, the objective of this study will be to investigate the effectiveness, cost-effectiveness and implementation process of a running-related injury prevention program (RunIn3). METHODS: This is the protocol of a pragmatic hybrid type 1 randomized controlled trial. There will be 530 runners over 18 years old, without running-related injuries in the last 3 months from São Paulo, Brazil. This program will be delivered online with two broad actions: (1) to provide feedback on individual training characteristics and running-related injury risk; and (2) providing/enhancing knowledge, skills and self-efficacy on running-related injury preventive behaviors. The primary outcome will be the proportion of runners reporting running-related injuries. The secondary outcomes will be preventive behaviors, direct and indirect costs, and implementation outcomes. The main effectiveness analysis on the primary outcome will be performed using linear probability mixed models in order to allow outcome changes over time and to yield the absolute risk reduction between-groups. DISCUSSION: The main hypothesis of this study is that the RunIn3 program will be effective in reducing the running-related injury risk and in promoting preventive behavior, either by increasing the frequency of healthy behaviors or by reducing the frequency of risk behaviors. Moreover, if the RunIn3 program is effective in reducing the running-related injuries risk, we believe that this effect would go alongside with a reduction of societal costs. TRAIL REGISTRATION: Clinicaltrials.gov (NCT03892239) Registered 5 February 2019 - Prospectively registered, https://clinicaltrials.gov/ct2/show/NCT03892239.

Understanding and interpreting confidence and credible intervals around effect estimates.

INTRODUCTION: Reporting confidence intervals in scientific articles is important and relevant for evidence-based practice. Clinicians should understand confidence intervals in order to determine if they can realistically expect results similar to those presented in research studies when they implement the scientific evidence in clinical practice. The aims of this masterclass are: (1) to discuss confidence intervals around effect estimates; (2) to understand confidence intervals estimation (frequentist and Bayesian approaches); and (3) to interpret such uncertainty measures. CONTENT: Confidence intervals are measures of uncertainty around effect estimates. Interpretation of the frequentist 95% confidence interval: we can be 95% confident that the true (unknown) estimate would lie within the lower and upper limits of the interval, based on hypothesized repeats of the experiment. Many researchers and health professionals oversimplify the interpretation of the frequentist 95% confidence interval by dichotomizing it in statistically significant or non-statistically significant, hampering a proper discussion on the values, the width (precision) and the practical implications of such interval. Interpretation of the Bayesian 95% confidence interval (which is known as credible interval): there is a 95% probability that the true (unknown) estimate would lie within the interval, given the evidence provided by the observed data. CONCLUSIONS: The use and reporting of confidence intervals should be encouraged in all scientific articles. Clinicians should consider using the interpretation, relevance and applicability of confidence intervals in real-world decision-making. Training and education may enhance knowledge and skills related to estimating, understanding and interpreting uncertainty measures, reducing the barriers for their use under either frequentist or Bayesian approaches.

Natação e aspectos morfológicos do músculo esquelético em processo de reparo após criolesão / Swimming and morphology of skeletal muscle repair process after cryoinjury

O objetivo do estudo foi investigar a influência da natação sobre as alterações morfológicas do músculo esquelético em processo de reparo após criolesão. Foram usados 45 ratos divididos em cinco grupos: controle (n=5); sham (n=5), adaptação (n=5), criolesionados e tratados com natação sacrificados após 7, 14 e 21 dias (n=15); criolesionados e sem tratamento aquático sacrificados após 7, 14 e 21 dias (n=15). As sessões de natação foram realizadas 6 vezes por semana com 90 min de duração cada. Ao término do protocolo os animais foram sacrificados e a análise morfológica da área da lesão foi realizada. A análise morfológica semiquantitativa demonstrou que os músculos do grupo controle apresentaram aspecto histológico normal. O grupo sham apresentou edema, mionecrose e infiltrado inflamatório em grau 1. Nos grupos 7, 14 e 21 dias, não existiram diferenças estatisticamente significativas nas 4 etapas de remodelamento tecidual avaliadas (infiltrado inflamatório, edema, necrose e fibras musculares imaturas) entre os grupos lesionados quando comparados aos grupos com lesão e tratamento aquático. Em conclusão, foi possível verificar que a natação não causou alterações morfológicas durante o reparo do músculo esquelético após criolesão.

The aim of study was investigate the influence of swimming on the morphological changes in skeletal muscle repair process following cryoinjury. There were used 45 rats divided into 5 groups: control (n=5), sham (n=5), adaptation (n=5), cryolesioned treated with swimming and sacrificed after 7, 14 and 21 days (n=15), untreated and cryolesioned sacrificed after 7, 14, and 21 days (n=15). Animals swan for 90 min/ each session and 6 times a week. At the end of the protocol, the animals were sacrificed and morphological analysis of the lesion area was performed. The semi-quantitative morphological analysis showed that the muscles in the control group exhibited normal histological aspects while the sham group exhibited edema, myonecrosis and inflammatory infiltrate grade 1. In groups 7, 14, and 21 days, the results indicated that there were no statistically significant differences in four stages of tissue remodeling evaluated (inflammatory infiltration, edema, necrosis, and immature muscle fibers) between the injured groups compared to groups with lesion and treated with swimming. In conclusion, it was verified that swimming did not alter morphological aspects of skeletal muscle during the repair process following cryoinjury.