Classification of Players Across the Australian Rules Football Participation Pathway Based on Physical Characteristics.

ABSTRACT: Haycraft, JAZ, Kovalchik, S, Pyne, DB, and Robertson, S. Classification of players across the Australian Rules football participation pathway based on physical characteristics. J Strength Cond Res 36(3): 702-709, 2022-This study investigated the utility of physical fitness and movement ability tests to differentiate and classify players into Australian Football League (AFL) participation pathway levels. Players (n = 293, age 10.9-19.1 years) completed the following tests; 5-m, 10-m, and 20-m sprint, AFL planned agility, vertical jump (VJ), running VJ, 20-m multistage fitness test (MSFT), and athletic ability assessment. A multivariate analysis of variance between AFL participation pathway levels was conducted, and a classification tree determined the extent to which players could be allocated to relevant levels. The magnitude of differences between physical fitness and movement ability were level-dependent, with the largest standardized effect size (ES) between Local U12, Local U14, and older levels for most physical fitness tests (ES: -4.64 to 5.02), except the 5-m and 10-m sprint. The 20-m, 5-m, AFL agility, 20-m MSFT, overhead squat, and running VJ (right) contributed to the classification model, with 57% overall accuracy reported (43% under cross-validation). National U16 players were easiest to classify (87%), while National U18 players were most difficult (0%). Physical fitness tests do not seem to differentiate between players after selection into AFL talent pathway levels. Other attributes (i.e., skill, psychological, and sociocultural) should be prioritized over physical fitness and movement attributes by selectors/coaches when considering selection of talented players.

Do magnetic resonance imaging abnormalities of the non-dominant wrist correlate with ulnar-sided wrist pain in elite tennis players?

OBJECTIVE: Ulnar-sided injuries of the non-dominant wrist are common in elite tennis players that use the double-handed backhand technique. This study aimed to define the relationship between ulnar-sided wrist pain in symptomatic and asymptomatic elite tennis players, and the presence of abnormalities on magnetic resonance imaging (MRI). MATERIALS AND METHODS: Fourteen symptomatic tennis players, 14 asymptomatic tennis players, and 12 healthy controls who did not play tennis, were analyzed prospectively, after undergoing MRI of their non-dominant wrist. Five anatomical regions were analyzed, thought to relate to ulnar-sided wrist pain. These consisted of the triangular fibrocartilage complex (TFCC), ulnar collateral ligament (UCL), extensor carpi ulnaris tendon (ECU), osseous-articular structures, and ganglia. Images were independently reviewed by two blinded musculoskeletal radiologists. RESULTS: Non-dominant, ulnar-sided, wrist pain in elite tennis players was not statistically significantly associated with an increased number of MRI abnormalities when compared with asymptomatic tennis players (p > 0.05). However, some evidence of statistical association was seen with an increased prevalence of ECU tendon abnormalities (OR = 8.0, 95% CI = (0.74, 20.00), p = 0.07). A statistically significant increase in MRI abnormalities of osseous structures (OR = 15.1, 95% CI = (1.56, 656.05), p = 0.02) and the dorsal radioulnar ligament (DRUL) (OR = 12.5, 95% CI = (2.15, 111.11), p = 0.03), was observed in symptomatic players compared with controls. CONCLUSIONS: Non-dominant, ulnar-sided, wrist pain in a subgroup of elite tennis players using a double-handed backhand technique is not associated with a statistically significant increased prevalence of MRI abnormalities when compared with asymptomatic tennis players, other than some evidence of statistical association with ECU tendon abnormalities. Therefore, significance of MRI abnormalities should be interpreted in the context of clinical findings.

Is running associated with a lower risk of all-cause, cardiovascular and cancer mortality, and is the more the better? A systematic review and meta-analysis.

OBJECTIVE: To investigate the association of running participation and the dose of running with the risk of all-cause, cardiovascular and cancer mortality. DESIGN: Systematic review and meta-analysis. DATA SOURCES: Journal articles, conference papers and doctoral theses indexed in Academic Search Ultimate, CINAHL, Health Source: Nursing/Academic Edition, MasterFILE Complete, Networked Digital Library of Theses and Dissertations, Open Access Theses and Dissertations, PsycINFO, PubMed/MEDLINE, Scopus, SPORTDiscus and Web of Science. ELIGIBILITY CRITERIA FOR SELECTING STUDIES: Prospective cohort studies on the association between running or jogging participation and the risk of all-cause, cardiovascular and/or cancer mortality in a non-clinical population of adults were included. RESULTS: Fourteen studies from six prospective cohorts with a pooled sample of 232 149 participants were included. In total, 25 951 deaths were recorded during 5.5-35 year follow-ups. Our meta-analysis showed that running participation is associated with 27%, 30% and 23% lower risk of all-cause (pooled adjusted hazard ratio (HR)=0.73; 95% confidence interval (CI) 0.68 to 0.79), cardiovascular (HR=0.70; 95% CI 0.49 to 0.98) and cancer (HR=0.77; 95% CI 0.68 to 0.87) mortality, respectively, compared with no running. A meta-regression analysis showed no significant dose-response trends for weekly frequency, weekly duration, pace and the total volume of running. CONCLUSION: Increased rates of participation in running, regardless of its dose, would probably lead to substantial improvements in population health and longevity. Any amount of running, even just once a week, is better than no running, but higher doses of running may not necessarily be associated with greater mortality benefits.

A machine learning approach for automatic detection and classification of changes of direction from player tracking data in professional tennis.

The purpose of this study was to develop an automated method for identifying and classifying change of direction (COD) movements in professional tennis using tracking data. Three sport science and strength and conditioning experts coded match-play footage of nineteen professional tennis players (9 male and 10 female) from the Australian Open Grand Slam for COD of medium and high intensity. A total of 1,494 changes were identified and aligned with 2D player position sampled at 25 Hz based on camera tracking data. Several machine learning classifiers were trained and tested on a set of 1,128 time-motion features. A random forest algorithm was found to have the best out-of-sample performance, classifying medium and high intensity changes with an F1-score of 0.729. This research offers a novel and applicable way for utilising player tracking data and machine learning techniques to automatically identify and classify COD movements in professional tennis.

Validation of ball spin estimates in tennis from multi-camera tracking data.

The purpose of this study was to examine the accuracy of methods to estimate ball spin rate and spin axis direction using multi-camera tracking technology. The method implemented by Hawk-Eye and a theoretical ball trajectory model were assessed for their accuracy against high speed-vision. A theoretical ball trajectory model was found to estimate ball spin rate and the spin axis direction more accurately than the method used by Hawk-Eye. The superior performance of the ball trajectory model was indicated by a lower mean bias and standard deviation (2.92 ± 222.76) than Hawk-Eye (-100.01 ± 542.44), as well as narrower limits of agreement and a 327.63 RPM lower RMSE. Spin rates less than 4500 RPM were estimated with the highest accuracy, but neither method was able to consistently and accurately estimate spin rates >4500 RPM. The application of a trajectory model to multi-camera ball tracking data provides a practical and non-invasive method to accurately estimate the spin imparted when hitting, allowing for large-scale collection of spin rates during matches.

Implications of Nine Risk Prediction Models for Selecting Ever-Smokers for Computed Tomography Lung Cancer Screening.

Background: Lung cancer screening guidelines recommend using individualized risk models to refer ever-smokers for screening. However, different models select different screening populations. The performance of each model in selecting ever-smokers for screening is unknown. Objective: To compare the U.S. screening populations selected by 9 lung cancer risk models (the Bach model; the Spitz model; the Liverpool Lung Project [LLP] model; the LLP Incidence Risk Model [LLPi]; the Hoggart model; the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial Model 2012 [PLCOM2012]; the Pittsburgh Predictor; the Lung Cancer Risk Assessment Tool [LCRAT]; and the Lung Cancer Death Risk Assessment Tool [LCDRAT]) and to examine their predictive performance in 2 cohorts. Design: Population-based prospective studies. Setting: United States. Participants: Models selected U.S. screening populations by using data from the National Health Interview Survey from 2010 to 2012. Model performance was evaluated using data from 337 388 ever-smokers in the National Institutes of Health-AARP Diet and Health Study and 72 338 ever-smokers in the CPS-II (Cancer Prevention Study II) Nutrition Survey cohort. Measurements: Model calibration (ratio of model-predicted to observed cases [expected-observed ratio]) and discrimination (area under the curve [AUC]). Results: At a 5-year risk threshold of 2.0%, the models chose U.S. screening populations ranging from 7.6 million to 26 million ever-smokers. These disagreements occurred because, in both validation cohorts, 4 models (the Bach model, PLCOM2012, LCRAT, and LCDRAT) were well-calibrated (expected-observed ratio range, 0.92 to 1.12) and had higher AUCs (range, 0.75 to 0.79) than 5 models that generally overestimated risk (expected-observed ratio range, 0.83 to 3.69) and had lower AUCs (range, 0.62 to 0.75). The 4 best-performing models also had the highest sensitivity at a fixed specificity (and vice versa) and similar discrimination at a fixed risk threshold. These models showed better agreement on size of the screening population (7.6 million to 10.9 million) and achieved consensus on 73% of persons chosen. Limitation: No consensus on risk thresholds for screening. Conclusion: The 9 lung cancer risk models chose widely differing U.S. screening populations. However, 4 models (the Bach model, PLCOM2012, LCRAT, and LCDRAT) most accurately predicted risk and performed best in selecting ever-smokers for screening. Primary Funding Source: Intramural Research Program of the National Institutes of Health/National Cancer Institute.

Enhancing skill transfer in tennis using representative learning design.

Representative Learning Design advocates that practice should simulate the demands of competition. The effectiveness of increased task representativeness to improve serving skill of junior tennis players was assessed after a six-week intervention. Thirty-three participants (15.4 ± 1.9 years of age) were assigned to one of the three groups; "serve only" (participants served to no opponent), "serve return" (participants served to an opponent and hit no extra shots) or "serve +3rd" (participants served to an opponent and hit one extra shot). Using the validated representative practice assessment tool (RPAT) tasks were considered to be low, moderate and high in task representativeness, respectively. Participants hit 56 serves, twice weekly for 6-weeks. Pre and post serving performances were assessed via a skill test and in-situ matchplay using SportsCode and HawkEye ball tracking, respectively. Serve speed, landing locations, serve angle and positional advantage was obtained for 1st and 2nd serves. The relationship between increasing representativeness and increased skill acquisition was not linear, rather different behaviours emerged. For example, when hitting 2nd serves in matchplay, the low and moderate representative groups prioritised speed over placement while the high representative group prioritised placement over speed. Coaches therefore need to carefully individualise representativeness to an athletes' specific needs.

A shot taxonomy in the era of tracking data in professional tennis.

Shots are an essential part of the language of tennis yet little is known about the distinct types of shots in the professional game. In this study, we build a taxonomy of shots for elite tennis players using tracking data from multiple years of men's and women's matches at the Australian Open. Our taxonomy is constructed using model-based multi-stage functional data clustering, an unsupervised machine learning approach. Among 270,023 men's and 178,136 women's shots, we found 13 distinct types of serves to both the Ad and Deuce court for male players and 17 and 15 types to the Ad and Deuce for female players. More variety was found among serve returns and rally shots compared to the serve; with less variety on the backhand than forehand. There was also more overlap in the physical characteristics of groundstroke shots between male and female players than on serve. Shot type was strongly associated with winning points and shots in the highest speed and lowest net clearance categories tended to be the most effective. This data-driven dictionary of shots provides a framework for analysis of elite player performance, characterizing playing style, and designing more representative practice.

Injury epidemiology of tennis players at the 2011-2016 Australian Open Grand Slam.

AIM: To examine the epidemiology and in-event treatment frequency of injury at the 2011-2016 Australian Open tournaments. METHODS: Injury incidence was defined as a medical consultation by a tournament physician and in-event treatment frequency as the mean total number of follow-up medical/physiotherapy consultations (2013-2016 tournaments only). Data were collated by sex, injury region and type and reported as frequencies per 10 000 game exposures. Incidence rate s± 95% CI and rate ratios were used to test effects for injury, sex and year. RESULTS: Female players experienced more injuries than male players (201.7 vs 148.6). The shoulder (5.1±1.1 injuries per year), foot (3.2±1.1), wrist (3.1±1.5) and knee (3.1±1.1) were the most commonly injured regions among females. Knee (3.5±1.6), ankle (2.3±1.3) and thigh (2.3±1.5) were the most prevalent male injuries. Upper arm injuries and in-event treatment frequency increased by ≥2.4 times in both sexes over the 5-year period. Muscle injuries were most frequent. There was a greater than twofold increase in men and women with stress fractures over the 5-year period. The torso region, including the neck, thoracic spine, trunk and abdominal, lumbar spine, hip and groin, pelvis/buttock, attracted high in-event treatment frequencies in both sexes. CONCLUSION: Investigation of injury at the Australian Open suggests that females are more commonly injured than males. Upper and lower extremity injuries affected females while lower limb injuries were more prominent in males. There was an increasing rate of in-event treatments of upper limb and torso injuries as well as stress fractures during the observation period.

Getting to the top: an analysis of 25 years of career rankings trajectories for professional women's tennis.

Official rankings are the most common measure of success in professional women's tennis. Despite their importance for earning potential and tournament seeding, little is known about ranking trajectories of female players and their influence on career success. Our objective was to conduct a comprehensive study of the career progression of elite female tennis talent. The study examined the ranking trajectories of the top 250 female professionals between 1990 and 2015. Using regression modelling of yearly peak rankings, we found a strong association between the shape of the ranking trajectory and the highest career ranking earned. Players with the highest career peak ranking were the youngest when first ranked. For example, top 10 players were first ranked at age 15.5 years (99% CI = 14.8-15.9), 1.2 years (99% CI = 0.8-1.5) earlier than top 51-100 players. Top 10 players were also ranked in the top 100 longer than other players, holding a top 100 ranking until a mean age of 29.0 years (99% CI = 27.8-30.3) compared with age 24.4 years (99% CI = 23.7-25.2) for top 51-100 players. Ranking trajectories were more distinct with respect to player age than years from first ranking. The present study's findings will be instructive for players, coaches, and administrators in setting goals and assessing athlete development in women's tennis.

Comparing Matchplay Characteristics and Physical Demands of Junior and Professional Tennis Athletes in the Era of Big Data.

Differences in the competitive performance characteristics of junior and professional tennis players are not well understood. The present study provides a comprehensive comparative analysis of junior and professional matchplay. The study utilized multiple large-scale datasets covering match, point, and shot outcomes over multiple years of competition. Regression analysis was used to identify differences between junior and professional matchplay. Top professional men and women were found to play significantly more matches, sets, and games compared to junior players of an equivalent ranking. Professional players had a greater serve advantage, men winning 4 and women winning 2 additional percentage points on serve compared to juniors. Clutch ability in break point conversion was 6 to 8 percentage points greater for junior players. In general, shots were more powerful and more accurate at the professional level with the largest differences observed for male players on serve. Serving to the center of the court was more than two times more common for junior players on first serve. While male professionals performed 50% more total work in a Grand Slam match than juniors, junior girls performed 50% more work than professional women. Understanding how competitiveness, play demands, and the physical characteristics of shots differ between junior and professional tennis players can help set realistic expectations and developmentally appropriate training for transitioning players.

Targeting of low-dose CT screening according to the risk of lung-cancer death.

BACKGROUND: In the National Lung Screening Trial (NLST), screening with low-dose computed tomography (CT) resulted in a 20% reduction in lung-cancer mortality among participants between the ages of 55 and 74 years with a minimum of 30 pack-years of smoking and no more than 15 years since quitting. It is not known whether the benefits and potential harms of such screening vary according to lung-cancer risk. METHODS: We assessed the variation in efficacy, the number of false positive results, and the number of lung-cancer deaths prevented among 26,604 participants in the NLST who underwent low-dose CT screening, as compared with the 26,554 participants who underwent chest radiography, according to the quintile of 5-year risk of lung-cancer death (ranging from 0.15 to 0.55% in the lowest-risk group [quintile 1] to more than 2.00% in the highest-risk group [quintile 5]). RESULTS: The number of lung-cancer deaths per 10,000 person-years that were prevented in the CT-screening group, as compared with the radiography group, increased according to risk quintile (0.2 in quintile 1, 3.5 in quintile 2, 5.1 in quintile 3, 11.0 in quintile 4, and 12.0 in quintile 5; P=0.01 for trend). Across risk quintiles, there were significant decreasing trends in the number of participants with false positive results per screening-prevented lung-cancer death (1648 in quintile 1, 181 in quintile 2, 147 in quintile 3, 64 in quintile 4, and 65 in quintile 5). The 60% of participants at highest risk for lung-cancer death (quintiles 3 through 5) accounted for 88% of the screening-prevented lung-cancer deaths and for 64% of participants with false positive results. The 20% of participants at lowest risk (quintile 1) accounted for only 1% of prevented lung-cancer deaths. CONCLUSIONS: Screening with low-dose CT prevented the greatest number of deaths from lung cancer among participants who were at highest risk and prevented very few deaths among those at lowest risk. These findings provide empirical support for risk-based targeting of smokers for such screening. (Funded by the National Cancer Institute.).

Monoclonal gammopathy of undetermined significance and risk of lymphoid and myeloid malignancies: 728 cases followed up to 30 years in Sweden.

In 728 Swedish cases of monoclonal gammopathy of undetermined significance (MGUS), followed up to 30 years (median, 10 years), we estimated the cumulative risk of hematologic disorders originating from lymphoid and myeloid lineages. Using Cox regression models, we examined associations of demographic and laboratory factors with progression and determined the discriminatory power of 3 prediction models for progression. Eighty-four MGUS cases developed a lymphoid disorder, representing a cumulative risk of 15.4%. Multiple myeloma (MM) occurred in 53 patients, and the 30-year cumulative risk was 10.6%; an ∼0.5% annual risk. Three factors were significantly associated with progression: abnormal free light-chain (FLC) ratio (<0.26 or >1.65), M-protein concentration (≥1.5 g/dL), and reduction of 1 or 2 noninvolved immunoglobulin isotype levels (immunoparesis). A prediction model with separate effects for these 3 factors and the M-protein isotype had higher discriminatory power than other models, although the differences were not statistically significant. The 30-year cumulative risk for myeloid malignancies was <2%. Our study confirms that abnormal FLC ratio and M-protein concentration >1.5 g/dL, factors previously considered by Mayo Clinic researchers, are predictors for MM progression and suggests that separate consideration of immunoparesis and the Mayo Clinic risk factors could improve identification of MGUS patients at high risk for progression.

Development and Validation of Risk Models to Select Ever-Smokers for CT Lung Cancer Screening.

IMPORTANCE: The US Preventive Services Task Force (USPSTF) recommends computed tomography (CT) lung cancer screening for ever-smokers aged 55 to 80 years who have smoked at least 30 pack-years with no more than 15 years since quitting. However, selecting ever-smokers for screening using individualized lung cancer risk calculations may be more effective and efficient than current USPSTF recommendations. OBJECTIVE: Comparison of modeled outcomes from risk-based CT lung-screening strategies vs USPSTF recommendations. DESIGN, SETTING, AND PARTICIPANTS: Empirical risk models for lung cancer incidence and death in the absence of CT screening using data on ever-smokers from the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO; 1993-2009) control group. Covariates included age; education; sex; race; smoking intensity, duration, and quit-years; body mass index; family history of lung cancer; and self-reported emphysema. Model validation in the chest radiography groups of the PLCO and the National Lung Screening Trial (NLST; 2002-2009), with additional validation of the death model in the National Health Interview Survey (NHIS; 1997-2001), a representative sample of the United States. Models were applied to US ever-smokers aged 50 to 80 years (NHIS 2010-2012) to estimate outcomes of risk-based selection for CT lung screening, assuming screening for all ever-smokers, yield the percent changes in lung cancer detection and death observed in the NLST. EXPOSURES: Annual CT lung screening for 3 years beginning at age 50 years. MAIN OUTCOMES AND MEASURES: For model validity: calibration (number of model-predicted cases divided by number of observed cases [estimated/observed]) and discrimination (area under curve [AUC]). For modeled screening outcomes: estimated number of screen-avertable lung cancer deaths and estimated screening effectiveness (number needed to screen [NNS] to prevent 1 lung cancer death). RESULTS: Lung cancer incidence and death risk models were well calibrated in PLCO and NLST. The lung cancer death model calibrated and discriminated well for US ever-smokers aged 50 to 80 years (NHIS 1997-2001: estimated/observed = 0.94 [95%CI, 0.84-1.05]; AUC, 0.78 [95%CI, 0.76-0.80]). Under USPSTF recommendations, the models estimated 9.0 million US ever-smokers would qualify for lung cancer screening and 46,488 (95% CI, 43,924-49,053) lung cancer deaths were estimated as screen-avertable over 5 years (estimated NNS, 194 [95% CI, 187-201]). In contrast, risk-based selection screening of the same number of ever-smokers (9.0 million) at highest 5-year lung cancer risk (≥1.9%) was estimated to avert 20% more deaths (55,717 [95% CI, 53,033-58,400]) and was estimated to reduce the estimated NNS by 17% (NNS, 162 [95% CI, 157-166]). CONCLUSIONS AND RELEVANCE: Among a cohort of US ever-smokers aged 50 to 80 years, application of a risk-based model for CT screening for lung cancer compared with a model based on USPSTF recommendations was estimated to be associated with a greater number of lung cancer deaths prevented over 5 years, along with a lower NNS to prevent 1 lung cancer death.

Aggregate-data estimation of an individual patient data linear random effects meta-analysis with a patient covariate-treatment interaction term.

Individual patient-data meta-analysis of randomized controlled trials is the gold standard for investigating how patient factors modify the effectiveness of treatment. Because participant data from primary studies might not be available, reliable alternatives using published data are needed. In this paper, I show that the maximum likelihood estimates of a participant-level linear random effects meta-analysis with a patient covariate-treatment interaction can be determined exactly from aggregate data when the model's variance components are known. I provide an equivalent aggregate-data EM algorithm and supporting software with the R package ipdmeta for the estimation of the "interaction meta-analysis" when the variance components are unknown. The properties of the methodology are assessed with simulation studies. The usefulness of the methods is illustrated with analyses of the effect modification of cholesterol and age on pravastatin in the multicenter placebo-controlled regression growth evaluation statin study. When a participant-level meta-analysis cannot be performed, aggregate-data interaction meta-analysis is a useful alternative for exploring individual-level sources of treatment effect heterogeneity.

Population-based absolute risk estimation with survey data.

Absolute risk is the probability that a cause-specific event occurs in a given time interval in the presence of competing events. We present methods to estimate population-based absolute risk from a complex survey cohort that can accommodate multiple exposure-specific competing risks. The hazard function for each event type consists of an individualized relative risk multiplied by a baseline hazard function, which is modeled nonparametrically or parametrically with a piecewise exponential model. An influence method is used to derive a Taylor-linearized variance estimate for the absolute risk estimates. We introduce novel measures of the cause-specific influences that can guide modeling choices for the competing event components of the model. To illustrate our methodology, we build and validate cause-specific absolute risk models for cardiovascular and cancer deaths using data from the National Health and Nutrition Examination Survey. Our applications demonstrate the usefulness of survey-based risk prediction models for predicting health outcomes and quantifying the potential impact of disease prevention programs at the population level.

Assessing heterogeneity of treatment effect in a clinical trial with the proportional interactions model.

Understanding how individuals vary in their response to treatment is an important task of clinical research. For standard regression models, a proportional interactions model first described by Follmann and Proschan (1999) offers a powerful approach for identifying effect modification in a randomized clinical trial when multiple variables influence treatment response. In this paper, we present a framework for using the proportional interactions model in the context of a parallel-arm clinical trial with multiple prespecified candidate effect modifiers. To protect against model misspecification, we propose a selection strategy that considers all possible proportional interactions models. We develop a modified Bonferroni correction for multiple testing that accounts for the positive correlation among candidate models. We describe methods for constructing a confidence interval for the proportionality parameter. In simulation studies, we show that our modified Bonferroni adjustment controls familywise error and has greater power to detect proportional interactions compared with multiplcity-corrected subgroup analyses. We demonstrate our methodology by using the Studies of Left Ventricular Dysfunction Treatment trial, a placebo-controlled randomized clinical trial of the efficacy of enalapril to reduce the risk of death or hospitalization in chronic heart failure patients. An R package called anoint is available for implementing the proportional interactions methodology.

A general binomial regression model to estimate standardized risk differences from binary response data.

Estimates of absolute risks and risk differences are necessary for evaluating the clinical and population impact of biomedical research findings. We have developed a linear-expit regression model (LEXPIT) to incorporate linear and nonlinear risk effects to estimate absolute risk from studies of a binary outcome. The LEXPIT is a generalization of both the binomial linear and logistic regression models. The coefficients of the LEXPIT linear terms estimate adjusted risk differences, whereas the exponentiated nonlinear terms estimate residual odds ratios. The LEXPIT could be particularly useful for epidemiological studies of risk association, where adjustment for multiple confounding variables is common. We present a constrained maximum likelihood estimation algorithm that ensures the feasibility of risk estimates of the LEXPIT model and describe procedures for defining the feasible region of the parameter space, judging convergence, and evaluating boundary cases. Simulations demonstrate that the methodology is computationally robust and yields feasible, consistent estimators. We applied the LEXPIT model to estimate the absolute 5-year risk of cervical precancer or cancer associated with different Pap and human papillomavirus test results in 167,171 women undergoing screening at Kaiser Permanente Northern California. The LEXPIT model found an increased risk due to abnormal Pap test in human papillomavirus-negative that was not detected with logistic regression. Our R package blm provides free and easy-to-use software for fitting the LEXPIT model.