Learning to Prescribe Interventions for Tuberculosis Patients Using Digital Adherence Data

Killian, Jackson A.; Wilder, Bryan; Sharma, Amit; Choudhary, Vinod; Dilkina, Bistra; Tambe, Milind

doi:10.1145/3292500.3330777

Cited by 41 publications

(43 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although studies related to medication adherence predictive models are found in tuberculosis14 and heart failure,15 few studies have been retrieved that established and examined the predictive models in patients with type 2 diabetes mellitus (T2DM), for whom medication adherence is a key factor to treatment outcomes 5 21 22. Kumamaru H et al 23 established a logistic regression model for the prediction of future drug compliance.…”

Section: Discussionmentioning

confidence: 99%

“…Therefore, it is necessary to explore the main influencing factors of local patients’ medication compliance and to establish a model that can accurately predict therapeutic compliance in specific regions 12 13. Despite several compliance-related predictive models that have been reported recently in patients with tuberculosis14 and heart failure,15 studies on patients with T2D based on multivariate machine learning algorithms have not been retrieved.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Predictive models of medication non-adherence risks of patients with T2D based on multiple machine learning algorithms

Yang

Run-zhang

et al. 2020

BMJ Open Diab Res Care

View full text Add to dashboard Cite

ObjectiveMedication adherence plays a key role in type 2 diabetes (T2D) care. Identifying patients with high risks of non-compliance helps individualized management, especially for China, where medical resources are relatively insufficient. However, models with good predictive capabilities have not been studied. This study aims to assess multiple machine learning algorithms and screen out a model that can be used to predict patients’ non-adherence risks.MethodsA real-world registration study was conducted at Sichuan Provincial People’s Hospital from 1 April 2018 to 30 March 2019. Data of patients with T2D on demographics, disease and treatment, diet and exercise, mental status, and treatment adherence were obtained by face-to-face questionnaires. The medication possession ratio was used to evaluate patients’ medication adherence status. Fourteen machine learning algorithms were applied for modeling, including Bayesian network, Neural Net, support vector machine, and so on, and balanced sampling, data imputation, binning, and methods of feature selection were evaluated by the area under the receiver operating characteristic curve (AUC). We use two-way cross-validation to ensure the accuracy of model evaluation, and we performed a posteriori test on the sample size based on the trend of AUC as the sample size increase.ResultsA total of 401 patients out of 630 candidates were investigated, of which 85 were evaluated as poor adherence (21.20%). A total of 16 variables were selected as potential variables for modeling, and 300 models were built based on 30 machine learning algorithms. Among these algorithms, the AUC of the best capable one was 0.866±0.082. Imputing, oversampling and larger sample size will help improve predictive ability.ConclusionsAn accurate and sensitive adherence prediction model based on real-world registration data was established after evaluating data filling, balanced sampling, and so on, which may provide a technical tool for individualized diabetes care.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Predictive models of medication non-adherence risks of patients with T2D based on multiple machine learning algorithms

Yang

Run-zhang

et al. 2020

BMJ Open Diab Res Care

View full text Add to dashboard Cite

show abstract

“…Predicting engagement from user behavior provides the first critical step toward increasing the dosage, frequency and effectiveness of digital interventions. These prediction models can also be used to proactively direct resources toward settings at a higher risk of failure ( Killian et al, 2019 ).…”

Section: The Challenges and Opportunities In Low And Middle Income Comentioning

confidence: 99%

AI for Improving Children’s Health: A Community Case Study

Ganju¹,

Satyan²,

Tanna³

et al. 2021

Front. Artif. Intell.

View full text Add to dashboard Cite

The Indian health care system lacks the infrastructure to meet the health care demands of the country. Physician and nurse availability is 30 and 50% below WHO recommendations, respectively, and has led to a steep imbalance between the demand for health care and the infrastructure available to support it. Among other concerns, India still struggles with challenges like undernutrition, with 38% of children under the age of five being underweight. Despite these challenges, technological advancements, mobile phone ubiquity and rising patient awareness offers a huge opportunity for artificial intelligence to enable efficient healthcare delivery, by improved targeting of constrained resources. The Saathealth mobile app provides low-middle income parents of young children nflwith interactive children’s health, nutrition and development content in the form of an entertaining video series, a gamified quiz journey and targeted notifications. The app iteratively evolves the user journey based on dynamic data and predictive algorithms, empowering a shift from reactive to proactive care. Saathealth users have registered over 500,000 sessions and over 200 million seconds on-app engagement over a year, comparing favorably with engagement on other digital health interventions in underserved communities. We have used valuable app analytics data and insights from our 45,000 users to build scalable, predictive models that were validated for specific use cases. Using the Random Forest model with heterogeneous data allowed us to predict user churn with a 93% accuracy. Predicting user lifetimes on the mobile app for preliminary insights gave us an RMSE of 25.09 days and an R2 value of 0.91, reflecting closely correlated predictions. These predictive algorithms allow us to incentivize users with optimized offers and omni-channel nudges, to increase engagement with content as well as other targeted online and offline behaviors. The algorithms also optimize the effectiveness of our intervention by augmenting personalized experiences and directing limited health resources toward populations that are most resistant to digital first interventions. These and similar AI powered algorithms will allow us to lengthen and deepen the lifetime relationship with our health consumers, making more of them effective, proactive participants in improving children’s health, nutrition and early cognitive development.

show abstract

“…It is important to note that Peetluk et al [23] do not classify LR as machine learning in their review as the LR analysis was used as a statistical methodology to understand the relationship between attributes and their prevalence. In the few machine learning studies identified, it was used primarily for predicting treatment completion [39] or unfavourable outcomes [40,41].…”

Section: Related Workmentioning

confidence: 99%

Benchmarking Machine Learning Models to Assist in the Prognosis of Tuberculosis

et al. 2021

View full text Add to dashboard Cite

Tuberculosis (TB) is an airborne infectious disease caused by organisms in the Mycobacterium tuberculosis (Mtb) complex. In many low and middle-income countries, TB remains a major cause of morbidity and mortality. Once a patient has been diagnosed with TB, it is critical that healthcare workers make the most appropriate treatment decision given the individual conditions of the patient and the likely course of the disease based on medical experience. Depending on the prognosis, delayed or inappropriate treatment can result in unsatisfactory results including the exacerbation of clinical symptoms, poor quality of life, and increased risk of death. This work benchmarks machine learning models to aid TB prognosis using a Brazilian health database of confirmed cases and deaths related to TB in the State of Amazonas. The goal is to predict the probability of death by TB thus aiding the prognosis of TB and associated treatment decision making process. In its original form, the data set comprised 36,228 records and 130 fields but suffered from missing, incomplete, or incorrect data. Following data cleaning and preprocessing, a revised data set was generated comprising 24,015 records and 38 fields, including 22,876 reported cured TB patients and 1139 deaths by TB. To explore how the data imbalance impacts model performance, two controlled experiments were designed using (1) imbalanced and (2) balanced data sets. The best result is achieved by the Gradient Boosting (GB) model using the balanced data set to predict TB-mortality, and the ensemble model composed by the Random Forest (RF), GB and Multi-Layer Perceptron (MLP) models is the best model to predict the cure class.

show abstract

Learning to Prescribe Interventions for Tuberculosis Patients Using Digital Adherence Data

Cited by 41 publications

References 24 publications

Predictive models of medication non-adherence risks of patients with T2D based on multiple machine learning algorithms

Predictive models of medication non-adherence risks of patients with T2D based on multiple machine learning algorithms

AI for Improving Children’s Health: A Community Case Study

Benchmarking Machine Learning Models to Assist in the Prognosis of Tuberculosis

Contact Info

Product

Resources

About