Using Similarity Metrics on Real World Data to Recommend the Next Treatment

Haas, Kyle; Mahoui, Malika; Gupta, Saloni; Morton, Stuart

doi:10.1145/3233547.3233647

Cited by 2 publications

(6 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our dataset contains diverse types of variables, while past studies mainly worked on datasets that only included either merely diagnosis data [ 31 ] or limited demographics and physiological data without medication information [ 15 , 16 , 21 ]. Even for those studies that worked on datasets containing medication data, the medication dosage information was ignored [ 18 , 19 ].…”

Section: Discussionmentioning

confidence: 99%

“…However, it remains a challenge to analyze and derive insights from the huge volume of EHR data, which are multivariate, heterogeneous, and sparse. These analyses involve finding similar patients for patient stratification [ 11 , 12 , 13 ], diagnosis prediction [ 14 , 15 ], medical prognosis [ 16 , 17 ], or treatment recommendations [ 18 , 19 , 20 ]. With patient similarity analytics, personalized models can be built based on the retrieved cohort of similar patients, thus furthering the development of personalized medicine.…”

Section: Introductionmentioning

confidence: 99%

“…Existing methods to find similar patients require computing the distances between patients using information such as demographics, diagnoses, relevant laboratory test results, and medications. These methods can be broadly classified into neighborhood-based [ 15 , 16 , 21 ] and cluster-based approaches [ 18 , 22 ]. For neighborhood-based approaches, Lee et al [ 16 ] utilized a cosine similarity metric to select patients similar to index patients, while Ng et al [ 21 ] used Mahalanobis distance and formulated the retrieval of similar patients as an optimization problem.…”

Section: Introductionmentioning

confidence: 99%

“…A new patient is then assigned to one of the pre-defined groups that has the highest similarity score. Haas et al [ 18 ] employed unsupervised clustering approaches with Gower similarity coefficient [ 23 ] to recommend the next treatment for patients with lung cancer. However, cluster-based algorithms often do not work well for patients with complex health conditions and co-morbidities [ 6 ].…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Using Domain Knowledge and Data-Driven Insights for Patient Similarity Analytics

Oei

Fang

Tan

et al. 2021

JPM

View full text Add to dashboard Cite

Patient similarity analytics has emerged as an essential tool to identify cohorts of patients who have similar clinical characteristics to some specific patient of interest. In this study, we propose a patient similarity measure called D3K that incorporates domain knowledge and data-driven insights. Using the electronic health records (EHRs) of 169,434 patients with either diabetes, hypertension or dyslipidaemia (DHL), we construct patient feature vectors containing demographics, vital signs, laboratory test results, and prescribed medications. We discretize the variables of interest into various bins based on domain knowledge and make the patient similarity computation to be aligned with clinical guidelines. Key findings from this study are: (1) D3K outperforms baseline approaches in all seven sub-cohorts; (2) our domain knowledge-based binning strategy outperformed the traditional percentile-based binning in all seven sub-cohorts; (3) there is substantial agreement between D3K and physicians (κ = 0.746), indicating that D3K can be applied to facilitate shared decision making. This is the first study to use patient similarity analytics on a cardiometabolic syndrome-related dataset sourced from medical institutions in Singapore. We consider patient similarity among patient cohorts with the same medical conditions to develop localized models for personalized decision support to improve the outcomes of a target patient.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Using Domain Knowledge and Data-Driven Insights for Patient Similarity Analytics

Oei

Fang

Tan

et al. 2021

JPM

View full text Add to dashboard Cite

show abstract

“…Gower distance was previously explored to analyze treatment selection in an RWD cohort with patients with non‐small cell lung cancer. 34 The RSF feature importance was assigned as covariate contribution in the distance processed by the unsupervised algorithm.…”

Section: Methodsmentioning

confidence: 99%

A novel analytical framework for risk stratification of real‐world data using machine learning: A small cell lung cancer study

Marzano

Darwich

Tendler

et al. 2022

Clinical Translational Sci

View full text Add to dashboard Cite

In recent studies, small cell lung cancer (SCLC) treatment guidelines based on Veterans’ Administration Lung Study Group limited/extensive disease staging and resulted in broad and inseparable prognostic subgroups. Evidence suggests that the eight versions of tumor, node, and metastasis (TNM) staging can play an important role to address this issue. The aim of the present study was to improve the detection of prognostic subgroups from a real‐word data (RWD) cohort of patients and analyze their patterns using a development pipeline with thoracic oncologists and machine learning methods. The method detected subgroups of patients informing unsupervised learning (partition around medoids) including the impact of covariates on prognosis (Cox regression and random survival forest). An analysis was carried out using patients with SCLC (n = 636) with stage IIIA–IVB according to TNM classification. The analysis yielded k = 7 compacted and well‐separated clusters of patients. Performance status (Eastern Cooperative Oncology Group‐Performance Status), lactate dehydrogenase, spreading of metastasis, cancer stage, and CRP were the baselines that characterized the subgroups. The selected clustering method outperformed standard clustering techniques, which were not capable of detecting meaningful subgroups. From the analysis of cluster treatment decisions, we showed the potential of future RWD applications to understand disease, develop individualized therapies, and improve healthcare decision making.

show abstract

Using Similarity Metrics on Real World Data to Recommend the Next Treatment

Cited by 2 publications

References 1 publication

Using Domain Knowledge and Data-Driven Insights for Patient Similarity Analytics

Using Domain Knowledge and Data-Driven Insights for Patient Similarity Analytics

A novel analytical framework for risk stratification of real‐world data using machine learning: A small cell lung cancer study

Contact Info

Product

Resources

About