scholarly journals Comparison of Conventional Statistical Methods with Machine Learning in Medicine: Diagnosis, Drug Development, and Treatment

Medicina ◽  
2020 ◽  
Vol 56 (9) ◽  
pp. 455
Author(s):  
Hema Sekhar Reddy Rajula ◽  
Giuseppe Verlato ◽  
Mirko Manchia ◽  
Nadia Antonucci ◽  
Vassilios Fanos
Keyword(s):  
Machine Learning ◽  
Drug Development ◽  
Statistical Methods ◽  
Patient Reported Outcomes ◽  
A Priori ◽  
Response To Treatment ◽  
Patient Reported ◽  
Need To Evaluate ◽  
Diagnosis And Classification ◽  
Treatment Integration

Futurists have anticipated that novel autonomous technologies, embedded with machine learning (ML), will substantially influence healthcare. ML is focused on making predictions as accurate as possible, while traditional statistical models are aimed at inferring relationships between variables. The benefits of ML comprise flexibility and scalability compared with conventional statistical approaches, which makes it deployable for several tasks, such as diagnosis and classification, and survival predictions. However, much of ML-based analysis remains scattered, lacking a cohesive structure. There is a need to evaluate and compare the performance of well-developed conventional statistical methods and ML on patient outcomes, such as survival, response to treatment, and patient-reported outcomes (PROs). In this article, we compare the usefulness and limitations of traditional statistical methods and ML, when applied to the medical field. Traditional statistical methods seem to be more useful when the number of cases largely exceeds the number of variables under study and a priori knowledge on the topic under study is substantial such as in public health. ML could be more suited in highly innovative fields with a huge bulk of data, such as omics, radiodiagnostics, drug development, and personalized treatment. Integration of the two approaches should be preferred over a unidirectional choice of either approach.

scholarly journals Electronic patient-reported outcomes and machine learning in predicting immune-related adverse events of immune checkpoint inhibitor therapies

2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Sanna Iivanainen ◽  
Jussi Ekstrom ◽  
Henri Virtanen ◽  
Vesa V. Kataja ◽  
Jussi P. Koivunen
Keyword(s):  
Machine Learning ◽  
Early Detection ◽  
Adverse Events ◽  
Cancer Patients ◽  
Immune Checkpoint ◽  
Patient Reported Outcomes ◽  
Prediction Models ◽  
Accuracy Score ◽  
Patient Reported ◽  
Auc Value

Abstract Background Immune-checkpoint inhibitors (ICIs) have introduced novel immune-related adverse events (irAEs), arising from various organ systems without strong timely dependency on therapy dosing. Early detection of irAEs could result in improved toxicity profile and quality of life. Symptom data collected by electronic (e) patient-reported outcomes (PRO) could be used as an input for machine learning (ML) based prediction models for the early detection of irAEs. Methods The utilized dataset consisted of two data sources. The first dataset consisted of 820 completed symptom questionnaires from 34 ICI treated advanced cancer patients, including 18 monitored symptoms collected using the Kaiku Health digital platform. The second dataset included prospectively collected irAE data, Common Terminology Criteria for Adverse Events (CTCAE) class, and the severity of 26 irAEs. The ML models were built using extreme gradient boosting algorithms. The first model was trained to detect the presence and the second the onset of irAEs. Results The model trained to predict the presence of irAEs had an excellent performance based on four metrics: accuracy score 0.97, Area Under the Curve (AUC) value 0.99, F1-score 0.94 and Matthew’s correlation coefficient (MCC) 0.92. The prediction of the irAE onset was more difficult with accuracy score 0.96, AUC value 0.93, F1-score 0.66 and MCC 0.64 but the model performance was still at a good level. Conclusion The current study suggests that ML based prediction models, using ePRO data as an input, can predict the presence and onset of irAEs with a high accuracy, indicating that ePRO follow-up with ML algorithms could facilitate the detection of irAEs in ICI-treated cancer patients. The results should be validated with a larger dataset. Trial registration Clinical Trials Register (NCT3928938), registration date the 26th of April, 2019

scholarly journals Linguistic validation of the simplified Chinese version of the US National Cancer Institute’s patient-reported outcomes version of the common terminology criteria for adverse events (PRO-CTCAE™)

BMC Cancer ◽  
2020 ◽  
Vol 20 (1) ◽  
Author(s):  
Cheng KKF ◽  
S. A. Mitchell ◽  
N. Chan ◽  
E. Ang ◽  
W. Tam ◽  
...  
Keyword(s):  
Adverse Events ◽  
Cancer Patients ◽  
Patient Reported Outcomes ◽  
A Priori ◽  
Cognitive Interviews ◽  
Simplified Chinese ◽  
Patient Reported ◽  
The Us ◽  
The Common ◽  
Response Choices

Abstract Background The aim of this study was to translate and linguistically validate the U.S. National Cancer Institute’s Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE™) into Simplified Chinese for use in Singapore. Methods All 124 items of the English source PRO-CTCAE item library were translated into Simplified Chinese using internationally established translation procedures. Two rounds of cognitive interviews were conducted with 96 cancer patients undergoing adjuvant treatment to determine if the translations adequately captured the PRO-CTCAE source concepts, and to evaluate comprehension, clarity and ease of judgement. Interview probes addressed the 78 PRO-CTCAE symptom terms (e.g. fatigue), as well as the attributes (e.g. severity), response choices, and phrasing of ‘at its worst’. Items that met the a priori threshold of ≥20% of participants with comprehension difficulties were considered for rephrasing and retesting. Items where < 20% of the sample experienced comprehension difficulties were also considered for rephrasing if better phrasing options were available. Results A majority of PRO-CTCAE-Simplified Chinese items were well comprehended by participants in Round 1. One item posed difficulties in ≥20% and was revised. Two items presented difficulties in < 20% but were revised as there were preferred alternative phrasings. Twenty-four items presented difficulties in < 10% of respondents. Of these, eleven items were revised to an alternative preferred phrasing, four items were revised to include synonyms. Revised items were tested in Round 2 and demonstrated satisfactory comprehension. Conclusions PRO-CTCAE-Simplified Chinese has been successfully developed and linguistically validated in a sample of cancer patients residing in Singapore.

scholarly journals Patient-reported outcomes in drug development for hematology

Hematology ◽  
2015 ◽  
Vol 2015 (1) ◽  
pp. 496-500 ◽  
Author(s):  
Catherine Acquadro ◽  
Antoine Regnault
Keyword(s):  
Clinical Trials ◽  
Drug Development ◽  
Patient Reported Outcomes ◽  
Clinical Decision Making ◽  
Well Being ◽  
New Drugs ◽  
Specific Situation ◽  
Treatment Benefit ◽  
Hematologic Diseases ◽  
Patient Reported

Abstract Patient-reported outcomes (PROs) are any outcome evaluated directly by the patient himself and based on the patient's perception of a disease and its treatment(s). PROs are direct outcome measures that can be used as clinical meaningful endpoints to characterize treatment benefit. They provide unique and important information about the effect of treatment from a patient's view. However, PROs will only be considered adequate if the assessment is well-defined and reliable. In 2009, the FDA has issued a guidance, which defines good measurement principles to consider for PRO measures intended to give evidence of treatment benefit in drug development. In hematologic clinical trials, when applied rigorously, they may be used to evaluate overall treatment effectiveness, treatment toxicity, and quality of patient's well-being at short-term and long-term after treatment from a patient's perspective. In situations in which multiple treatment options exist with similar survival outcome or if a new therapeutic strategy needs to be evaluated, the inclusion of PROs as an endpoint can provide additional data and help in clinical decision making. Given the diversity of the hematological field, the approach to measurement needs to be tailored for each specific situation. The importance of PROs in hematologic diseases has been highlighted in a number of international recommendations. In addition, new perspectives in the regulatory field will enhance the inclusion of PRO endpoints in clinical trials in hematology, allowing the voice of the patients with hematologic diseases to be taken into greater consideration in the development of new drugs.

Psychosocial profiles and their predictors in epilepsy using patient‐reported outcomes and machine learning

Epilepsia ◽  
2020 ◽  
Vol 61 (6) ◽  
pp. 1201-1210 ◽  
Author(s):  
Colin B. Josephson ◽  
Jordan D. T. Engbers ◽  
Meng Wang ◽  
Kevin Perera ◽  
Pamela Roach ◽  
...  
Keyword(s):  
Machine Learning ◽  
Patient Reported Outcomes ◽  
Patient Reported ◽  
Psychosocial Profiles

scholarly journals Benefits of patient-reported outcomes in Dermatology Drug Development

2015 ◽  
Vol 18 (3) ◽  
pp. A184 ◽  
Author(s):  
C. Copleymerriman ◽  
S. Zelt ◽  
M. Clark ◽  
A. Gnanasakthy
Keyword(s):  
Drug Development ◽  
Patient Reported Outcomes ◽  
Patient Reported

Patient-Reported Outcomes in Sinonasal and Skull Base Malignancy: An Assessment of Study Quality and Clinical Relevance

2020 ◽  
Vol 34 (6) ◽  
pp. 822-829
Author(s):  
Kody G. Bolk ◽  
Kelly A. Roth ◽  
Arun Sharma ◽  
Dana L. Crosby
Keyword(s):  
Quality Of Life ◽  
Skull Base ◽  
Patient Reported Outcomes ◽  
A Priori ◽  
High Quality ◽  
Exclusion Criteria ◽  
Future Studies ◽  
Patient Reported ◽  
Specific Symptoms

Background Sinonasal and skull base malignancies can cause significant adverse effects on functional status and survival. Objective The goal of this study was to systematically review the published literature of patient-reported outcomes pertaining to treatment of sinonasal and skull base malignancy. Methods A systematic literature search of Medline was conducted with PubMed to identify studies that assessed patient-reported outcomes in patients with sinonasal or skull base malignancy. Patient-reported outcomes studies with at least 10 patients published in English from January 2000 to April 2017 were included. Criteria from International Society for Quality of Life guidelines and criteria unique to sinonasal and skull base malignancies were used to calculate a composite score for each article. Studies with the top 33% of scores were categorized as high quality articles. Results Twenty-two articles met inclusion/exclusion criteria. Three studies (14%) reported a priori hypothesis. Eleven (50%) assessed specific quality of life domains and 10 studies (45%) performed statistical analysis on these domains. Specific symptoms were assessed in up to 32% of studies. Eight studies were characterized as high quality; these studies had higher sample sizes and more often assessed patient-reported outcomes prior to treatment compared to low quality studies. Conclusions The goal of the current study was to evaluate the quality of the current patient-reported outcomes literature on sinonasal and skull base malignancies. Areas of improvement for future studies include analysis of individual domains and disease-specific symptoms, reporting a priori hypotheses, and collecting preoperative and longitudinal patient-reported outcomes data.

Augmenting machine learning algorithms to predict mortality using patient-reported outcomes in oncology.

2021 ◽  
Vol 39 (15_suppl) ◽  
pp. 1510-1510
Author(s):  
Ravi Bharat Parikh ◽  
Jill Schnall ◽  
Manqing Liu ◽  
Peter Edward Gabriel ◽  
Corey Chivers ◽  
...  
Keyword(s):  
Quality Of Life ◽  
Machine Learning ◽  
Mortality Risk ◽  
Patient Reported Outcomes ◽  
Machine Learning Algorithms ◽  
Measurement Information ◽  
Patients With Cancer ◽  
Patient Reported ◽  
Mortality Table

1510 Background: Machine learning (ML) algorithms based on electronic health record (EHR) data have been shown to accurately predict mortality risk among patients with cancer, with areas under the curve (AUC) generally greater than 0.80. While patient-reported outcomes (PROs) may also predict mortality among patients with cancer, it is unclear whether routinely-collected PROs improve the predictive performance of EHR-based ML algorithms. Methods: This cohort study included 8600 patients with cancer who had an outpatient encounter at one of 18 medical oncology practices in a large academic health system between July 1st, 2019 and January 1st, 2020. 4692 (54.9%) patients completed assessments of symptoms, performance status, and quality of life from the PRO version of the Common Terminology Criteria for Adverse Events and the Patient-Reported Outcomes Measurement Information System Global v.1.2 scales. We hypothesized that ML models predicting 180-day all-cause mortality based on EHR + PRO data would improve AUC compared to ML models based on EHR data alone. We assessed univariate and adjusted associations between each PRO and 180-day mortality. To train the EHR-only model, we fit a Least Absolute Shrinkage and Selection Operator (LASSO) regression using 192 EHR demographic, comorbidity, and laboratory variables. To train the EHR + PRO model, we used a two-phase approach to fit a model using EHR data for all patients and PRO data for those who completed assessments. To test our hypothesis, we compared the bootstrapped AUC, area under the precision-recall curve (AUPRC), and sensitivity at a 20% risk threshold for both models. Results: 464 (5.4%) patients died within 180 days of the encounter. Decreased quality of life, functional status, and appetite were associated with greater 180-day mortality (Table). Compared to the EHR-only model, the EHR + PRO model significantly improved AUC (0.86 [95% CI 0.85-0.86] vs. 0.80 [95% CI 0.80-0.81]), AUPRC (0.40 [95% CI 0.37-0.42] vs. 0.30 [95% CI 0.28-0.32]), and sensitivity (0.45 [95% CI 0.42-0.48] vs. 0.33 [95% CI 0.30-0.35]). Conclusions: Routinely collected PROs augment EHR-based ML mortality risk algorithms. ML algorithms based on EHR and PRO data may facilitate earlier supportive care for patients with cancer. Association of PROs with 180-day mortality.[Table: see text]

329 Generation of patient-reported outcomes during early drug development

2015 ◽  
Vol 51 ◽  
pp. S64
Author(s):  
C. Cleeland ◽  
B. Cuffel ◽  
K. Keating ◽  
B. Childs ◽  
D. Walsh ◽  
...  
Keyword(s):  
Drug Development ◽  
Patient Reported Outcomes ◽  
Patient Reported ◽  
Early Drug

scholarly journals Measuring patients' views: a bifactor model of distinct patient-reported outcomes in psychosis

2010 ◽  
Vol 41 (2) ◽  
pp. 277-289 ◽  
Author(s):  
U. Reininghaus ◽  
R. McCabe ◽  
T. Burns ◽  
T. Croudace ◽  
S. Priebe
Keyword(s):  
Therapeutic Relationship ◽  
Treatment Satisfaction ◽  
Patient Reported Outcomes ◽  
A Priori ◽  
Bifactor Model ◽  
Subjective Quality Of Life ◽  
Needs For Care ◽  
Care Treatment ◽  
Patient Reported ◽  
Sum Scores

BackgroundPatient-reported outcomes (PROs) are widely used for evaluating the care of patients with psychosis. Previous studies have reported a considerable overlap in the information captured by measures designed to assess different outcomes. This may impair the validity of PROs and makes an a priori choice of the most appropriate measure difficult when assessing treatment benefits for patients. We aimed to investigate the extent to which four widely established PROs [subjective quality of life (SQOL), needs for care, treatment satisfaction and the therapeutic relationship] provide distinct information independent from this overlap.MethodAnalyses, based on item response modelling, were conducted on measures of SQOL, needs for care, treatment satisfaction and the therapeutic relationship in two large samples of patients with psychosis.ResultsIn both samples, a bifactor model matched the data best, suggesting sufficiently strong concept factors to allow for four distinct PRO scales. These were independent from overlap across measures due to a general appraisal tendency of patients for positive or negative ratings and shared domain content. The overlap partially impaired the ability of items to discriminate precisely between patients from lower and higher PRO levels. We found that widely used sum scores were strongly affected by the general appraisal tendency.ConclusionsFour widely established PROs can provide distinct information independent from overlap across measures. The findings may inform the use and further development of PROs in the evaluation of treatments for psychosis.

Sign in / Sign up

Export Citation Format

Share Document