Artificial neural networks outperform linear regression in estimating 9-month patient-reported outcomes after upper extremity fractures with increasing number of variables
Abstract. Objective:. To compare performance between linear regression (LR) and artificial neural network (ANN) models in estimating 9-month patient-reported outcomes (PROs) after upper extremity fractures using various subsets of early mental, social, and physical health variables. Methods:. We stu...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wolters Kluwer
2024-01-01
|
| Series: | OTA International |
| Online Access: | http://journals.lww.com/10.1097/OI9.0000000000000284 |
| _version_ | 1797341005133905920 |
|---|---|
| author | Niels Brinkman, BS Romil Shah, MD Job Doornberg, MD, PhD David Ring, MD, PhD Stephen Gwilym, DPhil, FRCS(Orth) Prakash Jayakumar, MD, PhD |
| author_facet | Niels Brinkman, BS Romil Shah, MD Job Doornberg, MD, PhD David Ring, MD, PhD Stephen Gwilym, DPhil, FRCS(Orth) Prakash Jayakumar, MD, PhD |
| author_sort | Niels Brinkman, BS |
| collection | DOAJ |
| description | Abstract. Objective:. To compare performance between linear regression (LR) and artificial neural network (ANN) models in estimating 9-month patient-reported outcomes (PROs) after upper extremity fractures using various subsets of early mental, social, and physical health variables.
Methods:. We studied 734 patients with isolated shoulder, elbow, or wrist fracture who completed demographics, mental and social health measures, and PROs at baseline, 2–4 weeks, and 6–9 months postinjury. PROs included 3 measures of capability (QuickDASH, PROMIS-UE-PF, PROMIS-PI) and one of pain intensity. We developed ANN and LR models with various selections of variables (20, 23, 29, 34, and 54) to estimate 9-month PROs using a training subset (70%) and internally validated them using another subset (15%). We assessed the accuracy of the estimated value being within one MCID of the actual 9-month PRO value in a test subset (15%).
Results:. ANNs outperformed LR in estimating 9-month outcomes in all models except the 20-variable model for capability measures and 20-variable and 23-variable models for pain intensity. The accuracy of ANN versus LR in the primary model (29-variable) was 83% versus 73% (Quick-DASH), 68% versus 65% (PROMIS-UE-PF), 66% versus 62% (PROMIS-PI), and 78% versus 65% (pain intensity). Mental and social health factors contributed most to the estimations.
Conclusion:. ANNs outperform LR in estimating 9-month PROs, particularly with a larger number of variables. Given the otherwise relatively comparable performance, aspects such as practicality of collecting greater sets of variables, nonparametric distribution, and presence of nonlinear correlations should be considered when deciding between these statistical methods. |
| first_indexed | 2024-03-08T10:12:31Z |
| format | Article |
| id | doaj.art-464676695a56435ba234a29752d07ac1 |
| institution | Directory Open Access Journal |
| issn | 2574-2167 |
| language | English |
| last_indexed | 2024-03-08T10:12:31Z |
| publishDate | 2024-01-01 |
| publisher | Wolters Kluwer |
| record_format | Article |
| series | OTA International |
| spelling | doaj.art-464676695a56435ba234a29752d07ac12024-01-29T07:02:03ZengWolters KluwerOTA International2574-21672024-01-0171S10.1097/OI9.0000000000000284OI90000000000000284Artificial neural networks outperform linear regression in estimating 9-month patient-reported outcomes after upper extremity fractures with increasing number of variablesNiels Brinkman, BS0Romil Shah, MD1Job Doornberg, MD, PhD2David Ring, MD, PhD3Stephen Gwilym, DPhil, FRCS(Orth)4Prakash Jayakumar, MD, PhD5a Department of Surgery and Perioperative Care, The University of Texas at Austin, Dell Medical School, Austin, TXa Department of Surgery and Perioperative Care, The University of Texas at Austin, Dell Medical School, Austin, TXb Department of Orthopaedic & Trauma Surgery, University Medical Center, Groningen, the Netherlandsa Department of Surgery and Perioperative Care, The University of Texas at Austin, Dell Medical School, Austin, TXd The Nuffield Department of Orthopaedics, Rheumatology, and Musculoskeletal Sciences, The University of Oxford, Oxford, United Kingdoma Department of Surgery and Perioperative Care, The University of Texas at Austin, Dell Medical School, Austin, TXAbstract. Objective:. To compare performance between linear regression (LR) and artificial neural network (ANN) models in estimating 9-month patient-reported outcomes (PROs) after upper extremity fractures using various subsets of early mental, social, and physical health variables. Methods:. We studied 734 patients with isolated shoulder, elbow, or wrist fracture who completed demographics, mental and social health measures, and PROs at baseline, 2–4 weeks, and 6–9 months postinjury. PROs included 3 measures of capability (QuickDASH, PROMIS-UE-PF, PROMIS-PI) and one of pain intensity. We developed ANN and LR models with various selections of variables (20, 23, 29, 34, and 54) to estimate 9-month PROs using a training subset (70%) and internally validated them using another subset (15%). We assessed the accuracy of the estimated value being within one MCID of the actual 9-month PRO value in a test subset (15%). Results:. ANNs outperformed LR in estimating 9-month outcomes in all models except the 20-variable model for capability measures and 20-variable and 23-variable models for pain intensity. The accuracy of ANN versus LR in the primary model (29-variable) was 83% versus 73% (Quick-DASH), 68% versus 65% (PROMIS-UE-PF), 66% versus 62% (PROMIS-PI), and 78% versus 65% (pain intensity). Mental and social health factors contributed most to the estimations. Conclusion:. ANNs outperform LR in estimating 9-month PROs, particularly with a larger number of variables. Given the otherwise relatively comparable performance, aspects such as practicality of collecting greater sets of variables, nonparametric distribution, and presence of nonlinear correlations should be considered when deciding between these statistical methods.http://journals.lww.com/10.1097/OI9.0000000000000284 |
| spellingShingle | Niels Brinkman, BS Romil Shah, MD Job Doornberg, MD, PhD David Ring, MD, PhD Stephen Gwilym, DPhil, FRCS(Orth) Prakash Jayakumar, MD, PhD Artificial neural networks outperform linear regression in estimating 9-month patient-reported outcomes after upper extremity fractures with increasing number of variables OTA International |
| title | Artificial neural networks outperform linear regression in estimating 9-month patient-reported outcomes after upper extremity fractures with increasing number of variables |
| title_full | Artificial neural networks outperform linear regression in estimating 9-month patient-reported outcomes after upper extremity fractures with increasing number of variables |
| title_fullStr | Artificial neural networks outperform linear regression in estimating 9-month patient-reported outcomes after upper extremity fractures with increasing number of variables |
| title_full_unstemmed | Artificial neural networks outperform linear regression in estimating 9-month patient-reported outcomes after upper extremity fractures with increasing number of variables |
| title_short | Artificial neural networks outperform linear regression in estimating 9-month patient-reported outcomes after upper extremity fractures with increasing number of variables |
| title_sort | artificial neural networks outperform linear regression in estimating 9 month patient reported outcomes after upper extremity fractures with increasing number of variables |
| url | http://journals.lww.com/10.1097/OI9.0000000000000284 |
| work_keys_str_mv | AT nielsbrinkmanbs artificialneuralnetworksoutperformlinearregressioninestimating9monthpatientreportedoutcomesafterupperextremityfractureswithincreasingnumberofvariables AT romilshahmd artificialneuralnetworksoutperformlinearregressioninestimating9monthpatientreportedoutcomesafterupperextremityfractureswithincreasingnumberofvariables AT jobdoornbergmdphd artificialneuralnetworksoutperformlinearregressioninestimating9monthpatientreportedoutcomesafterupperextremityfractureswithincreasingnumberofvariables AT davidringmdphd artificialneuralnetworksoutperformlinearregressioninestimating9monthpatientreportedoutcomesafterupperextremityfractureswithincreasingnumberofvariables AT stephengwilymdphilfrcsorth artificialneuralnetworksoutperformlinearregressioninestimating9monthpatientreportedoutcomesafterupperextremityfractureswithincreasingnumberofvariables AT prakashjayakumarmdphd artificialneuralnetworksoutperformlinearregressioninestimating9monthpatientreportedoutcomesafterupperextremityfractureswithincreasingnumberofvariables |