1
|
Patel RH, Fan L, Kelly NR, Gelsey F, Hertzberg JK, Brnabic AJM. A machine learning-based algorithm to identify U-500R insulin candidates among adults with type 2 diabetes mellitus in US retrospective databases. Curr Med Res Opin 2024; 40:367-375. [PMID: 38259227 DOI: 10.1080/03007995.2023.2293116] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 02/16/2023] [Accepted: 12/06/2023] [Indexed: 01/24/2024]
Abstract
OBJECTIVE To develop a machine learning-based predictive algorithm to identify patients with type 2 diabetes mellitus (T2DM) who are candidates for initiation of U-500R insulin (U-500R). METHODS A retrospective cohort of patients with T2DM was used from a large US administrative claims and electronic health records (EHR) database affiliated with Optum. Predictor variables derived from the data were used to identify appropriate supervised machine learning models including least absolute shrinkage and selection operator (LASSO) and extreme gradient boosted (XGBoost) methods. Predictive performance was assessed using precision-recall (PR) and receiver operating characteristic (ROC) area under the curve (AUC). The clinical interpretation of the final model was supported by fitting the final set of variables from the LASSO and XGBoost models to a traditional logistic regression model. Model choice was determined by comparing Akaike Information Criterion (AIC), residual deviances, and scaled Brier scores. RESULTS Among 81,242 patients who met the study eligibility criteria, 577 initiated U-500R and were assigned to the positive class. Predictors of U-500R initiation included overweight/obesity, neuropathy, HbA1c ≥9% and 8%-9%, BUN 23.8 to <112 mg/dl, ALT 35.9-2056.2 U/L, no radiological chest exams, no GFR labs, and gait/mobility abnormalities. The best performing model was the LASSO model with an ROC AUC of 0.776 on the hold-out test set. CONCLUSION This study successfully developed and validated a machine learning-based algorithm to identify U-500R candidates among patients with T2DM. This may help health care providers and decision-makers to understand important characteristics of patients who could use U-500R therapies which in turn could support policies and guidelines for optimal patient management.
Collapse
Affiliation(s)
| | - Ludi Fan
- Eli Lilly and Company, Indianapolis, IN, USA
| | | | | | | | | |
Collapse
|
2
|
Bali V, Turzhitsky V, Schelfhout J, Paudel M, Hulbert E, Peterson-Brandt J, Hertzberg J, Kelly NR, Patel RH. Machine learning to identify chronic cough from administrative claims data. Sci Rep 2024; 14:2449. [PMID: 38291064 PMCID: PMC10828499 DOI: 10.1038/s41598-024-51522-9] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/02/2023] [Accepted: 01/06/2024] [Indexed: 02/01/2024] Open
Abstract
Accurate identification of patient populations is an essential component of clinical research, especially for medical conditions such as chronic cough that are inconsistently defined and diagnosed. We aimed to develop and compare machine learning models to identify chronic cough from medical and pharmacy claims data. In this retrospective observational study, we compared 3 machine learning algorithms based on XG Boost, logistic regression, and neural network approaches using a large claims and electronic health record database. Of the 327,423 patients who met the study criteria, 4,818 had chronic cough based on linked claims-electronic health record data. The XG Boost model showed the best performance, achieving a Receiver-Operator Characteristic Area Under the Curve (ROC-AUC) of 0.916. We selected a cutoff that favors a high positive predictive value (PPV) to minimize false positives, resulting in a sensitivity, specificity, PPV, and negative predictive value of 18.0%, 99.6%, 38.7%, and 98.8%, respectively on the held-out testing set (n = 82,262). Logistic regression and neural network models achieved slightly lower ROC-AUCs of 0.907 and 0.838, respectively. The XG Boost and logistic regression models maintained their robust performance in subgroups of individuals with higher rates of chronic cough. Machine learning algorithms are one way of identifying conditions that are not coded in medical records, and can help identify individuals with chronic cough from claims data with a high degree of classification value.
Collapse
Affiliation(s)
- Vishal Bali
- Center for Observational and Real-World Evidence (CORE), Merck & Co, Rahway, NJ, USA.
| | - Vladimir Turzhitsky
- Center for Observational and Real-World Evidence (CORE), Merck & Co, Rahway, NJ, USA
| | - Jonathan Schelfhout
- Center for Observational and Real-World Evidence (CORE), Merck & Co, Rahway, NJ, USA
| | - Misti Paudel
- Health Economics and Outcomes Research (HEOR), Optum Insight, Eden Prairie, MN, USA
| | - Erin Hulbert
- Health Economics and Outcomes Research (HEOR), Optum Insight, Eden Prairie, MN, USA
| | | | | | | | | |
Collapse
|
3
|
Eby EL, Kelly NR, Hertzberg JK, Blodgett MC, Stubbins C, Patel RH, Meadows ES, Benneyworth BD, Faries DE. Predicting Response to Bolus Insulin Therapy in Patients With Type 2 Diabetes. J Diabetes Sci Technol 2023; 17:1573-1579. [PMID: 35596567 PMCID: PMC10658685 DOI: 10.1177/19322968221098057] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 11/15/2022]
Abstract
BACKGROUND The aim of this study was to develop a predictive model to classify people with type 2 diabetes (T2D) into expected levels of success upon bolus insulin initiation. METHODS Machine learning methods were applied to a large nationally representative insurance claims database from the United States (dNHI database; data from 2007 to 2017). We trained boosted decision tree ensembles (XGBoost) to assign people into Class 0 (never meeting HbA1c goal), Class 1 (meeting but not maintaining HbA1c goal), or Class 2 (meeting and maintaining HbA1c goal) based on the demographic and clinical data available prior to initiating bolus insulin. The primary objective of the study was to develop a model capable of determining at an individual level, whether people with T2D are likely to achieve and maintain HbA1c goals. HbA1c goal was defined at <8.0% or reduction of baseline HbA1c by >1.0%. RESULTS Of 15 331 people with T2D (mean age, 53.0 years; SD, 8.7), 7800 (50.9%) people met HbA1c goal but failed to maintain that goal (Class 1), 4510 (29.4%) never attained this goal (Class 0), and 3021 (19.7%) people met and maintained this goal (Class 2). Overall, the model's receiver operating characteristic (ROC) was 0.79 with greater performance on predicting those in Class 2 (ROC = 0.92) than those in Classes 0 and 1 (ROC = 0.71 and 0.62, respectively). The model achieved high area under the precision-recall curves for the individual classes (Class 0, 0.46; Class 1, 0.58; Class 2, 0.71). CONCLUSIONS Predictive modeling using routine health care data reasonably accurately classified patients initiating bolus insulin who would achieve and maintain HbA1c goals, but less so for differentiation between patients who never met and who did not maintain goals. Prior HbA1c was a major contributing parameter for the predictions.
Collapse
|
4
|
Kelly NR, Harding JT, Fulton JE, Kozinetz CA. A randomized controlled trial of a video module to increase U.S. poison center use by low-income parents. Clin Toxicol (Phila) 2013; 52:54-62. [PMID: 24299276 DOI: 10.3109/15563650.2013.863328] [Citation(s) in RCA: 4] [Impact Index Per Article: 0.4] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/13/2022]
Abstract
CONTEXT U.S. poison centers decrease medical visits by providing telephone advice for home management of potential poisonings, but are underutilized by low-income African-American and Latino parents, and those with limited English proficiency, due to lack of knowledge and misconceptions about poison centers. OBJECTIVES To assess the effectiveness of a poison prevention video module in improving knowledge, behavior, and behavioral intention concerning use of poison centers in a population of low-income, language-diverse adults attending parenting courses offered by a community organization. METHODS A randomized, blinded, controlled trial was conducted at 16 parenting course sites of a community organization and included 297 participants. The organization's instructors presented the video module (intervention) or the usual class curriculum (control). Participants completed questionnaires at baseline and a telephone interview 2-4 weeks later. Changes from baseline to follow-up were compared between the intervention and control groups using analysis of variance and Chi-square tests. Intervention group participants were stratified by English proficiency and compared to assess baseline and follow-up responses by language. RESULTS After the intervention, participants in the intervention group had a significantly greater increase in knowledge about the poison center, were more likely to have the correct poison center phone number at home, and had greater behavioral intention to use the poison center compared to control group participants. At baseline, Spanish-primary-language participants with limited English proficiency had less knowledge about the poison center, were less likely to have the poison center number at home, and had lower behavioral intention to use the poison center than English proficient participants, but significantly improved after the intervention. CONCLUSIONS This video module, when presented by a community organization's instructors, was highly effective in improving knowledge, behavior, and behavioral intention concerning use of poison centers within a low-income, language-diverse population.
Collapse
Affiliation(s)
- N R Kelly
- Department of Pediatrics, University of TX Southwestern Medical Center , Dallas, TX , USA
| | | | | | | |
Collapse
|
5
|
Abstract
The nurse who graduates from a nurse practitioner program leaves a comfort zone of nursing practice to enter a new position where feelings of insecurity and stress are common. Because the role of the nurse practitioner (NP) continues to evolve and is influenced by many environmental issues, the preparation of the NP by the academic institution needs to be assessed on a frequent basis. The purpose of this research was to obtain a better understanding of the transitional phase to the first position as NP after graduation. The perceptions of preparation, gains, losses, barriers, facilitators, and strategies for adjustment were explored. A qualitative approach using focus groups was developed in which 21 recent NP graduates from a large university participated in one of four focus groups. Peer debriefing and participant verification were techniques used to ensure credibility and trustworthiness of the data and subsequent analysis. The themes identified were: loss of personal control of time and privacy; changes and losses in relationships; feelings of isolation and uncertainty in establishing the NP role; and a special bonding with clients. Although the participants perceived they were adequately prepared for their role, they also described feelings of guilt and uncertainty from not knowing information they believed they should know. They concluded that they functioned differently than the physician as well as other nurses and found it necessary to distance themselves from the role of other health care providers. This study has implications for colleagues, academic institutions, and the individual nurse practitioner.
Collapse
Affiliation(s)
- N R Kelly
- Peoria Regional Program, University of Illinois at Chicago, 61656-1649, USA
| | | |
Collapse
|
6
|
Kelly NR, Groff JY. Exploring barriers to utilization of poison centers: a qualitative study of mothers attending an urban Women, Infants, and Children (WIC) Clinic. Pediatrics 2000; 106:199-204. [PMID: 10888692] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [MESH Headings] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 02/17/2023] Open
Abstract
OBJECTIVE Prior research regarding poison center utilization identified risk factors for underutilization including race/ethnicity and acculturation. The purpose of this study was to understand factors contributing to underutilization of poison centers by low-income and minority mothers. METHODS Focus groups were conducted with women attending an urban Women, Infants and Children clinic. Transcripts were analyzed for themes and key points. RESULTS Twenty-two English-speaking mothers and 21 Spanish-speaking mothers participated in 7 groups. Participants viewed poisoning as a serious problem to which all children are susceptible. English-speaking mothers had heard of the poison center but were unaware of services provided. They preferred to use the 911 system, which was viewed as immediate medical assistance and was an easy number to remember. Women questioned the credentials of the poison center staff. Spanish-speaking mothers had limited knowledge of poison centers and were concerned about language barriers. CONCLUSIONS To increase utilization of poison centers, educational interventions must address these specific needs and misconceptions and should be produced in Spanish and English versions.
Collapse
Affiliation(s)
- N R Kelly
- Department of Pediatrics, Baylor College of Medicine and Family Practice and Community Medicine, University of Texas-Houston Medical School, Houston, Texas, USA.
| | | |
Collapse
|
7
|
Abstract
Retirement exemplifies one of the most important transitions in life. A job provides structure to daily life, adds to a sense of self-worth and self-esteem, and is a means of self-expression, all of which define and identify the person. This work activity links individuals to a society and to a social network. When an individual retires, there is a loss of one role and a beginning of a new one. The purpose of this research was to explore the preretirement period of nurses to better understand the transitional process. A qualitative approach was chosen to enable the researchers to explore in-depth perceptions and opinions of this subjective experience. Interview questions were developed from the literature to determine the perceptions of preparation, barriers, and facilitators in the process, losses, gains, and coping strategies that were helpful. A purposive sample of female nurses from academic institutions and hospitals were invited to participate in one of three focus groups. All were invited to a fourth focus group session during which the findings from the researchers were presented and the participants verified the results. Of the 19 participants, 11 had retired, whereas 8 were in the preretirement phase. Four nurses had accepted an incentive retirement offered at their institution. The conversations were taped, transcribed, and analyzed. Thematic analysis identified both positive and negative aspects of the experience with concerns related to finances, timing of retirement, discretionary time, and relationships. Participants frequently expressed feelings of inadequacy with the numerous decisions related to finances and wished the institution had provided more information. They perceived the most enjoyable gain to be the time to enjoy and do the things they wished. The prevalent feeling among the groups was that the loss of work friends was very difficult. Old relationships would need to be fostered and new ones developed in the retiree role. Based on the findings, support services, including advisers and counselors, are recommended to ease the movement through this transitional process. The assumption is that if individuals have a better understanding of the retirement process and their new role, they will prepare and adjust better.
Collapse
Affiliation(s)
- N R Kelly
- Peoria Regional Graduate Nursing Program, University of Illinois at Chicago 61656-1649, USA
| | | |
Collapse
|
8
|
Abstract
The purpose of this study was to identify and characterize caretakers who fail to utilize the poison center for unintentional poisonings involving children. We interviewed 210 caretakers of children evaluated for unintentional poisoning in the emergency center of an urban, university-based teaching hospital to determine (1) whether demographic differences exist between those caretakers who contacted a poison center prior to the emergency center visit and those who did not and (2) whether differences exist in prevalence of poison prevention knowledge and behaviors between the two groups. Ninety-six (46%) of caretakers did not contact the poison center prior to the emergency center visit. Significant differences were found between the two groups for the following caretaker variables: race/ethnicity, language preference, age, level of education, country in which schooling occurred, and type of insurance coverage for the child. When logistic regression was used to control for confounding, the two variables associated with failure to use the poison center were black race and schooled outside the United States (primarily in Mexico). Poison center callers reported a higher prevalence of poison prevention knowledge and behaviors than noncallers. Educational interventions should be targeted to the groups of caretakers identified who do not use the poison center.
Collapse
Affiliation(s)
- N R Kelly
- Department of Pediatrics, Baylor College of Medicine, Texas, USA
| | | | | | | | | | | |
Collapse
|
9
|
Kelly NR, Ellis MD, Kirkland RT, Holmes SE, Kozinetz CA. Effectiveness of a poison center: impact on medical facility visits. Vet Hum Toxicol 1997; 39:44-8. [PMID: 9004468] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [MESH Headings] [Subscribe] [Scholar Register] [Indexed: 02/03/2023]
Abstract
To estimate the effectiveness of a poison center by determining the number of potential medical facility visits for childhood poisonings which are prevented, and to determine how callers learned about the poison center in order to help guide future public educational efforts, a cross-sectional telephone survey was done. A systematic sample of caretakers of children < 6-y-of-age living in Harris County, TX, and who called the poison center about a poisoning incident between February 1993 and January 1994, was taken. One hundred sixty-six/197 eligible caretakers (84%) completed the survey. Of the 166 caretakers, only 5 (3%) were referred by the poison center to a medical facility for treatment, although 6 (3.6%) actually sought medical attention. The remaining 160 (96.4%) were successfully managed at home. The majority of these (74%) required no intervention and the remaining 26% required only minimal treatment, including ipecac, dilution, irrigation or observation. Of the 160 caretakers managed at home, 69 (43%) would have sought medical attention for the child at a health care facility if the poison center did not exist, and of those 81% would have gone to an emergency center or hospital for evaluation. Callers learned about the poison center most often from family/ friends, doctors, Mr Yuk stickers, television, previous use, or pharmacists. Poison centers decrease health care costs by preventing unnecessary medical facility visits for minor childhood poisonings which can be successfully managed at home. Methods effective in educating the public about poison centers should be continued and other methods explored to increase poison centers' use.
Collapse
Affiliation(s)
- N R Kelly
- Department of Pediatrics, Baylor College of Medicine, Houston, TX 77030, USA
| | | | | | | | | |
Collapse
|
10
|
Abstract
Although schools of nursing market their programs to all segments of society, men are underrepresented in schools of nursing. The purpose of this study was to identify male students' perceptions of the motivational factors, barriers and frustrations encountered in becoming a nurse. Eighteen students attended one of four focus groups representing one associate program, two baccalaureate programs and one diploma program. The students' belief that society perceived nursing as a feminine profession was an underlying thread that related to many of their perceptions and feelings. The most influential support people noted were immediate family, especially wives. The participants also believed that high school counselors were of no assistance in choosing nursing as a career. Even though the schools of nursing were perceived as supportive, the participants had feelings of isolation and self-doubt.
Collapse
Affiliation(s)
- N R Kelly
- University of Illinois College of Nursing, Peoria Regional Program 61656-1649, USA
| | | | | |
Collapse
|
11
|
Kelly NR, Ellis EN. Proteinuria in children. J Ark Med Soc 1991; 88:219-23. [PMID: 1804838] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [MESH Headings] [Subscribe] [Scholar Register] [Indexed: 12/28/2022]
Affiliation(s)
- N R Kelly
- Department of Pediatrics, Arkansas Children's Hospital, Little Rock
| | | |
Collapse
|
12
|
Magee JS, Steele RW, Kelly NR, Jacobs RF. Tularemia transmitted by a squirrel bite. Pediatr Infect Dis J 1989; 8:123-5. [PMID: 2704604] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [MESH Headings] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 01/02/2023]
Affiliation(s)
- J S Magee
- Department of Pediatrics, University of Arkansas for Medical Sciences, Little Rock
| | | | | | | |
Collapse
|
13
|
|