Abstract:Purpose:
This study aimed to identify cases of developmental stuttering and associated comorbidities in de-identified electronic health records (EHRs) at Vanderbilt University Medical Center, and, in turn, build and test a stuttering prediction model.
Methods:
A multi-step process including a keyword search of medical notes, a text-mining algorithm, and manual review was employed to identify stuttering cases in the EHR. Confirmed cases were compared to matched controls … Show more
“…Model validation testing in an independent dataset containing manually reviewed records resulted in a positive prediction rate of 83.3% ( Table 1 ). 21 Applying this model in BioVU resulted in a higher proportion of individuals with imputed developmental stuttering (∼10%) than were observed by diagnostic code (0.15%) or manual chart review. 21 Figure 1 Outline of PheML development and application Within a set of 3.1 million deidentified electronic health records (A), we first identified a small pool of subjects (B) with developmental stuttering through expert manual review.…”
Section: Introductionmentioning
confidence: 90%
“… 21 Applying this model in BioVU resulted in a higher proportion of individuals with imputed developmental stuttering (∼10%) than were observed by diagnostic code (0.15%) or manual chart review. 21 Figure 1 Outline of PheML development and application Within a set of 3.1 million deidentified electronic health records (A), we first identified a small pool of subjects (B) with developmental stuttering through expert manual review. We selected these patients and their demographically matched controls to identify comorbidities as predictive features and develop and test a machine-learning model (C) that would impute stuttering in BioVU (D), an independent EHR dataset linked to genetic data.…”
Section: Introductionmentioning
confidence: 90%
“… 20 However, the billing codes traditionally used to assess patient status in the electronic health record are heavily underreported for developmental stuttering. 21 …”
Section: Introductionmentioning
confidence: 99%
“…Even if a patient were to seek evaluation and treatment for developmental stuttering, speech evaluations are typically performed by speech-language pathologists, usually outside of a hospital context (e.g., schools or private clinics). 21 There are also currently no FDA-approved medications or medical procedures to treat developmental stuttering, making it significantly less likely to be noted in a medical setting. Finally, although treatment exists for stuttering in the form of therapy and even though this is a chronic condition for many adults, this condition is not considered a parity diagnosis, and as such most government and private insurance plans do not cover treatment costs for stuttering.…”
Section: Introductionmentioning
confidence: 99%
“…In our previous research, we identified individuals affected by stuttering by applying a phenotype-driven machine-learning algorithm (PheML) that uses commonly reported phenotypes significantly associated with clinically diagnosed developmental stuttering as predictor variables to impute a developmental stuttering phenotype in BioVU (individuals with this phenotype are predicted by PheML to be affected by stuttering). 21 Our model takes a series of binary proxy parameters to impute developmental stuttering in a patient population by using a Gini impurity-based classification and regression tree classifier. 22 The PheML algorithm was built and tested with an initial pool of manually reviewed records from individuals affected by stuttering in Vanderbilt University’s EHR (no subjects within BioVU were used) across all ancestries ( Figure 1 ).…”
Developmental stuttering is a speech disorder characterized by disruption in the forward movement of speech. This disruption includes part-word and single-syllable repetitions, prolongations, and involuntary tension that blocks syllables and words, and the disorder has a life-time prevalence of 6-12%. Within Vanderbilt's electronic health record (EHR)-linked biorepository (BioVU), only 142 individuals out of 92,762 participants (0.15%) are identified with diagnostic ICD9/10 codes, suggesting a large portion of people who stutter do not have a record of diagnosis within the EHR. To identify individuals affected by stuttering within our EHR, we built a PheCode-driven Gini impurity-based classification and regression tree model, PheML, by using comorbidities enriched in individuals affected by stuttering as predicting features and imputing stuttering status as the outcome variable. Applying PheML in BioVU identified 9,239 genotyped affected individuals (a clinical prevalence of $10%) for downstream genetic analysis. Ancestry-stratified GWAS of PheML-imputed affected individuals and matched control individuals identified rs12613255, a variant near CYRIA on chromosome 2 (B ¼ 0.323; p value ¼ 1.31 3 10 À8 ) in European-ancestry analysis and rs7837758 (B ¼ 0.518; p value ¼ 5.07 3 10 À8 ), an intronic variant found within the ZMAT4 gene on chromosome 8, in African-ancestry analysis. Polygenic-risk prediction and concordance analysis in an independent clinically ascertained sample of developmental stuttering cases validate our GWAS findings in PheML-imputed affected and control individuals and demonstrate the clinical relevance of our population-based analysis for stuttering risk.
“…Model validation testing in an independent dataset containing manually reviewed records resulted in a positive prediction rate of 83.3% ( Table 1 ). 21 Applying this model in BioVU resulted in a higher proportion of individuals with imputed developmental stuttering (∼10%) than were observed by diagnostic code (0.15%) or manual chart review. 21 Figure 1 Outline of PheML development and application Within a set of 3.1 million deidentified electronic health records (A), we first identified a small pool of subjects (B) with developmental stuttering through expert manual review.…”
Section: Introductionmentioning
confidence: 90%
“… 21 Applying this model in BioVU resulted in a higher proportion of individuals with imputed developmental stuttering (∼10%) than were observed by diagnostic code (0.15%) or manual chart review. 21 Figure 1 Outline of PheML development and application Within a set of 3.1 million deidentified electronic health records (A), we first identified a small pool of subjects (B) with developmental stuttering through expert manual review. We selected these patients and their demographically matched controls to identify comorbidities as predictive features and develop and test a machine-learning model (C) that would impute stuttering in BioVU (D), an independent EHR dataset linked to genetic data.…”
Section: Introductionmentioning
confidence: 90%
“… 20 However, the billing codes traditionally used to assess patient status in the electronic health record are heavily underreported for developmental stuttering. 21 …”
Section: Introductionmentioning
confidence: 99%
“…Even if a patient were to seek evaluation and treatment for developmental stuttering, speech evaluations are typically performed by speech-language pathologists, usually outside of a hospital context (e.g., schools or private clinics). 21 There are also currently no FDA-approved medications or medical procedures to treat developmental stuttering, making it significantly less likely to be noted in a medical setting. Finally, although treatment exists for stuttering in the form of therapy and even though this is a chronic condition for many adults, this condition is not considered a parity diagnosis, and as such most government and private insurance plans do not cover treatment costs for stuttering.…”
Section: Introductionmentioning
confidence: 99%
“…In our previous research, we identified individuals affected by stuttering by applying a phenotype-driven machine-learning algorithm (PheML) that uses commonly reported phenotypes significantly associated with clinically diagnosed developmental stuttering as predictor variables to impute a developmental stuttering phenotype in BioVU (individuals with this phenotype are predicted by PheML to be affected by stuttering). 21 Our model takes a series of binary proxy parameters to impute developmental stuttering in a patient population by using a Gini impurity-based classification and regression tree classifier. 22 The PheML algorithm was built and tested with an initial pool of manually reviewed records from individuals affected by stuttering in Vanderbilt University’s EHR (no subjects within BioVU were used) across all ancestries ( Figure 1 ).…”
Developmental stuttering is a speech disorder characterized by disruption in the forward movement of speech. This disruption includes part-word and single-syllable repetitions, prolongations, and involuntary tension that blocks syllables and words, and the disorder has a life-time prevalence of 6-12%. Within Vanderbilt's electronic health record (EHR)-linked biorepository (BioVU), only 142 individuals out of 92,762 participants (0.15%) are identified with diagnostic ICD9/10 codes, suggesting a large portion of people who stutter do not have a record of diagnosis within the EHR. To identify individuals affected by stuttering within our EHR, we built a PheCode-driven Gini impurity-based classification and regression tree model, PheML, by using comorbidities enriched in individuals affected by stuttering as predicting features and imputing stuttering status as the outcome variable. Applying PheML in BioVU identified 9,239 genotyped affected individuals (a clinical prevalence of $10%) for downstream genetic analysis. Ancestry-stratified GWAS of PheML-imputed affected individuals and matched control individuals identified rs12613255, a variant near CYRIA on chromosome 2 (B ¼ 0.323; p value ¼ 1.31 3 10 À8 ) in European-ancestry analysis and rs7837758 (B ¼ 0.518; p value ¼ 5.07 3 10 À8 ), an intronic variant found within the ZMAT4 gene on chromosome 8, in African-ancestry analysis. Polygenic-risk prediction and concordance analysis in an independent clinically ascertained sample of developmental stuttering cases validate our GWAS findings in PheML-imputed affected and control individuals and demonstrate the clinical relevance of our population-based analysis for stuttering risk.
ImportanceDevelopmental language disorder (DLD) is a common (with up to 7% prevalence) yet underdiagnosed childhood disorder whose underlying biological profile and comorbidities are not fully understood, especially at the population level.ObjectiveTo identify clinically relevant conditions that co-occur with DLD at the population level.Design, Setting, and ParticipantsThis case-control study used an electronic health record (EHR)–based population-level approach to compare the prevalence of comorbid health phenotypes between DLD cases and matched controls. These cases were identified using the Automated Phenotyping Tool for Identifying Developmental Language Disorder algorithm of the Vanderbilt University Medical Center EHR, and a phenome enrichment analysis was used to identify comorbidities. An independent sample was selected from the Geisinger Health System EHR to test the replication of the phenome enrichment using the same phenotyping and analysis pipeline. Data from the Vanderbilt EHR were accessed between March 2019 and October 2020, while data from the Geisinger EHR were accessed between January and March 2022.Main Outcomes and MeasuresCommon and rare comorbidities of DLD at the population level were identified using EHRs and a phecode-based enrichment analysis.ResultsComorbidity analysis was conducted for 5273 DLD cases (mean [SD] age, 16.8 [7.2] years; 3748 males [71.1%]) and 26 353 matched controls (mean [SD] age, 14.6 [5.5] years; 18 729 males [71.1%]). Relevant phenotypes associated with DLD were found, including learning disorder, delayed milestones, disorders of the acoustic nerve, conduct disorders, attention-deficit/hyperactivity disorder, lack of coordination, and other motor deficits. Several other health phenotypes not previously associated with DLD were identified, such as dermatitis, conjunctivitis, and weight and nutrition, representing a new window into the clinical complexity of DLD.Conclusions and RelevanceThis study found both rare and common comorbidities of DLD. Comorbidity profiles may be leveraged to identify risk of additional health challenges, beyond language impairment, among children with DLD.
Making a specific diagnosis in neurodevelopmental disorders is traditionally based on recognizing clinical features of a distinct syndrome, which guides testing of its possible genetic etiologies. Scalable frameworks for genomic diagnostics, however, have struggled to integrate meaningful measurements of clinical phenotypic features. While standardization has enabled generation and interpretation of genomic data for clinical diagnostics at unprecedented scale, making the equivalent breakthrough for clinical data has proven challenging. However, increasingly clinical features are being recorded using controlled dictionaries with machine readable formats such as the Human Phenotype Ontology (HPO), which greatly facilitates their use in the diagnostic space. Improving the tractability of large-scale clinical information will present new opportunities to inform genomic research and diagnostics from a clinical perspective. Here, we describe novel approaches for computational phenotyping to harmonize clinical features, improve data translation through revising domain-specific dictionaries, quantify phenotypic features, and determine clinical relatedness. We demonstrate how these concepts can be applied to longitudinal phenotypic information, which represents a critical element of developmental disorders and pediatric conditions. Finally, we expand our discussion to clinical data derived from electronic medical records, a largely untapped resource of deep clinical information with distinct strengths and weaknesses.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
