IADR Abstract Archives
ICD-9 and ICD-10 Code Mapping: Diagnosis to Dental Treatment
Objectives: Coding of procedures in the medical (ICD-9 & ICD-10) and dental (ADA) records are not integrated to assist dentists in decision-making. The objective of this study was to map diagnosis codes in the medical record to inform dentists whether treatment is prescribed or not prescribed, needing referrals to specialists.
Methods: 12 articles were obtained using the American Diabetes Association, EBSCOhost, PubMed, and Google. Ten papers were obtained, reviewed, and analyzed. Codes were mapped using Neo4j graph database principles. For logistics reasons, the study focused on the condition Diabetes. Codes, edge-probabilities, and Glycemic index were used to construct the major traversal nodes to outcomes: Treatment, yes, and Treatment, no. When data was evidenced, probabilities were averaged using Shapiro-Wilks at p=.05 significance level.
Results: The map began with the Patient node and traversed to the Diabetes node. 3 nodes were mapped at this point: Diabetes Mellitus Type I, Type II, and Conditions associated with Diabetes, child nodes Type I and Type II. Attribute nodes with edge-probabilities were mapped. These included ICD-9 or ICD-10 codes and age: Type I, <20 years of age (US population), .35%. Of the adolescent, diabetic population, 71% had Type I, 29% had Type II. Subnodes were mapped according to the Glycemic index as: Low (Hypoglycemic), test measure range 55 or less; Normal, test measure 55-70; and High (Hyperglycemic), test measure 55-70. Treatment, yes node was applicable to only the Normal test measure range. Treatment, no was associated with Low and High test measure ranges.
Conclusions: Mapping of medical codes, ICD-9 and ICD-10, may be useful to dentists for enhancing decision-making. Coding of a condition and or test measure may assist dentists in proceeding to treatment plan formulation or be alerted to the need for specialist consult prior to initiating treatment.
Methods: 12 articles were obtained using the American Diabetes Association, EBSCOhost, PubMed, and Google. Ten papers were obtained, reviewed, and analyzed. Codes were mapped using Neo4j graph database principles. For logistics reasons, the study focused on the condition Diabetes. Codes, edge-probabilities, and Glycemic index were used to construct the major traversal nodes to outcomes: Treatment, yes, and Treatment, no. When data was evidenced, probabilities were averaged using Shapiro-Wilks at p=.05 significance level.
Results: The map began with the Patient node and traversed to the Diabetes node. 3 nodes were mapped at this point: Diabetes Mellitus Type I, Type II, and Conditions associated with Diabetes, child nodes Type I and Type II. Attribute nodes with edge-probabilities were mapped. These included ICD-9 or ICD-10 codes and age: Type I, <20 years of age (US population), .35%. Of the adolescent, diabetic population, 71% had Type I, 29% had Type II. Subnodes were mapped according to the Glycemic index as: Low (Hypoglycemic), test measure range 55 or less; Normal, test measure 55-70; and High (Hyperglycemic), test measure 55-70. Treatment, yes node was applicable to only the Normal test measure range. Treatment, no was associated with Low and High test measure ranges.
Conclusions: Mapping of medical codes, ICD-9 and ICD-10, may be useful to dentists for enhancing decision-making. Coding of a condition and or test measure may assist dentists in proceeding to treatment plan formulation or be alerted to the need for specialist consult prior to initiating treatment.