Discrepancy in diagnoses of diabetes and prediabetes using fasting plasma glucose and glycosylated hemoglobin and the underdiagnosis by ICD-10 coding: data from a tertiary hospital in Thailand.

Background: Early detection of prediabetes and diabetes better prevents long-term health complications. FPG and HbA1c levels are some common laboratory tests utilized as tools to diagnose diabetes and prediabetes, but the agreement rate between these two diagnostic tests varies, which could lead to underdiagnosis and thus undertreatment. This study aimed to analyze the agreement rate between FPG and HbA1c, as well as the physicians' accuracy of using these results to make a prediabetes or diabetes diagnosis through ICD-10 coding at a tertiary care hospital in Bangkok, Thailand. Methods: A cross-sectional descriptive study was conducted using secondary data collected in a tertiary hospital's check-up clinic from August 16, 2019 to June 30, 2022 to study the prevalence and diagnosis of diabetes and prediabetes, determined through FPG and HbA1c laboratory results. We analyzed the two laboratory tests' diagnosis agreement rate and the physicians' accuracy of diagnosing diabetes and prediabetes in ICD-10 coding using the FPG and HbA1c results. Results: Among 8,024 asymptomatic participants, the period prevalence diagnosed through laboratory results was 5.8% for diabetes and 19.8% for prediabetes. Diabetes and prediabetes prevalence based on laboratory data differs from that of ICD-10 coding data. Specifically, 79.6% of diabetes patients and 32.3% of prediabetes patients were coded using the ICD-10 coding system. 4,094 individuals had both FPG and HbA1c data. The agreement rate for diagnosing diabetes and prediabetes between the two laboratory results is 89.5%, with Kappa statistics of 0.58. Using only one of the two laboratory results would have missed a substantial number of patients. Conclusion: Our findings highlight screening test discrepancies and underdiagnosis issues that impede diagnostic accuracy enhancement and refined patient management strategies. Early diagnoses of prediabetes and diabetes, especially before symptoms arise, could increase health consciousness in individuals, thereby enabling the implementation of lifestyle modifications and prevention of serious health complications. We emphasize the importance of diagnosing these conditions using both FPG and HbA1c, along with subsequent accurate ICD-10 coding. Even though some hospitals lack certified HbA1c testing, we suggest enhancing the availability of HbA1c testing, which could benefit many people in Thailand.