Incidence of Clinical Coding Errors and Implications on Casemix Reimbursement in a Teaching Hospital in Malaysia
Last reviewed: March 2026
Key Findings
- An audit of 464 patient medical records found that 89.4% (415/464) contained at least one clinical coding error in diagnosis or procedure code assignment.
- Secondary diagnosis coding errors were most prevalent at 81.3%, followed by secondary procedures (58.2%), principal procedures (50.9%), and primary diagnoses (49.8%).
- Coding errors resulted in different MY-DRG® code assignments in 74.0% of affected cases, with 52.1% receiving a lower hospital tariff than warranted.
- The estimated potential loss of hospital income due to coding errors was RM654,303.91 — a substantial financial impact with implications for sustainable hospital funding.
Background
Clinical coding is the systematic process of translating medical diagnoses, procedures, and services into standardised alphanumeric codes using internationally recognised classification systems. In Malaysia, the International Classification of Diseases 10th Revision (ICD-10) is used for diagnosis coding, while the International Classification of Diseases 9th Revision Clinical Modification (ICD-9-CM) is used for procedure coding. These codes form the foundation of the Malaysia Diagnosis Related Group (MY-DRG®) casemix system, which classifies patients into clinically and economically homogeneous groups for the purposes of resource allocation, hospital budgeting, and performance monitoring.
The MY-DRG® casemix system was first officially implemented in Malaysian teaching hospitals in 2002 and has since been extended across the Ministry of Health hospital network. In the casemix framework, diagnosis and procedure codes are processed through a software grouper that assigns each patient episode to a specific DRG code. This DRG code is then linked to a cost weight and hospital tariff that determines the reimbursement level for that episode of care. The accuracy of clinical coding is therefore directly connected to the financial sustainability of hospitals operating under casemix-based funding models.
Clinical coding errors can occur at multiple levels: incorrect assignment of principal diagnosis codes, omission or misassignment of secondary diagnosis codes (including comorbidities and complications), and errors in procedure coding. These errors propagate through the casemix grouper, potentially resulting in incorrect DRG assignment and inappropriate hospital tariff allocation — either too high (over-reimbursement) or too low (under-reimbursement).
Study Design and Methods
This retrospective study was conducted at Hospital Canselor Tuanku Mukhriz (HCTM), a teaching hospital affiliated with Universiti Kebangsaan Malaysia Medical Centre. A total of 464 coded patient medical records were selected for audit. An independent senior coder (ISC) was appointed to re-examine and re-code the selected records. The ISC’s codes were compared against the original hospital coder’s codes, and where disagreements existed, the ISC’s codes were taken as the accurate reference standard. The re-coded cases were then re-grouped using the MY-DRG® grouper to assess changes in DRG assignment and hospital tariff. All outcomes were verified by a casemix expert to ensure the validity of the audit findings. Ethical approval was obtained from the Ethics Committee at Universiti Kebangsaan Malaysia (Reference Number: UKM.1.5.3.5/224/UNU-003-2014).
Coding Error Rates
| Coding Category | Error Rate | Records with Errors |
|---|---|---|
| Any coding error (overall) | 89.4% | 415 out of 464 |
| Secondary diagnoses | 81.3% | 377 out of 464 |
| Secondary procedures | 58.2% | 270 out of 464 |
| Principal procedures | 50.9% | 236 out of 464 |
| Primary diagnoses | 49.8% | 231 out of 464 |
The overall coding error rate of 89.4% was strikingly high, indicating systemic deficiencies in the clinical coding process at the study hospital. The finding that secondary diagnosis coding was the most error-prone category (81.3%) is particularly significant because secondary diagnoses — which capture comorbidities and complications — directly influence the severity classification within the MY-DRG® system. Accurate coding of comorbidities is essential for correct severity adjustment, which in turn determines the appropriate cost weight and tariff assignment.
The high error rate in principal procedure coding (50.9%) and primary diagnosis coding (49.8%) was equally concerning, as these codes determine the initial DRG group assignment before severity adjustment. Errors at this level can result in patients being classified into entirely inappropriate DRG categories, with correspondingly inappropriate tariff assignments.
Financial Impact
The coding errors resulted in changes to the MY-DRG® code assignment in 74.0% (307/415) of the cases with coding errors. Among the cases with changed DRG assignments, 52.1% (160/307) were assigned a lower hospital tariff than would have been appropriate based on the ISC’s corrected codes. The cumulative potential loss of hospital income attributable to coding errors was calculated at RM654,303.91. This figure represents a substantial financial impact for a single hospital over the audit period and, if extrapolated across the Malaysian public hospital system, suggests that coding errors may result in millions of ringgit in annual revenue shortfalls.
It is important to note that coding errors can result in both under-reimbursement and over-reimbursement. The predominance of under-reimbursement in this study (52.1% of changed cases receiving lower tariffs) suggests that the hospital was systematically under-coding the severity and complexity of patient episodes, resulting in DRG assignments that did not fully capture the resources consumed during care delivery.
Causes of Coding Errors
The study identified several contributing factors to the high coding error rates. Inadequate medical documentation by treating physicians was a primary driver — clinical coders can only code what is documented in the medical record, and incomplete or ambiguous clinical documentation inevitably leads to coding inaccuracies. Insufficient coder training and experience, particularly in complex clinical scenarios requiring nuanced code selection, was another important factor. The complexity of the ICD-10 and ICD-9-CM classification systems, which contain thousands of codes with fine distinctions between related conditions, creates inherent challenges even for experienced coders.
The lack of regular coding audits and quality assurance programmes at the hospital level also contributed to the persistence of coding errors. Without systematic feedback mechanisms, coders had limited opportunities to identify and correct recurring error patterns. The absence of standardised coding guidelines for common clinical scenarios further compounded the problem, as individual coders may interpret ambiguous documentation differently.
Recommendations
The study recommended several strategies to improve clinical coding quality. Intensive re-training programmes for clinical coders, with emphasis on secondary diagnosis coding and procedure coding, should be implemented as an immediate priority. Regular coding audits — ideally on a quarterly basis — should be established as a routine quality assurance activity, with feedback provided to individual coders to support continuous improvement. Medical documentation improvement programmes targeting treating physicians should be developed, emphasising the importance of documenting all relevant diagnoses, comorbidities, and procedures with sufficient specificity for accurate coding.
The development of standardised coding guidelines for high-volume clinical scenarios — drawing on international best practices and adapted for the Malaysian context — would reduce coding variability and improve consistency. Investment in computer-assisted coding (CAC) technologies, which can flag potential coding discrepancies and suggest appropriate codes based on clinical documentation, may also help reduce error rates as these technologies mature.
Significance
This study represents one of the most comprehensive assessments of clinical coding quality in the Malaysian casemix system. The findings have been widely cited in subsequent research on healthcare financing and hospital management in Malaysia. The study’s documentation of the financial implications of coding errors provided concrete evidence to support investment in coding quality improvement — demonstrating that the cost of re-training programmes and audit infrastructure is likely to be offset many times over by the recovery of under-captured hospital revenue. A companion paper published in BMC Health Services Research (2018) extended these findings with additional statistical analysis.
Limitations
The study was conducted at a single teaching hospital, which may have coding patterns and resource profiles that differ from district hospitals and specialist centres. Teaching hospitals typically manage more complex cases, which may contribute to higher coding error rates. The use of a single ISC as the reference standard, while verified by a casemix expert, introduces the possibility of reviewer bias. The study did not assess the root causes of coding errors in detail (e.g., whether errors arose from documentation deficiencies, coder knowledge gaps, or time pressure), which would be valuable for designing targeted interventions.
Zafirah SA, Amrizal MN, Wan Puteh SE, Aljunid SM. Incidence of Clinical Coding Errors and Implications on Casemix Reimbursement in a Teaching Hospital in Malaysia. Malaysian Journal of Public Health Medicine. 2017;17(2):19–28.
License: Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)