A clinical audit on International Classification of Primary Care coding in 4 primary care clinics and common errors in coding International Classification of Primary Care - 2

Wing-Kwun Lam 林永, King-Yip Ho 何敬業, Kwok-Keung Ng 吳國強, Kon-Hung Kwok 郭冠雄, Luke C Y Tsang 曾昭義

HK Pract 2005;27:344-352

Summary

The International Classification of Primary Care - (revised version) (ICPC-2) is a useful classification system, which allows classification of the patient's reason for encounter (RFE), the problems/diagnoses managed, interventions, and the ordering of these data in an episode of care structure. In this clinical audit, we identified deficiencies in the correctness and completeness of coding of diagnosis and health problems in our clinics and noted common errors in coding ICPC-2. We implemented changes to improve our coding practice. In this project, we found that clinical audit can be useful in improving the correctness and completeness of coding.

摘要

國際基層醫療分類(ICPC)-2是一個有效的分類系統，能夠分類病人的求診原因、問題或診斷、 治療，並將每次的資料按次序加以整理。本次臨床審計中， 我們發現了診所工作中分類編碼診斷及健康問題的正確性及全面性方面的缺陷和常見的錯誤， 所以採取了方法加以改善。於本課題中，我們發現臨床審計能有效改善編碼的正確性及全面性。

Introduction

The International Classification of Primary Care (ICPC), developed by the ICPC Working Party, broke new ground in the world of classification when it was published in 1987 by the World Organisation of National Colleges, Academies, and Academic Associations of General Practitioners/Family Physicians (WONCA), now known more briefly as the World Organisation of Family Doctors; or just Wonca.¹ In 1998 Wonca published a revised version of ICPC (ICPC-2) with inclusion and exclusion criteria attached to the classification rubrics, and a mapping to International Classification of Disease 10th Edition (ICD-10).¹ ICPC-2 allows classification of the patient's reason for encounter (RFE), the problems/diagnoses managed, interventions, and the ordering of these data in an episode of care structure.²

Our Professional Development and Quality Assurance unit of the Department of Health has four primary care clinics. ICPC coding has been implemented in two of our clinics for some years now. We believe that it is useful in many aspects. The coding of health problems and diagnoses of patients help us improve the quality of our work through clinical audits, administrative monitoring and other continuous improvement processes. Our database also provides data for research on morbidity patterns and planning of resource allocation. To this end, we plan to implement coding of health problems and diagnoses in all four clinics. We decided to use ICPC-2 as it is the new international standard and it also allows mapping to ICD-10 directly.

From our experience, the quality of our earlier ICPC database was not satisfactory. Many of our coding of health problems and diagnoses were incomplete as many health problems were managed but not coded. There was a lack of standardization in using the ICPC coding system, which resulted in incorrect coding. Most of our doctors received no training in coding, and so were deficient in their knowledge of and confidence in doing the coding. Therefore, we decided to do a clinical audit on our ICPC-2 coding after the implementation of ICPC-2 in all our clinics.

Objectives

1. To audit the correctness of coding of diagnoses or health problems in our four clinics.
2. To audit the completeness of coding of diagnoses or health problems in our four clinics.
3. To identify common errors in coding in ICPC-2 in our clinics.
4. To standardize and to improve the coding practices in our clinics.

Method

We searched several databases including Medline, Embase, Cochrane library of databases and Google, and we found very few publications which could help us to write "evidence-based" criteria for our clinical audit. Therefore we wrote to the Wonca International Classification Committee (WICC) for their expert advice. We also contacted the editor of the ICPC-2 Chinese version published by WICC and the Hong Kong College of Family Physicians for his advice.

The ICPC-2 book is a very informative resource on ICPC-2. An electronic version of the ICPC-2 is also available in the WICC website of WONCA (http://www.globalfamilydoctor.com/WICC/) which helps us to look up the criteria of each code automatically using computers. The "rules" for coding diagnoses/health problems in the ICPC-2 book and the criteria of ICPC-2 were used as the gold standard for assessing the "correctness" of our coding.

We formed an ICPC-2 working group comprising four experienced family physicians each with special interest in ICPC-2 coding from each clinic to standardize the coding practice and to settle any discrepancies in coding. The four experienced family physicians are all Fellows of the Hong Kong College of Family Physicians and Australian College of General Practitioners, and three of them are also Fellows of Hong Kong Academy of Medicine (Family Medicine).

Although the completeness of coding diagnoses and health problems could be best assessed by video taping the consultations or by direct observation of consultations, we found that it is not practical to do so. Therefore, the "completeness" of coding was assessed by the completeness of coding of diagnoses and health problems as recorded in our medical records.

We decided to use two "Must Do" Criteria:

1. All ICPC-2 codes of health problems or diagnoses recorded in individual Patients' Records of a clinical encounter should be correct in accordance with the rules and criteria set in the ICPC-2 book.
2. All ICPC-2 codes of health problems or diagnoses recorded in individual Patients' Records of a clinical encounter should be completely coded.

The ICPC-2 coding record of the clinic from a one-week period in late 2004 was collected. We used a systematic sampling in which every third record in the coding sheet from each consultation room was retrieved.We aimed at a standard of 90% for the "completeness" and "correctness" of coding and opted for a 10% limit of confidence interval in measuring the standard. The sample size was calculated using a statistical formula by Samuel et al (1993).

144 patient encounters from each clinic were targeted and the actual sample size collected was greater than 144 as we wished to assess enough medical records from every doctor clinic for individualized feedback. The members of the ICPC-2 working group assessed the medical records for the completeness and correctness of coding based on the two criteria and meet regularly to settle any discrepancies in coding by consensus. Common errors in coding were also identified during the process.

Results of first data collection

Results of the first data collection in our four clinics are summarized in Table 1.
Clinic A and Clinic B newly implemented ICPC-2 in their clinics in late 2004 while Clinic C and Clinic D upgraded their system from ICPC to ICPC-2 during the same period.
Common error in coding ICPC-2 is tabulated in Table 2.
Psychological and social problems were frequently not coded, which might lead to unreliable morbidity pattern data. Many doctors coded the symptoms of the encounter only while it should be more specific to code the diagnoses. Some doctors found it difficult to decide on which chapter of ICPC-2 to use, haemorrhoid (a circulatory problem) and sleep apnoea (a psychological problem) being typical examples.

Problems encountered

The first audit results were presented at the clinic meeting within each clinic. The doctors showed awareness for the need to improve. During the meetings, their concerns and difficulties were expressed. The common problems included:

1. Insufficient time to enter detailed problem lists in the busy clinics.
2. Difficulties in coding vague complaints without definite diagnoses.
3. Individual differences in interpreting those unclear codes in ICPC-2.
4. Lack of consulting channels when difficulties in coding were encountered.
5. Lack of effective feedback system on the correctness of coding.
6. Lack of communication among colleagues to standardize the coding practice.

Implementation of changes

Changes were implemented in the two months which followed in all our four clinics. The changes included:

1. Meetings among change-facilitators from each clinic were held to adopt improvement suggestions.
2. There was sharing and feedback of the audit result during clinic lunch meetings in the form of small group discussion.
3. Individualized feedback and educational outreaching of coordinators of the ICPC-2 working group to individual doctors.
4. A guideline book on ICPC-2 coding which included the common errors in coding and criteria of all the ICPC-2 codes was distributed to all doctors.
5. Coordination through the ICPC group coordinators of all ICPC-2 related activities.
6. Standardization of coding practices and resolution of unclear codes by the ICPC-2 working group by having regular monthly meetings.
7. Provision of resource persons in each clinic to resolve difficulties in coding.
8. Invitation of local experts on ICPC-2 to deliver a workshop on ICPC-2 for our unit.

Results of further data collection

Clinics A, C and D decided to perform a second data collection after the first one and implementation of changes. Clinic B, however, decided that little could be gained from a second data collection within a short period of time as the 95% confidence interval of the correctness and completeness of coding did not have a significant difference from the target standard of 90%.

We are glad to see improvements in correctness and completeness of coding in the second data collection in most clinics. The results are summarized in Table 4.

Third data Collection

Clinic D was still not satisfied with the correctness of the coding and decided to do a third data collection after reinforcement of the changes in the following month. The results are summarized in Table 5.

Discussion

The coding of health problems and diagnoses of patients helps the primary care physicians to improve the quality of their work through clinical audits, administrative monitoring and other continuous improvement processes. The quality of the coding database is important to achieve these goals. Sharing of this project may help other primary care physicians to improve their own coding practice and perhaps to propagate the use of ICPC-2 by other clinics especially the primary care doctors in private practice.

A Cochrane systematic review concluded that audit and feedback can be effective in improving professional practice.³ In our experience, we found that clinical audit is useful in improving patient care in chronic diseases. For example, in a multicenter clinical audit on diabetes care in our clinics, there were improvements in all process performance, including diet review, exercise advice, smoking habit assessment, hypoglycaemic attacks assessment, complications record, feet examination, fundi examination, urine check for albumin, glycosylated haemoglobin check and blood lipids check.⁴ In this project, we found that clinical audit leads to improvement in our ICPC-2 coding. Although coding does not directly lead to improvement in patient care, it provides an infrastructure to facilitate future clinical audits and other quality improvement projects.

However, there are limitations in our methodology. The completeness of coding was assessed by medical record review, which could be more valid by video taping and transcribing the consultations and by direct observation of the consultations. Furthermore, one clinic did not complete the whole audit cycle as the target was met in the first cycle already. Clinical audit is a spiral process, which means that the target could be higher in subsequent data collection.

At the other end, one clinic completed three rounds of data collection as the target was not reached in the first two rounds. We have adopted a more pragmatic approach appropriate to the clinical setting as resource for doing clinical audit is limited.

Our unit has workshops on evidence-based clinical audit on a regular basis. Participants report increase in confidence in their ability to conduct audit, use the Cochrane Library, develop review criteria, collect data and implement change.⁵ The present project was shared at one of the workshops. We hope to convey the idea that clinical audit can be simple and achievable by busy clinicians. It does not require sophisticated statistical analysis and it leads to improvement in practice.

Clinical audit is a very useful tool to monitor the standard of care for chronic illnesses.⁶ In our project, it shows that clinical audit can be applied to other aspects of our daily practice. We are planning to continuously monitor our coding and provide feedback to our doctors to ensure integrity of the database for use in future clinical audit, epidemiological and administrative work. Other frontline doctors who use ICPC-2 may find from our errors in ICPC-2 coding useful hints for improving their own practice. Although using ICPC-2 effectively involves painstaking training and practice, we believe that it is well compensated by its potential to improve quality of our work through clinical audits, administrative monitoring and other continuous improvement processes.

Conclusion

This clinical audit project proved to be a valuable tool for understanding the common problems in ICPC-2 coding and bringing about improvement. The objectives of this project were met. Through this project, the coding practice of our unit was standardized through individualized feedback, educational outreaching of resource persons, guideline development and formation of a working group to continuously monitor the process

Acknowledgement

The authors would like to thank all colleagues of Professional Development and Quality Assurance of the Department of Health, especially doctors and clerical staff for their selfless contribution to the project. We would also like to give our special thanks to Dr Y T Wun for his guidance and teachings on ICPC-2. We would also want to thank Professor Helena Britt and Professor Niels Bentzen of Wonca International Classification Committee for their assistance in setting up our coding project in the initial stage. Finally, we would like to thank Ms Joyce Tsang and Ms Angie Chan for their support in the project.

Key messages

1. The International Classification of Primary Care (revised version) (ICPC-2) is a useful classification system, which allows classification of the patient's reason for encounter (RFE), the problems/diagnoses managed, interventions, and the process of care.
2. The coding of health problems and diagnoses of patients helps the primary care physicians to improve quality of their work through facilitation of clinical audits, administrative monitoring and other continuous improvement processes.
3. The correctness and completeness of coding were used as "must do" criteria in this ICPC-2 audit project.
4. Factors that may affect correctness and completeness of coding include insufficient time, difficulties in coding vague complaints, misinterpretation of codes, lack of consultation channels in difficult codes, lack of feedback for wrong coding and lack of communications among doctors to standardize the coding practice.
5. Some clinical encounters were incompletely coded and in particular, psychological and social problems were frequently not coded.
6. Coding practice could be standardized and improved through individualized feedback to doctors, educational outreaching of resource persons, guideline development and formation of a working group to continuously monitoring the process.
7. Clinical audit is a very useful tool to monitor the standard of care for chronic illnesses. In this project, it also proved to be a useful tool for understanding the common problems in ICPC-2 coding and bringing about improvement.