1. Introduction

In family medicine, it is important to code not only the diseases but also the health-related data of the individuals who apply for whatever reason. The recorded data needs to be evaluated, archived, and when necessary, coded as a standard, especially for research. It is important that the coding is suitable for both primary healthcare institutions and can be used in integration with other health institutions. Today, there is a need for these codings to comply with international and global standards. In 1988, the World Organization of Family Doctors (WONCA) published the International Classification of Primary Care (ICPC-2) coding parallel to the existing International Statistical Classification of Diseases and Related Health Problems, briefly ICD (ICD-10). However, ICPC-2 was first published in 1998, with a mapping to the ICD-10. Like the ICD-10, the World Health Organization (WHO) accepted the ICPC-2 as a reason for encounter classification of diagnoses in primary healthcare services [1, 2].

While ICD-10 is more suitable for hospitals that involve a single department and apply to more specific problems, ICPC-2 is more suitable for primary healthcare institutions where individuals apply to complex health problems involving many departments. The holistic, comprehensive, and continuous care feature of family medicine necessitates the simple but systematic registration of symptoms/diagnoses/interventions in individuals who come with many complex problems. The classification to be used in family medicine should be specific to primary care, and WONCA’s recommendations are important in determining its standards [3].

The ICD, which was adopted and entered into force in 1948 under the auspices of WHO, is constantly updated until today. The ICD-10 version provides encoding of health-related data within 5 levels and 21 divisions (Table 1). With these coded data, evaluation of health services, policy development, cost calculations and payments, and especially scientific research and planning can be done [4]. On the other hand, ICPC-2 has a biaxial structure and consists of 17 sections and 7 components (Table 2) on one axis [5, 6].

ICD-10 is used for the classification of diseases and disorders in primary healthcare practice in different countries and Turkey. In a study in America, it is stated that although records are made with diagnosis codes, they should also be evaluated in terms of the use of ICPC-2, which is more practical in international criteria. It has been stated that a technical conversion between ICPC-2 and ICD-10 since they started to be used is practically always possible and can be applied by primary care physicians [7].

In a study, it is emphasized that the common concepts of the 3 different classifications of disease, disability, and health interventions in the family of international classifications of the WHO should be reconciled [8]. For a similar purpose, the WHO is working on a model that will include the interaction between the individual’s health status, and environmental, and personal factors. This model focuses on the International Classification of Functioning, Disability, and Health (ICF) using a standard and unified language [9]. It also mentions the necessity of coding and classification for specific target groups, data collection and comparisons as well as classification of diseases and disorders. A standardized set of procedures is recommended for this purpose [10]. In another study, it is stated that the common use of different applications in the international classification of diseases by the WHO should be encouraged. For this purpose, a sample model is proposed for a uniform application [11]. These latest studies mentioned show that there is a need for some definitions and coding for health and health-related conditions. Efforts to code definitions by different institutions and clinics in different countries are certainly useful. However, today, as a result of globalization, there is a greater need for these efforts to be standardized by a common center and relevant academic institutions. For this reason, it would be appropriate for the WHO to work in coordination with the academies of relevant health clinics and gather these classifications under a single roof. It is important to act together with WONCA regarding primary healthcare services.

Family medicine, which provides holistic medicine practice in health systems and is the first point of contact in both urban and rural areas, is the most important step in terms of healthcare records [12]. In this study, we aimed to evaluate the data obtained with the ICD-10 coding of the individuals registered in the family health center in the last 5 years by comparing them with the ICPC-2 codes.

2. Materials and Methods

3. Results

The most common diagnoses in the study were J00-J99 at 16.5%, M00-M99 at 10.1%, Z00-Z99 at 18%, and E00-E90 at 8.5%. While the diagnoses of Q00-Q99, and I00-I99 were mostly made in male patients, the diagnoses of D50-D89, L00-L99, N00-N99, O00-O99, P00-P96, and U00-Y98 were mostly made in female patients (p < 0.05). It was observed that the age of the patients diagnosed as I00-I99 was 52.62 ± 14.32, higher than the other patient groups. The mean age in the diagnoses of R00-R99, D50-D89, S00-T98, U00-Y98, and Z00-Z99 was lower than the other groups (p < 0.05) (Table 1).

According to ICPC-2 diagnoses, A 19.7%, L 5.8%, and R 16.6% are among the most common diagnostic codes. According to ICPC-2 diagnoses, F and Y diagnoses were found to be higher in male participants, and B, U, W, and X codes were higher in females. It was observed that the mean age level in Y, T, and K diagnoses was higher than in other diagnosis groups. It can be stated that the mean age of patients diagnosed with Z is lower than the other groups (p > 0.05) (Table 2).

There may not be any diseases or diagnoses in some blocks in the coding order of both classifications. For example, while there is a disease code in block A98 in the ICD-10 classification, there is no disease coding in block A97. Similarly, there is no C chapter alphabetically in ICPC-2. It was determined that 86.2% of ICPC-2 A diagnoses corresponded to ICD-10 codes Z00-Z99, 6.6% corresponded to A00-B99, and 6% corresponded to E00-E90 ICD-10 codes. It was observed that ICPC-2 A diagnoses were included in codes corresponding to 6 different ICD 10 codes. It was observed that 87.1% of ICPC-2B diagnoses corresponded to ICD-10 codes D50-D89, 12.8% to C00-D48, and 0.1% to I00-I99. It was observed that ICPC-2B diagnoses were included in codes corresponding to 3 different ICD-10 codes.

It has been determined that the ICPC-2 codes F (eye), H (ear), and Z (social problems) correspond 100% to the codes in ICD-10.

It was determined that W diagnoses (Pregnancy, childbirth, family planning) in ICPC-2 were included in codes corresponding to 2 different ICD-10 codes, and X diagnoses (female genital system and breast) were included in codes corresponding to 3 different ICD-10 codes. On the other hand, Y diagnoses (male genital system) in ICPC-2 were included in codes corresponding to 12 different ICD-10 codes (Table 3).

In our study, a total of 42,782 diagnoses were determined according to the ICD-10 coding, when the data provided by the family physician in the practices in the family health center in the last 5 years were examined. These diagnoses were evaluated by comparison according to ICPC-2 coding. There were 43,027 diagnoses in ICPC-2 compared to 42,782 diagnoses in ICD-10. A total of 218 diagnoses with codes in ICD-10 did not have a counterpart in ICPC-2. On the other hand, a total of 463 (1.1%) diagnoses in ICD-10 were found to have two or more codes in ICPC-2.

All diagnoses in ICD-10 originating from diseases of the eye and its appendages (H00-H59), diseases of the ear and mastoid process (H60-H95), and the section on health status (Z00-Z99) were found in ICPC-2 (Supporting Table 1).

4. Discussion

According to the results of our study, the equivalents of ICD-10 and ICPC-2 codes were different in number and percentage. It was determined that 43,027 diagnoses in ICPC-2 corresponded to 42,782 diagnoses in ICD-10. Of the 42,782 diagnoses with codes in ICD-10, 218 (0.51%) did not have an equivalent in ICPC-2. On the other hand, it was determined that a total of 463 diagnoses (1.08%) in ICD-10 had two or more codes in ICPC-2 (Supporting Table 1).

In our study, although the diagnosis codes for the female gender had similar names according to the sections in both classifications, they were not equal in total. The codes O00-O99 (pregnancy, childbirth, and the puerperium) and P00-P99 (certain conditions originating in the perinatal period) in ICD-10 corresponded to the codes W (pregnancy, childbirth, family planning) and X (female genital system and breast) in ICPC-2. Despite this similarity, the total number of diagnoses was not equal in both classifications. Interestingly, the male gender was detected at a rate of 0.6% in the X (female genital system and breast) code in ICPC-2. It was unclear whether this coding result was due to the individual’s incorrect anamnesis, the physician’s coding error, or the program’s recording error. Or, were the codes for diagnoses belonging to men with real breast disease recorded in this section? This situation was unclear. Another notable deficiency in the classifications of this condition is that while there was Y (male genital system) in ICPC-2, there was no corresponding coding section in ICD-10 (Tables 1, 2, and 3).

Our study findings show the importance of the integration and synchronization of the classifications of diseases and health conditions that provide the necessary data for scientific research and especially extraordinary situations (natural disasters, epidemics, etc.) in our daily clinical practice.

In our study, among the codes in ICD-10, only the codes H00-H59 (diseases of the eye and its appendages) and H60-H95 (diseases of the ear and mastoid process) completely matched the F (eye) and H (ear) codes in ICPC-2 in terms of number and percentage. This match was perfect because they are anatomically and functionally single organs. However, in primary healthcare, family physicians often approach the individual holistically in ear, nose, and throat (ENT) diseases (e.g., epistaxis). As a result, the individual can be referred to the ENT polyclinic with the diagnosis determined according to his/her condition. The diagnosis determined in primary care may be the same or different diagnosis by the relevant ENT or other specialists in secondary and tertiary care. In this case, the family physician wants to know what the code provided in the primary care corresponds to in the secondary care, and it is important to know. Most physicians want to be able to see and evaluate the reciprocal of these diagnoses at once. Therefore, it is important that the codes in ICD-10 and ICPC-2 match each other and that the differences are classified.

In a study conducted in Korea, it is stated that the benefit of ICPC-2 has not yet been investigated for the country. For this purpose, ICPC-2 was compared with ICD-10 in order to evaluate its applicability in primary care. It was reported that the diagnoses in the same section were higher in number and percentage in ICD-10. In conclusion, it is stated that ICPC can be a complementary tool in the detection of common diseases [13]. There are some differences for ICPC in Africa. In a study, 220 disease diagnoses were compared in terms of ICD-10 and ICPC-2 codes. Although few differences were found, ICPC-2 was reported to be a valid and reliable classification for primary care data collection [14].

In our study, diagnoses differed in both codings in terms of numbers and percentages. While ICPC-2 classification is based primarily on localization, ICD-10 is based on etiology. Therefore, comparing both classifications at the chapter level is a limitation of the study. However, it has been evaluated in order to see the differences between the chapters at a glance and to raise awareness. In our study, the most common disease diagnosis was respiratory system diseases (J00-J99) with 16.5% in ICD-10 and 16.6% with R (respiratory) code in ICPC-2. The second most common disease diagnosis code is musculoskeletal and connective tissue diseases with 10.1%. This group is L (musculoskeletal) code with 5.8% in ICPC-2. The codes with the most common diagnoses in both classifications were the groups under the general headings. These groups were Z00-Z99 (health status and factors affecting utilization of health services) with 18% in ICD-10 and A (general and unspecified) code with 19.7% in ICPC-2.

In a study conducted on patients’ complaints, categorizing the complaints using ICPC-2 and analyzing the results shows that ICPC-2 is suitable for primary care research [15]. In a study conducted with the use of antipsychotic drugs, disease diagnosis codes could be used as diagnostic indicators thanks to the transformations in ICD-10 and ICPC-2 [16].

In our study, Z00-Z99 codes (factors affecting health status and contact with health services) covered all the diagnoses in the Z code (social problems) in ICPC-2 in numbers and percentages. In addition, the diagnosis codes D (digestive), N (neurological), P (psychological), R (respiratory), T (endocrine, metabolic, and nutritional), and U (urology) in ICPC-2 received only 1 code in ICD-10. Because they are gender-specific, W (pregnancy, childbirth, family planning), X (female genital system and breast), and Y (male genital system) codes are distributed at different rates in both classifications. On the other hand, A (general and unspecified), B (blood, blood from organs, lymphatics, spleen), K (circulatory), L (musculoskeletal), and S (skin) resulted in different codes as they are associated with many organs and tissues.

In another study, diagnoses of symptoms recorded by family physicians and physicians working out of hours were compared with diagnoses after discharge. Diagnoses can change after being hospitalized, for instance a family physician can refer a patient with abdominal pain and the discharge code is appendicitis. While ICD-10 diagnoses in abdominal pain and chest pain referrals were mostly symptom diagnoses, ICD-10 diagnoses in pneumonia, appendicitis, acute myocardial infarction, and stroke referrals were mostly disease diagnoses [17].

In one study, ICD-10 codes used in secondary care were converted to ICPC-2 to compare injuries treated in primary and secondary care. In one of the results of the study, it was determined that injuries were more common in men in general, but injuries were more common in women in advanced ages. This difference according to age was detected only in ICD-10 in the secondary care [18]. With respect to age, the age of those diagnosed with circulatory system diseases in our study was higher than those in other patient groups in both ICD-10 (I00-I99) and ICPC-2 (K).

A study in Germany reported that ICD-10-GM (GM = German Modification) is insufficient in coding some rare diseases in inpatients and that additional coding is needed [19]. In another study, it was reported that ICD-10 codes for cough, fever, and shortness of breath were not sensitive during COVID-19 testing [20]. Coding modifications made in different countries and the inadequacy of coding health-related conditions and diseases in special situations such as pandemics should not be ignored.

In a study conducted in Norway which has 16 years of experience using ICPC, the codes in the period when ICPC was not used (Transitional period) and the periods when ICPC-2 was used (Normal phase) were compared. As a result of the study, some codes were missing in problematic cases that did not match the ICPC-2 classification (pneumonia, anemia, tonsillitis, diabetes, etc.). While only missing codes were found during the transition phase, it was determined that incorrect and inappropriate codes were used throughout the entire process. Physicians could not make these diagnoses because some diagnoses were missing in the ICPC. With these results, the study reports that ICPC is not suitable for clinical practice in primary care [21]. On the other hand, a study conducted in Norway reported that a morbidity index based on ICPC-2 could be used as a correction variable in epidemiological research conducted in primary care [22].

In a study conducted in Australia, ICD-10, ICPC-2, and the Systematized Nomenclature of Medicine-Clinical Terms-Australia (SNOMED CT-AU) codings were compared for the diagnosis of low back pain (LBP). According to current guidelines, authors were asked to develop subcategories for LBP with these three encodings. There were quite a lot of codes for the same clinical condition in all 3 codings. As a result, it has been reported that all 3 codings are insufficient and need revision according to the current LBP guidelines [23].

In a study in Belgium, it was reported that the classification of diagnoses and conditions in the family physician’s electronic patient records has become standard, double-tagged records for codes are mandatory, and as a result, a Dictionary of Synonyms has been developed [24]. This double-labeled record is then given different codes in clinics within the hospital, which may create doubt in the family physician and even the other hospital clinician as to which of these diagnoses is more appropriate. This may lead to unnecessary information load in the database. Taking this point into consideration, our study aims to use permanent data, if possible, thanks to the code that will be given with a single and final diagnosis. Therefore, we recommend that both ICPC-2 and ICD-10 have a code that can be evaluated synchronously by all clinicians. This common code will save clinicians time from performing unnecessary examinations, tests, etc. during subsequent visits of the patient and solving the health problem.

5. Conclusion

According to our results in family medicine practice, the equivalents of ICD-10 and ICPC-2 codes were different in number and percentage. It is ideal for individual health and research that the diagnosis codes in family medicine are the same as the codes in other secondary and tertiary care clinics. Since health requires holisticity, we recommend that both classifications be integrated and revised to be globally understandable and provide complementary coding in different clinical applications.

Ethics Statement

This retrospective study was performed after receiving the approval of the Clinical Research Ethics Committee of Bursa Uludağ University (Reference no: 2023-15/15, dated: 04.07.2023) and by the Declaration of Helsinki.

Conflicts of Interest

The author declares no conflicts of interest.

Author Contributions

Olgun Göktaş was responsible for manuscript designing, data collection, statistical analysis, writing and editing of the manuscript, review, and final approval of the manuscript.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Supporting Information

Supporting Table 1: The comparison of ICD-10 and ICPC-2 diagnoses.