The Reason-for-Encounter mode of the ICPC: reliable, adequate, and feasible

Scand J Prim Health Care 1989; 7: 99-103 The Reason-for-Encounter mode of the ICPC: reliable, adequate, and feasible FRANS VAN DER HORST, JOB METSEMAKERS, FRANS VISSERS, GERHART SAENGER*, CEES DE GEUS Faculty of Medicine, University of Limburg, Maarhicht, The Netherlands. Department of General Practice, Project: Regirtration Network Family Practices, Chaimn Prof. Dr. J. A. Knottnerus. a Department of Medical Sociology Scand J Prim Health Care Downloaded from informahealthcare.com by 23.106.241.99 on 05/20/14 van der Horst F, Metsemakers J, V irs F, Saenger G, de Gem C. The Reason-for-Encounter mode of the ICPC: reliable, adequate, and feasible. Scand J Prim Health Care 1989; 7: 99-103. For three consecutive months, eight general practitioners reported and classified all reasons for encounter presented by their patients. They were taking part in an international trial of the WHO-ICPC Committee, utilizing a trial version of the Reasons-for-Encounter (RFE-C). This scale has now been integrated with the newly developed International Classifwtion of Primary Care (ICPC). Attention was given to reliabiiity, adequacy, and feasibility. Results indicated that the trial version of this classification system is highly reliable; the participating doctors found it both adequate for their purposes and feasible for use in their practices. Key words: international classifcation primary care, general practice, reason-for-encounter, registration, reliability, adequacy, feasibility. Frans van der Horst, Department of General Practice, Faculty of Medicine, Rijksuniversiteit Limburg P.O.B. 6166200 MD Maasticht, The Netherlands. INTRODUCTION Registration of patient encounters in general practice with the aid of classifications is not new. Wellknown systems are the 1963 E-list (Eimerl) and the 1979 International Classification of Health in Primary Care (ICHPPC-2)3, both derived from the International Classification of diseases (ICD). In contrast to the earlier E list, the ICHPPC classified not only diagnoses but also the less well-defined problems with which primary health care is confronted. Unlike the ICHPPC and ICD, the more recently developed International Classification of Primary Care (ICPC) makes it possible to classify all four elements of the so-called problem-solving process (S, 0, A, and P*). When completed, the total SOAP classification will offer a unique framework for describing the total sequence of the problemsolving process. The present study was restricted to the Reasonfor-Encounter mode of the ICPC.. Its purpose was to ascertain whether the classification was reliable, feasible, and adequate for the classification of the presented reason for encounter. Data were gathered in a field trial of RFE-C in 1983 under the auspices of the WHO Working Party on Classifications of Primary Health Care, and they constitute the Dutch contribution to this trial. Results of the international field trial, in which 138 general practitioners (GPs) and nurses in 9 different countries all over the world took part, were presented by Lamberts et a16. * S = Subjective (patient complaints and problems) 0 = Objective (examination) A = Assessment (evaluation or diagnosis) P = Plans (for treatment or referral). 7 Scand J Prim Health Care 1989; 7

100 Frans van der Horst et al. THE ICPC AS A CLASSIFICATION INSTRUMENT A detailed description of the ICPC can be found elsewherezj*. B,riefly, it is based on two axes. The horizontal axis contains 17 chapters, of which 14 are concerned with organic (body) systems such as the digestive system. One chapter contains unspecified complaints and problems, while the two remaining chapters deal with psychological and social items. The vertical axis includes seven components containing categories which relate to the encounter process. Five of them include identical items for each chapter, such as medication or surgery minor. Only components l and 7 are chapter-specific, with complaintdsymptoms such as chest pain being placed in the first component and diagnoses stated in the seventh. THE PURPOSE OF A CLASSIFICATION The very large number of entries used by different GPs to record what their patients come to see them for does not lend itself to statistical analysis or research. Since different physicians may report the same complaint or problem offered by their patient in quite different ways, this practice impedes communication between GPs and other health professionals. Problems can be avoided to a large extent by providing systematic, comparable information through a reliable system of classification. A limited number of categories now replaces the infinite number of complaints and problems recorded on the case notes. The individual doctor s idiosyncracies are thus avoided. RELIABILITY, ADEQUACY, AND FEASIBILITY We considered the RFE mode of the ICPC list to be reliable if different independent observers assigned the same code number to the same complaint or problem reported by the patient; reliability implies agreement. Two important conditions for reliability are: 1. mutual exclusion of the different classes of disorders (e.g. organ systems) and individual codes 2. satisfactory operationalization (coding instructions) in the case of registration, formulating clear instructions which accurately define how the vari- Scand J Prim Health Care 1989; 7 ous reasons-for-encounter presented to the doctor are to be recorded, in an effort to avoid unsystematic errors. Unsystematic errors can be due to the participating GPs being insufficiently acquainted with the classification system, being unmotivated or fatigued when working with the codes, distractions by third persons, etc. We considered the ICPC to be adequate if the participating doctors agreed that they could use the classification system of the FWE-C list, i.e. found it possible to fit into it any and all complaints and problems presented to them by their patients as reasons-for-encounter. We agreed that this judgement could only be made after the participating GPs had used the classification for some time and after difficulties in the coding process had been resolved in discussion sessions with the doctors. To establish the adequacy of the classification we relied on the judgement of the participating GPs. We felt that their own practice situations would provide the widest spectrum of encounter reasons. All GPs had to participate in several training sessions in which the structure and the use of the classification were explained. We instructed the GPs that all reasons-for-encounter which were difficult to classify should be reported in the weekly project meeting. During these meetings GPs discussed these problems and tried to find the appropriate code. Adequacy in no way implies reliability, although the more adequate the system of classification, the greater the chance that it will be a reliable instrument. The reason-for-encounter mode was regarded as feasible if classification of complaints and problems offered to the GP would require only little extra time during the encounter. To assess the feasibility we asked the GPs to record how much time they spent on classifying the encounters. GPs did not classify during the encounter, but separately at the end of their working day. STUDY DESIGN Eight GPs took part in the study, one running a solo practice, two being associated in a group practice, and the remaining five working in health centres. The field trial lasted three months and covered

Mode of the ICPC 101 Scand J Prim Health Care Downloaded from informahealthcare.com by 23.106.241.99 on 05/20/14 Table I. The reliability (kappa-values) of the RFE-C chapters. Chapters A B (n=1690) (n=1740) A. General and unspecified B. Blood, blood forming organs, lymphatics, spleen D. Digestive F. Eye H. Ear K. Circulatory L. Musculoskeletal N. Neurological P. Psychological R. Respiratory S. Skin T. Endocrine, metabolic and nutritional U. Urinary X. Female genital system Y. Male genital system 2. Social problems Total 0.70 0.80 0.69 0.86 0.82 0.91 0.96 0.96 0.96 0.95 0.92 0.94 0.91 0.93 0.90 0.94 0.93 0.90 0.96 0.96 0.89 0.92 0.88 0.87 0.89 0.91 0.81 0.93 0.70 0.61 0.82 0.75 0.87 0.91 10046 patient encounters and 15195 reasons-for-encounter. (Individual patients could present more than one problem per contact.) The actual registration was done on so-called day forms designed by the WHO. The GPs wrote down the reason-for-encounter, preceded by the ICPC code. Problems with codes, if any, were noted on the back of the forms. All data were centrally collected and processed. In order to determine the reliability of the RFE list, two random samples of about equal size were drawn from all registration forms collected during the field trial (1690 and 1740 reasons-for-encounter, groups A and B respectively). Each sample was coded by four participating GPs, who were given the patient s age and sex per reason-for-encounter. This made it possible to calculate the agreement between GPs coding. A larger random sample size is not necessary to measure agreement. We took two samples because in this way we could prevent faults due to coding fatigue by GPs. The kappa value (K) indicates by how much agreement exceeds that which can be expected on the basis of pure chance. A kappa value of, say, 0.80 Indicates an agreement which is 80% in excess of Table 11. The reliability (kappa-values) of the WE-C components. Components A B (n=1690) (n=1740) 1. 2. 3. 4. 5. 6. 7. Symptoms and complaints 0.86 0.85 Diagnostic, screening and preventive procedures 0.69 0.72 Treatment procedures and medication 0.81 0.84 Results 0.89 0.96 Administrative 0.83 0.85 Other 0.68 0.34 Diagnosiddiseases 0.65 0.70 Total 0.78 0.79 that which can be expected on the basis of pure chance. The kappa-value can range from 0 (no agreement at all) to 1.00 (total agreement). RESULTS Reliability Table I presents data on the reliability of RFE-C codes for complaints and problems. Total scores for both samples were high, 0.87 and 0.91 respectively. Scores for different organ and problem systems ranged from a minimum of 0.69 to a maximum of 0.96 for sample A, and from a minimum of 0.61 to a maximum of 0.96 for sample B. Relatively low reliability scores were obtained for social problems, problems pertaining to the male genital system, and blood, blood forming organs, lymphatics, spleen ; highest scores were obtained for circulatory disorders, respiratory disorders, and disorders of the eye and ear. No striking differences between samples were found. Table 111. Causes of low scores. Number Reason for encounter not fully written out (doctor-related) 63 44 Wrong code due to coder error (doctor-related) 51 35 Wrong code due to deficiency of instruction and classification list 34 21 (list-related) Percentage Scund I Prim Health Cure 1989: 7

102 Frans van der Hoist et al. Table I1 likewise shows high degrees of agreement (0.78 and 0.79 respectively) between components. Scores were slightly lower than those in Table I. In view of the rigorous criteria set to establish reliability, the use of kappa, and agreement between four coders, reliability can again be seen as quite good. For separate components, eight of the fourteen scores were 0.80 or over. Again, the scores of A and B hardly differed, with the exception of component 6. Table 111 indicates that most errors were not listrelated but doctor-related, i.e. the reasons-for-encounter were not fully written down, or errors were due to coding errors rather than to the RFE-list itself. Adequacy There were hardly any problems regarding the adequacy of the WE-C: the division of the list into chapters and components, and the subdivision into codes were generally compatible with the (medical) frame of reference of the participating GPs. A few practitioners indicated that the P and Z chapters were not really specific enough and were less adequate. However, the participating GPs could not agree upon the necessary adjustments. Feasibility In the initial phase the coding of the reason-forencounter took an average of two minutes per encounter; later on, an average of 1 minute proved to be sufficient for virtually all participants. It may thus safely be concluded that the use of the RFE-C is quite feasible in general practice. CONCLUSIONS AND DISCUSSION At the onset of this trial we had stipulated that in order for the coding system to be usable by GPs, it must be reliable (independent coders assign the same code number to the same complaint), adequate (cover all individual complaints and problems presented by their patients), and feasible (requiring a minimum of, time and effort from busy practitioners). To summarize our findings: 1. The reliability can be considered good to excellent. It should be noted however that the reliability based on components is lower than that based on chapters. Scand J Prim Health Care 1989; 7 - We assume this to be a side-effect of our study design, in which GPs were unable to clarify the written-out reasons-for-encounter. Therefore it is probably more difficult to determine the individual component than the correct body system (i.e. chapter). - In this study design we were unable to determine the difference in agreement for component 6. We consider this to be a doctor-related effect, due to the higher number of coding errors in group B. It is our opinion that with improved instructions and better training of the participating GPs it will be possible to eliminate these errors. 2. If the ICPC in the reason-for-encounter mode was to be an adequate classification system in the sense that it corresponds to the medical frame of reference of the GP, it had to cover the total range of the specified reasons of contact presented to GPs. For the time being, P and Z were deemed acceptable, although further elaboration was thought necessary (especially in defining codes). 3. It appears to be quite feasible to use the list in the doctor s daily practice. After a short practice period, extra time needed to use the list was only one minute per reason for encounter. We conclude that the RFE-C is an adequate, reliable, and feasible instrument for routine use in general practice. Moreover, registration on the basis of adjusted lists can be expected to lead to higher reliability scores (above 0.90) and less loss of time (less than one minute per reason for encounter). It must be borne in mind, however, that the satisfactory results of this trial are partly due to the commitment and critical attitude of the participating GPs. Uniform registration and classification with the aid of a system that is adequate, reliable, and feasible for the GPs opens up new perspectives for interpeer auditing and consulting, research, and teaching. In any case, the ICPC provides a structure for the empirical description of the problem-solving process in general practice, and thus opens new roads to a further development of general practice. REFERENCES 1. Meads S. The WHO Reason-for-Encounter Classification WHO Chron. 1983; 37: 159-62. 2. Lamberts H, Wood M. eds. International CIassification

Mode of the ICPC 103 of Primary Care (ICPC). Oxford, Oxford University Press, 1987. 3. ICHPPC-2 (International Classification of Health Problems in Primary Care). Oxford, Oxford University Press, 1979. 4. World Health Organization Study Group on Classification of Diseases. Tenth Revision of the International Classification of Diseases, WHO/ICD 9ff 1.2. Geneva, World Health Organization, 1971. 5. Bentsen BG. International Classification of Primary Care. Scand J Prim Health Care 1986; 4: 43-50. 6. Lamberts H., Meads S., Wood M. Waarom gaat iemand naar de huisarts? Een internationale studie met de Reason-for-Encounter Classification. Huisarts en Wetenschap 1984; 27: 234-44. 7. Cohen J. Weighted kappa: nominal scale agreement with provision for scaled disagreement or partial credit. Psycho1 Bull. 1968; 70: 213. Received November 1987 Accepted January 1989 Scand I Prim Health Cure 1989: 7