Over the years, WONCA has developed the International Classification of Primary Care ICPC as the ordering principle of the family practice domain. The ICPC describes episodes of care by reasons for encounter (reflecting the patient’s perspective), diagnoses (reflecting the physician’s perspective), and interventions.3,15-19 On this basis, family practice databases can be created that allow international comparison.
Unfortunately, national representative databases fulfilling the requirements formulated by White and colleagues in 1961 are still not available.7 The goal of this study is to compare the content of family practice in different countries, using existing databases that (minimally) contain data on reasons for encounter, diagnoses and interventions that are coded with, or can be addressed by ICPC in an episode of care structure.7,18
Family physicians in the Netherlands, Japan, and Poland have been collecting episode of care data over several years in listed populations, with an ICPC -based electronic patient record for all encounters, for research purposes and under controlled conditions. Recently, Green and colleagues noted a serious lack of such data in the US, resulting in problems when estimating essential indicators from available sources.9 Since the publication of the IOM Report on Primary Care in 1996, pointing out that the available information in the US did not allow episode of care analysis, the increasing use of electronic patient records in family practice networks has not yet resulted in databases fulfilling all criteria for this study. However, the National Ambulatory Medical Care Survey (NAMCS) records reasons for visit and diagnoses, allowing an estimation of the family physician’s contribution to ambulatory care; no episodes of care could be identified from NAMCS data.20-23 It was decided to use these four databases in this study.
Obviously, comparative studies must take into account the major differences in the national health care systems; global data from these four countries indicate substantial differences in health care delivery, expenditure and health status Table 1.24-26 Although it is impossible to directly relate these differences to the available databases, and family practice does not have a major impact on all of these outcomes, they can be helpful to better understand the study’s results.
Dutch family physicians are gatekeepers for listed and relatively healthy practice populations with universal access. This contrasts sharply with the US, where far more is spent on health care with disappointing health status indicators, and without a central position for family practice. Dutch family practice data on reasons for encounter, diagnoses and interventions are, by their nature, a close proxy for the population’s demand, clinical need, and supply. Most Dutch family physicians use an electronic patient record, in which the use of ICPC for coding diagnoses is mandatory.27-29
In Poland, little is spent on a health care system with general access; health status is unsatisfactory. Over the past decade, Polish health care policy has strongly supported family practice, deploying a family practice retraining program for general internists, gynecologists, and pediatricians with an often-longstanding experience in hospitals.27-29
In Japan, health care contributes to a relatively long and healthy life at moderate costs; family practice has a weak position, being well developed in rural areas only. Family physician training is much like that in general internal medicine. Although the Japanese have freedom of choice and complete coverage, patients in the participating rural practices bring most health problems to their family physician, with the exception of practically all obstetric/gynecologic and most pediatric and psychiatric care for which they see specialists in the nearest cities.10,11,30-36
Methods
Data from the Netherlands, Japan, and Poland were collected identically with an electronic patient record (“Transhis”) as a part of the Transition Project of the Amsterdam University. For all face-to-face encounters, the reasons for encounter, diagnoses and interventions were coded with ICPC within an episode of care structure. Prescriptions were coded with the ICPC drug-code (derived from the Anatomical Therapeutic Classification [ATC]).29,37
In the Netherlands, 10 family physicians in 6 practices participated from 1995 to 2000. In Japan, 6 family physicians in rural health centers related to Jichi Medical School participated from 1996 to 1999. In Poland, the Family Practice Department of Katowice Medical School organized the study from 1997 to 1999) with 22 family physicians in 2 practices. Their population was assigned to them on the basis of census data. Therefore, families without 1 of its members having had at least 1 encounter with a participating family physician were excluded from the Polish data.
Since no such US data existed, we derived where possible comparable estimates using visit data from the NAMCS database.20 Sample physicians completed forms for a systematic random sample of office visits during a random 1-week period, coding up to 3 reasons for visit and diagnoses using the Reason for Visit Classification for Ambulatory Care (RVC) and the International Classification of Diseases-9th revision (ICD-9-CM).38 Prescribed drugs were classified with the National Drug Code Directory.39 Data included all ages, all races, and both sexes. The 1995-1997 data were used (91,395 visits), with 2955 ambulatory care visits per 1000 US citizens (26% with a family physician).20,22 Data were recoded with ICPC through mappings with RVC and ICD-9-CM, and ICPC drug codes were mapped with the major pharmaceutical groupings in NAMCS.37,39,40