Does it really matter? Evaluating the Continuous Morbidity Registration—The Nijmegen Project

The maintenance of a programme of continuous registration of morbidity over 40 years calls for the highest praise and, in particular, commendation of the foresight of Frans Huygen, the founding father. However, the Nijmegen team has taken a large, additional step in transferring the information into a database suitable for the analysis of clinical, social, and economic facets of illness in the community as it presents in primary care.

The introduction of computerized medical records has opened new opportunities for primary care research, but they are all dependent on high-quality, consistent and disciplined recording. It is important to note that this is a fully anonymized database, thus ensuring confidentially. The Steering Committee has been flexible, adapting to new developments over time, but, in the interest of consistency, has retained its original disease classification system (the E-book).

The fundamentals of the information system—a fixed registered population database, a composite score of socio-economic status, the recording of episode type by the reporting doctor, regular meetings with the recorders, the definition of continuing disease activity, and the assessment of severity—are all issues which have been similarly addressed in the programme of national morbidity surveys in England and Wales, and mostly resolved in similar ways [1], [2]. Experience from this project contributed to the recommendations made in the Electronic Health Indicator Data (eHID) Project funded by the European Commission and concerned with optimizing the electronic medical record for use in epidemiology [3].

The similarity of many of the findings concur with experience in England and Wales—increase in hypothyroidism but not hyperthyroidism, decrease in duodenal ulcer and in threadworms [4–6]. The ability to corroborate the findings of one information system in another is a strong argument for having independent databases, and the ability to do that internationally further supports the conclusions [7]. The trend data are particularly valuable, since they are less limited by constraints of sample size, and neatly illustrate the changing nature of primary care with the implied lessons for medical education.

New types of analysis applicable to general practice databases are reported and welcome; in particular, the analysis of relapse in irritable bowel disease and of the time between diagnosis and commencement of medication. The analysis of the management of hypertension in relation to guidelines is an example of the increasing use of databases to examine the quality of care. The use of the database to examine illness in families and links between different types of illness all illustrate the potential of primary care databases. However, the complexity of studying comorbidity should not be underestimated [8].

As van de Laar emphasized when discussing rare diseases, “Medical reasoning is not only based on knowledge of diseases, but also on a profound knowledge of epidemiology.” Particularly important points are made in the chapter by van den Dungen and colleagues, addressing the explanations for differences in the estimates of disease prevalence reported in general practice registration networks. In this regard, we cannot ignore the comparatively small population monitored, even though monitoring extended over many years. A combination of small populations in a geographically limited area and disease frequencies of less than 5% lead to wide confidence intervals around estimates and raise questions about the generalizability of the findings. Care is needed when interpreting the meaning of association. Coefficients of association may be statistically significant, but a value of 0.3 only explains 10% of the variance, which may have little clinical relevance. If socio-economic status is determined from occupational data, how can we differentiate between illness as the cause of low socio-economic status, and the converse, that low socio-economic status causes illness? Inability to work because of ill health leads to socio-economic disadvantage. Research in general practice is challenged by the selection of controls from practice registers. Can a person who never consults a doctor ever be a control for a person who does? Increasing prevalence of all chronic diseases was reported: some conditions are considered chronic on a permanent basis and form a cohort which can only increase year by year. Paradoxically, this is accentuated the more effective medical interventions become in extending life.

The most important contribution of this project concerns the methodology, which has stood for 40 years. It shows what can be done and hopefully hastens the day that it becomes what is routinely done. Primary care information systems offer enormous opportunities to monitor the impact of changes in lifestyle and advances in medicine and social conditions, and to quantify the quality of care. The trends they disclose illuminate the way for the next generation of healthcare providers.

In congratulating the Nijmegen team, we should pick up the words of another founder of general practice, van Es, whose belief in the value of continuity of care is seen as much in the quality of a continuing high-quality record as in the continuity provided by a personal doctor [9].