Prevalence of zebras in general practice: Data from the Continuous Morbidity Registration Nijmegen

Abstract

Objective: To explore the prevalence of rare diseases in the Continuous Morbidity Registration (CMR) Nijmegen, and to discuss methodological difficulties in the study of rare diseases in general practice. Methods: We selected all diseases with a prevalence <0.5/1000 patients/year between 1986 and 2006 in the CMR, and we compared the results with rare diseases in the Orphanet database of rare diseases. Results: We retrieved 71 codes of rare diseases: 15 referred to trauma or intoxications; the remaining 51 could be subdivided into malignancies, specific infections, diseases of short duration, and chronic diseases. Twenty (36%) of these 56 diseases were also listed in the Orphanet database (mostly malignancies and chronic diseases).

Conclusion: Although in the literature emphasis is placed on rare chronic diseases, the general practitioner is also confronted with many rare diseases of shorter duration. The CMR is not yet feasible for studying diseases with an extremely low prevalence.

• Rare diseases
• family practice
• prevalence
• incidence

Introduction

It is believed that Theodore Woodward (1914–2005), an American investigator of infectious diseases, coined the aphorism “if you hear hoofbeats behind you, don't expect to see a zebra” [1]. This aphorism adequately reflects the fact that the process of medical reasoning and decision making is not only based on knowledge of diseases but also on a profound knowledge of epidemiology: one is in general much more likely to encounter a common disease (e.g., a horse) than a rare one (e.g., a zebra). Woodward's saying has consequently led to the use of “zebra” as a symbol for rare disease.

A rare disease is defined as a condition affecting fewer than 5 in 10 000 people yearly in the community, and it is estimated that there are about 5000–8000 rare diseases in the world [2]. Based on these assumptions, and according to a WHO report, this might imply that 6–8% of the population suffers from a rare disease [3]. All known rare diseases are published in an online database [2], and prevalence data for each condition are estimated by a survey of the literature and online sources [4].

As a result of its low prevalence, a zebra is generally of little interest to general practice, and this is reflected in the absence of data concerning the prevalence of rare diseases in general practice. However, in a descriptive study of the records of 100 patients with a rare disorder in general practice, it was found that the general practitioner (GP) was involved in the identification, diagnosis, and provision of acute and chronic care in 54–89% of cases [5]. Other authors advocate that GPs are well placed to help patients with rare problems, and that the development of a generic general practice strategy may improve overall care [6]. From a patient's perspective, it is important to note that (parents of) patients with rare disease appreciate the help of a GP, but at the same time feel marginalized compared to patients with more common diseases [7]. Moreover, the healthcare needs of patients with rare (chronic) disorders are comparable to those with prevalent disorders [8].

In this paper, we explore rare diseases as encountered in the Continuous Morbidity Registration (CMR) Nijmegen, and we subsequently compare our findings with the rare diseases listed in the database of rare diseases. Some methodological difficulties in the study of rare diseases in general practice are discussed.

Methods

This is an explorative study in which we selected all disease codes with a reported prevalence <0.5/1000 patients/year, between 1986 and 2006, from the CMR [9]. The following codes were excluded: codes for symptoms; codes that did not refer to a medical condition; “rest categories”, i.e., codes in which different conditions related to one body system were merged; and codes related to conditions in pregnancy and labour. Next, based on the data, we categorized the remaining diagnoses into the following groups: trauma, malignancies, infectious diseases, and “other” diseases. The “other” diseases were subdivided into diseases of short duration and diseases of a chronic nature. “Short duration” was defined as when the condition lasted less than 1 year in more than 50% of cases. “Chronic” was defined as when the condition lasted longer than 1 year in 50% of cases or more. For all diseases, we checked whether they were included in the European Orphanet database of rare diseases [2].

Results

In the CMR, 222 codes had a reported prevalence ≤0.5/1000 patients/year (range 0 to 0.5). We excluded 151 codes for the following reasons: 62 codes were linked to symptoms (e.g., fever, weight loss), 25 categories did not refer to a medical condition (e.g., vaccination, social work), 41 codes related to a “rest category” (e.g., “other diseases thyroid gland”), and 23 codes were related to pregnancy and/or labour. From the remaining 71 codes, 15 referred to traumas or intoxications (e.g., skull fracture). The other 56 are listed in Table I. Twenty (36%) of these 56 conditions are listed in the Orphanet database.

Table I.  Rare diseases (prevalence per 1000 patient/year) in the Continuous Morbidity Registration Nijmegen.

Short duration		Chronic diseases		Specific infections		Malignancies
Salivary gland calculi	0.5	Cleft lip*	0.4	Lyme disease*	0.5	Squamous cell carcinoma	0.5
Pulmonary embolism	0.4	Nephrotic syndrome	0.4	Poliomyelitis*	0.4	Stomach cancer*	0.4
Cholecystitis	0.4	Spina bifida*	0.3	Mumps	0.4	Laryngeal cancer	0.4
Pelvic inflammatory disease†	0.4	Glomerulonephritis*	0.3	Helminthic infections	0.3	Hodgkin disease*	0.3
Iridocyclitis	0.3	Polycythaemia*	0.1	Rubella	0.3	Brain tumour*	0.3
Pilonidal sinus	0.2	Pneumoconiosis	0.1	Scabies	0.3	Oesophageal carcinoma*	0.2
Pneumothorax	0.2	Hydrocephalus*	0.1	Herpetic keratitis	0.3	Pancreatic carcinoma*	0.1
Osteomyelitis	0.2	Haemophilia*	0.1	Hepatitis B	0.3
Xanthelasma	0.2	Femoral hernia	0.1	Gonorrhoea	0.2
Mastoiditis	0.1	Congenital haemolytic anaemia*	0.1	AIDS/HIV	0.2
Ectopic pregnancy†	0.2	Rheumatic Fever*	<0.1	Tuberculosis*	0.1
Epidemic pleurodynia	<0.1			Amoebic dysentery*	0.1
Foetal erythroblastosis	<0.1			Toxoplasmosis	0.1
Nasal furuncle	<0.1			Hepatitis C	0.1
Neonatal sepsis	<0.1			Hepatitis A	<0.1
Umbilical sepsis	0			Syphilis	<0.1
				Malaria*	<0.1
				Measles	<0.1
				Shigella dysentery*	<0.1
				Meningococcal meningitis*	<0.1
				Trichomonas in male	<0.1
				Smallpox	0

*Included in Orphanet database of rare diseases (prevalence ≤0.5/1000 patients).

†Females only.

A disease of “short duration” was defined as a condition with a duration <1 year in more than 50% of cases. A “chronic disease” lasted 1 year or more in 50% or more of the cases.

Discussion

In this paper, we explored the occurrence of rare diseases in the CMR Nijmegen, defined as a prevalence ≤0.5/1000 patients/year. We found 56 rare diseases that could be categorized in terms of malignancies, infectious diseases, miscellaneous diseases of short duration, and diseases of a chronic nature. Of these 56 rare conditions, 20 were listed in the Orphanet database for rare diseases, which contains about 6000 rare diseases.

Why did we find only a few rare diseases in the CMR?

One obvious answer to this question is that, although CMR aims to register all morbidity encountered in general practice, priority is given to more common diseases. Rare diseases for which no specific code exists are registered in a “rest category” in the database and can only be retraced by manually searching the original files. Moreover, it is important to realize that most rare diseases have an extremely low prevalence (< 1/100 000 patients/year) and that, even if specific codes had existed for most rare diseases, the chances of such cases occurring in the CMR population, which consisted of 13 500 patients in 2006, are negligible.

Why are most rare conditions in the CMR not listed in the Orphanet database?

Most of the conditions found in both the CMR and the Orphanet database were in the malignancy category and the category of diseases of a chronic nature. None of the items in the “short duration” diseases category and only seven out of 22 items in the infections category were found in the Orphanet database. This points to the fact that, in general, “rare diseases” are seen as chronic conditions and that rare diseases of short duration might be understudied.

Other reasons for the discrepancy between the CMR and Orphanet data might be in the underreporting of disease. Some rare diseases are diagnosed and treated almost exclusively in secondary care, for example neonatal sepsis. For such patients, data collection depends on the quality and completeness of the reporting of the specialist. Moreover, some conditions will be less likely to be presented to the GP because they cause few complaints (e.g., xanthelasma) or can be self-diagnosed and treated by the patient (e.g., helminthic infections). Some diseases are rare in the Netherlands, but are more common in other parts of Europe or the world (e.g., hepatitis B). In addition, smallpox is considered to have been eradicated and is therefore not found in either the CMR or Orphanet.

In order to improve the study of the prevalence of rare diseases in general practice, changes have to be made in terms of methods of registration. Electronic patient files offer the opportunity to export more detailed information concerning diagnosis to the central database. Textual information could be exported and translated by investigators to, for example, ICD codes, which are much more detailed than e-List or ICPC codes. Moreover, this study could be used as a starting point for research into the question of whether the generalistic approach, as delivered by the GP, is effective for patients with rare disease [6].