Multimorbidity is common to family practice

METHODS

Bibliometry is a research method used in library and information science. It can be used to ascertain scientific activity quantitatively and qualitatively. Its applications include tracking the evolution of research topics in the literature. Bibliometric studies classify and count occurrences in databases.¹⁷ Bibliometric techniques are often used to assess the presence of particular topics in the medical literature.^18,19,20

Data sources

For this study, we consulted the MEDLINE database with the Ovid search engine for the reference period 1990 to the end of 2002. Since the term multimorbidity does not have an equivalent in the database’s thesaurus (Medical Subject Headings [MeSH]), it was used as a key word (or text word) with its various spellings. Comorbidity was searched using its MeSH term. For comparative purposes, asthma, hypertension, and diabetes were searched using their respective MeSH terms. The modifier “focus” was used with all MeSH terms to target documents in which the subject heading is considered the main point of the article. This simple strategy allows comparison between subjects, which was the purpose of this study. This modifier cannot be used when a term is searched as a key word. Only studies on human subjects were selected. To target the literature relating specifically to primary care, the following terms were used: general practice, family practice, and family medicine, as suggested by Rosser.²¹ Those terms were linked with the Boolean indicator “or” and with the main search terms using the indicator “and.”

Prevalence data were gathered from the 1998 Quebec Health Survey² and Statistics Canada (available at http://www.statcan.ca). The Quebec Health Survey was the only accessible source of data on the prevalence of multimorbidity in Canada. Statistics Canada data were used to measure the prevalence of asthma, hypertension, and diabetes in Canada.

Analyses

We compared the concept of multimorbidity with the three chronic conditions as to prevalence and number of articles published on them. We calculated the proportion of articles relating to primary care.

For the type and characteristics of the publications on multimorbidity, we reviewed the abstracts and analyzed the content using predetermined categories: type of publication and actual content of the abstract. A total of 353 articles were identified by the research strategy. Two authors (M.F. and L.L.) independently evaluated the first 30 abstracts. Controversial cases were discussed. After standardization, one author (L.L.) analyzed the remaining abstracts. Difficulties encountered during the classification were discussed by all the authors to obtain consensus.

Comparative analyses were done using SPSS version 8. Fisher’s exact test and the chi-square test were used bilaterally to compare proportions of publications on each topic.

RESULTS

Table 1 shows prevalence data on asthma, hypertension, and diabetes and compares them with prevalence data on multimorbidity. Multimorbidity had the highest prevalence, followed by hypertension. We found few articles in the medical literature dealing specifically with multimorbidity or comorbidity, but we found a great many articles on asthma, hypertension, and diabetes. For each article published on multimorbidity or comorbidity, 74 were published on asthma, 94 on hypertension, and 38 on diabetes. These differences, even if smaller for articles specific to primary care, are still substantial.

Table 2 shows the various categories of articles based on qualitative analysis of the abstracts. Articles found to be unrelated or unclassifiable were eliminated. An article was considered unclassifiable if the abstract was missing, and unrelated if the abstract showed no indication that the article dealt with multimorbidity or comorbidity. Only 42 articles (27.6%) identified a primary care context, and most of these (59.4%) were epidemiologic studies. We found no articles evaluating primary care interventions for managing or following up patients with multimorbidity. Articles relating to multimorbidity in primary care did not have significantly different classifications from those relating to the three other conditions.

Table 3 shows characteristics of the selected articles. Nearly 20% of the articles were published in languages other than English or French. A substantial number of published studies included or focused on elderly patients. Prevalence was estimated in 10 articles on primary care, but their definitions of multimorbidity were somewhat incomplete or unusual. The consequences of multimorbidity in a primary care context were addressed in most studies using various outcome measures including quality of life, length of hospitalization, mortality, and functional autonomy.

DISCUSSION

Despite the high prevalence of multimorbidity in the general population, few articles focus on multimorbidity or comorbidity in the general medical literature or in the literature dealing more specifically with primary care. In contrast, many publications focus on asthma, hypertension, and diabetes. Although clinical studies of these conditions probably included patients with multimorbidity, these studies would not have been identified by our search strategy. The proportion of primary care articles is similar for all four topics, although the number of articles on multimorbidity is much lower. In view of the high prevalence of multimorbidity, this is a cause for concern. It suggests that family physicians do not have access to enough information on caring for most of their patients.

Type of articles

From our analysis of the abstracts, we identified 42 articles relating to primary care (Tables 2 and 3); with the exhaustive strategy suggested by Rosser and associates,²¹ we identified only 10 (Table 1). We suggest caution when using this strategy because it might lack sensitivity. Classification of articles following content analysis showed that there is little diversity in publications. The greater number of epidemiologic studies implies, however, that multimorbidity is a growing concern for many researchers. Given the small number of indexed articles on multimorbidity, we cannot draw any conclusions about trends in publication over time. The fact that about one quarter of the articles were published during the last 2 years suggests a growing interest in this subject.

Characteristics of articles

When we look at the characteristics of the published articles in more detail (Table 3), we note that in 37 articles (seven in primary care), multimorbidity or comorbidity is not the primary focus but a secondary variable, which again reduces the amount of evidence on this subject.

Given the small number and lack of diversity of published articles about multimorbidity in the context of primary care, researchers and policy makers might wish to review their priorities. Research on elderly patients must continue, but middle-aged adults should be the focus of specific research efforts. Measuring the prevalence of multimorbidity in the population must be improved with the use of reliable indicators, validated in primary care.²² Quantitative studies on multimorbidity must be done in primary care using relevant outcome measures that could eventually be used in intervention studies.^23,24 Qualitative studies must also be undertaken²⁵ to develop a better understanding of multimorbidity from many perspectives, such as patients, caregivers, and health care workers.

Future research

Other information can arise from specific research into questions about causality. Is there a predisposition to acquire apparently unrelated diseases? To date, very few studies have addressed this question. The serious consequences and resulting health care burden require that more research be done to develop knowledge and understanding of multimorbidity. Primary care practices offer an ideal setting because of the diversity of patients and medical conditions they deal with every day. Studying family physicians’ patient populations from the perspective of multimorbidity is an innovative approach.

Future directions for studies include accurate measurement of multimorbidity and its various effects in the context of family practice, and later, proposing interventions. Special attention must be paid to including patients with comorbidity in clinical trials (providing their safety can be ensured). Adding the term “multimorbidity” to the MEDLINE database thesaurus should also be suggested to the relevant authorities in order to make literature reviews on this topic more effective.

Limitations

The main limitation of this type of study resides in its inability to find all the available documents. We consulted only MEDLINE and used only the Ovid search engine. Looking for articles in other databases or with other search engines might have generated more matches, but it is unlikely that more matches would have altered our findings substantially.

Our research strategy might also have limited the number of articles retrieved. The absence of a MeSH term for multimorbidity, a potentially limiting factor, has been partly circumvented by researching this term as a key word. Not being able to use the “focus” modifier with this strategy might have increased the number of articles on multimorbidity retrieved, but seemingly had little effect because the number of articles found was negligible. A research strategy using key words instead of MeSH terms would have identified more articles on comorbidity, asthma, hypertension, and diabetes. It would have had better sensitivity, but worse specificity without the “focus” modifier. We think that articles focusing on multimorbidity or comorbidity as a distinct entity would have been identified by the strategy we used. Various strategies were tested during this study; the one we used appeared to be the most accurate.

Finally, content analysis was limited to the abstracts of the articles found. Analyzing the complete articles might have increased the accuracy of the evaluation but would not have changed the number of articles found.

Conclusion

This study points up the mismatch between the prevalence of multimorbidity in the population and the number of articles published on it and indexed in MEDLINE. To date, the number and diversity or articles available on multimorbidity are insufficient to provide a strong scientific basis for evidence-based care of patients affected by multiple concurrent chronic medical conditions. It is essential to increase primary care research on multimorbidity in order to develop a better understanding of this important topic.