ReviewMeasuring comorbidity in older cancer patients
Introduction
Cancer is a disease that has an increasing incidence with age. Presently, 60% of cancers and two-thirds of cancer deaths occur beyond the age of 65 years in developed countries [1]. Due to the ageing of the population, this proportion is expected to increase markedly in the next decades. As a result, oncologists will be increasingly treating cancer in patients who have concomitant diseases. In a ‘typical' geriatric series, people 65 years of age and older suffer on average from three different diseases [2]. Similarly, older cancer patients present a high level of comorbidity, both in the general population and in oncology consultations (Fig. 1). Multiple studies have demonstrated that comorbidities are relevant to the prognosis of cancer patients, e.g. survival (reviewed in [3]). Therefore, comorbidity can become a major confounder in oncological practice and studies in the elderly. Indeed, geriatric oncology can be defined as “when the health status of a patient population begins to interfere with oncological decision-making guidelines”. A solution largely applied by oncological investigators in the past was either to exclude older cancer patients, or to consider comorbidities as an exclusion criterion. Therefore, randomised studies offer a mortality principally related to the cancer treated. However, given the high prevalence of comorbidity in older cancer patients, these patients are very under-represented, even in studies without an upper age limit 4, 5. In addition, as the comorbidity level of these selected older patients is seldom reported, clinicians are given few clues as to how to adapt the results from co-operative studies to patients with comorbid diseases.
Another approach would be to integrate comorbidities as a variable in the studies, in the same way as functional status presently is. Functional status does not appear to correlate closely with either tumour stage or comorbidity [6]. Therefore, comorbidity should be assessed independently.
However, contrary to functional status, comorbidity presents the unique challenge of being a multidimensional variable. Diseases influencing mortality may not be the same as diseases influencing function or tolerance to treatment. Several scoring systems with varying approaches to that problem have been proposed which will be reviewed in this article, with their characteristics and their validity. Comorbidity indexes can be used in different settings: epidemiological studies (e.g. tumour registries), clinical trials of chronic diseases (e.g. cooperative oncological studies) and clinical trials in acute care settings (such as Intensive Care Units). Each of these settings has a considerable influence on the way comorbidity is approached. The focus of this article is to provide measurement tools for use by oncologists in the clinical study setting. Four validated indexes applicable to this setting will be reviewed in careful detail, their metrological performances, the indexes themselves and rating references, and their practical implementation. Key research issues that need to be investigated are also addressed as well as a few pitfalls to avoid. The reader interested in a larger overview of the indexes available for different settings, as well as a comprehensive review of the prognostic impact of comorbidity is referred to a parallel article [3].
Section snippets
Why focus on four indexes?
Comorbidity is a rather young field of research, and as discussed in more detail in the research section below, much remains to be learned about how comorbidities cumulate to influence prognosis. Therefore, it may seem restrictive to some to review only four scales and suggest their use for clinical studies in oncology. Essentially two situations must be distinguished. The first is comorbidity as the focal point of a study. That kind of study necessitates large collectives of several hundred to
Review of the indexes
Although qualitative definitions are largely used in epidemiological studies, graded indexes tend overall to offer a higher risk discrimination and are preferred in clinical studies 3, 7, 11. All four indexes discussed here use a grading system. One index (the Charlson) weighs precise diseases differently, two indexes offer progressive severity grading within several categories of diseases (the Kaplan–Feinstein and the Index of Coexistent Diseases), one index finally offers progressive grading
Description
The Charlson is probably the most widely used comorbidity index to date [12]. It was designed by Mary Charlson and colleagues in 1987. They used data from an internal medicine inpatient service and analysed the mortality at 1 year as a function of various comorbidities. As a result, a list of 19 conditions (certain of them representing two degrees of severity of the same condition) was designed. Any disease generating a relative risk of death ⩾1.2 was retained and weighted. If the relative risk
Description
The CIRS was first designed by Linn and colleagues in 1968 [25]. It is aimed at a comprehensive recording of all the comorbid diseases of a patient. Its principle is to class comorbidities by organ system affected, and rate them according to their severity from 0 to 4, in a way similar to the Common Toxicity Criteria grading (none, mild, moderate, severe, extremely severe/life-threatening) [26]. Within each category, if two diseases are present, the disease with the highest severity is counted.
Description
The ICED was developed by Greenfield and colleagues in 1987 to address issues of intensity of care [34]. It consists of two subscales: physical and functional. In ICEDs present version, the physical subscale rates comorbidities from 0 to 4 in severity (the same principle as the CIRS), and regroups them in 14 categories: organic heart disease, ischaemic heart disease, primary arrhythmias and conduction problems, congestive heart failure, hypertension, cerebral vascular accident, peripheral
Description
The Kaplan–Feinstein was developed by these two authors in 1974 [43]. It consists of a list of conditions “that might be expected to impair a patient's long-term survival”. These conditions are regrouped in 12 categories (hypertension, cardiac, cerebral or psychic, respiratory, renal, hepatic, gastro-intestinal, peripheral vascular, malignancy, locomotor impairment, alcoholism and miscellaneous) and rated 0–3 according to severity. The severity criteria are well defined. The number and severity
Practical implementation
In the author's and other clinicians' experience, comorbidity indexes are usually easiest to fill on the basis of the initial study history and physical examination report, and the routine laboratory closest to that date. A direct attempt at implementation at bedside most often results in a doubling of services to the patient without clearly demonstrated benefit. Patient-filled comorbidity reports are fairly good for simple items, but their correlation with physician's history taking is not
Research issues
The measurement of the influence of comorbidity in (older) cancer patients is still in its infancy. Although there is now clear evidence of the high prevalence and the influence on survival of comorbidity, much remains to be explored. The first question is: Do only a few specific diseases matter, or is the overall burden of disease (a more ‘geriatric’ understanding) important? One may expect some diseases, such as myocardial infarction, dementia or stroke, to have a strong impact on prognosis.
Conclusions
Comorbidity is an important problem in older cancer patients, and validated tools are available to measure it. Much remains to be learned about the best way to measure and sum it, and the profile of its prognostic value, especially for endpoints other than mortality. A large part of this information will come through prospective studies and meta-analytical techniques, and therefore every effort should be made to use validated indexes, rather than ad hoc lists. This article reviews four such
References (50)
- et al.
Association of comorbidity with disability in older women: the Women's Health and Aging Study
J. Clin. Epidemiol.
(1999) - et al.
Comorbidity measurement in elderly female breast cancer patients with administrative and medical records data
J. Clin. Epidemiol.
(1997) - et al.
Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases
J. Clin. Epidemiol.
(1992) - et al.
A new method of classifying prognostic comorbidity in longitudinal studiesdevelopment and validation
J. Chron. Dis.
(1987) - et al.
Assessing illness severitydoes clinical judgement work?
J. Chron. Dis.
(1986) - et al.
Comorbidity measures for stroke outcome researcha preliminary study
Arch. Phys. Med. Rehabil.
(1997) - et al.
Validation of a comorbidity index
J. Clin. Epidemiol.
(1994) - et al.
The importance of comorbidity in HIV-infected patients over 55a retrospective case–control study
Am. J. Med.
(1996) - et al.
Rating chronic medical illness burden in geropsychiatric practice and researchapplication of the Cumulative Illness Rating Scale
Psychiatry Res.
(1992) - et al.
Reliability of a comorbidity measurethe Index of Coexistent Disease (ICED)
J. Clin. Epidemiol.
(1997)
Lack of association between patient's measured burden of disease and risk for hospital readmission
J. Clin. Epidemiol.
The importance of classifying initial co-morbidity in evaluating the outcome of diabetes mellitus
J. Chron. Dis.
A new clinical-anatomic staging system for evaluating prognosis and treatment of prostatic cancer
J. Chron. Dis.
Relating patient characteristics at the time of admission to outcomes of hospitalization
J. Clin. Epidemiol.
Cancer in older persons
Cancer
Entry and evaluation of elderly patients in European Organization for Research and Treatment of Cancer (EORTC) new-drug-development studies
Cancer
Representation of older patients in cancer treatment trials
Cancer
Comorbidity and functional status are independent in older cancer patients
J. Clin. Oncol.
Adapting a clinical comorbidity index for use with ICD-9-CM administrative datadiffering perspectives
J. Clin. Epidemiol.
A comparison of the Charlson Comorbidity Index derived from medical record data and administrative billing data
J. Clin. Epidemiol.
Comorbidity and age as predictors of risk for early mortality in male and female colon carcinoma patients
Cancer
Importance of functional measures in predicting mortality among older hospitalized patients
J. Am. Med. Assoc.
Comorbid illness is associated with survival and length of hospital stay in patients with chronic disability
Med. Care
Cited by (410)
Utility of a Patient-Reported Symptom and Functioning Assessment Tool for Geriatric Oncology Care in China
2022, Value in Health Regional IssuesImmune mechanisms in cancer patients that lead to poor outcomes of SARS-CoV-2 infection
2022, Translational Research