Prospective Comparison of Six Co-Morbidity Indices As Predictors of 5 Years Post Hospital Discharge Survival in the Elderly

Abstract

Older patients often suffer from multiple co-morbid conditions. Few co-morbidity indices are valid and reliable in elderly patients and comparison between them is rare. Our objective was to compare the performance, relevance, and abilities of six widely used and validated co-morbidity indices—the Charlson Cumulative Illness Rating Scale–Geriatrics (CIRS), Index of Co-Existent Disease, Kaplan Scale, Geriatrics Index of Co-morbidity (GIC), and Chronic Disease Score—to predict 5 years of survival after hospital discharge. Data came from a prospective study with yearly follow up, conducted 2004–2009 in 444 patients (mean age 85 years; 74% female) discharged from the acute geriatric hospital of the Geneva University Hospitals. In univariate analysis, mortality was significantly associated with age; each supplementary year added 7% of additional risk; and with sex, being male increased the risk by 1.5-fold. The best prognostic predictor was the GIC class 4 followed by the CIRS quartile 4 multiplying the risk of death by 4 and 3, respectively. After 1 year of discharge, for both scores approximately 50% of the high-score patients were already deceased and 80% were deceased after 5 years, compared with <5% in the lowest scores after 1 year and <40% after 5 years. When we entered all of the significant independent variables in a stepwise backward analysis, the best multiple regression model retained the CIRS quartile 4 as the strongest risk predictor followed by the GIC class 4. We conclude that the CIRS and the GIC may improve hospital discharge planning as being useful for clinical decision-making purposes and for clinical research in older patients.

Introduction

C o-morbidity, a measure of an individual's underlying health status, is a key variable and has an important role in health-care use, clinical management, treatment decisions, discharge plan, prognosis of adverse outcomes, and survival estimates.^1,2 As a consequence, co-morbidity indices could be useful for both clinical decision-making purposes and for clinical research.³ Elderly patients often suffer from multiple chronic conditions that individually and jointly increase the use of health services like hospitalization, as well as morbidity and mortality.⁴ Hospital discharge planning and prognosis can help physicians and family members plan for the care of patients who may be at increased risk of adverse outcomes in the coming years, especially regarding discussions of goals of care, treatment preferences, advance planning, and clinical therapeutic options.^5
–7 Therefore, hospitalization may present a key trigger point for identifying persons at greatest risk for mortality on the ensuing years after discharge.⁸ Older patients with higher life expectancy may be more likely to benefit from some disease screening, like cancer, because recent guidelines recommend that clinicians target screening in patients with life expectancies greater than 5 years.^5

–8

Although several co-morbidity indices have proven useful for both patient classification in clinical research and prognostication in medical care, only some of them are valid and reliable for use as a measure of co-morbidity in elderly patients.^9,10 In addition, clinicians and researchers are faced with a wide range of indices, with surprisingly little information on the relative strengths and weaknesses of these tools.³ Publications comparing the predictive performance of these indices are rare, especially in the elderly.

The aim of this prospective study was to compare the performance and the value of six validated, widely used co-morbidity indices predicting 5 years of mortality after discharge and to identify a practical and valuable risk-prediction tool that could be applied in routine clinical practice. The study population was derived from a cohort of very old, acutely ill geriatric inpatients.

Methods

Patients and data collection

We carried out a prospective study in a geriatric hospital with 300 acute beds that is a part of Geneva University Hospitals, Switzerland. Patients and data collection have been described elsewhere.^11,12 Briefly, patients were recruited by clinically trained staff. All patients over 75 years of age and consecutively admitted between January, 2004, and December, 2005, were included. We used a computer-generated randomization table to select daily a random sample of patients. The local ethics committee approved the protocol, and patients, their families, or legal representatives provided signed, written informed consent. Demographic data for the patients studied did not differ significantly from data for all patients admitted to the geriatric hospital in 2004–2005.¹¹ Therefore, our sample was representative of all patients admitted to this hospital, demonstrating the reliability of the randomization procedure used in this study.

Medical history was recorded on a standardized form, and the same geriatrician carried out a comprehensive geriatric assessment on all patients. Annual follow up was carried out with the same assessment over a 5-year period.

Sociodemographic data

The data recorded included age, sex, native language, marital status, living arrangement, and educational level.

Cognitive diagnosis

The same neuropsychologist assessed all subjects for clinical dementia. The Mini-Mental State Examination (MMSE)¹³ (scores 0–30) and the Short Cognitive Evaluation Battery were used.^14,15 On the basis of screening results, the neuropsychologist then carried out a comprehensive standardized neuropsychological assessment to determine the etiology and severity of clinical dementia, as previously described.¹¹

Assessment of co-morbidity

We chose the six validated co-morbidity indices that were most widely used to assess the elderly. At the baseline assessment, the same geriatrician calculated all six scores for each patient via an extensive review of the patient's medical records and administrative data for diagnoses established at or before enrollment in this study and by standardized interviews with patients and surrogates.

Charlson Co-morbidity Index (CCI)¹⁶

The CCI is a list of 19 conditions; each is assigned a weighting (1–6). Weightings reflect the ability of each condition to predict 1-year mortality, as originally reported for cancer patients. They are fixed for each diagnosis and range from 1 (for conditions, such as myocardial infarction or mild liver disease, with a relative risk ≥1.2 and <1.5) to 6 (assigned to metastatic cancer, with a relative risk ≥6). The CCI is the sum of the weightings for all conditions observed in a patient; higher scores indicated greater co-morbidity.

Cumulative Illness Rating Scale-–Geriatrics (CIRS)¹⁷

The CIRS identifies 14 items, corresponding to different systems. Each system is scored as follows: 1 (none), no impairment to that organ/system; 2 (mild), impairment does not interfere with normal activity; treatment may or may not be required; prognosis is excellent; 3 (moderate), impairment interferes with normal activity, treatment is needed, prognosis is good; 4 (severe), impairment is disabling, treatment is urgently needed, prognosis is guarded; 5 (extremely severe), impairment is life-threatening, treatment is urgent or of no avail; poor prognosis. The Illness Severity Index (summary score based on the average of all CIRS items, excluding psychiatric/behavioral factors) and the co-morbidity index (summary score based on a count of organ system with moderate or greater impairment, excluding psychiatric/behavioral) can then be calculated using these scores.

Index of Coexistent Diseases (ICED)¹⁸

The ICED is based on the presence and severity of 19 medical conditions and 11 physical impairments, using two scales: the Index of Disease Severity (IDS) and the Index of Physical Impairment (IPI). The final ICED score is determined by an algorithm combining the peak scores for the IDS and IPI. The ICED score ranges from 0 to 3 (four classes), reflecting increasing severity.

Kaplan Scale¹⁹

This index uses two forms of classification, focusing on the type of co-morbidity and the pathophysiologic severity of the co-morbid conditions present, respectively. The type of co-morbidity can be classified as vascular (hypertension, cardiac disorders, peripheral vascular disease, retinopathy, and cerebrovascular disease) or nonvascular (lung, liver, bone, and nondiabetic renal diseases). Pathophysiologic severity is rated on a four-point scale, ranging from 0 (co-morbidity is absent or easy to control) to 3 (recent full decompensation of co-morbid disease). The rating of the most severe condition determines the overall co-morbidity score. Scores for vascular and nonvascular co-morbidity can be calculated, based on the most severe condition in each subscale.

Geriatric Index of Co-morbidity (GIC)²⁰

In computing the GIC, each of the 15 more prevalent clinical conditions (ischemic or organic heart diseases, primary arrhythmias, heart diseases with a nonischemic or nonorganic origin, hypertension, stroke, peripheral vascular diseases, diabetes mellitus, anemia, gastrointestinal diseases, hepatobiliary diseases, renal diseases, respiratory diseases, parkinsonism and nonvascular neurologic diseases, musculoskeletal disorders, malignancies) is graded on a 0–4 disease severity scale on the basis of the following general framework 0 = absence of disease, 1 = asymptomatic disease, 2 = symptomatic disease requiring medication but under satisfactory control, 3 = symptomatic disease uncontrolled by therapy, and 4 = life-threatening or the most severe form of the disease. The GIC classifies patients into four classes of increasing somatic co-morbidity. Class 1 includes patients who have one or more conditions with a disease severity grade equal to or lower than 1. Class 2 includes patients who have one or more conditions with a disease severity grade of 2. Class 3 includes patients who have one condition with a disease severity of 3, other conditions having a disease severity equal to or lower than 2. Class 4 includes patients who have two or more conditions with a disease severity of 3 or one or more conditions with disease severity of 4.

Chronic Disease Score (CDS)²¹

This is a measure of co-morbidity obtained from a weighted sum of scores based on the use of 30 different classes of medication. An integer weight between 1 and 5 is given to each of the selected classes of medication; the overall score is then the sum of the weightings.

Outcome

The outcome of interest was death by December 31, 2009, which means there was 60 months (5 years) of follow up.

Statistical methods

We checked for the normal distribution of continuous scores (CCI, CIRS, Kaplan, and CDS) using skewness and kurtosis tests, and carried out standard transformations to normalize non-Gaussian variables. Because it was not possible to normalize these scores, they were categorized into quartiles to facilitate comparison with the four classes of the two other indices, ICED and GIC. Co-linearity among the six indexes was checked using Spearman rank correlation coefficient. Cox proportional hazards regression models was then carried out to take into account the time to the event using age, sex, and the six co-morbidity scores as independent variables and 5-year mortality as the dependent variable to identify the best predicting score for the outcome while adjusting for all the others. Hazard ratio (HR) and 95% confidence intervals (CI) were calculated. We then entered all the significant independent variables in a stepwise backward analysis to develop the best predictor model. We also used Kaplan–Meier survival curves to examine the performance of the six co-morbidity indices over time. The Cuzick nonparametric test for trend in survival across quartile of the indices was applied. Statistical analyses were performed with Stata software version 11 (College Station, TX).

Results

We included 444 patients in this study (mean age 85.3 ± 6.7; 74% female). A large number of different reasons for hospitalization were recorded, the most prevalent being falls and or fracture (139, 31%), pulmonary infection (55, 12.4%), cardiac failure (45, 10%), and delirium (39, 8.8%). Table 1 summarizes the frequency distribution of patients according to each co-morbidity score. Because there were no patients in the ICED classes 1 and 2, we considered only classes 3 and 4, providing binary data for the analyses. Likewise, only 2% of the patients were classified as class 1 by the GIC and were combined with class 2 for the analysis. For the other four indices, the distribution was almost equal between the four quartile ranges, with approximately 25% of the patients per range.

Table 1.

Quartile Range and Frequency of Six Co-Morbidity Scores

	Quartile/Classes
	1		2		3		4
Indices	Score	n (%)	Score	n (%)	Score	n (%)	Score	n (%)
CCI	0–3	165 (37)	4	91 (20)	5–6	91 (20)	7–14	97 (23)
CIRS	0–11	121 (27)	12–14	107 (24)	15–18	119 (27)	19–30	97 (22)
ICED^a		0		0		93 (21)		351 (79)
Kaplan	0–2	128 (29)	3–4	156 (35)	5	55 (12)	6–16	105 (24)
GIC^a		9 (2)		34 (8)		310 (70)		91 (20)
CDS	0–3	122 (28)	4–6	117 (26)	7–8	109 (24)	9–15	96 (22)

Quartile ranges do not apply to ICED and GIC, because these indices are predefined into four classes. Data are expressed as number of cases (%).

CCI, Charlson Co-morbidity Index; CIRS, Cumulative Illness Rating Scale–Geriatrics; ICED, Index of Coexistent Diseases; Kaplan, Kaplan Scale; GIC, Geriatrics Index of Comorbidity; CDS, Chronic Disease Score.

Univariate and multiple Cox proportional hazards modeling

Of the 444 patients, 264 died during the 5 years after discharge (59.5%). We first carried out a univariate Cox proportional hazard modeling analyses including age, sex, and the six co-morbidity indices tested predicting 5-year mortality after discharge. We then tested full multiple Cox proportional hazards models containing all the variables (Table 2). The Cuzick nonparametric test results for trend in survival across quartile were statistically significant (p < 0.001) for all indices except the CDS (p = 0.117).

Table 2.

Univariate and Stepwise Backward Multiple Cox Regression Predicting 5-Year Mortality (n = 444)

	Univariate Cox regression			Multiple Cox regression (full model)			Stepwise multiple Cox regression
	Independent variables
Outcome	Crude HR	95% CI	p	Adjusted HR	95% CI	p	Adjusted HR	95% CI	p
Death at 5 years
Age	1.07	1.04–1.09	<0.001	1.06	1.04–1.09	<0.001	1.06	1.04–1.08	<0.001
Male vs female	1.52	1.17–1.98	0.020	1.53	1.16–2.02	0.002	1.54	1.17–2.01	0.002
CCI
Quartile 1	1.00	—	—
Quartile 2	1.14	0.81–1.62	0.447	0.71	0.49–1.03	0.070
Quartile 3	1.46	1.05–2.04	0.024	0.79	0.54–1.16	0.228
Quartile 4	2.49	1.23–2.32	0.001	0.94	0.64–1.38	0.771
CIRS
Quartile 1	1.00	—	—	1.00	—	—	1.00	—	—
Quartile 2	1.15	0.79–1.69	0.464	1.00	0.66–1.50	0.990
Quartile 3	2.01	1.42–2.84	<0.001	1.47	0.98–1.20	0.063	1.58	1.17–2.12	0.003
Quartile 4	3.17	2.24–4.48	<0.001	2.00	1.27–3.15	0.003	2.19	1.59–3.01	<0.001
ICED
Class 1 + 2 + 3	1.00	—	—	—	—
Class 4	1.71	1.23–2.37	0.001	1.22	0.86–1.74	0.271
Kaplan
Quartile 1	1.00	—	—	1.00	—	—
Quartile 2	1.36	0.98–1.89	0.068	1.06	0.74–1.52	0.748
Quartile 3	2.04	1.36–3.05	0.001	1.31	0.82–2.12	0.259
Quartile 4	2.46	1.75–3.45	<0.001	1.28	0.82–1.99	0.270
GIC
Class 1 + 2	1.00	—	—	1.00	—	—	1.00	—	—
Class 3	1.63	1.00–2.66	0.047	1.12	0.65–1.95	0.675
Class 4	3.85	2.29–6.47	<0.001	1.75	0.93–3.28	0.082	1.71	1.27–2.31	<0.001
CDS
Quartile 1	1.00	—	—
Quartile 2	1.12	0.80–1.57	0.497	0.92	0.64–1.31	0.639
Quartile 3	1.16	0.83–1.64	0.384	0.95	0.66–1.38	0.808
Quartile 4	1.38	0.98–1.94	0.979	1.02	0.69–1.49	0.933

Bold entries indicate relevant results.

HR, Hazard ratio; CI, confidence interval; CCI, Charlson Co-morbidity index; CIRS, Cumulative Illness Rating Scale–Geriatrics; ICED, Index of Coexistent Diseases; Kaplan, Kaplan Scale; GIC, Geriatrics Index of Co-morbidity; CDS, Chronic Disease Score.

In univariate analysis, mortality was significantly associated with age, each supplementary year added 7% of additional risk of death (HR = 1.07; 95% CI, 1.05–1.09); with sex, being male increased the risk by 1.5 fold (95% CI, 1.17–1.98). The best prognostic predictor was the GIC class 4 (HR = 3.85; 95% CI, 2.29–6.47) followed by the CIRS quartile 4 (HR = 3.17; 95% CI, 2.24–4.48). The CIRS quartile 3 (HR = 2.01; 95% CI, 1.42–2.84); the Kaplan scale quartile 4 (HR = 2.46; 95% CI, 1.75–3.45) and quartile 3: HR = 2.04; 95% CI, 1.36–3.05); the ICED (HR = 1.71; 95% CI, 1.23–2.37); the GIC class 3 (HR = 1.63; 95% CI, 1.01–2.66) the CCI quartile 4 (HR = 1.69; 95% CI, 1.23–2.32) and quartile 3 (HR = 1.46; 95% CI, 1.05–2.04) performed in a similar manner increasing the risk of mortality by 1.5- to 2-fold. The CDS was not predictor of the outcome.

Kaplan–Meier survival curves of the six co-morbidity indices are shown in Fig. 1. The GIC class 4 and the CIRS quartile 4 were the best prognostic predictor of 5-year mortality, multiplying the risk of death by 4 and 3, respectively. For these two scores, after 1 year of discharge, approximately 50% of the high-score patients were already deceased, compared with <5% in the lowest scores. After 5 years, approximately 80% of the high-score patients were already deceased, compared with less than 40% in the lowest scores.

FIG. 1.

Kaplan–Meier survival curves according to the six co-morbidity indices. Marks represent censored observations. CCI, Charlson Co-morbidity Index; CIRS, Cumulative Illness Rating Scale–Geriatrics; Kaplan, Kaplan Scale; CDS, Chronic Disease Score (quartile 1, black line; quartile 2, long dashed line; quartile 3, dashed line; quartile 4, short dashed line); ICED, Index of Coexistent Diseases (class 0, black line; 1, long dashed line; GIC, Geriatrics Index of Comorbidity (class 1–2, black line; 3, long dashed line; 4, dashed line).

When all variables were included in the full model while adjusting for age and sex, the CIRS quartile 4 (HR = 2.00; 95% CI, 1.27–3.15) remained the best predictor independently associated with 5-year mortality. However, when we removed all the nonsignificant variables in a stepwise backward analysis, the best reduced multiple regression model retained the CIRS quartile 4 (HR = 2.19; 95% CI, 1.60–3.00) as the strongest risk predictor followed by the GIC class 4 (HR = 1.71; 95% CI, 1.28–2.30), the CIRS quartile 3 (HR = 1.58; 95% CI, 1.17–2.12); age (HR = 1.54; 95% CI, 1.17–2.01), and sex (HR = 1.06; 95% CI, 1.04–1.08).

Discussion

One of the main strengths of this study was the comprehensive and detailed assessment of the presence and extent of co-morbidity: The same medical doctor scored the six co-morbidity indices for all patients to ensure a high accuracy of scoring. To our knowledge, we carried out, for the first time, a prospective study in a very selected population of very old acutely ill geriatric inpatients. We compared the use of six co-morbidity indices, most of which have been widely used and validated in elderly subjects, for the prediction of 5-year mortality after discharge with a yearly follow up

In our prospective study, excluding the CDS, the others five co-morbidity indices were significantly associated with 5-year mortality in the univariate analysis. The CIRS provided the most accurate risk for 5-year mortality in the univariate as well as in the multiple regression model followed by the GIC. These results are consistent with other studies. Salvi et al. previously demonstrated the CIRS's ability to predict 18-month mortality and rehospitalization in a cohort of 387 patients aged 65 and older from an acute internal medicine ward.²² Parmelee et al. showed a significant association between the CIRS and mortality, acute hospitalization, medication usage, laboratory test results, and functional disability among frail elderly institutionalized patients.¹⁷ One advantage of the CIRS is its suitability for use in common clinical practice: It is based on measures of clinically relevant physiological systems and uses a clear and clinically sound ranking of severity. This index appears to be sufficiently reliable because it allows all of the co-morbid diseases from clinical examinations and medical files to be taken into account in a comprehensive manner.²³ The CIRS was based on the in-hospital mortality of a series of men in a southeastern U.S. veterans hospital in 1964. In the beginning, the CIRS was designed to estimate the total medical burden and survival capacity of elderly patients.²⁴ It has been converted to a co-morbidity index by removing the disease of interest. On the other hand, the CIRS rates the severity of any given disease also on the basis of its impact on function and not exclusively on the basis of biological, clinical, or prognostic considerations. Thus, estimates of the severity of co-morbidity based upon CIRS are somewhat confounded by disability estimates, which is far from ideal in older patients, especially when assessing the impact of co-morbidity on new or worsening disability.

Similar to our results, previous studies confirmed the impact of the GIC index on the prediction of 6-month survival in a population of 1,402 hospitalized elderly patients (age 80.1 ± 7.1 years; 68% female) with chronic disability and consecutively admitted to an acute care unit in Italy.²⁰ As observed in our study, patients with GIC class 1 and 2 scores were not representative in this acute geriatric ward. In a Cox regression analysis, adjusting for factors associated with mortality in univariate models (low levels of serum albumin and cholesterol, anemia, dementia, chronic obstructive pulmonary disease, coronary heart disease, renal diseases, gastrointestinal diseases, advanced cancer) and taking class 2 as a reference, patients with GIC scores in class 4 had a risk of death three times higher than patients with the lowest scores.

The CCI was predictive of 5-year mortality only in the univariate analysis, losing its significance in the Cox regression model. The CCI is the most extensively studied co-morbidity index. It was designed and scaled to predict mortality based on the mortality of 607 patients admitted to a general internal medicine service in a single New England hospital during 1 month in 1984.¹⁶ This index does not take into account the severity of certain major diseases, but only the presence of the disease. For example, in the case of congestive heart failure, patients with either a mild or a severe form of the disease will be assigned a score of 1. Therefore, this index may fail to identify important diseases, or their severity, in the elderly, which may otherwise act as predictors of adverse outcomes. On the contrary, certain pathological conditions, such as acquired immunodeficiency syndrome, are heavily weighted in the index yet rarely encountered in the elderly, whereas other highly elderly prevalent conditions, such as chronic heart failure, are probably underestimated. The CCI has previously been found to be limited in determining the full range of diseases in elderly patients, and very recently the CCI was not able to predict long-term mortality in elderly subjects with chronic heart disease.^23,25

The ICED and the Kaplan index performed well in our univariate analyses for 5-year mortality, but lost all significance when all variables were controlled for. The Kaplan index was specifically developed for use in diabetes research and has been mostly used in oncology. The ICED was created to demonstrate that illness due to diseases other than the primary illness may affect the outcome of interest over the period of observation and has been applied in renal disease. Probably both these scores are less useful to assess co-morbidity in the elderly. The negative predictive value of the CDS, a medication-based score, is consistent with earlier findings and may be due to the use of preventive treatments or treatment for benign conditions in healthier patients. We have to point out that recently Johnson et al. updated and extended the CDS, now renamed the Rx-Risk-V, by adding 26 additional disease categories. This evolution of the basic score was specifically created for the elderly. In a large American national cohort of 260,321 outpatients, the pharmacy co-morbidity score further improved the prediction of 1-year mortality compared to the Deyo diagnosis-based co-morbidity index.²⁶

Our data suggests that the CIRS and the GIC are the best 5-year mortality predictors. This is probably due to the fact that the CIRS as well as the GIC assumes that the impact of all diseases are additive. For these two scores, co-morbidity should take into account both the number of diseases and occurrence of very severe diseases as determinants of health.^20,22 The Kaplan Scale assumes that the single most severe illness will determine the prognosis. The ICED allows a high score in functional status severity to override a high disease severity score, and the CCI weights different severity categories having an impact on the patient's health. The type of scoring system based on additive effect is probably more appropriate to the elderly population.²⁷ As suggested by Fried, some diseases or conditions, in addition to having greater or lesser likelihoods of co-occurrence, may be synergistic in their effects.²⁸ Regarding the development of disability, Ettinger et al.²⁹ found that the risk posed by heart disease alone (odds ratio [OR] = 2.3) or by osteoarthritis alone (OR = 4.3) is considerably less than the risk posed by the combination of the two (OR = 13.6). Recently, the National Institute on Aging Comorbidity Task Force emphasized as central issues on co-morbidity tools on behalf of older persons the need to take into account coexisting and potentially synergistic diseases and their treatments, as well as the functional effect of each disease.^30
–32 Currently, none of the existent indices consider the impact of specific combinations of co-morbid illnesses. There are few prognostic indices available combining co-morbid conditions and functional measures for predicting mortality after discharge in hospitalized older adults. Some existing models are applicable only to specific patient populations and specific diseases.^33
–35 Others require the use of more lengthy formulas based on the knowledge of certain laboratory data and functional status,^36
–38 are based on common geriatric syndromes,³⁹ or are a multidimensional prognostic index based on the comprehensive geriatric assessment.⁴⁰ Although these tools have been developed to target high-risk patients, their use has not been incorporated yet into routine medical practice.

This study has some limitations. First, subjects were assessed at a single site in a university hospital setting where we had no participant in ICED class 1 and 2 and in GIC class 1. This reflects a much compromised cohort of very ill inpatients. Second, patient co-morbidity data were collected only once during hospitalization at the beginning of the follow-up period, and the subject's status could have change over time but it is common practice to take advantage of this specific encounter of a person with the health-care system to collect data otherwise not available. Third, in this large patient cohort, there were different causes for patients' hospitalization, and each cause may vary in severity. As a consequence, these two factors could impact survival independently. Therefore, the generalizability of these results needs to be tested in other locations with different groups of patients and subgroups having more homogeneous hospitalization causes and severity. In addition, the co-morbidity indices tested do not include laboratory data or functional status. Functional status has been shown to be a formidable predictor of health outcomes and a major health outcome by itself. Di Bari et al. have demonstrated in a longitudinal epidemiological survey in the entire population (n = 633) aged 65 and older living in Dicomano (mean age 74), a small rural town in Italy, that physical performance measures adds significantly to the prognostic value of any co-morbidity scale. They compared five different indices (four of which were the same as ours: CCI, ICED, GIC, and CDS), and all were shown to predict in these old community dwellers, although with different strengths, both incident disability (the original cohort was re-examined 4 years after the baseline) and mortality (9 years after the baseline by the city register office). The ICED performed better than the GIC and the CDS. The CIRS index was not applied in this study.⁴¹

In summary, the CIRS index provided an accurate method to identify older patients at high risk of 5-year mortality after a hospital stay followed by the GIC. Both indices are prognostic predictors and could be useful to guide clinicians regarding improvement on care and decisions at discharge. Given their validity and reliability, the GIC and the CIRS seem to provide useful measures of co-morbidity for clinical research and could assist researchers in selecting an effective index for the same outcome of interest.

Footnotes

Acknowledgments

We would like to thank the teams of Mrs. O. Baumer, L. Humblot, and M. Cos for technical assistance. This work was supported by grants from the Swiss National Science Foundation (3200B0-102069) and from the Swiss Foundation for Ageing Research (AETAS).

Author Disclosure Statement

No author received any consultancy fees or has any company holdings or patents. There are no conflicts of interest.

References

Guralnik

. Assessing the impact of comorbidity in the older population. Ann Epidemiol, 1996; 6:376–380.

Boyd

, Weiss

, Halter

, Han

, Ershler

, Fried

. Framework for evaluating disease severity measures in older adults with comorbidity. J Gerontol A Biol Sci Med Sci, 2007; 62:286–295.

Hall

. A user's guide to selecting a comorbidity index for clinical research. J Clin Epidemiol, 2006; 59:849–855.

Gijsen

, Hoeymans

, Schellevis

, Ruwaard

, Satariano

, van den Bos

. Causes and consequences of comorbidity: A review. J Clin Epidemiol, 2001; 54:661–674.

Christakis

. Death Foretold: Prophecy and Prognosis in Medical Care. University of Chicago Press: Chicago, 1999.

Lynn

. Serving patients who may die soon and their families: The role of hospice and other services. JAMA, 2001; 285:925–932.

Lynn

, Teno

, Harrell

. Accurate prognostications of death: Opportunities and challenges for clinicians. West J Med, 1995; 163:250–257.

Levine

, Sachs

, Jin

, Meltzer

. A prognostic model for 1-year mortality in older adults after hospital discharge. Am J Med, 2007; 120:455–460.

de Groot

, Beckerman

, Lankhorst

, Bouter

. How to measure comorbidity: A critical review of available methods. J Clin Epidemiol, 2003; 56:221–229.

10.

Rozzini

, Sabatini

, Barbisoni

, Trabucchi

. How to measure comorbidity in elderly persons. J Clin Epidemiol, 2004; 57:321–322.

11.

Zekry

, Herrmann

, Grandjean

, Meynet

, Michel

, Gold

, Krause

. Demented versus non-demented very old inpatients: The same comorbidities but poorer functional and nutritional status. Age Ageing, 2008; 37:83–89.

12.

Zekry

, Herrmann

, Grandjean

, Vitale

, De Pinho

, Michel

, Gold

, Krause

. Does dementia predict adverse hospitalization outcomes? A prospective study in aged inpatients. Int J Geriatr Psychiatry, 2009; 24:283–291.

13.

Folstein

, Folstein

, MacHugh

. Mini Mental State: A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res, 1975; 12:189–198.

14.

Solomon

, Hirschoff

, Kelly

, Relin

, Brush

, DeVeaux

, Pendlebury

. A 7 minute neurocognitive screening battery highly sensitive to Alzheimer's disease. Arch Neurol, 1998; 55:349–55.

15.

Robert

, Schuck

, Dubois

, Olié

, Lépine

, Gallarda

, Goni

, Troy

. Investigators' Group. Screening for Alzheimer's disease with the Short Cognitive Evaluation Battery. Dement Geriatr Cogn Disord, 2003; 15:92–98.

16.

Charlson

, Pompei

, Ales

, Mackenzie

. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis, 1987; 40:373–383.

17.

Parmelee

, Thuras

, Katz

, Lawton

. Validation of the cumulative illness rating scale in a geriatric residential population. J Am Geriatr Soc, 1995; 43:130–137.

18.

Greenfield

, Sullivan

, Dukes

, Silliman

, D'Agostino

, Kaplan

. Development and testing of a new measure of case-mix for use in office practice. Med Care, 1995; 33:AS47–AS55.

19.

Kaplan

, Feinstein

. The importance of classifying initial co-morbidity in evaluating the outcome of diabetes mellitus. J Chronic Dis, 1974; 27:387–404.

20.

Rozzini

, Frisoni

, Ferrucci

, Barbisoni

, Sabatini

, Ranieri

, Guralnik

, Trabucchi

. Geriatric Index of Comorbidity: Validation and comparison with other measures of comorbidity. Age Ageing, 2002; 31:277–285.

21.

Von Korff

, Wagner

, Saunders

. A chronic disease score from automated pharmacy data. J Clin Epidemiol, 1992; 45:197–203.

22.

Salvi

, Miller

, Grilli

, Giorgi

, Towers

, Morichi

, Spazzafumo

, Mancinelli

, Espinosa

, Rappelli

, Dessí-Fulgheri

. A manual of guidelines to score the modified cumulative illness rating scale and its validation in acute hospitalized elderly patients. J Am Geriatr Soc, 2008; 56:1926–1931.

23.

Harboun

, Ankri

. Comorbidity indexes: Review of the literature and application to studies of elderly population. Rev Epidemiol Sante Publique, 2001; 49:287–298.

24.

Linn

, Linn

, Lee

. Cumulative Illness Rating Scale. J Am Geriatr Soc, 1968; 5:622–626.

25.

Testa

, Cacciatore

, Galizia

, Della-Morte

, Mazzella

, Russo

, Ferrara

, Rengo

, Abete

. Charlson Comorbidity Index does not predict long-term mortality in elderly subjects with chronic heart failure. Age Ageing, 2009; 38:734–740.

26.

Johnson

, El-Serag

, Tran

, Hartman

, Richardson

, Abraham

. Adapting the Rx-Risk-V for mortality prediction in outpatient populations. Med Care, 2006; 44:793–797.

27.

Wieland

. From Bedside to Bench: Research in comorbidity and aging. Sci Aging Knowl Environ, 2005; 39:29–33.

28.

Fried

, Ferrucci

, Darer

, Williamson

, Anderson

. Untangling the concepts of disability, frailty, and comorbidity: Implications for improved targeting and care. J Gerontol A Biol Sci Med Sci, 2004; 59:255–263.

29.

Ettinger

, Davis

, Neuhaus

, Mallon

. Long-term physical functioning in persons with knee osteoarthritis from NHANES. I: Effects of comorbid medical conditions. J Clin Epidemiol, 1994; 47:809–815.

30.

Yancik

, Ershler

, Satariano

, Hazzard

, Cohen

, Ferrucci

. Report of the national institute on aging task force on comorbidity. J Gerontol A Biol Sci Med Sci, 2007; 62:275–280.

31.

Lash

, Mor

, Wieland

, Ferrucci

, Satariano

, Silliman

. Methodology, design, and analytical techniques to address measurement of comorbid disease. J Gerontol A Biol Sci Med Sci, 2007; 62A:281–285.

32.

Karlamangla

, Tinetti

, Guralnik

, Studenski

, Wetle

, Reuben

. Comorbidity in older adults: Nosology of impairment, diseases, and conditions. J Gerontol A Biol Sci Med Sci, 2007; 62A:296–300.

33.

Muers

, Shevlin

, Brown

. Prognosis in lung cancer: Physicians' opinions compared with outcome and a predictive model. Thorax, 1996; 51:894–902.

34.

Morita

, Tsunoda

, Inoue

, Chihara

. Survival prediction of terminally ill cancer patients by clinical symptoms: Development of a simple indicator. J Clin Oncol, 1999; 29:156–159.

35.

Fillenbaum

, Pieper

, Cohen

, Cornoni-Huntley

, Guralnik

. Comorbidity of five chronic health conditions in elderly community residents: determinants and impact on mortality. J Gerontol, 2000; 55:M84–M89.

36.

Walter

, Brand

, Counsell

, Palmer

, Landefeld

, Fortinsky

, Covinsky

. Development and validation of a prognostic index for 1-year mortality in older adults after hospitalization. JAMA, 2001; 285:2987–2993.

37.

Teno

, Harrell

, Knaus

, Phillips

, Wu

, Connors

Jr , Wenger

, Wagner

, Galanos

, Desbiens

, Lynn

. Prediction of survival for older hospitalized patients; the HELP survival model. J Am Geriatr Soc, 2000; 48,Suppl:S16–S24.

38.

Inouye

, Peduzzi

, Robison

, Hughes

, Horwitz

, Concato

. Importance of functional measures in predicting mortality among older hospitalized patients. JAMA, 1998; 279:1187–1193.

39.

Anpalahan

, Gibson

. Geriatric syndromes as predictors of adverse outcomes of hospitalization. Int Med J, 2008; 38:16–23.

40.

Pilotto

, Ferrucci

, Franceschi

, D'Ambrosio

, Scarcelli

, Cascavilla

, Paris

, Placentino

, Seripa

, Dallapiccola

, Leandro

. Development and validation of a multidimensional prognostic index for one-year mortality from comprehensive geriatric assessment in hospitalized older patients. Rejuvenation Res, 2008; 11:151–161.

41.

Di Bari

, Virgillo

, Matteuzzi

, Inzitari

, Mazzaglia

, Pozzi

, Geppetti

, Masotti

, Marchionni

, Pini

. Predictive validity of measures of comorbidity in older community dwellers: The Insufficienza Cardiaca negli Anziani Residenti a Dicomano Study. J Am Geriatr Soc, 2006; 54:210–216.

Prospective Comparison of Six Co-Morbidity Indices As Predictors of 5 Years Post Hospital Discharge Survival in the Elderly

Abstract

Introduction

Methods

Patients and data collection

Sociodemographic data

Cognitive diagnosis

Assessment of co-morbidity

Charlson Co-morbidity Index (CCI) 16

Cumulative Illness Rating Scale-–Geriatrics (CIRS) 17

Index of Coexistent Diseases (ICED) 18

Kaplan Scale 19

Geriatric Index of Co-morbidity (GIC) 20

Chronic Disease Score (CDS) 21

Outcome

Statistical methods

Results

Univariate and multiple Cox proportional hazards modeling

Discussion

Footnotes

Acknowledgments

Author Disclosure Statement

References

Charlson Co-morbidity Index (CCI)¹⁶

Cumulative Illness Rating Scale-–Geriatrics (CIRS)¹⁷

Index of Coexistent Diseases (ICED)¹⁸

Kaplan Scale¹⁹

Geriatric Index of Co-morbidity (GIC)²⁰

Chronic Disease Score (CDS)²¹