320.full.pdf

Peritoneal Dialysis International, Vol. 25, pp. 320332Printed in Canada. All rights reserved.

0896-8608/05 $3.00 + .00Copyright 2005 International Society for Peritoneal Dialysis

320

CHARACTERIZING COMORBIDITY IN DIALYSIS PATIENTS: PRINCIPLES OFMEASUREMENT AND APPLICATIONS IN RISK ADJUSTMENT AND PATIENT CARE

Dana Miskulin

Division of Nephrology, New England Medical Center, Boston, Massachusetts, USA

Correspondence to: D. Miskulin, Division of Nephrology,New England Medical Center, Box 391, 750 Washington Street,Boston, Massachusetts, 02111 USA.

[email protected] 16 February 2005; accepted 5 April 2005.

Comorbid conditions are highly prevalent in dialysis pa-tients and are significant predictors of mortality and otheradverse outcomes. Accordingly, it is important to accountfor differences in comorbid illness burden among groupsof dialysis patients being compared. At present, there isno consensus on what conditions matter, how each shouldbe defined, and what weights each carries when definingan individuals risk or case-mix severity. A number of co-morbidity instruments, generic or disease specific, havebeen employed in dialysis populations. They differ by therepresentation and definition of conditions as well as in-strument scoring. No instrument has been found to be su-perior to another in terms of predictive accuracy formortality, and accuracy across the board is low. Furtherstudies are needed to determine whether improvementswould be found with the use of more specifically defineditems and through assignment of item weights based onrelationships for outcomes specifically in a dialysis popu-lation. The roles of other factors in risk prediction, such asmarkers of nutritional status, inflammation, or otherphysiological parameters, relative to comorbid conditionsalso need to be defined. Outcomes other than mortality arelikely to identify different factors and/or different rela-tionships than those noted for mortality, which also requirestudy.

Comorbidity is important for risk adjusting comparativeanalyses in nonrandomized trials and quality of care as-sessments and may, in future, influence payment for di-alysis services. Efforts to improve the management ofcomorbid illnesses are needed. Comorbid conditions mustbe documented accurately and uniformly in all dialysis pa-tients to enable these applications.

Perit Dial Int 2005; 25:320332 www.PDIConnect.com

KEY WORDS: Case-mix severity; case-mix adjustment;survival analysis; questionnaires; quality of care.

Comorbid illnesses, which refer to medical conditionsother than the index disease, that is, kidney failure,are highly prevalent in the dialysis population, a reflec-tion of the primary disease that caused kidney failure orthe untoward effects of kidney failure on other organsystems. Cardiovascular disease in particular is commonand is a major cause of the striking increase in mortalityobserved across all age groups of patients with chronickidney disease (CKD) compared with controls (1,2). Theincident dialysis population in the USA in 2003 illustratesthe tremendous burden of comorbidity of these patients;over 60% had at least one cardiovascular comorbidity,with ischemic heart disease in 25%, congestive heartfailure (CHF) in 32%, and peripheral vascular disease in14% (3). Epidemiological studies have shown thatgrowth in new end-stage renal disease (ESRD) cases inthe USA in recent years has far outpaced CKD prevalence,likely reflecting the broader acceptance of individualsinto dialysis programs and the improved prevention andoutcomes of cardiovascular disease, resulting in in-creased longevity to reach dialysis (46). In keeping withthis, the number and extent of comorbid illnesses in theaverage patient initiating dialysis have increased overthe past two decades (710). These observations high-light the need for increased attention to comorbid ill-nesses, not simply in the context of risk-adjusting clinicalstudies, but for a number of reasons, including the day-to-day care of patients.

Comorbidity is a major determinant of morbidity andmortality, thereby defining an individuals risk for ad-verse outcomes, that is, case-mix severity (11). The othercase-mix factors involved will depend on the outcomeunder study but typically include demographic and so-cioeconomic factors, physiological parameters, treat-ment-related factors, and some that may be difficult todefine (e.g., cultural or religious factors) or measure(e.g., genetic factors). The usual context in which casemix is an issue is when groups of patients, defined by atreatment or parameter of interest, are compared for anoutcome but differ inherently by risk factors for that

321

PDI JULY 2005 VOL. 25, NO. 4 MEASURING COMORBIDITY IN DIALYSIS PATIENTS

outcome, not simply by chance but because the same fac-tors led to the treatment or parameter under study inthe first place (Figure 1). This confounding by case mixmust be accounted for in order to determine the true re-lationship of the treatment or other parameter with theoutcome (12). For example, the factors leading to theselection of peritoneal dialysis (PD) over hemodialysis(HD), such as young age, minimal comorbidity, greaterfunctional ability, and desire for independence in dailyliving (1315), also predispose to longer survival. With-out adequate control for the propensity to increased sur-vival in PD patients independent of the modality or anyother factor, comparisons of PD versus HD will be biasedin favor of PD.

A large number of potential case-mix factors may dif-fer across groups of dialysis patients, but how much case-mix adjustment is enough? In theory, adequacy ofcase-mix adjustment is measured as the extent to whichvariance in outcome is explained (16); in practice, it is afunction of the number of factors that can be reliablymeasured in all subjects. Groups of dialysis patients arecompared frequently in quality-monitoring programsand clinical studies, but, in most cases, case-mix differ-ences are ignored or inadequately controlled for. Nota-bly, comparisons of mortality rates across dialysisfacilities in the USA account only for differences in age,race, gender, and the primary diagnosis across popula-tions (17). These factors explain very little variability ofthe survival of dialysis patients (18,19). A major hurdlelimiting more routine measurement and adjustment forcase mix is a lack of knowledge about what comorbid con-ditions matter and how they should be defined andweighted.

For a host of reasons, not the least of which is thesheer number of comorbid conditions and combinationsof conditions encountered in a dialysis population that

are interrelated, comorbidity is one of the most challeng-ing aspects of case-mix severity to define. This reviewwill discuss general principles surrounding the measure-ment of comorbidity in dialysis patients, compare instru-ments used in past studies, discuss the role of case-mixfactors other than comorbidity in risk adjustment, andwill identify the many applications of comorbidity datathat emphasize the need for uniform comorbidity report-ing in all dialysis patients.

SOURCES OF COMORBIDITY DATA

An accurate and up-to-date source of information isnecessary to characterize the presence and extent of anindividuals comorbid illnesses. There are three generalsources of comorbidity data: the patient (self-report),administrative databases, and review of the medicalrecord, the latter varying in accuracy depending on theskill and knowledge of the person abstracting the data.The advantages and disadvantages of each are outlined.

Self-Report of Comorbid Illnesses: Self-reporting ofmedical conditions may be reliable in other populations,but reservations are warranted, considering the lengthyand complex medical histories typical of dialysis pa-tients as well as the high prevalence of cognitive defi-ciencies, which are estimated to affect more than 25%of patients (20,21). A single study compared self-re-porting of 8 comorbid conditions against data obtainedthrough a systematic medical record review in 959 in-cident dialysis subjects at baseline in the CHOICE(Choices for Healthy Outcomes in Caring for ESRD) Study(22). Specificity was high, meaning patients did notreport diseases that they did not have; however, sensi-tivity was low, even though these were major conditionssuch as CHF, myocardial infarction, stroke, and chronicobstructive pulmonary disease. Self-report of symptomsor limitations from disease, as opposed to diagnoses,are collected with health status assessments and maybe more promising, but also have drawbacks, as will bediscussed later.

Administrative Databases: Administrative databasesserve as a source of comorbidity information for largepopulations; the major advantage being that the dataalready exist, reducing time and expense that would beotherwise spent in chart review. Diagnoses and proce-dures stemming from hospitalizations and other encoun-ters are coded using uniform taxonomies such as theInternational Classifications of Disease, 9th revision.Some drawbacks relate to the coding systems themselvesin that they are not specific for a dialysis population,and where an exact code is not present, the closest sub-stitute may be inappropriate or too broad in definition

Figure 1 The relationship of case-mix factors to the treat-ment and outcome under study.

322

MISKULIN JULY 2005 VOL. 25, NO. 4 PDI

to accurately depict the true diagnosis. Other inaccura-cies may arise because personnel coding the medicalrecords are not medically trained and codes with higherreimbursements may be preferentially used. Nationalregistries, such as the United States Renal Data System,are another form of administrative database, nowpresent in several countries, that systematically collectdata to study epidemiology and outcomes of the ESRDpopulation over time. In the USA, no quality control pro-cedures govern the collection of comorbidity, under-taken in new Medicare-entitled patients using theMedical Evidence Form (Form 2728), which captures thepresence or absence of 18 major conditions and 2 physi-cal impairments (23). Longenecker et al. compared datafrom the Form 2728 against chart reviews in subjectsenrolled to the CHOICE Study and revealed grossunderreporting, even for major comorbid conditions(24). Efforts to improve the collection and quality of co-morbidity data within national data systems should bestrongly encouraged, as these data will enable standard-ized measurements and comparisons within and betweennational populations.

Medical Record Review: The medical record is the meansby which health care providers communicate with oneanother and is generally considered the most reliablesource of information about comorbid conditions (11).The latter is true if a complete record can be assembled,which is not a trivial task considering the potential forrecords to be dispersed among the offices of the nephrol-ogist, primary care physician, and various consultants,as well as one or more hospitals and the dialysis unit.Another source of variability stems from the reviewersfamiliarity with the diagnoses and procedures encoun-tered in dialysis patients. In theory, the nephrologist isthe ideal person to compile and maintain the list of pastand active medical problems, although this has not beenthe practice in the USA, perhaps because it is time-con-suming and benefits may not be readily apparent. Main-taining accurate data requires some investment of timebut is lessened if problem lists are updated as eventsoccur.

INSTRUMENT CONTENT: REPRESENTATION AND DEFINITIONSOF COMORBID CONDITIONS

A large number of comorbid conditions and/or com-binations of conditions are possible in a dialysis popula-tion, but only a finite list can be practicably and reliablycollected on a routine basis. A standardized instrumentis required to ensure consistency. Comorbidity instru-ments vary considerably in their representation and defi-nition of conditions. To enhance consistency in

interpretation across multiple reviewers, conditionsshould be defined as objectively as possible, using cri-teria appropriate for a dialysis population. The detail bywhich items should be defined is unclear. Given the highprevalence of many of the conditions in a dialysis popu-lation, it would seem especially important to delineatesubjects at one versus the other end of the severity spec-trum for each condition, as prognosis would be expectedto differ greatly between them. As shown in Table 1, whenthe broad entity any history of CHF is compared withthe more specific definitions collected in a large HDpopulation using the Index of Coexistent Diseases (ICED)(27), the 1-year mortality risk varies considerably, de-pending on the definition used. A substantial loss of in-formation and of discriminatory power would result ifwe considered only the term any CHF in defining anindividuals mortality risk. For risk measurement pur-poses, defining the severity of each condition however,would be unnecessary if those with a severe disease (e.g.,severe CHF) also consistently are afflicted with otherconditions (e.g., ischemic heart disease, diabetes, pe-ripheral vascular disease), in which case, the increasein risk is captured by the presence of the other condi-tions. The latter remains to be proven.

A final principle to consider when selecting a comor-bidity instrument is the availability of the data used todefine each condition. Non-routine test results in a di-alysis unit, such as electrocardiograms or chest x rays,take time to locate or may be missed. When reviewed ret-rospectively, the circumstances surrounding the test maynot be apparent and may not accurately reflect thechronic activity and prognosis of the condition (e.g., anechocardiogram immediately after an acute myocardialinfarction). Although the mere presence or absence of atest might in itself be considered a measure of severity,it may simply reflect differences in physician practicestyle, test availability, or that it is missing from the medi-cal record, all of which have nothing to do with the pres-ence or severity of disease (11).

ITEM WEIGHTS AND SCALING

There are two general ways to adjust for case-mix fac-tors. If there is a large number of outcomes, the comor-bidity and other case-mix factors can be treated asseparate covariates in the analysis, the advantage beingthat the relationships of each are estimated for this studypopulation as opposed to using predetermined weightsof a comorbidity instrument. More often, it is desirableto conserve statistical power for assessing the main fac-tor of interest with the outcome so information is sum-marized into a single covariate using a comorbidity

323


instrument. Depending on the instrument, the comor-bidity data may be summarized as the count of condi-tions per patient or the sum of weighted items. Itemweights are based on relationships with a specific out-come of interest, preferably determined in the popula-tion of interest. Weights effectively represent theprognostic significance of one condition relative to an-other. The summary score is usually divided into levelsto provide a simple and practical means for risk stratify-ing a population. The principles regarding instrumentcontent and scoring are illustrated in comparing fourcomorbidity instruments, two of which are generic, theCharlson Comorbidity Index (CCI) and the ICED, and twodesigned specifically for dialysis patients, the WrightKhan Index and the Davies Index.

COMORBIDITY INSTRUMENTS IN USE IN DIALYSISPOPULATIONS

Charlson Comorbidity Index: The major appeal of thisinstrument is its simplicity. The instrument consists of19 items, each weighted for mortality as determined fromits original development in a general medical inpatientpopulation in a New York hospital in the late 1980s, latervalidated in a population with breast cancer (28). Itemweights vary from 1 to 6 and the weighted items aresummed to a scale that ranges from 2 to 37 plus the ad-dition of 1 point for each decade after age 40 years. Thetime spent scoring a CCI through data abstraction fromchart review in an incident PD population ranged from10 to 20 minutes in one study (29).

A number of studies have shown increasing CCI scorecorrelates with increased mortality (2931) and costs(32). For example, the average CCI score in 268 incidentPD patients from a single USA center at the start of di-alysis was 5.4 2.2, and for each increase of 1 in the CCIscore, there was a 50% increase in the subsequent rela-tive risk (RR) of death (RR 1.54; 95% confidence inter-val 1.36 1.74) (29).

It is, however, reasonable to question whether an in-strument designed for a general medical population isapplicable to a dialysis population. First, the medicalconditions found to be most prognostic for mortalityamong inpatients on a medical ward in the 1980s are notlikely to be the same as for a present-day outpatient di-alysis population. Notably, there are only two items forcardiac disease, each of which receive the lowest weight,whereas 6 of 18 items (excluding moderate-to-severerenal disease) and more than 50% of the score relate toconditions that, together, account for

324


Table 2. As expected, cardiovascular conditions are givenmore prominence in the dialysis population. An inter-esting point, the condition diabetes with complica-tions received a new weight of 1, versus 2 in the originalCCI. Diabetes with complications is common in the di-alysis population and, defined as such, this conditionalone is unlikely to effectively discriminate with respectto survival. Also, the increased risk is captured in its com-plications peripheral vascular disease, ischemic heartdisease, etc. Oddly, diabetes without complications wasgiven a weight of 2 over 1, which the authors attributedto older age of these patients, although there may havebeen a problem with sample size as confidence intervalswere wide. Even using the reassigned weights, predic-tive accuracy was on the low side, leading us to question

whether the problem lies in the need for more specificdefinitions than those on the CCI. One distinct advan-tage of the CCI is that its widespread use enables directcomparisons across different populations (3335). Thesmall number of broadly defined items required to scorethis instrument, however, can easily be collected in ad-dition to those required of another instrument if anotherproves superior.

Davies Index: This instrument was developed from 97 PDpatients followed for 30 months and consists of 7 equallyweighted items: ischemic heart disease, peripheral vas-cular disease (including cerebrovascular), malignancy,left ventricular dysfunction, diabetes, systemic collagenvascular diseases, and other conditions. The basis forassigning equal weights to each condition is unclear, butthe relationships with mortality were not equivalent inits development (36). Furthermore, definitions in somecases are vague or subjective, for example left ventricu-lar dysfunction as clinical evidence of CHF not attribut-able to errors in fluid balance, requires clinical judgmentand does not define severity of the condition or frequencyof events, all of which leads to variability in scoring acrossmultiple reviewers. Nonetheless, several studies, includ-ing a multicenter study, the Netherlands CooperativeStudy on the Adequacy of Dialysis (NECOSAD), consistentlyshow a graded mortality risk for increasing levels of thescore divided into three levels, which persisted even afteradjustment for a number of demographic, physiological,and treatment-related factors (3638).

WrightKhan Index: The Wright-Khan Index was devel-oped in 375 dialysis patients followed at least 2 years anddivides the population into three risk levels based on ageand the presence of comorbid conditions. These consistof (1) age

325


morbid illnesses with the ICED is much more extensive,consisting of over 160 items that denote the presenceand severity of each condition. The instrument has twoparts: the first is based on data abstracted through chartreview and consists of 19 disease categories, definedfurther for three grades of disease severity; the secondpart of the instrument is an observer-based assessmentof physical impairments. The instrument was originallydesigned for a general medical population to predictfunctional status outcomes (40), but enhancements,particularly to the cardiovascular categories, were madeprior to its use in the HEMO Study (19).

The ICED has been used in a number of multicenter set-tings, including the Hemodialysis (HEMO) Study (41), theCHOICE study (14), and a not-for-profit national dialysisprovider in the USA (42). The instrument divides the popu-lation into three levels and each corresponds with increas-ing mortality r isk with little overlap among levels(19,26,43). It is the only instrument, to my knowledge,in which reliability of scoring has been tested, as carriedout in the HEMO Study where the average agreement inassigning an ICED score to 6 charts across nurses at the15 study sites and a physician (considered the gold stan-dard) was >90% (41). In addition, it is sensitive to change.Annual ICED measurements were performed as part of theCHOICE Study and the change or absence of change in ICEDscore was a strong predictor of mortality, independent ofthe baseline score (43). Another unique feature of theICED is the inclusion of physical impairments, which haveconsistently been shown to be strong predictors of mor-tality and adverse outcomes (23,4446).

The obvious tradeoff for the amount of data collectedis the time taken per assessment, which averages 50 min-utes per patient and makes it impractical for everyday use.Other drawbacks are that the wealth of information col-lected is lost by summary into only three levels, the prog-nostic signif icance of each condition is ef fectivelyconsidered equivalent because the final score is based onthe single peak disease score combined with the singlepeak impairment score, and there is no increment in scorefor patients with multiple severe comorbid conditions.

WHICH INSTRUMENT IS BEST AT PREDICTING MORTALITY?

All of the instruments above work in the sense thatincremental risk scores correlate with increased mortal-ity; however, chronological age divided into deciles wouldproduce a similar result but is a poor mortality predictoron its own. Measures of model fit (e.g., chi-square test)statistics are often reported and describe how well thecovariates explain the variability in the outcome but aredifficult to conceptualize in clinical terms. The area under

the receiver-operator-characteristic (ROC) curve providesa measure of discriminatory ability to compare instru-ments (47). Values for ROC range from 0 to 1 and may beinterpreted as the frequency with which the instrumentassigns a higher score to the person who dies versus sur-vives, given all possible pairs in which one subject diesand survives. An instrument with an ROC of 1 perfectlydiscriminates among those who die versus survive.

van Manen et al. compared the CCI, the Khan, and theDavies indices for their accuracy in predicting mortalityin 1205 incident PD patients with a 1-year survival rateof 87% from NECOSAD (31). In addition to the data col-lected to score these instruments, a separate aim of thisstudy was to assess whether consideration of diseaseseverity for four conditions (diabetes, cancer, angina,and CHF) improved accuracy over the broad categoriza-tions of diseases from the above instruments. Results ofconcordance statistics (analogous to ROC) for mortalitypredictions of each comorbidity instrument plus age wereroughly equivalent as follows: 0.72 for the Khan, 0.73for the Davies, 0.74 for the CCI, and 0.75 for the newindex that accounts for disease severity. The addition ofexplicit severity definitions for the comorbid conditionsdid not appear to improve predictive accuracy, althougha limited number of conditions were def ined morespecifically.

In another study, involving 2388 individuals from anational dialysis provider in the USA, with an annual sur-vival rate of 79%, the CCI, ICED, Khan, and Davies indi-ces were compared for 1-year mortality predictions (48).A comparison of survival models composed of each co-morbidity measure plus age is shown in Table 3. In termsof discriminatory ability, the ICED appeared superior tothe others when predictions were based simply on a co-morbidity instrument plus age. The addition of race, gen-der, cause of ESRD, vintage, and serum albumin to eachmodel improved discrimination for each model, and themargin in favor of the ICED lessened. An ROC of 0.77 isconsidered reasonably good for a predictive tool, butstill, 23% of the time, individuals are misclassified asbeing at higher risk than an individual who dies, whenthey in fact survive, or vice versa. The most widely usedprognostication system, the APACHE, usually yields anROC of 0.80 or higher for 30-day mortality (49,50); how-ever, this is probably not a fair benchmark as accuracywould be expected to be better with a shorter time spanbetween the measurement and the outcome. Anotherimportant measure of accuracy is calibration, which sum-marizes agreement between predictions and observedoutcomes. Calibration is particularly important for di-rect applications of the predictions, for example in re-porting an individuals predicted mortality to aid

326


decision-making. Results show the WrightKhan Indexmay not be calibrated, but the others were satisfactory.

The list of instruments above is by no means a com-plete list of comorbidity measures used in dialysis pa-tients, although, to my knowledge, the accuracy of theothers has not been tested. We can conclude that all ofthese instruments are roughly equivalent once othercase-mix factors are added, but none is sufficiently ac-curate to be used alone in making clinical decisions.

ALTERNATE APPROACHES TO COMORBIDITY AND/OR CASE-MIXMEASUREMENT

Physical Impairments and Health Status Measurements:In clinical practice, we gauge the severity of many con-ditions by the patients accounts of the intensity or fre-quency of symptoms or their impact on functionality.

Self-report of disease severity through symptom burdenis the basis of instruments successfully used in generalmedical, diabetic, and ischemic heart disease outpatientpopulations (51,52). In the dialysis population, a num-ber of studies have shown strong relationships of healthstatus or physical impairments with mortality, includ-ing through self-report with the SF-36 (53,54), or as re-ported by observers, for example with the KarnofskyIndex (44,46), the ICED (43), or Form 2728 (23). Whilethese approaches intuitively make sense and would beless costly than collection of comorbid conditions, majorconsideration should be given to the fact that the indi-viduals at greatest risk (and those most important tocapture), may be selectively omitted because they aretoo ill and/or cognitively impaired to respond. Observer-based assessments should also be viewed with cautiongiven the subjective nature of scoring, which may pose

TABLE 3Predicting One-Year Mortality: a Comparison of Comorbidity Indices + Age

Unadjusted Adjustedc

Subjects within 1-year Area under Area underInstrument (+ age)a each level [n (%)] mortality (%) ROCb curve 95% CI ROCb curve 95% CI

ICED 0.72 0.690.75 0.77 0.750.7901 545 (31) 9.72 500 (28) 23.23 734 (41) 36.1

CCI 0.67 0.650.70 0.74 0.720.771 181 (10) 6.12 116 (7) 10.33 767 (43) 21.24 507 (29) 30.35 208 (12) 45.2

WrightKhan 0.68 0.650.70 0.75 0.720.781 204 (12) 4.92 787 (33) 18.93 992 (56) 31.7

Davies 0.68 0.650.70 0.75 0.730.781 301 (17) 9.62 937 (53) 22.83 541 (30) 35.3

ICED = Index of Coexistent Disease; CCI = Charlson Comorbidity Index; ROC = receiver operator characteristic; CI = confidenceinterval.a Age was added to each model.b The ROC provides a measure of discriminatory ability and was highest for the ICED when a comorbidity measure was used on its

own to predict mortality. Upon the addition of other case-mix factors, the discriminatory ability of each model was equivalent,regardless of the comorbidity measure.

c Each model was adjusted for age, comorbidity measure, race, gender, cause of end-stage renal disease, and serum albumin.The proportion that died within a year increased with higher risk score for each comorbidity instrument. [Adapted from Ref. (48)]Source: Miskulin D, et al. Predicting one-year mortality in an outpatient hemodialysis population: a comparison of comorbidityinstruments. Nephrol Dial Transplant 2004; 19(2): p. 416. Copyright (2004). Reprinted with permission from the European RenalAssociationEuropean Dialysis and Transplant Association.

327


problems depending on the application. If, for example,resource allocation was tied to case mix or through riskadjustment, facility performance reports could be madeto look better, there would be incentive to overstate theillness burden of ones dialysis population.

FACTORS OTHER THAN COMORBIDITY AS RISK ADJUSTORS

Unique to dialysis populations over other outpatientpopulations is the amount of clinical data collected on aroutine basis. A number of studies have demonstratedrelationships of various laboratory and physiological pa-rameters, for example serum albumin, systolic bloodpressure, body mass index, serum phosphate, and par-athyroid hormone, with mortality and other outcomesin dialysis patients (23,5558). Markers of malnutritionand inflammation, such as interleukin-6 (IL-6), C-reac-tive protein (CRP), serum ferritin, total cholesterol, andothers also associate strongly with mortality (59,60).Although it may be unclear why a patient has a high CRPor a low serum albumin, from a risk stratification stand-point, that the factor is highly predictive for mortalityand is reliably measured is all that matters. Many of thesefactors are already collected as part of the monthly orquarterly blood work in HD and PD populations. Giventhe expense of collecting comorbidity information andthe potential for variability, at any point from compilingthe data record to scoring it, laboratory tests are moreand more appealing. For some of these laboratory mea-sures, factors affecting the reliability of measurement,such as diurnal variation, timing relative to dialysis, orupsets due to an acute illness in an otherwise low-riskindividual, need to be worked out before they can be pro-posed as risk stratification measures for widespread use(61). Further, whether comorbidity can be substituted,in part or wholly, by inflammatory and physiological pa-rameters for risk stratification purposes remains to betested. One study showed this was not the case in thatcomorbidity, as measured using the ICED, maintainedsignificance and contributed more to the survival modelthan either CRP or IL-6, or the combination (62), al-though further studies to address this are needed. Onething is for certain, these nonspecific inflammatory andnutritional markers cannot replace the need to charac-terize comorbid conditions for the purposes of patientcare.

STRATEGIES FOR IMPROVING PREDICTIVE ACCURACY

In reviewing studies relating a comorbidity instrumentand other factors with mortality, four issues need to beexplored as a means to improve risk predictions in dialy-

sis patients. First, the comorbid conditions in the aboveinstruments are defined broadly, or, as in the case of theICED, the information regarding disease severity is ef-fectively lost in the final scoring. The use of explicit defi-nitions for disease severity for each condition is likely toimprove discrimination. Second, the weights of condi-tions in the CCI are based on relationships with mortal-ity in a general medical population, and in the otherinstruments, purely on clinical judgment. Relationshipsof factors with mortality must be worked out specificallyin the dialysis population. Third, the contribution oflaboratory and physiological data is strong and much ofthe data is routinely available in a dialysis population.Such factors should be considered alongside the comor-bid conditions to identify and weight those of key prog-nostic signif icance as the physiological data andcomorbidity data are likely to be correlated. Fourth, co-morbidity and other selected factors will change withtime, yet these changes cannot be incorporated in a pre-dictive instrument because it is not known with certaintywho will survive to the next interval at which variablesare updated. To account for this, the maximum time span(between assessment and outcome) that ensures stabil-ity and accuracy of predictions needs to be worked out.It may be less than a year, but cannot be so short that itwould be impractical to update the variables at the re-quired frequency in routine practice. To conduct thesestudies, a large population with an extensive list of can-didate comorbid conditions and other case-mix factorsis needed.

USES OF COMORBIDITY DATA

Risk Adjustment in Clinical Research: Case-mix severityis generally considered a nuisance factor for non-randomized studies; there is no particular interest in thecase-mix factors themselves but, in order for the treat-ment effect to be measured accurately, differences instudy arms must be controlled for. Even in randomizedtrials when small differences that are strongly prognos-tic for the outcome exist, adjustment for case mix can in-crease statistical power to detect the treatment effect(63,64). The potential for error from lack of control overcase-mix bias is illustrated in a recent study of the effectsof cholesterol on mortality (65). It has long been ques-tioned whether measures to lower cholesterol should beapplied to the dialysis population, given past observa-tions of an increase in mortality risk with lower choles-terol and a decrease with higher cholesterol (66,67). Theterm reverse epidemiology has been coined to describeother instances where relationships of nutritional or car-diovascular risk factors with mortality are opposite to

328


those seen in the general population (6870). A recentstudy showed that, once the population is stratified bycase mix (into high and low groups), effectively control-ling for the fact that those with lower cholesterol wereinherently sicker, a different result emerged. In those atlow-risk, a decreasing cholesterol was associated withreduced mortality, consistent with relationships in thegeneral population. A similar effect of modification bycase mix was found with body mass index and mortality(71). Reverse epidemiology may be nothing more than in-adequate control over case-mix severity. Increased num-bers of clinical tr ials are being conducted in CKDpopulations and greater attention to case mix is neededto avoid erroneous conclusions.

Comorbid illnesses may lead to different responses intreatment. For example, PD may be postulated to confera survival advantage for patients with severe left ven-tricular dysfunction because of the continuous natureof ultrafiltration and avoidance of hemodynamic insultsof HD. When examining a treatment effect, these pa-tients need to be separated from the rest of the PD popu-lation as they may respond very differently. Recentstudies using administrative USA data show specific sub-groups, including new ESRD patients with CHF, ischemicheart disease, or large body habitus, have, if anything,reduced survival compared with their counterparts onHD and in contrast with the rest of the population(13,72,73). Further studies are needed to determinewhether these f indings are reproduced and, if so,whether different treatment targets or other modifica-tions are needed in the setting of specific comorbidillnesses.

Risk Adjustment in Quality Improvement Programs: TheClinical Performance Measures Program in conjunctionwith the National Kidney Foundation sets and monitorsadherence to various parameters of care and, in doingso, strives to improve the quality of care of patients re-ceiving dialysis at facilities throughout the USA (74).Case-mix adjustment may be integral to quality report-ing given that there are well-known differences in pa-tient populations across providers that may affectprocess and outcome measures (52,75). In a facility withhigher than average case mix, it is likely to be more dif-ficult to adhere to the designated quality benchmarks,resource use is higher, and survival rates lower. Withoutaccounting for such differences, irrespective of the caredelivered, units caring for sicker patients will be scruti-nized unfairly, while those caring for healthier popula-tions pass unnoticed (76). Providers are increasingly heldaccountable for their results by patients and payers. Withincreased pressure to achieve results, and with ever dwin-dling resources, dialysis facilities may be pressured into

cherry picking the younger and healthier patients,which is far from the goals of a quality improvement pro-gram (77). Control over case-mix differences is criticalto the assurance that the measures being monitored ac-tually represent quality (as opposed to inherent differ-ences in patient populations) and thereby promiseimproved care and outcomes.

Resource Allocation: Whether it be warfarin dosing,antibiotic administration, diabetes management, relay-ing care plans to caretakers in institutionalized patients,or the extra dialysis sessions required of patients proneto intradialytic hypotension, there are ample reasonswhy patients with greater burden of comorbid illness costmore to care for. This, unfortunately, is not reflected inthe current reimbursement for outpatient dialysis ser-vices in the USA. In order for a case-mix adjusted modelfor reimbursement to become operational, the key co-morbid conditions that are influential will need to beidentified and measured in all dialysis patients.

Management of Comorbid Conditions in Everyday Prac-tice: Studies have shown that fewer patients with CKD thanin the general population receive treatments consideredthe standard of care for cardiovascular disease, cardio-vascular risk factors, diabetes and lipid management, andother preventive measures (3,78,79). Accordingly, theNational Kidney Foundation and others are expandingclinical practice guidelines to the management of diabe-tes and lipid and cardiovascular disease in CKD patients(80). Disseminating these into practice poses a greaterchallenge. As a first step, documentation of past and cur-rent medical conditions in the dialysis unit medical recordneeds to improve (Figure 2). This alone will enhance com-munications among the multidisciplinary team of provid-ers and will lead to better treatment decisions. The use ofstandardized care algorithms for specific comorbid con-ditions encountered in a dialysis population is a novelapproach to improving comorbid illness management ina dialysis population (81). A recent study, the Dialysis RiskFactor Intervention Trial, developed and tested the useof standardized algorithms for management of bloodpressure, diabetes, lipids, ischemic heart disease, atrialfibrillation, and CHF in five dialysis units of a not-for-profit provider in the USA (82,83). The algorithms focusedon pharmacological and nonpharmacological treatmentsand were based on review of the medical evidence and,where they existed, clinical practice guidelines. A phar-macist reviewed a standardized comorbidity profile andthe medication records for each patient, and recommen-dations were communicated to the physician, who coulddecide to accept or reject them on a case-by-case basis.Recommendations were followed 80% of the time andwere found to improve care (defined as the addition of a

329


standard treatment without adverse effects), have noimpact, or worsen care 89.9%, 7.6%, and 2.4% of the time,respectively. Whether this program impacts on hard clini-cal end points requires further follow-up. An electronicmedical record system would enable automation of carealgorithms for implementation and monitoring in largepopulations.

In order for a population-based strategy such as theabove to be implemented, individuals with a specific con-dition or risk factor must be accurately identified andnot missed, another reason why a standardized assess-ment of comorbidity should be routine. The list of co-morbid conditions identif ied as key predictors ofmortality and used for risk-adjustment purposes will bemuch smaller and may be defined differently (more spe-cifically, probably) than those addressed by the care al-gorithms. A standardized assessment of comorbidityshould include the conditions as defined appropriatelyfor both applications.

SUMMARY

There are many reasons why comorbid conditionsshould be measured in a standardized fashion in all di-alysis patients: to enable unbiased comparisons in clini-cal tr ials and quality improvement programs, toimplement strategies for improving comorbid illnessmanagement, to provide projections about prognosisthat aid decision-making, to enable more equitable al-location of resources in order that the current standardsof care can continue to be achieved. A simple yet accu-rate instrument that collects the pertinent informationfor risk stratification purposes and disease managementinitiatives is required. Studies are needed to identify andweight the key conditions that relate to mortality, alsoafter consideration of other demographic and physi-ological variables that are strong predictors and would

be routinely available in a dialysis population. Differentfactors and weights may apply for different outcomes andalso deserve study. In the interim, local and nationalstrategies to improve the documentation and manage-ment of comorbid conditions are greatly needed, as suchefforts may yield improved care and outcomes.

REFERENCES

1. Sarnak M, Levey A. Cardiovascular disease and chronicrenal disease: a new paradigm. Am J Kidney Dis 2000; 35(4Suppl 11):S11731.

2. Foley R, Parfrey PS, Sarnak MJ. The clinical epidemiologyof cardiovascular disease in chronic renal disease. Am JKidney Dis 1998; 32(5 Suppl 3):S11219.

3. United States Renal Data System. Excerpts from the USRDS2004 annual renal data report. Am J Kidney Dis 2005;45(1 Pt 2):1280.

4. Xue J, Ma J, Louis T, Collins A. Forecast of the number ofpatients with end-stage renal disease in the United Statesto the year 2010. J Am Soc Nephrol 2001; 12:27538.

5. Muntner P, Coresh J, Powe N, Klag M. The contribution ofincreased diabetes prevalence and improved myocardialinfarction and stroke survival to the increase in treatedend-stage renal disease. J Am Soc Nephrol 2003; 14:156877.

6. Hsu C-Y, Vittinghoff E, Lin F, Shlipak MG. The incidence ofend-stage renal disease is increasing faster than theprevalence of chronic renal insufficiency. Ann Intern Med2004; 141:95101.

7. Port F. The end-stage renal disease program: trends overthe past 18 years. Am J Kidney Dis 1992; 20:37.

8. Mailloux LU, Napolitano B, Alessandro G, Bellucci G,Mossey RT, Vernace MA, et al. The impact of co-morbid riskfactors at the start of dialysis upon the survival of ESRDpatients. ASAIO J 1996; 42:1649.

9. Hylander B, Lundblad H, Kjellstrand C. Changing patientcharacteristics in chronic hemodialysis. Scand J UrolNephrol 1991; 25:5963.

Figure 2 Collection and uses of comorbidity data in practice.

330


10. Collins AJ, Hanson G, Umen A, Kjellstrand C, Keshaviah P.Changing risk factor demographics in ESRD patients en-tering hemodialysis and the impact on long-term mortal-ity. Am J Kidney Dis 1990; 15:42232.

11. Iezzoni L. Risk Adjustment for Measuring Health Out-comes. 2nd ed. Ann Arbor, MI: Health AdministrationPress; 1997.

12. Greenfield S, Aronow HU, Elashoff RM, Watanabe D. Flawsin mortality data: the hazards of ignoring comorbid dis-ease. JAMA 1988; 260:22535.

13. Stack A. Determinants of modality selection among inci-dent US dialysis patients: results from a national study.J Am Soc Nephrol 2002; 13:127987.

14. Miskulin DC, Meyer KB, Athienites NV, Martin AA, MarshJV, Fink NF, et al. Comorbidity and other factors associ-ated with modality selection in incident dialysis patients:the CHOICE Study. Am J Kidney Dis 2002; 39:32436.

15. Collins A, Weinhandl E, Snyder J, Shu-Cheng C, GilberstonD. Comparison and survival of hemodialysis and perito-neal dialysis in the elderly. Semin Dial 2002; 15:98102.

16. Greenfield S, Sullivan L, Silliman RA, Dukes K, Kaplan SH.Principles and practice of case-mix adjustment: applica-tions to end-stage renal disease. Am J Kidney Dis 1994;24:298307.

17. Wolfe R, Gaylin D, Port F, Held P, Wood C. Using USRDSgenerated mortality tables to compare local ESRD mor-tality rates to national rates. Kidney Int 1992; 42:9916.

18. Khan I, Campbell M, Cantarovich D, Catto G, Delcroix D,Edward N, et al. Comparing outcomes in renal replacementtherapy: how should we correct for case mix? [Publishederratum appears in Am J Kidney Dis 1998; 31:900]. Am JKidney Dis 1998; 31:4738.

19. Athienites NV, Miskulin DC, Fernandez GF, BunnapradistS, Simon G, Landa M, et al. Comorbidity assessment inhemodialysis and peritoneal dialysis using the Index ofCoexistent Disease (ICED). Semin Dial 2000; 13:3206.

20. Sehgal A, Grey S, DeOreo P, Whitehouse P. Prevalence, rec-ognition, and implications of mental impairment amonghemodialysis patients. Am J Kidney Dis 1997; 30:419.

21. Kurella M, Chertow G, Luan J, Yaffe K. Cognitive impair-ment in chronic kidney disease. J Am Geriatr Soc 2004;52:18639.

22. Merkin SS, Longenecker JC, Fink NE, Levey AS, Powe NR.Comparison of patients reports of comorbid disease withthe medical record [Abstract]. J Am Soc Nephrol 2000; 11:157A.

23. United States Renal Data System. USRDS 1991 annual datareport: comorbid conditions and correlations with mor-tality risk among 3,399 incident hemodialysis patients.Am J Kidney Dis 1992; 20:328.

24. Longenecker JC, Klag MJ, Coresh J, Levey AS, Martin AA,Fink NE, et al. Validation of comorbid conditions on theend-stage renal disease medical evidence report: theCHOICE Study. J Am Soc Nephrol 2000; 11:5209.

25. Khan I, Graeme R, Edward N, Fleming L, Henderson I,MacLeod A. Influence of coexisting disease on survival on

renal-replacement therapy. Lancet 1993; 341:41518.26. Nicolucci A, Cubasso D, Labbrozzi D, Mari E, Imicciatore P,

Procaccini DA, et al. Effect of coexistent diseases on sur-vival of patients undergoing dialysis. ASAIO J 1995; 41:M291M295.

27. Miskulin D, Martin A, Brown R, Levey A. Development andvalidation of a dialysis-specific comorbidity index [Ab-stract]. J Am Soc Nephrol 2002; 13:613A.

28. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A newmethod of classifying prognostic comorbidity in longitu-dinal studies: development and validation. J Chronic Dis1987; 40:37383.

29. Fried L, Bernardini J, Piraino B. Charlson ComorbidityIndex as a predictor of outcomes in incident peritonealdialysis patients. Am J Kidney Dis 2001; 37:33742.

30. Hemmelgarn BR, Manns BJ, Quan H, Ghali WA. Adaptingthe Charlson Comorbidity Index for use in patients withESRD. Am J Kidney Dis 2003; 42:12532.

31. van Manen J, Korevaar J, Dekker F, Boeschoten E, BossuytP, Krediet R. How to adjust for comorbidity in survival stud-ies in ESRD patients: a comparison of different indices.Am J Kidney Dis 2002; 40:829.

32. Beddhu S, Bruns FJ, Saul M, Seddon P, Zeidel ML. A simplecomorbidity scale predicts clinical outcomes and costs indialysis patients. Am J Med 2000; 108:60913.

33. Chin M, Goldman L. Correlates of early hospital readmis-sion or death in patients with congestive heart failure.Am J Cardiol 1997; 79:16404.

34. Extermann M. Measuring comorbidity in older cancer pa-tients. Eur J Cancer 2000; 36:45371.

35. Gabriel SE, Crowson CS, OFallon WM. A comparison of twocomorbidity instruments in arthritis. J Clin Epidemiol1999; 52:113742.

36. Davies S, Russell L, Bryan J, Phillips L, Russell G. Comor-bidity, urea kinetics and appetite in continuous ambula-tory peritoneal dialysis patients: their interrelationshipand prediction of survival. Am J Kidney Dis 1995; 26:35361.

37. Davies S, Phillips L, Naish P, Russell G. Quantifying comor-bidity in peritoneal dialysis patients and its relationshipto other predictors of survival. Nephrol Dial Transplant2002; 17:108592.

38. Jager K, Merkus M, Dekker F, Els W, Tijssen J, Stevens P,et al. Mortality and technique failure in patients startingchronic peritoneal dialysis treatment: results of the Neth-erlands Cooperative Study on The Adequacy of Dialysis.Kidney Int 1999; 55:147685.

39. Wright L. Survival in patients with end-stage renal dis-ease. Am J Kidney Dis 1991; 17:258.

40. Greenfield S, Blanco DM, Elashoff RM, Aronow HU. Devel-opment and testing of a new index of comorbidity [Ab-stract]. Clin Res 1987; 35:346A.

41. Miskulin D, Athienites N, Yan G, Martin A, Ornt D, Kusek J,et al. Comorbidity assessment using the Index of Coexist-ent Diseases (ICED) in a multicenter clinical trial: the He-modialysis (HEMO) Study. Kidney Int 2001; 60:1498510.

331


42. Levey AS, Martin AA, Miskulin DC, Athienites NV, MeyerKB. Comorbidity assessment in ESRD: methods and resultsin a national dialysis chain [Abstract]. J Am Soc Nephrol1999; 10:171(A).

43. Miskulin D, Meyer K, Martin A, Fink N, Coresh J, Powe N,et al. Baseline comorbidity and its change predict survivalin an incident dialysis population. Am J Kidney Dis 2003;41:14961.

44. McClellan W, Anson D, Birkeli K, Tuttle E. Functional sta-tus and quality of life: predictors of early mortality amongpatients entering treatment for end-stage renal disease.J Clin Epidemiol 1991; 44:839.

45. Miskulin DC, Meyer KB, Athienites NV, Martin AA, MarshJV, Fink NF, et al. The contribution of medical conditionsand physical impairments to a comorbidity index: theIndex of Coexistent Disease (ICED) [Abstract]. J Am SocNephrol 1999; 10:175(A).

46. Chanda S, Schulz J, Lawrence C, Greenwood R, FarringtonK. Is there a rationale for rationing chronic dialysis? Ahospital based cohort study of factors affecting survivaland morbidity. Br Med J 1999; 318:21723.

47. Hanley JA, McNeil BJ. The meaning and use of the areaunder a receiver operating characteristic (ROC) curve.Radiology 1983; 143:2936.

48. Miskulin D, Martin A, Brown R, Fink N, Coresh J, Powe N,et al. Predicting one-year mortality in an outpatient he-modialysis population: a comparison of comorbidity in-struments. Nephrol Dial Transplant 2004; 19:41320.

49. Cook D. Performance of the APACHE III models in an Aus-tralian ICU. Chest 2000; 118:17328.

50. Knaus W, Wagner D, Draper E, Zimmerman J, Bergner M,Bastos P, et al. The APACHE III prognostic system: risk pre-diction of hospital mortality for critically ill hospitalizedadults. Chest 1991; 100:161936.

51. Spertus JA, Winder JA, Dewhurst TA, Deyo RA, ProdzinskiJ, McDonell M, et al. Development and evaluation of theSeattle Angina Questionnaire: a new functional statusmeasure for coronary artery disease. J Am Coll Cardiol1995; 25:33341.

52. Greenfield S, Sullivan L, Dukes KA, Silliman R, DAgostinoR, Kaplan SH. Development and testing of a new measureof case mix for use in office practice. Med Care 1995; 33(4 Suppl):AS47AS55.

53. DeOreo PB. Hemodialysis patient-assessed functionalhealth status predicts continued survival, hospitalization,and dialysis-associated compliance. Am J Kidney Dis 1997;30:20412.

54. Mapes D, Lopes A, Satayathum S, McCullough K, GoodkinD, Locatelli F, et al. Health-related quality of life as a pre-dictor of mortality and hospitalization: the Dialysis Out-comes and Practice Patterns Study (DOPPS). Kidney Int2003; 64:33949.

55. Zager P, Nikolic J, Brown R, Campbell M, Hunt W, PetersonD, et al. U curve association of blood pressure and mor-tality in hemodialysis patients. Kidney Int 1998; 54:5619.

56. Fleischmann E, Teal N, Dudley J, May W, Bower J,

Salahudeen A. Influence of excess weight on mortality andhospital stay in 1346 hemodialysis patients. Kidney Int1999; 55:15607.

57. Leavey S, Strawderman R, Jones C, Port F, Held P. Simplenutritional indicators as independent predictors of mor-tality in hemodialysis patients. Am J Kidney Dis 1998; 31:9971006.

58. Block G, Hulbert-Shearon T, Levin N, Port F. Associationof serum phosphorus and calcium x phosphate productwith mortality risk in chronic hemodialysis patients: anational study. Am J Kidney Dis 1998; 31:60417.

59. Kalantar-Zadeh K, Kopple J, Block G, Humphreys M. Amalnutrition-inflammation score is correlated with mor-bidity and mortality in maintenance hemodialysis pa-tients. Am J Kidney Dis 2001; 38:125163.

60. Zimmermann J, Herrlinger S, Pruy A, Metzger T, WannerC. Inflammation enhances cardiovascular risk and mor-tality in hemodialysis. Kidney Int 1999; 55:64858.

61. Myers G, Rifai N, Russell P, Roberts W, Alexander R, BiasucciL, et al. CDC/AHA workshop on markers of inflammationand cardiovascular disease. Circulation 2004; 110:e5459.

62. Miskulin D, Coresh J, Klag M, Levin N, Fink N, Powe N, et al.Can serum CRP and IL-6 levels replace comorbidity in pre-dicting mortality in dialysis patients [Abstract]? J Am SocNephrol 2004; 15:382A.

63. Elwood M. Critical Appraisal of Epidemiological Studiesand Clinical Trials. 2nd ed. New York, NY: Oxford; 1998.

64. Greene T, Beck GJ, Gassman JJ, Gotch FA, Kusek JW, LeveyAS, et al. Design and statistical issues of the Hemodialy-sis (HEMO) Study. Control Clin Trials 2000; 21:50225.

65. Liu Y, Coresh J, Eustace J, Longenecker J, Jaar B, Fink N,et al. Association between cholesterol level and mortal-ity in dialysis patients. JAMA 2004; 291:4519.

66. Lowrie EG, Lew NL. Death risk in hemodialysis patients:the predictive value of commonly measured variables andan evaluation of death rate differences between facilities.Am J Kidney Dis 1990; 15:45882.

67. Iseki K, Yamazato M, Tozawa M, Takishita S. Hypocholes-terolemia is a significant predictor of death in a cohort ofchronic hemodialysis patients. Kidney Int 2002; 62:188793.

68. Coresh J, Longenecker J, Miller E. Epidemiology of car-diovascular risk factors in chronic renal disease. J Am SocNephrol 1998; 9(Suppl 1):S2430.

69. Kalantar-Zadeh K, Block F, Humphreys M, Kopple J. Re-verse epidemiology of cardiovascular risk factors in main-tenance dialysis patients. Kidney Int 2003; 63:793808.

70. Fleischmann E, Bower J, Salahudeen A. Risk factor para-dox in hemodialysis: better nutrition as a partial expla-nation. ASAIO J 2001; 47:7481.

71. Leavey SF, McCullough K, Hecking E, Goodkin D, Port F,Young E. Body mass index and mortality in healthier ascompared with sicker haemodialysis patients: resultsfrom the Dialysis Outcomes and Practice Patterns Study(DOPPS). Nephrol Dial Transplant 2001; 16:238694.

72. Stack A, Molony D, Rahman N, Dosekun A, Murthy B.

332


Impact of dialysis modality on survival of new ESRD pa-tients with congestive heart failure in the United States.Kidney Int 2003; 64:10719.

73. Stack A, Murthy B, Molony D. Survival differences betweenperitoneal dialysis and hemodialysis among large ESRDpatients in the United States. Kidney Int 2004; 65:2398408.

74. Centers for Medicare & Medicaid Services. 2001 ESRD Clini-cal Performance Measures Project. Am J Kidney Dis 2002;39(5 Suppl 3):S198.

75. Greenfield S, Nelson EC, Zubkoff M, Manning W, Rogers W,Kravitz RL, et al. Variations in resource utilization amongmedical specialties and systems of care. JAMA 1992; 267:162430.

76. Lacson Jr E, Teng M, Lazarus J, Lew N, Lowrie E, Owen W.Limitations of the facility-specific standardization mor-tality ratio for profiling health care quality in dialysisunits. Am J Kidney Dis 2001; 37:26775.

77. Chertow GM. Leveling the paying field in end-stage renaldisease. Am J Med 2000; 108:6668.

78. Tonelli M, Gill J, Pandeya S, Bohm C, Levin A, Kiberd B.Barriers to blood pressure control and angiotensin enzyme

inhibitor use in Canadian patients with chronic renal in-sufficiency. Nephrol Dial Transplant 2002; 17:142633.

79. Manley H, Garvin C, Drayer D, Reid G, Bender W, Neufeld T,et al. Medication prescribing patterns in ambulatory he-modialysis patients: comparisons of USRDS to a large not-for-profit dialysis provider. Nephrol Dial Transplant 2004;19:18428.

80. National Kidney Foundation. K/DOQI clinical practiceguidelines for managing dyslipidemias in chronic kidneydisease. Am J Transplant 2004; 4(Suppl 7):753.

81. Nissenson A, Collins J, Dickmeyer J, Litchfield T, MatternW, McMahill C, et al. Evaluation of disease-state manage-ment of dialysis patients. Am J Kidney Dis 2001; 37:93844.

82. Manley H, Drayer D, Muther R, Hebbar S, Miskulin D. Im-pact of multidisciplinary team on cardiovascular risk (CV)reduction in hemodialysis (HD) patients: Dialysis Risk Fac-tor Intervention Trial [Abstract]. J Am Soc Nephrol 2004;15:191A.

83. Manley H, Drayer D. Clinical pharmacy interventions re-duce ambulatory hemodialysis (HD) patients cardiovas-cular risk [Abstract]. J Am Soc Nephrol 2004; 15:3A.

320.full.pdf

Documents