 |
||||
| 2007 JASN IMPACT FACTOR 7.111 | HOME AUTHOR INFO EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP | |||
| CURRENT ISSUE | ARCHIVES | JASN Express | ONLINE SUBMISSION | |
|
||||
| 2007 JASN IMPACT FACTOR 7.111 | HOME AUTHOR INFO EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP | |||
| CURRENT ISSUE | ARCHIVES | JASN Express | ONLINE SUBMISSION | |
Division of Renal Diseases and Hypertension, Department of Internal Medicine, University of Texas Health Science Center, Houston, Texas; Kidney Epidemiology and Cost Center, University of Michigan, Ann Arbor, Michigan
Correspondence to: Dr. Austin G. Stack, Division of Renal Diseases and Hypertension, University of Texas Health Science Center at Houston, 6431 Fannin St, MSB 4.148, Houston, TX 77030. Phone: 713-500-6873; Fax: 734-998-6620; Email: austin.stack{at}uth.tmc.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
4 mo) and seen more frequently by a nephrologist (
2 visits versus <2 visits) in the pre-ESRD period had greater PD use. Of the factors listed, 25% of the variability (R2) in PD use was explained by demographic (4.1%), comorbid (1.2%), social/pre-ESRD (14.5%), and geographic (5.2%) factors. This study identifies several clinical, social, and pre-ESRD factors with the selection of PD, and it underscores the importance of patient education, autonomy, and a strong social support system in improving rates of PD use in the United States. As pre-ESRD patient care is an important contributor to PD use in the United States, greater efforts should be expended in improving its delivery earlier in the pre-ESRD period. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
There are few published studies available that have identified clinical predictors of modality choice among new ESRD patients. Rarer still are those that have addressed the relationship of nonclinical factors with modality assignment. Existing studies have identified differences in patient demographics, including age, gender, and race, in the use of PD over HD (15). In an analysis of data from the United States Renal Data System (USRDS), patients selected to PD were more likely to be younger in age and have fewer comorbid medical conditions than their HD counterparts. More recently, the relationship between modality assignment and several clinical and nonclinical factors was explored in two national surveys of practicing US nephrologists (13,16). Although the survey instruments differed in each of these studies, patient preference, compliance, quality of life, and a strong social support system were identified as the most important nonclinical factors in selecting PD from a nephrologist’s perspective. Although these studies have contributed greatly to our understanding of modality assignment, they have not adequately addressed the relative contribution of other potentially important psychosocial factors, including education, physical disability, insurance, employment, and several clinical and geographic factors in an adjusted analysis. Furthermore, it should be noted that these survey instruments merely recorded the opinions of nephrologists on modality selection; therefore, they did not necessarily reflect actual rates of PD use in the United States.
The purpose of this study was to (1) explore the relationship of clinical, social, and pre-ESRD factors with modality selection among new ESRD patients in the United States and to (2) determine the relative contribution of demographic, geographic, social, and pre-ESRD factors on modality assignment. The USRDS Dialysis Morbidity and Mortality Study Wave 2 (DMMS 2) allowed us a unique opportunity to investigate these associations in a nationally representative sample of new ESRD patients as data on each of the domains listed was captured through a medical and patient questionnaire.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Patient data on the presence of medical conditions and assignment to PD or HD was obtained through a Medical Questionnaire completed by personnel at each dialysis facility. Data on demographics, prior medical history, laboratory results, dialysis prescription, vascular access, and medications were obtained from review of patients’ medical records. To capture data on quality of life, modality selection, pre-ESRD care, and psychosocial aspects of ESRD, a Patient Questionnaire, developed through a collaboration between the RAND Corporation (Santa Monica, CA) and Amgen Inc, (Thousand Oaks, CA) was included as part of the data collection instrument. This used the Kidney Disease Quality of Life (KDQOL) Short Form, a validated kidney-specific quality of life questionnaire to capture several domains of interest (19). Follow-up questionnaires were administered 9 to 12 mo after enrollment, and the data collection was completed in early 1999.
Data Analyses
The primary outcome variable was modality assignment to PD (as compared with HD) in new ESRD patients. The modality assignment for patients on HD but who were training for PD on day 60 was deferred for 10 d. Multivariate models of increasing complexity were constructed to explore the relationships of demographic, clinical, psychosocial, pre-ESRD care, and geographic factors with modality assignment. Factors were included only if they were perceived to be clinically important in deciding modality type or if univariate analysis indicated a significant association.
The first multivariate model evaluated the relationship of demographic and medical conditions with assignment to PD as compared with HD at ESRD start. The demographic variables modeled included age at ESRD onset, gender, and race, as these have been shown to be associated with modality choice in smaller center studies. A total of 11 comorbid medical conditions were represented in this model, including cause of ESRD, coronary artery disease (CAD), congestive heart failure (CHF), peripheral and cerebral vascular disease, structural cardiac abnormalities, pericarditis, malignancy, indicators of nutrition, and tobacco use.
The second multivariate analysis explored the relationship of psychosocial, economic, and geographic factors with modality assignment in a series of models with adjustment for demographic and comorbid medical conditions listed above. In these series of models, we determined the association of employment status (full-time or part-time versus unemployed, retired, homemaker, never employed, or other), living status (living alone versus not living alone, institutionalized, and homeless), marital status (married versus single, widowed, divorced, or separated), level of education (high school, some college, and college graduate versus elementary), and insurance status (Medicare only, Medicaid only, Medicare and Medicaid only, Veterans Administration or other insurance, and private health insurance versus no insurance) with modality assignment.
To address the contribution of patient versus the medical team in deciding modality choice, the questionnaire included a question that specifically asked respondents "who played the key role in this decision?," to which the answers included patient, medical team, or both. This permitted an evaluation of the association of patient’s autonomy with modality selection. Given the importance of pre-ESRD care on ESRD outcomes, we also included variables in the model to represent these and examine their relative contribution to modality selection. These included the timing of nephrology referral and the number of visits to a nephrologist or dietitian before ESRD start, as we hypothesized that these might be important determinants of modality selection. Late referral was arbitrarily defined as the period when patients were first seen by a nephrologist within 4 mo of ESRD start. We also determined the contribution of geographic region to modality assignment by including all 18 ESRD network areas as additional covariates and compared the likelihood of PD selection over HD with that of the national average. Each factor or factor group was entered singularly into the model with adjustment for demographic and clinical variables, and the corresponding adjusted odds-ratio (AOR) was computed.
In the third and final model, all significant associations from previous models were included to identify the independent contribution of each factor to modality assignment while adjusting for all other factors in the model. Multivariate logistic regression analysis determined the relationship of each covariate in each model to the outcome variable using an AOR. As the response rate to the pre-ESRD and the patient preference component of the patient questionnaire was relatively low (65% and 62%, respectively) compared with all other components of the questionnaire (88%), analyses exploring these relationships with modality selection were limited only to those patients who had responded. For each model the coefficient of determination (R2) was computed. All statistical analyses were performed using SAS version 6.12 (SAS Institute, Cary, NC).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Patients who were started on PD had a lower prevalence of vascular comorbid conditions and structural cardiac abnormalities and a higher prevalence of smokers than their HD counterparts (Table 1). Serum hematocrit and calcium levels were higher and serum albumin levels were lower in patients selected to PD compared with HD patients. There were no significant differences in serum creatinine levels (8.2 ± 3.5 versus 8.3 ± 3.8 mg/dl) or in the estimated level of residual renal function (8.0 ± 2.9 versus 8.1 ± 3.1 ml/min) in those assigned to PD compared with HD.
|
4 mo before ESRD start) and for those who were seen by a nephrologist more frequently (2 visits) compared with less frequently (<2 visits) before dialysis initiation.
|
|
|
|
The relationship between nonclinical factors and modality assignment is illustrated in Table 4. For these analyses, each variable was modeled separately (a total of 11 models) with adjustment for all clinical factors listed. Patients at ESRD start who were either in full-time or part-time employment, compared with those who were not, were more likely to be assigned to PD in favor of HD (AOR, 2.24 [P < 0.0001] and 1.74 [P < 0.001], respectively). Similarly the selection of PD versus HD varied with level of education. Compared with patients who received elementary education only, those who completed high school, some college, or graduated from college had greater likelihood of placement on PD (AOR, 1.42 [P < 0.001], 1.67, and 2.74, respectively [P < 0.0001]). In this analysis, patients who were married or not living alone were more likely to receive PD therapy. Assignment to PD was also associated with patient’s level of physical functioning. Patients who were able to ambulate independently or eat without assistance had higher odds of placement on PD (AOR, 2.59 [P < 0.0001] and 2.08, respectively [P < 0.05]).
|
4 mo before ESRD versus <4 mo), and those who visited a nephrologist more frequently (>2 visits versus <2 visits) were more likely to be placed on PD (AOR, 1.80 and 1.73, respectively [P < 0.0001]). We found that the selection of PD over HD varied by geographic location, adjusting for several other clinical parameters. New ESRD patients starting dialysis in ESRD network areas 9, 11, and 16 were significantly more likely to receive PD (AOR, 2.73, 1.56, and 3.48, respectively), and those starting dialysis in networks 7, 8, and 14 were less likely to receive PD (AOR, 0.39, 0.51, and 0.40, respectively) compared with the national average for the United States.
The final multivariate analysis explored the independent relationship of all demographic, clinical, and nonclinical factors with modality assignment in a single model (Table 5). The relationships observed here were similar in magnitude and strength to those already mentioned. Of the factors listed, 25% of the variability in PD use was explained by demographic (4.1%), clinical (1.2%), geographic factors (5.2%), and social/pre-ESRD factors (14.5%).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The choice of PD over HD was negatively correlated with advancing age. With the increasing acceptance of elderly patients for renal replacement therapy in the United States and the preferential use of HD in this group, it is likely that rates of PD use will fall over time.
We also found a strong gender association that was modified by age. Older males and younger females were more likely to receive PD in favor of HD with adjustment for several other clinical factors. Our findings are at variance with the recently reported observations of Thamer et al. (13) and Little et al. (14). In the Thamer study, a national survey of practicing US nephrologists, men were almost twice as likely as women to be recommended for PD; whereas the Little study, a longitudinal study of over 254 patients in the United Kingdom, reported that PD use was more common in females. Our data suggest that the relationship between gender and modality type is more complex as a result of strong age interaction. Whether these findings reflect differences in attitudes to body image between males and females and for different age groups or point to differences in lifestyle or social characteristics between the sexes is unclear.
Another area of controversy, recently highlighted, is the relationship of race with choice of modality. Our results are in contrast with those of Thamer et al. (13), who failed to find any significant differences in nephrologists’ recommendation of PD for black patients over white patients. We found that white patients were more likely to be selected for PD than HD (AOR, 1.90; P < 0.0001), a finding that persisted despite a comprehensive adjustment for 11 preexisting comorbid conditions. Although Thamer et al. (13) suggests that lack of adjustment for differences in body size, typically greater in black patients, may account for the lower use of PD in this racial group, our analysis does not support this hypothesis. Adjusting for BMI did not decrease likelihood of PD use in whites patients. Additional adjustment for social and pre-ESRD patient care factors, however, caused the AOR to fall to 1.62, suggesting that nonclinical factors only partly explain the greater use of PD in white patients compared with black patients.
Our findings extend the observations of Thamer et al. (13), who found that nephrologists were less likely to recommend PD as an option for patients with body weight >91 kg (200 lbs.). We provide a more comprehensive assessment of this relationship and find a U-shaped correlation between body size and PD use. In this analysis, not only larger patients but also smaller patients were less likely to receive PD compared with HD. Although we cannot be certain, we speculate that these associations reflect nephrologists’ concerns in providing adequate dialysis to these susceptible subgroups.
The likelihood of receiving PD was lower for patients with selected comorbid conditions as well as for patients with low serum albumin. The lower use of PD over HD among patients with congestive heart failure was an unexpected finding and contrary to current opinion, despite the perceived potential benefits of greater hemodynamic stability and fewer electrolyte disturbances. Lower solute clearance and higher cardiac risk profile associated with PD may also prejudice the nephrology team toward HD selection in patients with preexisting cardiovascular disease (20,21). These data also confirm several selection biases toward modality assignment in new ESRD patients and underscore the importance for comprehensive adjustment of these factors in multivariate analyses.
A novel aspect of this study is the finding of different rates of PD use by geographic location. The likelihood of placement on PD was low in the south and southeast United States and highest in the north and northwest United States. One might speculate that the lower use of PD in ESRD Network Area 7 (Florida) is in part contributed to by an older prevalent population while higher use in Network 16 (Washington, Alaska, Idaho, Oregon, Montana) may driven by geographic location and distance to a HD dialysis facility.
There are several unique features in this observational study. First, data were available on several social factors, which have been conventionally perceived as important in deciding the optimal modality at ESRD start. New ESRD patients who were part of the labor force were more likely to be placed on PD over HD, presumably allowing greater flexibility in work schedules. Patients who had progressed further in educational training were more likely to undertake PD as a dialytic modality, possibly reflecting greater competence and a greater understanding of the complexities of this procedure. Consistent with conventional practice and recently reported studies, patients who were physically independent and those with a strong supportive system received greater rates of PD utilization than HD (13).
Second, unlike the Thamer et al. study (13), which explored the choice of modality from a nephrologist’s perspective, this study explored the decision on modality assignment from a patient perspective. We found that patients who had contributed to the decision-making process were more likely to be subsequently placed on PD, and those who felt they had not were more likely to receive HD. This suggests that patients who were more assertive and more autonomous in the decision-making process favored PD over HD.
Third, the association of pre-ESRD care and timing of nephrology referral with modality assignment is largely unexplored. Two recent studies have identified early nephrology referral and patient counseling before ESRD start as strong determinants of PD placement (14,22). We corroborate their findings with the observation that new ESRD patients were over twice as likely to be placed on PD if they were referred earlier to a nephrologist for pre-ESRD counseling. Similarly, frequent attendance at a nephrologist’s office was associated with a greater likelihood of placement on PD over HD. These findings highlight the importance of early pre-ESRD care in permitting satisfactory patient counseling on modality choice and facilitating timely placement of vascular or peritoneal access.
Finally, this study is unique in that it explored the relative contribution of demographic, geographic, clinical, social, and pre-ESRD care factors on initial modality assignment in a multivariate analysis. Although some nephrologists may believe that the selection of PD is mainly determined by clinical and geographic factors, this study suggests that social and pre-ESRD factors are also important factors in deciding initial modality assignment.
Although this study has strengths, its limitations must also be recognized. First, the complex interaction of clinical, social, and geographic factors, all of which may contribute to modality assignment, cannot be adequately explored in a cross-sectional study. To address the issue of modality choice in a more comprehensive fashion, a prospectively-designed cohort of pre-ESRD patients would be required. Second, although the overall response rate for most components of the patient questionnaire was in excess of 88%, lower response rates were recorded for patients who answered questions relating to pre-ESRD patient care (65%) and patient versus physician preference (62%) on modality assignment. It is possible that patients who did not respond to these questions may have behaved differently than those that did respond and altered the relationships that we observed. However, in an analysis comparing responders to nonresponders with respect to several measured characteristics, we did not detect any major differences. Third, we considered the potential for recall bias in this type of study, which might be present between subgroups of new ESRD patients. It is possible that "sicker" patients with less recall of their pre-ESRD care perhaps favored HD, whereas those with greater recall favored PD. We explored this possibility by modeling pre-ESRD care elements as a function of selected comorbidities (CHF, CAD, etc) and found that the odds of having pre-ESRD care were not greater for those with selected comorbidities compared with those without. Finally, our multivariate model only explained 25% of the total variance, suggesting that there are several other potentially important factors that were not included in the model that determine initial modality selection.
This study highlights the importance of demographic, geographic, clinical, social, and pre-ESRD factors with initial modality among new ESRD patients in the United States. It suggests that social and pre-ESRD patient care factors are strong contributors to this decision-making process. It also underscores the importance of education, patient autonomy and pre-ESRD care in improving rates of PD use in the United States. Furthermore, the presence of several selection biases with respect to modality assignment may confound survival analysis comparing PD with HD and stress the need for a more comprehensive adjustment of these factors in multivariate analysis.
|
Acknowledgments |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  R. A. Rodriguez, S. Sen, K. Mehta, S. Moody-Ayers, P. Bacchetti, and A. M. O'Hare Geography Matters: Relationships among Urban Residential Segregation, Dialysis Facilities, and Patient Outcomes Ann Intern Med, April 3, 2007; 146(7): 493 - 501. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. G. Kutner, R. Zhang, Y. Huang, and D. L. Bliwise Association of Sleep Difficulty with Kidney Disease Quality of Life Cognitive Function Score Reported by Patients Who Recently Started Dialysis Clin. J. Am. Soc. Nephrol., March 1, 2007; 2(2): 284 - 289. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Saxena and C. West Peritoneal dialysis: a primary care perspective. J Am Board Fam Med, July 1, 2006; 19(4): 380 - 389. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. K. Inrig, J. L. Sun, Q. Yang, L. P. Briley, and L. A. Szczech Mortality by Dialysis Modality among Patients Who Have End-Stage Renal Disease and Are Awaiting Renal Transplantation Clin. J. Am. Soc. Nephrol., July 1, 2006; 1(4): 774 - 779. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. G. Kutner, R. Zhang, H. Barnhart, and A. J. Collins Health status and quality of life reported by incident patients after 1 year on haemodialysis or peritoneal dialysis Nephrol. Dial. Transplant., October 1, 2005; 20(10): 2159 - 2167. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. G. Kutner, R. Zhang, and D. Brogan Race, Gender, and Incident Dialysis Patients' Reported Health Status and Quality of Life J. Am. Soc. Nephrol., May 1, 2005; 16(5): 1440 - 1448. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. M. Chertow A 43-Year-Old Woman With Chronic Renal Insufficiency JAMA, March 10, 2004; 291(10): 1252 - 1259. [Full Text] [PDF]
|
|
|
|
|
|
S. K. Ganesh, T. Hulbert-Shearon, F. K. Port, K. Eagle, and A. G. Stack Mortality Differences by Dialysis Modality among Incident ESRD Patients with and without Coronary Artery Disease J. Am. Soc. Nephrol., February 1, 2003; 14(2): 415 - 424. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
HOME
CURRENT ISSUE
ARCHIVES
JASN Express
ONLINE SUBMISSION
AUTHOR INFO
EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP |
