- Email: [email protected]
ESRD providers
10
ESRD providers I love people who harness themselves, an ox to a heavy cart, who pull like water buffalo, with massive patience, who strain in the mud and the muck to move things forward, who do what has to be done, again and again.
marge piercy “to be of use”
American Journal of Kidney Diseases, Vol 47, No , Suppl (January, 2006): pp s85–s98. 10 providers 05.indd 185
11/18/2005 7:37:52 AM
10 s186
ESRD PROVIDERS
Introduction
T
The growing number of patients with ESRD has led to increasing decentralization of the dialysis delivery system—once limited to major medical centers, and now spread to suburban and more rural communities. Restrictive reimbursement systems, in which only minor increases have occurred since 982, have led to major consolidations in the ESRD provider networks. Although this report centers on data through the end of 2003, the months since December of 2004 have brought continued consolidation of the dialysis chains into two large providers, both for-profit, publicly-traded corporations. Recent changes such as this are illustrated in Figure 0., which shows the continued growth of freestanding, for-profit dialysis units in the United States, with slight declines in the number of non-profit and hospital-based providers. Characteristics of the dialysis population across this large provider system have been the focus of this chapter for many years. In this year’s first spread, we present information on provider-level anemia at initiation, treatment with erythropoietin in the prevalent population, and the variation in the erythropoietin doses based on achieved hemoglobin groups and iron therapy. At the initiation of hemodialysis, units owned by Dialysis Clinics, Inc. and National Nephrology Associates have the greatest percentage of patients with hemoglobin levels less than g/dl.
Across all hemoglobin levels, however, the highest EPO doses are given by DaVita. Data on EPO doses and hemoglobin levels by unit affiliation show the considerable variation across providers in the last six months of 2003. Clear in these data is the fact that the high EPO doses used by DaVita moved patients from hemoglobin levels less than g/dl in June, 2003, to hemoglobins of at least 2 g/dl within three months, a level which then stabilized. Geographic variations in hemoglobin levels vary by profit status, as do variations in the percent of patients receiving intravenous iron. These differences in provider practices suggest that there is less consistent application of the target guidelines for treatment, something which needs careful consideration when addressing the cost-effectiveness of these interventions. Data on provider growth show a striking decline in the use of peritoneal dialysis across ESRD networks. This is particularly true in Networks , 2, 4, 9, and 0. Only in Network 6 did the number of peritoneal dialysis treatments increase between 999 and 2003. Interestingly, non-chain affiliated units dominate Network 2, which has seen the steepest decline in peritoneal dialysis treatments. Data on the demographic characteristics of the incident population show that the distribution of patients by age and gender is similar in the top 4–5 chains. The same is true for other
s188 · provider-level patient distribution by hemoglobin · anemia EPO dosing · geographic variations · management iron therapy
s194 · provider anemia treatment & dialysis compliance with adequacy · vascular access use · K/DOQI guidelines nutritional parameters
s190 · differences dialysis treatments · unit location in provider & growth · unit & patient counts · growth affiliation & profit status s192 · patient demographics & clinical parameters characteristics, by of incident & prevalent patients · unit unit affiliation & patient counts, by unit affiliation
10 providers 05.indd 186
s196 · preventive influenza, pneumonia, & hepatitis care, by unit B vaccinations · lipid monitoring · affiliation glycosylated hemoglobin testing s198 · summary
10/13/2005 8:44:20 AM
ESRD PROVIDERS
10
2005 Annual Data Report
demographic characteristics, though hospitalbased units have the highest percentage of patients with high albumin levels. We look next at provider compliance with guidelines for care advanced by K/DOQI. DaVita, for example, has the greatest percentage of patients with an average hemoglobin of 2 g/dl or more throughout the year, while Dialysis Clinics, Inc. has the greatest percentage whose hemoglobins meet the K/DOQI target of –2 g/ dl. Once again, there is considerable geographic variation in mean hemoglobin levels of patients treated in chain versus non-chain units, as well as in associated EPO dosing patterns. Vascular access use appears to be similar across providers, with catheter use highest in hospital-based units. Use of an internal access does, however, vary across the country by chain and non-chain status. And interestingly, albumin, creatinine, and hemoglobin levels in patients beginning ESRD therapy vary little by provider. The final spread in this chapter addresses preventive care. Influenza, pneumococcal pneumonia, and hepatitis B vaccination rates have changed considerably over the years, and have now reached a degree of stability. Chains have tended to vaccinate a greater percentage of pa-
3,500
s187
Freestanding for-profit Freestanding non-profit Hospital center
3,000 Number of units
tients for influenza and for pneumonia than have non-chain units. It is interesting to note that the 2002–2003 rate of pneumococcal pneumonia vaccinations in units owned by Renal Care Group far outpaces that seen in other providers. Diabetic care also varies considerably across providers. Lipid testing in diabetic patients, for example, appears to be used least in units owned by Renal Care Group and Dialysis Clinics, Inc., while recommended testing of hemoglobin AC levels is given most often in DaVita units. The quality of care given to the dialysis population differs between providers, particularly between the chain-affiliated units and their non-chain or hospital-based counterparts. Although the Clinical Performance Measures program of the Centers for Medicare and Medicaid Services has focused primarily on dialysis delivery, anemia treatment, and vascular access, other aspects of care are clearly a concern well. The largest dialysis provider groups show considerable room for improvement. These areas and others will be explored further in subsequent Annual Data Reports to help us better assess provider performance.
2,500 2,000 1,500
Hospital facility Transplant & dialysis center Transplant center
1,000 500 0
90
92
94
96
98
00
02
0. · Counts of dialysis & transplant units, by CMS certification type data obtained from the CMS annual End-Stage Renal Disease Facility Survey, CMS Independent Renal Facility Cost Reports, & the CMS “Dialysis Facility Compare” website. The leveling out of the number of freestanding, for-profit units in 2002 is due to changes in how CMS determines profit status, resulting in some units not being classified.
Chapter highlights
Figure 10.8 Geographic comparisons indicate that patients in for-profit units who have hemoglobins of 12 g/dl or greater are more likely to receive iron than those in non-profit units; mean percentages in the lower quintile differ by more than 16 percent. Figure 10.9 The total number of in-center hemodialysis treatments increased 24.3 percent between 1999 and 2003—from 33 to 41 million. Growth ranged from 15.3 percent in Network 4 to nearly 33 percent in Network 9. Figure 10.14 More than three-quarters of patients are
treated in freestanding, for-profit dialysis units, up from 65 percent in 1996. Figure 10.27 Patients in chain-owned units are more likely than those in non-chain units to receive an influenza vaccination, at 64 and 55 percent, respectively.
10 providers 05.indd 187
10/13/2005 8:44:24 AM
Provider-level anemia management
s188
10 providers 05.indd 188
10.2
Distribution of hgb providers of period groups, by unit affiliation prevalent HD pts.
10.3
100
Percent of providers
60 40 20 0
10.4
Distribution of hgb providers of period groups, by unit type prevalent HD patients
100 12+ 11-<12 <11
80
Percent of providers
A
cross dialysis providers, hemoglobin levels average less than g/dl in 23 percent of units, –<2 g/dl in 33 percent, and 2 g/dl or higher in 44 percent (Figure 0.2). The target set by NKF’s Kidney Dialysis Outcomes Quality Initiative (K/DOQI) is a hemoglobin of g/dl or greater. Levels are highest in DaVita and Renal Care Group units, at 2 g/dl or greater in 48–50 percent of units, and lowest in facilities owned by Dialysis Clinics Incorporated or National Nephrology Associates—in 27 and 38 percent of these units, respectively, hemoglobin levels average less than g/dl. Hemoglobin levels vary less by unit type (Figure 0.3). Regardless of profit status or hospital-based/freestanding status, average hemoglobin levels are less than g/dl in 23– 24 percent of units. The percent of units in which levels are 2 g/dl or greater is slightly higher in for-profit units compared to nonprofit units—44.2 and 40.6, respectively— but is the same in freestanding and hospital-based units. EPO dosing practices appear to vary by unit affiliation, with chain-affiliated units tending to give higher doses of EPO to patients with hemoglobins less than g/dl. Chain units, for example, have a greater proportion of patients (4–4 percent) with hemoglobins less than 0 g/dl in the top quartile of EPO dose compared to nonchain units; DaVita has the largest proportion of patients in the top quartile of EPO dosing for all hemoglobin levels. Assessment of anemia treatment over a six-month period shows similar patterns (Figure 0.5). Chain units again tend to give the highest amount of EPO to patients whose hemoglobins fall below g/dl. In patients with hemoglobins above 2 g/dl, differences in dosing patterns are much less apparent. Mean hemoglobins in the less than g/dl group rise in the early months of treatment and then plateau, in contrast to those in the greater than 2 g/dl group, where slight decreases occur in the initial months followed by a stable pattern. Geographic variations in mean weekly EPO dose by profit status show only slight differences between for-profit and nonprofit units (Figure 0.6). In patients with hemoglobins of 2 g/dl or greater, for example, the mean EPO dose in the top quintile of
All 1 2 3 4 5 6 NC HB Unit affiliation (see table below for codes)
12+ 11-<12 <11
80 60 40 20 0
Profit
Nonprofit
Hospital- Freebased standing
EPO dosing, by starting period prevalent hemoglobin & unit affiliation hemodialysis patients 10-<11 g/dl
100 <10 g/dl 18,000+ units 12,000-<18,000 units 6,000-<12,000 units <6,000 units
80 60 Percent of patients
10
ESRD PROVIDERS
40 20 0
12+ g/dl
100 11-<12 g/dl 80 60 40 20 0
All
1
2
3
4
All 1 2 3 5 6 NC HB Unit affiliation (see table below for codes)
non-profit units is less than 7 percent higher than the mean dose in for-profit units. It appears that iron therapy is used in some degree by all unit types to supplement the positive effects of EPO (Figures 0.7–8). NNA tends to give the largest amounts of iron in both hemoglobin groups and, with DaVita (Figure 0.5), the highest EPO doses to patients with hemoglobins less than g/ dl. Geographic comparisons indicate that patients in for-profit units who have hemoglobins of 2 g/dl or greater are more likely to receive iron than those in non-profit units; mean percentages in the lower quintile differ by more than 6 percent.
4
5
6
NC HB
Chain · Fresenius Chain 2 · Gambro Chain 3 · DaVita Chain 4 · Renal Care Group Chain 5 · Dialysis Clinics, Inc. Chain 6 · Nat’l Nephrology Assoc. NC · Non-chain units HB · Hospital-based units
10/13/2005 8:44:27 AM
ESRD PROVIDERS
10
2005 Annual Data Report
EPo therapy
iron therapy
10.5 Anemia treatment & hemoglobin period prevalent levels, by unit affiliation hemodialysis patients
10.7
30 25 20
95 June hgb <11 g/dl
June hgb >12 g/dl Fresenius Gambro DaVita RCG DCI Nat'l Neph. Assoc. Non-chain Hospital-based
15 10
Cumulative percent of patients
EPO dose (in 1,000s of units)
35 EPO dose/week: June hgb <11 g/dl
Patients receiving iron therapy, period prevalent by hemoglobin level & unit affiliation hemodialysis patients
85
75
Fresenius Gambro DaVita RCG DCI
Nat’l Neph. Assoc. Non-chain Hospital-based
July Aug Sept Oct Nov Dec
July Aug Sept Oct Nov Dec
65
55
13 Mean hemoglobin
June hgb >12 g/dl
s189
Hemoglobin (g/dl)
10.8 Geographic variations in the percent period prevalent
of patients receiving iron, by state hemodialysis patients
12
Non-profit: Hgb <11 g/dl
11
10
10.6
Profit: Hgb <11 g/dl
July Aug Sept Oct Nov Dec
July Aug Sept Oct Nov Dec
90.9 + (95.1) 85.7 to <90.9 80.4 to <85.7 74.4 to <80.4 below 74.4 (65.2) Insuff. data
Geographic variations in provider- period prevalent level mean weekly EPO dose, by state hemodialysis patients
Profit: Hgb <11 g/dl
Non-profit: Hgb <11 g/dl Profit: Hgb 12+ g/dl
29.1+ (33.4) 25.7 to <29.1 23.5 to <25.7 17.9 to <23.5 below 17.9 (8.5) Insuff. data
Profit: Hgb 12+ g/dl
13.6 + (14.6) 12.7 to <13.6 11.6 to <12.7 9.7 to <11.6 below 9.7 (9.1) Insuff. data
10 providers 05.indd 189
29.1 + (32.9) 25.7 to <29.1 23.5 to <25.7 17.9 to <23.5 below 17.9 (14.4) Insuff. data
Non-profit: Hgb 12+g/dl
13.6 + (15.6) 12.7 to <13.6 11.6 to <12.7 9.7 to <11.6 below 9.7 (7.4) Insuff. data
90.9 + (96.9) 85.7 to <90.9 80.4 to <85.7 74.4 to <80.4 below 74.4 (63.0) Insuff. data
Non-profit: Hgb 12+g/dl
90.9 + (94.0) 85.7 to <90.9 80.4 to <85.7 74.4 to <80.4 below 74.4 (68.7) Insuff. data
90.9 + (94.6) 85.7 to <90.9 80.4 to <85.7 74.4 to <80.4 below 74.4 (52.6) Insuff. data
{Figures 0.2–3} providers of period prevalent hemodialysis patients who have at least three months of dialysis prior to June, 2003, a valid EPO claim in June, 2003, & an identifiable provider. {Figure 0.4} period prevalent hemodialysis patients who have at least three months of dialysis prior to June 2003, a valid EPO claim in June & July of 2003, & an identifiable provider. Hemoglobin groups are determined from June claims, & EPO doses from July claims. {Figures 0.5–8} period prevalent hemodialysis patients who have at least three months of dialysis prior to June 2003, a valid EPO claim in June 2003, & an identifiable provider; who dialyze at the same provider from June–December 2003 & have a valid EPO claim in each of those months; & who are from a provider with at least ten such patients.
10/13/2005 8:44:35 AM
s190
Differences in provider growth 10.9 In-center treatments (in millions)
T
he total number of in-center hemodialysis treatments increased 24.3 percent between 999 and 2003— from 33 to 4 million (Figure 0.9). Growth ranged from 5.3 percent in Network 4 (Delaware and Pennsylvania) to nearly 33 percent in Network 9 (Indiana, Kentucky, and Ohio). The number of peritoneal dialysis treatments varies widely from year to year; 22,378 patients were reported to be on this therapy in 2003. Provider growth between 999 and 2003 was higher for units with chain affiliation than for non-chain units (Figure 0.0). During this period, 39 units were dropped from chain status and 758 units were added, a net gain of 69 units. Conversely, 4 units were dropped from non-chain status and 526 were added, a net gain of 5, demonstrating a continued trend towards private ownership within renal providers. In Network 6—Alaska, Idaho, Montana, Oregon, and Washington—the proportion of units that are chain-affiliated rose from 7 percent in 999 to 49 percent in 2003 (Figure 0.), the most dramatic change seen among the networks. Only in Networks 2, 3, , 2, and 6 do non-chain units account for more than half of those providing dialysis; in Network 2 (New
Dialysis treatments, by ESRD network & modality 4.0 Hemodialysis 1999 3.5 2003 3.0 2.5 2.0 1.5 1.0 0.5 0.0 45 Peritoneal dialysis 40
Treatments (in thousands)
10
ESRD PROVIDERS
35 30 25 20 15 10 5 0
1
2
3
4
5
6
7
8
Chain-affiliated & non-chain units
10.10 Unit growth between 1999 & 2003
Chain units
Non-chain units
Chain units dropped Chain units added Chain units unchanged
Non-chain units dropped Non-chain units added Non-chain units unchanged
10.11 Unit distribution, by ESRD network
100
Non-chain Chain-affiliated
60
All
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
40
99 03
Percent of units
80
20
0
10 providers 05.indd 190
9 10 11 ESRD network
12
13
14
15
16
17
18
York), more than 80 percent of units continue to be independently owned. Since the late 980s, ownership of dialysis units has changed dramatically (Figure 0.2). Counts of units that are nonchain, non-profit, or hospital-based have remained relatively stable or declined, as has the number of patients treated in these units. The number of chain-owned units, in contrast, has grown over eleven-fold since 988, and 3.5 times as many units are now run on a for-profit basis. From 999 to 2003, growth in the total number of both units and patients was 9.5 percent (Figure 0.3). Few individual networks, however, showed such consistent change. In Network 0 (Illinois), the 5 percent rise in the number of units far outpaced the increase of 8 percent seen in the patient population. The number of patients in Network 6 rose 23 percent, but was matched, in contrast, by only a 6 percent growth in the number of units. More than three-quarters of patients are treated in freestanding, for-profit dialysis units, up from 65 percent in 996 (Figure 0.4). The distribution of for-profit units, however, varies widely by network. In Networks 6, 7, 3, and 4—comprising many of the southern states and Network 8 in California—more than 80 percent of units operate on a for-profit basis, while in Network 2, nearly 60 percent remain nonprofit. Not surprisingly, the highest concentration of dialysis units is seen in the east-
10/13/2005 8:44:51 AM
ESRD PROVIDERS
10
2005 Annual Data Report Unit & patient counts, by unit type
10.14 Patient distribution, by
18
Units per 100,000 population, 2003, by HSA
100 80 60 Transplant & dialysis center Hospital facility Hospital center Freestanding non-profit Freestanding for-profit
40 20 0 96
97
98
99
00
01
02
2.04 + (2.71) 1.60 to <2.04 1.31 to <1.60 1.09 to <1.31 below 1.09 (0.88)
03
10.16 Distribution of for-profit & nonprofit units, by ESRD network
100 80 Unknown Non-profit Profit
60 40
All
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
20
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
All
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
99 03
99 03
99 03
99 03
99 03
99 03
0
99 03
{Figures 0.9–7} data obtained from the CMS annual End-Stage Renal Disease Facility Survey, CMS Independent Renal Facility Cost Reports, & the CMS “Dialysis Facility Compare” website. {Figure 0.5} For Figure 0.5, data 2003, by HSA, unadjusted. also obtained from estimates of the United States 2003 {Figure 0.9} census, based on the 2000 census. Transient treatments, which account for less than percent of all treatments, are not included. Hemodialysis includes outpatient hemodialysis & hemodialysis training treatments; peritoneal dialysis includes outpatient IPD treatments & IPD, CAPD, & CCPD training treatments. Figure 2.39, in Chapter Two, contains a map of the ESRD networks; a list of network contacts can be found on page 230 of Appendix A.
10.15
CMS certification type
0
ern half of the country (Figure 0.5). Areas along the Gulf Coast and eastern seaboard have more than twice the number of dialysis units per 00,000 population as some areas in the western third of the nation. Eighty-two percent of dialysis units in the U.S. are now freestanding, up 4 percent from 999 (Figure 0.7). By renal network, the distribution of units parallels that seen with profit status. In Networks 6, 7, and 8, more than 90 percent of all units are freestanding; in Network 2, in contrast, nearly half of the units are hospital-based.
17
99 03
88 90 92 94 96 98 00 02
16
99 03
0
15
99 03
1,000
HB: units HB: patients 100
14
99 03
2,000
13
99 03
200
12
99 03
3,000
11
s191
300
FS: units FS: patients
8 9 10 ESRD network
99 03
4,000 Freestanding/hospital-based
7
99 03
0
6
99 03
0
50
5
99 03
Non-profit: units Non-profit: patients
4
99 03
100
1,000
3
99 03
1,500
2
99 03
150
1
99 03
2,000
All
99 03
200
2,500
0
99 03
250
Profit: units Profit: patients
99 03
3,500 Profit status
10
99 03
0
20
99 03
0
50
30
99 03
500
100
40
99 03
Number of units
Non-chain: units Non-chain: patients
1,000
Units Patients
50
99 03
150
1,500
500
Percent change, 1999-2003
200
2,000
3,000
60
250
Chain: units Chain: patients
Percent of patients
2,500
Chain status
Percent of units
3,000
patients, 1999 to 2003, by ESRD network
10.17 Distribution of freestanding & hospitalbased units, by ESRD network
100 80 Percent of units
3,500
10.13 Percent change in the number of units &
Number of patients (in thousands)
10.12
60
Hospital-based Freestanding
40 20
10 providers 05.indd 191
99 03
0
10/13/2005 8:45:05 AM
10
ESRD PROVIDERS
Patient characteristics, by unit affiliation 10.18 Characteristics of incident dialysis incident
patients, by unit affiliation, 2003 dialysis patients Age
50
100
48 Percent female
64 Mean age (in years)
Gender: Female
62
60
Race
80 Percent of patients
66
46
44
60
40
42
20
40
0
Other/unknown Asian Native American Black White
s192 58
Hispanic ethnicity Other/unknown Non-Hispanic Hispanic-other Hispanic-Mexican
Diabetic status: Diabetics
100
46
60
40
44
42
60
40
20
40
20
0
38
0
10.2 Mean hemoglobin at initiation
100 Modality
Primary diagnosis
80 Percent of patients
80 Percent of patients
48
Percent diabetic
100
Other/unknown Cystic kidney Glomerulonephritis Hypertension Diabetes
40 Percent receiving EPO at initiation
60
40
20
9.6
30
65
BMI (kg/m2)
60
55
25
29 Mean BMI at initiation
11.0 Mean eGFR at initiation
28
10.5
eGFR (ml/min/1.73 m2)
70 Percent with albumin < test's lower limit
27
26
50
45
9.8
35
Peritoneal dialysis Hemodialysis
0
Percent of patients
Percent of patients
10.0
Hemoglobin (g/dl)
Percent of patients
80
All
1
2
3
4
5
6
NC HB U
25
All
1
2
3
4
5
6 NC HB U
10.0
9.5
9.0
All
1
2
3
4
5
6
NC HB U
Unit affiliation (see table at right for codes)
10 providers 05.indd 192
10/13/2005 8:45:11 AM
ESRD PROVIDERS
10
2005 Annual Data Report
T
10.19 Unit & patient counts, incident
by unit affiliation dialysis patients
1,200
Number of units
1,000
80 1999 2003
800
Number of patients (in thousands)
60
40
600 400
20 200 0
1
2
3
4
5
0 6 NC HB 1 2 3 4 Unit affiliation (see table below for codes)
5
6
NC
HB
10.20 Characteristics of prevalent dialysis December 31 point
patients, by unit affiliation, 2003 prevalent dialysis patients
66
Age
50 48 Percent females
Mean age (in years)
64
62
60
46 44 42
58 100
Gender: Female
40
Race
44
Diabetic status: Diabetics
Percent diabetic
Percent of patients
80 60 Other/unknown Asian Native American Black White
40 20 0 100
Primary diagnosis
100
Modality
80
60 40 Other/unknown Cystic kidney Glomerulonephritis
20
All
1
2
3
4
Hypertension Diabetes
5
6 NC HB U
Percent of patients
Percent of patients
40
38
80
0
42
60 40 20 0
Peritoneal dialysis Hemodialysis All
1
2
3
4
5
6 NC HB U
Unit affiliation (see table below for codes)
All · All units Chain · Fresenius Chain 2 · Gambro Chain 3 · DaVita
10 providers 05.indd 193
Chain 4 · Renal Care Group Chain 5 · Dialysis Clinics, Inc. Chain 6 · Nat’l Nephrology Assoc. NC · Non-chain units
HB · Hospital-based units U · Unknown affiliation
he mean age of the incident dialysis population is now 62.6, and is slightly higher in units that are owned by National Nephrology Associates or are not chain-affiliated (Figure 0.8). Fortytwo percent of patients treated in hospital-based units are female, compared to 46 percent overall. Differences in the distribution of patients by race and ethnicity are not dramatic; DaVita does, however, have the greatest proportion of Hispanic patients, at 7.5 percent of its population, compared to only 4.3 percent in units owned by DCI. Patients receiving therapy in units owned by Fresenius or Gambro have a mean hemoglobin at initiation that is just slightly lower than the overall average of 0. g/dl, while for those in DaVita and RCG units it is slightly higher; these levels do not seem related to the use of EPO prior to initiation. The percent of patients who begin therapy with an albumin lower than the test’s lower limit ranges from 57.6 in units owned by RCG to 67.7 in hospitalbased facilities. Mean BMIs and estimated GFRs at initiation are similar across chains, with an overall average of 27.7 kg/m2 and 9.8 ml/min/.73 m2, respectively. Growth in most dialysis chains has been quite aggressive since 999 (Figure 0.9). Fresenius, for example, has seen an increase of 36 percent in the number of both units and patients; at Renal Care Group, growth has reached 65–70 percent. The number of non-chain dialysis units has risen 9 percent since 999; there has been a slight fall, however, in the number of hospital-based units, accompanied by a decrease of nearly 0 percent in the number of patients treated in these units. Prevalent dialysis patients tend to be slightly younger than patients starting therapy, with a mean age of 60. (Figure 0.20). Across unit affiliations, diabetics account for 40–43 percent of the population, and patient distribution by primary diagnosis varies little. As noted in Chapter Four, corporate acquisitions in late 2004 and the spring of 2005 will bring significant change to this picture of the dialysis population. The use of peritoneal dialysis, for example, is likely to decline, as Gambro and Renal Care Group—chains with some of the greatest use of this therapy—have now been acquired by chains more likely to place their patients on hemodialysis.
s193
{Figures 0.8–9} incident dialysis patients, 2003. Facility data obtained from the CMS annual EndStage Renal Disease Facility Survey, the CMS Independent Renal Facility Cost Reports, & the CMS “Dialysis Facility Compare” website. The lower limit of albumins measured by bromcresol purple is 3.2 g/ dl, & by bromcresol green is 3.5 g/dl. {Figure 0.20} December 3 point prevalent dialysis patients, 2003. Facility data obtained from the CMS annual EndStage Renal Disease Facility Survey, the CMS Independent Renal Facility Cost Reports, & the CMS “Dialysis Facility Compare” website.
10/13/2005 8:45:15 AM
10
ESRD PROVIDERS
Provider compliance with K/DOQI guidelines 10.21
Anemia treatment & dialysis period prevalent adequacy, by unit affiliation, 2003 dialysis patients
Percent of patients
100
Hemoglobin (g/dl): hemodialysis
80 60 40
23,000+ 16,000<23,000 11,000<16,000 7,000<11,000 <7,000
20 0 100
Median URR (%): hemodialysis
Mean weekly Kt/V: peritoneal dialysis 3.6+ 3.0-<3.6 2.6-<3.0 2.0-<2.6 <2.0
75+ 70-<75 65-<70 60-<65 <60
80 60
s194
EPO dose (units): hemodialysis 12+ 11-<12 10-<11 9-<10 <9
40 20 0
All 1
2
3
4
5
6 NC HB Unk All 1 2 3 4 Unit affiliation (see table at right for codes)
5
6 NC HB Unk
10.22 Percent of patients meeting target URR period prevalent hemodialysis level, by chain status & state, 2003 patients; CPM data
Chain
Non-chain
95.1 + (98.3) 91.3 to <95.1 88.3 to <91.3
10.23
83.3 to <88.3 below 83.3 (80.8) Insuff. data
83.3 to <88.3 below 83.3 (75.9) Insuff. data
Geographic variations in anemia period prevalent treatment, by chain status & state, 2003 dialysis patients
Mean hemoglobin: chain
11.93 + (12.04) 11.82 to <11.93 11.76 to <11.82
Mean hemoglobin: non-chain
11.69 to <11.76 below 11.69 (11.63)
Mean EPO dose: chain
18.2 + (19.2) 17.3 to <18.2 16.2 to <17.3
10 providers 05.indd 194
95.1 + (99.1) 91.3 to <95.1 88.3 to <91.3
11.93 + (12.09) 11.82 to <11.93 11.76 to <11.82
11.69 to <11.76 below 11.69 (11.53)
Mean EPO dose: non-chain
14.3 to <16.2 below 14.3 (13.1)
18.2 + (20.1) 17.3 to <18.2 16.2 to <17.3
14.3 to <16.2 below 14.3 (11.9)
10/13/2005 8:45:22 AM
ESRD PROVIDERS
10
2005 Annual Data Report
T
he proportion of hemodialysis patients with an average hemoglobin at or above g/dl—the target set by K/DOQI—ranges from 83 percent in units owned by Dialysis Clinics Inc. to 9 percent in those owned by DaVita (Figure 0.2). Nearly six in ten patients treated in DaVita units have a mean hemoglobin of 2 g/dl or higher. Mean weekly EPO doses of 6,000 units or more are received by 36 percent of patients treated in non-chain or hospitalbased units, compared to a high of 45 percent in DaVita units. The K/DOQI target for the median urea reduction ratio in hemodialysis patients is 65 percent or higher—a target met, in patients tracked in the CPM dataset, by 88 percent of patients overall, and by 92 percent of those treated in units owned by Renal Care Group or National Nephrology Associates. Geographic variations in the percent of patients meeting the target URR level of 65 percent, along with differences in anemia treatment by unit affiliation, are depicted in Figures 0.22–23. Analyses by state indicate that therapy targets are reached in slightly more patients dialyzing in chainaffiliated units than in non-chain facilities. Across the nation, higher hemoglobins are more widespread in patients receiving care in chain-affiliated units, and such units
located east of the Mississippi tend to give higher doses of EPO compared to their non-chain counterparts. Guidelines of both K/DOQI and the HP200 project recommend increased use of arteriovenous fistulas; the HP200 goal is to have 50 percent of new patients using a fistula as their primary access. By unit affiliation, fistula use in prevalent hemodialysis patients is relatively consistent, ranging from 28 to 35 percent (Figure 0.24). Geographic patterns showing the percent of patients with internal accesses indicate that overall use of this access is comparable between chain-affiliated and non-chain units (Figure 0.25). In regions represented by the upper-quintile, approximately 5 percent more non-chain patients have internal accesses compared to the upper-quintile patients in chain-affiliated units. Fewer than one-third of incident dialysis patients begin therapy with an albumin level greater than the test’s lower limit; the proportion ranges from 27 percent in hospital-based units to 39 percent in those owned by Renal Care Group (Figure 0.26). Serum creatinine levels are less than 0 mg/ dl in 82–89 percent of patients. And only 27–3 percent of new patients initiate treatment with a hemoglobin at or above the level 0f g/dl recommended by K/DOQI.
10.24 Vascular access use, by prevalent
10.25 Geographic variations in the percent of patients with prevalent hemodialysis
unit affiliation, 2002 hemodialysis pts
s195
Non-chain
Non-cuffed catheters AV fistula Cuffed catheter Non-cuffed catheter
80
Percent of patients
{Figure 0.2} period prevalent dialysis patients, 2003. Hemoglobin graph includes only patients treated with EPO, & the mean hemoglobin represents the average hemoglobin value for the year across all patients. EPO dose adjusted for inpatient days. URR & Kt/V data obtained from 2003 CPM data, & include only patients who are in both the USRDS & CPM databases. {Figure 0.22} period prevalent hemodialysis patients from CPM data who are also in the USRDS database. URR obtained from 2003 CPM data. {Figure 0.23} period prevalent dialysis patients, 2003. Hemoglobin maps include only patients treated with EPO, & the mean hemoglobin represents the average hemoglobin value for the year across all patients. EPO dose adjusted for inpatient days. {Figure 0.24} prevalent hemodialysis patients. Data obtained from the CDC’s National Surveillance of Dialysis-Associated Diseases in the United States. {Figure 0.25} prevalent hemodialysis patients from the 2003 CPM data who are also in the USRDS database; current access determined from CPM data. {Figure 0.26} incident dialysis patients, 2003. The lower limit of albumins measured by bromcresol purple is 3.2 g/dl, & by bromcresol green is 3.5 g/dl.
an internal access, by chain status & state, 2003 patients; CPM data
Chain
100
Chain · Fresenius Chain 2 · Gambro Chain 3 · DaVita Chain 4 · Renal Care Group Chain 5 · Dialysis Clinics, Inc. Chain 6 · Nat’l Nephrology Assoc. NC · Non-chain units HB · Hospital-based units Unk · Unknown affiliation
60 40 20 0
78.8 + (85.3) 76.6 to <78.8 73.0 to <76.6
All 1 2 3 4 5 6 NC HB Unit affiliation (see table at right for codes)
68.7 to <73.0 below 68.7 (56.9) Insuff. data
68.7 to <73.0 below 68.7 (62.0)
78.8 + (90.1) 76.6 to <78.8 73.0 to <76.6
10.26 Nutritional & hemopoietic parameters, Medical Evidence form:
by unit affiliation, 2003 incident dialysis patients
100 Albumin > test’s lower limit
Serum creatinine <10 mg/dl
Hemoglobin ≥11 g/dl
All
All
Percent of patients
80 60 40 20 0
10 providers 05.indd 195
All
1
2
3
4
5
6
NC HB Unk
1 2 3 4 5 6 NC HB Unk Unit affiliation (see table above for codes)
1
2
3
4
5
6
NC HB Unk
10/13/2005 8:45:29 AM
10
ESRD PROVIDERS
Preventive care, by unit affiliation influenza vaccinations
pneumococcal pneumonia vaccinations
10.27 Patients receiving influenza vaccinations, point prevalent
10.29 Patients receiving pneumococcal pneumonia point prevalent
by unit affiliation, 2003 ESRD patients, 2003
vaccinations, by unit affiliation, 2002–2003 ESRD patients, 2002
30
80
Percent of patients
Percent of patients
60
40
20
0
All
s196
2 3 4 5 6 NC Unit affiliation (see table at right for codes)
Patients receiving influenza point prevalent vaccinations, by chain status & year ESRD patients 80
Percent of patients
20
95
96
97
98
99
00
01
02
03
Patients receiving hepatitis B point prevalent vaccinations, by unit affiliation, 2003 ESRD patients, 2003
Percent of patients
40
30
20
10
0
All
1
2 3 4 5 6 NC Unit affiliation (see table at right for codes)
HB
10.32 Probability of receiving a hepatitis B point prevalent
vaccination, by chain status, 2003 ESRD patients, 2003
Cumulative probability
0.3
Chain
0.2
Non-chain
0.1
0.0
0
10 providers 05.indd 196
1
2
3
HB
4
5
6 Months
7
8
9
Chain
0.1 Non-chain
0.0
0
2
4
6
8
10 12 14 Months
16
18
20
22
F
hepatitis b vaccinations
10.31
2 3 4 5 6 NC Unit affiliation (see table at right for codes)
0.2
40
94
1
vaccination, by chain status, 2002–2003 ESRD pts, 2002
60
93
All
10.30 Probability of receiving a pneumococcal pneumonia point prevalent
Chain Non-chain
0
10
0
HB
Cumulative probability
10.28
1
20
10
11
ifty-seven percent of patients receive an influenza vaccination; this rises to 65 percent in units owned by DaVita, and drops to 5 percent in hospital-based units (Figure 0.27). Patients in chain-owned units are more likely than those in non-chain units to receive this vaccination, at 64 and 55 percent, respectively. There is still considerable progress to be made before the HP200 goal of a 90 percent influenza vaccination rate is met. The rate of pneumococcal pneumonia vaccinations is only 3.6 percent overall, rising to 29 percent in units owned by Renal Care Group (RCG) (Figure 0.29). At the end of two years, the probability of being vaccinated is only 0.4 for patients in chain-owned units, and 0.2 in those treated by non-chain facilities (Figure 0.30). The overall rate of hepatitis B vaccinations is 2 percent; 20 percent of patients in hospital-based units receive this vaccination, compared to 34 percent in Gambro units (Figure 0.3). After one year, the cumulative probability of being vaccinated is 0.26 and 0.22 for patients in chain-owned and non-chain units, respectively. In the diabetic ESRD population, 44 percent of patients receive two or more lipid tests within a year (Figure 0.33). Only 7 percent of patients in RCG and DCI units meet this treatment guideline, compared to 80 percent of those receiving treatment in units owned by National Nephrology Associates (NNA). Clinical practices in these chains are illustrated as well by the cumulative probability of receiving a first lipid test, which reaches 0.74 in the first month for patients treated in NNA units (Figure 0.35). The percent of patients receiving at least one lipid test has grown since the early 990s, and, since 994–995, has remained highest in units that are not chainaffiliated (Figure 0.34).
10/13/2005 8:45:32 AM
ESRD PROVIDERS
10
2005 Annual Data Report lipid monitoring in diabetic patients
glycosylated hemoglobin testing in diabetic patients
10.33 Number of lipid tests received, point prevalent
10.36 Number of HbA1c tests received, point prevalent
by unit affiliation, 2003 ESRD patients, 2002
by unit affiliation, 2003 ESRD patients, 2002
80
80
Percent of patients
100
Percent of patients
100
60 40 4+ tests 3 tests 2 tests
20 0
10.34
All
2 3 4 5 6 NC Unit affiliation (see table below for codes)
40 20
All
1
2 3 4 5 6 NC Unit affiliation (see table below for codes)
HB
s197
10.37 Patients receiving an HbA1c point prevalent test, by chain status & year ESRD patients
80
Chain Non-chain Percent of patients
60
1 test 0 tests
60
0
HB
Patients receiving a lipid point prevalent test, by chain status & year ESRD patients
70
Percent of patients
1
1 test 0 tests
4+ tests 3 tests 2 tests
50 40 30 20
Chain Non-chain
60
40
20
10 0
0
92-93 93-94 94-95 95-96 96-97 97-98 98-99 99-00 00-01 01-02 02-03
10.35 Probability of receiving a lipid point prevalent
10.38 Probability of receiving an HbA1c point prevalent
test, by unit affiliation, 2003 ESRD patients, 2002
test, by unit affiliation, 2003 ESRD patients, 2002
1.0 Fresenius Gambro DaVita RCG DCI NNA Non-chain Hosp.-based All
0.8 0.6 0.4 0.2
0
1
2
3
4
5 6 7 Months
8
9
10 11
Differences in the fulfillment of guidelines for glycosylated hemoglobin testing are not as dramatic. Forty-two percent of diabetic patients overall receive the recommended four or more tests per year; by chain, rates range from 35 percent in hospital-based units to 58 percent in those owned by DaVita (Figure 0.36). Patients in units owned by NNA are again most likely to receive testing in the first month (Figure 0.38). The percent of patients receiving an HbAc test during the year has grown steadily over the last decade. In contrast to patterns seen with lipid testing, however, HbAc testing rates have, since 2000–200, been highest for patients receiving therapy in chain-owned units (Figure 0.37). {All figures} patients with Medicare Parts A & B primary payor coverage during the entire period. {Figures 0.27–28} ESRD patients initiating therapy 90 days before September , 2003, & alive on December 3, 2003. Vaccinations tracked between September & December 3 of each year. {Figure 0.29} point prevalent
10 providers 05.indd 197
Cumulative probability
Cumulative probability
1.0
0.0
92-93 93-94 94-95 95-96 96-97 97-98 98-99 99-00 00-01 01-02 02-03
Fresenius Gambro DaVita RCG DCI NNA Non-chain Hosp.-based All
0.8 0.6 0.4 0.2 0.0
0
1
2
3
4
5 6 7 Months
8
9
10 11
patients, 2002, with 90-day rule, alive on December 3, 2003. Vaccinations tracked in 2002 & 2003. {Figure 0.30} point prevalent patients, 2002, with 90-day rule. Vaccinations tracked in 2002 & 2003. {Figure 0.3} ESRD patients initiating therapy 90 days before January , 2003, & alive on December 3. Vaccinations tracked in 2003. {Figure 0.32} point prevalent patients, 2003, with 90-day rule. Vaccinations tracked in 2003. {Figures 0.33–34 & 0.36–37} point prevalent patients initiating ESRD 90 days prior to January of the first year, age 8–75 on December 3 of the second year, & alive through the end of the second year, with diabetes as the primary cause of ESRD or a comorbidity on the Medical Evidence form, or with diabetes diagnosed during the first year. Testing tracked in the second year; tests are at least 30 days apart. {Figures 0.35 & 0.37} point prevalent Chain · Fresenius patients, 2002, with 90-day rule, age Chain 2 · Gambro 8–75 on December 3, 2003, & alive Chain 3 · DaVita through the end of 2002, with diabeChain 4 · Renal Care Group tes as identified for Figure 0.33. First testing tracked in 2003. {Figures Chain 5 · Dialysis Clinics, Inc. 0.30, 0.32, 0.35, & 0.38} patients Chain 6 · Nat’l Nephrology Assoc. censored at death, end of plan, loss to NC · Non-chain units followup, & end of 2003.
HB · Hospital-based units
10/13/2005 8:45:36 AM
10 s198
ESRD PROVIDERS
Chapter summary
Figure 0.2 Hemoglobin levels are highest in DaVita and RCG units, at 2 g/ dl or greater in 48–50 percent of units, and lowest in facilities owned by DCI or National Nephrology Associates—in 27 and 38 percent of these units, respectively, hemoglobin levels average less than g/dl. Figure 0.5 Chain units tend to give the highest amount of EPO to patients whose hemoglobins fall below g/dl. Figure 0.8 Geographic comparisons indicate that patients in for-profit units who have hemoglobins of 2 g/dl or greater are more likely to receive iron than those in non-profit units; mean percentages in the lower quintile differ by more than 6 percent.
provider-level anemia management
Figure 0.9 The total number of in-center hemodialysis treatments increased 24.3 percent between 999 and 2003— from 33 to 4 million. Growth ranged from 5.3 percent in Network 4 to nearly 33 percent in Network 9. Figure 0. In Network 6, the proportion of units that are chain-affiliated rose from 7 percent in 999 to 49 percent in 2003. Figure 0.3 From 999 to 2003, growth in the total number of both units and patients was 9.5 percent. Figure 0.4 More than three-quarters of patients are treated in freestanding, for-profit dialysis units, up from 65 percent in 996.
differences in provider growth
Figure 0.8 Patients receiving therapy in units owned by Fresenius or Gambro have a mean hemoglobin at initiation that is just slightly lower than the overall average of 0. g/dl, while for those in DaVita and RCG units it is slightly higher; these levels do not seem related to the use of EPO prior to initiation. Figure 0.9 Growth in most dialysis chains has been quite aggressive since 999. Fresenius has seen an increase of 36 percent in the number of both units and patients; at RCG, growth has reached 65–70 percent.
patient characteristics, by unit affiliation
Figure 0.2 Mean weekly EPO doses of 6,000 units or more are received by 36 percent of patients treated in non-chain or hospital-based units, compared to a high of 45 percent in DaVita units. Figure 0.24 Fistula use in prevalent hemodialysis patients is relatively consistent across unit affiliations, ranging from 28 to 35 percent. Figure 0.26 The proportion of incident dialysis patients who begin therapy with an albumin greater than the test’s lower limit ranges from 27 percent in hospital-based units to 39 percent in those owned by RCG.
provider compliance with K/DOQI guidelines
Figure 0.27 Patients in chain-owned units are more likely than those in nonchain units to receive an influenza vaccination, at 64 and 55 percent, respectively. Figure 0.33 In the diabetic ESRD population, 44 percent of patients receive two or more lipid tests within a year. Only 7 percent of patients in RCG and DCI units meet this treatment guideline, compared to 80 percent of those receiving treatment in units owned by NNA.
preventive care, by unit affiliation
maps: national means & patient popUlations
Figure number
Overall value for all pts Total patients Overall value for pts mapped Missing HSA/state: pts dropped Figure number Overall value for all pts Total patients Overall value for pts mapped Missing HSA/state: pts dropped
10 providers 05.indd 198
10.6 <11/p 29,918 2,493 29,115 42
10.6 <11/n-p 25,794 493 25,788 2
10.23 10.23 hgb/ch hgb/n-ch 11.8 11.8 141,088 68,703 11.8 11.8 138,287 67,166
10.6 12+/p 12,669 2,594 12,667 44
10.6 12+/n-p 12,073 507 12,049 3
10.8 <11/p 85.2 2,489 85.2 42
10.8 <11/n-p 83.7 492 83.7 2
10.23 10.23 EPO/ch EPO/n-ch 17,797 16,208 141,088 68,703 17,759 16,241 141,088 67,166
10.25 chain 74.6 4,936 74.5 94
10.25 non-ch 71.2 2,809 71.2 71
10.8 12+/p 83.6 2,593 83.6 44
10.8 12+/n-p 82.9 507 83.0 3
10/13/2005 8:45:37 AM
ESRD providers I love people who harness themselves, an ox to a heavy cart, who pull like water buffalo, with massive patience, who strain in the mud and the muck to move things forward, who do what has to be done, again and again.
marge piercy “to be of use”
American Journal of Kidney Diseases, Vol 47, No , Suppl (January, 2006): pp s85–s98. 10 providers 05.indd 185
11/18/2005 7:37:52 AM
10 s186
ESRD PROVIDERS
Introduction
T
The growing number of patients with ESRD has led to increasing decentralization of the dialysis delivery system—once limited to major medical centers, and now spread to suburban and more rural communities. Restrictive reimbursement systems, in which only minor increases have occurred since 982, have led to major consolidations in the ESRD provider networks. Although this report centers on data through the end of 2003, the months since December of 2004 have brought continued consolidation of the dialysis chains into two large providers, both for-profit, publicly-traded corporations. Recent changes such as this are illustrated in Figure 0., which shows the continued growth of freestanding, for-profit dialysis units in the United States, with slight declines in the number of non-profit and hospital-based providers. Characteristics of the dialysis population across this large provider system have been the focus of this chapter for many years. In this year’s first spread, we present information on provider-level anemia at initiation, treatment with erythropoietin in the prevalent population, and the variation in the erythropoietin doses based on achieved hemoglobin groups and iron therapy. At the initiation of hemodialysis, units owned by Dialysis Clinics, Inc. and National Nephrology Associates have the greatest percentage of patients with hemoglobin levels less than g/dl.
Across all hemoglobin levels, however, the highest EPO doses are given by DaVita. Data on EPO doses and hemoglobin levels by unit affiliation show the considerable variation across providers in the last six months of 2003. Clear in these data is the fact that the high EPO doses used by DaVita moved patients from hemoglobin levels less than g/dl in June, 2003, to hemoglobins of at least 2 g/dl within three months, a level which then stabilized. Geographic variations in hemoglobin levels vary by profit status, as do variations in the percent of patients receiving intravenous iron. These differences in provider practices suggest that there is less consistent application of the target guidelines for treatment, something which needs careful consideration when addressing the cost-effectiveness of these interventions. Data on provider growth show a striking decline in the use of peritoneal dialysis across ESRD networks. This is particularly true in Networks , 2, 4, 9, and 0. Only in Network 6 did the number of peritoneal dialysis treatments increase between 999 and 2003. Interestingly, non-chain affiliated units dominate Network 2, which has seen the steepest decline in peritoneal dialysis treatments. Data on the demographic characteristics of the incident population show that the distribution of patients by age and gender is similar in the top 4–5 chains. The same is true for other
s188 · provider-level patient distribution by hemoglobin · anemia EPO dosing · geographic variations · management iron therapy
s194 · provider anemia treatment & dialysis compliance with adequacy · vascular access use · K/DOQI guidelines nutritional parameters
s190 · differences dialysis treatments · unit location in provider & growth · unit & patient counts · growth affiliation & profit status s192 · patient demographics & clinical parameters characteristics, by of incident & prevalent patients · unit unit affiliation & patient counts, by unit affiliation
10 providers 05.indd 186
s196 · preventive influenza, pneumonia, & hepatitis care, by unit B vaccinations · lipid monitoring · affiliation glycosylated hemoglobin testing s198 · summary
10/13/2005 8:44:20 AM
ESRD PROVIDERS
10
2005 Annual Data Report
demographic characteristics, though hospitalbased units have the highest percentage of patients with high albumin levels. We look next at provider compliance with guidelines for care advanced by K/DOQI. DaVita, for example, has the greatest percentage of patients with an average hemoglobin of 2 g/dl or more throughout the year, while Dialysis Clinics, Inc. has the greatest percentage whose hemoglobins meet the K/DOQI target of –2 g/ dl. Once again, there is considerable geographic variation in mean hemoglobin levels of patients treated in chain versus non-chain units, as well as in associated EPO dosing patterns. Vascular access use appears to be similar across providers, with catheter use highest in hospital-based units. Use of an internal access does, however, vary across the country by chain and non-chain status. And interestingly, albumin, creatinine, and hemoglobin levels in patients beginning ESRD therapy vary little by provider. The final spread in this chapter addresses preventive care. Influenza, pneumococcal pneumonia, and hepatitis B vaccination rates have changed considerably over the years, and have now reached a degree of stability. Chains have tended to vaccinate a greater percentage of pa-
3,500
s187
Freestanding for-profit Freestanding non-profit Hospital center
3,000 Number of units
tients for influenza and for pneumonia than have non-chain units. It is interesting to note that the 2002–2003 rate of pneumococcal pneumonia vaccinations in units owned by Renal Care Group far outpaces that seen in other providers. Diabetic care also varies considerably across providers. Lipid testing in diabetic patients, for example, appears to be used least in units owned by Renal Care Group and Dialysis Clinics, Inc., while recommended testing of hemoglobin AC levels is given most often in DaVita units. The quality of care given to the dialysis population differs between providers, particularly between the chain-affiliated units and their non-chain or hospital-based counterparts. Although the Clinical Performance Measures program of the Centers for Medicare and Medicaid Services has focused primarily on dialysis delivery, anemia treatment, and vascular access, other aspects of care are clearly a concern well. The largest dialysis provider groups show considerable room for improvement. These areas and others will be explored further in subsequent Annual Data Reports to help us better assess provider performance.
2,500 2,000 1,500
Hospital facility Transplant & dialysis center Transplant center
1,000 500 0
90
92
94
96
98
00
02
0. · Counts of dialysis & transplant units, by CMS certification type data obtained from the CMS annual End-Stage Renal Disease Facility Survey, CMS Independent Renal Facility Cost Reports, & the CMS “Dialysis Facility Compare” website. The leveling out of the number of freestanding, for-profit units in 2002 is due to changes in how CMS determines profit status, resulting in some units not being classified.
Chapter highlights
Figure 10.8 Geographic comparisons indicate that patients in for-profit units who have hemoglobins of 12 g/dl or greater are more likely to receive iron than those in non-profit units; mean percentages in the lower quintile differ by more than 16 percent. Figure 10.9 The total number of in-center hemodialysis treatments increased 24.3 percent between 1999 and 2003—from 33 to 41 million. Growth ranged from 15.3 percent in Network 4 to nearly 33 percent in Network 9. Figure 10.14 More than three-quarters of patients are
treated in freestanding, for-profit dialysis units, up from 65 percent in 1996. Figure 10.27 Patients in chain-owned units are more likely than those in non-chain units to receive an influenza vaccination, at 64 and 55 percent, respectively.
10 providers 05.indd 187
10/13/2005 8:44:24 AM
Provider-level anemia management
s188
10 providers 05.indd 188
10.2
Distribution of hgb providers of period groups, by unit affiliation prevalent HD pts.
10.3
100
Percent of providers
60 40 20 0
10.4
Distribution of hgb providers of period groups, by unit type prevalent HD patients
100 12+ 11-<12 <11
80
Percent of providers
A
cross dialysis providers, hemoglobin levels average less than g/dl in 23 percent of units, –<2 g/dl in 33 percent, and 2 g/dl or higher in 44 percent (Figure 0.2). The target set by NKF’s Kidney Dialysis Outcomes Quality Initiative (K/DOQI) is a hemoglobin of g/dl or greater. Levels are highest in DaVita and Renal Care Group units, at 2 g/dl or greater in 48–50 percent of units, and lowest in facilities owned by Dialysis Clinics Incorporated or National Nephrology Associates—in 27 and 38 percent of these units, respectively, hemoglobin levels average less than g/dl. Hemoglobin levels vary less by unit type (Figure 0.3). Regardless of profit status or hospital-based/freestanding status, average hemoglobin levels are less than g/dl in 23– 24 percent of units. The percent of units in which levels are 2 g/dl or greater is slightly higher in for-profit units compared to nonprofit units—44.2 and 40.6, respectively— but is the same in freestanding and hospital-based units. EPO dosing practices appear to vary by unit affiliation, with chain-affiliated units tending to give higher doses of EPO to patients with hemoglobins less than g/dl. Chain units, for example, have a greater proportion of patients (4–4 percent) with hemoglobins less than 0 g/dl in the top quartile of EPO dose compared to nonchain units; DaVita has the largest proportion of patients in the top quartile of EPO dosing for all hemoglobin levels. Assessment of anemia treatment over a six-month period shows similar patterns (Figure 0.5). Chain units again tend to give the highest amount of EPO to patients whose hemoglobins fall below g/dl. In patients with hemoglobins above 2 g/dl, differences in dosing patterns are much less apparent. Mean hemoglobins in the less than g/dl group rise in the early months of treatment and then plateau, in contrast to those in the greater than 2 g/dl group, where slight decreases occur in the initial months followed by a stable pattern. Geographic variations in mean weekly EPO dose by profit status show only slight differences between for-profit and nonprofit units (Figure 0.6). In patients with hemoglobins of 2 g/dl or greater, for example, the mean EPO dose in the top quintile of
All 1 2 3 4 5 6 NC HB Unit affiliation (see table below for codes)
12+ 11-<12 <11
80 60 40 20 0
Profit
Nonprofit
Hospital- Freebased standing
EPO dosing, by starting period prevalent hemoglobin & unit affiliation hemodialysis patients 10-<11 g/dl
100 <10 g/dl 18,000+ units 12,000-<18,000 units 6,000-<12,000 units <6,000 units
80 60 Percent of patients
10
ESRD PROVIDERS
40 20 0
12+ g/dl
100 11-<12 g/dl 80 60 40 20 0
All
1
2
3
4
All 1 2 3 5 6 NC HB Unit affiliation (see table below for codes)
non-profit units is less than 7 percent higher than the mean dose in for-profit units. It appears that iron therapy is used in some degree by all unit types to supplement the positive effects of EPO (Figures 0.7–8). NNA tends to give the largest amounts of iron in both hemoglobin groups and, with DaVita (Figure 0.5), the highest EPO doses to patients with hemoglobins less than g/ dl. Geographic comparisons indicate that patients in for-profit units who have hemoglobins of 2 g/dl or greater are more likely to receive iron than those in non-profit units; mean percentages in the lower quintile differ by more than 6 percent.
4
5
6
NC HB
Chain · Fresenius Chain 2 · Gambro Chain 3 · DaVita Chain 4 · Renal Care Group Chain 5 · Dialysis Clinics, Inc. Chain 6 · Nat’l Nephrology Assoc. NC · Non-chain units HB · Hospital-based units
10/13/2005 8:44:27 AM
ESRD PROVIDERS
10
2005 Annual Data Report
EPo therapy
iron therapy
10.5 Anemia treatment & hemoglobin period prevalent levels, by unit affiliation hemodialysis patients
10.7
30 25 20
95 June hgb <11 g/dl
June hgb >12 g/dl Fresenius Gambro DaVita RCG DCI Nat'l Neph. Assoc. Non-chain Hospital-based
15 10
Cumulative percent of patients
EPO dose (in 1,000s of units)
35 EPO dose/week: June hgb <11 g/dl
Patients receiving iron therapy, period prevalent by hemoglobin level & unit affiliation hemodialysis patients
85
75
Fresenius Gambro DaVita RCG DCI
Nat’l Neph. Assoc. Non-chain Hospital-based
July Aug Sept Oct Nov Dec
July Aug Sept Oct Nov Dec
65
55
13 Mean hemoglobin
June hgb >12 g/dl
s189
Hemoglobin (g/dl)
10.8 Geographic variations in the percent period prevalent
of patients receiving iron, by state hemodialysis patients
12
Non-profit: Hgb <11 g/dl
11
10
10.6
Profit: Hgb <11 g/dl
July Aug Sept Oct Nov Dec
July Aug Sept Oct Nov Dec
90.9 + (95.1) 85.7 to <90.9 80.4 to <85.7 74.4 to <80.4 below 74.4 (65.2) Insuff. data
Geographic variations in provider- period prevalent level mean weekly EPO dose, by state hemodialysis patients
Profit: Hgb <11 g/dl
Non-profit: Hgb <11 g/dl Profit: Hgb 12+ g/dl
29.1+ (33.4) 25.7 to <29.1 23.5 to <25.7 17.9 to <23.5 below 17.9 (8.5) Insuff. data
Profit: Hgb 12+ g/dl
13.6 + (14.6) 12.7 to <13.6 11.6 to <12.7 9.7 to <11.6 below 9.7 (9.1) Insuff. data
10 providers 05.indd 189
29.1 + (32.9) 25.7 to <29.1 23.5 to <25.7 17.9 to <23.5 below 17.9 (14.4) Insuff. data
Non-profit: Hgb 12+g/dl
13.6 + (15.6) 12.7 to <13.6 11.6 to <12.7 9.7 to <11.6 below 9.7 (7.4) Insuff. data
90.9 + (96.9) 85.7 to <90.9 80.4 to <85.7 74.4 to <80.4 below 74.4 (63.0) Insuff. data
Non-profit: Hgb 12+g/dl
90.9 + (94.0) 85.7 to <90.9 80.4 to <85.7 74.4 to <80.4 below 74.4 (68.7) Insuff. data
90.9 + (94.6) 85.7 to <90.9 80.4 to <85.7 74.4 to <80.4 below 74.4 (52.6) Insuff. data
{Figures 0.2–3} providers of period prevalent hemodialysis patients who have at least three months of dialysis prior to June, 2003, a valid EPO claim in June, 2003, & an identifiable provider. {Figure 0.4} period prevalent hemodialysis patients who have at least three months of dialysis prior to June 2003, a valid EPO claim in June & July of 2003, & an identifiable provider. Hemoglobin groups are determined from June claims, & EPO doses from July claims. {Figures 0.5–8} period prevalent hemodialysis patients who have at least three months of dialysis prior to June 2003, a valid EPO claim in June 2003, & an identifiable provider; who dialyze at the same provider from June–December 2003 & have a valid EPO claim in each of those months; & who are from a provider with at least ten such patients.
10/13/2005 8:44:35 AM
s190
Differences in provider growth 10.9 In-center treatments (in millions)
T
he total number of in-center hemodialysis treatments increased 24.3 percent between 999 and 2003— from 33 to 4 million (Figure 0.9). Growth ranged from 5.3 percent in Network 4 (Delaware and Pennsylvania) to nearly 33 percent in Network 9 (Indiana, Kentucky, and Ohio). The number of peritoneal dialysis treatments varies widely from year to year; 22,378 patients were reported to be on this therapy in 2003. Provider growth between 999 and 2003 was higher for units with chain affiliation than for non-chain units (Figure 0.0). During this period, 39 units were dropped from chain status and 758 units were added, a net gain of 69 units. Conversely, 4 units were dropped from non-chain status and 526 were added, a net gain of 5, demonstrating a continued trend towards private ownership within renal providers. In Network 6—Alaska, Idaho, Montana, Oregon, and Washington—the proportion of units that are chain-affiliated rose from 7 percent in 999 to 49 percent in 2003 (Figure 0.), the most dramatic change seen among the networks. Only in Networks 2, 3, , 2, and 6 do non-chain units account for more than half of those providing dialysis; in Network 2 (New
Dialysis treatments, by ESRD network & modality 4.0 Hemodialysis 1999 3.5 2003 3.0 2.5 2.0 1.5 1.0 0.5 0.0 45 Peritoneal dialysis 40
Treatments (in thousands)
10
ESRD PROVIDERS
35 30 25 20 15 10 5 0
1
2
3
4
5
6
7
8
Chain-affiliated & non-chain units
10.10 Unit growth between 1999 & 2003
Chain units
Non-chain units
Chain units dropped Chain units added Chain units unchanged
Non-chain units dropped Non-chain units added Non-chain units unchanged
10.11 Unit distribution, by ESRD network
100
Non-chain Chain-affiliated
60
All
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
40
99 03
Percent of units
80
20
0
10 providers 05.indd 190
9 10 11 ESRD network
12
13
14
15
16
17
18
York), more than 80 percent of units continue to be independently owned. Since the late 980s, ownership of dialysis units has changed dramatically (Figure 0.2). Counts of units that are nonchain, non-profit, or hospital-based have remained relatively stable or declined, as has the number of patients treated in these units. The number of chain-owned units, in contrast, has grown over eleven-fold since 988, and 3.5 times as many units are now run on a for-profit basis. From 999 to 2003, growth in the total number of both units and patients was 9.5 percent (Figure 0.3). Few individual networks, however, showed such consistent change. In Network 0 (Illinois), the 5 percent rise in the number of units far outpaced the increase of 8 percent seen in the patient population. The number of patients in Network 6 rose 23 percent, but was matched, in contrast, by only a 6 percent growth in the number of units. More than three-quarters of patients are treated in freestanding, for-profit dialysis units, up from 65 percent in 996 (Figure 0.4). The distribution of for-profit units, however, varies widely by network. In Networks 6, 7, 3, and 4—comprising many of the southern states and Network 8 in California—more than 80 percent of units operate on a for-profit basis, while in Network 2, nearly 60 percent remain nonprofit. Not surprisingly, the highest concentration of dialysis units is seen in the east-
10/13/2005 8:44:51 AM
ESRD PROVIDERS
10
2005 Annual Data Report Unit & patient counts, by unit type
10.14 Patient distribution, by
18
Units per 100,000 population, 2003, by HSA
100 80 60 Transplant & dialysis center Hospital facility Hospital center Freestanding non-profit Freestanding for-profit
40 20 0 96
97
98
99
00
01
02
2.04 + (2.71) 1.60 to <2.04 1.31 to <1.60 1.09 to <1.31 below 1.09 (0.88)
03
10.16 Distribution of for-profit & nonprofit units, by ESRD network
100 80 Unknown Non-profit Profit
60 40
All
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
20
99 03
99 03
99 03
99 03
99 03
99 03
99 03
99 03
All
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
99 03
99 03
99 03
99 03
99 03
99 03
0
99 03
{Figures 0.9–7} data obtained from the CMS annual End-Stage Renal Disease Facility Survey, CMS Independent Renal Facility Cost Reports, & the CMS “Dialysis Facility Compare” website. {Figure 0.5} For Figure 0.5, data 2003, by HSA, unadjusted. also obtained from estimates of the United States 2003 {Figure 0.9} census, based on the 2000 census. Transient treatments, which account for less than percent of all treatments, are not included. Hemodialysis includes outpatient hemodialysis & hemodialysis training treatments; peritoneal dialysis includes outpatient IPD treatments & IPD, CAPD, & CCPD training treatments. Figure 2.39, in Chapter Two, contains a map of the ESRD networks; a list of network contacts can be found on page 230 of Appendix A.
10.15
CMS certification type
0
ern half of the country (Figure 0.5). Areas along the Gulf Coast and eastern seaboard have more than twice the number of dialysis units per 00,000 population as some areas in the western third of the nation. Eighty-two percent of dialysis units in the U.S. are now freestanding, up 4 percent from 999 (Figure 0.7). By renal network, the distribution of units parallels that seen with profit status. In Networks 6, 7, and 8, more than 90 percent of all units are freestanding; in Network 2, in contrast, nearly half of the units are hospital-based.
17
99 03
88 90 92 94 96 98 00 02
16
99 03
0
15
99 03
1,000
HB: units HB: patients 100
14
99 03
2,000
13
99 03
200
12
99 03
3,000
11
s191
300
FS: units FS: patients
8 9 10 ESRD network
99 03
4,000 Freestanding/hospital-based
7
99 03
0
6
99 03
0
50
5
99 03
Non-profit: units Non-profit: patients
4
99 03
100
1,000
3
99 03
1,500
2
99 03
150
1
99 03
2,000
All
99 03
200
2,500
0
99 03
250
Profit: units Profit: patients
99 03
3,500 Profit status
10
99 03
0
20
99 03
0
50
30
99 03
500
100
40
99 03
Number of units
Non-chain: units Non-chain: patients
1,000
Units Patients
50
99 03
150
1,500
500
Percent change, 1999-2003
200
2,000
3,000
60
250
Chain: units Chain: patients
Percent of patients
2,500
Chain status
Percent of units
3,000
patients, 1999 to 2003, by ESRD network
10.17 Distribution of freestanding & hospitalbased units, by ESRD network
100 80 Percent of units
3,500
10.13 Percent change in the number of units &
Number of patients (in thousands)
10.12
60
Hospital-based Freestanding
40 20
10 providers 05.indd 191
99 03
0
10/13/2005 8:45:05 AM
10
ESRD PROVIDERS
Patient characteristics, by unit affiliation 10.18 Characteristics of incident dialysis incident
patients, by unit affiliation, 2003 dialysis patients Age
50
100
48 Percent female
64 Mean age (in years)
Gender: Female
62
60
Race
80 Percent of patients
66
46
44
60
40
42
20
40
0
Other/unknown Asian Native American Black White
s192 58
Hispanic ethnicity Other/unknown Non-Hispanic Hispanic-other Hispanic-Mexican
Diabetic status: Diabetics
100
46
60
40
44
42
60
40
20
40
20
0
38
0
10.2 Mean hemoglobin at initiation
100 Modality
Primary diagnosis
80 Percent of patients
80 Percent of patients
48
Percent diabetic
100
Other/unknown Cystic kidney Glomerulonephritis Hypertension Diabetes
40 Percent receiving EPO at initiation
60
40
20
9.6
30
65
BMI (kg/m2)
60
55
25
29 Mean BMI at initiation
11.0 Mean eGFR at initiation
28
10.5
eGFR (ml/min/1.73 m2)
70 Percent with albumin < test's lower limit
27
26
50
45
9.8
35
Peritoneal dialysis Hemodialysis
0
Percent of patients
Percent of patients
10.0
Hemoglobin (g/dl)
Percent of patients
80
All
1
2
3
4
5
6
NC HB U
25
All
1
2
3
4
5
6 NC HB U
10.0
9.5
9.0
All
1
2
3
4
5
6
NC HB U
Unit affiliation (see table at right for codes)
10 providers 05.indd 192
10/13/2005 8:45:11 AM
ESRD PROVIDERS
10
2005 Annual Data Report
T
10.19 Unit & patient counts, incident
by unit affiliation dialysis patients
1,200
Number of units
1,000
80 1999 2003
800
Number of patients (in thousands)
60
40
600 400
20 200 0
1
2
3
4
5
0 6 NC HB 1 2 3 4 Unit affiliation (see table below for codes)
5
6
NC
HB
10.20 Characteristics of prevalent dialysis December 31 point
patients, by unit affiliation, 2003 prevalent dialysis patients
66
Age
50 48 Percent females
Mean age (in years)
64
62
60
46 44 42
58 100
Gender: Female
40
Race
44
Diabetic status: Diabetics
Percent diabetic
Percent of patients
80 60 Other/unknown Asian Native American Black White
40 20 0 100
Primary diagnosis
100
Modality
80
60 40 Other/unknown Cystic kidney Glomerulonephritis
20
All
1
2
3
4
Hypertension Diabetes
5
6 NC HB U
Percent of patients
Percent of patients
40
38
80
0
42
60 40 20 0
Peritoneal dialysis Hemodialysis All
1
2
3
4
5
6 NC HB U
Unit affiliation (see table below for codes)
All · All units Chain · Fresenius Chain 2 · Gambro Chain 3 · DaVita
10 providers 05.indd 193
Chain 4 · Renal Care Group Chain 5 · Dialysis Clinics, Inc. Chain 6 · Nat’l Nephrology Assoc. NC · Non-chain units
HB · Hospital-based units U · Unknown affiliation
he mean age of the incident dialysis population is now 62.6, and is slightly higher in units that are owned by National Nephrology Associates or are not chain-affiliated (Figure 0.8). Fortytwo percent of patients treated in hospital-based units are female, compared to 46 percent overall. Differences in the distribution of patients by race and ethnicity are not dramatic; DaVita does, however, have the greatest proportion of Hispanic patients, at 7.5 percent of its population, compared to only 4.3 percent in units owned by DCI. Patients receiving therapy in units owned by Fresenius or Gambro have a mean hemoglobin at initiation that is just slightly lower than the overall average of 0. g/dl, while for those in DaVita and RCG units it is slightly higher; these levels do not seem related to the use of EPO prior to initiation. The percent of patients who begin therapy with an albumin lower than the test’s lower limit ranges from 57.6 in units owned by RCG to 67.7 in hospitalbased facilities. Mean BMIs and estimated GFRs at initiation are similar across chains, with an overall average of 27.7 kg/m2 and 9.8 ml/min/.73 m2, respectively. Growth in most dialysis chains has been quite aggressive since 999 (Figure 0.9). Fresenius, for example, has seen an increase of 36 percent in the number of both units and patients; at Renal Care Group, growth has reached 65–70 percent. The number of non-chain dialysis units has risen 9 percent since 999; there has been a slight fall, however, in the number of hospital-based units, accompanied by a decrease of nearly 0 percent in the number of patients treated in these units. Prevalent dialysis patients tend to be slightly younger than patients starting therapy, with a mean age of 60. (Figure 0.20). Across unit affiliations, diabetics account for 40–43 percent of the population, and patient distribution by primary diagnosis varies little. As noted in Chapter Four, corporate acquisitions in late 2004 and the spring of 2005 will bring significant change to this picture of the dialysis population. The use of peritoneal dialysis, for example, is likely to decline, as Gambro and Renal Care Group—chains with some of the greatest use of this therapy—have now been acquired by chains more likely to place their patients on hemodialysis.
s193
{Figures 0.8–9} incident dialysis patients, 2003. Facility data obtained from the CMS annual EndStage Renal Disease Facility Survey, the CMS Independent Renal Facility Cost Reports, & the CMS “Dialysis Facility Compare” website. The lower limit of albumins measured by bromcresol purple is 3.2 g/ dl, & by bromcresol green is 3.5 g/dl. {Figure 0.20} December 3 point prevalent dialysis patients, 2003. Facility data obtained from the CMS annual EndStage Renal Disease Facility Survey, the CMS Independent Renal Facility Cost Reports, & the CMS “Dialysis Facility Compare” website.
10/13/2005 8:45:15 AM
10
ESRD PROVIDERS
Provider compliance with K/DOQI guidelines 10.21
Anemia treatment & dialysis period prevalent adequacy, by unit affiliation, 2003 dialysis patients
Percent of patients
100
Hemoglobin (g/dl): hemodialysis
80 60 40
23,000+ 16,000<23,000 11,000<16,000 7,000<11,000 <7,000
20 0 100
Median URR (%): hemodialysis
Mean weekly Kt/V: peritoneal dialysis 3.6+ 3.0-<3.6 2.6-<3.0 2.0-<2.6 <2.0
75+ 70-<75 65-<70 60-<65 <60
80 60
s194
EPO dose (units): hemodialysis 12+ 11-<12 10-<11 9-<10 <9
40 20 0
All 1
2
3
4
5
6 NC HB Unk All 1 2 3 4 Unit affiliation (see table at right for codes)
5
6 NC HB Unk
10.22 Percent of patients meeting target URR period prevalent hemodialysis level, by chain status & state, 2003 patients; CPM data
Chain
Non-chain
95.1 + (98.3) 91.3 to <95.1 88.3 to <91.3
10.23
83.3 to <88.3 below 83.3 (80.8) Insuff. data
83.3 to <88.3 below 83.3 (75.9) Insuff. data
Geographic variations in anemia period prevalent treatment, by chain status & state, 2003 dialysis patients
Mean hemoglobin: chain
11.93 + (12.04) 11.82 to <11.93 11.76 to <11.82
Mean hemoglobin: non-chain
11.69 to <11.76 below 11.69 (11.63)
Mean EPO dose: chain
18.2 + (19.2) 17.3 to <18.2 16.2 to <17.3
10 providers 05.indd 194
95.1 + (99.1) 91.3 to <95.1 88.3 to <91.3
11.93 + (12.09) 11.82 to <11.93 11.76 to <11.82
11.69 to <11.76 below 11.69 (11.53)
Mean EPO dose: non-chain
14.3 to <16.2 below 14.3 (13.1)
18.2 + (20.1) 17.3 to <18.2 16.2 to <17.3
14.3 to <16.2 below 14.3 (11.9)
10/13/2005 8:45:22 AM
ESRD PROVIDERS
10
2005 Annual Data Report
T
he proportion of hemodialysis patients with an average hemoglobin at or above g/dl—the target set by K/DOQI—ranges from 83 percent in units owned by Dialysis Clinics Inc. to 9 percent in those owned by DaVita (Figure 0.2). Nearly six in ten patients treated in DaVita units have a mean hemoglobin of 2 g/dl or higher. Mean weekly EPO doses of 6,000 units or more are received by 36 percent of patients treated in non-chain or hospitalbased units, compared to a high of 45 percent in DaVita units. The K/DOQI target for the median urea reduction ratio in hemodialysis patients is 65 percent or higher—a target met, in patients tracked in the CPM dataset, by 88 percent of patients overall, and by 92 percent of those treated in units owned by Renal Care Group or National Nephrology Associates. Geographic variations in the percent of patients meeting the target URR level of 65 percent, along with differences in anemia treatment by unit affiliation, are depicted in Figures 0.22–23. Analyses by state indicate that therapy targets are reached in slightly more patients dialyzing in chainaffiliated units than in non-chain facilities. Across the nation, higher hemoglobins are more widespread in patients receiving care in chain-affiliated units, and such units
located east of the Mississippi tend to give higher doses of EPO compared to their non-chain counterparts. Guidelines of both K/DOQI and the HP200 project recommend increased use of arteriovenous fistulas; the HP200 goal is to have 50 percent of new patients using a fistula as their primary access. By unit affiliation, fistula use in prevalent hemodialysis patients is relatively consistent, ranging from 28 to 35 percent (Figure 0.24). Geographic patterns showing the percent of patients with internal accesses indicate that overall use of this access is comparable between chain-affiliated and non-chain units (Figure 0.25). In regions represented by the upper-quintile, approximately 5 percent more non-chain patients have internal accesses compared to the upper-quintile patients in chain-affiliated units. Fewer than one-third of incident dialysis patients begin therapy with an albumin level greater than the test’s lower limit; the proportion ranges from 27 percent in hospital-based units to 39 percent in those owned by Renal Care Group (Figure 0.26). Serum creatinine levels are less than 0 mg/ dl in 82–89 percent of patients. And only 27–3 percent of new patients initiate treatment with a hemoglobin at or above the level 0f g/dl recommended by K/DOQI.
10.24 Vascular access use, by prevalent
10.25 Geographic variations in the percent of patients with prevalent hemodialysis
unit affiliation, 2002 hemodialysis pts
s195
Non-chain
Non-cuffed catheters AV fistula Cuffed catheter Non-cuffed catheter
80
Percent of patients
{Figure 0.2} period prevalent dialysis patients, 2003. Hemoglobin graph includes only patients treated with EPO, & the mean hemoglobin represents the average hemoglobin value for the year across all patients. EPO dose adjusted for inpatient days. URR & Kt/V data obtained from 2003 CPM data, & include only patients who are in both the USRDS & CPM databases. {Figure 0.22} period prevalent hemodialysis patients from CPM data who are also in the USRDS database. URR obtained from 2003 CPM data. {Figure 0.23} period prevalent dialysis patients, 2003. Hemoglobin maps include only patients treated with EPO, & the mean hemoglobin represents the average hemoglobin value for the year across all patients. EPO dose adjusted for inpatient days. {Figure 0.24} prevalent hemodialysis patients. Data obtained from the CDC’s National Surveillance of Dialysis-Associated Diseases in the United States. {Figure 0.25} prevalent hemodialysis patients from the 2003 CPM data who are also in the USRDS database; current access determined from CPM data. {Figure 0.26} incident dialysis patients, 2003. The lower limit of albumins measured by bromcresol purple is 3.2 g/dl, & by bromcresol green is 3.5 g/dl.
an internal access, by chain status & state, 2003 patients; CPM data
Chain
100
Chain · Fresenius Chain 2 · Gambro Chain 3 · DaVita Chain 4 · Renal Care Group Chain 5 · Dialysis Clinics, Inc. Chain 6 · Nat’l Nephrology Assoc. NC · Non-chain units HB · Hospital-based units Unk · Unknown affiliation
60 40 20 0
78.8 + (85.3) 76.6 to <78.8 73.0 to <76.6
All 1 2 3 4 5 6 NC HB Unit affiliation (see table at right for codes)
68.7 to <73.0 below 68.7 (56.9) Insuff. data
68.7 to <73.0 below 68.7 (62.0)
78.8 + (90.1) 76.6 to <78.8 73.0 to <76.6
10.26 Nutritional & hemopoietic parameters, Medical Evidence form:
by unit affiliation, 2003 incident dialysis patients
100 Albumin > test’s lower limit
Serum creatinine <10 mg/dl
Hemoglobin ≥11 g/dl
All
All
Percent of patients
80 60 40 20 0
10 providers 05.indd 195
All
1
2
3
4
5
6
NC HB Unk
1 2 3 4 5 6 NC HB Unk Unit affiliation (see table above for codes)
1
2
3
4
5
6
NC HB Unk
10/13/2005 8:45:29 AM
10
ESRD PROVIDERS
Preventive care, by unit affiliation influenza vaccinations
pneumococcal pneumonia vaccinations
10.27 Patients receiving influenza vaccinations, point prevalent
10.29 Patients receiving pneumococcal pneumonia point prevalent
by unit affiliation, 2003 ESRD patients, 2003
vaccinations, by unit affiliation, 2002–2003 ESRD patients, 2002
30
80
Percent of patients
Percent of patients
60
40
20
0
All
s196
2 3 4 5 6 NC Unit affiliation (see table at right for codes)
Patients receiving influenza point prevalent vaccinations, by chain status & year ESRD patients 80
Percent of patients
20
95
96
97
98
99
00
01
02
03
Patients receiving hepatitis B point prevalent vaccinations, by unit affiliation, 2003 ESRD patients, 2003
Percent of patients
40
30
20
10
0
All
1
2 3 4 5 6 NC Unit affiliation (see table at right for codes)
HB
10.32 Probability of receiving a hepatitis B point prevalent
vaccination, by chain status, 2003 ESRD patients, 2003
Cumulative probability
0.3
Chain
0.2
Non-chain
0.1
0.0
0
10 providers 05.indd 196
1
2
3
HB
4
5
6 Months
7
8
9
Chain
0.1 Non-chain
0.0
0
2
4
6
8
10 12 14 Months
16
18
20
22
F
hepatitis b vaccinations
10.31
2 3 4 5 6 NC Unit affiliation (see table at right for codes)
0.2
40
94
1
vaccination, by chain status, 2002–2003 ESRD pts, 2002
60
93
All
10.30 Probability of receiving a pneumococcal pneumonia point prevalent
Chain Non-chain
0
10
0
HB
Cumulative probability
10.28
1
20
10
11
ifty-seven percent of patients receive an influenza vaccination; this rises to 65 percent in units owned by DaVita, and drops to 5 percent in hospital-based units (Figure 0.27). Patients in chain-owned units are more likely than those in non-chain units to receive this vaccination, at 64 and 55 percent, respectively. There is still considerable progress to be made before the HP200 goal of a 90 percent influenza vaccination rate is met. The rate of pneumococcal pneumonia vaccinations is only 3.6 percent overall, rising to 29 percent in units owned by Renal Care Group (RCG) (Figure 0.29). At the end of two years, the probability of being vaccinated is only 0.4 for patients in chain-owned units, and 0.2 in those treated by non-chain facilities (Figure 0.30). The overall rate of hepatitis B vaccinations is 2 percent; 20 percent of patients in hospital-based units receive this vaccination, compared to 34 percent in Gambro units (Figure 0.3). After one year, the cumulative probability of being vaccinated is 0.26 and 0.22 for patients in chain-owned and non-chain units, respectively. In the diabetic ESRD population, 44 percent of patients receive two or more lipid tests within a year (Figure 0.33). Only 7 percent of patients in RCG and DCI units meet this treatment guideline, compared to 80 percent of those receiving treatment in units owned by National Nephrology Associates (NNA). Clinical practices in these chains are illustrated as well by the cumulative probability of receiving a first lipid test, which reaches 0.74 in the first month for patients treated in NNA units (Figure 0.35). The percent of patients receiving at least one lipid test has grown since the early 990s, and, since 994–995, has remained highest in units that are not chainaffiliated (Figure 0.34).
10/13/2005 8:45:32 AM
ESRD PROVIDERS
10
2005 Annual Data Report lipid monitoring in diabetic patients
glycosylated hemoglobin testing in diabetic patients
10.33 Number of lipid tests received, point prevalent
10.36 Number of HbA1c tests received, point prevalent
by unit affiliation, 2003 ESRD patients, 2002
by unit affiliation, 2003 ESRD patients, 2002
80
80
Percent of patients
100
Percent of patients
100
60 40 4+ tests 3 tests 2 tests
20 0
10.34
All
2 3 4 5 6 NC Unit affiliation (see table below for codes)
40 20
All
1
2 3 4 5 6 NC Unit affiliation (see table below for codes)
HB
s197
10.37 Patients receiving an HbA1c point prevalent test, by chain status & year ESRD patients
80
Chain Non-chain Percent of patients
60
1 test 0 tests
60
0
HB
Patients receiving a lipid point prevalent test, by chain status & year ESRD patients
70
Percent of patients
1
1 test 0 tests
4+ tests 3 tests 2 tests
50 40 30 20
Chain Non-chain
60
40
20
10 0
0
92-93 93-94 94-95 95-96 96-97 97-98 98-99 99-00 00-01 01-02 02-03
10.35 Probability of receiving a lipid point prevalent
10.38 Probability of receiving an HbA1c point prevalent
test, by unit affiliation, 2003 ESRD patients, 2002
test, by unit affiliation, 2003 ESRD patients, 2002
1.0 Fresenius Gambro DaVita RCG DCI NNA Non-chain Hosp.-based All
0.8 0.6 0.4 0.2
0
1
2
3
4
5 6 7 Months
8
9
10 11
Differences in the fulfillment of guidelines for glycosylated hemoglobin testing are not as dramatic. Forty-two percent of diabetic patients overall receive the recommended four or more tests per year; by chain, rates range from 35 percent in hospital-based units to 58 percent in those owned by DaVita (Figure 0.36). Patients in units owned by NNA are again most likely to receive testing in the first month (Figure 0.38). The percent of patients receiving an HbAc test during the year has grown steadily over the last decade. In contrast to patterns seen with lipid testing, however, HbAc testing rates have, since 2000–200, been highest for patients receiving therapy in chain-owned units (Figure 0.37). {All figures} patients with Medicare Parts A & B primary payor coverage during the entire period. {Figures 0.27–28} ESRD patients initiating therapy 90 days before September , 2003, & alive on December 3, 2003. Vaccinations tracked between September & December 3 of each year. {Figure 0.29} point prevalent
10 providers 05.indd 197
Cumulative probability
Cumulative probability
1.0
0.0
92-93 93-94 94-95 95-96 96-97 97-98 98-99 99-00 00-01 01-02 02-03
Fresenius Gambro DaVita RCG DCI NNA Non-chain Hosp.-based All
0.8 0.6 0.4 0.2 0.0
0
1
2
3
4
5 6 7 Months
8
9
10 11
patients, 2002, with 90-day rule, alive on December 3, 2003. Vaccinations tracked in 2002 & 2003. {Figure 0.30} point prevalent patients, 2002, with 90-day rule. Vaccinations tracked in 2002 & 2003. {Figure 0.3} ESRD patients initiating therapy 90 days before January , 2003, & alive on December 3. Vaccinations tracked in 2003. {Figure 0.32} point prevalent patients, 2003, with 90-day rule. Vaccinations tracked in 2003. {Figures 0.33–34 & 0.36–37} point prevalent patients initiating ESRD 90 days prior to January of the first year, age 8–75 on December 3 of the second year, & alive through the end of the second year, with diabetes as the primary cause of ESRD or a comorbidity on the Medical Evidence form, or with diabetes diagnosed during the first year. Testing tracked in the second year; tests are at least 30 days apart. {Figures 0.35 & 0.37} point prevalent Chain · Fresenius patients, 2002, with 90-day rule, age Chain 2 · Gambro 8–75 on December 3, 2003, & alive Chain 3 · DaVita through the end of 2002, with diabeChain 4 · Renal Care Group tes as identified for Figure 0.33. First testing tracked in 2003. {Figures Chain 5 · Dialysis Clinics, Inc. 0.30, 0.32, 0.35, & 0.38} patients Chain 6 · Nat’l Nephrology Assoc. censored at death, end of plan, loss to NC · Non-chain units followup, & end of 2003.
HB · Hospital-based units
10/13/2005 8:45:36 AM
10 s198
ESRD PROVIDERS
Chapter summary
Figure 0.2 Hemoglobin levels are highest in DaVita and RCG units, at 2 g/ dl or greater in 48–50 percent of units, and lowest in facilities owned by DCI or National Nephrology Associates—in 27 and 38 percent of these units, respectively, hemoglobin levels average less than g/dl. Figure 0.5 Chain units tend to give the highest amount of EPO to patients whose hemoglobins fall below g/dl. Figure 0.8 Geographic comparisons indicate that patients in for-profit units who have hemoglobins of 2 g/dl or greater are more likely to receive iron than those in non-profit units; mean percentages in the lower quintile differ by more than 6 percent.
provider-level anemia management
Figure 0.9 The total number of in-center hemodialysis treatments increased 24.3 percent between 999 and 2003— from 33 to 4 million. Growth ranged from 5.3 percent in Network 4 to nearly 33 percent in Network 9. Figure 0. In Network 6, the proportion of units that are chain-affiliated rose from 7 percent in 999 to 49 percent in 2003. Figure 0.3 From 999 to 2003, growth in the total number of both units and patients was 9.5 percent. Figure 0.4 More than three-quarters of patients are treated in freestanding, for-profit dialysis units, up from 65 percent in 996.
differences in provider growth
Figure 0.8 Patients receiving therapy in units owned by Fresenius or Gambro have a mean hemoglobin at initiation that is just slightly lower than the overall average of 0. g/dl, while for those in DaVita and RCG units it is slightly higher; these levels do not seem related to the use of EPO prior to initiation. Figure 0.9 Growth in most dialysis chains has been quite aggressive since 999. Fresenius has seen an increase of 36 percent in the number of both units and patients; at RCG, growth has reached 65–70 percent.
patient characteristics, by unit affiliation
Figure 0.2 Mean weekly EPO doses of 6,000 units or more are received by 36 percent of patients treated in non-chain or hospital-based units, compared to a high of 45 percent in DaVita units. Figure 0.24 Fistula use in prevalent hemodialysis patients is relatively consistent across unit affiliations, ranging from 28 to 35 percent. Figure 0.26 The proportion of incident dialysis patients who begin therapy with an albumin greater than the test’s lower limit ranges from 27 percent in hospital-based units to 39 percent in those owned by RCG.
provider compliance with K/DOQI guidelines
Figure 0.27 Patients in chain-owned units are more likely than those in nonchain units to receive an influenza vaccination, at 64 and 55 percent, respectively. Figure 0.33 In the diabetic ESRD population, 44 percent of patients receive two or more lipid tests within a year. Only 7 percent of patients in RCG and DCI units meet this treatment guideline, compared to 80 percent of those receiving treatment in units owned by NNA.
preventive care, by unit affiliation
maps: national means & patient popUlations
Figure number
Overall value for all pts Total patients Overall value for pts mapped Missing HSA/state: pts dropped Figure number Overall value for all pts Total patients Overall value for pts mapped Missing HSA/state: pts dropped
10 providers 05.indd 198
10.6 <11/p 29,918 2,493 29,115 42
10.6 <11/n-p 25,794 493 25,788 2
10.23 10.23 hgb/ch hgb/n-ch 11.8 11.8 141,088 68,703 11.8 11.8 138,287 67,166
10.6 12+/p 12,669 2,594 12,667 44
10.6 12+/n-p 12,073 507 12,049 3
10.8 <11/p 85.2 2,489 85.2 42
10.8 <11/n-p 83.7 492 83.7 2
10.23 10.23 EPO/ch EPO/n-ch 17,797 16,208 141,088 68,703 17,759 16,241 141,088 67,166
10.25 chain 74.6 4,936 74.5 94
10.25 non-ch 71.2 2,809 71.2 71
10.8 12+/p 83.6 2,593 83.6 44
10.8 12+/n-p 82.9 507 83.0 3
10/13/2005 8:45:37 AM