INTRODUCTION

While many aspects of renal replacement therapy for stage 5 chronic kidney disease (CKD) have been standardized in internationally accepted guidelines such as the Kidney Disease Outcomes Quality Initiative (K/DOQI),1 significant differences are seen in the literature in the degree of compliance with the proposed standards both between centers and between the different countries.2-4

One of the current challenges is to reduce the variability in the care provided. Though some differences could be justified because of the different demographic characteristics of patients dialyzed at the centers, there are other differences due to the different procedures used.5-8

Systematic, planned measurement of quality indicators has been shown to help improve patient monitoring and outcomes of treatments applied, because it allows us for understanding our situation, introducing improvement activities, and systematically and continuously checking their effectiveness. 9 The true rationale for attempting to achieve compliance with such indicators is the recently shown evidence that joint achievement of some of them (anemia, dialysis dose, calcium-phosphorus metabolism, albumin, or type of vascular access) has an impact on patient survival, morbidity, and costs.10,11

Other improvement tools include feedback and benchmarking.12-14 Feedback consists of pooling individual results to obtain joint results providing a populational perspective, systematically and regularly reporting such results to the staff in charge of care. Benchmarking is based on the assumption that open knowledge of our own results and their comparison with those achieved by other similar centers represents a powerful stimulus able to modify behaviors and hence results.

The main objective of this study was to assess the impact of systematic measurement of quality indicators and implementation of feedback and benchmarking procedures on the results achieved in three hemodialysis units, and to provide readers with a new tool for improving results in daily work.

MATERIALS AND METHODS

Definitions of interest

Indicator: A measure of performance of an organization that is used to assess the efficiency, efficacy, and quality of a given action.

Quality objective: A formulation of what is intended to be achieved in the field of quality. Quality objectives should be formulated in a simple way (to make it easier to understand), and should also be realistic (achievable), attractive (to be a source of motivation), and measurable (so that each person may verify at any time how he/she is performing). Deadlines for meeting such objectives should also be established.

Study design
This was an uncontrolled, pre-post, or intervention study assessing response after definition of an objective and implementation of feedback and benchmarking procedures in the same group of subjects. Evaluations were made before and after intervention, with each subject acting as his/her own control.

Methods
The results obtained during the 2006-2007 period at the three hemodialysis units belonging to Hospital Perpetuo Socorro were analyzed. These are units working for the social security under an agreement, subject to the same quality management system and using common protocols.

A total of 317 and 311 patients underwent dialysis in all three units during 2006 and 2007 respectively. Results of quality indicators refer to these patients (207 patients in unit A, 53 patients in unit B, and 51 patients in unit C).
Eighty-three incident patients (51 patients in unit A, 13 patients in unit B, and 19 patients in unit C) were received
during 2006, and 60 incident patients (45, 6, and 9 patients in units A, B, and C respectively) were received during
2007.

The quality management system was implemented in 2000, and has been certified by AENOR since January 2001. All three units were recently accredited by the Valencian model of accreditation for healthcare activities (INACEPS) specific for hemodialysis, being rated as excellent.

Performance measurement was started in 2000 and showed good results, with an improvement in most indicators monitored. 9 The other previously described improvement tools, feedback and benchmarking, were implemented in 2006, and indicator results were more widely disseminated to the centers, so that each center knew its own results and could compare itself to the other centers. On the other hand, all three centers participate since January 2007 in the Multicenter study of quality indicators in hemodialysis of the Quality Management Group of the SEN. The Nefrosoft® 3.0 (Visual-limes) database is used to record the whole clinical history of patients in an electronic format, and the indicator module (specific quality software (Visual-limes) that works coupled to such database and
allows for easily estimating indicators) has been incorporated into routine work. This facilitates direct calculation of indicators at each center, that also receives every three months the results of the other centers participating in the study. Moreover, all three centers have the specific quality objectives set for centers belonging to Hospital Perpetuo Socorro. Table I shows the main areas for which quality indicators have been defined, and table II lists the indicators defined for the areas in which objectives were subsequently established.

Definition of objectives
In January 2006, an assessment was made of the results of indicators in all three centers in the different treatment areas defined, from which objectives for 2006 and 2007 were defined (table III).

Quality objectives were established separately from quality indicators, and were focused on areas where results were unsatisfactory or where an improvement over the proposed standard was intended. This allowed us for focusing our attention on such aspects. Based on these results, objectives were established for treatment of iron deficiency and control of phosphorus, calcium, and PTH. No objectives were defined for control of hemoglobin or suitability of dialysis, as these showed optimum levels that were difficult to improve.

Planning included definition of the objective, that established a higher goal than achieved in January 2006. An action
plan was devised, the staff, resources, and deadlines for its implementation were established, and an indicator that allowed for monitoring achievement of the objective was defined. Achievement of these objectives was evaluated every three months.

The objectives and the actions intended to promote their achievement are described below:

1. To increase over 70% the proportion of patients with serum calcium levels < 9.5 mg/dL.

Staff in charge of implementation: unit nephrologists. Measurement and control of objectives, as well as setting of common objectives: head of department.

Action plan:

¿ Monthly measurement of serum calcium levels in patients on dialysis according to the previously defined laboratory test plan.
¿ Monthly follow-up of quality indicators defined for monitoring calcium levels to detect downward deviations.
¿ Early modification of vitamin D doses, use of dialysis baths with different calcium levels, and use of calciumbased
phosphate binders in the event of calcium levels higher than 9.5 mg/dL.
¿ Use of calcimimetic drugs in patients with secondary hyperparathyroidism who developed hypercalcemia due to
high vitamin D doses.
¿ Preferential use of bath with a calcium content of 2.5 mEq/L in patients with calcium levels higher than 9.5 mg/dL. Decrease in use of dialyzate containing 3.5 mEq/L of calcium.
¿ Use of combinations of phosphate binders that do not allow for phosphate control with a calcium provision as
phosphate binders under 1500 mg/day.

Resources: Cinacalcet, calcium-free phosphate binders (sevelamer, aluminium hydroxide), calcium-based phosphate binders (calcium acetate), calcium content in dialyzate 2.5 mEq/L.

Timeline: 1 year

2. To increase over 80% the proportion of patients with serum phosphate levels < 5.5 mg/dL.

Staff in charge of implementation: unit nephrologists. Measurement and control of objectives, as well as setting of common objectives: head of department.

Action plan:

¿ Monthly measurement of serum phosphate levels in patients on dialysis according to the previously defined laboratory test plan.
¿ Early modification of vitamin D doses and adequate use of phosphate binders.
¿ Use of combinations of phosphate binders that would allow us for reaching the goal.
¿ Specific action in non-compliant patients by searching for the reasons for non compliance, analysis of preferences
in use of binders (type, scheme), dietary education, and advice on how drugs should be taken.
¿ Establishment of an individual follow-up and encouragement plan, with special motivation of patients not adhering
to treatment.

Resources: Cinacalcet, calcium-free phosphate binders (sevelamer, aluminium hydroxide), calcium-based phosphate binders (calcium acetate), calcium content in dialyzate 2.5 mEq/L.

Timeline: 1 year

3. To increase over 40% the proportion of patients with serum PTH levels ranging from 150 300 pg/dL.

Staff in charge of implementation: unit nephrologists. Measurement and control of objectives, as well as setting of common objectives: head of department.

Action plan:

¿ Two-monthly measurement of PTH and monthly measurements of phosphate and calcium in patients on dialysis
according to the previously defined laboratory test plan.
¿ Maintenance of phosphate levels < 5.5 mg/dL by adequate use of phosphate binders, and emphasis on dietary
education in particularly problematic patients.
¿ Adequate use of new drugs accepted for management of bone and mineral metabolism (calcimimetics).

Resources: Cinacalcet, calcium-free phosphate binders (sevelamer, aluminium hydroxide), calcium-based phosphate binders (calcium acetate), calcium content in dialyzate 2.5 mEq/L.

Timeline: 1 year

4. To decrease under 10% the proportion of iron-deficient patients (ferritin < 100 ng/mL) in unit B, that did not
meet the objective.

Staff in charge of implementation: unit nephrologists. Measurement and control of objectives, as well as setting of common objectives: head of department.

Action plan:

¿ Two-monthly measurement of ferritin levels in patients on dialysis according to the previously defined laboratory
test plan.
¿ Maintenance of an adequate balance of iron metabolism using ferritin controls every three months and intravenous administration of iron every 1, 2, or 4 weeks based on ferritin levels and transferrin saturation index.
Resources: Intravenous administration of iron.

Timeline: 1 year

Feedback and benchmarking
Every three months, the results of all three centers were sent to each unit as a «traffic lights» table showing the results obtained in different colors: adequate results in green, intermediate results close to the goal in yellow, and results far from the goal in red.

Statistical analysis
All data were analyzed using SPSS 12 software.

For quantitative variables, the normality hypothesis was analyzed using a distribution study and a non-parametric
Kolmogorov-Smirnov test. To describe these variables, the arithmetic mean was used as a measure of central tendency, and the standard deviation as a dispersion measure. To study variables, a Student¿s t test was used to compare means when the normality assumptions and homogeneity of variances were met. When the normality assumption could not be verified, a Mann-Whitney¿s U test was used. The comparative analysis of percentages was performed using a Chi-square test for linear trend in proportions, which allowed for analyzing whether improvements in percentages resulted from a trend or were relatively arbitrary. A value of p < 0.05 was considered significant in twosided tests.

RESULTS

Global indicators and demographic data
Table IV shows the demographic and treatment characteristics of patients at study start. Table V gives the results of outcome and treatment indicators as regards anemia and bone and mineral metabolism at study end.

Figures 1, 2, 3, and 4 show the quarterly changes in the goals set. In all three units, a significant increase was seen
in the proportion of patients with serum calcium levels under 9.5 mg/dL, and the goal was reached in all of them.

Compliance increased by 33.5% in unit A, by 26.7% in unit B, and by 19.5% in unit C (fig. 1). The same occurred with the proportion of patients with phosphate levels under 5.5 mg/dL, for which units A, B, and C showed compliance increases of 3.7%, 4.4%, and 19% respectively (fig. 2), and with the proportion of patients with PTH values ranging from 150-300 pg/mL, for which increases in compliance were 14.5%, 2.4%, and 13.1% in units A, B, and C respectively (fig. 3).

In unit B, the proportion of iron-deficient patients (ferritin < 100 ng/mL) decreased to less than 10%, similar to the results found in the two other units (fig. 4). At study start, this unit did not reach the defined standard (ferritin levels ranging from 100-800 pg/mL: 59.3%; ferritin < 100: 39.0%; and ferritin > 800: 0.9%); after the follow-up period, optimum overall results were found (proportion of patients with ferritin levels ranging from 100-800 pg/mL: 94.1%; ferritin < 100: 5.9%; and ferritin > 800: 0%), similar to those in the two other units (unit A: ferritin < 100: 4.3%, and unit C: ferritin < 100: 4.6%).

Mean EPO use in 2005 was 145.5.5 ± 13.2 U/kg/week in unit A, 226.2.2 ± 39.8 U/kg/week in unit B, and 175.5.5 ±
13.9 U/kg/week in unit C. At the end of 2007, mean EPO use significantly decreased in unit B (144.2.2 ± 15 U/kg/week) to values similar to those in the two other units (140 ± 14.2 in unit A and 135.1.1±13.8 in unit C). The erythropoietin resistance index (ERI) decreased from 26.1 to 11.3 in unit B.

DISCUSSION

This study shows that use of quality management tools such as monitoring of quality indicators, setting of goals with
their associated improvement plans, provision and receipt of information about the results achieved (feedback),
knowledge by the centers of their own results as compared to those of other centers (benchmarking), and the receipt of «warnings» to remind deviations from the established goals may be helpful tools to promote achievement of improved results and a decreased variability between centers. The type of study design allowed for assessing the effectiveness of the measure taken, but it is difficult to say what was the actual cause of change because this probably resulted from the sum of all these factors.

Between-center variability in the outcome of hemodialysis treatment is an actual fact, that has been shown in multiple studies.15,16 Such variability may be explained by multiple and complex reasons:5-8 organizational problems; limited resources; lack of quality management tools; conceptual errors of healthcare professionals; different preferences; or differences in the demographic characteristics of the population studied. In addition, while the different therapeutic strategies are within the reach of everybody, they are not used in the same way in all centers. Even the goals and standards established for the same healthcare process are different depending on the
country; for instance, guidelines for vascular access of the SEN recommend that an arteriovenous fistula should be placed in 80% of prevalent patients on hemodialysis,17 while the goal established in the K&