Predictive modeling of cardiovascular complications in incident hemodialysis patients

The administration of hemodialysis (HD) treatment leads to the continuous collection of a vast quantity of medical data. Many variables related to the patient health status, to the treatment, and to dialyzer settings can be recorded and stored at each treatment session. In this study a dataset of 42 variables and 1526 patients extracted from the Fresenius Medical Care database EuCliD was used to develop and apply a random forest predictive model for the prediction of cardiovascular events in the first year of HD treatment. A ridge-lasso logistic regression algorithm was then applied to the subset of variables mostly involved in the prediction model to get insights in the mechanisms underlying the incidence of cardiovascular complications in this high risk population of patients.

Increased use of catheters as vascular access: is it justified by patients' clinical conditions?

BACKGROUND

Over the last years many technical improvements have been made in hemodialysis treatment. Vascular access (VA) still remains an important problem. Although the use of indwelling vascular catheters is discouraged, in Europe there is an increasing use of them. The K/DOQI Guidelines recommend a native arteriovenous fistula (AVF) as VA of choice. As reported by DOPPS, there is considerable geographic variation in the distribution of type of VA used amongst hemodialysis patients. The aim of this study was to evaluate the time patients in four European countries have to wait before undergoing their first surgery for VA (AVF or graft).

METHODS

All incident patients admitted to HD clinics located in Turkey, Italy, the UK and Portugal of the European FME clinics network between October 1, 2002 and September 30, 2004 were considered. Data were gained from the Clinical Database EuCliD.

RESULTS

2,152 patients (males 55.9%, mean age 62.5+/-15.7 years, diabetics 27%) were selected. Italy and Portugal had a higher proportion of elderly patients. At time of admission, the proportion of patients starting dialysis with AVF ranged between 23% and 60% from Turkey to Italy respectively. Patients with an indwelling catheter at admission are expected to undergo VA surgery as soon as possible. After 3 months of follow-up, about 75% of all patients had undergone surgery, however in the UK less than 50% of the patients had had a VA procedure. Overall, males have significantly higher probability of undergoing surgery, whilst elderly patients have a lower probability (27% and 14% respectively).

CONCLUSION

Significant differences exist between countries in the time interval from referral to creation of VA. Health care system related problems seem to be the major reason to explain such differences. Patients in the UK have longer waiting times than the other countries studied.

Satisfaction with care in peritoneal dialysis patients

Patient satisfaction is an important aspect of dialysis care, only recently evaluated in clinical studies. We developed a tool to assess peritoneal dialysis (PD) customer satisfaction, and sought to evaluate and validate the Customer Satisfaction Questionnaire (CSQ), quantifying PD patient satisfaction. The CSQ included questions regarding administrative issues, Delivery Service, PD Training, Handling Requests, and transportation. The study was performed using interviews in all Hungarian Fresenius Medical Care dialysis centers offering PD. CSQ results were compared with psychosocial measures to identify if patient satisfaction was associated with perception of social support and illness burden, or depression. We assessed CSQ internal consistency and validity. Factor analysis explored potential underlying dimensions of the CSQ. One hundred and thirty-three patients treated with PD for end-stage renal disease for more than 3 months were interviewed. The CSQ had high internal consistency. There was high patient satisfaction with customer service. PD patient satisfaction scores correlated with quality of life (QOL) and social support measures, but not with medical or demographic factors, or depressive affect. The CSQ is a reliable tool to assess PD customer satisfaction. PD patient satisfaction is associated with perception of QOL. Efforts to improve customer satisfaction may improve PD patients' quantity as well as QOL.

[Health behaviours of students at the University Institute of Technology in Poitiers]

Better knowledge and understanding of students' health attitudes and behaviours are necessary to more effectively plan targeted prevention strategies. A survey conducted among 617 students at the University Institute of Technology in Poitiers showed that although some indicators affirm good health (relationships with friends and family, appropriate contact with care providers in the health system and regular visits to prevention centres, sexual life, and sports and physical fitness), other indicators reveal a certain amount of suffering among the student population. Students reported suicidal thoughts, symptoms of poor well-being (fatigue, sleeping disorders, concentration problems, sadness or depression) and poor eating habits and/or eating disorders. Consumption of both legal and illegal products seems to be a means utilised to facilitate the social integration process for some students. Therefore, this underlines a significant challenge for prevention campaigns to develop an alternative role model, which promotes the image and profile of a healthy student, with a strong sense of self-esteem, who is well-integrated with his peers without having to engage in substance abuse.

EuCliD--a medical registry

OBJECTIVES

The European Clinical Database EuCliD small star, filled has been developed as a tool for supervising selected quality indicators of about 200 European dialysis centers. Major efforts had to be made to comply with local and European laws regarding data security.

METHOD

EuCliD is a Lotus Notes based flat-file database currently containing medical data of more than 14,000 dialysis patients from 10 European countries. Another 15,000 patients from 150 centers in 4 South-American countries will be added soon. Data are entered either manually or by means of interfaces to existing local data managing systems. This information is transferred to a central Lotus Notes Server. Data evaluation was performed with statistical tools like SPSS.

RESULTS

EuCliD is used as a part of the CQI (Continuous Quality Improvement) management system of Fresenius Medical Care (FMC) dialysis units. Each participating dialysis center receives (currently every half year) benchmarking reports at a regular interval. The benchmark for all quality parameters is the weighted mean of the corresponding data of all centers.

CONCLUSIONS

An obvious impact of data sampling and data evaluation on the quality of the treatments could be observed within the first one and a half years of working with EuCliD. This also concerns important outcome predictors like Kt/V and hemoglobin concentration as the outcome itself expressed in hospitalization days and survival rates. With the help of EuCliD the user is able to sample clinical data, identify problems, search for solutions with the aim of improving the dialysis treatment quality and guarantee a high-class treatment quality for all patients.

[Anemia management in haemodialysis. EuCliD database in Spain]

We present the results on Anaemia Management in Fresenius Medical Care Spain dialysis centres as reported by EuCliD (European Clinical Database), evaluating a population of 4,426 patients treated in Spain during the year 2001. To analyse the erythropoietin dose and the haemoglobin levels we divided the population in two groups according to the time with dialysis treatment: patients treated less than six months and patients between six months, and four years on therapy. We compared our results with the evidence based recommendations Guidelines: the European Best Practice Guidelines (EBPG) and the US National Kidney Foundation (NKF-K/DOQI). We also compared our results with those presented by the ESAM2 on 2,618 patients on dialysis in Spain carried out in the second half of the year 2000. We observed that 70% of the population reaches an haemoglobin value higher that 11 g/dl, with a mean erythropoietin (rHu-EPO) dose of 111.9 Ul/kg weight/week (n = 3,700; SD 74.9). However, for those patients on treatment for less than six months, the mean Haemoglobin only reaches 10.65 g/dl (n = 222; SD 1.4). The rHu-EPO was administrated subcutaneously in 70.2% of the patients. About the iron therapy, 86% of the patients received iron treatment and the administration route was intravenous in 93% of the population. The ferritin levels were below 100 micrograms/dl in 10% of the patients and 26.4% showed a transferrin saturation index (TSAT) below 20%. The erythropoieting resistance index (ERI), as rHu-EPO/haemoglobin, has been used to evaluate the response to rHu-Epo, according to different variables. It was observed that the following factors lead to a higher rHu-EPO resistance: intravenous rHu-EPO as administration route, the presence of hypoalbuminemia, increase of protein C reactive, Transferrin saturation below 20% and starting dialysis during the last six months.

Results from EuCliD (European Clinical Dialysis Database): impact of shifting treatment modality

BACKGROUND

The use of biocompatible high-flux membranes is more efficient than low-flux membranes in controlling a number of hemodialysis-related diseases. The aim of this cooperative study was to evaluate the 6-month effect of a switch from low- to high-flux dialysers on patients treated in 39 Spanish dialysis centres.

METHODS

The clinical data used in this analysis were prospectively collected by the EuCliD database, developed to monitor the quality of treatment delivered in a large network of European Dialysis Centres. Inclusion criteria for the study were the condition of end-stage renal disease (ESRD) on chronic hemodialysis and low-flux dialysis for at least six months before the switch to high-flux dialysis. Of 1,543 patients enrolled in the study between 2000 and 2001, 1,046 patients were considered for the analysis. 497 patients were excluded because they did not complete the follow-up. Outcome measures were the reduction of pre-dialysis beta-2 microglobulin, the improvement of anemia or reduction in rHu-EPO dose required to maintain best correction of anemia, reduction of inflammatory parameters (CRP), improvement in lipid profile (Total and HDL cholesterol, tryglycerides), maintenance of nutritional status. Albumin and "dry" (post-hemodialysis) body weight were both evaluated as nutritional indexes.

RESULTS

During the six months of high-flux hemodialysis, there was a significant increase in hemoglobin (from 11.55 +/- 1.41 to 11.88 +/- 1.43 g/L; p < 0.001). Considering the temporarily untreated patients on a 0 U/week dose, erythropoietin remained stable (from 5,670 +/- 4,199 to 5,657 +/- 4,411 U/week). During the second part of the follow-up, the lipid profile significantly improved (Fig. 3). Total cholesterol and triglycerides decreased significantly (p < 0.001), while HDL cholesterol increased (p = 0.006). Calculated levels of LDL cholesterol also significantly decreased (p = 0.001). Dry body weight remained stable (64.7 +/- 11.9 vs. 64.7 +/- 12.0 kg) as well as in albumin levels (3.93 +/- 0.43 vs. 3.94 +/- 0.43 g/dL) between the two modalities of treatment. The level of beta2-microglobulin significantly decreased during high-flux dialysis (33.5 +/- 14.4 vs. 26.3 +/- 8.6 mg/dL, p < 0.001).

CONCLUSION

All above mentioned results may have as a common denominator an improved blood purification from uremic toxins and a reduced level of chronic sub-clinical inflammation. All together, these results seem to confirm the superiority of high-flux dialysis in terms of clinical and physiological outcomes.

EuCliD (European Clinical Database): a database comparing different realities

Quality and variability of dialysis practice are generally gaining more and more importance. Fresenius Medical Care (FMC), as provider of dialysis, has the duty to continuously monitor and guarantee the quality of care delivered to patients treated in its European dialysis units. Accordingly, a new clinical database called EuCliD has been developed. It is a multilingual and fully codified database, using as far as possible international standard coding tables. EuCliD collects and handles sensitive medical patient data, fully assuring confidentiality. The Infrastructure: a Domino server is installed in each country connected to EuCliD. All the centres belonging to a country are connected via modem to the country server. All the Domino Servers are connected via Wide Area Network to the Head Quarter Server in Bad Homburg (Germany). Inside each country server only anonymous data related to that particular country are available. The only place where all the anonymous data are available is the Head Quarter Server. The data collection is strongly supported in each country by "key-persons" with solid relationships to their respective national dialysis units. The quality of the data in EuCliD is ensured at different levels. At the end of January 2001, more than 11,000 patients treated in 135 centres located in 7 countries are already included in the system. FMC has put the patient care at the centre of its activities for many years and now is able to provide transparency to the community (Authorities, Nephrologists, Patients.....) thus demonstrating the quality of the service.