|Year : 2016 | Volume
| Issue : 7 | Page : 1-11
|The dialysis outcomes and practice patterns study phase 5 in the Gulf Cooperation Council countries: Design and study methods
Ronald L Pisoni1, Brian A Bieber1, Jamal Al Wakeel2, Sameer Al Arrayed3, Naser Alkandari4, Mohamed Hassan5, Ayman Karkar6, Nabil M Al Lawati7, Fadwa Al Ali8, Justin M Albert1, Bruce M Robinson9, GCC-DOPPS 5 Study Group10
1 Arbor Research Collaborative for Health, Ann Arbor, MI, USA
2 King Saud University, Riyadh, Saudi Arabia
3 Salmaniya Medical Complex; Manama, Bahrain
4 Mubarak Alkabeer Hospital; Kuwait City, Kuwait
5 Sheikh Khalifa Medical City, Abu Dhabi, United Arab Emirates
6 Ministry of Health, Riyadh, Saudi Arabia
7 Sultan Qaboos University, Muscat, Oman
8 Hamad General Hospital, Doha, Qatar
9 Arbor Research Collaborative for Health; Department of Internal Medicine, Division of Nephrology, University of Michigan, Ann Arbor, MI, USA
10 List of Study Group in Acknowledgment
Click here for correspondence address and email
|Date of Web Publication||1-Dec-2016|
| Abstract|| |
The Dialysis Outcomes and Practice Patterns Study (DOPPS) is an international prospective cohort study of the relationships between hemodialysis (HD) care practices and HD patient outcomes. The DOPPS began in 1996, in the United States, and has since expanded to 21 countries, collecting detailed data from >75,000 HD patients, with >200 scientific publications, focused on describing HD practices associated with improved HD patient outcomes. The goal of DOPPS is to help HD patients "live better and live longer." Starting in 2012, the DOPPS was able to expand to all six of the Gulf Cooperation Council (GCC) countries, namely, Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the United Arab Emirates. The DOPPS study design consists of selecting HD facilities for study participation in each country to represent the different types of HD facilities and geographic regions within each GCC country. Within each study site, HD patients were randomly selected for detailed data collection to represent the HD practices within each participating HD facility. Altogether, 41 HD facilities have participated in the GCC-DOPPS Phase 5 study including 20 facilities from Saudi Arabia, nine from the United Arab Emirates, four each from Kuwait and Oman, two from Qatar, and one from Bahrain. Herein, we provide a detailed description of the study design and methods, data collection, study management, scientific investigator oversight and guidance, and study governance and support for the GCCDOPPS Phase 5 study.
|How to cite this article:|
Pisoni RL, Bieber BA, Al Wakeel J, Al Arrayed S, Alkandari N, Hassan M, Karkar A, Al Lawati NM, Al Ali F, Albert JM, Robinson BM, GD. The dialysis outcomes and practice patterns study phase 5 in the Gulf Cooperation Council countries: Design and study methods. Saudi J Kidney Dis Transpl 2016;27, Suppl S1:1-11
|How to cite this URL:|
Pisoni RL, Bieber BA, Al Wakeel J, Al Arrayed S, Alkandari N, Hassan M, Karkar A, Al Lawati NM, Al Ali F, Albert JM, Robinson BM, GD. The dialysis outcomes and practice patterns study phase 5 in the Gulf Cooperation Council countries: Design and study methods. Saudi J Kidney Dis Transpl [serial online] 2016 [cited 2017 Jan 20];27, Suppl S1:1-11. Available from: http://www.sjkdt.org/text.asp?2016/27/7/1/194878
| Introduction|| |
The Dialysis Outcomes and Practice Patterns Study (DOPPS) is an international prospective cohort study of in-center hemodialysis (HD) patients and practices, ongoing since 1996, with detailed patient-level data collected from >75,000 HD patients to date with 21 countries currently participating in the DOPPS [Figure 1]. The DOPPS is an observational study of how HD is routinely practiced with no study interventions or additional tests required. There have been five DOPPS study phases thus far with some special features applied within each DOPPS study phase to complement the overall DOPPS study design and data collection. The goal of DOPPS is to identify HD practices associated with improved HD patient outcomes as a means to help patients to live longer and live better on HD. The DOPPS carries out this endeavor by: (a) describing variation and monitoring trends in HD practice in nationally or regionally representative samples of HD facilities in numerous countries, (b) identifying the best practices through rigorous analyses of the relationships between HD practices and patient outcomes, and (c) communicating key study findings to the worldwide HD community to improve clinical care. Key outcomes examined in the DOPPS include patient survival, hospitalization, quality of life and other patient-reported outcomes, vascular access, and guideline achievement for many different HD practice indicators. More than 200 published papers have been contributed from the DOPPS thus far, in addition to numerous presentations internationally each year, as a means to help inform the broad worldwide HD community of HD practices related to the best outcomes for benefiting patients. In addition, summaries of DOPPS publications with downloadable graphics are available on the DOPPS website at www.DOPPS.org.
A distinguishing feature of the DOPPS study design is that by collecting detailed data at many study sites across numerous countries, much greater variation in HD practice is observed than in a single country. Additionally, this study design allows relationships between facility HD practices and patient outcomes to be investigated using instrumental variable statistical methods which reduce confounding with the possibility of a more reliable understanding of treatment effects. The overall study design, nationally representative sampling approach, and data collection used in the DOPPS have been described previously. ,,
Here, we provide a detailed description of the study design and methods used for DOPPS 5 (2012-2015) data collected from the six Gulf Cooperation Council (GCC) countries: Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the United Arab Emirates. The same data collection protocol was used in numerous other countries in DOPPS 5 and was designed to collect important longitudinal and cross-sectional HD patient-level data with decreased data collection work burden.
| Methods|| |
Random selection of HD facility study sites with the goal of nationally representative study samples
HD facility sampling frame:
The overall approach for selecting HD facilities in each GCC country in DOPPS 5 was to obtain a list of all HD facilities within each country which included an estimate of the number of patients treated at each facility, the geographic location (city) of each facility, and the type (governmental or private facility) of HD facility. This latter information was provided by nephrologists in each country who have served as DOPPS country investigators in each GCC country, and they are listed in [Table 1]. For the GCC in DOPPS phase 5, we list in [Table 2] the total number of HD facilities and patients, and types of HD facilities per country, as well as how sampling was stratified by geographic region within each country. In addition, the number of study facilities enrolled in GCCDOPPS 5 by country, and their distribution by facility type and geographic region are shown. The DOPPS study design requires a minimum number of 20-25 in-center HD patients to be treated in a facility in order for a facility to be eligible for participation in the DOPPS which makes it feasible to reliably estimate a facility's practice based on at least this number of study patients within each participating HD unit. A target of 40 study facilities was established for the GCC-DOPPS 5 study to provide representative results for the GCC as a whole and as representative as possible for each country. The target of 40 facilities for the GCC was based on prior DOPPS experience and funding support level for the GCC-DOPPS. Confidence intervals for the estimates are expected to be larger in countries having a smaller number of participating facilities. Data collection was initiated in quarter-3 of 2012 and finished in quarter-4 of 2015.
|Table 1: DOPPS 5 country investigators from the Gulf Cooperation Council by country.|
Click here to view
Once the total number of study-eligible HD facilities and the number of study sites to be recruited were established for each country, HD units were randomly selected for study participation from among all study-eligible HD facilities within each country. The random selection of study sites was stratified to provide approximately proportional representation of HD units in each country according to geographic location (if relevant) and facility types as shown in [Table 2]. The random selection of study sites was carried out at the DOPPS Coordinating Center at Arbor Research Collaborative for Health in Ann Arbor, MI, USA. The list of randomly selected facilities was provided to the country investigators in each GCC country who then contacted the medical director at each randomly selected HD unit to inquire of the interest and feasibility for the HD unit to participate in the DOPPS. The study sites were financially compensated for their participation in the DOPPS, with compensation in accordance to the number and type of study questionnaire completed. Facilities that were approached for study participation but did not agree to participate were randomly replaced by approaching the next facility on the randomized list of studyeligible facilities, and this process was repeated until the target of 40 participating GCC study sites meeting the facility type strata requirements in each country was achieved. Overall, more than 75% of facilities which were approached for study participation agreed to participate in DOPPS 5.
Random selection of prevalent and incident HD patient samples at each study site
Each HD facility participating in the DOPPS initiated the study by compiling a census list indicating all in-center HD patients being treated at the unit at that point in time and updating this list throughout the study by indicating all HD patients who depart or enter the facility for HD therapy during DOPPS 5 participation. This facility census serves as the sampling frame of in-center HD patients within each facility for random selection of two different HD patient samples at each study site: (a) prevalent (cross-sectional) patient sample and (b) an incident patient sample selected over time during DOPPS 5. The prevalent patient sample was obtained by randomly selecting 20-30 patients from among all patients receiving maintenance HD therapy within each selected dialysis unit at the time the dialysis unit initiated DOPPS 5 data collection. A greater number of sample patients was selected from larger dialysis units within each country as defined by a target size algorithm. Sampled prevalent patients who did not consent to the study participation or departing from the study for any reason (e.g., death, transfer to another HD unit, switch to peritoneal dialysis, kidney transplantation, and regain kidney function) were randomly replaced annually from among patients new to the HD unit since the prior random selection and were still receiving HD treatment in the facility at the time of the new random selection. Enrollment of the incident patient sample began at each site approximately four months after the study initiation. Incident HD patients who had initiated chronic HD within 30 days of being listed on the study census were sequentially selected for study participation at fourmonth intervals in establishing the incident patient cohort in each country. Incident patients were selected for study participation until a total of 15 incident HD patients were enrolled in GCC-DOPPS 5 from each study site. However, some of the smaller GCC study sites were not able to achieve participation of 15 incident HD patients within the time period of DOPPS 5. The above sampling approach has resulted in detailed data collection from 1139 prevalent HD patients and 180 incident patients in GCCDOPPS 5. Data were reported only for patients who consented to participate in the study, with an overall con-sent rate of 86% for the GCCDOPPS 5 study. DOPPS was approved by the relevant Ethics Committee in each country and for each participating study site as required.
Patient and facility-level data
A broad spectrum of different types of patient-level and facility-level data were collected for each patient [Table 3]. These data included relevant, detailed information for sampled patients, including demographics, more than 70 measures of comorbidity, socioeconomic status, primary cause of ESRD, laboratory values, prescribed medications, hospitalization and outpatient events, vascular access use, dialysis prescription, delivered dialysis dose, residual renal function, nutritional measures, aspects of medical care prior to ESRD, kidney transplant waitlisting, patient quality of life, physiciandiagnosed and patient-reported measures of depression, and date and cause(s) of death for patients who died during the study or who transferred to or from peritoneal dialysis or kidney transplantation. These patient data served to represent the practices and patient mix of each participating dialysis unit, and allowed these characteristics and practices to be related to a wide variety of key outcomes (e.g., mortality, hospitalization, patient quality of life, achievement of guideline targets).
In addition to the collection of detailed patient data, a major focus in the DOPPS was the collection of information on many different aspects of facility-level HD practice as indicated by the medical directors and nurse managers of the HD facilities participating in DOPPS 5. [Table 4] highlights some of the major facility practice areas for which data were collected from the study site medical directors in DOPPS 5.
|Table 4: Facility-level data collection items collected from the study site medical directors and nurse managers in GCC-DOPPS 5.|
Click here to view
Data collection, study management, and scientific oversight
At all GCC-DOPPS 5 study sites, all data, except for patient questionnaires, were reported to the DOPPS Coordinating Center through the DOPPS link web-based data collection system which is a customized data entry system created by Arbor Research Collaborative for Health. Patient data were reported in an anonymous fashion and thus, a patient's identity and participation in the DOPPS will not be known outside of the clinic in which the patient is treated. Typically, patient-level data were abstracted from the patient's medical record and reported anonymously from each study site by a study coordinator who was a nurse or physician working at the participating HD unit. In addition, the study coordinators approach patients for consent to participate in the DOPPS and distribute paper patient questionnaires to be completed by study patients during the time of study entry and annually thereafter.
Data collection and data quality activities for GCC-DOPPS 5 were coordinated at the country level, by one or more clinical research associates (CRAs) within each country [Table 5], and coordinated across the international DOPPS program by the DOPPS project team members from the DOPPS Coordinating Center. The CRAs played a very important key role for the success of GCC-DOPPS 5. The CRAs were trained on the use of the DOPPSLink webbased data collection system and in turn, the CRAs trained each of the GCC-DOPPS 5 study site coordinators regarding how to input data for the study patients according to the DOPPS 5 study protocol. Moreover, the CRAs maintained a close contact with each study site in the given country throughout DOPPS 5 to encourage achievement of study data collection goals and address study site data collection questions, difficulties in reporting study data according to the study timelines, missing data, data quality-related questions, study site payments, and reporting changes or long-term absence of study site medical directors or study coordinators. An important aspect of this coordinated effort to maintain high-quality data collection was close interaction of the country CRAs with the study staff at the DOPPS Coordinating Center in Ann Arbor, MI, USA, who provided ongoing communication with each CRA regarding the quality of the data being received from each study site over the course of the study. Depending on the nature of the problem, the study site data collection problems were addressed by a combination of efforts by CRAs, DOPPS Coordinating Center staff, and local DOPPS country investigators within the GCC.
Data reported for GCC-DOPPS 5 were routinely run through quality control programs (e.g., range checks, units of measure, evaluation of nonresponse, missing data) when data were first received by the DOPPS Coordinating Center. Queries generated from these database quality control procedures were brought to the attention of the primary data collectors at each study site to verify the quality of specific data elements and address the issues of nonresponsiveness. In addition, statistics calculated from GCC-DOPPS 5 data were compared to registry reports or data where available, as part of validating the representativeness of collected data [e.g., compared with the Saudi Center for Organ Transplantation Data Dialysis in the Kingdom of Saudi Arabia 2014, Saudi J Kidney Dis Transplant 2014;25(4):918-926].
Scientific oversight of GCC-DOPPS 5 was achieved through the devoted involvement of 22 GCC-DOPPS 5 country investigators along with investigators from the DOPPS Coordinating Center [Table 1]. In addition, research priorities were established by the GCC-DOPPS 5 Scientific Advisory Committee (SAC), in communication with other country investigators, with the SAC consisting of one country investigator from each GCC country [Dr. Ali Alsahow (Kuwait), Dr. Mona Al Rukhaimi (UAE), Dr. Issa Alsalmi (Oman), Dr. Sumaya Al Ghareeb (Bahrain), Dr. Faissal Shaheen (Saudi Arabia), and Dr. Fadwa Al Ali (Qatar)], and Dr. Ronald Pisoni and Dr. Bruce Robinson from the DOPPS Coordinating Center.
Beyond general statistical descriptive analyses, associations between predictors and outcomes (e.g., patient quality of life measures, achievement of guideline practice indicators, and clinical outcomes) were analyzed at the subject level and, when feasible, at the facility level. Analyses used Poisson, logistic, proportional hazards, log-linear, or linear regression models, as appropriate, controlling for subject demographic and comorbidity differences. Subject-level left-truncated proportional hazards models were used to model time to death or first observed hospitalization during the study (and other primary events). Subjectlevel analyses also were modeled to account for the effects of facility-level clustering on the variability between patients through either random effects models with linear, logit, log links, or robust standard estimates based on the sandwich estimator for proportional hazards models. Analyses of longitudinal trends in subject characteristics, practice indicators, and outcomes used repeated measures models to account for serial measures over successive years (time-dependent covariates).
All analyses attempted to identify the evidence of potential biases, and diagnostic tools were used to assess goodness of fit and appropriateness of models according to the accepted statistical practices. Most analyses were repeated using different levels of adjustment to increase clinical understanding and assess the stability of findings.
When feasible, depending on sample size, practice variation, and effect estimates, instrumental variable analyses of treatment or practice with outcomes were considered as a possible approach for minimizing potential confounding. ,,,,,
In analyses included in each of the manuscripts for this special issue, descriptive national statistics for each country were derived from the initial cross-section of selected patients and weighted by the facility sampling fraction calculated as the number of sampled patients divided by the total number of patients at the facility.
Data use and sharing
External researchers were encouraged to utilize DOPPS data and submit requests for the secondary use of GCC-DOPPS 5 to the DOPPS Coordinating Center (via www.DOPPS.org). Requests for data use were made through a standardized data request form available on the DOPPS website. Researchers were required to submit a proposal detailing the nature of the study research question, data elements needed, planned analyses, ability and skills in carrying out statistical analyses, data security protocols, and publication plans, as appropriate. Proposals were reviewed for scientific merit and appropriateness of the release by the members of the GCC-DOPPS SAC and by the DOPPS Steering Committee. If the proposal was approved, researchers were required to complete a data use agreement with Arbor Research, which outlines the terms of the data use, confidentiality, publications, and destruction of the data, once it has been used for the stated purpose.
Protection of human beings
A patient's treatment was not affected by participation in the DOPPS, as the DOPPS is purely an observational study with no additional interventions as a result of study participation. A very high level of confidentiality and security was maintained with all data collected in the DOPPS as evidenced over 20 years of carrying out the DOPPS. Furthermore, patient-level data were collected in an anonymous fashion such that the identity of patients was known only to the facility staff members (their nurses and doctors). A patient's name or identifying number was not reported within DOPPS data. To participate in the DOPPS, study sites were required to meet pertinent ethics committee requirements, as well as each study patient providing written consent for study participation.
| Acknowledgments|| |
The authors wish to thank the devoted efforts of GCC-DOPPS 5 Study group members for their numerous contributions to this study including country investigators (Dr. Sameer Al-Arrayed, Dr. Sumaya Al Ghareeb, Dr. Bassam Al Helal, Dr. Naser Alkandari, Dr. Ali Alsahow, Dr. Anas AlYousef, Dr. Mohammad AlAzmi, Dr. Yaqob Ahmed Almaimani, Dr. Issa Alsalmi, Dr. Nabil Salmeen Mohsin, Dr. Fadwa Al Ali, Dr. Mohamed El-Sayed Abdel Fatah, Dr. Ashraf Fawzy, Dr. Abdulla Hamad, Dr. Saeed M. G. Al-Ghamdi, Dr. Mohammed Al Ghonaim, Dr. Jamal Al Wakeel, Dr. Fayez Hejaili, Dr. Ayman Karkar, Dr. Faissal Shaheen, Dr. Samra Abouchacra, Dr. Ali Abdulkarim Al Obaidli, Dr. Mohamed Hassan, Dr. Mona Nasir Al Rukhaimi, Dr. Abdul Kareem Saleh, Dr. Sylvia Ramirez, Dr. Bruce Robinson, and Dr. Ronald Pisoni), Clinical Research Associates and Project Co-ordinators/Managers (Dr. Amgad El-Baz El-Agroudy, Dr. Balaji Dandi, Ms. Cherry Flores, Mrs. Ph. Fatma Al Raisi, Ms. Ibtisam Al-Hasni, Dr. Ashraf Fawzy, Dr. Haroun Zakaria Ahmed, and Mr. Dan Santiago from the Saudi Center for Organ Transplantation, Mr. Muhammad Awwad, Ms. Roberta Al Housani-Blakely, Christina Pustulka, Anne Vandermade, Melissa Fava, Anna Hogan, Michelle Maxim, Justin Albert), Biostatisticians (Brian A. Bieber), and Study Site Medical Directors, and Study Co-ordinators: Dr. Hamid Ali Alyousif, Ms. Sohair Abdulroof Albably, Dr. Ahmed Eid Alkady, Dr. Samir Al Muielo, MD, Dr. Fakreldein Elamien Ali, Dr. Ayman El Monger, Dr. Sameh Mohammed Al-sammad, Dr. Baraa Rajab, Dr. Atif Hanan, Tarek Abdelfattah Ahmed Ali, Dr. Ayman Karkar, Dr. Mohammed Abdelrahman, Dr. Mohammed Hussein, Dr. Wael Abdlah, Dr. Faisal Mushtaq, Dr. Salah Eldin El Sheik Beshir, Dr. Medhat Abdelmonem Shalaby, Dr. Mustafa Khaleel Al Obeid, Dr. Mohamad El Hadary, Dr. Alaa Al Shamy; Dr. Aboud Mamari, Dr. Mohammed Raily, Dr. Mustafa Ahmed; Ramzi AbouAyache, Dr. Hormoz Dastoor, Dr. Mohamed Hassan, Dr. Bassam Bernieh, Dr. Hareth Muthanna Mohammed Saaed, Dr. Mustafa Kamel, Prof. Abdel Basset Hassan, MD, PhD, FASN, Dr. Hassan Khammas, Dr. Balaji Dandi, Dr. Sumaya Al-Ghareeb, Dr. Fadwa Al Ali, Dr. Anas Alyousef, Dr. Rouben George, Dr. Nasser AlKandari, Dr. Ali Alsahow, Dr. Bassam AlHelal, Dr. Ahmed Mandour, Dr. Yaqob Ahmed Almaimani, Dr. Amer Ahmed Alaamri, Dr. Ibrahim Sayed Ibrahim, Dr. El Badri Abdelgadir, Dr. Mohammed Al-Sayed Seleem, Dr. Ali Hassan Hakami; Dr. Samir Beshay, Dr. Nayer Morsy, Mr. Menwer Al-ofi, Ms. Sohair Abdulroof Albably, Dr. Huda Mohammed Saeed Ahmed, Ola Al-Marhoon; Dr. Samir Al Muielo, Ms. Joynalyn Barrios, Ms. Fatima Cruza; Ligaya Battad, Mr. Abdullah A. K. Al Harbi, Dr. Sameh Mohammed Al-sammad, Ms. Manal Ahmed Hamdan, Jennifer Samson, Mr. Ahmed Mousa Khawaji, Ms. Eman Al-Hejji, Ms. Rosalinda Lamis, Ms. Lagrimas Codotco, Dr. Wael Abdlah, Mrs. Priyanka Prasad, RN, Mr. Bander Faisal Justinia, Mr. Faisal Al Enazy, Sarah Brazil, Dr. Naemah Abdullah, Dr. Alaa Al Shamy, Dr. Aboud Mamari, Mely Gari Piling, Dr. Fakhriya Al Alwai, Rusmina Binti Sudin, Dr. Chandra Mauli Jha, MD; Katheryn Jamilano, Basima Khaddah, Hilal Al Rasbi, Muhy Eddin Hashem Hasan, Sr. Ekram Awadh Salem, Walid Gouiaa, Mini Issac, Jiby Mammen, Mary Ellen Monday, Mrs. Afrah Al Jamri, Mr. Yasser Khalil Abbas, Mrs. Rania Abd El-Aziz, Rania Abd El-Aziz, Cherry Flores, Mini Kumari Ramakrishnan, Shanty Mannaraprayil, Maria Charisse Bernardo, Bassam AlHelal, Evangeline Valdez, Rosily Joseph, Mrs. Ghaniya Al-Shukaili, Adel Mohamed Bayoumy, Hamid Alshahri, Amira Balkhair, Amina Said Al Shezawi, Sultan Saif Ali Alroshdi, Dr. Shaninaz Faisal Bashir, Dr. Ahmed Mousa Dawood, Keirin Porras, Dr. Anthonimuthu Victor, Mr. Sultan Al-Toqi.
Source of Support
The DOPPS Phase 5 Study in the GCC has been supported by Amgen without restrictions on publications. The DOPPS Program is supported by Amgen, Kyowa Hakko Kirin, AbbVie, Sanofi Renal, Baxter Healthcare, and Vifor Fresenius Medical Care Renal Pharma. Additional support for specific projects and countries is provided by Keryx Biopharmaceuticals, Merck Sharp & Dohme, Proteon Therapeutics, Relypsa, and F. Hoffmann-LaRoche; in Canada by Amgen, BHC Medical, Janssen, Takeda, and the Kidney Foundation of Canada (for logistics support); in Germany by Hexal, DGfN, Shire, and the WiNe Institute; and for PDOPPS in Japan by the Japanese Society for Peritoneal Dialysis (JSPD). All support is provided without restrictions on publications.
Conflict of interest: None declared.
| References|| |
Young EW, Goodkin DA, Mapes DL, et al. The Dialysis Outcomes and Practice Patterns Study (DOPPS): An international hemodialysis study. Kidney Int Suppl 2000;74:S74-81.
Pisoni RL, Gillespie BW, Dickinson DM, Chen K, Kutner MH, Wolfe RA. The Dialysis Outcomes and Practice Patterns Study (DOPPS): Design, data elements, and methodology. Am J Kidney Dis 2004;44 Suppl 2:7-15.
Robinson B, Fuller D, Zinsser D, et al. The Dialysis Outcomes and Practice Patterns Study (DOPPS) Practice Monitor: Rationale and methods for an initiative to monitor the new US bundled dialysis payment system. Am J Kidney Dis 2011;57:822-31.
Li Y, Lee Y, Wolfe RA, et al. On a preference-based instrumental variable approach in reducing unmeasured confounding-by-indication. Stat Med 2015;34:1150-68.
Pisoni RL, Arrington CJ, Albert JM, et al. Facility hemodialysis vascular access use and mortality in countries participating in DOPPS: An instrumental variable analysis. Am J Kidney Dis 2009;53:475-91.
Pisoni RL, Bragg-Gresham JL, Fuller DS, et al. Facility-level interpatient hemoglobin variability in hemodialysis centers participating in the Dialysis Outcomes and Practice Patterns Study (DOPPS): Associations with mortality, patient characteristics, and facility practices. Am J Kidney Dis 2011;57:266-75.
Ramirez SP, Albert JM, Blayney MJ, et al. Rosiglitazone is associated with mortality in chronic hemodialysis patients. J Am Soc Nephrol 2009;20:1094-101.
Tentori F, Albert JM, Young EW, et al. The survival advantage for haemodialysis patients taking Vitamin D is questioned: Findings from the Dialysis Outcomes and Practice Patterns Study. Nephrol Dial Transplant 2009;24:963-72.
Lopes AA, Tong L, Thumma J, et al. Phosphate binder use and mortality among hemodialysis patients in the Dialysis Outcomes and Practice Patterns Study (DOPPS): Evaluation of possible confounding by nutritional status. Am J Kidney Dis 2012;60:90-101.
Ronald L Pisoni
Arbor Research Collaborative for Health, 340 East Huron Street, Suite 300, Ann Arbor, MI 48104
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]