Real-Life Example of Pragmatic Clinical Trials in End-Stage Renal Disease: The TiMe Trial Michel Chonchol, MD Professor of Medicine University of Colorado Denver Aurora, CO, USA Randomized Controlled Trials in Nephrology Palmer S et al. Am J Kidney Dis 2011; 58:335-337 Clinical Practice Guidelines KDIGO Clinical Practice Guideline for the Diagnosis, Evaluation, Prevention, and Treatment of Chronic Kidney Disease-Mineral and Bone Disorder (CKD-MBD) Strength of recommendation RECOMMENDATIONS 1 = strong 5D, we suggest lowering elevated phosphorus levels the normal range (2C). 2 =toward weak • 4.1.1In patients with CKD stages 3-5, we suggest maintaining serum phosphorus in the normal range (2C). In patients with CKD stage • 4.1.2In patients with CKD stages 3-5D, we suggest maintaining serum calcium in the normal range (2D). Quality of evidence mEq/l) (2D). A= high 4.1.4In patients with CKD stages 3-5 (2D) andB= 5D (2B), we suggest using phosphate-binding agents in the treatment of moderate hyperphosphatemia. It is reasonable that the choice of phosphate binder takes into account CKD stage, presence of other components of CKD-MBD, concomitant therapies, and side-effect (not graded). Cprofile = low 4.1.5In patients with CKD stages 3-5D and hyperphosphatemia, we recommend restricting the dose of calcium-based phosphate binders D in=thevery and/or the dose of calcitriol or vitamin D analog presencelow of persistent or recurrent hypercalcemia (1B).In patients with CKD • 4.1.3In patients with CKD stage 5D, we suggest using a dialysate calcium concentration between 1.25 and 1.50 mmol/l (2.5 and 3.0 • • stages 3-5D and hyperphosphatemia, we suggest restricting the dose of calcium-based phosphate binders in the presence of arterial calcification (2C) and/or adynamic bone disease (2C) and/or if serum PTH levels are persistently low (2C). • 4.1.6In patients with CKD stages 3-5D, we recommend avoiding the long-term use of aluminum-containing phosphate binders and, in patients with CKD stage 5D, avoiding dialysate aluminum contamination to prevent aluminum intoxication (1C). • 4.1.7In patients with CKD stages 3-5D, we suggest limiting dietary phosphate intake in the treatment of hyperphosphatemia alone or in combination with other treatments (2D). Learning Health System Every clinical care encounter is an opportunity for learning and for producing generalizable knowledge • Favorable developments ‒ Consolidation of health systems ‒ Increased use of electronic health records • Goal is to move beyond observational studies and conduct clinical trials within the clinical care setting NIH Health Care Systems Research Collaboratory • “The overall goal …. is to strengthen the national capacity to implement cost-effective large-scale research studies that engage health care delivery organizations as research partners.” • RFA Requirements – Conduct a large pragmatic trial demonstration project – Perform trial in at least two health care systems – Rely on hard clinical outcomes and/or patient reported outcomes – Ascertain outcomes from clinically available data Pragmatic Trial Does this therapy work under usual care conditions? Explanatory versus Pragmatic Trials • Explanatory trials evaluate intervention under ideal, experimental conditions in order to test a causal hypothesis (assessment of efficacy) • Pragmatic trials evaluate intervention under “real-world” conditions to inform choices between treatment options (assessment of effectiveness) • Tradeoff between high internal validity (explanatory) and high generalizability (pragmatic) Most trials fall somewhere in between two extremes (continuum rather than dichotomy) PRECIS Criteria (Pragmatic-Explanatory Continuum Indicator Summary) Explanatory Pragmatic Eligibility Criteria • Restrictive: highest risk for outcome, most likely to respond, most likely to comply • All individuals with condition of interest regardless of risk, comorbidities, adherence, language Intervention (& comparator) • Strict implementation • Expert practitioners • Close monitoring of dose, adverse effects with adjustment or treatment, respectively • • • • Follow-up • High intensity • More f/u than usual care • Data collection for trial • Low intensity • No study visits • Administrative databases Thorpe KE J Clin Epidemiol 2009 Flexible implementation No expertise needed Full range of clinical settings Comparator is often usual practice PRECIS Criteria Explanatory Pragmatic Outcome • Direct and immediate consequence of intervention • May be surrogate of downstream outcome • Specialized training for ascertainment • May require adjudication Adherence to intervention • Close monitoring of • Unobtrusive or no participants and centers measurement • Adherence may be • No strategies to improve requirement for participation adherence • Strategies employed Analysis • Intention to treat often supplemented by per protocol • Attempt to answer narrowest, mechanistic question Thorpe KE J Clin Epidemiol 2009 • • • • Clinically meaningful Objectively measured No adjudication Assessed under usual conditions • Pure intention to treat • Noise is accepted (embraced?) Features of Pragmatic Trials • Evaluate “simple” interventions • Require large sample • Often use cluster randomization • Use simplified approach to consent • Use electronic health records (recruitment, data collection, communication, monitoring) • Relatively inexpensive Pragmatic Trials are Not New 1950s: Polio vaccine 1980s - 1990s: Acute treatment for MI Hemodialysis is an Ideal Setting for Pragmatic Trials • Highly accessible study population with frequent clinical encounters • Highly granular and uniform data collection as part of routine clinical care • Infra-structure allows for centralization of activities and ability to conduct trial in large number of facilities across a large geographic area Dialysis Facility is the Principal Source of Health Care for Many Patients with ESRD • Patients have frequent contact with multidisciplinary team members • Dialysis facilities perform/provide laboratory studies, blood pressure measurements, QOL assessments, vaccinations, nutritional supplements, pharmacy services • Burdensome for patients to go elsewhere for care • Primary care providers often relinquish care to nephrologists and dialysis unit personnel Nespor SL ASAIO 1992; Holley JL AJKD 1993; Bender FH AJKD 1996; Zimmerman DL, NDT 2003; Shah N, Int Urol Nephrol 2005; Nissenson AR AJKD 2012. Dialysis is Already a Learning Health System • United States Renal Data System (USRDS) • Dialysis Outcomes and Practice Patterns Study (DOPPS) • Dialysis provider organization data • Quality improvement initiatives But very little data from randomized clinical trials! Hemodialysis Session Length in the 9-15 hr/wk range in 3/week schedules (Rationale behind the TiMe Trial) Dialysis Session Length for a 3/week Schedule • Dialysis used to be given for 6-8 hours per session, 3x/week • NCDS study suggested that Kt/V was a key measure of adequacy • More efficient dialyzers, higher blood flow rates • Kt/V > 1.2 can be given to small patients and women over about 3.0 hours • Is session length per se important beyond Kt/V? • KDOQI – at least 3 hours; European BPG – at least 4 hours • What is the evidence for session length per se in the 9-15 hr/week range and how strong is it? “Physiologic” Rationale to Increase Time per se • Urea not representative uremic solute • Phosphate, middle molecules, protein-bound substances may depend more on weekly time • Urea not a good model solute • Surface-area scaling results in longer dialysis times for small patients and women • Volume control and ultrafiltration rate depend on time. • Shorter time patients may be more volume overloaded • Shorter-time patients may have higher UF rate Lines of Evidence for Benefits of Longer Session-Length • Randomized trials (NCDS) • Cross-country comparisons (US vs. Europe vs. Japan) • Within-country observational datasets • Volume overload and hard outcomes • High UF rates and hard outcomes Time in 3/week Dialysis Markedly different in Japan vs. Europe vs. N. America Tentori et al, NDT, 27:4180, 2012 Is The Session Length Data From Japan Based On Biology? Shinzato T et al, NDT, 1996;2139-2151 Should minimum td > 4 hours? Miller et al, AJKD, 2009 Is There Something Special About 4.0 Hours? What About 3.5? Flytheet al, Kidney Int. 2013; 83:104-113 Time > 4 hrs and Survival 3 Different Methods; 3 Different Answers Brunelli et al, Kidney Int. 2010; 77:630-636 Dose-targeting bias Daugirdas, Kidney Int 2013;83:9-13 High UF Rate and Mortality Flythe J et al, Kidney Int., January, 2011 Movilli E et al, NDT 2007, 22:3547. Summary • Some evidence from randomized trials (primary analysis of NCDS that time per se may be important) • May be due to surface-area-related scaling of solute removal • May be due to better volume control • Observational data is all over the map; country bias, dose-targeting bias (DTB), confounding with body size • Results in the US are not consistent; more prominent effect at 4 hours than at 3.5, suggesting DTB. TiME Trial Team Academic Investigators Fresenius TiME University of Pennsylvania DaVita NIDDK, OD TiME Trial: Team Effort from the Start • Trial question • Trial design • Grant application • Protocol development • Trial conduct Time to Reduce Mortality in End-stage Renal Disease (TiME) Trial Hypothesis Thrice weekly hemodialysis with session durations of at least 4 hours and 25 minutes improves outcomes compared with usual care. TiME Trial Design Usual Care Facilities Enroll and Randomize Facilities (session duration not driven by trial) Enroll and follow incident patients Intervention Facilities ≥4.25 hour sessions Primary outcome: All-cause mortality Secondary outcomes: Hospitalizations & Quality of Life Follow-up Period: 2-3 years Rationale for Cluster Randomization • Logisitical efficiencies around recruitment • As importantly, greater ability to implement intervention if it is a facility approach Eligibility Criteria Facility • Willingness of nephrologists and facility leadership • Capacity to accommodate 4 hr, 25 minute treatments for incident patients Patient • Age >18 years • Ability to provide consent for dialysis care Sample Size and Power • 402 facilities, 6432 patients (approximately 50% from each dialysis provider organization) • Average of 16 patients per facility • 80% power for HR 0.85 • Assumptions • Mortality rate 18% per year • Intra-class correlation for mortality of 0.03 • 5% loss to f/u per year Stratified Randomization • Central venous catheters • Black race >20% vs ≤20% >50% vs ≤50% Data Acquisition • Clinical and administrative data transmitted electronically from individual facilities and centralized laboratory to dialysis providers’ data warehouses • De-identified data elements transmitted from dialysis provider data warehouses to Data Coordinating Center (Penn) Pragmatic Features of the TiME Trial • All patients starting dialysis are eligible unless they are not competent to sign consent to clinical care • Intervention is delivered by clinical providers • Outcomes: • ascertained from routine clinical data • derived from data elements common to all sites • Adherence to intervention at the patient level will be promoted using systems already in use • Highly centralized implementation approach • Single IRB of record • Testing effectiveness rather than efficacy Approach to Consent • Patients starting dialysis at participating facilities are given a brief information document with: ‒ Purpose of the trial ‒ Effect of trial on patient’s session duration ‒ Toll-free telephone number to obtain additional information from the research team and to opt-out of participation • Trial information and toll-free telephone number to contact research staff are posted in patient care areas of participating dialysis facilities throughout the duration of the trial. Why Use Opt-Out Approach? • Goal is to evaluate effectiveness rather than efficacy • Implementation of intervention will be more successful if facility-wide approach • Consent requirement will result in imbalance in patient characteristics between treatment groups since randomized assignment is known prior to enrollment Criteria for Waiving Consent 1. The research involves no more than minimal risk to the subjects 2. The waiver or alteration will not adversely affect the rights and welfare of the subjects 3. The research could not practicably be carried out without the waiver or alteration 4. Whenever appropriate, the subjects will be provided with additional pertinent information after participation. Key Factors for Minimal Risk Determination • Physician and patient autonomy are maintained • Research does not change care for patients in Usual Care arm (which is hypothesized to have inferior outcomes) Challenges • Ethical and regulatory hurdles regarding consent approach • Multiple layers of buy-in required • Facility personnel education • Potential for contamination across Intervention and Usual Care facilities • Changing practices over conduct of trial • Patient acceptance • Physician acceptance Ethical and Regulatory Vetting • IRB • Collaboratory / NIH • OHRP • FDA • CMS Down side of using waiver of consent • Less able to push adherence to intervention at patient level Update Year 1 IRB approval Office of Human Subjects Protections review FDA review, exempt from IDE regulations/oversight Contracts signed DSMB review IT systems established Years 2 - 5 Facility enrollment and randomization Patient enrollment and follow-up Many Unanswered Questions in Dialysis about Fundamental Aspects of Care • Duration of hemodialysis sessions? • Dialysis solution potassium concentration? • Blood pressure target? • Phosphorus target? • Hemoglobin target? • Preventive health care? • Anticoagulation for atrial fibrillation?