Trial of Feedback on Blood Use

Trial of Feedback on Blood Use (TOFU)

Background \& Rationale Many studies have demonstrated substantial variability in blood use among both physicians and hospitals, suggesting the need to standardize transfusion practices and reduce unnecessary transfusions.To that end, national regulatory agencies in many countries require that hospitals monitor local blood use. However, the utility of this monitoring is unclear. A small number of prospective studies aimed at improving physician transfusion behavior have been published, but well-supported, generalizable findings are currently lacking. The proposed study will test the hypothesis that providing feedback of individual blood use data to transfusing physicians will reduce unnecessary RBC transfusions. A limited number of studies conducted in a range of settings have examined whether audit and feedback can improve physician practice and clinical outcomes. Results have been mixed. The proposed study would represent the first attempt to rigorously assess the impact of audit/feedback on changing transfusion practice.

The recently completed FOCUS trial presents a unique opportunity in that it provides high-quality data on which to base transfusion practice recommendations for a specific patient population. In the FOCUS trial, 2016 adult patients with a history of or risk factors for cardiovascular disease undergoing surgical hip fracture repair were randomized to be transfused with RBCs for a Hgb <10 g/dL (liberal group) or for symptoms of anemia and/or a Hgb <8 g/dL (restrictive group). The primary end point was the ability to ambulate independently at 60 days, and key secondary endpoints included 30-day and long-term mortality, myocardial infarction, and stroke. No significant difference was observed in any end point, demonstrating that -- as in the earlier TRICC trial -- a restrictive strategy was as safe as a liberal transfusion strategy. An earlier multicenter RCT similarly showed no difference in the rate of silent myocardial ischemia among 260 knee and hip surgery patients randomized to be transfused for a Hgb <8 g/dL or <10g/dL. With no demonstrable benefit of transfusing at a higher Hgb, and given the high costs and multiple risks of transfusion, adoption of the lower (8 g/dL) transfusion trigger in hip surgery patients seems logical. A small preliminary study was conducted at 9 BEST institutions. Each participating institution contributed data on a minimum of 10 consecutive patients who underwent total hip replacement surgery during 2010-11. Out of 186 patients, 63 (34%) were transfused postoperatively. 41/122 (34%) of the postoperative RBC units were transfused at a pretransfusion hemoglobin of > 8 g/dL, suggesting that a practice improvement opportunity exists.

Research Question In orthopedic surgeons performing adult elective hip surgery, does providing individual feedback data on transfusion practice in the context of the FOCUS trial data reduce postoperative RBC utilization compared with providing no feedback?

Primary Hypothesis Providing individual feedback on transfusion practice to orthopedic surgeons performing adult hip arthroplasty will reduce elective RBC transfusions in the postoperative period.

Study Population:

The unit of study is orthopedic surgical attending physicians who perform hip arthroplasty.

There is no restriction by underlying causal orthopedic factors (e.g. arthritis or traumatic fracture), or whether surgery is primary or secondary/revision.

Clinical Trial Design:

Study type: Two-arm, multicenter, cluster-randomized controlled clinical trial.

Subjects: Each cluster will be comprised of the set of orthopedic surgeons that perform hip surgery at a single study site.

Procedures: Baseline blood use in hip arthroplasty cases will be assessed for each surgeon for the 3-month period just prior to the start of the intervention phase. Baseline blood use measurements on each surgeon will consist of:

1. Hip arthroplasty cases (n).
2. Total postoperative RBC units (n) transfused to hip surgery patients. (Here, "postoperative" is defined as the interval beginning the day after surgery ("postoperative day 1") and ending when the patient is discharged from the hospital or at postoperative day 14, whichever comes first.
3. Pretransfusion hemoglobin levels.

Baseline characteristics for each orthopedic center will also be collected:

Facility type (e.g. academic hospital, community hospital) Orthopedic surgeons performing hip surgery (n) Annual number of hip surgery procedures (n)

Prior to randomization, the site PI will provide a brief educational presentation to the orthopedic surgeons at each site. The presentation will consist of a standardized 10-minute PowerPoint presentation reviewing the data from the FOCUS study and the recommended RBC transfusion trigger of 8 g/dL (or symptomatic anemia) based on that data. The presentation will also cover published data on hemoglobin levels and postoperative rehabilitation/quality of life measures8. The surgeons will be advised that a study of blood use in hip surgery will be commencing, and that they may be randomized to receive feedback about their individual postoperative. Clusters of orthopedic surgeons will then be randomized to either the Control arm (no feedback) or Intervention arm (monthly feedback).

The intervention phase will be conducted for 12 months. Surgeons randomized to the Intervention arm will be given feedback at the end of each month. The feedback will be provided in the form of a standardized email report sent by the site PI to each surgeon individually. The monthly feedback report will show:

The number of hip surgery cases performed and postoperative RBCs transfused by each surgeon for the month.

The percentage of postoperative RBCs transfused at a Hgb above or below 8 g/dL by each surgeon for the previous month.

The percentage of postoperative RBCs transfused at a Hgb below 8 g/dL by each surgeon, trended over time.

In the reports, surgeon names will be anonymized (e.g. Surgeon A, B, C etc.) However, individual surgeons will be told which data is their own, so that they may compare themselves with their peers. No blood utilization data will be reported to the Control arm surgeons.

Measurements: For data collection purposes, each attending orthopedic surgeon will be assigned a letter code. Surgeon identity will be kept confidential within each site, and will not be shared with other sites or the Data Coordinating Center.

Data collection: The following information will be collected for each patient:

Patient Study Number Age/Gender Surgical procedure (primary or reoperative hip replacement, hip fracture) Surgeon study number Date of surgery Date of discharge Postoperative RBC transfusions (number of units and date) Pretransfusion Hgb levels (g/dL) Hgb (g/dL) at admission/discharge

Data will be collected using electronic CRFs completed by each site. If desired, sites may first record data on an optional paper CRF, then upload the data electronically. The feedback reports will be generated by a central data coordinating center. The feedback reports will be emailed to each site PI monthly for subsequent email distribution to the local orthopedic surgeons. Surgeon names and patient identifiers will be maintained locally at each site and will not be shared with the data coordinating center or with other study sites. Read receipts will be used for emailed reports to the surgeons; site PIs will track the number of feedback reports read or not read each month by surgeon.