Adding clinical notes to radiographs increases the consistency of clinical study

In clinical practice, orthopaedists commonly use radiographs and clinical notes to determine whether a fracture has healed. But in orthopaedic trials, outcomes assessors often only use plain radiographs to evaluate the extent of healing.

When Mohit Bhandari, MD, and his fellow researchers conducted a clinical trial in Canada that evaluated healing among patients with tibial shaft fractures, they found that reviewers revised their initial impression regarding extent of healing in a significant number of cases after reading clinical notes with radiographs.

“We found that the addition of clinical notes changed the outcome decision in a substantial number of cases, which suggests that the use of plain films in isolation may have limitations,” said Dr. Bhandari, who presented the results of his study at the 2010 Annual Meeting.

“Our findings suggest that orthopaedic trials should consider adding clinical notes to radiographs in studies of fracture healing to enhance the ability to make generalizations based on the results,” he added.

Evaluating fracture healing
From July 2005 to June 2007, 51 patients (39 males and 12 females; mean age, 39.5 years) with 14 open and 37 closed tibial fractures participated in a multicenter, clinical trial.

All of the patients were treated with reamed intramedullary nailing. Researchers conducted follow-up visits at 6 weeks, as well as at 3, 4, 5, 6, 9, and 12 months to take standardized anteroposterior and lateral radiographs. In addition, clinical examinations were performed, and clinical notes from the visit were obtained.

“Fracture healing was assessed centrally and independently by three blinded orthopaedic trauma surgeons,” said Dr. Bhandari. “They evaluated whether the fracture was healed based on a general review of serial radiographs alone. In the study, radiographic healing was defined as ‘full bridging of at least 3 out of 4 cortices.’”

After the surgeons reached a consensus regarding fracture healing based on the follow-up radiographs, they reviewed clinical notes from each follow-up visit.

“Based on the combination of clinical notes and radiographic information, each adjudicator maintained or revised the date of fracture healing, and the group achieved consensus through discussion,” said Dr. Bhandari.

Notes changed decisions
Researchers found that clinical notes changed the adjudicators’ consensus in 19 patients (Table 1). Of these changed decisions, 9 were altered to support an earlier time to healing, and 10 were changed to support a later time to healing. Most of the changed decisions occurred at 3-month follow-up.

Clinical notes that indicated significant pain and failure to resume weight bearing, work, or activities were most likely to cause a change in decisions from “healed” to “not healed.” When decisions were changed to “healed,” clinical notes indicated that the patient’s pain was resolving, and/or weight bearing, work, or activities were resumed.

“We found that clinical notes of follow-up visits with unchanged decisions contained fewer positive or negative comments than clinical notes that changed decisions,” said Dr. Bhandari.

A potential limitation of the study was that many follow-up visits had to be excluded because of insufficient data on radiographs and/or clinical notes due to patients’ poor compliance with subsequent follow-up visits after they were declared radiographically healed.

The study has a number of methodological strengths, continued Dr. Bhandari, including that the same adjudicators decided on both the radiographic as well as the combined radiographic and clinical fracture healing time, which “provides reassurance that the shift in time to healing was solely due to the additional information from clinical notes.”

Dr. Bhandari concluded that “comments in clinical notes regarding weight-bearing ability, pain, and return to work or activities, which an adjudicator could not possibly obtain from radiographs alone, appear to have influenced their decisions.

“The adjudication process should be consistent among trials to increase comparability of multiple studies’ results,” concluded Dr. Bhandari. “We recommend including comments on weight-bearing ability and pain to maximally approximate the clinical information available in clinical practice.”

Dr. Bhandari’s coauthors on “The Impact of Clinical Data on the Evaluation of Tibial Fracture Healing – TRUST Investigators” were Bernadette G. Dijkman, BSc; Jason W. Busse, DC, PhD; Stephen D. Walter, PhD; and Gordon H. Guyatt, MD, MSc. Dr. Bhandari reports the following conflicts: Stryker; Smith & Nephew; Amgen Co; Pfizer; Baxter; King Pharmaceuticals; Wyeth; Canadian Institutes of Health Research; Canadian Orthopaedic Foundation; AO; DePuy, A Johnson & Johnson Company; and National Institutes of Health (NIAMS & NICHD).

Disclosure information for coauthors: Ms. Dijkman and Dr. Walter report no conflicts. Dr. Busse reports the following conflict: Smith & Nephew. Dr. Guyatt reports the following conflicts: Eli Lilly and American Medical Association Publications.

Jennie McKee is a staff writer for AAOS Now. She can be reached at mckee@aaos.org