Background: It is essential to adjust for the
level of preoperative pain and functional status when measuring
the outcome of total knee arthroplasty. Some study designs rely
on postoperative patient recall to derive preoperative status. In
this study, we compared prospectively collected preoperative data
with data derived from patient recall of preoperative status three
months after total knee arthroplasty.
Methods: Patients were recruited as part of a prospective
observational study of the outcome of primary total knee arthroplasty
for osteoarthritis at four centers in the United States, six centers in
the United Kingdom, and two centers in Australia. Independent research
assistants recruited patients and collected data with use of a uniform
documentation system preoperatively and three months postoperatively.
Preoperative data included the findings of a clinical history and
physical examination, demographic information, socioeconomic status,
and scores from two health-status instruments: the Western Ontario
and McMaster University Osteoarthritis Index (WOMAC) and the Medical
Outcomes Study Short Form-36 Health Survey (SF-36).
Postoperative data included the WOMAC and SF-36 scores
and patient recall of preoperative status on selected items from
these health-status measures.
Results: A total of 862 patients were recruited,
and recall data were available for 770 patients (89%).
The mean age was seventy years (range, thirty-eight to
ninety years), and 59% of the patients were women. Comparisons
of prospective and recall data on individual pain and function items
showed poor-to-fair agreement (weighted kappa, 0.20 to 0.41). Patients recalled
significantly more pain than they had reported preoperatively (p < 0.001),
but there were random recollection errors for the function items.
There was only moderate correlation between the prospective and
recalled summary scores for pain (Spearman r = 0.53) and
function (Spearman r = 0.48). In addition, 61% of
the recalled pain scores and 50% of the recalled function
scores differed from the prospective scores by more than 10 points
(10% of the total range).
Conclusions: Patients’ recall of preoperative
pain and functional status three months after total knee arthroplasty
demonstrated only moderate agreement with what the patients had
reported prospectively. Researchers who use recall data to derive
preoperative status must recognize these limitations when drawing conclusions
about the effectiveness of total knee arthroplasty.