Background: Anchoring is a cognitive bias for an individual to rely too heavily on an initial piece of information offered (known as the "anchor") when making decisions and estimates. The anchoring effect, while understood to influence judgement in multiple domains outside of medicine, has not been extensively studied in the context of ED patients. We conducted a prospective study to test the null hypothesis that there would be no difference between pain scores for ED patients anchored to lower vs. higher numbers.
Methods: This was a prospective, randomized, interventional trial at an academic, urban ED. A convenience sample of patients age >17 years presenting to the ED with complaints of pain that were located in the head, abdomen, chest, extremities, back, and/or generalized, respectively, were included. Participants were randomized to receive a written survey with either an anchor number of “2” (G2) or “9” (G9) and were asked if their pain score for this visit was “greater than” or “less than” the anchor number provided on a visualized analog scale. Categorical data are presented as frequency of occurrence and analyzed by chi-square; continuous data are presented as mean(SD) and analyzed by t-tests. The primary outcome parameter was to compare mean pain scores between G2 and G9.
Results: 522 patients were eligible, 31 patients declined to participate, study group 491; 55% female, 50% age >54 years, 56% Hispanic, 61% income < $20,000, 29% high school, 40% private insurance, 22% abdominal pain, 14% chest pain, 10% extremity. 235(48%) patients were randomized to the G2 condition. Groups were similar with respect to the following characteristics: % female (50% vs 58%; p=0. 06), age < 55 years (49% vs 51%; p=0. 68), Hispanic ethnicity (59% vs 54%; p=0. 24), income < $20,000 (62% vs. 60%; p=0.59), < high school (31% vs. 27%; p=0.27), private insurance (39% vs. 41%; p=0.61), and type of chief complaint (p=0. 06). Mean pain scores G2 vs G9 (6. 31+/-3. 09 vs. 6. 55+/-3. 23; p=0. 41) were similar with respect to both statistical and clinical significance.
Conclusion: Patients in our study group did not show evidence of a significant anchoring effect on their reports of pain scores.