Subjects and Methods

A HIPAA-compliant institutional review board–approved prospective single-institution nonrandomized survey study without blinding was completed between December 2010 and June 2011 at our institution. We used a non–psychometrically validated original survey instrument because no validated survey existed at the time of the study (Appendixes 1 and 2). Adult (older than 18 years) English-speaking outpatients undergoing CT of the torso, abdominopelvic CT, or nonobstetric, nonvascular laboratory ultrasound examinations (to be interpreted by our abdominal radiology section) were eligible to participate. Patients undergoing imaging examinations for evaluation of known malignancy and those with musculoskeletal indications were excluded. Those undergoing screening examinations for malignancy were eligible.

Patients were recruited Monday through Friday on the day of their radiology examination at one of two sites in our radiology department where outpatient CT and ultrasound examinations are performed. Day of the week of enrollment was based on the availability of participating investigators when not assigned to clinical duties. The outpatient schedule and submitted history were available for review, and one of three board-certified abdominal radiology fellows determined eligibility, obtained consent before the examination, and conducted the review session with the patient.

Consenting patients who met the eligibility criteria were prospectively but not consecutively enrolled. To avoid bias regarding knowledge of a radiologist, name badges were inverted, and coinvestigators introduced themselves only as a member of the radiology department conducting a study. Before undergoing an examination, patients who enrolled completed the preexamination survey (Appendix 1), which was checked for completeness by the investigator.

The radiologic examinations were completed according to our normal department protocol. All ultrasound examinations were performed by an onsite sonographer, who reviewed the study for completeness with the onsite radiologist or radiology fellow. Real-time scanning by a radiologist was performed at the reviewing radiologist’s discretion (for study patients this process was completed by the participating investigator reviewing the examination). For CT examinations, the CT technologist paged the investigator when the examination was complete.

The time from completion of the examination to investigator-patient review was recorded. An attending radiologist was available to review the examination before the investigator-patient consultation when the fellow deemed such a review necessary, and whether such a review took place was recorded. Ordering providers were contacted before or after the investigator-patient consultation when deemed necessary according to our department protocol (e.g., new malignancy detected or other unexpected or critical finding that would warrant immediate action by the ordering provider).

After initial review of the examination by the participating investigator, patients were escorted to a private reading area where the examination images were reviewed directly by the investigator with the patient on the PACS display system. Basics of the examination, overview of patient anatomy, and any pertinent findings were discussed. Any questions raised were answered to the best ability of the investigator and referred to the ordering physicians when deemed appropriate. Patients were given the preliminary results orally and were reminded to follow up with their ordering provider, who would receive a final interpretive report, according to department protocol. The duration of the encounter was recorded, and patients completed the second survey (Appendix 2) at the conclusion of the consultation session. All examinations were reviewed by an attending radiologist with the participating fellow on the day of the examination, and all final reports included a statement that a preliminary report was provided directly to the patient.

The final reports of all examinations in the study were reviewed after study closure by one board-certified abdominal radiologist with more than 20 years of experience, who did not participate in any of the consultation sessions. This radiologist classified the importance of the findings discussed in the consultation with the patient. A modified categorization system based on the CT colonography extracolonic reporting and data system was used. Category 1 included normal findings and anatomic variants; category 2 included clinically unimportant findings (e.g., simple cyst); category 3 included findings likely unimportant but incompletely characterized (e.g., complex renal cyst, thyroid nodule, pancreatic cyst, fatty liver); and category 4 included potentially important findings. Category 4 findings were communicated to the referring physician on the basis of either American College of Radiology guidelines or the pertinence of the finding according to the indication for the examination (e.g., potentially malignant mass, acute diverticulitis in patient with abdominal pain) 21. In addition, a limited cost analysis was performed by completion of a retrospective analysis of all billed procedures and CT and ultrasound examinations interpreted over a 20-day block (4 weeks) during April and May 2011 for one of the participating fellows during normal clinical practice while working on our CT and ultrasound services. No patients were enrolled by the fellow during this time. Current Procedural Terminology codes for each examination were extracted, and relative value units assigned for a calculation of estimated revenue completed by our billing department. Revenue per minute of fellow time was calculated on the basis of our standard 8-hour workday.

Statistical analysis was completed with SAS software (version 9.1.3, SAS Institute). The Fisher exact test was used to assess for statistically significant differences in responses with respect to examination type, sex, and race. The Wilcoxon signed rank test was used to assess for within-subject differences in anxiety level before and after the consultation session, and the Wilcoxon rank sum test was used to compare wait time and encounter duration. The McNemar test was used to assess for significance in patient preference about how to receive test results (question 5 on both surveys). A value of p < 0.05 was used to determine statistical significance.

Results

Of 108 patients approached for enrollment, 86 (80%) agreed to participate (43 men, 43 women; mean age, 52 years; range, 26–83 years). Of these, 37 underwent CT and 49 underwent ultrasound examinations. There was no statistically significant difference in any of the results stratified by sex or examination type (CT vs ultrasound). Sixty-four patients (74%) described themselves as white, 10 (12%) as black or African American, four (5%) as Asian, and five (6%) as other. By race, only preexamination survey question 5, regarding preference about how patients would prefer to hear results, and postconsultation survey question 3, regarding the option of reviewing future examinations with the radiologist, had statistically significant differences (p = 0.01, p = 0.03) (Table 1). Race was not a statistically significant variable (p = 0.7) in the reassessment of patient preferences about how they would prefer to hear results after they completed the radiologist consultation session (postconsultation survey question 5).

Among the 86 patients enrolled, 77 (90%) stated they fully understood the reason for the examination, and nine (10%) stated they partly understood. In 85 cases (99%), the reason for the examination stated by the patient and the reason listed on the order history were deemed concordant by the participating investigator. In the assessment of knowledge of a radiologist, 48 patients (56%) correctly identified a radiologist as a physician who interprets images; 33 (38%) believed a radiologist was a technician; four (5%) believed a radiologist was a nurse; and one (1%) did know what a radiologist was.

After consultation, there was a statistically significant change (p = 0.03) in patients’ stated preferences on how they wanted to hear results. Before the consultation session, 70 patients (81%) indicated a preference to hear results from both the ordering provider and the radiologist; this number increased to 78 (91%) after consultation (Table 2). Six of the eight patients who responded on the preexamination survey that they wanted to hear results from the primary care or ordering provider changed their preference to hearing results from both the radiologist and the primary care or ordering provider.

Before the examination, 84 patients (98%) indicated they would be comfortable hearing normal or abnormal results, and two (2%) wanted to hear only normal results. After consultation, these numbers changed to 85 (99%) and one (1%) (p > 0.05). As Figure 1 shows, 99% of patients agreed or strongly agreed that reviewing their examination and results with the radiologist was helpful or beneficial. Results for assessing of the option of reviewing future examinations on the same day as the examination even if it meant waiting longer for the results (postconsultation survey question 3) yielded the following: 71 of the 86 patients (83%) reported a desire to always review future examinations, 13 (15%) wanted to review results but only when they asked, and two (2%) did not want to review future examinations. Seventy-seven patients (90%) stated having the option to review future examinations with the radiologist would make them more likely to return to our institution for their examinations; the other 10% stated it would have no effect on where they would go.

Patient anxiety levels before and after consultation exhibited a statistically significant change (p < 0.0001), decreasing in 41 of the 86 patients (48%), increasing in 13 (15%), and remaining the same in 32 (37%). The mean reported anxiety score decreased from 2.5 to 1.9 after consultation. No statistically significant differences were found in the correlations between anxiety level or change in anxiety level and examination result category, race, sex, examination type, number of previous visits with the ordering provider, or appointment length. The CT and ultrasound results were assigned category 1 for 18 of the 86 patients (21%), category 2 for 32 patients (37%), category 3 for 29 patients (34%), and category 4 for seven patients (8%). No statistically significant correlation was found between findings category and any of the survey questions, examination type, race, sex, or change in anxiety level after consultation. The participating fellow needed the attending radiologist for consultation before reviewing the examination with the patient in seven cases (8%); four of these cases were assigned category 4 findings (p < 0.001). In the seven examinations listed as category 4 (potentially important or pertinent to examination), one patient had an increase in reported anxiety, two remained the same, and four reported decreased anxiety after consultation with the radiologist.

Table 3 shows the results of preexamination survey questions 1 and 2, which detailed the number of previous visits and length of the most recent visit with the ordering provider. There was no statistical correlation between the number of previous appointments with the ordering provider and patient preference regarding from whom they preferred to hear results. Seventeen of the 86 patients (20%) stated they had a previous radiologic test discussed with them (preexamination survey question 6), ultrasound (chosen by 10 patients) and radiography (chosen by eight patients) being the most common. After a CT examination, the average wait time was 9.9 minutes (range, 1–21 minutes) and the consultation time 10.4 minutes (range, 3–22 minutes). After an ultrasound examination, the average wait time was 1.2 minutes (range, 0–10 minutes) and the average consultation time 7.1 minutes (range, 2–19 minutes). No statistically significant correlation was found between duration of radiologic consultation and duration of last appointment, number of previous appointments with the ordering provider, or participating in a review of a radiologic examination in the past. No significant differences in wait or encounter time were noted when the data were stratified by examination finding category or need for consultation with an attending radiologist.

Estimated revenue based on relative value unit analysis of all billed CT and ultrasound procedures and interpreted diagnostic examinations over 4 weeks (20 work days) for one participating fellow yielded $29,254 of estimated revenue. In our typical 8-hour workday, this amount would yield approximately $1463/day or $3.05/minute of fellow time. The estimated revenue of $3.05/minute of clinical time and average consult time of 10.4 minutes for CT and 7.1 minutes for ultrasound yielded an estimated cost of $31.72 and $21.66 per examination.

Discussion

In radiology, the degree of direct patient contact is extremely variable, depending on the type of examination, degree of involvement of the interpreting or operating radiologist, and the geographic region of practice 8. Restructuring the method of delivering results to patients may enhance the daily practice of radiology. The value of a radiologist also may be augmented through direct patient communication 9. In addition, a blended practice pattern of establishing patient relationships via direct communication of results and imaging review may help alter patient behavior 22 23. Arguments for increased interaction with patients and direct communication have ranged from patients’ desire to hear results after examinations, legal and ethical responsibility, and increasing promotion of the field of radiology and its role in healthcare 5–7 9 18. Direct communication of results by radiologists may also improve patient care by closing the communication loop 6 7 19 24. Previous analysis of direct electronic communication with referring physicians regarding important but nonurgent findings encountered during examination interpretation showed that 15.4% of these notifications went unread 24. In addition, increased direct communication may help improve public awareness of radiologists. Our data showed only 56% of patients recognized a radiologist as a physician who interprets imaging examinations. This finding is similar to that in survey work conducted by the American College of Radiology in 2008, which showed only 48% of respondents knew that a radiologist was a licensed physician who interprets results of imaging tests 22.

Many opposed to direct communication of results have raised important points ranging from increased legal liability, possible harm by communicating results without full patient history and results, lack of an established relationship with the patient, lack of time due to high clinical volume (especially in private practice), increasing work for referring physicians, and discomfort in communicating bad news such as a new diagnosis of malignancy 4 6 7 10 13 25. A survey conducted in 1992 13 showed that as results become more complex or severe, both the radiologist’s desire and the referring physician’s acceptance of result disclosure by the radiologist decrease.

The preferences of patients receiving care in the United States have been investigated in previous survey studies, which showed a preference for direct disclosure of results by the interpreting radiologist for both mammography and other imaging examinations 1 4 14 17. Our results differ from those in previous studies in that 91% of our patients expressed a preference for hearing results from both the radiologist and ordering provider after completing our consultation session. This percentage is much higher than that in another recent survey-only study 1, in which 14% of patients chose this method for normal results and 27% for abnormal results. Our data also showed a much higher rate of acceptance (99%) after the consultation session for hearing both normal and abnormal results from the interpreting radiologist. This discrepancy likely is related to closer to uniform acceptance once a patient meets the radiologist and reviews the examination in person, as in our study.

Our results revealed a statistically significant change in reported anxiety levels, most patients reporting decreased anxiety after radiologic consultation. A trend in patient preference to hear results quicker found in previous work was confirmed in this study. Most of the patients (98%) chose the opportunity to always review or have the option of reviewing future examinations with a radiologist the same day even if it meant waiting longer 1. The benefit of immediate communication would eliminate the additional waiting to hear from the ordering provider in normal current practice, a time described by many patients as a period of high anxiety 7 25. However, it should be noted that even the patients in our study with clinically significant or important findings (assigned category 4) showed no statistical difference in desire to review future examinations with the radiologist or perceived benefit of reviewing the examination with the radiologist. In addition, most respondents in this category also reported a decrease in anxiety after consultation.

An important issue to consider is the time that consultation encounters take and the effect on the workflow of a radiology practice. We found that the overall time for reviewing ultrasound examinations was shorter than that for CT, which likely relates to the number of images in each examination. Because ultrasound examinations tend to be shorter, they can be reviewed quicker. At our institution (and many other academic medical centers) the examinations are always reviewed by a radiologist before patient dismissal, so no additional time was needed for this portion of the study protocol. The duration of the consultation session was also shorter for ultrasound than for CT. The range of session time varied, largely according to the complexity of the examination and number of images, the examination findings, and patient questions. We acknowledge that instituting direct communication may temporarily diminish revenue. A previous assessment 14 of direct communication in mammography concluded that “talk isn’t cheap.” According to our data analysis, a radiologist can expect a cost associated with the time needed to review results directly with the patient, which is not reimbursed.

Because this study was a pilot study not previously performed at our institution, it was new to both investigators and patients. Many of our patients had never seen their images or had them formally reviewed. Hence they had many basic questions that necessitated overview of the examination and review of basic anatomy before review of actual results. We would expect that the time to review studies would improve as practices and patients become more accustomed to this process.

Unique patient preferences based on cultural and geographic background were suggested in a study conducted in Germany that showed a slight preference for receiving results from the referring physician as opposed to the radiologist 26. Although our data are limited by small sample size, a similar preference before the consultation session was found among patients in the race categories Asian and other (Table 1). In addition as shown in Table 1, a higher percentage of patients in these race categories opted to review future examinations with a radiologist only when they asked, whereas most white and African American patients opted to always review future examinations with the radiologist. This difference may be related to a partial language barrier, especially if the ordering or primary care provider speaks to the patient’s primary language.