Abstract
This randomized clinical trial examines the effect of implementing a default order option in the electronic medical record on daily imaging utilization for patients with cancer undergoing radiotherapy during palliative care.
Annually in the United States, about 250 000 patients with advanced cancer receive palliative radiotherapy to lessen pain, control bleeding, or improve quality of life. To ensure reproducible positioning, patients are immobilized on laser-aligned treatment tables and have standard weekly imaging during treatment. Daily imaging, using radiography or computed tomography, can augment positioning. Although daily imaging is often used for curative radiotherapy, national guidelines consider it unnecessary for palliative radiotherapy.1,2 Unnecessary imaging can increase treatment time and expense for patients in distress.
Default options, which leverage insights from behavioral economics, can change physician behavior but have focused less on deadoption of unnecessary care.3,4 We conducted a stepped-wedge cluster randomized clinical trial to test the effectiveness of introducing a default imaging order in the electronic health record (EHR) vs usual practice to reduce unnecessary daily imaging during palliative radiotherapy.
Methods
The trial (NCT03110692) was conducted in the University of Pennsylvania Health System from February 10, 2016, to February 9, 2018, including a 1-year preintervention period. The trial protocol is available in Supplement 1. The sample comprised physicians from 5 radiation oncology practices (1 university practice in Philadelphia and 4 community practices in Pennsylvania and New Jersey). Eligible physicians prescribed at least 10 courses of palliative radiotherapy during the trial. Eligible patients were aged 18 years or older with bone, soft tissue, or brain metastases receiving 3-dimensional conformal radiotherapy. Single fraction radiotherapy was excluded. This study was approved as a quality improvement project by the University of Pennsylvania institutional review board and informed consent was waived.
The intervention introduced a default imaging order in the EHR that specified no daily imaging during palliative radiotherapy. Physicians could opt out, selecting another imaging frequency. Practices were classified into 2 groups: university or community based. Groups were randomly assigned by coin flip to cross over to the intervention in two 4-month predefined wedges (analyses included 1-month washout periods).
The primary outcome was a binary indicator of radiotherapy courses with daily imaging (defined as imaging during ≥80% of treatments). In intention-to-treat primary analyses, we fit models using generalized estimating equations clustering on physicians, using group and period (4-month increments) fixed effects and adjusting for monthly temporal trends. In secondary analyses, we adjusted for age, sex, race, performance status, insurance type, fraction count, dose per fraction, prior radiotherapy, and target. We examined effects at university and community practices by interacting group with the intervention periods. We bootstrapped to obtain adjusted differences in percentage points. Analyses were conducted in SAS statistical software (version 9.4; SAS Institute Inc).
Results
The sample comprised 21 radiation oncologists and 1019 patients who received 1188 palliative radiotherapy courses (n = 747 at the university practice; n = 441 at the community-based practices) to bone (52.2%), soft tissue (19.9%), brain (15.7%), or multiple sites (12.3%). Table 1 shows the flow of patients through the trial. Daily imaging was used in 68.2% (463/679) of courses in preintervention periods and 32.4% (165/509) of courses in intervention periods. The default intervention led to a significant reduction in daily imaging (adjusted odds ratio, 0.43; 95% CI, 0.24-0.77; adjusted difference in percentage points, −18.6; 95% CI, −34.1 to −2.1; P = .004) (Table 2). These findings were similar in analyses also adjusted for patient and treatment characteristics and across both university and community practices.
Table 1. Flow of Patients Who Received Palliative Radiotherapy Courses Through the Trial.
| Variable | Prestep Observation Period: February 10, 2016, to February 9, 2017, No. | Step 1: February 10, 2017, to June 9, 2017 (Intervention at University-Based Practice), No. | Step 2: June 10, 2017, to October 9, 2017 (Intervention at Community-Based Practices), No. | Poststep Observation Period: October 10, 2017, to February 9, 2018, No. |
|---|---|---|---|---|
| University-based practice | ||||
| Radiotherapy courses screened | 441 | 173 | 150 | 194 |
| Eligible | 375 | 105 | 123 | 144 |
| Ineligible | 66 | 68 | 27 | 50 |
| Single fraction | 40 | 21 | 19 | 31 |
| Physician with less than 10 radiotherapy courses | 26 | 6 | 8 | 19 |
| Washout month | 41 | |||
| Community-based practices | ||||
| Radiotherapy courses screened | 241 | 88 | 87 | 76 |
| Eligible | 220 | 84 | 67 | 70 |
| Ineligible | 21 | 4 | 20 | 6 |
| Single fraction | 12 | 4 | 4 | 6 |
| Physician with less than 10 radiotherapy courses | 9 | 0 | 0 | 0 |
| Washout month | 16 | |||
Table 2. Analyses of the Effect of a Default Order Option on Daily Imaging During Palliative Radiotherapy.
| Variable | Preintervention Periods, No./Total (%) | Intervention Periods, No./Total (%) | Adjusted Odds Ratio (95% CI) | Adjusted Percentage Point Difference (95% CI) | P Value |
|---|---|---|---|---|---|
| Main analyses | |||||
| Primary model | 463/679 (68.2) | 165/509 (32.4) | 0.43 (0.24 to 0.77) | −18.6 (−34.1 to −2.1) | .004 |
| Also adjusted for patient and treatment characteristics | NA | NA | 0.37 (0.19 to 0.72) | −18.8 (−34.2 to −2.4) | .003 |
| Heterogeneity of Treatment Effects by Groupa | |||||
| University-based practice | 252/375 (67.2) | 107/372 (28.8) | 0.33 (0.14 to 0.76) | −22.3 (−44.0 to −5.9) | .01 |
| Community-based practices | 211/304 (69.4) | 58/137 (42.3) | 0.45 (0.22 to 0.89) | −27.5 (−46.5 to −11.1) | .02 |
Abbreviation: NA, not applicable.
Primary model also adjusted for patient and treatment characteristics.
Discussion
In a network of 5 radiation oncology practices, introducing a default order in the EHR reduced unnecessary daily imaging during palliative radiotherapy. There was potential for spillover to community practices during the university intervention period; however, this would bias results toward the null. Our findings suggest that simple nudges, such as setting default orders, can meaningfully reduce unnecessary care.
Trial Protocol
Data Sharing Statement
References
- 1.eviCore healthcare Clinical Guidelines: Radiation Therapy. 2016. Bluffton, SC: CareCore National, LLC d/b/a eviCore healthcare (eviCore). https://www.evicore.com/-/media/files/evicore/clinical-guidelines/solution/radiation-oncology/healthplan/mvp-radiation-therapy_v102019_eff03012019_pub-1122018.pdf. Accessed December 4, 2018.
- 2.Nabavizadeh N, Elliott DA, Chen Y, et al. Image guided radiation therapy (igrt) practice patterns and igrt’s impact on workflow and treatment planning: results from a national survey of American Society for Radiation Oncology members. Int J Radiat Oncol Biol Phys. 2016;94(4):850-857. doi: 10.1016/j.ijrobp.2015.09.035 [DOI] [PubMed] [Google Scholar]
- 3.Ojerholm E, Halpern SD, Bekelman JE. Default options: opportunities to improve quality and value in oncology. J Clin Oncol. 2016;34(16):1844-1847. doi: 10.1200/JCO.2015.64.8741 [DOI] [PubMed] [Google Scholar]
- 4.Patel MS, Day SC, Halpern SD, et al. Generic medication prescription rates after health system-wide redesign of default options within the electronic health record. JAMA Intern Med. 2016;176(6):847-848. doi: 10.1001/jamainternmed.2016.1691 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Trial Protocol
Data Sharing Statement