Skilled Nursing Facilities, Behavioral Economics
Conditions
Keywords
Default Options, Order of Options, Choice Architecture, Video, Decision Aid, Mobile Application
Brief summary
This study tests whether changing the relative order of presentation of post acute care skilled nursing facilities (SNFs) in the Repisodic application, i.e., presenting preferred SNFs (P-SNFs) first, increases selection of, and discharges to, P-SNFs and decreases lengths of stay (LoS) for Geisinger patients discharged to SNFs. Along with the order of P-SNFs, the study also assesses whether a short video reviewing the benefits of care coordination in P-SNFs increases patient election to transfer to one.
Detailed description
Patients in acute-care facilities (hospitals) who are being discharged to post-acute care skilled nursing facilities (SNFs) currently struggle, in the absence of decision aids, to obtain practical information about SNFs to which the patients must be discharged. The same difficulties affect the attending case managers, social workers, and discharge planners. Moreover, although preferred SNF (P-SNF) networks offer patients the potential for improved coordination of care over non-preferred SNFs (NP-SNFs)-via shared staff across facilities and electronic sharing of information-this aspect of SNF selection is not highlighted or shared with patients according to any established protocol in the post-acute care setting. The Repisodic application is a technology platform, accessed via laptop, smart phone, or tablet, intended to provide a user-friendly approach to rapid SNF facility identification and selection, with information provided on facility amenities and details, quality ratings, distance, and fit for the patient's needs. It functions in a manner analogous to online consumer search/sort/filter/quality applications, allowing the discharge planner to build a curated list of discharge options matched to each patient's needs. The purpose of the current study is to test whether highlighting and changing the organization of, or information about, P-SNFs in the Repisodic app increases the selection of, and placements in, P-SNFs, while decreasing length of stay (LoS) at the discharging Geisinger acute care hospital. Our hypothesis is that presenting the P-SNF options at the top of the list (for case managers) and as the only options on the first screen or default map (for patients) will increase selection of and placement in P-SNFs, compared with the default ranking of SNFs (according to fit) in Repisodic. Note that patients presented P-SNF options on the first screen will always have the ability to expand the list to see all options they have been given, thus preserving choice. We also hypothesize that the P-SNF focused presentations will reduce LoS thanks to easier coordination between Geisinger hospitals and P-SNFs. An introductory video describing the SNF selection and discharge process to patients, will be available for optional viewing by patients across conditions. The optional link to the video will be further highlighted and the video will include sections describing the advantage of P-SNFs for care coordination, in the P-SNF focused (patient) conditions. The video, when viewed, may therefore further enhance the appeal of P-SNFs to patients in the conditions in which P-SNFs are highlighted. Secondarily, we hypothesize an overall reduction in LoS (at the acute care hospital) due to the quicker decision making, selection, and transmission processes facilitated by the Repisodic application. In a proportion of cases, patients and/or their caregivers may only have access to the SNF list via paper printout, in which case all relevant SNFs and their key characteristics will be included in the printout, but the map and video will not be available. The cluster-randomized trial study design will randomize case managers to one of two conditions (Default List vs. P-SNF Top-Sorted List) and patients to one of two conditions (Full List/Full Map/Default Video vs. P-SNF List/P-SNF Map/P-SNF Video). Primary outcomes include P-SNF selection and P-SNF placement (post-introduction of Repisodic) and LoS (i.e., the change in LoS from prior to Repisodic to post-Repisodic introduction). Secondary outcomes include patient time spent selecting a SNF on Repisodic. Generalized linear mixed models will be used to examine outcomes as a function of the conditions, accounting for patients nested within case managers and controlling for potential covariates and moderators at the patient, case manager, and discharging unit level. The findings will enhance our understanding of how patients' informed choice, as facilitated by a bedside decision tool, affects SNF choice. Note: random assignment of case managers to conditions will occur upon launch of the application (anticipated: January, 2021) and any patients will be randomly assigned starting at that time. However, due to the COVID-19 pandemic, SNFs are experiencing aberrant operational issues and bed availabilities. Furthermore, many patients who would be discharged to SNFs under normal circumstances are being sent home instead. In order to improve generalizability of the present study to post-pandemic circumstances, formal data collection to assess the impact of the intervention on primary outcomes will begin after a pandemic alleviation benchmark has been met. This benchmark specifies 25 or fewer average cases per 100,000 over a contiguous two-week period in the top 12 volume counties of Geisinger's catchment area (Centre, Mifflin, Juniata, Perry, Cumberland, Union, Snyder, Northumberland, Montour, Columbia, Luzerne, and Lackawanna).
Interventions
BEHAVIORALOrder of Options
Relevant providers who are part of the Geisinger preferred provider network will be presented at the top of the otherwise ranked list of SNFs.
BEHAVIORALDefault Option
Both the list and the map will present only the P-SNFs, with a button allowing patients to view all results only if they opt to do so. The optional P-SNF advantage video will be included and the link to it highlighted.
Sponsors
Geisinger Clinic
Study design
Allocation
RANDOMIZED
Intervention model
FACTORIAL
Primary purpose
HEALTH_SERVICES_RESEARCH
Masking
SINGLE (Subject)
Masking description
Participants (i.e., patients/caretakers and case managers) will not be informed specifically of their assignment to different arms throughout the study, but case managers will be informed that Repisodic is continuing to undergo development and testing and that they may therefore notice different patients and colleagues receiving slightly different versions of Repisodic over time.
Intervention model description
Case managers (\ 145) will be randomly assigned to conditions in a cluster-randomized trial design: 1. Default List: all eligible SNF options presented in Repisodic in rank order. 2. P-SNF Top-Sorted List: same as Default List, except relevant P-SNFs will be shown at the top of the otherwise ranked list. Patients (\ 5,000) will independently be randomized to conditions and tracked: 1. Full List/Full Map/Default Video: all SNF options preselected by case managers will be presented to patients ordered by Repisodic's ranking algorithm. A map will present SNFs on the list if selected, and a link to an introductory video on SNF selection and discharge will be available. 2. P-SNF List/P-SNF Map/P-SNF Video: same as Full List, except that both the list and map will present only the P-SNFs, with an opt-in button allowing patients to view all results. Moreover, the link to view the optional SNF video will be highlighted and the advantage of P-SNFs for care coordination will be presented.
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
No
Inclusion criteria
* Discharging units at Geisinger (e.g., Geisinger Medical Center, Geisinger Wyoming Valley Medical Center, and Geisinger Community Medical Center) * Patients placed on adult medical/surgical wards * Patients selected for discharge to a SNF * Case managers using Repisodic to discharge patients (under the criteria above) to a SNF * Discharge date occurs after pandemic alleviation benchmark has been met (see Study Description for further details)
Exclusion criteria
* Discharges in which the full list presented to the case manager contains no P-SNFs. Note: If a patient consistent with the above criteria has a caretaker that is interacting with the Repisodic application for the patient's post-acute care decision, the caretaker's behavior will be used.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| P-SNF selection (patient/caretaker level) | 24 months or as long as it takes to reach N=5,000, whichever occurs first | Patient selected a P-SNF for discharge: yes/no. This is a more direct reflection of patient behavior, not yet moderated by bed availability and other real-world considerations that affect Outcome 7. |
| P-SNF pre-selection (case manager level) | 24 months or as long as it takes to reach N=5,000, whichever occurs first | Number of P-SNFs selected by case managers for inclusion in the patients' lists, to assess the direct impact of Repisodic changes on case manager behavior; this is the first decision in the process, before patients' selections and SNF availability factor into the final discharge decision and placement. |
| P-SNF selection (case manager level) | 24 months or as long as it takes to reach N=5,000, whichever occurs first | Number of P-SNFs selected by patients to help assess the downstream impact of Repisodic changes on patient behavior following case manager pre-selection, again omitting the influence of SNF availability. |
| P-SNF percent selected (case manager level) | 24 months or as long as it takes to reach N=5,000, whichever occurs first | Percentage of patients who selected at least one P-SNF, to help assess the downstream impact of Repisodic changes on patient behavior following case manager pre-selection. |
| LoS (patient/caretaker level) | 24 months or as long as it takes to reach N=5,000, whichever occurs first | LoS for patients (in days). |
| P-SNF placement (patient/caretaker level) | 24 months or as long as it takes to reach N=5,000, whichever occurs first | Patient was discharged to a P-SNF: yes/no. |
| Overall LoS (system/case manager level) | 24 months or as long as it takes to reach N=5,000, whichever occurs first | Difference in LoS (in days) for patients as a function of case manager study condition. |
| P-SNF placement (system/case manager level) | 24 months or as long as it takes to reach N=5,000, whichever occurs first | Difference in proportion of discharges of patients to P-SNFs (out of P-SNFs + NP-SNFs) as a function of case manager study condition. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Time to pre-select a SNF (case manager level) | 24 months or as long as it takes to reach N=5,000, whichever occurs first | Time to pre-select a SNF/SNF list in Repisodic (in minutes), from the time that the case manager first signs into the iPad to when the case manager saves the list to send to the patient. This measure excludes cases in which the task was not completed. |
| Time to select a SNF (patient/caretaker level) | 24 months or as long as it takes to reach N=5,000, whichever occurs first | Time to select a SNF in Repisodic (in minutes), from the time that the patient first signs or is signed into the iPad to when the patient clicks to send the list to the case manager. This measure excludes cases in which the task was not completed. |
Other
| Measure | Time frame | Description |
|---|---|---|
| LoS for patients who had a P-SNF as an option (patient/caretaker level) | 24 months or as long as it takes to reach N=5,000, whichever occurs first | LoS for patients (in days). Excluding patients who did not have at least 1 P-SNF on the full list they were provided. |
| P-SNF placement for patients who had a P-SNF as an option (patient/caretaker level) | 24 months or as long as it takes to reach N=5,000, whichever occurs first | Patient was discharged to a P-SNF: yes/no. Excluding patients who did not have at least 1 P-SNF on the full list they were provided. |
| P-SNF selection for patients who had a P-SNF as an option (patient/caretaker level) | 24 months or as long as it takes to reach N=5,000, whichever occurs first | Patient selected a P-SNF for discharge: yes/no. Excluding patients who did not have at least 1 P-SNF on the full list they were provided. This is a more direct reflection of patient behavior, not yet moderated by bed availability and other real-world considerations that affect Outcome 12. |
Countries
United States
Outcome results
None listed