A Patient-Centered Perioperative Mental Health Intervention Bundle: A Multi- and Mixed-Method Adaptation Study

Abstract Background Anxiety and depression are common among older adults and can intensify during perioperative periods, but few mental health interventions are designed for older surgical patients’ unique needs. We developed and adapted a perioperative mental health (PMH) bundle for older patients comprised of behavioral activation (BA) and medication optimization (MO) to ameliorate anxiety and depressive symptoms before, during, and after cardiac, orthopedic, and oncologic surgery. Methods We used mixed-methods including workshop studios with patients, caregivers, clinicians, researchers, and interventionists; intervention refinement and reflection meetings; patient case review meetings; intervention session audio-recordings and documentation forms; and patient and caregiver semi-structured interviews. We used the results to refine our PMH bundle. We used multiple analytical approaches to report the nature of adaptations, including hybrid thematic analysis and content analysis informed by the Framework for Reporting Adaptations and Modifications – Expanded. Results Adaptations were categorized by content (intervention components), context (how the intervention is delivered, based on the study, target population, intervention format, intervention delivery mode, study setting, study personnel), training, and evaluation. Of 51 adaptations, 43.1% involved content, 41.2% involved context, and 15.7% involved training and evaluation. Several key adaptations were noted: 1) Intervention content was tailored to patient preferences and needs (e.g., rewording elements to prevent stigmatization of mental health needs; adjusting BA techniques and documentation forms to improve patient buy-in and motivation). 2) Cohort-specific adaptations were recommended based on differing patient needs. 3) Compassion was identified by patients as the most important element. Conclusions We identified evidence-based mental health intervention components from other settings and adapted them to the perioperative setting for older adults. Informed by mixed-methods, we created an innovative and pragmatic patient-centered intervention bundle that is acceptable, feasible, and responsive to the needs of older surgical populations. This approach allowed us to identify implementation strategies to improve the reach, scalability, and sustainability of our bundle, and can guide future patient-centered intervention adaptations.

recovering from surgery without complete and accurate medication lists from patients. Clinicians worried about patients' limited understanding of their medications and suggested that inpatient social work and psychiatric consultation services could help.
These challenges highlighted the need for a mental health intervention for older surgical patients to alleviate symptoms of anxiety and depression before, during, and after surgery. Both patients and clinicians voiced the need for mitigating patients' fears and uncertainties across the perioperative care continuum, supporting both behavioral changes and psychiatric medication management, especially during the hospital stay; exibility to address each patient's characteristics, contexts, and surgical procedures; and, hiring a dedicated perioperative mental health interventionist (e.g., social worker) to deliver the intervention.
Towards this end, we proposed incorporating psychotherapeutic and pharmacotherapeutic treatments into a perioperative mental health (PMH) intervention bundle for older surgical patients [29][30][31]. Our PMH intervention bundle (Fig. 1) includes behavioral activation (BA), an evidence-based intervention for improving mood by increasing engagement in enjoyable activities [32], and medication optimization and deprescription (MOD) [26,33]. MOD includes a review and evaluation of patients' medications to determine if any are eligible for optimization and deprescription. BA and MOD are effective across medically ill populations in several settings and patients [34], especially older patients [35][36][37]. [38]. MOD can simplify polypharmacy, especially among older patients whose prescription lists grow along with the complexity of their chronic illnesses and whose risk for prescribing cascades increases over time [39].
In this paper, we report on our systematic and multi-and mixed-method, multi-stakeholder tracking and assessment of adaptations to our PMH intervention bundle, in a study running parallel to our feasibility study (NCT05110690) ( Table 1). We de ne adaptation as thoughtful or deliberate modi cations made to the intervention to improve its t within a given context [40]. Table 1 Details on adapted perioperative mental health bundle components ( [41]).

Behavioral Activation Medication Optimization and Deprescription
Interventionist Trained social worker Trained social worker and pharmacy team consisting of pharmacists and a geriatric psychiatrist Description Behavioral psychotherapy that helps depressed and anxious patients through identifying and tracking enjoyable and meaningful activities guided by personal goals and priorities [52].
Pharmacotherapy that helps to adjust suboptimal psychotropic dosages, deprescribe unnecessary or harmful medications, and ensure psychotropic continuation across the perioperative period [53,54].

Conceptual Framework
Our work was guided by ADAPT [42], a framework and step-by-step approach for working with stakeholders; selecting suitable interventions; undertaking and reporting adaptations; and evaluating and implementing these adaptations. Both proactive and reactive adaptations were assessed with stakeholder feedback (Fig. 2).

Study Setting
Page 4/32 The study was conducted at a large academic medical center in St. Louis, Missouri, with approximately 18,000 adult inpatient surgeries annually.

Study Design
Our intervention adaptations were tracked and assessed across two phases: pre-implementation (i.e., before the feasibility study) and post-implementation (i.e., during the feasibility study). Twenty-four patients from three surgical cohorts (orthopedic, oncological, and cardiac) were enrolled in the feasibility studies. These surgical procedures were selected due to their complex nature and high risk for postoperative complications and other poor outcomes (e.g., delirium, falls [43,44]). We also documented the delity of our bundle [45] (i.e., the degree to which its core components were delivered as intended). During both pre-and post-implementation adaptation phases, we aimed to maintain the core components of the bundle (BA and MOD) while adapting the exible components based on patient preferences and priorities (see protocol for details on intervention bundle [41]).
Given that this study focuses on describing the intervention bundle adaptation process and the use of multiple methods to track these systematically, it does not include the evaluation results from the feasibility study (see protocol [41]). The Institutional Review Board approved the study at Washington University (IRB#202101103).

Study Partners and Participants
This study included study partners in both pre-implementation and post-implementation phases, and participants (enrolled patients and caregivers) from the feasibility study in the post-implementation phase only.

Study partners
In the pre-implementation phase, we organized an internal advisory board (IAB) of study partners from the community and collaborators of the research team.
The IAB was comprised of patients and caregivers from each targeted surgical specialty; surgeons and nurses from each surgical specialty; community social workers/interventionists (masters-level clinicians trained in BA and MOD); pharmacists; a health IT administrator; Barnes Jewish Hospital patient experience representatives; and research team members (e.g., treatment developers, informatician). Patients on our IAB have a surgical and mental health history.
Patients and caregivers were recruited to the IAB through word of mouth and advertisements at the academic medical center.

Study participants
In the post-implementation phase, we included patient participants and caregivers from the feasibility study. Patients were at least 60 years of age, scheduled for one of the three identi ed surgical specialties, with clinically signi cant depressive or anxiety symptoms (see protocol for further details [41]). Patients' caregiver(s) were also invited to participate. Caregivers included patient-identi ed family members or friends who supported their health, safety, and recovery. Patients and caregivers enrolled in the feasibility study were contacted via telephone at the end of the study to gather feedback and suggestions for adaptations.

Data Collection
Data collection methods, participants, and ndings from each method are presented in Table 2. Table 2 Data collection methods. Pre-implementation refers to time-period prior to intervention implementation before the feasibility study; post-implementation refers to time-period after intervention implementation during and after the feasibility study. Weekly intervention re nement meetings: To identify pre-implementation adaptations necessary for successful PMH intervention bundle use among interventionists and patients.

Preimplementation
Interventionists, social workers, pharmacists, psychiatrists, behavioral scientists, and research team members 12 We assessed progress in intervention bundle design We identi ed preimplementation barriers to intervention delivery and brainstormed adaptations to make accordingly.
Periodic intervention re ection meeting: To re ect upon interventionists' experiences, to collect contextual data and triangulate data for a richer understanding [49].
We also identi ed barriers to intervention delivery and brainstormed adaptations to make accordingly.
Weekly patient case review meetings: To review and discuss patient intervention sessions and to document adaptations and challenges to intervention implementation.

Postimplementation
Interventionists, social workers, pharmacists, psychiatrists, behavioral scientists, and research team members 33 We assessed study progress and intervention bundle use among patients. We also identi ed postimplementation barriers to intervention delivery and brainstormed adaptations to make accordingly.
Audio-recordings of intervention sessions and collection of session documentation forms completed by interventionists: To capture data on progress towards MOD (adherence to medication changes, side effects) and BA (goals, values, activity scheduling and assessment) components; to also assess intervention delity through intervention delivery (delivering PMH intervention bundle consistently), intervention receipt (re ection of patients' receipt and understanding of the PMH intervention bundle and their capacity to use skills taught), and intervention enactment (patients' actual performance of MOD and BA skills and implementation of core intervention components) [55,56].

Postimplementation
Patients and interventionists 226 We assessed intervention delity to core components of MOD and BA and recorded any adaptations made during each session.
Patient interviews and caregiver interviews: To assess perspectives on the intervention and study overall.

Postimplementation
Patients and caregivers from feasibility study 19 We identi ed patient suggestions for future improvement to study content and implementation (for adaptation evaluation only).

Data Analysis
IAB workshop studios and periodic intervention re ection meetings IAB workshop studios, weekly intervention re nement meetings, weekly case review meetings, and periodic intervention re ection meetings were audiorecorded and transcribed.
We performed a hybrid inductive-deductive thematic analysis for each data source [46]. First, an experienced researcher in qualitative methods (JA) read transcripts multiple times for familiarity. JA then openly coded transcripts using data-driven codes (e.g., individuals involved in suggestions, rationale for adaptation) and created an intervention and research adaptation log. Next, JA analyzed adaptations thematically, organizing codes by themes (e.g., intervention bundle component adaptations, study adaptations) and sub-themes (e.g., design and implementation requirements for BA). Discrepancies in analysis were resolved through discussion until 100% consensus was reached with the interventionist team.
Weekly intervention re nement meetings and weekly case review meetings We conducted thematic and content analysis on weekly intervention re nement and case review meetings. Following transcription and review, JA openly coded transcripts using data-driven codes. Then, JA and another researcher (AM) determined coding categories based on the Framework for Reporting Adaptations and Modi cations -Expanded (FRAME [40]), a framework to track intervention and implementation strategy adaptations to re ne codes based on what was being adapted and the nature of the adaptations; when did the adaptations happen; who suggested the adaptations; and why the adaptations were needed. JA and AM coded the content of each meeting in accordance with coding categories (e.g., who includes patients, caregivers, interventionists, etc.) and tallied the frequencies of FRAME-based codes, identifying the most commonly suggested types of adaptations. Table 3 lists a speci c example of how we tracked and analyzed adaptations, including sub-theme de nitions. How did the intervention, study, or training and evaluation change?
Tailoring/rewording/re ning: a change to the intervention that leaves all of the major intervention principles and techniques intact Integrating intervention into another framework: another treatment approach is the starting point, but elements of the intervention are brought into the treatment Integrating another treatment into the intervention: the intervention is the starting point, but aspects of different therapeutic approaches or evidence-based practices are also used Removing/skipping elements: intervention baseline or standard treatment is based on the evidence-based practice, but particular elements are dropped Lengthening/extending (pacing/timing): a longer amount of time than prescribed by the manual is spent to complete the intervention or intervention sessions Adjusting the order of intervention components: intervention modules or concepts are presented in a different order than originally described in the manual

Patient and caregiver interviews
Patient and caregiver interviews were audio-recorded and transcribed for hybrid thematic analysis. A qualitative research team (JA and AM or FL) rst read transcripts multiple times for familiarity, and then both openly coded transcripts using data-driven codes (e.g., physical challenges, pain, patient perceptions of BA and MO). Second, they identi ed similar and overlapping codes and factors and categorized them into sub-themes, which were compared within and across transcripts to identify higher-level themes (e.g., barriers to successful BA implementation). These higher-level themes were translated into adaptation suggestions, which were coded using FRAME, similar to how the team analyzed IAB workshop studios, with weekly intervention re nement meetings, case review meetings, and periodic intervention re ection meetings. Discrepancies in coding were resolved through discussion and consensus with the interventionist team.

Intervention session audio-recordings and session documentation forms
Audio-recordings of intervention sessions and documentation forms were analyzed for adaptations by our interventionists across patients and for delity to the core components of our intervention bundle. Session documentation forms completed by the perioperative wellness team were analyzed using a deductive thematic analysis approach: adaptations noted on these forms were analyzed at the patient level and then higher-level themes on adaptations across patients were organized using FRAME.
Second, the intervention bundle delity was assessed using a structured delity rating checklist (Appendix S1) developed by our interventionist team and evaluated for language and clarity by researchers (JA and KF). The delity rating checklist mirrored the core components of the intervention bundle. After piloting the rating checklist, an undergraduate researcher (trained by JA) listened to the audio-recordings and documented how well the interventionists were delivering the intervention bundle as intended in the standard operating procedures (SOPs) or deviating from the intervention bundle SOPs. A PhD-level social worker (KH) randomly selected 20% of these intervention session recordings and completed the delity rating checklist. Fidelity was rated based on four core components: personalized rationale; values and goals assessment; activity scheduling; and activity tracking. Personalized rationale and values and goals assessment were considered completed if were discussed with the participant during at least one BA session. Activity scheduling and tracking were considered complete if participants engaged in scheduling activities during at least 80% of the BA sessions after session 3 (out of a maximum of 10). The IRR was calculated between the two researchers for all items on the delity rating checklist (Cohen's k = 0.76, consistent with a high level of agreement).

Results
Data were systematically collected between June 2021 and September 2022. 29 IAB study partners (including 15 patients and caregivers) and 19 patients and caregivers from our feasibility study participated in our adaptation assessment (Table 4). Prefer not to answer N=13 N=0 N=1 Table 5 displays our adaptation evaluation ndings, divided into pre-implementation and post-implementation themes and data. For example, one adaptation involved simplifying BA activity forms -this adaptation was coded as a pre-implementation content adaptation that involved tailoring, rewording, or re ning.
As a planned adaptation at the target intervention group level, we con rmed that the adaptation adhered to the core components of the intervention bundle and served to increase its effectiveness. Based on a framework 4 1 5 Based on knowledge and experience working with patients 0 9 9 Based on practical considerations 8 18 26 Based on nancial incentives/payments 0 0 0 Based on feedback or suggestions 0 1 1 WHY? What is the purpose of the adaptation?
Increase reach, participation, access 3 12 15 Increase effectiveness 6 4 10 Make intervention more aligned with organization goals 1 4 5 Increase implementation/ability of staff to deliver intervention successfully

Pre-implementation adaptations
One example of a content-based adaptation includes renaming interventionists. During IAB Studio #2, study partners discussed intervention language and expressed that the term "perioperative wellness partner (PWP)" could better re ect the intervention bundle deliverer's holistic training while emphasizing a comfortable environment for patients to improve their wellness after surgery. Other IAB members agreed, stressing that someone trained to speak reassuringly with patients and serve as a mental health advocate was necessary, as the patient would rely on the bond formed throughout the entire perioperative process.
In a similar discussion about language and patient acceptability, the term "medication optimization and de-prescription" was re ned to "medication optimization (MO)." This phrase felt less intimidating to patients, who were previously wary about stopping any of the medications they already took. This planned adaptation occurred at the target intervention group level and was intended to align the intervention bundle with organization goals better.
In another example, across IAB Studios #2 and #3, several IAB members suggested that it could be di cult for patients to connect with and trust strangers with personal issues during their rst session, especially over the phone. One PWP also emphasized that there was a "need for rapport building [ rst], so that we actually can personalize it. It's kind of… hard to personalize it when you don't know the patient that well and you're kind of working to get to know them through that." Thus, building a relationship with the PWP was crucial and was recommended prior to beginning BA. Following further discussion, the research team decided to modify the content of the rst session to focus on building trust and rapport and introducing the patient to the intervention and its core components (e.g., personalized rationale). This planned adaptation occurred at the individual patient/practitioner level and was intended to increase the effectiveness of the intervention bundle.

Post-implementation adaptations
During a periodic intervention re ection meeting, PWPs noted that patients had di culty following the activity scheduling and tracking documentation forms that they were assigned. One PWP, for example, stated, "Having [patients] strictly write stuff down… they don't really seem to need that." Similarly, Cardiac-Patient-3's documentation form noted that they did not track or schedule activities according to their PWP's instructions but remained very active and talked to their PWP about their recovery period activities. Similarly, Orthopedic-Patient-2 voiced that they "had the same problem when [they] went through trauma therapy.
[They] just don't write things down." Orthopedic-Patient-60 declined to log their activities, so their PWP proposed that they review their calendar at each session and recall activities without writing them down. Thus, BA documentation forms were reduced in detail and emphasized meeting the patient where they were, suggesting but not requiring activity documentation, with PWPs encouraged to offer exible methods of activity documentation (e.g., journaling). Loosening the structure of the BA documentation form was a reactive adaptation that occurred at the individual patient/practitioner level to increase patient reach, participation, and access to the intervention bundle.

Pre-implementation adaptations
One signi cant pre-implementation contextual adaptation involved forming separate teams for each intervention bundle component, as feedback indicated that the PWP would require real-time assistance during MO to correctly identify which medication changes could bene t the patient. At IAB Studios #2 and #3, patients and caregivers remarked that they wanted to see clinicians handling their medications directly rather than through consultation with ancillary staff and researchers. Several clinicians for leading MO were suggested, including primary care physicians and pharmacists. Ultimately, the research team established that the new surgical wellness program would comprise two teams: the PWP team and the pharmacy team. The PWP team managed BA and supervised all sessions, while the pharmacy team (pharmacists and a geriatric psychiatrist) managed MO. This planned addition to the intervention bundle occurred at the target intervention group level and served to increase the ability of staff to deliver MO successfully.

Post-implementation adaptations
Across periodic intervention re ection meetings and interviews, PWPs and patients noted that their overall success with the intervention bundle relied heavily on building trust and warm relationships during sessions. interactive and patient-sensitive. PWPs were instructed to reassure patients that their sessions were exible and personalized for their preferences, schedules, and needs. Adding these elements to the intervention bundle was a planned adaptation at the target intervention group level and served to align the intervention bundle with organization goals.
Additionally, patients suggested that during recruitment, before scheduling sessions, PWPs should provide more detailed explanations of the intervention bundle using more straightforward language. Cardiac-Patient-2 explained, "[Details about the intervention bundle] should be told to people before the surgery, and they need to know what bene ts are [and] what to do. And who to talk to for help." Orthopedic-Patient-2 and others felt that they went into the study not fully understanding what they needed to do and how the intervention bundle would help them, and only realized partway through the study. This was also evident in our session documentation form analysis: some patients did not understand what they were supposed to do before study participation and were ultimately not interested in the intervention bundle upon nding out more details throughout sessions. For example, Orthopedic-Patient-6's documentation form on their 5th session indicated that "the patient opted to withdraw from the study.
[They] stated that this is something that doesn't interest [them]," after two missed sessions and two sessions where they declined to complete the BA instructions. Thus, the research team modi ed consenting language to include a more thorough description of what to expect from the intervention bundle. This reactive adaptation on an individual patient/practitioner level served to increase reach, participation, and access to the study for patients.
Furthermore, we noted three key differences between surgical cohorts throughout the feasibility study: pre-operative timelines, session schedules, and patient needs. First, we observed substantial differences among the cohorts in perioperative timelines. Orthopedic patients typically scheduled their surgeries over 3 months in advance, oncologic patients scheduled their surgeries about 2 weeks in advance, and cardiac patients scheduled their surgeries about 2-3 days in advance. Therefore, orthopedic patients typically had more time preoperatively to start the intervention and plan pre-operative sessions, while cardiac patients had little pre-operative preparation time. This was noted for future implementation considerations to better shape intervention plans for each cohort.
Second, differences in postoperative schedules and medical treatment were observed between cohorts, resulting in changes to session frequency. For example, oncologic patients often required continued chemotherapy and thus could not attend BA sessions as frequently; they needed sessions every 2-3 weeks (vs. 1-2 weeks). Similarly, orthopedic patients were often busy with physical therapy following surgery, resulting in sessions scheduled every 2-3 weeks.
Third, each cohort had speci c surgery-based needs and priorities and utilized different activities and techniques. For example, oncologic patients often had trouble sleeping due to extensive discomfort and had di culties with physical recovery. As such, PWPs adapted their recommendations to provide sleep hygiene suggestions. Oncologic-Patient-2 explained that their PWP aided them in sleep hygiene strategies and felt that the BA components helped them with "incorporating [techniques] into the evening and the morning routine," which bene tted them. In another example, orthopedic patients typically had a physical therapist and received exercise instructions to strengthen replacement joints. Thus, PWPs established physical goals more frequently for them.

Pre-implementation adaptations
Before intervention bundle implementation, study partners agreed during IAB Studios #1, #2, and #3 that training interventionists were essential to obtain patient buy-in and trust in the intervention bundle. Patients were apprehensive about social workers, since they were skeptical about the potential lack of intervention training or experience in delivering mental health interventions. Thus, all PWPs recruited for this study came with training and prior experience in mental health. They were also trained in BA by Puspitasari and colleagues (using BASA training modules) [47]. Training sessions consisted of four 1-hour weekly sessions covering 4 core BA strategies and oriented the trainers to BA SOP content and resources. In addition to discussing core BA strategies, PWPs were taught how to model BA for patients, lead BA sessions, and provide feedback.
Weekly intervention meetings also helped PWP and pharmacy team to review the SOPs and materials with BA and MO treatment developers and receive continuous feedback. This planned addition to the study was conducted at an individual practitioner level and increased the ability of staff to deliver the intervention bundle successfully.

Post-implementation adaptations
Several training sessions were incorporated across intervention implementation as both refreshers and to introduce new elements to the study. First, intervention lead coordinators (EL and KH) provided several refresher training sessions and materials to train all PWPs on introducing the study, introducing the intervention bundle, and working through each form with patients. Training sessions were held over four 1-2 hour video conference meetings and included a mix of didactic and interactive content (e.g., role-playing). The sessions provided an orientation to the revised intervention manual and the objectives for each session, along with a review of unchanged core components and instructions on future work adaptations. This added element was planned at an individual practitioner level and was based on practical considerations to increase staff's ability for successful intervention delivery.
Another post-implementation adaptation to training involved adding pharmacy students to the pharmacy team to provide further expertise regarding medication optimization and education for patients. Training sessions were led by the two study team clinical pharmacy specialists with multiple weekly sessions. Session content included good clinical practices; review of the MO SOPs; electronic health record access and navigation; intervention database navigation; and demonstrating compassion and empathy during patient communication. Students were given supplemental readings about antidepressant dosing and potentially harmful medications. This reactive addition to study training occurred at an individual practitioner level and was based on practical considerations to increase staff's ability for successful intervention delivery.
In addition, several evaluation adaptations were made to the study as reactive responses to data and outcome collection di culties. For example, patients were evaluated originally at 1-month, 2-month, and 3-month follow-ups. However, the 2-month follow-up was removed from the SOP due to di culty in following up with patients. This reactive removal of an element occurred at the target intervention group level and served to increase reach, participation, and access to the study.

Intervention Fidelity
Personalized rationale was discussed for 87% of patients (n = 20), and values and goals assessment were discussed for 78% of patients (n = 18). Only 26% of participants engaged in activity scheduling (n = 6) and 17% in activity tracking (n = 4) for at least 80% of sessions after session 3.
A few patients (especially those who withdrew from the study) did not follow the core components of the intervention bundle. For example, Oncologic-Patient-2 was uninterested in BA and MO, and their sessions mainly consisted of their PWPs checking in on their recovery and activities. This content-based removal of BA and MO elements increased participation in the study but lacked delity to the original intervention bundle's purpose and functions.
However, most patients followed the core components of the intervention bundle while altering or skipping some BA and MO steps. For instance, Cardiac-Patient-5 did not schedule any activities, but was very motivated to recover from their procedure, so their PWP worked with them to identify goals, priorities, and strategies for meeting them. This re ning of the BA SOP let the patient decide how to utilize BA in a way that worked for them. Even though they did not schedule activities, they planned to do them on their own terms throughout the week, increasing the effectiveness of the intervention bundle and adhering to its components.

Discussion
We reported on the systematic adaptation process of a PMH intervention bundle for older adults to ensure intervention bundle feasibility and sustainability across three different surgical settings.

Use of Implementation Science Methods and Implications
Our comprehensive and robust approach to adaptation led to the development and re nement of our PMH intervention bundle that we anticipate will be acceptable, appropriate, and feasible for patients and PWPs in our full trial. Using the ADAPT framework to capture process [42], and FRAME to capture adaptations [40], our multi-pronged, multi-method approach identi ed methods to develop an adaptable intervention bundle for older patients in the perioperative context in three different cohorts (orthopedic, cardiovascular, and cancer patients).
Our data revealed differences in the types of adaptation between the pre-implementation and the post-implementation phases. In the pre-implementation phase, we reported more planned adaptations, mostly around content, to t the intervention to the perioperative setting. In the post-implementation phase, the number of reactive content and contextual adaptations increased. Several elements (e.g., adding compassionate modules) were added in the postimplementation phase. These data indicate the importance of feasibility trials to develop adaptable interventions to increase the probability of tting evidencebased interventions in new settings and/or for new populations. Speci cally, pre-implementation adaptation work has historically supported interventioncontext t and has contributed to intervention sustainability [48]. Spending time adapting the bundle to increase the t with the contexts hopefully will increase the probability of success and future sustainment of the intervention.
One of the main challenges in the eld of adaptation is examining how to adapt and track adaptations along the process [40] as the literature has scant examples of adaptations done in different phases of implementation [49]. We used multiple methods to triangulate the needs of patients and PWP with feedback from the advisory board and research team -this allowed us to ensure the t of the bundle in the three settings [50]. While our approach was timeconsuming and exhaustive with multiple iterations of feedback, discussion, and adaptation over the feasibility study and the planning period before implementation, it provided a robust understanding of needed adaptations. However, others may nd our approach not feasible and may employ a simpli ed version of our methods. Nevertheless, this adaptation assessment and tracking process can guide future patient-centered intervention adaptations while ensuring that they remain consistent with the original design and goals.

Adapted PMH Intervention Bundle
Our PMH intervention bundle (Figure 3), dubbed the "surgical wellness program," is pragmatic and collaborative, with both reproducible, generalizable core components (e.g., a dedicated pharmacy team, simpli ed BA documentation forms, and emphasis on compassionate care) and adapted, patient-driven components (e.g., varied activity scheduling and tracking methods and surgery-speci c preferences for BA activities). We describe signi cant adaptations below, with details of the intervention bundle according to the TIDieR checklist (Appendix S2) [51].
First, activity BA scheduling and tracking forms are now exible and based on patients' comfort and preference in entering necessary information; for example, some patients may prefer writing their activities down in a journal, while others might note their activities in their mobile phone. Second, BA activities now depend heavily on each individual patient's preferences and surgical recovery. Furthermore, PWPs demonstrate exibility in scheduling and session agenda planning according to each patient's mood during sessions. Examples of supplemental behavioral activities include activity plans for self-directed mindfulness practice, sleep hygiene exercises, and evidence-based cognitive training.
Third, surgery-based protocols are adjusted according to the different priorities, pre-operative timelines, session schedules, and patient needs of each surgical cohort. For example, orthopedic patients can schedule more pre-operative sessions, while oncologic patients can schedule one or two sessions before their surgeries. Cardiac patients can either schedule one session prior to surgery or do postoperative sessions only.
Fourth, our PMH intervention bundle takes on a shared decision-making approach with patients and is exible for each patient's needs and types of surgeries.
Our data found the importance of the PWP establishing trust with their patient and assessing patient needs and preferences while approaching their situation with empathy and compassion. Our data also showed the importance of giving the patient the option of using MO to reduce problematic medications and increase sub-therapeutic doses of antidepressant medications to therapeutic levels through a collaborative approach. In the immediate postoperative period, giving the option for the PWP and pharmacy team to visit the patient in the hospital is important. Following discharge, patients can choose how to utilize BA activity scheduling and tracking to support an individualized, active recovery. Patients can also schedule more or fewer sessions depending on their physical recovery progress and level of stress.
In the updated PMH intervention bundle, two main components are integrated to prepare cardiac, oncologic, and orthopedic patients for surgery and to promote enhanced recovery. MO reduces the use of problematic medications and increase sub-therapeutic doses of antidepressant medications to therapeutic levels. Through a collaborative approach, medication adjustments are made with the patient and pharmacy team. Additionally, BA engages patients in activities that are personally rewarding, supporting an individualized, active recovery from surgery, and encouraging patients to gradually re-engage in the important activities and cope with perioperative stress.
The PMH intervention bundle is carried out remotely via phone or secure web conference, with optional in-person visit/s while the patient is hospitalized. Each patient is assigned a PWP, who works with the patient prior to surgery and approximately 3 months after surgery. Preoperatively, the PWP establishes trust with their patient, assessing patient needs and preferences while approaching their situation with compassion. Meanwhile, the pharmacy team assesses patient medications and discusses recommendations with the patient and inpatient care teams, ensuring any changes are maintained in-house following surgery. Within the hospital, the pharmacy team conducts MO again if any further changes needed to be made. Finally, at home post-discharge, wellness partners work with patients on BA (e.g., activity planning, activity logging). Further MO guidance can be provided at the patient's request.
One-on-one sessions occur on a weekly basis initially and then are reduced to every two weeks or according to the patient's needs and preferences for a total of up to 10 sessions. Session duration is approximately 40 minutes per session but can be adjusted depending on patient needs and treatment goals.
All adaptations to the intervention bundle ensure personalized perioperative mental health care delivery. Although we have made several changes, we maintain the core components of the original intervention bundle and ensure its delity. In other words, the underlying functions of targeting behavioral change and medication optimization remain constant, preparing older patients for surgery through BA to build resilience and through MO to improve medication regimens and reduce risk of complications.

Future Work
We will conduct a Hybrid Type 1 Effectiveness-Implementation RCT to assess the effectiveness and implementation-potential of our adapted PMH intervention bundle in 300 older surgical patients across cardiac, orthopedic, and oncologic cohorts. Our control condition will receive printed mental health resources from our team, while the intervention condition will receive our PMH bundle. Our primary outcome of interest is depression/anxiety. Exploratory outcomes include quality of life, medication list, delirium, length of stay, rehospitalization, falls, PMH intervention bundle reach, implementation potential (acceptability, appropriateness, feasibility), and overall patient experience.

Limitations
We acknowledge study limitations. First, the approach was resource-intensive, requiring iterative data collection, analysis, and integration from several sources and stakeholders. Nevertheless, this can also be considered as a strength of the approach as it allowed us to conduct a thorough examination of necessary adaptations to suit our target surgical population's needs, priorities, and preferences, thereby improving the rigor in our intervention adaptation process. Second, our adaptations to the bundle were informed by the needs of our participants and advisory board members who may not be representative of diverse backgrounds (e.g., racial/ cultural differences). However, we are currently conducting a follow-up study to investigate this particular aspect, which is supported by our Diversity, Equity, and Inclusion Community Advisory Board. Third, given that this is a single-site study, feedback gathered might not have been representative of the overall target population's needs and preferences; additionally, results may not be generalizable from academic to community hospital settings. We plan for future multi-site evaluation that will help us re ne the bundle to meet the needs of patients across both rural and urban settings.

Conclusions
Mental health symptoms are a signi cant issue in the perioperative setting and can worsen adverse surgical outcomes. Across the literature, several studies have reported on interventions to address perioperative depression and anxiety, but often for general adult surgical populations, not speci cally older adults. Additionally, few studies have utilized mental health interventions along the entire perioperative timeline from pre-operative preparation to postoperative athome recovery. In response to a pressing need for perioperative mental health interventions adapted for an older surgical population, we identi ed evidencebased mental health intervention components from other settings and adapted them to the perioperative setting for older adults in a novel study. Tracking and assessing adaptations through multiple methods, we have created a pragmatic and patient-centered intervention bundle comprised of BA and MO components that ts the needs of older surgical populations.  ADAPT guidance for PMH intervention bundle adaptations. Blue boxes indicate stages of step-by-step guidance; grey boxes indicate potential outcomes; directional arrows indicate recommendations for moving between stages.