A multi- and mixed-method adaptation study of a patient-centered perioperative mental health intervention bundle

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. As part of the feasibility trial, 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.

Clinical trials Registration

NCT05110690 (11/08/2021).

Background

Approximately 14.4 million inpatient and 12 million major ambulatory surgeries are performed annually in the US [1]. Nearly half of these involve patients over age 65 [2]. Older surgical patients can experience anxiety and depression and have an increased risk of post-operative falls, venous thrombosis, delirium, short-term functional dependence, nausea, and vomiting [3,4,5,6,7,8,9]. Furthermore, older age combined with perioperative anxiety, distress, worry, or depression can lead to poor outcomes such as morbidity and mortality, pain, and decreased quality of life [8,9,10,11,12,13,14,15].

Psychotherapeutic interventions, or psychological and behavioral treatments aimed at changing behaviors [16], have been used to help patients manage symptoms of anxiety and depression, thereby maintaining overall mental health across the perioperative trajectory. Pre-operative psychological care (e.g., pre-operative assessment, guidance, and family support) and post-operative psychological care (e.g., timely feedback, standardized pain management, and psychological counseling) have improved mental health, including anxiety, hostility, paranoia, depression, and psychosis [17]. For example, cognitive behavioral therapy (CBT), including psychoeducation and reviewing behavioral goals, have alleviated symptoms of pre-operative anxiety and depression and is also associated with a speedier post-operative recovery process with early hospital discharges [18, 19].

In addition to psychotherapy, pharmacotherapy interventions to address mental health symptoms [20] show significant promise in reducing post-operative anxiety and pain. For example, medication-based treatments such as patient-controlled midolazam has shown to reduce pre-operative anxiety in middle-aged patients undergoing hysterectomies [21]. Similarly, escitalopram from preanesthetia to day 6 post-operatively relieved symptoms of depression in older patients undergoing knee arthroplasty [22].

However, most psychotherapeutic and pharmacotherapeutic interventions have primarly been developed and tested within younger adult populations and may be challenging to directly apply and may prove less effective for older surgical patients due to differences in risks (e.g., types of social determinants [23]) and symptoms (e.g., frailty, multimorbidity [24, 25]). For example, older patients with depression might not respond to selective serotonin reuptake inhibitors or serotonin-norepinephrine reuptake inhibitors, -which can result in worse physical and cognitive outcomes such as increased disability, cognitive decline, and increased risk of dementia [26]. Additionally, biases among older patients against mental health care can present challenges to successful psychotherapy [27]. Hence, it is critical to ascertain the perioperative experiences of older surgical patients with anxiety and depression, specifically focusing on their barriers and unique needs for a perioperative mental health intervention.

Towards this end, we conducted a multi-stakeholder qualitative interview study with 22 patients over age 65 who underwent major orthopedic, oncologic, or cardiac/thoracic surgeries and 18 perioperative clinicians [28]. Patients reported fear and stress about their surgery, difficulty with medication management, fragmented care transitions, lack of mental health assessment and treatment in the perioperative period, and limited clinician-patient communication about mental health treatment. Clinicians reported concerns about restarting psychotropic medications while patients were recovering from surgery without complete and accurate medication lists from patients and also worried about patients’ current psychiatric medications that were either at a sub-optimal dose or inappropriate and toxic for patients. This study also 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; flexibility to address each patient’s characteristics, contexts, and surgical procedures; and need to hire a dedicated perioperative mental health interventionist (e.g., social worker) to deliver the intervention.

To address this gap, we proposed a perioperative mental health (PMH) intervention bundle incorporating psychotherapeutic and pharmacotherapeutic treatments for older surgical patients [29,30,31]. Our proposed PMH intervention bundle (Fig. 1) includes two evidence-based treatments: behavioral activation (BA) and and medication optimization and deprescription (MOD). BA focuses on improving mood of patients by increasing their engagement in enjoyable activities [32,33,34]. 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 [35]. MOD entails a review and evaluation of patients’ medications to determine if any are eligible for optimization/escalation and deprescription. BA and MOD are effective across medically ill populations in several settings and patients [36], especially older patients [37,38,39,40].

However, in order to mitigate the mental health challenges faced by older surgical patients and address their unique and complex care needs across the perioperative period, we need to adapt the proposed PMH intervention bundle such that it is acceptable, usable, and approporiate for older surgical patient populations.

Fig. 1

Proposed PMH intervention bundle

In this paper, we report on the systematic and multi- and mixed-method, multi-stakeholder tracking and assessment of adaptations to our PMH intervention bundle (Table 1). We define adaptation as thoughtful or deliberate modifications made to the intervention to improve its fit within a given context [41].

Table 1 Details on adapted perioperative mental health bundle components ( [42])

Conceptual Framework

Our work was guided by ADAPT [46], a step-by-step approach for working with stakeholders; selecting suitable interventions; undertaking and reporting adaptations; and evaluating and implementing these adaptations. All adaptations were assessed with stakeholder feedback (Fig. 2).

Fig. 2

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 phases

Methods

Study setting

This adapatation study is part of the planning and feasibility study [47] (NCT05110690) 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-three older surgical patients from three surgical cohorts (orthopedic, oncological, and cardiac) were enrolled in the feasibility (please see [47] for CONSORT flowchart). These surgical procedures were selected due to their complex nature and high risk for post-operative complications and other poor outcomes (e.g., delirium, falls [48, 49]). We also documented the fidelity of our bundle [50] (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 flexible components based on patient preferences and priorities (see protocol for details on intervention bundle [42]).

In this paper, we report on the intervention bundle adaptation process and the use of multiple methods to track these systematically. The results from the evaluation of the bundle are published elsewhere (see [42] and [47]). The Institutional Review Board approved the study at Washington University (IRB#202,101,103).

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; health information technology administrators; hospital patient experience representatives; and research team members (e.g., treatment developers, informatician). Patients on our IAB have experience with surgery and a history of depression and/or anxiety diagnoses. Patients and caregivers were recruited to the IAB through word of mouth and advertisements at the academic medical center. Our patient and caregiver IAB members were compensated $100 per hour for their time.

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 identified surgical specialties (cardiac, oncologic, orthopedic surgeries), with clinically significant depressive or anxiety symptoms (see further details in [42, 47]). Patients were recruited by phone following clinician referral, self-referral, or screening through the electronic health record. Patients’ caregiver(s) were also recruited through patient referral. Caregivers included adults who are at least 18 years of age, patient-identified family members or friends who support the patient’s 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. Patients were consented via paper collected by mail, in person, or via an electronic REDCap® link to e-consent; caregivers consented verbally. Patients and caregivers who participated in the feasibltiy study were eligible for up to $125 and $25, respectively, as compensation.

Data collection

Several forms of data were collected to inform and track intervention adaptations across the two phases of study. Data collection methods, participants, and findings 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

Data analysis

IAB workshop studios and periodic intervention reflection meetings

IAB workshop studios and periodic intervention reflection meetings were audio-recorded and transcribed. We performed a hybrid inductive-deductive thematic analysis for each data source [54]. 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).

Weekly intervention refinement meetings and weekly case review meetings

Additionally, we conducted similar thematic and content analysis on weekly intervention refinement 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 Modifications – Expanded (FRAME [41]), a framework to track intervention and implementation strategy adaptations to refine 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 specific example of how we tracked and analyzed adaptations, including sub-theme definitions. Discrepancies in coding were resolved through research team peer debriefings and analysis discussions with the interventionist team until 100% consensus in coding was achieved.

Table 3 FRAME definitions and example of tracking document

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) first 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 identified 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 intervention-related themes were translated into adaptation suggestions, which were coded using FRAME, similar to how the team analyzed IAB workshop studios, alongside weekly intervention refinement meetings, and case review meetings. JA and AM independently and iteratively coded data, creating a codebook and refining it through multiple rounds of team discussion to reach 100% consensus in codes. The themes and sub-themes were informed by the framework, JA and AM reviewed individual themes and finalized those via peer-review and discussion.

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 fidelity to the core components of our intervention bundle. Session documentation forms completed by the interventionists 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 fidelity was assessed using a structured fidelity rating checklist (Appendix S2) developed by our interventionist team and evaluated for language and clarity by researchers (JA and KF). The fidelity 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 to assess and completed the fidelity rating checklists. Fidelity of BA 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 they 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). Fidelity of MOD was assessed based on whether our interventionists followed the core components: medication list review, determination of medications eligible for optimization or deprescription, discussion about medications of interest with the patient, buy-in and communication of recommendations to patient’s provider (i.e., medication prescriber), make adjustments to medications after receiving patient/prescriber buy-in, follow-up on treatment response, and weekly review of medications. The IRR was calculated between the two researchers for all items on the fidelity rating checklist (Cohen’s k = 0.76, consistent with a high level of agreement).

Results

Data were systematically collected between June 2021 and December 2022. 29 IAB study partners (including 15 patients and caregivers) and feasibility study participants (23 patients and 5 caregivers) participated in our adaptation assessment (Table 4).

Table 4 Study partner and participant demographics. *Caregivers were not asked to respond with sex, race, or ethnicity for the study

Table 5 displays our adaptation evaluation findings, 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 refining. As a planned adaptation at the target intervention group level, we confirmed that the adaptation adhered to the core components of the intervention bundle and served to increase its effectiveness.

Table 5 Adaptations identified across implementation timepoints

Out of 51 adaptations, content adaptations represented 43.1% (n = 22), contextual adaptations 41.2% (n = 21), and training and evaluation adaptations 15.7% (n = 8). The most common nature of adaptations was tailoring/rewording/refining adaptations (n = 18, 35.3%), while the most common level of adaptation was at the individual patient or practitioner-level (n = 36, 70.6%). Additionally, most adaptations were based on practical considerations (n = 26, 51.0%), were reactive (n = 28, 54.9%), and served to increase implementation or the ability of staff to deliver the intervention bundle successfully (n = 21, 41.2%). In the sections below, we describe these adaptations across pre-implementation and post- implementation phases.

Content adaptations

We identified 22 content adaptations (n = 22/51; 43.1%): 9 were pre-implementation (17.6%), and 13 were post-implementation (25.5%) (Table 6).

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 reflect 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 deprescription” was refined 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 difficult for patients to connect with and trust strangers with personal issues during their first session, especially over the phone. One PWP also emphasized that there was a “need for rapport building [first], 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 first 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 reflection meeting, PWPs noted that patients had difficulty 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 flexible 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.

Table 6 Content adaptations

Contextual adaptations

We identified 21 contextual adaptations (n = 21/51; 41.2%): 6 were pre-implementation (11.8%) and 15 were post-implementation (29.4%) (Table 7).

Pre-implementation adaptations

One significant 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 benefit 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. Subsequently, the research team established that the new perioperative wellness program would be led by the PWP team and the MO will be led by 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 reflection 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. Cardiac-Patient-1 noted that their PWP was “very sensitive… Very caring.” They elaborated, “I feel with all my heart that [my PWP] really helped me through a tough time… and cared about… my health and my well-being.” Similarly, Oncologic-Patient-5 said, “[My PWP] wasn’t judgmental. [They were] totally understanding.” This also aligned with findings from weekly patient case review meetings, as other PWPs agreed that including elements of emotional validation and warmth was important for intervention bundle success. Thus, to maintain the standard of empathetic care, the research team opted to incorporate elements of compassion into the intervention bundle protocol, adapting each session to be more interactive and patient-sensitive. PWPs were instructed to reassure patients that their sessions were flexible 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 benefits 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 finding 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 modified 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 pre-operatively 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 post-operative 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 specific 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 difficulties 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 benefitted 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.

Table 7 Contextual adaptations

Training and evaluation adaptations

We identified 1 pre-implementation (n = 1/51; 2.0%) and 7 post-implementation training and evaluation adaptations (n = 7/51; 13.7%) (Table 8).

Pre-implementation adaptations

Before intervention bundle implementation, study partners agreed during IAB Studios #1, #2, and #3 that training interventionists was 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 research. They were also trained in BA by Puspitasari and colleagues (using BASA training modules) [33]. 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 the PWPs 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 h 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 MO 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 difficulties. 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 difficulty 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.

Table 8 Training and evaluation adaptations

Intervention fidelity

With regards to fidelity to the BA core components, we found that 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.

With regards to fidelity to the MO core components, we found that medication review and dose adjustments were discussed with all patients with medications of interest eligible for MO (100%). Although the fidelity to other components was high, it varied among patients based on their buy-in. Across 23 patients, 16 patients were eligible for MO: there were 14 patients with medications eligible for deprescription and 5 patients with medications eligible for dose escalation, 3 of these participants had medications eligible for both deprescription and dose escalation. Of the 23 medications that were eligible for deprescription, 7 were deprescribed, 12 were not deprescribed and 4 were decreased or discontinued but not necessarily as a result of our intervention. Of the 8 medications eligible for dose optimization, 3 were optimized, 3 were not increased, 1 was not increased due to clincal reasons and 1 was still in consultation with the prescriber when the study ended. Some patients with medications eligible for MO deprescribing were not willing to discontinue medications that they felt helped them. For example, Oncologic-Patient-7 refused to stop taking zolpidem, explaining that it is “very important” to them.

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 fidelity 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 refining 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 report on a systematic adaptation process of a PMH intervention bundle for older adults to deliver a personalized pathway to optimize mental and cognitive health of surgical patients belonging to three different surgical cohorts. In addition, the adaptations help to ensure intervention bundle reach, uptake, and sustainability.

Use of implementation science methods and implications

Our comprehensive and robust approach to adaptation led to the development and refinement of our PMH intervention bundle that we anticipate will be acceptable, appropriate, and feasible for patients and interventionists (e.g., community social workers, pharmacists) in our ongoing clinical trials (NCT05575128, NCT05685511, NCT05697835). Using the ADAPT framework to inform the pre- and post-implementation adaptation processes and decision points [46], and the FRAME to characterize the nature and type of adaptations [41], we were able to plan, develop and iteratively adapt an intervention bundle for older patients within the perioperative context for three different cohorts (orthopedic, cardiovascular, and cancer patients).

Our findings elucidated differences in the types of adaptations between the pre-implementation and the post-implementation phases. In the pre-implementation phase, we reported more planned adaptations, mostly around content, to fit 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 post-implementation phase. These data indicate the importance of feasibility trials to develop adaptable interventions to increase the probability of fitting evidence-based interventions in new settings and/or for new populations. Specifically, pre-implementation adaptation work has historically supported intervention-context fit and has contributed to intervention sustainability [57]. Spending time adapting the bundle to increase the fit with the contexts hopefully will increase the probability of success and future sustainment of the intervention.

One of the main challenges in the field of adaptation is examining how to adapt and track adaptations along the process [41] as the literature has scant examples of adaptations done in different phases of implementation [51]. We used a multi-pronged, multi-method approach to triangulate the needs of patients and interventionists with feedback from the internal advisory board and research team – which allowed us to ensure the fit of the bundle in the three settings [58]. While our approach was time-consuming and resource-intensive with multiple iterations of feedback, discussion, and adaptation before and after implementation, it allowed us to refine the bundle further to optimize the mental and cognitive needs of older surgical patients. Such an 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

In the adapted PMH intervention bundle, two main components are integrated to prepare cardiac, oncologic, and orthopedic patients for surgery and to promote enhanced recovery. The bundle (Fig. 3) is pragmatic and collaborative, with both reproducible, generalizable core components (e.g., a dedicated pharmacy team, simplified BA documentation forms, and emphasis on compassionate care and care coordination support) and adapted, patient-driven components (e.g., varied activity scheduling and tracking methods and surgery-specific preferences for BA activities). 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 activities that are important to them and help to cope with perioperative stress. MO reduces the use of problematic medications and increase sub-therapeutic doses of antidepressant medications to therapeutic levels. Our data also showed the importance of giving the patient the option of using MO, and through a collaborative approach, medication adjustments are made with the patient and pharmacy team. In addition, by establishing a mutual understanding and collaborative approach to MO, we employ a shared decision-making approach with patients and their primary care providers. We labeled the psychological and psychosocial pieces underlying BA and compassionate care under the umbrella of a “perioperative wellness program.” We describe significant adaptations below, with details of the intervention bundle according to the TIDieR checklist (Appendix S3) [59].

First, activity BA scheduling and tracking forms are now flexible 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 flexibility 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 cardiac and oncologic patients can schedule one or two sessions before their surgeries or do post-operative sessions only depending on their pre-operative runways leading up to their surgeries.

Fig. 3

Adapted perioperative mental health intervention bundle

Fourth, our PMH intervention bundle is flexible 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.

Each patient is assigned a PWP, who works with the patient prior to surgery and approximately 3 months after surgery. Pre-operatively, 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, PWPs work with patients on BA (e.g., activity planning, activity logging). Further MO guidance can be provided at the patient’s request.

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. In the immediate post-operative 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. 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 min per session but can be adjusted depending on patient needs and treatment goals.

All adaptations to the intervention bundle were made to 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 fidelity. In other words, the underlying functions of targeting behavioral change and medication optimization remain constant, preparing older patients for surgery and recovery after surgery – with our overall perioperative wellness program and MO.

Ongoing work

We are currently conducting three Hybrid Type 1 Effectiveness-Implementation RCTs to assess the effectiveness and implementation-potential of our adapted PMH intervention bundle in a total of 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 also 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. Others may find our approach not feasible and may employ a simplified version of our methods. 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 refine the bundle to meet the needs of patients across both rural and urban settings.

Conclusions

Mental health symptoms are a significant 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 specifically older adults. Additionally, few studies have utilized mental health interventions along the entire perioperative timeline from pre-operative preparation to post-operative recovery. In response to a pressing need for perioperative mental health interventions adapted for an older surgical population, we identified evidence-based 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 comprised of a perioperative wellness program component supported by principles of BA and compassion-centered care; and a MO component, to optimize mental and cognitive health of older surgical populations.