Two novel prehabilitation apps to help patients stop smoking and risky drinking prior to hip and knee arthroplasty

Purpose Daily smoking or risky drinking increases the risk of complications after surgery by ~50%. Intensive prehabilitation aimed at complete cessation reduces the complication rate but is time-consuming. The purpose of this study was to carry out preoperative pilot tests (randomized design) of the feasibility (1A) and validation (1B) of two novel prehabilitation apps, habeat® (Ha-app) or rehaviour® (Re-app). Methods Patients scheduled for hip or knee arthroplasty with daily smoking, risky drinking, or both were randomised to one of the two apps. In part 1A, eight patients and their staff measured feasibility on a visual analog scale (VAS) and were interviewed about what worked well and the challenges requiring improvement. In part 1B, seven patients and their staff tested the improved apps for up to two weeks before validating the understanding, usability, coverage, and empowerment on a VAS and being interviewed. Results In 1A, all patients and staff returned scores of ≥5 for understanding the apps and mostly suggested technical improvements. In 1B, the scores varied widely for both apps, with no consensus achieved. Two of four patients (Ha-app) and one-third of the patients (Re-app) found the apps helpful for reducing smoking, but without successful quitting. The staff experienced low app competencies among patients and high time consumption. Specifically, patients most often needed help for the Ha-app, and the staff most often for Re-app; however, the staff reported the Re-app dashboard was more user-friendly. Support and follow-up from an addiction specialist staff member were suggested to complement the apps, thereby increasing the time consumption for staff. Conclusions This pilot study to test prototype apps generated helpful feedback for the app developers. Based on the patient and staff comments, multiple improvements in functionality seem required before scaling up the evaluation for effect on prehabilitation and postoperative complications.


Introduction
Unhealthy lifestyle behaviors are important and preventable risk factors for patients undergoing surgery.Among these behaviours, smoking and risky drinking add about 50% to the classic risk factors for developing complications [1,2].Recently, daily smoking and social alcohol drinking have been included in the American Society of Anesthesiology (ASA) international preoperative risk evaluation (the ASA score).Thus, a healthy patient in the lowest risk group is now defined as "healthy, non-smoking, no or minimal alcohol" [3].
Patient groups with smoking and risky drinking develop the same type of complications as all other patients, but they develop them much more frequently [1,2,4,5].The pathophysiology behind the increased risk involves the suppression of multiple organ systems [6].Organ dysfunction also exists in patients without alcoholinduced liver disease or smoking-induced lung disease [6].The organ damage is often subclinical and occurs at the cellular level, thus reducing the extra capacity that usually supports the patient during surgery and recovery.The result is a significantly increased risk of postoperative complications.However, complete abstinence from smoking and drinking for one to two months leads to sufficient recovery of organ functions, thereby reducing the complication rate [7].
Recently, strong evidence has shown significant risk reduction (leading to an approximate halving of postoperative complications after surgery) following patient participation in a four to eight week intensive program aiming at the complete cessation of smoking and drinking before surgery [8,9].This is also the case for arthroplasties, which have been the subject of several randomized trials on perioperative smoking and alcohol intervention [10][11][12].In contrast, shorter intervention programs do not affect postoperative complications [8,9,13].Likewise, there is no evidence of an effect on complications of only reducing the amount of smoking and intake preoperatively rather than ensuring complete cessation [12].
The intensive programs for preoperative cessation of smoking and alcohol consumption include pharmacological support, patient education, and motivational support.They are followed by high quit rates for smoking and alcohol and a significant reduction in complication rates [8,9].Implementing new evidence in the clinical setting is often slow and fragmented [14] and could benefit from effective modern communication and new technology.
A recent systematic review of perioperative digital behavior change interventions identified only one fullscale randomized trial [15], performed by Lemanu and colleagues on 2 × 44 patients [16].They evaluated the efficacy of a daily text message for four to six weeks preoperatively to encourage physical activity concerning bariatric surgery.However, significantly better perioperative adherence to exercise advice in the group receiving text messages was not followed by a change in the distance walked during the six minute walk test or in postoperative complications [16].More randomized trials are required.An important step toward that would be to pilot-test two newly developed apps and related dashboards aimed at prehabilitation involving successfully quitting smoking and risky drinking prior to surgery.
Here, we evaluate the functionality of these two apps in terms of the feasibility of the IT solutions (IA) and as validation regarding understanding, applicability, and support (IB) of the two new apps by the patients and their staff in the clinical setting.
For IA, the hypothesis was that at least one of the prototypes would score at least 5 on a visual analog scale for feasibility.For IB, we hypothesized that a substantial agreement (i.e., Kappa > 0.6) [17] on the validation of the improved prototypes would be obtained among the patient groups and the staff members.

Materials and methods
This H2020-supported project was initiated by researchers at the surgical and IT departments at three European university hospitals.

Pre-clinical steps
Before this clinical pilot on evaluation of the two IT solutions, development and selection processes regarding app development were carried out.This included two earlier phases with sprint tests, reports, and approval by the project group.These first steps were completed before the introduction to the clinical setting; in this last phase, the two most promising IT solutions of the original 15 proposals were further developed to be relevant for the clinical setting.

Hierarchical design
We used a hierarchical design for this randomized pilot, beginning with the first part (IA) and allowing for some last adjustments before the next pilot part IB: • IB would begin only if IA was completed successfully (i.e., the prototype worked well in the clinical set-up as measured by a score of at least 5 on a 0-10 visual analog scale) • A final and sizable randomized trial of the efficacy on lifestyle behaviors and postoperative complications would be triggered only if a substantial agreement was achieved among the patient and the staff groups in IB (defined as Kappa above 0.6 for understanding, usability, covering, and managing lifestyle improvement).
Risky drinking was defined as an intake ranging from above 14 drinks per week or at least 25 g ethanol per day on average.Daily smoking was defined as any everyday tobacco smoking, with or without dependence symptoms.

Criteria and recruitment
The recruitment for convenience samples took place from April 2020 to September 2021 at the orthopaedic departments with stratification at the national level and with blocks of unknown size (ranging from 1 to 3).Recruitment was closed due to the lockdown of elective surgery by the COVID-19 pandemic.For the feasibility test (IA), eight patients fulfilling the inclusion criteria (and meeting none of the exclusion criteria) were included with their six staff members.However, in case of major app adjustments, it would be necessary to repeat the 1A part and include more participants.For the validation (1B), we aimed to include ten other patients for the apps in each country; altogether, 30 patients and their staff members.
All staff had two to four of training sessions each per app.The inclusion criteria were adult patients scheduled for hip or knee arthroplasty with daily smoking, risky drinking, or both.Exclusion criteria were not being able to give informed consent (e.g., being less than 18 years old, having a severe mental illness, not understanding the national language), acute surgery, withdrawal of consent, being pregnant or lactating, not having or being able to press the buttons, or otherwise use a smartphone/tablet.

The IT solutions
The evaluation included two apps with related dashboards and the working titles Habeats® (Ha-app) from 8Wires® and Rehavior® (Re-app) from Eurecat®, both aiming to support patients quit daily smoking and risky drinking in the perioperative period, thereby reducing the risk of complications.The apps and dashboards were not connected to the medical record systems, as they used cloud solutions for data storage, which is not allowed for safety reasons; thus, anonymous data collection at the source was used for the IT solutions (see below).

The intervention
The IT intervention in the Re-app was based on the abovedescribed, highly effective face-to-face intensive smoking and alcohol cessation interventions, which also inspired the Ha-app [8,9].The intensive interventions include an education program for the patient, motivational support, feedback tools, and recommendations for supportive medicine (over the counter), such as nicotine replacement therapy, based on individual dependency measurements and preferences.Medication was not part of the apps.Both apps used artificial intelligence with videos or charts for information, recommendations, intervention, and praising of good results in patient feedback.During the project, a daytime hotline to the project staff was available to the participants.

Data collection and handling
Data were stored at the secure server for sensitive data in Region H.After informed consent, the clinical project group collected data on lifestyle behaviours and changes via interviews, the apps, and the dashboard.Data on the date and type of surgery were collected from the medical records.The IT companies extracted data on the duration of app use, the number of visits, and the opening and completion of games/chares.
The data handling conformed to the GDPR agreement.There was no contact between IT companies and patients.The project staff had contact with each IT company and did not share any personally identifiable information.Only the clinical project group had access to personally identifiable data.
Due to the cloud storage of data, there was no personally identifiable information in the apps or the dashboard, and the IT companies had no access to personally identifiable data.Anonymous pre-paid cards or smartphones were handed out.Only a random number or nickname was used in the app and dashboard, thereby making it visible to the IT companies.

Statistics
In IA, the feasibility was documented by describing the process from the log registration and response using a VAS with a cut-off value of 5.The time spent on the test was compared between the groups by a Mann-Whitney test; p < 0.05 was considered significant.
In IB, using Kappa statistics, the validity was analyzed for inter-variability among patients and staff.A substantial agreement was considered acceptable, i.e., a result of Kappa > 0.60.However, 0.81-1.00would be considered an almost perfect agreement.
The pilot results were expected to be used for the following power calculation regarding a possible final randomized trial on efficacy (part II).

Ethics, risks, side effects, and inconvenience
All participants were included after informed consent and could withdraw their consent and decide to make their data deleted at any time without further explanation and with no impact on the treatment and care in the department.The Scientific Ethical Committee (VEK-number H-21018467) in Denmark and the local hospital-based IRB approved the pilot and the final project before starting.
Inclusion in the project involved few and low risks, mainly originating from the collection and storage of sensitive data for patients, which was lowered even further by carefully following the guidelines for each procedure.In addition, changing lifestyle behaviours could be temporarily uncomfortable for patients, but the possible effect might balance this over time of reduced postoperative complications and better health.
The participants had access to a hotline (e-mail or phone) for the clinical project staff during the weekday.The participants were free to use other healthcare services to support smoking and alcohol cessation.

Results
Overall, the randomized pilot was doable for evaluating the app prototypes.It collected useful data on what worked well and identified several problems and challenges requiring solutions.Both apps required improvement during the feasibility test (1A) to prepare for the following validation (1B).There were no patients lost during the intervention and follow-up.
The project and, thereby, the patient recruitment were canceled from 2 April 2020 due to hospital reorganization to meet the acute needs during the COVID-19 pandemic.Afterward, elective surgery was only partly re-opened.However, part IA has been completed with the inclusion of eight patients and their staff, but only seven patients were recruited for part IB; see the consort diagram in Fig. 1 and characteristics in Table 1.None of the included patients were lost during the intervention and follow-up.

Feasibility
The evaluation in IA showed similar scores for both Ha-app and Re-app among patients and staff, respectively (Table 2).This was also the case for several comments regarding what worked well and identifying challenges and problems.Most of the challenges and problems during IA were solved by the patients themselves and the staff, but some required help from the IT companies to be solved.The clinical time consumption was relatively high, but significantly lower for the Re-app (Table 2).
Only one patient in 1A scored lower than 5 on the preand post-improvement tests.The patient felt the language stigmatizing concerning the risk of drinking in the Ha-app and scored 2.4 on the VAS at the test.Several other problems were mentioned, which were well solved, but not the language challenge, and the patient still scored below the value 5, which was the lower limit for the app to continue to part 1B (Table 2).However, another patient reporting risky drinking did not comment upon the language in that app.
Considering the small numbers in the pilot, there seemed to be a tendency for the Re-app to require a shorter time for testing the functions, showing fewer problems and challenges for the patients, and managing to develop solutions for all challenges and problems within the test period compared to the Ha-app (Tables 2 and 3).

Validation
After improvements, the new versions were evaluated among seven patients who were daily smokers (Fig. 1) and their staff members.Overall, feasibility evaluation was possible using the VAS score for all four outcomes for patients and five for staff.From a clinical point of view, the required level of involvement for solution of problems identified during the study is of importance; i.e. patients, staff and/or the company.All patients and staff scored 5 or more regarding understanding how to use the two improved prototypes, but the scores varied widely for both apps, and no substantial agreements were obtained (Tables 2 and 3).In total, three expressed that the app helped them to reduce smoking, 2/4 for the Ha-app and 1/3 for the Re-app.No patients had successfully quit smoking during the test.
The staff VAS scores and comments were rather similar for both apps.Overall, the challenges and problems experienced were mainly that the patients had minimal experience with using apps and did not have the necessary time to use the apps due to work reasons.The staff found the Re-app-related dashboard the most user-friendly.
Both apps were time-consuming: the Ha-app for the patients who often needed help and the Re-app for the clinical staff.The staff recommended support and follow-up from an addiction specialist as complementary to the apps.

Discussion
This study showed that the feasibility test identified several technical challenges requiring help from patients, staff, and IT companies.The time spent on the feasibility test was the longest for the Ha-app despite training workshops for the staff.The validity test results showed no substantial agreement for any of the apps.Some reduction in the number of cigarettes was reported in both apps after testing them for up to two weeks, although no participants managed to quit.
Recruitment, continuity, and time consumption are general challenges in digital behaviour change intervention studies concerning surgery [15].In this study, the time consumption for introduction and use and clinical involvement is worth considering concerning the intensive faceto-face interventions, which have been proven effective regarding both lifestyle behaviour changes and consequent reduction of complications.These intensive interventions are defined by four face-to-face sessions of ten minutes each or more, psycho-social and health education, pharmacotherapy, and motivational support individualized to personal needs regarding health literacy and other conditions of importance for the outcomes [18].
Despite a great potential for using effective IT solutions for perioperative lifestyle behavior risk reduction via lifestyle interventions, at the time of this study only four products have been evaluated in randomized trials: three pilot studies and one full-scale trial.Another six feasibility studies have been performed with only one arm and mainly using text messages.As most have used convenience samples, it is not surprising that the patient groups, in general, were positive.In addition, many of those, similar to this study, showed a need for optimizing functionality, IT literacy requirements, and time consumption reduction [19][20][21][22][23][24].
In addition to the earlier mentioned randomized trial by Lemanu and colleagues [16], three pilot studies have been performed in a randomized design [25][26][27].They included 145 patients in total, and all reported some tendencies toward effects on lifestyle improvement.DeMartini and colleagues evaluated the effect of text messages aiming at alcohol abstinence concerning liver transplantation.At the eight week follow-up, zero of eight in the intervention group compared to two of six in the standard care group had positive bio-markers indicating alcohol intake [25].Krebs and colleagues used gamification for an intervention aiming at smoking abstinence concerning lung and gastrointestinal cancer surgery.After one month, the confirmed abstinence was four of 13 versus two of 11 [26].
Van der Velde and colleagues tested a smartphone app 14 days preoperatively for 86 patients undergoing major surgery.Only 11 patients were frequent alcohol users, and seven were smokers.The self-reported results showed that two of two vs. one of nine patients reduced their alcohol intake, three of the seven smokers were followed up, and one in each group had reduced or successfully quit smoking [27].
Further to a possible positive tendency, the sparse randomized studies hitherto indicate that objectively measurable outcomes or other validations are necessary in addition to self-reported effects.

Strengths limitations
It was a strength that the estimated number of patients was included in part (IA), but the low number of participants biased the validation part (IB), and only smokers participated.The preliminary results of IB are probably not representative of the apps.Furthermore, the recruitment was strongly impacted by the COVID-19 pandemic, and participants were not included consecutively, which could lead to a selection bias.
Table 2 Functionality (1A) and validation (1B) of the apps, given as median (range) 2 One participant being risky drinking identified stigmatizing language, which was not solved by the IT company *P = 0.05

1A
Ha-app (n = 4) Re-app (n = 4) Patient responses to the apps After you have opened the app and put in your data, how well do you think it works on a VAS scale from 0 to 10 (0 = not at all; 10 = fantastic good)?recording system.This may have increased the already high need for staff resources and competencies to introduce the apps to the patients, use the dashboard, and follow up, all of which may be a limitation.In general, pharmacotherapy is recommended to increase the successful quit rate in the perioperative period [13], but this was not part of the intervention integrated into the apps, which could bias the results and lower the usefulness in the clinical setting.The length of education and level of health literacy of the patients may impact the compliance with programs aiming at lifestyle intervention.However, in previous face-to-face studies, the intensive programs for lifestyle intervention used in this project have an overall major effect on smoking and drinking across the patient groups [6,8,9].All the patients received similar information before inclusion.In addition, they received the educational material handed out as part of the clinical routines.
This study should be generalized only with caution beyond Europe.Even across Europe, the cultural differences include differences in IT literacy and language.Thus, phrases may be stigmatizing by some persons and not by others; this could even be the case within the same country or municipality.Language may be considered more stigmatizing in an app without staff to adapt a dialog to the patients' different vulnerabilities and situations.This could limit outreach via apps and other IT solutions to surgical patients needing perioperative risk reduction.

Perspectives
In contrast to the seldom performed effect studies on publicly accessible apps targeting lifestyle improvement, increasing evidence has been gathered for apps aiming at disease management, early identification of exacerbations, and symptom control for non-communicable diseases.This should also be the case for IT-supported prehabilitation, as the consequences of ineffective preoperative risk reduction have tremendous negative effects on individual patients, their families, and society at large.Given the updated European guidelines for medical devices to include apps in higher risk classification, hopefully, more IT solutions will be investigated properly.
In conclusion, it seems doable to pilot the prototype of apps in a hierarchical and randomized design and to improve the functionality based on the comments from the users, both patients and staff.The next steps are to run randomized controlled trials on the effect on lifestyle and postoperative complications before possible implementation in the surgical setting.

7. 5 9 7 7
(2.4 2 -10.0) 9.0 (5.0-10.0)After solving (or discussing) problems or challenges, how well do you think the app works now on the VAS scaleAfter introducing the app to one or more patients, how well do you think it works on the VAS scale?6.5 (4.7-10.0)7.5 (6.5-10.0)After solving (or discussing) problems or challenges, how well do you think the app works now on the VAS scale?7.0 (6.0-10.0)8.5 (7.0-10.0)Patient responses to the apps Do you understand how to use the app on a VAS scale from 0 to 10 (0 = not at all; 10 = fantastic good)?9.0 (5.0-10.0)10.0 (10.0-10.0)Do you think the app can be used to change your personal lifestyle on the VAS scale?5.5 (3.0-10.0)3.0 (2.0-8.0)Does the app cover your personal situation on the VAS scale?5.5 (3.0-10.0)3.0 (2.0-10.0)Do you feel the app makes you capable of taking control over the lifestyle change on the VAS scale?7.0 (4.0-10.0)5.5 (3.0-8.0)Time spent on the interview (minutes) 7 (7-10) 10 (7-10) Number of visits to the app -22 (8-35) The total duration of the visits -47 (21-73) Number of clicks to the app (minutes) 114 (0-182) -The total duration of clicks (minutes) 122 (0-216) -Tasks/charts and videos completed No data yet 22% (3-40) Staff responses: Do you understand how to use the app on the VAS scale? 10 10 Do you understand how to use the dashboard on the VAS scale?7 Do you think the app can be used for your patient group on the VAS scale?7 Does the app cover your patient group on the VAS scale?7 Does the app make your patients capable of taking control over the lifestyle change themselves on the VAS scale? 5 5

Table 1
Characteristics