Abstract
Background:
Early postoperative ambulation reduces length of stay and prevents postoperative complications after bariatric surgery. Rarely is postoperative inpatient activity objectively measured despite readily available commercial activity trackers.
Objectives:
Evaluate the impact of using activity trackers to record number of inpatient steps taken after bariatric surgery and assess how patient characteristics may affect the number of steps recorded.
Setting:
University Hospital, United States.
Methods:
Using an activity tracker, the number of steps taken during the postoperative hospital stay was recorded for 235 patients undergoing either laparoscopic Roux-en-Y gastric bypass or laparoscopic sleeve gastrectomy. Patients were randomly assigned to either being informed about the devices’ ability to record the number of steps taken or blinded to the purpose of the devices. Descriptive statistics were used to summarize study sample, a t test was used to compare number of steps recorded between groups, and a multivariate regression model was used to examine the effect of age, sex, preoperative body mass index (BMI), types of surgery, and length of stay on number of steps recorded.
Results:
One hundred twenty-five patients (52.8%) were randomized to the blinded group while 111 (47.2%) were informed that the device would record the number of steps taken. There were no differences in the number of steps recorded between the 2 groups. Patients with prolonged length of stay recorded lower numbers of steps taken on postoperative day 0. Increasing age was seen to reduce the number of steps recorded on postoperative day number 1. There were no significant differences in number of steps recorded based on sex, preoperative BMI, or surgery type.
Conclusion:
The present study found that knowledge of an activity tracker being used did not affect the patient’s activity level as measured by steps recorded. Increasing age correlated to reduced number of steps recorded on postoperative day 1 after bariatric surgery.
Keywords: Bariatric surgery, Roux-en-Y gastric bypass, Sleeve gastrectomy, Activity tracker, Fitbit, Enhanced recovery after surgery (ERAS)
The practice of encouraging early activity in the form of ambulation in postoperative patients has been recognized by surgeons since the 19th century [1]. In the modern day, early postoperative activity is an essential component of Enhanced Recovery after Surgery (ERAS) protocols, a multimodal approach to postoperative care that has been proven to reduce hospital stays and prevent postoperative complications [2]. Though early postoperative activity is best seen as only a part of the holistic care of an operative patient, ambulation itself has been shown to improve postoperative outcome measures and decrease the incidence of postoperative complications such as delirium and pneumonia [3,4]. Early initiated activity is especially important in the realm of bariatric and metabolic surgery, as this patient population has a higher risk of developing postoperative complications, including cardiopulmonary complications and venous thromboembolism (VTE), compared to patients with a body mass index (BMI) <30 kg/m2 [5]. The link between excess weight and VTE is well-established [5] and has been shown to be mitigated by early ambulation [6].
Despite the importance of inpatient ambulation, patient compliance often remains low [7]. While constant monitoring by staff might elicit compliance, this approach is labor-intensive and frequently not practical in modern hospital settings. Investigations into this problem gained a powerful tool in recent years with the advent of commercially available activity trackers. Such devices, when fitted with a wireless transmitter, allow effortless collection of data regarding a patient’s activity and ambulation as recorded in number of steps taken. In a prior study, Daskivich et al. used activity monitors to objectively quantify postoperative ambulation in patients undergoing major surgery [8]. Their study found an inverse correlation between number of recorded steps taken on postoperative day 1 and length of hospital stay after surgery. Another study by Twiggs et al. used activity monitors to record the number of steps taken after total knee replacement. Their results found an inverse correlation between the number of recorded postoperative steps by the activity counter and patients’ BMI [9]. Activity trackers have been used to show that preoperative patient characteristics predict the number of steps that patients take postoperatively. Patients with severe obesity have been shown to take fewer steps than patients with lower BMIs and about a quarter of bariatric surgery patients were reported as less active after surgery than before their operation [10].
It is still unknown whether activity trackers might be used to increase patient ambulation efforts, especially among those with severe obesity. The purpose of this study was to investigate, among patients undergoing bariatric surgery, whether telling patients that their steps were being recorded with activity trackers would affect the number of steps taken during their postoperative hospital stay. The study also examined how patient and surgery characteristics affect the number of steps recorded after bariatric surgery by an activity tracker. We believe this is the first study to measure postoperative inpatient activity with an activity tracker in patients undergoing bariatric surgery.
Methods
Sample and procedures
This prospective study received approval from the institutional review board (IRB) at West Virginia University (Morgantown, WV), IRB#1710829121. Included patients were all patients undergoing laparoscopic sleeve gastrectomy or laparoscopic Roux-en-Y gastric bypass, performed from August 1, 2018 through December 31, 2019, at a large, tertiary university hospital with an accredited Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program. Patients were excluded if they were having an additional procedure during their bariatric surgery, most commonly a hiatal hernia repair. This was done to reduce potentially confounding factors that may alter the degree of postoperative activity. Consent for the study was attained during the patients’ final preoperative clinic appointment.
Activity trackers were placed on all consenting patients in the postanesthesia care unit (PACU). Patients were randomly assigned to being either informed about the devices’ ability to record the number of steps taken or blinded to the purpose of the devices. Those in the control group were told that the device was a monitoring device, but were not told what parameters were being monitored or for what reason. Those that were informed of the devices’ intent were told that the device would record every step they took while in the hospital. In addition to this information being discussed in the PACU, patients were also reeducated with the same conversation in the morning on postoperative day one (POD#1).
Beyond the discussion regarding the intent of the device, all patients were given the same postoperative instructions, including the recommendation to ambulate around the unit at least 3 times daily. This education to encourage ambulation was provided at the preoperative education appointment, at the postoperative bedside nursing education given upon admission to the hospital floor unit, and by the surgeon on the POD#1 morning visit. Our practice protocol is to have all patients ambulate within 4 to 6 hours of completing the operation. Each patient was given an ambulation tracking sheet to self-report the number of times they ambulated around the hospital floor unit in addition to self-tracking the amount of oral intake. The activity sheet was reviewed twice daily by the patients’ providers to ensure practice ambulation guidelines were being met. With multiple educational reminders, all patients in both groups followed the recommendation, ambulating within 4 to 6 hours of completion of the operation and at least 3 times daily.
The activity tracker chosen for this study was the FitBit™ Flex produced by Fitbit, Inc (San Francisco, CA). The device has a small central unit that was utilized without the wristband by attaching it to the patient’s hospital gown. Each unit was tested before the beginning of the study to ensure it had the same accuracy whether attached to the wristband or pinned to clothing. The device was pinned to the hospital gown in a consistent location for all patients: the upper anterior chest region in the midclavicular line. This allowed the device to be positioned over the hospital gown pocket that is designed for telemetry packs. The device has a 3-axis accelerometer which records accelerations and proprietary software (FitBit™ application version 2.82) translated acceleration measurements into a step count [11]. While the software can be used to calculate distance ambulated and caloric expenditure, we used number of steps to avoid errors in comparisons secondary to differences in ambulatory stride length and metabolic rates that influence accuracy in software calculated distances and caloric expenditure.
A maximum of 6 activity trackers were utilized at a time for this study. The 6 activity trackers were each placed in a small manila envelope which was then sealed to prevent the patients in the study from seeing the device. This envelope was then placed in a transparent, hospital-sanctioned pharmacy bag and attached to the patient’s hospital gown using a safety pin. Activity trackers remained on the patient until discharge and were only removed when showering. Patients who removed their activity tracker before discharge were excluded from the data set. Upon discharge, the activity tracker was collected by the patient’s nurse and placed in a centrally located study box. Activity trackers were collected and charged weekly. Data were stored in a research electronic data capture (REDCap, Vanderbilt University, TN) database, which provides a secure web application for managing research databases. Patients’ demographic and operative information were collected through retrospective chart review of the electronic medical record.
Measures and analyses
Number of Steps Recorded
The number of steps taken was measured starting on postoperative day 0 (POD#0, the day of surgery) and continued until the time of discharge. Historically, most patients undergoing bariatric surgery in our practice are discharged home on the morning of postoperative day 2 (POD#2). Because of varying times of discharge, the 24-hour period of greatest consistency when comparing patients was the POD#1 data. This represented a complete 24-hour period of data for each subject in which ambulation was encouraged and unrestricted.
Subject and surgery characteristics
Patients’ characteristics included age (yr), sex (male versus female), and preoperative BMI (kg/m2), along with type of surgery (laparoscopic sleeve gastrectomy versus laparoscopic Roux-en-Y gastric bypass), length of stay (LOS) measured in days, and case order (first, second, third). First case surgeries were scheduled for 7 AM with second and third cases occurring consecutively. The order of the cases was determined by the surgeon.
Analyses
Descriptive statistics were used to summarize the sample including mean and standard deviation (SD). A t test was conducted to examine whether telling a patient that their steps were being recorded with an activity tracker would affect the number of steps recorded postoperatively. A multivariate Ordinary Least Squares regression model was used to examine the effect of age, sex, preoperative BMI, types of surgery, and LOS on number of steps recorded after checking residuals (e.g., quantile-quantile plot, residuals against fitted values) for main assumptions. There were no signs of violations of normality, linearity, homoscedasticity, or independence. Additionally, an ANOVA was used to compare the number of steps recorded on POD#0 between different case orders. All analyses were performed using Stata/MP version 15 software (StataCorp, College Station, TX).
Results
Two hundred sixty-6 patients were consented for the study. Thirteen patients were removed from the data set because of noncompliance with wearing the activity tracker, while 18 patients were excluded from the data set because of incomplete data extraction from the device. Two hundred and 35 patients (88% of included patients) successfully completed the study by maintaining the use of the activity tracker during the entirety of their hospital stay after bariatric surgery. Of the 235 patients, 185 (78.7%) were female sex and 204 (86.8%) underwent laparoscopic sleeve gastrectomy. The patients’ average age was 44 years (SD = 11) with an average preoperative BMI of 46 mg/kg2 (SD = 6). Among the total sample of 235 patients, 125 patients (52.8%) were randomized into the blinded group (unaware of the device’s intent) while 111 (47.2%) were informed that the device would record the number of steps taken. The average hospital stay was 2.7 days (SD = 0.1).
There were no demographic or operative differences between the blinded and unblinded groups (Table 1). As seen in Table 2, there were no statistically significant differences in the number of steps recorded on POD#0 and POD#1 between patients that knew they were wearing an activity tracker versus those blinded to the function of the device. No patient in the study suffered a documented deep vein thrombosis or venous thromboembolic event.
Table 1.
Descriptive statistics of sample
|
Mean (SD) or n (%) |
Statistics (Blinded versus known) | |||
|---|---|---|---|---|
| Overall (n = 235) | Blinded group (n = 124) | Known group (n = 111) | ||
| Age (yr) | 44.0 (11.1) | 44.8 (10.8) | 43.1 (11.3) | t = 1.2, P = .2 |
| Female sex | 185 (78.7%) | 95 (76.6%) | 90 (81.1%) | χ2 = .7, P = .4 |
| Preoperative BMI (kg/m2) | 46.5 (6.2) | 46.7 (6.4) | 46.4 (6.0) | t = .4, P = .7 |
| Sleeve gastrectomy | 204 (86.8%) | 112 (90.3%) | 92 (82.9%) | χ2 = 2.8, P = .1 |
| Length of hospital stay (d) | 2.7 (.1) | 2.7 (.1) | 2.7 (.1) | t = −.5, P = .6 |
SD = standard deviation; n = sample size; BMI = body mass index.
Table 2.
Comparison between blinded and activity tracker function known groups
|
Mean (SD) |
Statistics | ||
|---|---|---|---|
| Blinded group (n = 124) | Known group (n = 111) | ||
| Postoperative day 0 (steps recorded) | 1170.4 (1191.3) | 1752.9 (1463.1) | t = .1, P = .9 |
| Postoperative day 1 (steps recorded) | 1858.7 (1417.5) | 1926.8 (1700.3) | t = −.3, P = .7 |
SD = standard deviation; n = sample size.
Table 3 summarizes the results from multivariate regression analysis. After controlling for age, sex, preoperative BMI, surgery type, and knowledge of activity tracker function, the LOS data found a statistically significant correlation with POD#0 number of recorded steps. Patients with prolonged LOS recorded lower numbers of steps taken on POD#0. Increasing age was seen to reduce the number of steps recorded on POD#1. Of note, there were no significant differences in the number of steps recorded based on sex, preoperative BMI, or surgery type.
Table 3.
Multivariate regression models for patients recorded steps on postoperative days
|
Postoperative day 0 |
Postoperative day 1 |
|
|---|---|---|
| B (SE) | B (SE) | |
| Age (yr) | −9.8 (7.9) | −23.1 (9.4)* |
| Female | 148.7 (7.9) | −413.9 (251.3) |
| Preoperative BMI (kg/m2) | −25.2 (14.0) | −19.9 (16.6) |
| Sleeve gastrectomy | −202.4 (254.7) | −157.4 (299.9) |
| Length of hospital stay | −417.6 (117.7)† | −179.1 (144.2) |
| Pedometer known | −33.2 (171.2) | 82.3 (202.3) |
B = coefficient of variable; SE = standard error; BMI = body mass index.
P < .01
P < .001.
Table 4 summarizes the results from ANOVA of POD#0 recorded steps between subject groups depending on if their surgery was performed first, second, or third on the day of surgery. Five patients were excluded from this analysis because they were the fourth case for the day. The order in which the case was performed did not affect the number of steps taken on POD#0.
Table 4.
Comparison between case order and postoperative day 0 recorded steps
|
Mean (SD) |
Statistics | |||
|---|---|---|---|---|
| First case (n = 93) | Second case (n = 81) | Third case (n = 56) | ||
| Postoperative day 0 (steps recorded) | 1702.5 (1074.1) | 1709.1 (1232.2) | 1966.8 (1791.3) | F =.7, P =.6 |
SD = standard deviation; n = sample size.
Discussion
In the present study, we examined postoperative activity recorded in number of steps taken among patients undergoing bariatric surgery. Patients’ steps were recorded with the use of a wireless activity tracker that was attached to the patients’ hospital gown in a consistent location, and patients were randomized by whether the function of the device was revealed to the patient (treatment group) or concealed from the patient (control group). We found no difference in postoperative recorded step counts between the treatment and control groups. Use of a multivariate regression model revealed that LOS had a statistically significant inverse association with number of recorded steps on POD#0 (though the relationship was not significant for POD#1). The model also found a statistically significant inverse association between age and number of steps recorded on POD#1 (though the relationship was not significant for POD#0). We found no statistically significant relationship between BMI and number of steps recorded postoperatively.
The negative finding regarding our primary outcome suggests that recording postoperative activity with wearable commercial activity trackers is not likely to increase postoperative ambulation among patients undergoing bariatric surgery. However, it is possible, given the high variance in steps recorded across both the treatment and control groups, that our study was not able to detect such a difference. Nevertheless, there are several possible explanations for a negative result. First, the days immediately after bariatric surgery can be overwhelming. In this critical time, patients must balance provider expectations for nutrition and ambulation with postoperative pain and nausea, as well as expectations for weight loss. It is possible that adding activity trackers to encourage increased ambulation was simply not enough to overcome the cognitive noise of the postoperative period. Conversely, it is possible that ambulation was maximally emphasized by physicians, midlevel providers, and nursing staff, such that the addition of a wearable activity tracker made no difference. It is likely that activity counseling, which all patients in our study received, is more effective than the use of an activity tracker alone which has been shown previously in bariatric surgery patients in the outpatient setting [12].
Regression analysis found a statistically significant correlation between increased recorded step count on POD#0 and decreased hospital stay. There are any number of explanations for this finding: healthier patients may be both more likely to ambulate and to be discharged earlier; or ambulating may decrease hospital stay by decreasing postoperative complications, even subtle complications like atelectasis or nausea. Associations between ambulation and LOS has been shown in prior research [8]. Increasing age was also inversely associated with the number of recorded steps on POD#1. Prior research has shown that postoperative ambulation is inversely related to overall physical health, so this result is not unexpected [9]. In fact, decreased ambulation may be one reason for the increasing risk of deep vein thromboembolism and VTE with increasing age. The fact that this effect was shown for POD#1, rather than POD#0, may have to do with greater consistency of the data for POD#1.
Did the timing of the surgery effect the number of recorded steps on POD#0? To determine if this had an effect on the patients recorded step count, we compared POD#0 steps depending on case order. First cases were scheduled for 7 AM while second and third cases occurred later in the day. Interestingly, the timing of the surgery did not affect the number of steps taken on POD#0. Patients that had the operation early in the morning (first case) or later in the day (second or third case) had statistically similar recorded step counts on POD#0. Possible explanations for this finding include POD#0 ambulation being independent on duration of time within the measured range. Alternatively, confounding variables including a surgeon unconscious bias on the ordering of the cases may have influenced this result. While patients in the third case order had less available time for ambulation on POD#0, they may represent a healthier group compared to first start cases.
Surprisingly, this study did not show a statistically significant correlation between postoperative ambulation and BMI, a relationship that has appeared in prior studies [9]. This may be in part due to the low numbers of patients with a BMI >55 kg/m2 (n = 27, 11% of study patients), which may have made the study insufficiently powered to detect an association. Alternatively, differences in ambulation might not appear between different classes of obesity, but rather between patients with obesity and those without. Since this study included only patients with class II and III obesity, such a relationship would not appear in this study population.
Overall, this study was well-designed to ask and answer its narrowly defined question regarding the impact of using activity trackers to record the number of steps taken during the inpatient stay after bariatric surgery. However, this leaves broader questions surrounding postoperative ambulation unaddressed, which is reflected in a low R-squared (R2) value. The focus of the present study, effects of knowing an activity tracker is recording as measured by the number of steps taken, is one potentially important variable effecting patient behavior. A low R2 value is common for studies of human behaviors in humanities or social science because human behaviors cannot be predicted or explained often by a few variables captured in a research study. The low R2 of the present study can be improved by expanding more surrounding information.
Activity tracking ceased on discharge, raising the question of whether in-hospital activity correlates with outpatient activity in the postoperative period, and what the importance of home ambulation is for the postoperative patient. The study also leaves unaddressed the relationship of level of activity to postoperative complications such as deep vein thromboembolism or VTE. As these are rare complications (no patient in this study suffered from thrombus or thromboembolism), even large, well-powered studies have had trouble detecting this relationship in the past. Another limitation to be noted is that we excluded 18 patients because of incomplete data extraction from the activity tracker. There were no indications that the 18 patients were systematically different from the rest of sample, and we tried to reduce any potential effect related to inaccurate information by excluding them. However, we caution that the excluded patients might be different from the study sample although its potential unmeasured effect would be minimal with the small number of excluded patients.
Finally, though the study failed to show that activity trackers increased patient ambulation, there may yet be a role for activity trackers in the postoperative inpatient setting. Activity trackers have varying degrees of accuracy [13] and are often inexpensive. Though prior research has indicated that patients are wary of sharing data from wearable devices with their healthcare providers, compliance with activity trackers was high among our included patients [14]. This opens the door for further innovation: perhaps through real-time tracking with targeted intervention for low-ambulators; perhaps through hospital- or ward-level ambulation benchmarks and incentives; perhaps through improved prognostication by using activity tracker data with validated models. The applications of this technology in the field of bariatric surgery have only begun to be investigated. More studies about the wearable devices are necessary. Because the device is relatively new in this area, more rigorous testing on its validity and reliability is required. In addition to measurement testing, it would be beneficial to conduct a qualitative study with patients who wear the device. Through a focus group, future studies can capture participants’ experiences about the device with its strengths and weaknesses. Valid and reliable wearable devices established through further research will likely contribute to the field of bariatrics. Moreover, future studies including multi-informants (patient self-report, device measurements, and staff-report) and their potential different effects on ambulation would have important clinical implications. Unfortunately, the self-reported ambulation tracking data from the patients in this study was not retained for analysis to compare self-reported data versus activity tracker collected data.
Because of the narrow focus of the present study, we did not include other variables surrounding postoperative ambulation. Activity trackers can be used to better study the effects of co-morbidities, surgery duration, time of surgery (morning versus afternoon), and known orthopedic related disease on postoperative ambulation.
Conclusion
The average number of steps recorded in the inpatient setting with an activity tracker after bariatric surgery did not differ in patients that were aware versus unaware that their steps were being tracked. The study did not show a difference in the recorded number of steps with increasing BMI. Increasing age did correlate to reduced number of steps recorded on postoperative day one after bariatric surgery.
Acknowledgments
Eric Lundstrom, Michael DeRosa, and Nathan Cook participated in pedometer deployment and data collection for the study while they were students in the Master’s program in Health Sciences at West Virginia University. They have since graduated and we wish them well in their health sciences careers and continued interests in research.
Footnotes
Disclosures
The authors have no commercial associations that might be a conflict of interest in relation to this article.
References
- [1].Brieger GH. Early ambulation. A study in the history of surgery. Ann Surg 1983;197(4):443–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Ljungqvist O, Scott M, Fearon KC. Enhanced recovery after surgery: a review. JAMA Surg 2017;152(3):292–8. [DOI] [PubMed] [Google Scholar]
- [3].Oldmeadow LB, Edwards ER, Kimmel LA, Kipen E, Robertson VJ, Bailey MJ. No rest for the wounded: early ambulation after hip surgery accelerates recovery. ANZ J Surg 2006;76(7):607–11. [DOI] [PubMed] [Google Scholar]
- [4].Kamel HK, Iqbal MA, Mogallapu R, Maas D, Hoffmann RG. Time to ambulation after hip fracture surgery: relation to hospitalization outcomes. J Gerontol A Biol Sci Med Sci 2003;58(11):M1042–5. [DOI] [PubMed] [Google Scholar]
- [5].Allman-Farinelli MA. Obesity and venous thrombosis: a review. Semin Thromb Hemost 2011;37(8):903–7. [DOI] [PubMed] [Google Scholar]
- [6].Hensley L, Romano PS, White RH, et al. Mechanical and suboptimal pharmacologic prophylaxis and delayed mobilization but not morbid obesity are associated with venous thromboembolism after total knee arthroplasty: a case-control study. J Hosp Med 2012;7(9). [DOI] [PubMed] [Google Scholar]
- [7].Callen BL, Mahoney JE, Grieves CB, Wells TJ, Enloe M. Frequency of hallway ambulation by hospitalized older adults on medical units of an academic hospital. Geriatr Nurs 2004;25(4):212–7. [DOI] [PubMed] [Google Scholar]
- [8].Daskivich TJ, Houman J, Lopez M, et al. Association of wearable activity monitors with assessment of daily ambulation and length of stay among patients undergoing major surgery. JAMA Netw Open 2019;2(2):e187673. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Twiggs J, Salmon L, Kolos E, Bogue E, Miles B, Roe J. Measurement of physical activity in the pre- and early post-operative period after total knee arthroplasty for osteoarthritis using a Fitbit Flex device. Med Eng Phys 2018;51:31–40. [DOI] [PubMed] [Google Scholar]
- [10].King WC, Hsu JY, Belle SH, et al. Pre- to postoperative changes in physical activity: report from the Longitudinal Assessment of Bariatric Surgery-2 (LABS-2). Surg Obes Relat Dis 2012;8:522–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].FitBit™. How does my Fitbit device calculate my daily activity?. [cited April 13, 2020]. Available from: https://help.fitbit.com/articles/en_US/Help_article/1141#steps.
- [12].Creel DB, Schuh LM, Reed CA, et al. A randomized trial comparing two interventions to increase physical activity among patients undergoing bariatric surgery. Obesity 2016;24(8):1660–8. [DOI] [PubMed] [Google Scholar]
- [13].Freehan LM, Geldman J, Sayre EC, et al. Accuracy of Fitbit devices: systematic review and narrative syntheses of quantitative data. JMIR Mhealth Uhealth 2018;6(8):e10527. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Pevnick JM, Fuller G, Duncan R, Spiegel BMR. A large-scale initiative inviting patients to share personal fitness tracker data with their providers: initial results. PLoS One 2016;11(11). [DOI] [PMC free article] [PubMed] [Google Scholar]