This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR mhealth and uhealth, is properly cited. The complete bibliographic information, a link to the original publication on http://mhealth.jmir.org/, as well as this copyright and license information must be included.
Rehabilitation is crucial for postoperative patients with low back pain (LBP). However, the implementation of traditional clinic-based programs is limited in developing countries, such as China, because of the maldistribution of medical resources. Mobile phone–based programs may be a potential substitute for those who have no access to traditional rehabilitation.
The aim of this study was to examine the efficacy of mobile phone–based rehabilitation systems in patients who underwent lumbar spinal surgery.
Patients who accepted spinal surgeries were recruited and randomized into 2 groups of rehabilitation treatments: (1) a mobile phone–based eHealth (electronic health) program (EH) or (2) usual care treatment (UC). The primary outcomes were (1) function and pain status assessed by the Oswestry Disability Index (ODI) and (2) the visual analog scale (VAS). Secondary outcomes were (1) general mental health and (2) quality of life (Likert scales, EuroQol-5 Dimension health questionnaire, and 36-item Short-Form Health Survey). All the patients were assessed preoperatively and then at 3, 6, 12, and 24 months postoperatively.
A total of 168 of the 863 eligible patients were included and randomized in this study. Our analysis showed that the improvement of primary outcomes in the EH group was superior to the UC group at 24 months postoperatively (ODI mean 7.02, SD 3.10,
A subgroup analysis showed that the improvements of the primary outcomes were more significant in those who completed 6 or more training sessions each week throughout the trial (the highest compliance group) compared with the UC group at 6 months (ODI mean 17.94, SD 5.24,
This research demonstrated that a mobile phone–based telerehabilitation system is effective in self-managed rehabilitation for postoperative patients with LBP. The effectiveness of eHealth was more evident in participants with higher compliance. Future research should focus on improving patients’ compliance.
Chinese Clinical Trial Registry ChiCTR-TRC-13003314; http://www.chictr.org.cn/showproj.aspx?proj=6245 (Archived by WebCite at http://www.webcitation.org/766RAIDNc)
Low back pain (LBP) is a common health problem with a point prevalence of 15% and a lifetime incidence as high as 85%. LBP is related to disability, work loss, and also accounts for high economic costs in society. For example, total annual back pain–related costs in the United States exceed US $200 billion and are still increasing [
Surgical intervention is an important treatment for LBP [
Recent evidence suggests that patients with low back pain usually have inherent muscle dysfunction, which may be exacerbated by surgeries. Thus, postoperative rehabilitation becomes very critical [
The rapid development of mobile phone–based programs provides a new option to promote health and prevent diseases [
This trial was conducted to investigate whether a mobile phone–based program (electronic health; eHealth), designed to provide telerehabilitation for patients with LBP, would reduce pain-related disability and improve prognosis among postoperative patients who have no access to traditional clinic-based rehabilitation.
This study was a multicenter, prospective, randomized controlled trial, approved by the Ethics Committees of the Sun Yat-sen Memorial Hospital. All the 3 hospitals that participated in this study were affiliated to the Sun Yat-sen University, where the surgery could be carried out safely and skillfully. All the patients were assessed for postoperative functional ability, pain, and general mental and health status at baseline and 3, 6, 12, and 24 months.
The researchers were required to explain the purposes, procedures, and possible risks of the trial in detail to the patients before inclusion. Written informed consents were obtained from all patients. The inclusion and exclusion criteria are shown in
On the basis of previous studies, we anticipated that to have a 90% chance of detecting a between-group difference of 8 points on the Oswestry Disability Index (ODI) and declaring it statistically significant using a two-sided alpha=.05, an enrollment of 168 patients was required. This calculation allowed for a loss to follow-up of 23% [
After completing the baseline survey, each participant was randomly allocated in a 1:1 ratio to the mobile phone–based eHealth program (EH) group or usual care treatment (UC) group according to a computer-generated randomization list. The allocation was stratified by a surgeon, operative procedures, and preoperative diagnosis. An email was sent to participants to inform them about their group assignment. The allocation sequence was concealed from the researchers enrolling and assessing patients.
Inclusion criteria
Aged between 18 and 64 years
Agreed to receive lumbar spinal surgery and that the surgical intervention involved no more than 3 columns
Diagnosed as lumbar disc herniation, spinal stenosis, or lumbar spondylolisthesis with imaging support
Living at least 100 km or a 2 hours’ drive away from the hospitals
Signed the informed consent
Exclusion criteria
Diagnosed as tuberculosis and tumor patients
Those who accepted lumbar surgeries before this trial
Patients with rheumatoid arthritis or ankylosing spondylitis
Pregnancy
Those who could not sign the informed consent or complete the rehabilitation exercise because of mental retardation or other reasons
No specific rehabilitation program was provided to patients randomized to the UC control group. The relevant surgeons’ usual practice was still provided, including advices to keep physically active and simple instructions to train the back muscles. Analgesia and other symptomatic treatments were also provided when necessary. All the postoperative regimes were documented.
Besides the relevant surgeons’usual practice, patients randomized to the intervention group received telerehabilitation provided by eHealth, a mobile phone–based system developed by our group.
Moreover, eHealth was designed based on the user-centered theory, aimed to provide a platform for the delivery of self-management interventions [
The exercises included in this software were designed based on core stability exercise principles, which were all aimed to restore normal muscle strength and mobility, to activate the deep core musculature and to promote balance and coordination of the patients’ daily movements. The detailed plan of rehabilitation is shown in
The validation study was conducted with 10 healthy adults. Then, the information for the usability of the system was collected through paper-based questionnaires, 1 day after the tryout. The validation study confirmed that our system was well designed and easy to use, and the rehabilitation guidance was easy for users to understand. Combined with the results of previous studies and user preferences [
The software was installed into the patients’ phones 3 months after the surgery. Two meetings were held to show the patients how to use this software and how to conduct the exercises. The patients were also evaluated to make sure they can conduct the rehabilitation exercise correctly. They were required to complete at least 2 months of training. After 2 months, the patients could still log on to the system, and those who completed 5 or more training sessions each week were considered as high adherence, 3 to 5 training sessions as medium adherence, and 2 training sessions and less as low adherence.
The primary outcome measures were the ODI, a disease-specific questionnaire documenting the function of known validity and reliability, and the visual analog scale (VAS) to record back pain [
At 12 months postoperative, an open survey was also conducted to detect the factors that affect patient compliance. All patients with medium and low compliance were asked to list 3 of the most important factors that they thought affected their compliance to the system.
The eHealth system contained 2 interfaces: mobile phone–based interface and Web-based interface. Through mobile phone–based interface, patients were able to view the rehabilitation plans and conduct their rehabilitation following the video instructions. Daily reports and an alert were sent to prompt them to return to this system. They could also communicate with doctors through this system. Through Web-based interface, the doctors could adjust rehabilitation plans for patients and view reports about their daily exercise. All data were synchronized and stored in a remote server.
The analyst assessing trial outcomes was blinded to the assignments. All analyses were conducted using an intent-to-treat approach with participants analyzed according to original group assignment. The baseline data for those lost to follow-up were included. Baseline characteristics were compared between the groups using chi-square tests for categorical data and a 2-sample
For analyses of primary and secondary outcomes, a paired
All the analyses were conducted using Stata version 23.0 (StataCorp LLC) and a
Recruitment occurred between August 2013 and November 2014 at 3 hospitals, and 845 patients were assessed for eligibility. Of those, 428 patients were excluded for not meeting the inclusion criteria or meeting the exclusion criteria. Of the 417 eligible patients, 92 were not approached, 135 declined to participate, and 22 consented patients withdrew before randomization. The final 168 consenting patients were then randomized in this study.
All the randomized patients received operation treatments and completed the required baseline assessments. However, during the study, 2 patients in the EH group and 4 patients in the UC group dropped out at 3 months. From the EH group, 82 patients entered the treatment phase, of which 77 finished the treatment and follow-up was done at 6 months. In the UC group, 80 and 74 patients were met for follow-up at 3 months and 6 months, respectively. The follow-up rate in the EH group was 97.62% (82/84) at 3 months, 91.67% (77/84) at 6 months, 85.71% (72/84) at 12 months, and 71.43% (60/84) at 24 months. In UC group, the follow-up rate was 95.24% (80/84) at 3 months, 88.10% (74/84) at 6 months, 83.33% (70/84) at 12 months, and 72.62% (61/84) at 24 months (see
Both the clinical and demographic characteristics of the patients were similar in the 2 groups (
Flowchart.
Median eHealth attendance was 5 times per week (interquartile range, IQR, 4-6),5 times per week (IQR 3-6), and 5 times per week (IQR 4-6) for 6, 12, and 24 months postoperatively. A total of 50, 37, and 38 patients were considered as high compliance at 6, 12, and 24 months, respectively, postoperatively. Although the high compliance rate was higher at 24 months (63.33%) compared with that at 12 months postoperatively (51.39%), it was not statistically significant (
To determine the reason for low compliance, we carried out a brief survey focusing on the medium and low compliance group, asking them to list out the top 3 factors that affected their adherence at 12 months. A total of 33 out of 35 patients considered lack of communication with their doctors as the important factor. Through our record, we found that mean communication frequency was 2.54 (SD 0.89) for patients with medium or low compliance and 4.46 (SD 1.35) for patients with high compliance. The frequency of responses from doctors was significantly higher in patients with high compliance (
The ODI and VAS for the EH and UC group were similar at baseline and 3 months postoperatively. At 6 and 12 months, the mean for change of the ODI from baseline was –7.27 (SD 5.31) and –18.43 (SD 23.92),respectively, for the EH,while it was –7.90 (SD 4.53) and –14.39 (SD 4.64), respectively, for the UC group (see
Demographics and baseline characteristics of all participants.
Characteristics | UCa (n=84) | EHb (n=84) | ||
Female, n (%) | 42 (50) | 48 (57) | .35 | |
Age (years), mean (SD) | 49.36 (9.52) | 51.11 (9.54) | .24 | |
High school or lower | 63 (75) | 60 (71) | .60 | |
College degree or higher | 21 (25) | 24 (29) | .60 | |
Currently employed | 62 (74) | 58 (69) | .50 | |
Married | 76 (90) | 73 (87) | .72 | |
Divorced | 5 (6) | 6 (7) | .72 | |
Single | 3 (4) | 5 (6) | .72 | |
1 disc | 37 (44) | 36 (43) | .96 | |
2 discs | 41 (49) | 41 (49) | .96 | |
3 discs | 6 (7) | 7 (8) | .96 | |
Mean ODIc score (SD)d | 55.40 (14.78) | 54.14 (15.18) | .59 | |
Mean VASe score (SD)f | 60.11 (15.99) | 57.71 (14.91) | .32 | |
Mean Likert score (SD)g | 59.14 (14.86) | 59.71 (16.49) | .40 | |
Mean EQ-5Dh score (SD)i | 35.75 (15.37) | 34.26 (14.84) | .32 | |
Mean SF-36 GHj score (SD)k | 13.55 (5.58) | 13.60 (6.02) | .96 | |
Mean SF-36 PFl score (SD)m | 21.11 (8.36) | 21.52 (8.72) | .75 |
aUC: usual care.
bEH: eHealth program.
cODI: Oswestry Disability Index.
dRated to assess the patient’s level of disability because of low back pain. Ranged from 0 to 100, with higher scores indicating more disability.
eVAS: visual analog scale.
fRated between 0 and 100 with 100 representing worst pain possible.
gMeasured using an 11-point numerical rating scale for average difficulty for movement in the previous week, where 0 indicated no difficulty and 10 indicated most difficulty.
hEQ-5D: EuroQol 5-Dimension health questionnaire.
iRated between 0 and 100 with 100 representing a perfect health-related quality of life.
jSF-36 GH: General health for 36-item Short-Form Health Survey.
kRanging from 0 to 100, with higher scores indicating better health-related quality of life.
lSF-36 PF: Physical functioning for 36-item Short-Form Health Survey.
mRanging from 0 to 100, with higher scores indicating better health-related quality of life.
Compliance status in different follow-ups.
Follow-up (month) | Low compliance, n (%) | Medium compliance, n (%) | High compliance, n (%) | |
6 | 11 (14.47) | 16 (19.74) | 50 (65.79) | .23 |
12 | 8 (11.11) | 27 (37.50) | 37 (51.39) | .23 |
24 | 6 (11.48) | 16 (26.23) | 38 (62.29) | .23 |
Primary outcomes change from baseline and between-group difference.
Measurements and follow-up (month) | UCa change from baseline | EHb change from baseline | Difference UC versus EH, mean (SD) | ||||
Participants | Mean (SD) | Participants | Mean (SD) | ||||
3 | 80 | –7.90 (4.53) | 82 | –7.27 (5.31) | –0.63 (0.78) | .42 | |
6 | 74 | –14.39 (4.64) | 77 | –18.43 (23.92) | 4.0 (2.83) | .16 | |
12 | 70 | –22.07 (5.56) | 72 | –21.58 (24.64) | –0.49 (2.98) | .87 | |
24 | 61 | –23.41 (6.65) | 60 | –30.43 (23.75) | 7.02 (3.18) | .03 | |
3 | 80 | –7.61 (5.15) | 82 | –7.02 (4.45) | –0.59 (0.76) | .44 | |
6 | 74 | –14.19 (5.11) | 77 | –17.49 (25.48) | 3.30 (2.96) | .27 | |
12 | 70 | –21.94 (5.8) | 72 | –20.55 (25.92) | –1.39 (3.13) | .66 | |
24 | 61 | –22.36 (6.90) | 60 | –29.95 (25.60) | 7.59 (3.42) | .03 |
aUC: usual care.
bEH: eHealth program.
cODI: Oswestry Disability Index.
dVAS: visual analog scale.
However, at 24 months, improvement in the ODI was more significant in the EH group compared with the UC group (
No difference in the Likert scale for movement was found at 3, 6, and 12 months postoperatively in the EH and UC. At 24 months, patients in the EH displayed superior results of Likert scale (mean of change: EH –32.51, SD 25.94; UC –22.54, SD 5.81;
As for the EuroQol-5 Dimension (EQ-5D), the change was similar for EH and UC at 3 months. At 6 months, the improvement for EQ-5D was 0.23 (SD 0.03) for the EH and 0.13 (SD 0.08) for the UC. The patients in the EH got a significantly superior result over that in the UC. This advantage was sustained at subsequent time points (
For the SF-36, the improvement was more significant in the EH compared with the UC at 3, 6, and 24 months (
In the EH group, 24 patients completed all the follow-ups, with average eHealth attendance no less than 6 times per week, considered as the highest compliance. Thus, we conducted a subgroup analysis between the HC group with UC group.
Both the clinical and demographic characteristics were consistent between the 2 groups at baseline (see
Adverse events, mostly mild, self-limited joint and back pain, were reported in 9 EH and 6 UC participants. It did not differ significantly in frequency or severity of adverse events in these 2 groups.
Secondary outcomes change from baseline and between-group difference outcomes.
Measurements and follow-up (month) | UCa change from baseline | EHb change from baseline | Difference UC versus EH, mean (SD) | ||||||
Participants | Mean (SD) | Participants | Mean (SD) | ||||||
3 | 80 | –7.79 (4.96) | 82 | –7.20 (4.74) | 0.20 (0.78) | .80 | |||
6 | 74 | –13.51 (5.39) | 77 | –19.66 (26.47) | 6.15 (3.08) | .05 | |||
12 | 70 | –21.09 (5.68) | 72 | –23.00 (27.12) | 1.91 (3.31) | .56 | |||
24 | 61 | –22.54 (5.81) | 60 | –32.51 (25.94) | 9.98 (3.43) | .01 | |||
3 | 80 | 0.09 (0.02) | 82 | 0.09 (0.02) | 0.00 (0.00) | .62 | |||
6 | 74 | 0.13 (0.08) | 77 | 0.23 (0.03) | –0.10 (0.01) | <.001 | |||
12 | 70 | 0.17 (0.03) | 72 | 0.24 (0.04) | –0.05 (0.01) | .003 | |||
24 | 61 | 0.22 (0.04) | 60 | 0.35 (0.03) | –0.12 (0.01) | .001 | |||
3 | 80 | 38.16 (2.43) | 82 | 40.01 (3.37) | –1.85 (0.46) | .004 | |||
6 | 74 | 45.85 (3.43) | 77 | 54.75 (4.59) | –8.90 (0.66) | .002 | |||
12 | 70 | 55.53 (3.86) | 72 | 56.25 (5.31) | –0.72 (0.78) | .36 | |||
24 | 61 | 57.98 (5.26) | 60 | 62.80 (6.61) | –4.82 (1.09) | .002 | |||
3 | 80 | 30.58 (2.29) | 82 | 40.76 (3.05) | –2.18 (0.42) | .004 | |||
6 | 74 | 46.35 (3.62) | 77 | 56.12 (4.48) | –9.77 (0.66) | .003 | |||
12 | 70 | 56.13 (4.79) | 72 | 56.74 (5.83) | –0.61 (0.89) | .49 | |||
24 | 61 | 59.07 (5.89) | 60 | 62.45 (5.78) | –3.38 (1.06) | .02 |
aUC: usual care.
bEH: eHealth program.
cEQ-5D: EuroQol 5-Dimension health questionnaire
dSF-36 GH: General health for 36-item Short-Form Health Survey
eSF-36 PF: Physical functioning for 36-item Short-Form Health Survey
Subgroup analysis of primary outcomes change from baseline and between-group difference outcomes.
Measurements and follow-up (month) | UCa change from baseline | HCb change from baseline | Difference UC versus HC, mean (SD) | ||||
Participants | Mean (SD) | Participants | Mean (SD) | ||||
3 | 80 | –7.90 (4.53) | 24 | –7.71 (5.29) | –0.19 (1.10) | .86 | |
6 | 74 | –14.39 (4.64) | 24 | –32.33 (25.56) | 17.94 (5.24) | <.001 | |
12 | 70 | –22.07 (5.56) | 24 | –35.46 (25.88) | 13.39 (5.32) | .02 | |
24 | 61 | –23.41 (6.65) | 24 | –42.21 (25.26) | 18.80 (5.22) | .01 | |
3 | 80 | –7.61 (5.15) | 24 | –6.42 (4. 91) | –1.20 (1.19) | .32 | |
6 | 74 | –14.19 (5.11) | 24 | –33.75 (25.67) | 19.56 (5.27) | .01 | |
12 | 70 | –21.94 (5.8) | 24 | –36.29 (25.38) | 14.35 (5.23) | .01 | |
24 | 61 | –22.36 (6.90) | 24 | –43.92 (25.50) | 21.56 (5.28) | .001 |
aUC: usual care.
bHC: highest compliance.
cODI: Oswestry Disability Index.
dVAS: visual analog scale.
Much research had been done to explore the effect of rehabilitation on postoperative patients [
In this randomized controlled trial, we compared postoperative patients (EH group) with low back pain treated by eHealth, a mobile phone–based telerehabilitation system, with those that received nonspecific rehabilitation (UC group). We found that primary outcomes (ODI and VAS) in the EH group were superior to the UC group at 24 months postoperatively. However, no significant difference was found at all the other time points during follow-up. Furthermore, we compared 24 patients having an average eHealth attendance of no less than 6 times per week (HC group) with the UC group. Subgroup analysis showed that the improvements of the primary outcomes were more significant in the HC group compared with the UC group at 6, 12, and 24 months. These results suggest that patients with a higher compliance with our telerehabilitation system tend to have a better prognosis.
Adherence to postoperative rehabilitation in clinical practice is a serious problem [
This study also found that the 24 patients who completed 6 or more training sessions each week throughout the trial (HC group) had a better prognosis. This may be because the rehabilitation exercise needs to reach a certain length of time to achieve a more significant effect [
The abovementioned features could be achieved by optimizing the designs of a mobile phone–based system. Previous studies have revealed that the mobile phone is an effective tool in improving health behaviors of patients [
The main limitation of this study was the high loss to follow-up rate. Compared with previous studies, the loss to follow-up rates were similar to the loss to follow-up rates in this study [
The reason for the high loss to follow-up rate was probably because of the long distance from the patients’ home to the hospital, which was 1 of the inclusion criteria. These inclusion criteria are consistent with our aim, which focused on remote self-rehabilitation. However, it also adds to the difficulty for following up. As our study was conducted with paper-based questionnaires, patients included in our study had to travel a long distance back to hospitals or mail back the questionnaires, causing more to be lost to follow-up. In order to overcome this problem, electronic surveys have great potential to improve data collection [
In conclusion, eHealth, a mobile phone–based telerehabilitation system, may be an effective rehabilitation tool for postoperative patients with LBP, especially for those who have no access to traditional clinic-based rehabilitation. The effectiveness of eHealth was more evident in patients with higher adherence. However, more studies are still needed to find optimal methods to improve compliance.
The detailed plan of rehabilitation.
Results of the validation study.
Supplementary tables.
CONSORT‐EHEALTH checklist (V 1.6.1).
eHealth program
electronic health
EuroQol 5-Dimension health questionnaire
highest compliance
interquartile range
low back pain
Oswestry Disability Index
36-item Short-Form Health Survey
General health for 36-item Short-Form Health Survey
Physical functioning for 36-item Short-Form Health Survey
usual care treatment
visual analog scale
This study was financially supported by the Guangdong Medical Research Fund Project (C2015048) and Guangdong Natural Science Foundation (2015A030310321). Phei Er Saw proofread and improved the language of this paper.
None declared.