Low Back Pain (LBP) is a leading cause of disability globally, making it a serious public health concern. In this paper we present the initial results and analysis of a longitudinal self-tracking study on sleep and LBP using a custom built mobile application. We designed and deployed a mobile app for a period of 7 months to collect daily and monthly sleep and low back pain data using custom and standardized questionnaires. We discuss the feasibility of our approach for longitudinal data collection. Our data analysis reveals heterogeneity in user perceptions of factors that afect their sleep and LBP. Combining our quantitative and qualitative analyses, we contribute to literature on sleep and LBP.
custom-built mobile application “Sleep Better with Back
The global population of people aged 60 and above has Pain” to investigate if, and in what ways, the felt pain
been growing steadily. Aided by advances in technology, afects the quality of sleep among people living with LBP.
healthcare delivery and the medical field, the average life Our main contributions are:
expectancy seems to be increasing. The longer people
live, the more susceptible they become to conditions such 1. A contribution to the literature on sleep and LBP,
as Low Back Pain (LBP) [
Sleep is a vital biological function essential for the deployments.
overall health of mammalian organisms [6] and allows 4. We discuss how factors such as sleep time, wake
for psychological recuperation and memory consolida- up time, sleep disruptions, and pain intensity at
tion [
We conducted a long term study that ofered daily cerning the number of completed daily surveys about observations from the perspectives of people living with sleep and LBP for several months, which is promising LBP to investigate the efects of LBP on sleep: how does considering our future deployment of the next version back pain afect one’s sleep quality? In this vein, we also of the application. consider how factors such as time of sleep, wake up time, and sleep interruptions among others afect the perceived 2. Related Work sleep quality and/or felt pain intensity of people living with LBP.
Low Back Pain (LBP) adversely afects the quality of life
of people [11] and produces annual global costs
exceeding $100 billion [12]. Although there are numerous pain
conditions, back pain carries the greatest societal burden,
afecting nearly every adult at least once in their lifetime
[12]. The Global Burden of Disease (GBD) study esti- 3.1. On-boarding and Daily Use
mated that among 306 conditions, LBP ranked highest as
the most significant condition with respect to the number
of years lived with disability [
users. It has a welcome screen (Fig. 1a) that on-boards
users by asking preliminary questions about user
demographics and past experience with pain. This welcome
2.2. Sleep screen also displays a consent checkbox which the user
Recent studies have reported that sleep dificulties such must agree to in order to proceed. A consent statement
as initiating and maintaining sleep, as well as experienc- is provided beside the checkbox and contains links to
ing inadequate sleep, can afect 20-30% of the western the “user agreement” and “privacy policy” screens.
Afpopulation on a daily to weekly basis [
The “Sleep Better with Back Pain” app includes three in-app questionnaires: daily survey, sleep survey, and quality of life survey. The “daily” survey is taken daily while the “sleep survey” and “quality of life survey” are taken monthly.
The exact mechanisms underpinning the sleep and pain
relationship are unclear but likely to involve multiple
contributors [19], such as depression [20], and daytime
napping [21]. The comorbidities between chronic pain and
sleep impairments have been shown in various studies
[
We designed a cross-platform mobile application to collect periodical questionnaire data. The application was built with Flutter1 and runs on both Android and iOS mobile operating systems. Flutter is a cross-platform mobile application development framework that makes it easy to build apps for the Android and iOS mobile operating systems using the same code base. • Sleep time and wake up time, 5-point Likert-style item: 1 (Earlier than normal) - 5 (Later than normal) • Time to fall asleep, 5-point item: 1 (Less time than
normal) - 5 (More time than normal) • Number of times waking up at night, 5-point item:
1 (Less than normal) - 5 (More than normal) • How rested you feel after waking up, 5-point item: 1 (Less rest than normal) - 5 (More rest than normal) • Severity of LBP pain, 1-point item: 0 (No pain)
10 (Worst imaginable pain) • How pain afected sleep and/or how sleep afected pain: Free-form text (a) On-boarding (b) Home screen (c) Settings screen (d) Daily survey (e) Sleep survey (f) Quality of life 3.1.2. Monthly Questionnaires sleep duration, sleep eficiency, sleep disturbances, sleep medication, and daytime dysfunction.
The monthly sleep questionnaire is a standardized Pitts- The quality of life survey is a Patient-Reported Outburgh Sleep Quality Index (PSQI) questionnaire devel- comes Measurement Information System (PROMIS-10) oped by Buysee et al. [26] as a self-report on subjective standardized questionnaire comprising of 10 global items sleep quality over a period of four weeks. It consists of that each represent a domain of health [27]. Global phys19 items from which seven component scores that re- ical and mental health component scores are computed veal the severity of various sleep problems are formed. from the global items. Results from PROMIS-10 can be These components include: sleep quality, sleep latency, used to assess patients’ perceptions of their health. The physical health component scale comprises of four items on: physical health, physical functioning, pain intensity, and fatigue. The mental health component scale includes items on: quality of life, mental health, satisfaction with social activities and relationships, and emotional problems. Two of the global items (general health and social roles) are not used in the calculation of the physical and mental component scores.
A total of 35 people participated in our field study. Participants were mainly recruited from our local campus through our campus-wide email lists. However, we got additional participants via Google’s Play Store as the app was published publicly there. Participant ages ranged between 20-68 (M = 38.03, SD = 12.94) with 20 females and 15 males. Participants had diverse academic qualiifcations and employment statuses. 20 hold a Master’s degree or higher, 8 hold a Bachelor’s degree, 2 hold a 3.1.3. Reminder Notifications Vocational degree, and 3 hold a High School diploma Notifications are delivered from our custom-built back- while 3 did not disclose their educational information. end application through OneSignal 2 APIs. Data from the Concerning employment, 21 participants identified as answered surveys are saved in an online Google Sheets having a full-time job, 7 have a part-time job while 3 are document. User data is saved in a database hosted on our unemployed, 2 are retired, and 2 did not disclose their own servers. The “Sleep Better with Back Pain” mobile employment status. None of the participants have had app is hosted on Google Play. their back surgically operated by a doctor and the number of years lived with LBP ranged from 0-38 (M = 9, SD 3.2. Pilot Experiment = 9.48). 20 participants have been clinically diagnosed with LBP, 8 have chronic sciatica, 5 non-chronic sciatica, Before embarking on the longitudinal deployment, we and 20 have no sciatica. validated the feasibility of the app by conducting usabil- We also inquired about how active participants’ ity tests with four users within our research premises. lifestyle was using a single 0-10 item (0 - not at all acDuring this session, we discovered some usability issues: tive to 10 - extremely active). The mean value was 5.51 1) mandatory questions were not marked with an asterisk (SD = 1.93). For inquiries about treatments and self(*), 2) after completing a survey, the user is automatically management techniques used in managing their pain, directed to the main screen without any message to indi- participants shared a variety of actions they take. Some cate the survey was complete, and 3) the choice of font of these include: “exercises”, “yoga”, “massages”, “stretchtype and color made visualization dificult. We also no- ing”, and “taking medication”. A more nuanced analysis ticed technical issues such as 1) users finding it dificult of these is outside the scope of this article. to select 0 on the 0-10 slider scale, and 2) the keyboard not disappearing after users have finished typing an answer. These insights were useful in improving the app 5. Results design and functionality before embarking on the longer deployment.
After making modifications to the app based on feedback from the usability testing session, we launched the app on the Google Play Store and published the study to our campus wide email lists. We did not specify any reward for participating in the study as we wanted to reach people who really wanted to use the app or were not motivated by any form of reward but were interested in understanding their sleep and LBP. We also reached out and recommended the study to previous participants of studies conducted by co-authors. The study run for a period of seven (7) months, from May to November 2021.
In the following sections, we present our quantitative and qualitative findings, focusing primarily on the efect of low back pain on the quality of sleep experienced by people living with LBP.
5.1.1. Relationship between sleep quality and pain intensity We analysed the daily surveys (N = 1030) and computed Spearman’s rank correlation to assess the relationship between the collected variables and “well-restedness”. We noticed a negligible negative correlation between pain intensity and well-restedness (r(1028) = -.12, p < .005) and a negligible correlation between wake up time and well-restedness ( r(1028) = .08, p = .01). Thus, the felt pain was not strongly associated with how well-rested participants felt the next morning. We also discovered a weak negative correlation between sleep time and wellrestedness (r(1028) = -.22, p < .005). However, there was a strong negative correlation between sleep disruption and well-restedness (r(1028) = -.41, p = < .005). Thus, and as can be expected, sleep disruptions are associated with how well-rested people feel in the morning. While these ifndings are not surprising, they are indicative of the app’s feasibility in overall data collection. 5.1.2. PSQI The global PSQI score from the responses (N = 36) of our study, which is the sum of all seven component scores has a range of 0-21 points with actual scores ranging from 4-16, an overall group mean of 8.4 (SD = 2.83, = .68). For the individual components, each with a possible range of 0-3, the ranges were 0-3 except sleep disturbance which had a range of 0-2. For each component, as well as the global PSQI score, higher scores are an indication of worse sleep quality.
In Table 1, we examine the correlation of the PSQI components in our cohort using Spearman rank correlation. We found strong correlations between sleep quality and use of sleep medication (r = .48, p < 0.005), sleep eficiency and sleep duration (r = .61, p < 0.005), a moderate correlation between sleep quality and sleep duration (r = .36, p < 0.05), sleep disturbance and daytime dysfunction (r = .34, p < 0.05). Sleep quality had little to no efect on daytime dysfunction (r = -.06, p = 0.71) however. 5.1.3. PROMIS-10
PROMIS-10 domains of the results (N = 36) of our study in Table 2. The mean physical health score of participants was 13.67 (SD = 3.01) and the physical health T-score of participants was lower (mean = 44.63, SD = 8.43) relative to the standardized mean T-score of 50. Similarly, the mean mental health score was 11.92 (SD = 3.23) and the mental health T-score of participants was lower (mean = 43.33, SD = 8.24) relative to the standard score. Thus our participants have significantly lower physical and mental function compared to the general population.
We analyzed 1030 responses to the open-ended daily questionnaire item (i.e. users’ own elaboration on how pain afected sleep or how sleep afected pain) following a deductive thematic analysis process [28]. In using this analytical method, the first author first coded the most informative responses, generating the coding scheme which was then shared with the other co-authors. The authors held online meetings to discuss and resolve disagreements with the coding. They further identified aspects of the responses that would be most interesting to present and discuss. In subsequent sections, we present a subset of the findings from the analysis that we believe best fits the scope of this article. 5.2.1. Pain before, during, and after sleep Participants expressed various degrees of pain felt prior to going to sleep, during sleep hours, and when they woke up in the morning. While there were reported feelings of pain prior to going to bed on some days, pain before bed did not always result to a painful sleep. One participant notes it was “...dificult to fall asleep due to the pain but once I did, it didn’t bother my sleep” (Female, 42).
In a similar vein, another participant exclaimed, “...it was worst when going to bed, it did not bother me during the night”(Male, 30).
On the contrary, some participants were quick to highlight the efect their LBP had on their sleep such as forcing them to stay awake at night for hours. One participant notes, “I stayed up two hours at night because of pain in my hip” (Male, 49) and “I couldn’t rest well due to pain at my back” (Male, 30). There were days however where participants simply reported their LBP not having an effect on their sleep despite the presence of the pain. One participant stating that the “...pain did not disturb my sleep” (Female, 48).
It is also worth highlighting that good quality sleep positively afected participants’ perception of pain in the morning, “...sleeping helped pain” (Female, 48). This was further supported by another who acknowledges to be 5.2.3. Exercises and active lifestyle “t.h..ainn ausbuitalle”s(sFpeaminalteh, i6s7m)ionrdniicnagtibnegcaaupseotIesnletpiatlalybpitobsiettitveer Some participants highlight being engaged in an active efect sleep can have on one’s LBP. lifestyle and sports. To this, we noticed a report of an
While we see several instances of sleep aefcting pain increase in pain or sleeplessness due to pain after an positively and pain having a negative efect on partici- intense or strenuous workout, “In the morning I felt some pants’ sleep, we also report that there were days where lower back pain and stifness. It could be result from gym participants were confident that neither sleep nor pain exercise” (Male, 38). Interestingly, sufering from LBP had an efect on the other. One participant sums it all did not stop participants from indulging in an active up, “I didn’t feel the efect from both sides ” (Female, 56). lifestyle (e.g. cycling, swimming, gardening, etc) or even Participants could simply not tell if and how their sleep participating in sports competitions although sometimes afected their pain or how their pain influenced the qual- they have had to resort to medication to ease the pain ity of their sleep. As such, experiencing pain did not afterwards as one participant notes, “Yesterday I took part stop them from having a good night’s sleep, neither did in an athletics competition (discus, javelin and hammer having good quality sleep improve their perception of throw) and needed painkillers straight after” (Female, 67). pain. 5.2.2. Weather, mattress, and sleep position
It was clear from the responses that other factors beside application along with some preliminary evidence speakLBP afected participants’ sleep. We noted that finding ing in favour of people being able to use it for providing a “...good position to sleep” (Male, 40), a good support data about sleep, LBP and their relationship. from the bed or mattress, and the need to “...turn during The exploratory nature of our study is aimed at shedsleep” (Male, 38) were some of the concerns of a lot of ding more light on the LBP and sleep relationship coparticipants. Participants also noted that the weather nundrum. Sleep Better with Back Pain helps capture afected their sleep. This was particularly noticeable important self-reports on sleep and LBP that provides reduring the summer months where participants expressed searchers with data and insights on this important health a lack of or poor sleep due to hot weather. To some, issue. The long term plan is to build a community powthe weather rather than their LBP afected their sleep, ered application for improving the lifestyles of people “...my sleep last night was afected by the humid weather living with LBP through timely reminders, recommendabut not by the back pain” (Female, 56). Another drew tions, and coaching to motivate people to make healthier a connection between the weather and their felt pain, life choices that could improve both their low back pain outlining how the heat could possibly be exacerbating and sleep quality. the intensity of the felt pain, “...it’s mainly the heat that is making other things harder to deal with” (Female, 67).
Thus, to efectively manage one’s pain, environmental 6.1. Heterogeneity of Poor Sleep Factors factors must also be considered.
Overall, our results indicate the existence of a negligible relationship between felt pain and how well-rested participants felt in the morning. Indeed, this result is interesting as it could be an indication that people living with LBP have become so accustomed to their pain such that it does not matter anymore. It is also an in- source treatments for LBP and people sufering from LBP dication that the precise mechanisms that underpin the can suggest and view common LBP treatments used by sleep and pain relationship remain unclear but are likely community members. We envision the app to become an to involve multiple contributors such as depression [20], enabler of various studies with real users. physical discomfort, neurotropic factors [29], as well as psychological factors [19] among others. Despite the pain felt prior to or during sleep, participants on average 7. Conclusion sleep well. On the contrary, we notice the emergence Our results show that the well-restedness of people livof factors such as “sleep distruption”, “weather”, “active ing with LBP is influenced not only by pain at night but leifefecsttoynle”h,o“wmewdeiclla-rtieosnte”,dapnedo“pslleeefepltpionsitthioenm”oarllnhinagv.inOgnean other factors such as sleep time, wake up time, and sleep disruptions. We also indicate that people are willing to such factor worth expounding on is “sleep distruption”.
Sleep disruption afected significantly how well-rested donate their data toward sleep and LBP related research participants felt in the morning despite the intensity of purposes. This study introduces a preliminary investitheir pain during the previous night. Thus, while partic- gation into sleep tracking and low back pain using the ipants’ perceived sleep quality was not afected by the “Sleep Better with Back Pain” app and communicates our intensity of their pain at night, it was afected by the initial results to the academic community in hopes to disruption of their sleep, disruptions which they identify spark interest and discussion within the community. to be caused non-exhaustively by e.g. sleep movements, bed or mattress, and environmental factors such as the Acknowledgments weather.
However, the connection of some of these factors to This research is connected to the GenZ strategic profilpain is evident, e.g. sleep movements or uncomfortable ing project at the University of Oulu, supported by the sleep positions lead to sleep disruptions because partici- Academy of Finland (project number 318930), and CRITpants’ express feeling pain while moving in bed or laying ICAL (Academy of Finland Strategic Research, 335729). in a particular position. Although participants often refer Part of the work was also carried out with the support of to a factor such as sleep disruption as a major cause of Biocenter Oulu, spearhead project ICON. their poor quality sleep, one possible underlying reason for this disruption is their pain. The existence of and possibly, interconnection of these heterogeneous factors References in assessing the sleep and pain relationship calls for a more holistic approach to investigating this issue.
did not provide any form of reports, analytics, etc. nor incentives for participation to participants of the study, a possible reason for the low number of participants. Our focus in this study is to ascertain the feasibility of the core features of the app, and our results confirm that. While this afects the generalizability of our results, we argue that people have clear interest in LBP and sleep and are willing to contribute their data towards research in this sphere even when all they do is donate their data without getting any insights in return.
Back Pain” app, with new features and user-interface (UI) redesign. One such feature is the integration of daily, weekly, and monthly analytics for all survey items (daily, sleep, and quality of life surveys). We also foresee other features like an in-app community where we could crowd[6] J. Vinstrup, M. D. Jakobsen, L. L. Andersen, Poor and sleep disorders in adults living with chronic sleep is a risk factor for low-back pain among pain: a meta-analysis, Sleep medicine 52 (2018) healthcare workers: Prospective cohort study, In- 198–210. ternational Journal of Environmental Research and [20] K. Harman, R. Pivik, J. L. D’Eon, K. G. Wilson, Public Health 17 (2020) 996. J. Swenson, L. Matsunaga, Sleep in depressed and [7] J. M. Siegel, Clues to the functions of mammalian nondepressed participants with chronic low back sleep, Nature 437 (2005) 1264–1271. pain: electroencephalographic and behaviour find[8] J. McBeth, R. Wilkie, J. Bedson, C. Chew-Graham, ings, Sleep 25 (2002) 47–55.
R. J. Lacey, Sleep disturbance and chronic [21] M. T. Smith, J. A. Haythornthwaite, How do sleep widespread pain, Current rheumatology reports disturbance and chronic pain inter-relate? insights 17 (2015) 1. from the longitudinal and cognitive-behavioral clin[9] L. Shi, S.-J. Chen, M.-Y. Ma, Y.-P. Bao, Y. Han, Y.-M. ical trials literature, Sleep medicine reviews 8 (2004) Wang, J. Shi, M. V. Vitiello, L. Lu, Sleep disturbances 119–132. increase the risk of dementia: a systematic review [22] T. A. Roehrs, Does efective management of sleep and meta-analysis, Sleep medicine reviews 40 (2018) disorders improve pain symptoms?, Drugs 69 (2009) 4–16. 5–11. [10] P. H. Finan, B. R. Goodin, M. T. Smith, The asso- [23] R. M. Benca, S. Ancoli-Israel, H. Moldofsky, Special ciation of sleep and pain: an update and a path considerations in insomnia diagnosis and manageforward, The Journal of Pain 14 (2013) 1539–1552. ment: depressed, elderly, and chronic pain popu[11] M. L. Andersen, P. Araujo, C. Frange, S. Tufik, Sleep lations., Journal of Clinical psychiatry 65 (2004) disturbance and pain: a tale of two common prob- 26–35.
lems, Chest 154 (2018) 1249–1259. [24] T. Roth, Challenges in the comorbid condition, [12] R. Przkora, M. S. Wallace, L. Doan, A. D. Wasan, Sleep medicine 8 (2007) S1–S2.
M. A. Ashburn, J. Mao, A major step to improve [25] T. Roehrs, T. Roth, Sleep and pain: interaction pain research infrastructure: The nih early phase of two vital functions, in: Seminars in neurology, pain investigation clinical network (eppic-net), ASA volume 25, Copyright© 2005 by Thieme Medical Monitor 84 (2020) 38–39. Publishers, Inc., 333 Seventh Avenue, New . . . , 2005, [13] T. Vos, C. Allen, M. Arora, R. M. Barber, Z. A. Bhutta, pp. 106–116.
A. Brown, A. Carter, D. C. Casey, F. J. Charlson, [26] D. J. Buysse, C. F. Reynolds III, T. H. Monk, S. R. A. Z. Chen, et al., Global, regional, and national in- Berman, D. J. Kupfer, The pittsburgh sleep quality cidence, prevalence, and years lived with disability index: a new instrument for psychiatric practice for 310 diseases and injuries, 1990–2015: a system- and research, Psychiatry research 28 (1989) 193– atic analysis for the global burden of disease study 213.
2015, The lancet 388 (2016) 1545–1602. [27] R. D. Hays, J. B. Bjorner, D. A. Revicki, K. L. [14] D. R. Hillman, L. C. Lack, Public health implica- Spritzer, D. Cella, Development of physical and tions of sleep loss: the community burden, Medical mental health summary scores from the patientJournal of Australia 199 (2013) S7–S10. reported outcomes measurement information sys[15] W. D. Killgore, T. J. Balkin, N. J. Wesensten, Im- tem (promis) global items, Quality of Life Research paired decision making following 49h of sleep de- 18 (2009) 873–880.
privation, Journal of sleep research 15 (2006) 7–13. [28] V. Braun, V. Clarke, Using thematic analysis in [16] J. Yin, X. Jin, Z. Shan, S. Li, H. Huang, P. Li, X. Peng, psychology, Qualitative research in psychology 3 Z. Peng, K. Yu, W. Bao, et al., Relationship of sleep (2006) 77–101. duration with all-cause mortality and cardiovascu- [29] P. A. Boakye, C. Olechowski, S. Rashiq, M. J. Verrier, lar events: a systematic review and dose-response B. Kerr, M. Witmans, G. Baker, A. Joyce, B. D. Dick, meta-analysis of prospective cohort studies, Journal A critical review of neurobiological factors involved of the American Heart Association 6 (2017) e005947. in the interactions between chronic pain, depres[17] E. F. Afolalu, F. Ramlee, N. K. Tang, Efects of sleep sion, and sleep disruption, The Clinical Journal of changes on pain-related health outcomes in the Pain 32 (2016) 327–336. general population: a systematic review of longitudinal studies with exploratory meta-analysis, Sleep medicine reviews 39 (2018) 82–97. [18] D. Hillman, S. Mitchell, J. Streatfeild, C. Burns,
D. Bruck, L. Pezzullo, The economic cost of inadequate sleep, Sleep 41 (2018) zsy083. [19] J. Mathias, M. Cant, A. Burke, Sleep disturbances