Teut 2016.

Study characteristics
Methods	Randomized controlled parallel‐group trial.
Participants	176 adults aged ≥ 65 years with chronic LBP. Settings: based at a university and the yoga classes were held either at the university or retirement homes. Country: Germany. Recruitment: distribution of brochures and handouts, holding information events in retirement homes, and advertising in newspapers. Planned half of participants would live on their own and half would live in retirement homes. Participants in retirement homes were recruited directly at the retirement homes to ensure that they could participate. Inclusion criteria: adults ≥ 65 years, chronic LBP for ≥ 6 months; intensity of back pain according to pain item of the Functional Rating Index ≥ 2 over past 7 days; providing written informed consent. Exclusion criteria: acute disk prolapse or protrusion with acute neurologic symptoms within past 3 months; severe organic or psychiatric disease precluding participation in trial; pain due to cancerous effects on bones; use of pain medication that works through the central nervous system (e.g. opioids); drug or alcohol addiction (or both); participation in another clinical trial within past 6 months; participation in yoga or qigong training within past 12 months; preplanned start of a physiotherapy within study duration.
Interventions	Yoga group: (n = 61) 24 × 45‐min viniyoga yoga classes over 3 months. Home practice: none. Qigong group: (n = 58) 12 × 90‐min qigong classes provided over 3 months. Home practice: none. Waiting list group: (n = 57) no specific intervention and were allowed to participate in either yoga or qigong at end of follow‐up (6 months). Common interventions: all groups continued to receive their usual medical care and medications. Co‐interventions: all groups were allowed concomitant health care except for physiotherapy and central nervous system pain medication. Duration and follow‐up: interventions provided for 3 months and follow‐up ended at 6 months.
Outcomes	Back‐specific function (Hannover Functional Ability Questionnaire for measuring back pain‐related disability) at 3 and 6 months. Back pain (mean pain for the previous 7 days on a 100 mm VAS) at 3 and 6 months. Health‐related quality of life – physical (SF‐36 Physical Health component) at 3 and 6 months. Health‐related quality of life – mental (SF‐36 Mental Health component) at 3 and 6 months. Depression (Geriatric Depression Scale) at 3 and 6 months. Other outcomes collected: mean pain intensity over past 7 days on the Functional Rating Index, pain medication, frequency of falls, risk of falls measured using the Tinetti Test, body self‐efficacy, the handgrip strength test, credibility and treatment satisfaction.
Notes	Adverse events: reported to be assessed in the methods section but the occurrence of adverse effects was not mentioned in the results or discussion sections. Measurement of expectations or treatment preferences at baseline: expectations of yoga and qigong were assessed at baseline. Unpublished data: none. Funding: non‐profit. Quote: "This study was performed as part of the grant for the professorship for complementary medicine funded by the Karl and Veronica Carstens‐Foundation. The Yoga intervention was funded by Berliner Yoga Zentrum." Clinical trial registration: NCT01303588.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization was carried out using the random function of SAS software (version 9.1; SAS Inc, Cary, NC). The allocation ratio of yoga:qigong:control group was 1:1:1. The randomization was stratified according to the participant housing situation (living in a retirement home or living alone) and blocked with a fixed block length, which was unknown to the study staff."
Allocation concealment (selection bias)	Low risk	Quote: "The randomization list was transferred to a generated Microsoft Access 2003 (Microsoft Corp, Redmond, CA) secured database, where it was hidden in the background and was not accessible to anyone involved in randomization or treatment to ensure allocation concealment." Quote: "Patients who fulfilled the eligibility criteria and provided informed consent were registered in the preface of the database and randomized by clicking on a button. The group allocation was then revealed to the participants. The group allocation could not be changed or deleted, which ensured allocation concealment."
Blinding of participants	High risk	No blinding.
Blinding of personnel/providers	High risk	No blinding.
Blinding of outcome assessors	High risk	Participants not blinded and outcomes based on self‐report.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: proportion missing from follow‐up was not high, did not differ substantially between groups, and reasons for loss to follow‐up were given.
Selective reporting (reporting bias)	Low risk	Comment: all primary and secondary outcomes corresponded to those in the trial registration.
Group similarity at baseline	Low risk	Quote: "Table 1 [of publication] shows the sociodemographic and clinical characteristics of the patients at baseline. The mean age was similar in all groups. Furthermore, most of the participants were women. The mean duration of the back pain was between 18 and 20 years for all groups, and >60% of the patients used pain medication. In the yoga class, there was a lower percentage of patients with >10 years of education (24.6% compared with 43.1% and 43.9%). In the qigong group, there was a lower percentage of single households (59.9% compared with 78.0% and 71.4%). In the control group, the period elapsed since the last physical therapy session was shorter than the period in either of the intervention groups (1.1 ± 2.4 years for the control group compared with 2.2 ± 3.7 years for the yoga group and 2.7 ± 4.8 years for the qigong group). All other baseline data measurements and clinical outcomes were nearly equally distributed among the groups." Comment: almost all demographic characteristics matched. A sensitivity analysis for period since last physical therapy session did not change the results.
Co‐interventions	Low risk	Quote: "Participants in both groups were allowed to use concomitant health care, but the use of physiotherapy and the intake of pain medication that works over the central nervous system (e.g., opioids) were not permitted." Comment: mean medication use at baseline was similar between groups, and restriction of physiotherapy and opioids during the trial is mentioned.
Compliance	Low risk	Quote: "Of all patients in the yoga group, 74.1% participated in more than 75% of the yoga lessons, another 12.9% in 50 to 75% of the lessons. From 58 patients allocated to the qigong group 3 refused further participation (medical reasons: n = 1, lack of time: n = 1, unknown reasons: n = 1). Of the patients in the qigong group, 72.7% participated in more than 75% of the classes, and another 18.2% participated in 50 to 75% of the classes."
ITT analysis	Low risk	Quote: "Outcomes were analyzed for the full analysis set, on the basis of the intention‐to‐treat principle. The full analysis set includes each randomized patient regardless of whether he or she adhered to the assigned treatment, complied with the protocol, or provided a complete set of all data. If any outcome data were missing, these values were not imputed. Thus, for each outcome the respective available cases were used in the analysis."
Timing of outcome assessments	Low risk	Timing of outcome assessment appeared to be the same for all groups.
Other bias	Low risk	No other biases identified.