Recommendations for Determining the Validity of Consumer Wearables and Smartphones for the Estimation of Energy Expenditure : Expert Statement and Checklist of the INTERLIVE Network

This is a sel -a chi ed e sion o an o iginal a icle. This e sion may di e om he o iginal in pagina ion and ypog aphic de ails. Au ho (s): Ti le: Yea : Ve sion: Copy igh : Righ s: Righ s u l: Please ci e he o iginal e sion: CC BY 4.0 h ps://c ea i ecommons.o g/licenses/by/4.0/ Recommenda ions o De e mining he Validi y o Consume Wea ables and Sma phones o he Es ima ion o Ene gy Expendi u e : Expe S a emen and Checklis o he INTERLIVE Ne wo k © Au ho s, 2022 Published e sion A gen , Rob; He he ing on-Rau h, Megan; S ang, Julie; Ta p, Jakob; O ega, F ancisco B.; Molina-Ga cia, Pablo; Schumann, Mo i z; Bloch, Wilhelm; Cheng, Sulin; G øn ed, Ande s; B ønd, Jan Ch is ian; Ekelund, Ul ; Sa dinha, Luis B.; Caul ield, B ian A gen , R., He he ing on-Rau h, M., S ang, J., Ta p, J., O ega, F. B., Molina-Ga cia, P., Schumann, M., Bloch, W., Cheng, S., G øn ed, A., B ønd, J. C., Ekelund, U., Sa dinha, L. B., & Caul ield, B. (2022). Recommenda ions o De e mining he Validi y o Consume Wea ables and Sma phones o he Es ima ion o Ene gy Expendi u e : Expe S a emen and Checklis o he INTERLIVE Ne wo k. Spo s Medicine, 52(8), 1817-1832. h ps://doi.o g/10.1007/s40279-022- 01665-4 2022 Vol.:(0123456789) Spo s Medicine h ps://doi.o g/10.1007/s40279-022-01665-4 SYSTEMATIC REVIEW Recommenda ions o De e mining heValidi y o Consume Wea ables andSma phones o  heEs ima ion o Ene gy Expendi u e: Expe S a emen andChecklis o  heINTERLIVE Ne wo k RobA gen 1,2,3 · MeganHe he ing on‑Rau h4· JulieS ang5· JakobTa p5· F anciscoB.O ega6,7· PabloMolina‑Ga cia6· Mo i zSchumann8,9· WilhelmBloch8· SulinCheng8,9,10· Ande sG øn ed11· JanCh is ianB ønd11· Ul Ekelund5· LuisB.Sa dinha4· B ianCaul ield1,2 Accep ed: 15 Feb ua y 2022 © The Au ho (s) 2022 Abs ac Backg ound Consume wea ables and sma phone de ices commonly o e an es ima e o ene gy expendi u e (EE) o assis in he objec i e moni o ing o physical ac i i y o he gene al popula ion. Alongside consume s, heal hca e p o essionals and esea che s a e seeking o u ilise hese de ices o he moni o ing o aining and imp o ing human heal h. Howe e , he me hods o alida ion and epo ing o EE es ima ion in hese de ices lacks igou , nega i ely impac ing on he abili y o make compa isons be ween de ices and p o ide anspa en accu acy. Objec i es The Towa ds In elligen Heal h and Well-Being Ne wo k o Physical Ac i i y Assessmen (INTERLIVE) is a join Eu opean ini ia i e o six uni e si ies and one indus ial pa ne . The ne wo k was ounded in 2019 and s i es owa ds de eloping bes -p ac ice ecommenda ions o e alua ing he alidi y o consume wea ables and sma phones.This expe s a emen p esen s a bes -p ac ice alida ion p o ocol o consume wea ables and sma phones in he es ima ion o EE. Me hods The ecommenda ions we e de eloped h ough (1) a sys ema ic li e a u e e iew; (2) an uns uc u ed e iew o he wide li e a u e discussing he po en ial ac o s ha may in oduce bias du ing alida ion s udies; and (3) e idence-in o med expe opinions om membe s o he INTERLIVE ne wo k. Resul s The sys ema ic li e a u e e iew p ocess iden i ied 1645 po en ial a icles, o which 62 we e deemed eligible o he inal da ase . Based on hese s udies and he wide li e a u e sea ch, a alida ion amewo k is p oposed encompassing six key domains o alida ion: he a ge popula ion, c i e ion measu e, index measu e, es ing condi ions, da a p ocessing and he s a is ical analysis. Conclusions The INTERLIVE ne wo k ecommends ha he p oposed p o ocol, and checklis s p o ided, a e used o s and- a dise he es ing and epo ing o he alida ion o any consume wea able o sma phone de ice o es ima e EE. This in u n will maximise he po en ial u ili y o hese echnologies o clinicians, esea che s, consume s, and manu ac u e s/ de elope s, while ensu ing anspa ency, compa abili y, and eplicabili y in alida ion. T ial Regis a ion PROSPERO ID: CRD42021223508. * Rob A gen ob.a [email p o ec ed] Ex ended au ho in o ma ion a ailable on he las page o he a icle 1 In oduc ion Wo ldwide, he numbe o consume wea able de ices sold in 2020 is es ima ed o ha e eached almos 400 million uni s [1]. The g owing popula i y o hese de ices, com- bined wi h he numbe o sma phone use s oday su passing 3 billion [2], means ha mo e people han e e a e able o sel -moni o hei physical heal h and ac i i y. Addi ion- ally, consume wea ables a e being inc easingly u ilised in heal hca e and esea ch se ings [3]. These de ices can assess a a ie y o me ics associa ed wi h physical ac i i y including s ep coun , hea a e, maximal oxygen consump- ion (VO2max), and ene gy expendi u e (EE). T adi ionally, he measu emen o EE has been conduc ed ia indi ec calo ime y (whe e gaseous exchange is meas- u ed and con e ed o EE using s anda d o mulae such as he Wei Equa ion [4]), di ec calo ime y ( he measu emen o hea p oduc ion by he body), o doubly labelled wa e R.A gen e al. Key Poin s This sys ema ic li e a u e e iew o alida ion s udies o consume wea ables and sma phone applica ions has highligh ed a he e ogenei y be ween alida ion me hod- ologies in key domains, pa icula ly in he a ge popula- ions, da a p ocessing and s a is ical app oaches, gi ing ise o alida ion bias. The lack o ee-li ing alida ion leads o limi ed abili- ies o use s o unde s and he accu acy o wea able de ices and sma phones in es ima ing ene gy expendi- u e (EE) in he in ended use case. In his a icle, he INTERLIVE ne wo k p o ides bes - p ac ice ecommenda ions o be used in u u e p o o- cols o mo e owa ds a mo e accu a e, anspa en and compa able alida ion o EE es ima ion de i ed om consume wea ables and sma phone applica ions. di icul o consume s and clinicians o na iga e he ma ke and unde s and o compa e he accu acy o such de ices [13]. Whe he a consume moni o can be conside ed accu- a e o es ima ing EE is de e mined by a combina ion o he a ge popula ion, he in ended ou pu (daily TEE, daily AEE, aining session EE, aining session AEE, e c.) and he ans e abili y o he alida ion p o ocol o he in ended use. Fo example, a consume moni o wi h epo ed high accu acy du ing an inc emen al walking o unning p o ocol may no be he igh choice o an indi idual who expends ene gy abo e es , p ima ily du ing less o mal ac i i ies. As consume moni o s will be used o a ange o pu poses, a comp ehensi e e alua ion du ing disc e e ac i i ies and ee-li ing se ings is needed o maximise applicabili y. The e ha e been calls o s anda disa ion o he e alua- ion o consume wea able de ices, wi h eques s o ma - ke ing claims om manu ac u e s o be e idence-based and independen ly e i ied in a anspa en manne wi h app o- p ia e analy ical app oaches [14, 15]. While amewo ks ha e been p oposed o e alua e wea ables ac oss he mos commonly measu ed me ics, such as s ep coun , hea a e, EE and VO2 max [13, 16], he majo i y o hese ela e o esea ch-g ade de ices. To add ess he challenges discussed, he Towa ds In elligen Heal h and Well-Being Ne wo k o Physical Ac i i y Assessmen (INTERLIVE) ne was es ab- lished o in eg a e and build upon p e ious wo k o p opose bes -p ac ice p o ocols o he alida ion o exe cise/ac i - i y me ic measu emen capabili ies in consume wea ables and sma phones. In his case, INTERLIVE is add essing he measu emen o EE. The ne wo k i s conduc ed a sys- ema ic li e a u e e iew o iden i y he me hods p e iously used in he alida ion o EE es ima ion in consume wea a- bles and sma phones. Following his, o he ele an li e a- u e discussing ac o s ha may in oduce bias in alida ion s udies was hen consul ed, alongside e idence-in o med ne wo k discussion o de elop bes -p ac ice alida ion ec- ommenda ions. This pape p esen s a comp ehensi e epo on he a iables o conside when designing and conduc - ing alida ion p o ocols o he es ima ion o EE and makes ecommenda ions o he bes p ac ice me hodologies and epo ing o such s udies. 2 Expe S a emen P ocess 2.1 The INTERLIVE Ne wo k INTERLIVE is a join ini ia i e o he Uni e si y o Lisbon (Po ugal), Ge man Spo Uni e si y (Ge many), Uni e - si y o Sou he n Denma k (Denma k), No wegian School o Spo Sciences (No way), Uni e si y College Dublin (I e- land), Uni e si y o G anada (Spain) and Huawei Technolo- gies, Finland. The ne wo k was ounded in 2019, comp ising (DLW; whe e u ine samples a e analysed using iso ope a io mass spec ome y) [5, 6]. While DLW is conside ed he gold s anda d o measu ing ee-li ing EE [6, 7], i only p o ides a measu e o o al EE (TEE), and he ac i e EE (AEE) is de e mined by sub ac ing an es ima e o he es - ing EE. Addi ionally, DLW is an expensi e and complex app oach, equi ing specialis skills and esou ces. The e- o e, indi ec calo ime y has been he p edominan c i e ion measu e o AEE in ecen yea s [6]. Howe e , hese indi ec calo ime y me hodologies a e limi ed o mo e con olled p o ocols equi ing he use o me abolic ca s, which a e expensi e and equi e echnical skills o adminis e . The widesp ead a ailabili y, ela i ely low cos , and wea abili y o consume ac i i y moni o s means ha da a ha could only be cap u ed in specialis se ings a e now a ailable o he gene al popula ion, and hese de ices a e inc easingly being used o p o ide measu es o EE. Mos use s seek his in o ma ion o ei he AEE in moni o ing calo ie consump ion du ing a speci ic exe cise ac i i y, o o suppo hem wi h hei TEE ac oss a day, possibly in he managemen o die o weigh loss. A a ie y o inpu s, including use demog aphics such as age, sex, weigh and heigh , and da a om on-boa d senso s, including accele - ome e s and/o pho ople hysmog aphy hea a e senso s, a e used o make es ima es o EE [8–10]. Howe e , he alidi y and eliabili y o consume wea ables o es ima e g oup- and indi idual-le el EE is subjec o deba e, wi h nume ous alida ion s udies being conduc ed, o e ing a y- ing me hodologies and esul s [11, 12]. The inconsis ency in alida ion p o ocols used in he li e a u e, along wi h a lack o anspa ency in epo ing s anda ds means ha i is De e mining Validi y o Wea ables o Es ima ion o Ene gy Expendi u e o he au ho s o his pape , plus addi ional expe s wi hin each ins i u ion, and s i es owa ds de eloping bes -p ac ice p o ocols o e alua ing he alidi y o consume wea a- bles wi h ega d o measu emen o exe cise/ac i i y me - ics. Mo eo e , we a e aiming o inc ease awa eness o he ad an ages and limi a ions o di e en alida ion me hods and o in oduce no el heal h- ela ed me ics, os e ing a widesp ead use o physical ac i i y indica o s. To da e, bes - p ac ice alida ion p o ocols o consume wea able hea a e moni o ing [17] and s ep-coun ing [18] ha e been p o- posed by he ne wo k. In his pape , we de ail INTERLIVE’s wo k in espec o he measu emen o EE. 2.2 Expe Valida ion P o ocol De elopmen 2.2.1 Expe Valida ion P ocess The ne wo k used he same p ocess o de elop he bes - p ac ice alida ion p o ocols o he consume wea able de ice-de i ed me ics as was used o he p e ious HR moni o ing and s ep-coun ing pa ame e s [17, 18]. The i s s ep o de eloping he expe ecommenda ions consis ed o a sys ema ic e iew o he alida ion p o ocols used in he scien i ic li e a u e. This in o ma ion was hen used as he ounda ion o discussions and o p o ide ecommenda ions on he op imal and mos easible p o ocol o assessing he alidi y o consume wea ables. Wo king g oup mee ings we e held o discuss he aspec s o he alida ion p o ocols used in he s udies iden i ied in he sys ema ic sea ch. A se o key domains o bes -p ac ice ecommenda ions we e p oposed based on he ou comes o he sys ema ic li e a u e e iew, he a p io i knowledge ela ing o esea ch g ade de ice alida ion [13, 16, 19, 20], and he e idence-in o med expe opinion o he INTERLIVE membe s. The syn he- sised da a we e hen e iewed wi h espec o hese domains, and expe alida ion p o ocols o he consume wea able- de i ed EE me ic was hen de eloped by he wo king g oup. 2.2.2 Sys ema ic Re iew P ocess The p ima y aim o he sys ema ic li e a u e e iew was o de e mine which me hods and p o ocols a e cu en ly being used in he scien i ic li e a u e o alida e he consume wea able o sma phone-de i ed me ic o EE. I was beyond he scope o his e iew o explo e he esul s o he included s udies, al hough a ecen sys ema ic e iew and me a-anal- ysis was published, wi h accu acy a ying dependen on ac i i y ype [11]. The sea ch was conduc ed wi h espec o he P e e ed Repo ing I ems o Sys ema ic Re iews and Me a-Analyses (PRISMA) s a emen and egis e ed wi h he in e na ional da abase o p ospec i ely egis e ed sys ema ic e iews in heal h and social ca e (PROSPERO ID: CRD42021223508). To iden i y pee - e iewed jou nal a icles, we sea ched he PubMed, Web o Science, and Sco- pus elec onic da abases using speci ic sea ch e ms, which a e lis ed in elec onic supplemen a y Table1. To be included o e iew, he s udies had o (1) be pub- lished p io o 4 Decembe 2020; (2) examine he alidi y o consume -based wea able and sma phone applica ions o es ima ing EE compa ed wi h he gold-s anda d c i e ion measu e o ei he di ec o indi ec calo ime y, me abolic chambe , o DLW; and (3) be w i en in English. S udies we e excluded i he index measu e was no de i ed om da a acqui ed using a consume wea able de ice o sma - phone, did no con ain a alida ion p o ocol, he e was no gold-s anda d c i e ion measu e, he ou come assessed by he de ice was no EE, o i he ull ex was no a ailable. Fo ou pu poses, wea ables we e es ic ed o only hose wi h a consume pu pose a he han a esea ch o clinical pu pose, meaning ha he de ices a e p ima ily ma ke ed o he consume and a e eadily a ailable o pu chase and use by membe s o he public. No exclusion c i e ia we e se o he s udy popula ion (i.e., heal hy, clinical pa ien s, child en). The sys ema ic li e a u e e iew p ocess was con- duc ed by RA, MHR, JS, and JT. Ti le/abs ac and ull- ex sc eening we e comple ed by wo independen e iewe s and con i med by a hi d independen e iewe using Co idence so wa e (Ve i as Heal h Inno a ion) [21]. A cus om-designed sp eadshee was used o da a ex ac- ion and included he ollowing headings: sample size included/analysed, sex dis ibu ion, ype o popula ion, body mass index (BMI), heigh , weigh , age, condi ion ype (labo a o y; semi- ee-li ing; ee-li ing), c i e ion measu e (desc ip ion, con igu a ion and placemen ), index measu e (desc ip ion, con igu a ion and placemen ), es ing p o ocol, signal p ocessing, da a synch onisa ion, s a is ical analysis, esul s and conclusions. Valida ion p o ocols we e di ided in o h ee ca ego ies based on he ollowing de ini ions: Labo a o y: Well-con olled condi ions, p edominan ly asks such as walking o unning on a eadmill/indoo /ou - doo en i onmen , o cycling on a s a iona y cycle e gome e a p e-de ined o sel -selec ed speeds. Semi ee-li ing: Semi-con olled condi ions, including ‘simula ed’ ac i i ies o daily li ing o he pu pose o epli- ca ing ‘ ee-li ing’ condi ions (e.g. sweeping, cooking, old- ing laund y, compu e use, spo -speci ic ac i i y). F ee-li ing: Uncon olled me hodologies ha in ol e pa - icipan s wea ing he index de ice du ing ‘no mal’ daily li e, ou side o a con olled labo a o y o simula ed en i onmen . 3 Cu en S a e o Knowledge The sea ch s a egy o his sys ema ic e iew yielded 1645 po en ial a icles o inclusion. Following he emo al o 371 duplica es, 1275 i les and abs ac s we e sc eened, esul ing R.A gen e al. in 254 a icles ha me he c i e ia o ull- ex sc eening. O e all, 192 s udies we e excluded, esul ing in a inal o al o 62 s udies accep ed o da a ex ac ion. O he 192 exclu- sions, 174 we e excluded as hey did no e alua e consume wea ables and ins ead assessed clinical and esea ch-g ade de ices. The PRISMA lowcha o he s ages o he sys- ema ic sea ch and sc eening p ocess, including easons o exclusion, is illus a ed in Fig.1. The ull ex ac ion da a p esen ing he me hodologies employed ac oss he included s udies can be ound in elec- onic supplemen a y Table2 (labo a o y p o ocols), elec- onic supplemen a y Table3 (semi- ee-li ing p o ocols), and elec onic supplemen a y Table4 ( ee-li ing p o ocols). Eigh y- i e di e en models om 35 manu ac u e s we e epo ed in a icles included in his sys ema ic e iew, he majo i y being wea able manu ac u e s (n = 30), wi h only i e sma phone applica ions e alua ed. The mos s udied manu ac u e was Fi bi , wi h a leas one model e alua ed in 33 a icles, ollowed by he Apple Wa ch (n = 14), Ga min (n = 12), Jawbone (n = 11) and Pola (n = 10). Simila o he p e ious s a emen s o he INTERLIVE ne wo k, he da a a e p esen ed in six key alida ion p o ocol and epo ing domains comp ising o he a ge popula ion, c i e ion meas- u e, index measu e, es ing condi ions, da a p ocessing and s a is ical analysis (Fig.2). Wi hin each domain, aspec s deemed c i ical o e ec i e alida ion and epo ing we e iden i ied and a e ou lined wi h espec o he in o ma ion colla ed du ing he s a emen p ocess. 3.1 Ta ge Popula ion In he measu emen o TEE, a numbe o a iables con- ibu e o indi idual a ia ion in de e mining bo h es ing me abolic a e (RMR) and AEE, including age, sex, body size and composi ion, e hnici y, gene ics and i ness le el [6, 13]. Consume wea ables a e ma ke ed o a b oad global demog aphic, om adolescen s o olde adul s as well as om a hle es o hose wi h seden a y li es. Thus, unless in es iga ing a speci ic clinical applica ion, alida ion s ud- ies o EE es ima ion should e lec he he e ogenous na u e o he a ge popula ion in he sample, and no be con ined o a niche sec o . Twen y-se en (44%) o he s udies iden i- ied in he sys ema ic li e a u e e iew ec ui ed hei sample om a uni e si y communi y. While his may be con eni- en , we ecommend ha i he aim o he s udy is alida- ion o he wea able in he gene al popula ion, hen a mo e he e ogenous sample is equi ed wi h a wide a ia ion in EE es ima es. No ably, 13 o he 62 s udies (21%) ailed o p o ide any epo o whe e he pa icipan s we e sampled om, and 34 s udies (55%) did no epo how pa icipan s we e ec ui ed. This in o ma ion should be a p e equisi e in he epo ing o any alida ion s udy. RMR is well documen ed o decline wi h age [22–24], and while age is ac o ed in o many EE es ima ion equa- ions [13], pa icipan s in alida ion s udies should ep esen a spec um o ages, wi h due conside a ion o s a i ica- ion acco ding o age anges. The cu en li e a u e ends o alida e wea ables in he younge adul popula ion, wi h 35 s udies epo ing a sample wi h an a e age age in he 20s. Fig. 1 PRISMA lowcha o he sys ema ic e iew p ocess. PRISMA P e e ed Repo ing I ems o Sys ema ic Re iews and Me a-Analyses De e mining Validi y o Wea ables o Es ima ion o Ene gy Expendi u e Olde adul s (> 60yea s) a e pa icula ly poo ly ep esen ed. Alongside a ep esen a i e age g oup, i is impo an ha alida ion p o ocols include an equal sample ac oss bo h sexes, gi en he a iance in me abolic a e [6, 25], and epo on body size and composi ion. Finally, skin one has p e i- ously been no ed as ha ing an e ec on he accu acy o hea a e measu es [17]. Gi en ha many consume wea ables use hea a e in hei EE es ima ion algo i hm, i he de ice includes pho ople hysmog aphy, skin one should also be epo ed using he Fi zpa ick scale [26]. The a e age sample size ac oss all he s udies in he sys- ema ic e iew was 34 pa icipan s ( ange 12–100), wi h only nine a icles (15%) conduc ing a sample size calcula- ion and jus i ica ion. Despi e he lack o sample size calcu- la ions, many s udies conduc ed equi alence o di e ence es s, which a e dependen on sample size. The e o e, i he objec i e o he s udy was o de e mine a minimum accep - able le el o accu acy, a sample size calcula ion should be conduc ed based on pilo es ing o p e iously published mean and s anda d de ia ion o he di e ences be ween he index and c i e ion measu es [27]. I his is no he p ima y ocus, o homogenous samples we ecommend a minimum o 45 pa icipan s [28], al hough conside a ion should be gi en o he pa icipan cha ac e is ics, and a la ge sample size will likely be equi ed o mo e he e ogenous g oups. 3.2 C i e ion Measu e I is widely ega ded ha he gold-s anda d app oach o measu ing TEE is by using DLW [6, 7, 29, 30]. This is assessed o e a pe iod o se e al days in a ee-li ing con ex , and a e age daily TEE can be calcula ed o e a speci ied ime pe iod ia he collec ion o u ine samples. In o de o de e mine AEE, RMR is calcula ed ia indi ec calo ime y o p edic ion equa ions and hen sub ac ed om he TEE measu ed wi h DLW [31]. Howe e , i is no pos- sible o measu e ac i i y ype, in ensi y o du a ion wi hin he TEE o each ime pe iod [32], and his in o ma ion may be ele an o many use s and clinicians. Equally, he use o DLW is expensi e and equi es in-dep h specialis analy ical acili ies, which places a es ic ion on he easibili y o i s deploymen in many s udies [30, 32]. The e o e, while he use o DLW is ecommended in ee-li ing p o ocols, o he measu es such as calo ime y a e needed o he calcula ion o AEE in speci ic ac i i ies. Calo ime y in ol es he measu emen o ei he he mal change (di ec calo ime y) o he measu emen o gaseous exchange (indi ec calo ime y), and can ake he o m o whole- oom me abolic chambe s, en ila ed hoods, o a acemask/mou hpiece [33–35]. Indi ec calo ime y is mos commonly used o measu e EE in he labo a o y, as i is highly easible o measu e oxygen consump ion and cal- cula e ene gy cos , ye his is no p ac ical in a ee-li ing Fig. 2 Six domains and co esponding a iables o in e es iden i ied as being o impo ance in he alida ion o consume wea able and sma - phone es ima ion o EE. EE ene gy expendi u e R.A gen e al. con ex [6]. The esul s o a ecen sys ema ic e iew and me a-analysis sugges ha he accu acy o esea ch-g ade wea ables o es ima ing EE is no a a sa is ac o y le el o hese de ices o be used in he alida ion o comme - cial wea ables as a c i e ion measu e [11]. The e o e, based on he cu en a ailable e idence, we ecommend ha he c i e ion measu e o he alida ion o speci ic ac i i ies o exe cise p o ocols be indi ec calo ime y, wi h DLW used o ee-li ing s udies. All i e s udies in he sys em- a ic li e a u e e iew alida ing consume wea ables in a ee-li ing con ex u ilised he DLW me hod [10, 36–39], while 3 (10%) o he 28 semi- ee-li ing p o ocols u ilised a me abolic chambe [10, 37, 40]. The emaining s udies wi h a semi- ee-li ing p o ocol (25) and all 52 labo a o y p o ocols selec ed indi ec calo ime y in he o m o a po - able sys em o me abolic ca . The c i e ion de ice should be placed acco ding o manu ac u e ’s ins uc ions and epo ed in he s udy w i e-up, alongside labo a o y-speci ic da a on he quali y o he c i e ion measu e whe e possible (i.e. pe - cen age coe icien o a ia ion [ eliabili y]), and any cali- b a ion de ails. In he case o using DLW in ee-li ing p o- ocols and epo ing on AEE, au ho s should p o ide a ull accoun o assump ions and equa ions used [6], o example whe he RMR is calcula ed ia indi ec calo ime y o by equa ion. We would ecommend ha he bes -p ac ice o measu ing RMR in a ee-li ing p o ocol is wi h indi ec calo ime y, a he han es ima ed using equa ions; howe e , we ecognise his may no always be easible o esea che s, and in hose cases, he exac equa ion used should be s a ed. 3.3 Index Measu e In o de o ensu e he ecological alidi y o a s udy p o ocol, he placemen o he wea able de ice mus be conside ed, especially in s udies e alua ing mul iple de ices a he same ime. All comme cial wea ables should be e alua ed wi h he de ice placed acco ding o he manu ac u e ’s ins uc ions. Ac i i y moni o s a e commonly wo n on he w is , al hough in some cases may be a ached o he wais o he use [41]. Less used placemen s include he o so [42] and he shoes [43]. Fi een s udies iden i ied in he sys ema ic e iew e alua ed mul iple w is -wo n de ices on he same a m in he p o ocol, wi h up o h ee de ices wo n on one a m. I is unclea wha e ec his placemen migh ha e on he ou - pu s o he de ice, and, a guably, while s ill placed on he lowe o ea m, a leas one o he h ee de ices is no being wo n acco ding o he manu ac u e ’s ins uc ions. The e is a nega i e impac o he easibili y o s udies i esea che s a e equi ed o only place one de ice on each w is du ing a es ing p o ocol, he e o e we would ecommend ha whe e possible, esea che s a oid placing mul iple de ices on he same w is , wi h a maximum o wo w is -wo n de ices being es ed on one a m simul aneously. In addi ion, when using wo w is -wo n de ices on he same a m, we ecom- mend ha he e should be a andom coun e balanced place- men o he de ices be ween pa icipan s; howe e , he o de o de ice placemen wi h espec o dis ance om he w is join should be clea ly epo ed. Equally, sma phone applica ions may no ha e an explici manu ac u e /de elope -speci ied placemen . O he ou s udies in he sys ema ic e iew ha assessed sma phone measu ed EE, one placed he sma phone a he lowe back [44], one on he bel nex o a wea able a he mid-axilla y line [45], one in a pocke (al hough he loca ion o he pocke was no speci ied) [46], and he las s udy used wo sma - phones, one in he hand and one in he igh pocke , wi h a compa ison made be ween he wo posi ions [47]. I has been epo ed ha in day- o-day li e, 60% o emales place hei phone in hei bag, while 60% o males ca y hei s in hei pocke [48]. As such, we would conside clo hing pocke s, handbags o ca ying in he hand as he mos sui able posi- ions o e alua e sma phone es ima ion o EE. The phone should no be moun ed o he body in an unna u al way such as o he ches o low back. As well as epo ing he placemen p o ocol in de ail, he exac model and e sion o he index de ice should be clea ly speci ied, o bo h he ha dwa e and so wa e, includ- ing no ing i any i mwa e upda es ook place du ing he da a cap u e pe iod. Mos s udies epo ed he b and and de ice name (e.g. Fi bi Cha ge HR); howe e , wi h egula upda es o ha dwa e and so wa e, i is di icul o ully ep- lica e he s udy wi hou speci ied e sion numbe s. Fu he - mo e, he de ice should be ese back o baseline be ween pa icipan s o ensu e he e is no in luence on de ice EE algo i hms po en ially based on p e ious use. Finally, any demog aphic and an h opome ic de ails inpu ed in o he de ice, and he use o any speci ic exe cise modes, should be epo ed. The use o exe cise modes has been highligh ed as an a ea o u he examina ion, wi h esea che s and use s cu en ly unawa e o he impac o using di e en modes on he measu emen o EE [49, 50]. 3.4 Tes ing Condi ions Consume wea ables a e designed o p o ide a measu e o EE in uncon olled en i onmen s such as daily EE, wi h limi ed use in e ac ion wi h he de ice, and du ing mo e con olled se ings whe e EE is es ima ed du ing a speci ic exe cise/ac i i y. The esul s om he sys ema ic e iew show ha many esea che s a e a ou ing he use o labo a- o y o semi- ee-li ing p o ocols. O he 62 s udies included in he sys ema ic e iew, 51 (82%) included a labo a o y p o ocol, 27 s udies (44%) conduc ed some o m o semi- ee-li ing in es iga ion, and only i e s udies (8%) exam- ined ee-li ing condi ions. De e mining Validi y o Wea ables o Es ima ion o Ene gy Expendi u e Labo a o y p o ocols seek o alida e he wea able in measu ing EE du ing asks ha a e well con olled. The accu acy can be examined ac oss a ying in ensi ies and include a combina ion o p ede ined walking, unning, and cycling ac i i ies. The s udies iden i ied in he sys ema ic li - e a u e sea ch wi h labo a o y p o ocols highligh ed a ying le els o accu acy o a wea able in measu ing EE depend- en on he in ensi y o he ac i i y. Addi ionally, in he case o walking and unning, he accu acy o he index measu e a ied dependen on he le el o incline o he ac i i ies in a numbe o s udies [42, 51–53]. In con as , semi- ee-li ing e alua ion inco po a es simula ed ac i i ies o daily li ing and ac i i ies emula ing spo s o exe cises ha a e mo e speci ic han walking, unning o cycling. Wi hin hese p o- ocols, pa icipan s a e encou aged o comple e he asks as hey would do du ing no mal daily li e, a a sel -selec ed pace. The pu pose o hese ype o alida ion s udies is o cap u e (1) speci ic spo s o complex ac i i ies, and/o (2) a di e se ange o ac i i ies emula ing ac i i ies o daily li - ing, such as elaxa ion, housekeeping, o o he physically ac i e leisu e pu sui s. We a gue ha he accu a e measu e- men o TEE and AEE du ing speci ic spo s o complex ac i i ies is impo an because i allows end use s o de e - mine he alidi y o hei de ice du ing ac i i ies o pa icu- la ele ance o hem. Howe e , because o he cons ain s o labo a o y equipmen and he labo a o y se ing, many ac i i ies/spo s, and pa icula ly ac i i ies o daily li ing, canno be pe o med in a manne ha is compa able o he way hese ac i i ies o exe cise modes a e pe o med on a daily basis. I his is no possible, we a gue ha he in o - ma ion acqui ed om he ypical semi- ee-li ing s udy has e y li le alidi y in de e mining he p ecision o EE om consume de ices du ing ac ual ee-li ing beha iou s o use s. In con as o ou p e ious ecommenda ions on he alida ion o consume wea ables o measu e hea and s ep coun [17, 18], we ecommend ha alida ion o TEE and AEE om wea ables should be pe o med wi hin wo es ing condi ions: (1) du ing speci ic ac i i ies o exe cises/spo s, and (2) in a ee-li ing con ex . In o de o accu a ely assess EE du ing a speci ic exe cise o ac i i y, a dis inc ion be ween s eady s a e and a iable in ensi y ac i i ies mus be made. Du ing s eady-s a e ac i i- ies, s eady s a e mus be achie ed whe e hea a e emains s able du ing a con inuous ask and ca diac ou pu is a a su icien le el o suppo oxygen anspo in o de o mee ene gy cos [54–56], be o e measu emen s a e included in he analysis. The e is some deba e in he li e a u e abou when s eady s a e is achie ed, bu he gene al consensus is ha he e is a 2- o 5-min pe iod whe e VO2 and VCO2 a ies by up o 15% om he EE o he pe o med wo k, depending on he ac i i y and he change in in ensi y g adi- en [55–59]. The e o e, we ecommend alida ion o s eady- s a e ac i i ies be pe o med ac oss di e en in ensi ies o de e mine alidi y a di e en absolu e le els, wi h esea ch- e s p o iding jus i ica ion o he achie emen o s eady s a e. As a gene al ule o humb, bu dependen on he in ensi y g adien and he absolu e in ensi y pe o med, each o he ac i i ies/in ensi ies wi hin an exe cise/ac i i y-speci ic p o ocol is ecommended o las a leas 6min o acili a e s able s eady-s a e measu emen s. We ecommend conduc - ing e alua ion wi h as wide a ange o in ensi ies as easible and a conside a ion o he ele ance o g aded ac i i ies (e.g. an incline ≥ 5% du ing walking o unning asks). Equally, i conduc ing walking/ unning assessmen on a eadmill, we would ecommend all es ing is conduc ed wi h a 1% incline o accoun o he educed me abolic cos o eadmill p opulsion [60]. Due conside a ion should be gi en o he po en ial need o a eco e y pe iod be ween ac i i y bou s when highe -in ensi y exe cises a e examined o allow su - icien eco e y o app oxima ely RMR. The choice o all in ensi ies should ake in o conside a ion he cha ac e is- ics o he sample, wi h calcula ed in ensi ies based on he pa icipan ’s maximal hea a e o VO2max being he mos p e e able o e selec ing absolu e alues, o lesse s ill, sel -selec ed in ensi ies. Howe e , while hese pe sonalised in ensi ies a e ecommended, he e may be cases whe e i is no easible o calcula e pe sonalised i ness le els, and, in hese ins ances, conduc ing and epo ing absolu e exe cise in ensi ies such as speed o wa s is ecommended. Du ing hese ac i i ies, conside a ion o a ailable wo k- ou modes on de ices a e needed as he mode selec ed dic a es wha senso s a e enabled on he de ice in o de o p ese e ba e y, o example indoo ac i i ies disable he GPS and cycling does no equi e s ep de ec ion, and hese modi ica ions a e likely o impac he alidi y o he meas- u emen s. These use -de ined ac i i ies o assessmen o EE end o ela e o he mos common wo kou s, including, bu no limi ed o, unning, walking, cycling and swimming. A emp ing o alida e de ices in o he asks ha a e no de ined by he manu ac u e , such as eading, acuuming o sweeping, is mos app op ia ely assessed in he alidi y o TEE es ima ion in he ee-li ing en i onmen . Addi ionally, a u he issue ha mus be add essed is he need o co - ec o inaccu acies in he measu emen o EE a hea y o maximal in ensi y ac i i y, whe e he anae obic componen will no be cap u ed by indi ec calo ime y. Resea che s conduc ing high-in ensi y ac i i ies should ecognise his, adjus o i using an app op ia e me hod such as inco po- a ing blood lac a e measu emen s [61, 62] o in e pola ion based on exe cise in ensi y [63], and epo hese me hods app op ia ely. To assess EE es ima ion o consume wea ables in he eal-wo ld, ee-li ing p o ocols should cap u e a ange o ac i i ies om pa icipan s in uncon olled condi ions. Gi en he ecommenda ion o he use o DLW in ee-li ing alida ion, he p o ocol pe iod should las 7–14days, and R.A gen e al. pa icipan s should no be cons ained in hei ac i i ies in any way [30, 64]. The epo ing o such p o ocols should include in de ail he ins uc ions gi en o pa icipan s, including whe he hey we e asked o use o did use any speci ic exe cise modes on he wea ables du ing he pe iod. Finally, i is p e e able o eco d and epo on he adhe ence o pa icipan s o wea ing he de ice as ins uc ed o high- ligh any disc epancies be ween de ices due o non-adhe - ence, and epo on any missing da a. In he case o speci ic ac i i y EE es ima ion, i is ecommended ha da a mus be cap u ed wi h he pa icipan wea ing he de ice o he en i e y o he p o ocol, and i he e is non-adhe ence o his o he e a e missing da a, hen ha da a cap u e be disca ded. Fu he mo e, o ee-li ing e alua ions, any me hods used o in e pola e missing da a, such as du ing cha ging o e - nigh , should be epo ed clea ly. 3.5 P ocessing The da a p ocessing and epo ing ha is p o ided in ali- da ion s udies is impo an bu is o en o e looked, wi h nume ous a icles in he sys ema ic e iew lea ing he eade le o make assump ions on he p ocessing me hods o he s udy. Au ho s mus p o ide clea de ail on how s eady-s a e o anae obic EE (i ele an ) was managed in hei p o o- col, and how he da a collec ion, including synch onisa ion, was managed o cap u e he measu es o he c i e ion and index de ice pe ac i i y. Gi en he 2- o 5-min pe iod ha is equi ed o achie e s eady s a e, we would ecommend ha he a e age o he las 3min o he exe cise p o ocols o s eady-s a e exe cises o ac i i ies be used o calcula e EE. The e is a iance in he epo ing o he ype o EE being measu ed, wi h lack o cla i y in pape s ega ding whe he he measu e is TEE o AEE, and in he uni s used, wi h kilocalo ies pe minu e (kcals/min), kilocalo ies pe ac i - i y (kcals/ac i i y) and me abolic equi alen s (METs) all being desc ibed in he li e a u e. Recen li e a u e epo s METminu es o be he uni o choice o EE alida ion [15], ye while he ime-phased na u e o hese uni s a e desi able o compa abili y, we do no eel ha he consume ma - ke is cu en ly able o unde s and METminu es. Fu he - mo e, mos consume de ices epo EE in kcals. As such we ecommend ha he p e e ed me ic should be kcals pe uni o ime (minu e/day). Fu he mo e, esea che s should speci y which equa ion is used o de e mine EE om gas- eous consump ion, epo all ela ed assump ions, and a e clea in epo ing whe he he s udy assessed TEE o AEE. Equally wi h ega d o he index measu e, as much in o - ma ion as possible should be p o ided on he inpu s o he EE calcula ion o he de ice. Repo ing ha a p op ie a y algo i hm was used should be deemed as he minimum c i- e ia, wi h a emp s made o no e any o he inpu s o ha speci ic EE algo i hm, including hea a e, accele ome e da a, use demog aphics and subjec i e epo ing such as a e o pe cei ed exe ion. This is somewha challenging in many cases, gi en he ‘black-box’ na u e o hese p op i- e a y algo i hms ha manu ac u e s do no wish o di ulge. Addi ionally, whe e possible, esea che s a e encou aged o ollow open science p inciples in making aw da a a ailable using app op ia e da a eposi o ies o acili a e combined analyses in he u u e. Finally, he managemen o synch onisa ion be ween de ices is ano he po en ial sou ce o e o . The app oach o his will a y depending on he s udy design and de ices used. As such, esea che s mus be cognisan o da a syn- ch onisa ion when designing he s udy, and he synch o- nisa ion p ocess unde aken should be desc ibed in as much de ail as possible in he epo o allow o adequa e eplica ion. 3.6 S a is ical Analysis When assessing he alidi y o an index de ice o a c i e ion measu e, he le els o accu acy o he index de ice should be calcula ed wi h ag eemen be ween he wo de ices [65, 66]. The Bland–Al man app oach o limi s o ag eemen has become widely used in he cu en medical li e a u e [67], including 38 s udies (61%) in he sys ema ic e iew con- duc ing his app oach in he analysis. Howe e , he e was la ge a iance in he s a is ical app oaches used, wi h many s udies u ilising measu emen s o co ela ion (e.g. Pea - son, Spea man) [n = 38, 61%], compa ison o means (e.g. es ) [n = 22, 36%], and measu es o ela i e eliabili y (e.g. in aclass co ela ion coe icien ) [n = 13, 21%]. These app oaches ha e been highligh ed as inapp op ia e when assessing ag eemen be ween a measu emen ool and he e e ence s anda d in medical ins umen alida ion s udies [65]. To comp ehensi ely assess he alidi y o he index de ice in es ima ing EE, we ecommend he use o Bland–Al man limi s o ag eemen analysis in combina- ion wi h mean absolu e pe cen age e o (MAPE). The Bland–Al man me hod p o ides a measu e o he ag ee- men be ween he c i e ion and index de ice; esea che s should s a e i he assump ions o alid limi s o ag eemen analysis we e ul illed [67], and, in addi ion, should inco - po a e leas -p oduc s eg ession o assess o p opo ional o ixed bias, as desc ibed by Ludb ook [66]. The use o MAPE allows o compa ison be ween de ices and es ing condi ions and is he a e age o he absolu e e o o a ool. I is commonly used in desc ibing he e o o a p edic- ion [68], and 29 s udies (47%) included in ou sys ema ic e iew used his app oach in hei analysis. The use o bo h me hods and ela ed isualisa ions p o ides a comp ehen- si e assessmen o he g oup- and indi idual-le el alidi y De e mining Validi y o Wea ables o Es ima ion o Ene gy Expendi u e 7. Schoelle DA, Van San en E. Measu emen o ene gy expendi- u e in humans by doubly labeled wa e me hod. J Appl Physiol. 1982;53:955–9. 8. Düking P, Giessing L, F enkel MO, e al. W is -wo n wea ables o moni o ing hea a e and ene gy expendi u e while si ing o pe o ming ligh - o- igo ous physical ac i i y: alida ion s udy. JMIR Mheal h Uheal h. 2020;8:e16716. 9. Kendall B, Bello a y B, Go he NP. Validi y o wea able ac i i y moni o s o acking s eps and es ima ing ene gy expendi u e du - ing a g aded maximal eadmill es . J Spo s Sci. 2019;37:42–9. 10. Mu akami H, Kawakami R, Nakae S, e al. Accu acy o 12 wea - able de ices o es ima ing physical ac i i y ene gy expendi u e using a me abolic chambe and he doubly labeled wa e me hod: alida ion s udy. JMIR Mheal h Uheal h. 2019;7:e13938. 11. O’D iscoll R, Tu icchi J, Beaulieu K, e al. How well do ac i i y moni o s es ima e ene gy expendi u e? A sys ema ic e iew and me a-analysis o he alidi y o cu en echnologies. B J Spo s Med. 2020;54:332–40. 12. Fulle D, Colwell E, Low J, e al. Reliabili y and alidi y o com- me cially a ailable wea able de ices o measu ing s eps, ene gy expendi u e, and hea a e: sys ema ic e iew. JMIR Mheal h Uheal h. 2020;8:e18694. 13. Welk GJ, McClain J, Ainswo h BE. P o ocols o e alua ing equi alency o accele ome y-based ac i i y moni o s. Med Sci Spo Exe c. 2012;44(1 Suppl 1):S39-49. 14. Spe lich B, Holmbe g H-C. Wea able, yes, bu able…? I is ime o e idence-based ma ke ing claims! B J Spo s Med. 2017;51:1240. 15. Welk GJ, Bai Y, Lee J-M, e al. S anda dizing analy ic me hods and epo ing in ac i i y moni o alida ion s udies. Med Sci Spo s Exe c. 2019;51:1767. 16. Keadle SK, Lyden KA, S a h SJ, e al. A amewo k o e alu- a e de ices ha assess physical beha io . Exe c Spo Sci Re . 2019;47(4):206–14. 17. Mühlen JM, S ang J, Sko gaa d EL, e al. Recommenda ions o de e mining he alidi y o consume wea able hea a e de ices: expe s a emen and checklis o he INTERLIVE Ne wo k. B J Spo s Med. 2012;55(14):767–79. 18. Johns on W, Judice PB, Ga cía PM, e al. Recommenda ions o de e mining he alidi y o consume wea able and sma phone s ep coun : expe s a emen and checklis o he INTERLIVE ne wo k. B J Spo s Med. 2021;55(14):780–93. 19. Basse DR J , Rowlands AV, T os SG. Calib a ion and alida ion o wea able moni o s. Med Sci Spo s Exe c. 2012;44:S32. 20. F eedson P, Bowles HR, T oiano R, e al. Assessmen o physical ac i i y using wea able moni o s: ecommenda ions o moni o calib a ion and use in he ield. Med Sci Spo Exe c. 2012;44(1 Suppl 1):S1-4. 21. Ve i as. Co idence Sys ema ic Re iew So wa e. Ve i as Heal h Inno a ion; 2021. www. co id ence. o g. 22. Fukagawa NK, Bandini LG, Young JB. E ec o age on body composi ion and es ing me abolic a e. Am J Physiol Me ab. 1990;259:E233–8. 23. Poehlman ET, Go an MI, Ga dne AW, e al. De e minan s o decline in es ing me abolic a e in aging emales. Am J Physiol Me ab. 1993;264:E450–5. 24. Van Pel RE, Jones PP, Da y KP, e al. Regula exe cise and he age- ela ed decline in es ing me abolic a e in women. J Clin Endoc inol Me ab. 1997;82:3208–12. 25. A cie o PJ, Go an MI, Poehlman ET. Res ing me abolic a e is lowe in women han in men. J Appl Physiol. 1993;75:2514–20. 26. Fi zpa ick TB. The alidi y and p ac icali y o sun- eac i e skin ypes I h ough VI. A ch De ma ol. 1988;124:869–71. 27. Lu M-J, Zhong W-H, Liu Y-X, e al. Sample size o assess- ing ag eemen be ween wo me hods o measu emen by Bland−Al man me hod. In J Bios a . 2016. h ps:// doi. o g/ 10. 1515/ ijb- 2015- 0039. 28. Ca s ensen B. Compa ing clinical measu emen me hods: a p ac i- cal guide. New Yo k: Wiley; 2011. 29. Bu le PJ, G een JA, Boyd IL, e al. Measu ing me abolic a e in he ield: he p os and cons o he doubly labelled wa e and hea a e me hods. Func Ecol. 2004;18:168–83. 30. Wes e e p KR. Doubly labelled wa e assessmen o ene gy expendi u e: p inciple, p ac ice, and p omise. Eu J Appl Physiol. 2017;117:1277–85. 31. Da idsson L. Assessmen o body composi ion and o al ene gy expendi u e in humans using s able iso ope echniques. Vienna: In e na ional A omic Ene gy Agency; 2009. 32. DeLany JP, Lo ejoy JC. Ene gy expendi u e. Endoc inol Me ab Clin N Am. 1996;25:831–46. 33. Delsoglio M, Acham ah N, Be ge MM, e al. Indi ec calo ime y in clinical p ac ice. J Clin Med. 2019;8:1387. 34. M aweh H, Tui a L, Floh AA, e al. Indi ec calo ime y: his o y, echnology, and applica ion. F on Pedia . 2018;6:257. 35. Leona d WR. Labo a o y and ield me hods o measu ing human ene gy expendi u e. Am J Hum Biol. 2012;24:372–84. 36. Kinnunen H, Häkkinen K, Schumann M, e al. T aining-induced changes in daily ene gy expendi u e: Me hodological e alua ion using w is -wo n accele ome e , hea a e moni o , and doubly labeled wa e echnique. PLoS ONE. 2019;14:e0219563. 37. Mu akami H, Kawakami R, Nakae S, e al. Accu acy o wea - able de ices o es ima ing o al ene gy expendi u e: compa ison wi h me abolic chambe and doubly labeled wa e me hod. JAMA In e n Med. 2016;176:702–3. 38. Rami ez-Ma e o FA, Smi h BA, She man WM, e al. Compa ison o me hods o es ima e physical ac i i y and ene gy expendi u e in A ican Ame ican child en. In J Spo Med. 2005;26:363–71. 39. Siddall A, Powell S, Needham-Beck S, e al. Validi y o ene gy expendi u e es ima ion me hods du ing 10 days o mili a y ain- ing. Scand J Med Sci Spo . 2019;29:1313–21. 40. Dannecke K, Sazono a N, Melanson E, e al. A compa ison o ene gy expendi u e es ima ion o se e al physical ac i i y moni- o s. Med Sci Spo Exe c. 2013;45:2105–12. 41. E enson KR, Go o MM, Fu be g RD. Sys ema ic e iew o he alidi y and eliabili y o consume -wea able ac i i y acke s. In J Beha Nu Phys Ac . 2015;12:1–22. 42. Mon oye AHK, Mi zyk JR, Molesky MJ. Compa a i e accu acy o a w is -wo n ac i i y acke and a sma shi o physical ac i i y assessmen . Meas Phys Educ Exe c Sci. 2017;21:201–11. 43. Kane NA, Simmons MC, John D, e al. Validi y o he Nike plus de ice du ing walking and unning. In J Spo Med. 2010;31:101–5. 44. Nolan M, Mi chell J, Doyle-Bake P. Validi y o he Apple iPhone®/iPod Touch® as an accele ome e -based physical ac i i y moni o : a p oo -o -concep s udy. J Phys Ac Heal. 2014;11:759–69. 45. Maddison R, Gemming L, Monede o J, e al. Quan i ying Human Mo emen Using he Mo n Sma phone App: Valida ion and Field S udy. JMIR Mheal h Uheal h. 2017;5(8):e122. 46. Xie J, Wen D, Liang L, e al. E alua ing he alidi y o cu en mains eam wea able de ices in i ness acking unde a ious physical ac i i ies: compa a i e s udy. JMIR Mheal h Uheal h. 2018;6:e94. 47. Johnson M, Tu ek J, Do n eld C, e al. Validi y o he Samsung Phone S Heal h applica ion o assessing s eps and ene gy expend- i u e du ing walking and unning: Does phone placemen ma e ? Digi Heal. 2016;2:2055207616652747. 48. Cui Y, Chipchase J, Ichikawa F. A c oss cul u e s udy on phone ca ying and physical pe sonaliza ion. In: In e na ional con e ence on usabili y and in e na ionaliza ion. Sp inge , Be lin; 2007. pp. 483–92. R.A gen e al. 49. Chowdhu y E, Wes e n M, Nigh ingale T, e al. Assessmen o labo a o y and daily ene gy expendi u e es ima es om con- sume mul i-senso physical ac i i y moni o s. PLoS ONE. 2017;12:e0171720. 50. Thiebaud R, Funk M, Pa on J, e al. Validi y o w is -wo n con- sume p oduc s o measu e hea a e and ene gy expendi u e. Digi Heal h. 2018;4:2055207618770322. 51. Adam Noah J, Spie e D, Gu J, e al. Compa ison o s eps and ene gy expendi u e assessmen in adul s o Fi bi T acke and Ul a o he Ac ical and indi ec calo ime y. J Med Eng Technol. 2013;37:456–62. 52. Alsubheen S, Geo ge A, Bake A, e al. Accu acy o he i o i ac i i y acke . J Med Eng Technol. 2016;40:298–306. 53. Machač S, P ocházka M, Rad anský J, e al. Valida ion o physical ac i i y moni o s in indi iduals wi h diabe es: ene gy expendi u e es ima ion by he mul isenso SenseWea A mband P o3 and he s ep coun e Om on HJ-720 agains indi ec calo ime y du ing walking. Diabe es Technol The . 2013;15:413–8. 54. McA dle WD, Ka ch FI, Ka ch VL. Exe cise physiology: nu i- ion, ene gy, and human pe o mance. New Yo k: Lippinco Wil- liams & Wilkins; 2010. 55. McCla e SA, Spain DA, Skolnick JL, e al. Achie emen o s eady s a e op imizes esul s when pe o ming indi ec calo ime y. J Pa en e En e Nu . 2003;27:16–20. 56. Plowman SA, Smi h DL. Exe cise physiology o heal h i ness and pe o mance. New Yo k: Lippinco Williams & Wilkins; 2013. 57. Popp CJ, Tisch JJ, Saka can KE, e al. App oxima e ime o s eady-s a e es ing ene gy expendi u e using indi ec calo ime y in young, heal hy adul s. F on Nu . 2016;3:49. 58. Bo ges JH, Lange RD, Ci olini VX, e al. Minimum ime o achie e he s eady s a e and op imum abb e ia ed pe iod o es i- ma e he es ing ene gy expendi u e by indi ec calo ime y in heal hy young adul s. Nu Clin P ac . 2016;31:349–54. 59. Ree es MM, Da ies PSW, Baue J, e al. Reducing he ime pe iod o s eady s a e does no a ec he accu acy o ene gy expendi- u e measu emen s by indi ec calo ime y. J Appl Physiol. 2004;97:130–4. 60. Jones AM, Dous JH. A 1% eadmill g ade mos accu a ely e lec s he ene ge ic cos o ou doo unning. J Spo s Sci. 1996;14:321–7. 61. Sco CB, Leigh on BH, Ahea n KJ, e al. Ae obic, anae obic, and excess pos exe cise oxygen consump ion ene gy expendi u e o muscula endu ance and s eng h: 1-se o bench p ess o muscula a igue. J S eng h Cond Res. 2011;25:903–8. 62. Sco CB. Con ibu ion o blood lac a e o he ene gy expendi u e o weigh aining. J S eng h Cond Res. 2006;20:404. 63. Medbo JI, Mohn A-C, Taba a I, e al. Anae obic capaci y de e - mined by maximal accumula ed O2 de ici . J Appl Physiol. 1988;64:50–60. 64. Speakman J. Doubly labelled wa e : heo y and p ac ice. New Yo k: Sp inge Science & Business Media; 1997. 65. Zaki R, Bulgiba A, Ismail R, e al. S a is ical me hods used o es o ag eemen o medical ins umen s measu ing con inuous a i- ables in me hod compa ison s udies: a sys ema ic e iew. PLoS ONE. 2012;7:e37908. 66. Ludb ook J. S a is ical echniques o compa ing measu e s and me hods o measu emen : a c i ical e iew. Clin Exp Pha macol Physiol. 2002;29:527–36. 67. Bland JM, Al man DG. S a is ical me hods o assessing ag ee- men be ween wo me hods o clinical measu emen . Lance . 1986;327:307–10. 68. De My enae e A, Golden B, Le G and B, e al. Mean abso- lu e pe cen age e o o eg ession models. Neu ocompu ing. 2016;192:38–48. 69. Bennie JA, Shakespea -D ue y J, De Cocke K. Muscle-s eng h- ening Exe cise Epidemiology: a New F on ie in Ch onic Disease P e en ion. Spo Med. 2020;6:1–8. 70. Fe o Magosso R, Campanholi Ne o J, de Ca li JPC, e al. A Re iew o E gogenesis and E ec o T aining Va iables on Ene gy Expendi u e in Resis ance T aining Exe cises. J Exe c Physiol Online. 2017;20(2). 71. Sco C. Misconcep ions abou ae obic and anae obic ene gy expendi u e. J In Soc Spo s Nu . 2005;2:1–6. Au ho s and A ilia ions RobA gen 1,2,3 · MeganHe he ing on‑Rau h4· JulieS ang5· JakobTa p5· F anciscoB.O ega6,7· PabloMolina‑Ga cia6· Mo i zSchumann8,9· WilhelmBloch8· SulinCheng8,9,10· Ande sG øn ed11· JanCh is ianB ønd11· Ul Ekelund5· LuisB.Sa dinha4· B ianCaul ield1,2 1 Insigh Cen e o Da a Analy ics, Uni e si y College Dublin, Dublin, I eland 2 School o Public Heal h, Physio he apy andSpo Science, Uni e si y College Dublin, Dublin, I eland 3 School o Pha macy andBiomolecula Sciences, Royal College o Su geons inI eland, Dublin, I eland 4 Exe cise andHeal h Labo a o y, CIPER, Faculdade de Mo icidade Humana, Uni e sidade de Lisboa, Lisbon, Po ugal 5 Depa men o Spo Medicine, No wegian School o Spo Sciences, Oslo, No way 6 PROFITH (PROmo ing FITness andHeal h Th ough Physical Ac i i y) Resea ch G oup, Depa men o Physical Educa ion andSpo s, Facul y o Spo Sciences, Uni e si y o G anada, G anada, Spain 7 Depa men o Bioscience andNu i ion, Ka olinska Ins i u e , Solna, Sweden 8 Ins i u e o Ca dio ascula Resea ch andSpo s Medicine, Depa men o Molecula andCellula Spo s Medicine, Ge man Spo Uni e si y, Cologne, Ge many 9 Exe cise T ansla ional Medicine Cen e, heKey Labo a o y o Sys ems Biomedicine, Minis y o Educa ion, andExe cise, Heal h andTechnology Cen e, Depa men o Physical Educa ion, Shanghai Jiao Tong Uni e si y, Shanghai, China 10 Facul y o Spo andHeal h Sciences, Uni e si y o Jy äskylä, Jy äskylä, Finland 11 Depa men o Spo s Science andClinical Biomechanics, Resea ch Uni o Exe cise Epidemiology, Cen e o Resea ch inChildhood Heal h, Uni e si y o Sou he n Denma k, OdenseM, Denma k