Computers & Electronics

A pilot study of adherence with light treatment for seasonal affective disorder

A pilot study of adherence with light treatment for seasonal affective disorder
of 6
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Brief report  A pilot study of adherence with light treatment for seasonal affective disorder  Erin E. Michalak   a, ⁎ , Greg Murray  b , Clare Wilkinson  c ,Chris Dowrick   d , Raymond W. Lam  a  a   Department of Psychiatry, University of British Columbia, 2255 Wesbrook Mall Vancouver, BC, Canada V6T 2A1  b  Faculty of Life and Social Sciences, Swinburne University of Technology, Hawthorn, Australia c  Department of General Practice, University of Wales College of Medicine, Wrexham, United Kingdom d  Department of Primary Care, University of Liverpool, Liverpool, United Kingdom Received 15 November 2005; received in revised form 21 February 2006; accepted 3 May 2006 Abstract  Non-adherence with antidepressant medication regimens is now recognised as a substantial problem when evaluating depressionoutcome. Given the behavioural demands of light treatment (LT), it might be expected that non-adherence would be even more pronounced in LT, a form of intervention for seasonal affective disorder (SAD). However, little research has focused upon the extent towhich patients in light treatment protocols adhere to set regimens. Nineteen patients with SAD were allocated to either treatment with brightwhitelight(intervention)ordimredlight(controlcondition)inafour-weekprotocol.Lightexposurewasestimatedautomatically(withoutparticipants'knowledge)withelapsedtimemetersbuiltintothelightbox.Dailydiarieswerealsousedtomeasureself-reportedlightboxuse.Participantswereinstructedtousethelightboxfor30mineachdayduringweek1,45duringweek2andonehourduringweeks3 and 4 (total duration ofprescribed light exposure1365min).The resultsindicatedthatmean duration oflightboxoperation for the entire sample was 59.3% of the prescribed 1365 min. Six of nineteen (31.6%) patients dropped out of treatment. Amongst thosecompleting treatment, adherence to the prescribed duration of exposure averaged 83.3% (S.D.=31.4). A trend was found for theintervention condition to generate a lower dropout rate, as well as a trend for the degree of adherence to be greater in the interventioncondition. Importantly, there was no association between adherence as measured automatically and the higher rates of self-reportedadherence as measured in diaries. In summary, the results of this pilot study suggest that adherence with light treatment is of a similar order of magnitude to antidepressant medication adherence. Patient self-report was found to be unrelated to objectively estimatedduration of light box use, a finding with significant research and clinical implications. Future research studies should routinely measureand evaluate adherence with light therapy and evidence-based techniques for maximising treatment adherence should be incorporatedinto routine clinical practice.© 2006 Elsevier Ireland Ltd. All rights reserved.  Keywords:  Seasonal affective disorder; Light therapy; Adherence 1. Introduction Seasonal affective disorder (SAD) is a variant of recurrent major depression characterised by typical andatypical depressive symptomatology with a distinct  Psychiatry Research 149 (2007) 315 – ⁎  Corresponding author. Division of Clinical Neuroscience, Depart-ment of Psychiatry, University of British Columbia, 2255 Wesbrook Mall Vancouver, BC,Canada V6T2A1.Tel.: +1 6048273393;fax:+1604 822 7792.  E-mail address: (E.E. Michalak).0165-1781/$ - see front matter © 2006 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.psychres.2006.05.005  seasonal pattern (Rosenthal et al., 1984; Lam and Levitt,1999). More than seventy controlled studies of light therapy for SAD have been conducted, and three sys-tematic reviews incorporating meta-analyses have alsosupported the efficacy of the treatment intervention (Leeand Chan, 1999; Thompson, 2001; Golden et al., 2005).This evidence resulted in the recommendation of light therapy as a first-line treatment for SAD in expert andconsensus clinical guidelines (Lam and Levitt, 1999;American Psychiatric Association, 2000; Kennedy et al.,2001; Bauer et al., 2002). However, certain clinical ques-tions remain to be addressed with respect to light treat-ment. In particular, the behavioural restrictions of thetreatment intervention are substantial; patients are typi-cally required to sit in front of a light box for at least 30 min each morning, sometimes for upwards of an hour (seePartonenandMagnusson,2001).Apriori,itmightbeexpected that such regimens would present a significant challenge to patient adherence. 1  Non-adherence is receiving growing attention as afactorinantidepressanttreatmentoutcome(Demyttenaereet al., 2004; Pampallona et al., 2004). In a review of theliterature,LingamandScottfoundamedianprevalenceof non-adherence to be 53% (Lingam and Scott, 2002). Incontrolled therapeutic studies, dropout rates ranging from21%to30%havebeenreported(Pampallonaetal.,2002),and a 30% dropout rate was identified in a meta-analysis(Bollini et al., 1999). Although dropout rate is the most common measure of adherence, Lingam and Scott alsoemphasise that this dichotomous measure provides anincomplete estimate of adherence, and patients may be partly or intermittently adherent. For example, Demytte-naere and colleagues have assessed adherence quantita-tively as the proportion of days on which treatment wasfollowed (Demyttenaere et al., 1998).In comparison, very little research has examined theissue of adherence with light therapy, or tried to ascertainhow these rates may compare to those obtained for anti-depressants, which are recognised as an alternative formoftreatmentforSAD(Kasperetal.,2001).Inaconference proceeding, Desan and colleagues have reported on datafrom seven participants enrolled in a controlled study of light therapy (intervention) vs. deactivated negative iongenerator (control) for SAD in which light therapy adhe-rence was measured by microprocessors concealed in thelight boxes that recorded time and duration of treatment (Desan et al., 2004). Although the majority of their parti-cipantsusedtheirlightboxesonadailybasis,theydidnot accurately report missed days, duration of treatment or timeoftreatment,failingtocompletetreatmentby8amasinstructed on 41% of days. Adherence in the control armof the study was not measured, and could not thereforeshed any light on the on-going debate concerning plausible control interventions for bright light (Eastman,1990). Most controlled trials of light therapy have usedeitherdimredlightorlow-densitynegativeiongeneratorsas a  ‘  placebo ’  intervention. Bright white light has beenshown to be more efficacious than red light, but given themedia attention on SAD, doubtsremain about red light asa plausible placebo (Eastman et al., 1998), and in turnadherence rates with red light, which may be reduced if expectations of treatment efficacy are lower (Masand,2003).We expand here on pilot data on adherence in a com-munity-based controlled trial of light therapy that we previously published in a letter format (Michalak et al.,2002).Thisextendedpaperexpandsonourpreviousbrief  publication byincorporating moredetailedstatistical ana-lyses of the data and taking into account literature onadherence with antidepressant medications that was not available at the time of the srcinal publication. As aninitial attempt to illuminate the relationship between ad-herence, expectations and intervention, we comparedadherence in the intervention versus control conditions,and tested whether adherence was associated with expec-tationsheldbypatientsinbothgroups. 2 Threepredictionswere set to focus analyses. First, as assessed in dropout rate, it was predicted that adherence with light therapywould be poorer than the approximate one third non-adherence observed across antidepressant studies. Adher-encewasalsoexploredasa quantitativevariable,intermsof the recorded light exposure as a percentage of the total prescribed light exposure. Second, it was expected that more positive expectations of treatment would be asso-ciated with better adherence. Third, it was expected that adherence would be poorer under the red light controlconditionincomparisontothebrightwhitelighttreatment condition,possiblyduetolowertreatmentexpectationsor experienced inefficacy. Finally, self-reported light boxusage was assessed in daily logs permitting calculation of the correlation between patient-derived and objectivelyderived rates of adherence. 2. Methods Participants were recruited as part of the Outcomes of DepressionInternationalNetwork(ODIN)project,alarge 2 The relationship between these variables and clinical outcomes ismore complex again (see, for example, Murphy and Coster, 1997). 1 The term adherence is considered preferable to the related termcompliance, because it emphasises active participation from the patient (Lingam and Scott, 2002).316  E.E. Michalak et al. / Psychiatry Research 149 (2007) 315  –  320  study designed to assess the prevalence, risk factors andoutcomes of depressive disorders in rural and urban areaswithin the European Community (Dowrick et al., 1998);additional work examining the prevalence of SAD wasconducted in conjunction with standard ODIN method-ology at the project's North Wales site in the UnitedKingdom. Ethical approval for the study was providedfrom the South Clwyd Ethics Committee. Residents(  N  =1999) of a rural area of North Wales were screenedfor SAD by mail using a sub-scale of the Seasonal Pat-ternsAssessmentQuestionnaire(SPAQ)(Rosenthaletal.,1987) (see Michalak et al., 2001) for the fullresults of the screeningphaseofthestudy).Participantshavingaglobalseasonality score (GSS, range 0 – 24) of 11 or greater onthe SPAQ subsequently underwent diagnostic interviewover two consecutive winters, being required to meet minimum Hamilton Depression Rating Scale (HDRS)scores and DSM-IV criteria during both winters. Eligible participants (  N  =19) entered a four-week community based protocol. Mean age of the sample was 45.2 yr (S.D.=8.55, range 30 – 63), and 15 of the 19 participants(78.9%) were female. The SPAQ global seasonality score(GSS) ranged from 12 to 20 (  M  =14.11, S.D.=2.54).Interventionandcontrolgroupsdidnotdiffersignificantlyon any of these variables as assessed by relevant non- parametric analyses.Intervention and control light was provided for a4-week intervention period by bright white (Outside InSunray Max 10,000 lx @ 51 cm, 3500 K triphosphor lamps) and dim red (modified Outside In Sunray Max430lx@ 56cm720nMredacrylic filter) light boxes.Alllight boxes were fitted with simple elapsed time metersconcealedbehindthereflectivepanelattherearofthebox.The meters recorded the total number of minutes the light  box had been switched on for, but not time of use.Treatment expectation was assessed on a four-item scaleadapted from Borkovec and Nau (Borkovec and Nau,1972). Pre- and post-treatment, participants rated on a7-pointscale:1)howlogicalthe treatmentseemed2)howuseful they thought it would be 3) how confident theywere it would be successful and 4) how confident theywouldbeinrecommendingthetreatmenttoafriend.Inthe present sample, responses to the four items were stronglyassociated and hence formed a reliable scale (Cronbach'salpha=0.87 and 0.81 at pre- and post treatment,respectively).Scalescoresforexpectationswerethereforeused in analyses. Participants completed the expectationquestionnaire at baseline, after seeing their allocated light  box switched on, and again at the end of the trial.Participants received verbal and written instructionsfromablindedresearcheraboutthepositioninganduseof the light box, and the prescribed timing of treatment. Asregards the positioning, they were instructed to place thelightboxonaflatsurfaceatanangleofapproximately30°to their body, with their eyes at mid-fixture level. Theywereinstructednot tostare directlyatthe light but togazeacross at it once or twice a minute. As regards the timingof the intervention, participants were instructed to use thelight box for 30 min each day during week 1, 45 minduring week 2 and one hour during weeks 3 and 4. Theywereadvisedthatthemostbeneficialtimeofusewouldbemornings, but that use before 7 PM would be acceptable.The total duration of light exposure prescribed for theregimen was 1365 min. Participants were also required tocomplete daily diaries recording the length of time thelight box was used for. Research staff did not systemat-ically contact patients during the 4-week treatment pro-tocol regarding their adherence with the treatment intervention. 2.1. Data analysis To permit comparison with the existing literature onantidepressant adherence, the central analysis involvedadherence measured dichotomously in dropout rates.Overt dropout was defined as patients informing the ex- perimenter that they had discontinued treatment. UsingthecategorizationofferedbyDemyttenaere(1997), ‘  poor  ’ adherence was defined as using the light box for 25% or less of the prescribed duration, and participants showing poor adherence under this definition were deemed covert dropouts. In light of the small sample size, nonparametricstatistics were calculated throughout. 3. Results 3.1. Dropout rates and the extent of adherence Two of the nineteen participants (10.5% of the totalsample) were overt dropouts. Light meter readings sho-wed that their light boxeshad been in operation for 7 and28 min. Four participants (21.1% of the total sample)were considered dropouts on the basis of less than 25%adherence (43, 98, 182 and 224 min of light box ope-ration). In total, therefore, six of nineteen patients weredeemed dropouts (31.6%). When analyses were con-ducted excluding the six dropouts described above,quantitatively assessed adherence was remarkably good.In the sub-sample completing the regimen (analysis n =13), mean adherence was 83.3% (S.D.=31.4). Whenthe four covert dropouts were retained in analyses(analysis  n =17), mean adherence was 66.1% (S.D.=42.1). When both covert and overt dropouts wereincluded in analyses (analysis  n =19, the equivalent of  317  E.E. Michalak et al. / Psychiatry Research 149 (2007) 315  –  320  an intention-to-treat analysis), mean adherence fell to59.3% (S.D.=44.7). 3.2. Moderators of adherence: expectations and group Bothbeforeandaftertreatment,patientexpectationsinthe whole sample were high in relation to the maximum possiblemeanof7(  M  =5.81,S.D.=1.08,  M  =5.96,S.D.=1.00, for pre- and post, respectively). Wilcoxon signedranks test found that expectations were not significantlydifferent pre- versus post treatment (  Z  = − 28,  P  N 0.05).Analyses of expectations as a moderator of adherencewere therefore calculated on a score averaged across pre-and post treatment reports (mean of this aggregate vari-able was 5.92, S.D.=0.75).Patient expectations were found to be unrelated toadherence. Comparison of expectations in dropouts (meanrank=11.00) versus completers (mean rank=8.92) foundno significant relationship (Mann – Whitney  U  =25.00,  Z  = − 74,  P  N 0.05). Similarly, nonparametric correlation betweenexpectationsandthedegreeofadherencefoundnoassociation (Kendall's  τ  b =0.09,  P  N 0.05). The impact of group membership (intervention vs. control) on adherencewas assessed both dichotomously and quantitatively. Anon-significant trend was found for the intervention con-dition to generate a lower dropout rate ( χ 2 (1)=3.32,  P  N 0.05). Both of the overt dropouts were from the controlgroup, as were three of the four covert dropouts. A trendwas also found for the extent of adherence to be higher inthe intervention (  M  =78.61%, S.D.=43.23) as opposed tothe control group (  M  =41.86%, S.D.=40.23), but this dif-ference again did not reach significance (Mann – Whitney U  =27.0,  Z  = − 1.47,  P  N 0.05). To complete the investiga-tion of moderators of adherence, it was useful to test the potential association between expectations and groupmembership. Patient expectations did not differ in the bright white light condition (mean rank=8.22) versus thedimredlightcontrolcondition(meanrank=10.78,Mann – Whitney  U  =29.00,  Z  = − 1.02,  P  N 0.05). 3.3. Self-report vs. objective estimate of adherence Finally,theassociationbetweenobjectivelymeasuredand self-reported light box use was investigated. Asnoted above, mean duration of light box operation in thewhole sample was 59.3% of the prescribed 1365 min.Self-reported light box use (  M  =85.2%, S.D.=23.1) wassignificantly inflated relative to the objective estimate(  Z  = − 2.03,  P  b 0.05). Furthermore, the nonparametriccorrelationbetweenself-reportandobjectiveestimatesof adherence was not significant (Kendall's  τ  b =11,  P  N 0.05). This pattern of significant mean differenceand absent correlation was not affected by removingdropouts from the analyses. 4. Discussion This pilot investigation of adherence with light treat-ment in the context of a clinical trial generated a series of findingsthatwarrant replication and refinementinfurther research. Most importantly, the overall dropout rate for light treatment was found to be 31.6%, a nonadherencerate which is no higher than the approximately one thirdreported across studies of antidepressant adherence.Whenonlyovertdropoutswerecounted,a smallminorityof the sample (10.5%) were deemed nonadherent under the dichotomous definition. Furthermore, amongst thesubgroup that completed the regimen, mean adherencecould be described as  ‘ good ’  against the 75% criterionnoted in the literature (Demyttenaere, 1997). Thus, light therapy regimens do not appear to present a significant  behavioural burden compared to antidepressant medica-tion regimens, an alternative form of treatment for SAD.Light treatment involves a not insignificant beha-viouralchangeonthepartofthepatient,sohowshouldweunderstand the relatively high adherence found here?First,medicationsideeffects are a factorinantidepressant non-adherence (Myers and Branthwaite, 1992; Le Pen et al.,1994).Whilstlighttherapymaycausesideeffectssuchas eye strain, headaches and insomnia (Levitt et al.,1993;Kogan and Guilford, 1998) these symptoms tend to berelatively mild, and the therapy appears to be welltolerated by the majority of patients. Second, the general population holds quite negative views of pharmacother-apy for depression (Angermeyer and Matschinger, 1996),while there may be a lay attraction to the notion of seasonalmoodpatternsandthe consequentlinkwithlight treatment(Murray,2001).Similarly,inthepresentsample patient expectations of light treatment efficacy were re-markably high both before and after treatment. Relativelygood adherence in light therapy regimens may therefore be the result of lower side effect burden, or holding positiveattitudesandexpectationstowardsthe potentiallyless invasive light therapy.As expected, in both categorical and quantitativemeasures of adherence, there was a trend for adherenceto be better in the intervention than the controlcondition. As discussed by Demyttenaere (1997), thereis a complex relationship between adherence, clinicalefficacy and expectations, and only some elements of this picture could be tested here. Expectations did not differ between the two groups, either upon first viewingthe light box or at completion of four weeks' treatment,so the trend for increased adherence in the intervention 318  E.E. Michalak et al. / Psychiatry Research 149 (2007) 315  –  320  condition cannot be reduced to expectancy effects. It therefore can be hypothesised that adherence tended to be better in the intervention condition because the bright white light condition was experienced as clinicallysuperior during the protocol. Although beyond thescope of the present study, future research could usefullymeasure symptoms repeatedly during the protocol to test the hypothesis that adherence in the interventioncondition is improved by patient experience of symptomamelioration. The finding that expectations did not differ between the two conditions is also incidentallyimportant insomuch as it constitutes some evidence that dim red light is a plausible placebo in studies of light treatment for SAD (cf. (Eastman, 1990)). More problematic for ongoing research into light treatment for SAD is the finding that self-reports of light box usewere significantly exaggerated relative to objectivelymeasured light box use, a finding echoed in other  preliminary research examining treatment adherence tolight therapy regimes (Desan et al., 2004).The present study had a number of limitations. Most significantly, duration of operation of the light box is anindirectmeasureoflightexposureandwould,ifanything,tend to inflate estimates. Given that our measure of adhe-rencemustbeattheupperlimitoftruevalues,itisstrikingthat in the present study, self-report estimates of durationwere significantly larger than automatic estimates basedon the duration of operation of the box. Future researchcould take advantage of portable wrist-worn light meters(see, for example, Espiritu et al., 1994) to provide moredirect estimates of the duration of light exposure. Increa-singlysophisticatedmeasuresofpatientadherenceinlight therapy trialsshouldthenallow us toaddress the questionof whether better adherence in turn results in better thera- peutic result. Second, only large experimental effectscould be observed with the small sample studied here.Type II error may explain, for example, the failure of thegroup difference in adherence to reach two-tailed signi-ficance. Third, participants were not systematically con-tactedbytheresearchteamregardingtheiradherencewiththe treatment regimen. Regular contact with the partici- pants (as would be expected in good clinical practice)mighthavehadapositiveimpactuponadherencerates,asmight the use of daily telephone log-ins to record adhe-rence, rather than daily diaries. Fourth, the instrument used to measure treatment expectations for light therapyin the study was relatively simplistic. Fifth, the doseincrease light therapy protocol utilised in the present study would not now be considered a standard light therapy regimen  —  clinical guidelines for the treatment of SAD recommend 30 min of morning light exposurewith a 10,000 lx light box (Lam and Levitt, 1999).Finally, although the present protocol involvedadministration of light treatment in the community, adhe-rence data were collected in the context of a controlledclinical trial, and thus may not generalise to routine cli-nical practice. Regardless, it seems appropriate to incor- porate evidence-based techniques for maximisingtreatment adherence into routine clinical practice when prescribing light therapy, and to routinely measure andevaluate adherence with the treatment. Recent theorisingand research into antidepressant adherence underscoresthe importance of engaging with the patient as an activecollaborator in treatment planning (Lingam and Scott,2002), and is likely also to be a key factor in the optimalmanagement of SAD. 5. Conclusion Perhaps because SAD research is underpinned by anunalloyed biological paradigm, the complex issue of light treatment adherence has received almost no attention todate. The present study provides provisional evidence that adherence with a light treatment regimen and adherencewith antidepressant medication are of similar magnitude,viz. approximately one third of participants do not ade-quately complete treatment. Further research is needed toclarifythe importance of relatively positive expectations of light treatment, active vs. control conditions, and clinicaloutcomes in moderating adherence. Future studies will benefit from using more direct operationalisations of light exposure and forming hypotheses based on developingknowledge in relation to antidepressant adherence. Acknowledgments The North Wales arm of the ODIN project receivedfinancial support from the EC Biomed 2 Programme(Contract BMH-4-CT96-1681)andthe WalesOfficeofR and D (Contract RC092). The light boxes used in the trialwere provided by Outside In light box company, Cam- bridge. E. Michalak is supported by a Michael SmithScholar Award from the Michael Smith Foundation for Health Research and a CIHR New Investigator Award. References American Psychiatric Association, 2000. Practice guideline for thetreatment of patients with major depressive disorder (revision).American Journal of Psychiatry 157, 1 – 45.Angermeyer, M.C., Matschinger, H., 1996. Public attitude towards psychiatric treatment. Acta Psychiatrica Scandinavica 94, 326 – 336.Bauer, M., Whybrow, P.C., Angst, J., Versiani, M., Moller, H.-J., 2002.WFSBP Task Force on Treatment Guidelines for Unipolar Dep-ressive Disorders World Federation of Societies of Biological319  E.E. Michalak et al. / Psychiatry Research 149 (2007) 315  –  320
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks