When and How to Treat Subthreshold Depression
Depressive disorders are presentin about 10% of pri-mary care patients and account for more years lived withdisability than any single disease.1,2Nearly three-quartersof all outpatient visits fordepression are to primarycare clinicians rather than tomental health specialists.3Collaborative care is a therapeutic intervention in whichbehavioral health is integrated into primary care, most com-monly using a nurse care manager to monitor depressivesymptoms in depressed patients and adjust treatmentunder the supervision of a psychiatrist. Many of the nursecontacts are conducted by telephone, thereby increasing theefficiency of collaborative care. Although collaborative carehas been demonstrated to improve depression in more than80 randomized clinical trials,4,5most trials have targetedmajor depression.Major depression requires the presence of at least 5of 9 criterion symptoms of depression for 2 weeks orlonger, with at least 1 of the symptoms being depressedmood or anhedonia.6In comparison, subthreshold depres-sion (also called minor or subsyndromal depression) isthe presence of 2 to 4 criterion symptoms of depression for2 weeks or longer with at least 1 of the core symptoms(depressed mood or anhedonia).6,7Alternative definitionsof subthreshold depression use a severity cut point on adepression scale or vary in duration and core symptomrequirements but typically require the absence of majordepression. Only a small number of therapeutic trials havetargeted subthreshold depression, and results have beenmixed.6,8Identifying effective therapies for subthresholddepression is important because many patients with sub-threshold depression have persistent depressive symptomsat 12-month follow-up, a third to half report moderate func-tional impairment, and at least 10% to 20% progress to majordepression.6,7,9,10In this issue of JAMA, the CASPER trial by Gilbody et al11provides the first evidence that collaborative care may ben-efit patients with subthreshold depression. In this prag-matic clinical trial conducted in the United Kingdom, theauthors randomized 705 adults aged 65 years or older withsubthreshold depression to either a collaborative care inter-vention or usual primary care. The collaborative care treat-ment consisted of 8 weekly 30-minute sessions of behav-ioral activation administered by a care manager with abackground in mental health nursing or psychology whowas supervised by a mental health professional. Behavioralactivation is a psychological intervention that encouragesincreased social interactions and engagement in pleasur-able, rewarding activities. The first session was face-to-face,and subsequent sessions were delivered by telephone.Several important findings emerged from the CASPERtrial. First, improvement on the 9-item Patient Health Ques-tionnaire depression score (PHQ-9, which ranges from 0-27)at 4 months was 1.3 points greater in the collaborative caregroup compared with the usual care group (the PHQ-9 scorewas 7.8 in both groups at baseline and declined to 5.4 in thecollaborative care group and 6.7 in the usual care group).This difference was sustained at 12-month follow-up andrepresents an effect size of 0.3, consistent with a small tomoderate clinical improvement and comparable with theaverage effect size demonstrated in collaborative care trialsfor major depression.4,5This is notable because the meanbaseline PHQ-9 score among patients in this trial was only7.8 (median, 7), meaning that there was a limit to how muchimprovement could be achieved given that a score of lessthan 5 corresponds to minimal depressive symptoms.A second important finding was that progression tothreshold-level depression (defined in this trial as a PHQ-9score ≥10) was not different between the 2 groups at 4months (17.2% vs 23.5%) but was significantly less frequentin the collaborative care group compared with usual care at12 months (15.7% vs 27.8%). This represents a relative risk of0.65 (95% CI, 0.46-0.91). Third, the collaborative care inter-vention improved anxiety, functional status, and severalother secondary outcomes compared with usual care. Theeffect size for anxiety was 0.23 at 4-month follow-up, a sig-nificant between-group difference that was sustained at12-month follow-up and comparable with the average effectsizes of 0.30 to 0.33 reported in other studies examiningcollaborative care for anxiety.4,5The reduction in anxiety isespecially important because anxiety has a similar preva-lence as depression in primary care practice and is presentin a third to half of patients with depression.1Another important aspect of the CASPER interventionwas that it was brief and relatively inexpensive. The inter-vention consisted of 8 telephone sessions lasting 30 min-utes each delivered by care managers with less psychologi-cal training than psychologists or other mental healthprofessionals. Other study strengths included the largesample size; the low rate of antidepressant therapy, whichwas balanced among groups (thus minimizing the effect ofantidepressants as a cointervention and potential con-founder); and a sensitivity analysis showing that effective-ness did not vary among care managers. Although attritionwas higher in the collaborative care group than in the usualcare group, imputed analyses accounting for this differencedid not change the results
The core treatment in the CASPER trial was behavioralactivation, for which a previous meta-analysis of 26 trialsinvolving 1524 depressed patients also showed a reductionin depression.12Strong evidence for the effectiveness ofbehavioral activation was provided by the recent COBRAtrial, in which 440 adults with major depression were ran-domized to either behavioral activation or cognitive behav-ioral therapy.13Each group received an average of 12 one-hour in-person sessions, and behavioral activation wasfound to be noninferior to cognitive behavioral therapy forthe outcome of depression. A meta-analysis of brief psycho-logical therapies (typically ≤8 sessions) found that cognitivebehavioral therapy (13 trials), problem-solving therapy(12 trials), and counseling (8 trials) were all associated withimproved depression outcomes in primary care, with effectsizes ranging from 0.21 to 0.33.14Thus, behavioral activa-tion can now be added to the list of brief therapies thatmight be considered for treating depressed patients in pri-mary care.A recent German trial also found benefits in treatingsubthreshold depression: 406 adults with subthresholddepression were randomized to either a web-based guidedintervention (cognitive behavioral and problem-solvingtherapy supported by an online trainer) or a web-based psy-choeducation control program.15Fewer patients in the inter-vention group progressed to major depression at 12-monthfollow-up (27% vs 41%), resulting in a number needed totreat of 5.9 to avoid 1 new case of major depression (com-pared with a number needed to treat of 8.3 in CASPER). Cost-effectiveness results for the CASPER trial are planned but notyet available. Notably, collaborative care interventions formajor depression are cost-effective for the outcome ofquality-adjusted life-years.5,16Whether this is also true forcollaborative care interventions that target subthresholddepression, for which the absolute amount of improvementis less, needs to be evaluated.Do results of these recent trials mean that cliniciansshould expand therapy beyond major depression to includeactive treatment of subthreshold depression? There are sev-eral factors to consider. First, the way in which subthresholddepression is defined varies across studies.7The lack of aconsensus on the definition of subthreshold depression cancomplicate comparisons across treatment trials that definesubthreshold depression differently.Second, the rate at which subthreshold depression pro-gresses to major depression varies. Although the progressionrate in the control groups of the CASPER and German trialsranged from 28% to 41% over 12 months,11,15other studiessuggest a lower progression rate of 10% to 20%.6,9The higherprogression rates in the CASPER and German studies may bebecause trial participants had a moderate degree of func-tional impairment. Previous research has shown that the sub-group of patients with subthreshold depression and func-tional impairment is the most likely to benefit from activetreatment.6,17It is likely that if all patients with subthresholddepression were actively treated regardless of level of impair-ment, the magnitude of benefit would be less and the num-ber needed to treat large
Third, patient preferences must be considered. Onestudy of 1025 consecutive patients presenting to 19 generalpractices in New Zealand found that adding the question todepression screening “Is this something for which you wouldlike help?” markedly decreased the number of false-positivedepression diagnoses.18Although this study focused on diag-nosis rather than treatment, it is possible that patients withsubthreshold depression who do not desire treatment fortheir symptoms are less likely to benefit.Fourth, the duration of subthreshold depression and thecontext in which it occurs should be considered. Forexample, individuals who experience adjustment reactionsto adverse life events often have short-duration depressivesymptoms that resolve without active therapy. Thus, there isa role for watchful waiting in subthreshold depression ofrecent onset. Patients could be asked to monitor their symp-toms (possibly by completing a home-based PHQ-9 or otherdepression scale) and, if they fail to improve within a reason-able period (eg, 1-3 months), to make a follow-up appoint-ment. Alternatively, they could be scheduled for a follow-uptelephone call or, with advances in automated monitoring, beprompted to complete a web-based depression scale.The need to carefully weigh the benefits of treat-ing individuals who have symptoms but fall short of meet-ing criteria for major depression is not unique to sub-threshold depression but is applicable to prediabetes, mildobesity, statins for primary prevention, and numerousother health care decisions in individuals with mild dis-ease. In determining eligibility for the CASPER trial, theratio of patients with subthreshold to major depression was4:1. Before expanding treatment beyond major depressionto the larger number of patients with subthreshold depres-sion, further research to identify those most likely to benefitis warranted.Fifth, further study also is needed to determine whetherantidepressants have any role as an alternate therapy forsubthreshold depression. While antidepressants havefrequently not been better than placebo in randomizedtrials of subthreshold depression,8access to evidence-basedpsychological therapies varies across health care settings.Individual patient preferences for medication vs nonphar-macological treatments also vary. Limited evidence sug-gests that antidepressants might be beneficial in patientswith more severe subthreshold depression, such as thosewith functional impairment or suicidal ideation.6,17,19Cur-rently, evidence is stronger for brief psychological therapieswhen a treatment intervention is considered for subthresh-old depression.In summary, CASPER provides new evidence that collab-orative care improves outcomes for at least some patientswith subthreshold depression. A principal component oftreatment in CASPER was behavioral activation. Other briefpsychological therapies such as cognitive behavioral therapyand problem-solving therapy have also proven effectivein some studies. Accessibility of these psychological treat-ments can be enhanced by telephone or web-based delivery.Patients with persistent symptoms, functional impairment,and a desire for treatment may particularly benefit.
REFERENCES1.
Löwe B, Spitzer RL, Williams JB, Mussell M,Schellberg D, Kroenke K. Depression, anxiety andsomatization in primary care: syndrome overlapand functional impairment.Gen Hosp Psychiatry.2008;30(3):191-199.2. Murray CJ, Atkinson C, Bhalla K, et al; US Burdenof Disease Collaborators. The state of US health,1990-2010: burden of diseases, injuries, and riskfactors.JAMA. 2013;310(6):591-608.3. Olfson M, Blanco C, Marcus SC. Treatment ofadult depression in the United States.JAMA InternMed. 2016;176(10):1482-1491.4. Archer J, Bower P, Gilbody S, et al. Collaborativecare for depression and anxiety problems.CochraneDatabase Syst Rev. 2012;10:CD006525.5. Reed SJ, Shore KK, Tice JA. Effectiveness andvalue of integrating behavioral health into primarycare.JAMA Intern Med. 2016;176(5):691-692.6. Kroenke K. Minor depression: midway betweenmajor depression and euthymia.Ann Intern Med.2006;144(7):528-530.7. Rodríguez MR, Nuevo R, Chatterji S,Ayuso-Mateos JL. Definitions and factorsassociated with subthreshold depressiveconditions: a systematic review.BMC Psychiatry.2012;12:181.8. Baumeister H. Inappropriate prescriptions ofantidepressant drugs in patients with subthresholdto mild depression: time for the evidence tobecome practice.J Affect Disord. 2012;139(3):240-243.9. Lyness JM, Heo M, Datto CJ, et al. Outcomes ofminor and subsyndromal depression among elderlypatients in primary care settings.Ann Intern Med.2006;144(7):496-504.10. Jaffe A, Froom J, Galambos N. Minordepression and functional impairment.Arch FamMed. 1994;3(12):1081-1086.11. Gilbody S, Lewis H, Adamson J, et al. Effect ofcollaborative care vs usual care on depressivesymptoms in older adults with subthresholddepression: the CASPER randomized clinical trial.JAMA. doi:10.1001/jama.2017.013012. Ekers D, Webster L, Van Straten A, Cuijpers P,Richards D, Gilbody S. Behavioural activation fordepression: an update of meta-analysis ofeffectiveness and sub group analysis.PLoS One.2014;9(6):e100100.13. Richards DA, Ekers D, McMillan D, et al. Costand outcome of behaviour
0 Comments