How to manage depression and anxiety: when self-help works, when you need a clinician

When searching "how to manage" is the right question

Each month 16,000+ people in Russia type "how to manage anxiety" or "how to manage depression" into search engines. This is not "weak motivation" or "help avoidance" — it is a normal first step for most people. Not every case of depressive or anxious symptoms requires immediate clinician contact. There is a zone where self-help empirically works.

But there is also a zone where self-help no longer works, and delaying clinician contact worsens prognosis. This article is about a decision tree that distinguishes those zones without a psychology degree. It is not "simplified therapy in an article" — it is a methodological reference: when self-help can work, what to do structurally if so, and which red flags mean "reach out now".

What empirically works in self-help

Self-help is a broad term covering very different interventions with very different evidence bases. It is worth distinguishing three categories.

Guided self-help — structured programs (books, CBT apps, online courses) with minimal clinician support (≤ 2–3 brief contacts). Cuijpers et al. (2019) in the largest NMA to date (155 RCTs, N=15,191): SMD 0.87–1.02 for guided self-help / individual / group / telephone CBT vs waitlist in depression. Cuijpers et al. (2010) in a direct comparison: the difference between guided self-help and face-to-face CBT was d=−0.02 (effectively equivalent). Hedman-Lagerlöf et al. (2023, *World Psychiatry*, 31 RCT): g=0.02 for iCBT vs face-to-face. This is the most evidence-supported zone for self-help.

Unguided digital interventions — fully self-directed apps and programs without human guidance. Karyotaki et al. (2017, *JAMA Psychiatry*, IPD-MA, N=3,876) found g=0.27 for self-guided iCBT — substantially lower than guided, NNT = 8. Cuijpers et al. (2023, *World Psychiatry*, N=52,702) reported g=0.45 for unguided self-help format in broader samples. They work, but less reliably: dropout rate in digital self-guided programs is often 30–60% (Karyotaki 2021).

"Pop-psychology self-help" — bestselling books without structured methodology, motivational blogs, popular podcasts. The evidence base is weak or absent. Not "harmful", but not an intervention in the evidence-based sense — closer to psycho-educational content. Cuijpers 1997 in the classic bibliotherapy meta-analysis showed effect sizes for traditional bibliotherapy lower than for modern structured guided self-help, but still significant for mild symptoms.

The core conclusion from the research literature — self-help works in the mild-to-moderate range with a structured approach. With a program (not "random advice from the internet"), with regularity (not "when I have energy"), with a measurable progress tracker (PHQ-9 / GAD-7). In this zone, effect size is comparable to face-to-face short-term CBT, at substantially lower cost and higher accessibility. Karyotaki et al. (2021, *JAMA Psychiatry*, IPD-NMA, N=9,751) showed moderation by severity: the advantage of guided iCBT over unguided grows with baseline severity (MD −0.8 PHQ-9 at PHQ-9 > 9, smaller difference at PHQ-9 5–9).

Decision tree: 5 criteria — self-help vs clinician

Five criteria for a structured assessment of "is self-help enough or is it time for a clinician". This is not a diagnostic algorithm — it is a guide for your own decision.

• 1. Symptom duration. Up to 4 weeks — self-help zone. More than 2 months of persistent symptoms — reason to reach out. Between 1 and 2 months — transitional zone: try structured self-help, reassess at 4 weeks via PHQ-9 / GAD-7.
• 2. Trend. Are symptoms improving with self-help attempts? If after 4 weeks of regular effort PHQ-9 does not move or rises — self-help is not working in this zone, a clinician is needed.
• 3. Functional impact. Can you continue working, maintain relationships, take care of daily tasks? Functioning preserved — self-help applies. Serious disruption to work, relationships, or basic self-care — this is no longer "mild", it is moderate-severe.
• 4. Suicidal ideation. Any active suicidal thoughts (plan, means, intent) — automatic red flag, not a self-help zone. Passive thoughts ("I wish I wasn't here") with preserved functioning are a gray zone — requires conversation with a clinician or close person, but not necessarily immediate crisis intervention.
• 5. History. Previous depression or anxiety episodes, especially ones that required medical help — a risk factor for recurrence. Symptoms recurring in someone with this history make clinician contact methodologically more sound than "let's try self-help again".

Decision rule: 4–5 "green" criteria (short duration, positive trend, functioning preserved, no suicidal ideation, no heavy history) — self-help zone. 2 or fewer "green" — time for a clinician. 3 "green" — transitional zone: try structured self-help with strict 4-week tracking, then reassess.

Red flags for immediate clinician contact

There is a set of situations where self-help is not appropriate in principle — immediate contact with a clinician or emergency service is needed.

• Active suicidal ideation with plan or intent — crisis line (in Russia: 8-800-2000-122 for children and adolescents; for adults — regional services) or emergency department. C-SSRS assessment (Columbia Suicide Severity Rating Scale, Posner et al. 2011) is mandatory and is administered by a clinician, not as self-report
• Psychotic symptoms — auditory / visual hallucinations, marked reality-testing disturbance, paranoid ideation exceeding anxious norm — psychiatrist within 72 hours
• Mania or hypomania — markedly elevated mood with impulsivity, decreased need for sleep ≥ 4 days, grandiose thinking — psychiatrist (bipolar disorder requires different tactics than unipolar depression)
• Substance use out of control — especially with comorbid depression or anxiety when "self-medicating with alcohol" becomes regular
• Significant deterioration within 1–2 weeks — abrupt drop in functioning (loss of work due to depression, isolation from family, unable to manage basic tasks) — not a self-help zone, requires clinical assessment

These situations are not "more severe depression / anxiety" — they are qualitatively different categories. Self-help fails here not because "self-help isn't strong enough", but because of a mismatch between the type of intervention and the nature of the condition.

How to self-track progress — PHQ-9 / GAD-7

The main difference between "emotional self-monitoring" and structured measurement is that the latter gives you a number you can compare to your own past result. This is the basic methodology of measurement-based care (MBC) applied to individual self-help.

Two tools you can rely on without a psychology degree. PHQ-9 (Spitzer et al. 1999) — 9 items, 5 minutes, public domain. Score 0–27, severity bands: 0–4 minimal, 5–9 mild, 10–14 moderate, 15–19 moderately severe, 20–27 severe. Reliable Change Index ≈ 5 points — a change of 5+ points between measurements is statistically significant movement, not noise. PHQ-9 is validated: sensitivity 88% / specificity 85% at cutoff ≥ 10 (Levis et al. 2019 IPDMA, n=17,357).

GAD-7 (Spitzer et al. 2006) — 7 items, 3 minutes, public domain. Score 0–21, severity bands: 0–4 minimal, 5–9 mild, 10–14 moderate, 15–21 severe. RCI ≈ 4 points. Plummer et al. (2016, *Gen Hosp Psychiatry*, n=5,223) confirmed sensitivity 0.83 and specificity 0.84 at cutoff ≥ 8 for GAD-7 screening of generalized anxiety. It also partially covers panic-related symptoms.

• Step 1 — baseline. Complete PHQ-9 + GAD-7 once at the start (e.g., Sunday evening). Record the values — this is your reference point
• Step 2 — repeat every 2–4 weeks. Same day of the week, same time. Log numbers in a simple table: date, PHQ-9, GAD-7, brief note about the week
• Step 3 — interpretation. Change ≥ RCI (5 for PHQ-9, 4 for GAD-7) downward — real improvement. Change in the opposite direction — worsening. Change below RCI — statistical noise, no conclusion
• Step 4 — escalation criterion. If 4–8 weeks of structured self-help produces no ≥ RCI reduction — self-help is not working in your zone. Time to reach out to a clinician

This methodology is what Shimokawa, Lambert & Smart (2010) in a mega-analysis on n=6,151 showed substantially increases recovery rate in signal-alarm cases when feedback-informed practice is added to therapy. The same logic applies in parallel to structured self-help: the number increases the chance of noticing deterioration or stagnation that otherwise wouldn't be reacted to until a critical point. For more on MBC methodology — "Evidence-Based Psychology in Practice" and "Supervision for Psychologists" (full Shimokawa 2010 figures there).

When self-help has stopped working: signs to escalate

The most common self-help mistake is to continue after the point where escalation is methodologically due. A few markers of "self-help is not working in this case".

• Marker 1 — RCI not reached in 8 weeks. If PHQ-9 / GAD-7 has not dropped by ≥ RCI across 8 weeks of regular structured self-help — empirical evidence that this approach is not working. Doesn't mean "therapy doesn't work in general", but means a different intervention or a clinician is needed
• Marker 2 — suicidal thoughts have become active. Emergence of plan, means, intent — transition into red flag zone
• Marker 3 — functioning has started to break down. If at the start of self-help work and relationships were preserved, and after 4–6 weeks problems appeared — self-help is not coping
• Marker 4 — new symptoms have emerged. Especially: psychotic signs, mania / hypomania, marked substance use
• Marker 5 — feel-stuck pattern. Several weeks of the feeling "I'm doing everything right but nothing is changing" — independent of objective PHQ-9 numbers. This subjective stuckness is a standalone signal to revise the approach

Escalation does not equal "failure of self-help". It is a methodologically correct step: this intervention is not working in this zone, a different one is needed. Just as in medicine — if over-the-counter ibuprofen doesn't help with a headache for two weeks, you go to a doctor, you don't increase the dose on your own. This is the basic logic of stepped care (Gaynes/Rush 2009 STAR*D framework).

How to find a clinician if you decide to reach out

A short section on the practical next step if the decision tree points to "clinician needed".

• Type of clinician. For psychological symptoms without need for medication — a clinical psychologist or psychotherapist with verified qualifications. With suspected mania / hypomania, psychosis, or marked severity (PHQ-9 ≥ 20, GAD-7 ≥ 15) — psychiatrist, not just psychologist
• What to ask at the first meeting. What approach (CBT, ACT, schema therapy, EMDR for trauma); how the clinician tracks progress (questionnaires / outcome measures — marker of evidence-based practice); contract length (short 6–12 sessions vs long-term)
• What to watch for. The clinician's own professional supervision (see "Supervision for Psychologists"); whether structured outcome measures are used — a marker of MBC practice empirically tied to better outcomes (Shimokawa-Lambert 2010)
• Where to turn in a crisis. With active suicidal ideation — not "search for a therapist for a week", but call a crisis line or ER. Therapy work follows acute stabilization, not precedes it

Soveria is a measurement-based care platform with 42 validated instruments for clinicians. If a client already works with a therapist who uses Soveria, structured outcome monitoring is built into their shared workflow. If looking for a clinician from scratch — ask whether they use regular outcome measures. That is a structural marker of evidence-based practice. For depth: "Valid Psychological Tests Online" on distinguishing validated screening from quizzes, and "Types of Depression" on which instruments fit which clinical context.