Illness Anxiety Disorder: Psychotherapy-First or Medication-First?
Two patients with the same diagnosis and genuinely comparable trial evidence for either modality. The disagreement isn't about which treatment works — both do — it's about whether access and severity are ever reasons to abandon a philosophy of psychotherapy first.
A.T., a 29-year-old woman, married her college boyfriend fourteen months ago and still keeps their wedding photo as her phone's lock screen, which is how her husband first noticed how often she was checking her phone for something else entirely — searching her GI symptoms at two and three in the morning. It began with intermittent bloating after meals that a colonoscopy six months ago found nothing to explain, but the reassurance didn't take: she now reads every twinge as the colon cancer she is convinced the colonoscopy somehow missed, and has started declining dinner invitations rather than risk a symptom flaring in front of people. She still works full time as a middle-school teacher and hasn't missed a day, but her husband describes watching her disappear into her phone mid-conversation as soon as her stomach makes a sound.
Her insurance covers a CBT-specialized anxiety clinic twelve minutes from her apartment with a two-week wait for a first appointment, and she is, by her own account, motivated to try therapy specifically — she has watched a close friend do CBT for a different anxiety diagnosis and wants to try the same route before adding a medication to her routine. That preference matters clinically, not just personally: Greeven and colleagues' randomized trial in hypochondriasis found paroxetine and CBT produced statistically indistinguishable improvement, and Barsky and Ahern's earlier trial found CBT's benefit for hypochondriasis held at both six and twelve months against usual care. With real access, real motivation, and evidence that neither modality outperforms the other, nothing about her case argues for overriding what she is already asking for.
First visit, deciding where to start
Start with CBT alone. She has real, fast access to a clinic that specializes in exactly this, she's asking for it directly, and Greeven's trial found no efficacy gap between CBT and paroxetine in this diagnosis — Barsky and Ahern's earlier trial showed the same CBT benefit holding out to a year. There's no efficacy argument for adding a medication she hasn't asked for when the therapy she wants is two weeks away and has equivalent trial support.
I don't disagree that CBT alone is defensible here — the evidence really is that close. My only hesitation is two weeks is still two weeks of three-in-the-morning symptom searches, and starting an SSRI in parallel rather than sequentially costs her nothing while she waits, given the two drugs' equivalent efficacy profile.
But she told us directly she wants to try therapy on its own first, and overriding a clearly stated preference in a patient with real access and real motivation, purely because a faster combined start is theoretically available, treats her wish as an obstacle rather than as relevant clinical information. I'll defer to that.
Agreed: CBT alone, first appointment in two weeks, medication held entirely unless symptoms worsen before then or she asks to reconsider.
Not fully agreed: whether that decision would still be right if her wait had been eight weeks instead of two — the psychiatrist's parallel-start argument was accepted as reasonable in the abstract and set aside here specifically because the access gap was short enough not to matter.
W.D., a 58-year-old man, has lived on the same forty acres outside a small town for thirty years, ten miles from the nearest paved road and over ninety minutes from the regional medical center where his cardiologist works. His wife died of a sudden cardiac event three years ago, and since then he has become convinced that every episode of chest tightness — real, benign, musculoskeletal tightness his primary care physician has evaluated and cleared four separate times — is the same thing that killed her. He has missed eleven days of work this year checking himself into the regional ED, twice leaving mid-shift, and has started sleeping in a recliner because lying flat "feels like it's compressing something." The nearest CBT-trained therapist accepting his insurance has a four-month waiting list, and telehealth options are limited by broadband that drops out for hours at a time during storms.
The evidence that made A.T.'s choice a genuinely open one doesn't resolve his case the same way, because the choice it describes isn't actually available to him: Greeven's parity finding assumes both arms are reachable, and his realistically isn't for months. Hedman and colleagues' 2011 randomized trial of internet-delivered CBT for severe health anxiety offers a partial bridge — two-thirds of the treated group no longer met criteria for health anxiety at post-treatment, against an online discussion-forum control, with gains holding at six months — but it still depends on a connection stable enough to complete structured sessions, which his own home cannot currently guarantee. Eleven missed workdays and a recliner he now sleeps in every night are not the same starting point as A.T.'s two-week wait with full function intact.
Four months is the wait, not the plan
Start sertraline today. Whatever parity Greeven's trial found between CBT and medication, it assumed a patient could actually get either one — his four-month wait means the choice A.T. had isn't available to him, and eleven missed workdays this year say the cost of waiting is already real, not hypothetical.
I agree medication starts today — I'm not arguing to wait four months for anything. My concern is narrower: starting a drug can quietly become the entire plan once the acute crisis eases, and the access problem underneath it never gets solved. He needs an actual next step toward therapy, not just a prescription that makes the wait more bearable.
Medication treating his symptoms well enough that nobody circles back to the access gap in three months would be its own kind of failure — relief without ever addressing why the modality he might have preferred was never reachable in the first place.
There's a concrete way to hold both of your points at once. Hedman and colleagues' 2011 randomized trial put two-thirds of its internet-CBT group below criteria for severe health anxiety at post-treatment — and I'd note that was against an active discussion-forum control, not a waiting list, so it isn't just an attention effect. It's not a perfect substitute for his broadband situation, but sessions can be timed around clearer daytime connectivity and downloaded for offline review, rather than requiring live video the way some telehealth options do.
Sertraline starts today for the reason the psychiatrist gave. iCBT gets referred today too, not held until the medication either works or doesn't — that's the concrete step the psychologist is asking for, and it doesn't require the four-month in-person wait list at all.
Agreed: sertraline started today, iCBT referral placed the same day rather than held in reserve, and the in-person wait list kept active rather than cancelled.
Not agreed, and named explicitly rather than left to resolve itself:
The psychiatrist's worry is that iCBT engagement quietly lapses once he feels better — a three-month check-in was set specifically to ask about it, not assume it happened.
The four-month in-person wait becomes the real plan, not a fallback — medication alone was never accepted as sufficient on its own by any voice at the table.