Reading of the Week: Is ‘New’ Overrated? Antipsychotics in the Real World

From the Editor

Is new better?

You may be reading this on an iPhone 7, having driven to work this morning in a 2017 Hybrid Prius. So should your patients be taking a medication that became available four-and-a-half decades ago – when people drove gus-gusling 8-cylinder Oldsmobiles and smartphones didn’t even exist in science fiction novels.

This week, we look at a just-published JAMA Psychiatry paper which promises to look at the “real-world” effectiveness of antipsychotics. The authors tapped Swedish databases to consider outcomes for nearly thirty thousand people with schizophrenia.

Sweden: elaborate welfare state, beautiful historic buildings, and – yes – rich databases

Spoiler alert: the authors found that new wasn’t better. That is, newer antipsychotics tended to underperform clozapine and depot medications.

We also look at similar “real-world” work drawing from a Finnish database considering treatment of depression.



Antipsychotics and Outcomes

“Real-World Effectiveness of Antipsychotic Treatments in a Nationwide Cohort of 29 823 Patients With Schizophrenia”

Jari Tiihonen, Ellenor Mittendorfer-Rutz, Maila Majak, Juha Mehtälä, Fabian Hoti, Erik Jedenius, Dana Enkusson, Amy Leval, Jan Sermon, Antti Tanskanen, Heidi Taipale

JAMA Psychiatry, 7 June 2017 Online First


“The comparative effectiveness of antipsychotic treatments for patients with schizophrenia has remained controversial despite extensive research. Results from randomized clinical trials (RCTs) suggest that clozapine, olanzapine, and amisulpiride are superior to other antipsychotic medications in terms of efficacy. However, the most efficacious drugs such as clozapine and olanzapine frequently induce adverse effects, such as weight gain and dyslipidemia, which may result in severe deterioration of health after long-term treatment. Investigation of these adverse effects or associated outcomes such as hospitalization and death requires thousands of patients and several years of follow-up to achieve enough statistical power, which is not possible for RCTs.


“Another major issue in RCTs is the selection of patients. Those included in RCTs represent an atypical minority of the patient population because up to 80% to 90% of patients are excluded because of refusal, substance abuse, suicidal or antisocial behavior, or mental or physical comorbidity. Especially problematic is the comparison of oral antipsychotic medications vs long-acting injections of antipsychotic medications because patients with the poorest adherence (ie, those who would receive the greatest benefit from long-acting injectable antipsychotic medications) are excluded from RCTs because participation is fully voluntary. Because RCTs include only an atypical fraction of the most adherent patients, they do not provide information on the real-world effectiveness of the antipsychotic treatments.”

Jari Tiihonen

So opens a new paper by Tiihonen et al. In this paper, the authors make an observational study, drawing on national databases. This approach isn’t unique – the authors acknowledge past work shows better outcomes for clozapine, olanzapine and long-acting depot medications – but such work has had the problem of selection bias.

They attempt to address this:

“We aimed to overcome this problem by using within-individual analysis, in which each person is his or her own control. In this approach, the exposure periods of each individual are compared with the nonexposure periods of the same individual. Therefore, the only factors that need to be adjusted are those that change as a function of time, such as time since cohort entry, temporal order of exposure periods, and concomitant medications.”

Here’s what they did:

  • “We used nationwide register-based data to conduct a prospective population-based cohort study of patients with schizophrenia…”
  • Drawing on Swedish databases, they looked at people with a diagnosis of schizophrenia between July 1, 2006, to December 31, 2013. Inclusion criteria included “all individuals residing in Sweden who were 16 to 64 years of age in 2006.”
  • Data on medication use was drawn from the Prescribed Drug Register – which includes outpatient medication, though no inpatient prescriptions. (!)
  • They considered outcomes as follows: “psychiatric rehospitalization” and “treatment failure” (defined as rehospitalization, discontinuation or switch to other antipsychotic medication, or death).
  • They did a more complicated statistical analysis – that is, they used within-individual Cox proportional hazards regression model. “The within-individual model is a stratified Cox proportional hazards regression model in which each individual forms his or her own stratum.” They also looked at covariables.


Here’s what they found:

  • There were 29,823 patients.
  • Demographically: more men than women (12,822 women and 17,001 men). The mean age was 44.9.
  • “13,042 of 29,823 patients (43.7%) experienced psychiatric rehospitalization and 20,225 of 28,189 patients (71.7%) had treatment failure.”
  • In terms of drugs used: Oral olanzapine was the most frequently used drug, and zuclopenthixol the most frequently used as a long-acting injectable antipsychotic medication.”
  • In terms of rehospitalization: “The lowest risk of rehospitalization was observed for once-monthly long-acting injectable paliperidone (HR, 0.51), long-acting injectable zuclopenthixol (HR, 0.53), clozapine (HR, 0.53), long-acting injectable perphenazine (HR, 0.58), and long-acting injectable olanzapine (HR, 0.58).” See figure below.
  • In terms of treatment failure: “The lowest risk of treatment failure was observed for clozapine (HR, 0.58), and the second lowest was seen for all long-acting injectable antipsychotic medications (HRs, 0.65-0.80), whereas the highest risk was seen for levomepromazine (HR, 1.15).” See figure below.


Adjusted Hazard Ratios (HRs) and 95% CIs for Psychiatric Rehospitalization During Monotherapy Compared With No Use of Antipsychotic in Within-Individual Analyses in the Prevalent Population


Adjusted Hazard Ratios (HRs) and 95% CIs for Treatment Failure During Each Monotherapy Compared With Oral Olanzapine Use

“Our results from a large nationwide cohort show that clozapine and long-acting injectable antipsychotic medications are substantially more effective than other antipsychotics in reducing the risk of rehospitalization or any treatment failure. The most consistent findings were observed for clozapine, being the first in rank order in most of the analyses. These results are in line with those of previous cohort studies using traditional between-individual analyses, although the effect sizes differed to some extent, especially for comparisons between long-acting injectable antipsychotic medications and corresponding oral formulations. Our results showed that the risk of rehospitalization was 22% lower during treatment with long-acting injectable antipsychotic medications compared with treatment with equivalent oral formulations in the total cohort and 32% lower in the incident cohort of newly diagnosed patients.”


A few thoughts:

  1. This is a good study, drawing on a huge dataset – not dozens of people with schizophrenia, or even hundreds, but tens of thousands.


  1. The findings are strong. Let’s not mince our words: new isn’t necessarily better. Of the five best performing medications for treatment failure that were studied, four were old. And, yes, clozapine topped that list. Depot medications were very strong in terms of rehospitalizations, with robust results for drugs that pre-date Atari’s Pong (the first and only video game in 1972), like perphenazine; though it should be added that paliperidone (new) did the best.


  1. There are implications here for practice – how many newly diagnosed patients are on depot medications? There are also implications in terms of health systems – does funding and billing schedules reward depot choices over non-depot choices? Dare I ask about clozapine?!?


  1. Has similar work been done for depression? Actually, Lancet Psychiatry has a solid paper with a very similar analysis. That’s not exactly surprising since Tiihonen is the first author. (Wow, he’s having a good month, at least compared to Theresa May.) In this study, he and his co-authors tap Finnish databases. I will quickly summarize “Pharmacological treatments and risk of readmission to hospital for unipolar depression in Finland: a nationwide cohort study”: they looked at the risk of readmission for all patients who had at least one hospitalization for depression, with data from almost 125,000; exclusion criteria included schizophrenia and bipolar. They found: “Lithium use was associated with a lower risk of re-admission to hospital for mental illness than was no lithium use.” Yes, lithium – speaking of older medications, this one is as old as the earth.6
  1. In the accompanying editorial, Allan H. Young of King’s College is enthusiastic about the findings, but calls for more investigation into lithium:


“Replication of these findings is needed, and should be possible given that similar databases exist in other countries (e.g., Denmark and Taiwan). These data could be easily assessed to establish whether they replicate the Finnish findings or not. The findings of Tiihonen and colleagues are particularly noteworthy because of recent disquiet about the use of antidepressants in unipolar mood disorders, and they suggest that lithium monotherapy might be the best long-term prophylactic drug.”


  1. Big data is changing psychiatry.


Reading of the Week. Every week I pick articles and papers from the world of Psychiatry. 

Dr. David Gratzer



Reading of the Week: Better Treatment, Safer Roads? The New JAMA Psychiatric Paper on ADHD & Driving

From the Editor

How can we reduce the number of car accidents?

We often speak about treating mental illness in terms of reducing personal suffering. Recent selections have looked at the economic cost of mental illness. But what are the implications to public health?

This week, we look at a new JAMA Psychiatry paper; this national cohort study involved more than 2.3 million people with ADHD, and considered motor vehicle crashes (as measured by emergency department visits) and whether or not they were taking medications.

Yes, he has a plaid shirt, but should he be taking his prescription meds?

Spoiler alert: The authors find “medication use for the disorder was associated with a significantly reduced risk” of car accidents.

We also look at an editorial that finds “clinical pearls” in this paper.


Driving and ADHD

 “Association Between Medication Use for Attention-Deficit/Hyperactivity Disorder and Risk of Motor Vehicle Crashes”

Zheng Chang, Patrick D. Quinn, Kwan Hur, Robert D. Gibbons, Arvid Sjolander, Henrik Larsson, Brian M. D’Onofrio

JAMA Psychiatry, 10 May 2017 Online First


“Approximately 1.25 million people die each year globally as a result of motor vehicle crashes (MVCs). In the United States, more than 33 700 individuals died from MVCs in 2014 alone, with an additional 2.4 million visiting the emergency department as a result. In addition, MVCs are a major cause of the gap in life expectancy between the United States and other high-income countries.

“Attention-deficit/hyperactivity disorder (ADHD) is a prevalent neurodevelopmental disorder comprising symptoms that include poor sustained attention, impaired impulse control, and hyperactivity. The disorder affects 5% to 7% of children and adolescents and persists into adulthood in a substantial proportion of affected individuals. Previous studies have demonstrated that individuals with ADHD are more likely to experience MVCs. However, the magnitude of this association has varied substantially because of differences in outcome measures, sample selection, and confounding adjustment.

“Pharmacotherapy is considered the first-line treatment for ADHD in many countries, and rates of ADHD medication prescription have increased significantly during the last decade in the United States and other countries. Evidence from controlled trials has shown that pharmacotherapy has marked beneficial effects on core symptoms of ADHD; to some extent, it also improves driving performance in virtual reality driving simulators. The use of population-based health record data and self-controlled designs provides an innovative and informative approach to evaluate the effect of medication use on important outcomes in real-world situations. A Swedish register-based study found that ADHD medication use was associated with lower risk of traffic crashes in men. However, the association in women was not clear. Moreover, there are cross-national differences in ADHD treatment practices and rates of MVCs between Sweden and the United States. In addition, it is unclear whether ADHD medication treatment will change the long-term course of the patients and lower the risk of MVCs. Therefore, additional population-based studies in the United States are needed to evaluate the effect of ADHD medication use on MVCs.

“In the present study, we followed up a national cohort of patients with ADHD between January 1, 2005, and December 31, 2014, using data from commercial health care claims in the United States.”

Zheng Chang

 Here’s what they did:

  • The data was drawn from the Truven Health Analytics MarketScan Commercial Claims and Encounters databases – “one of the largest collections of deidentified patient data and includes inpatient, outpatient, and filled prescription claims for more than 100 insurers in the United States.” This covers 146 million people.
  • The study period was January 1, 2005, to December 31, 2014.
  • Patients were 18 and older, and had received an ADHD diagnosis (on an inpatient or outpatient basis) or a filled ADHD medication (the list of medications included amphetamine salt combination and methamphetamine hydrochloride).
  • The outcome event was an ED visit for an MVC, defined by the appropriate code (ICD-9 codes E810-E825).
  • Statistical analysis was done to find the risk of at least one MVC between patients with ADHD and controls. Additionally, to understand the association between medication use for those with ADHD and MVCs, the authors made “a monthly person-time data set” that considered the use of medications (if they filled a prescription in that month or there was a carryover prescription from a prior month). They also looked at 2-year follow ups.


Here’s what they found:

  • “The study cohort consisted of 2,319,450 patients with ADHD… observed for a total of 50,667,665 person-months.”
  • Demographically: the median age was 32.5, with a roughly equal gender distribution (1,121,053 men and 1,198,397 women). 83.9% (1,946,198) of those with ADHD had received at least one prescription for an ADHD medication.
  • “Patients with ADHD had a significantly higher risk of an MVC than their matched controls (OR, 1.49) and untreated patients with ADHD had the highest risk of an MVC compared with medicated patients with ADHD and controls…”
  • “At the population level, months with ADHD medication were associated with a 12% (OR,0.88) lower risk of MVCs in male patients with ADHD relative to unmedicated months and a 14% (OR, 0.86) lower risk of MVCs in female patients with ADHD… More important, the within-individual analyses showed that men with ADHD were 38% (OR, 0.62) less likely to have MVC events during medicated months relative to unmedicated months, suggesting that, within an individual (i.e., after controlling for all unmeasured static and measured time varying confounding factors), ADHD medication use was associated with a significant reduction in the risk of MVCs. Our PAF estimated that 22.2% of the MVCs among male patients with ADHD were attributable to lack of medication treatment, assuming that the association was causal.” The results were similar for female patients.
  • “At the population level, there were no significant associations between ADHD medication use and MVC events 2 years later. However, the within-individual analyses showed that ADHD medication use was associated with a 34% (OR, 0.66) lower risk ofMVCs 2 years later inmate patients with ADHD and a 27% (OR, 0.73) lower risk of MVCs in female patients with ADHD.”


“In this large, nationwide cohort study over 10 years, patients with ADHD had a higher risk of MVCs compared with controls without ADHD. However, in male and female patients with ADHD, medication use for the disorder was associated with a significantly reduced risk of MVCs. Similar reductions were found across all age groups, across multiple sensitivity analyses, and when considering the long-term association between ADHD medication use and MVCs.”

 The paper runs with a short and readable editorial by the University of Virginia Health System’s Vishal Madaan and Daniel J. Cox.

You can find the editorial here:

 Vishal Madaan

 The Editorial opens:

 “While driving is a ubiquitous functionality and an important activity of independent daily living, it also represents a complex neurobehavioral task involving an interplay of cognitive, motor, perceptual, and visuospatial skills. As a result, patients with neurodevelopmental disorders often have limitations in such skills. Although there has been a recent interest in understanding driving concerns in individuals with other neurdevelopmental disorders such as autism spectrum disorders, substantial research has reviewed the influence of attention-deficit/hyperactivity disorder (ADHD) on driving safety, especially given how pharmacotherapy may affect inattention, impulsivity, and executive dysfunction.”

 They praise the study, noting that: “The findings in the study by Chang et al in this issue of JAMA Psychiatry confirm and extend existing experimental studies and have impressive implications for judicious use of ADHD medication.”

 Madaan and Cox see “clinical pearls” for clinicians. The “management of ADHD is not limited to one’s school, college, or workplace: it extends to several other aspects of life, such as driving, which may be ignored to the clinician’s and patient’s peril.” They also comment that “health care professionals should be aware that MVCs in individuals with ADHD often happen later in the evening when their medications may have worn off.”

 A few thoughts:

  1. This is a good study.
  2. The topic is important.
  3. The findings are a gentle reminder of the importance of our work – literally keeping our patients safe. “The within-individual analyses showed that men with ADHD were 38% (OR, 0.62) less likely to have MVC events during medicated months relative to unmedicated months…” Wow.
  4. The paper isn’t without limitations, of course. The authors make some big assumptions: they equate medication compliance with pharmacy compliance (that is, if the prescription is filled, the patient is taking the medication); they only look to ED visits (post-MVC, a patient may seek care with his or her primary care physician); medication compliance doesn’t mean proper medication management (as Madaan and Cox note in their editorial, many stimulants have shorter half-lives, leaving patients under-medicated during evening and night-time driving). Still, it’s difficult not to be impressed with the incredible database that Chang et al. have drawn from – involving millions of people.
  5. If public health and psychiatry catches your fancy, the American Economic Review has an interesting paper on using CBT for crime reduction in Liberian men. Spoiler alert: it worked.


You can find that study here:

(Nice surprise: the paper on CBT and its public health implications ran in an economics journal.)

  1. Mental illness casts a long shadow over our society. Yes, we can see this in terms of absenteeism and presenteeism. Yes, we can see this in terms of personal tragedy and loss. But, as the authors of this paper argue, yes, we can see this in terms of car accidents and general safety.

Reading of the Week. Every week I pick articles and papers from the world of Psychiatry. 

-Dr. David Gratzer