“More Treatment but No Less Depression: The Treatment-Prevalence Paradox”, Johan Ormel, Steven D. Hollon, Ronald C. Kessler, Pim Cuijpers, Scott M. Monroe2021-12-11 (depression, scientific bias; similar)⁠:

The puzzling paradox of more treatment but no less depression requires answers.

First incidence has probably not increased and offset treatment-driven prevalence drops.

The published trial literature substantially overestimates efficacy of treatments.

In addition, treatment-quality gaps in routine care reduce effectiveness further.

Long-term outcome, under-treatment of recurrence, iatrogenicity need further study.

Treatments for depression have improved, and their availability has markedly increased since the 1980s. Mysteriously the general population prevalence of depression has not decreased. This treatment-prevalence paradox (TPP) raises fundamental questions about the diagnosis and treatment of depression.

We propose and evaluate 7 explanations for the TPP. First, 2 explanations assume that improved and more widely available treatments have reduced prevalence, but that the reduction has been offset by an increase in:

misdiagnosing distress as depression, yielding more “false positive” diagnoses; or

an actual increase in depression incidence.

Second, the remaining 5 explanations assume prevalence has not decreased, but suggest that:

treatments are less efficacious and

less enduring than the literature suggests;

RCT trial efficacy doesn’t generalize to real-world settings;

population-level treatment impact differs for chronic-recurrent versus non-recurrent cases; and

treatments have some iatrogenic consequences.

Any of these 7 explanations could undermine treatment impact on prevalence, thereby helping to explain the TPP.

Our analysis reveals that there is little evidence that incidence or prevalence have increased as a result of error or fact (Explanations 1 & 2), and strong evidence that (a) the published literature overestimates short-term and long-term treatment efficacy, (b) treatments are considerably less effective as deployed in “real world” settings, and (c) treatment impact differs substantially for chronic-recurrent cases relative to non-recurrent cases.

Collectively, these 4 explanations likely account for most of the TPP. Lastly, little research exists on iatrogenic effects of current treatments (Explanation 7), but further exploration is critical.

[Keywords: depression, treatment, prevalence, more treatment but not less depression, explanations treatment-prevalence paradox]

Possible Explanations for the TPP Evidence and (Preliminary) Conclusions

Have prevalence estimates been spuriously inflated due to increasing societal recognition of depression and associated diagnostic practices?

People have probably become more willing to admit depressive symptoms and to present for treatment, where they may receive false-positive diagnoses of MDD.
But since epidemiologic surveys are conducted by well-trained interviewers using structured interviews to generate well-standardized diagnoses, it is unlikely that systematic drift at the population level of ‘caseness’ has occurred. Thus, an increase in “false positives” diagnoses would not mask a treatment-driven drop in “true” epidemiological prevalence.

Have first incidence rates increased and offset a “true” treatment-driven reduction in point-prevalence?

Post-1980 first incidence studies and information on trends in causal risk factors are few, and too inconsistent to provide a conclusive answer. The handful of incidence studies, though, do not hint at any substantial rise in incidence since the 1980’s, and some even suggest a decrease. The evidence is sparse and uncertain, though, and ends around 2010.
It seems unlikely that an true increase in depression incidence offsets any treatment-driven prevalence reduction.

Do RCTs overestimate Acute-Phase treatment efficacy? Might biases both within the trials and across the larger literature on medication and psychotherapy inflate these short-term benefits of treatment? | R

RCTs do yield inflated acute-phase efficacy estimates. Adjusted for bias, efficacy drops by a third to half to-modest effect-sizes at best (about 0.30 for medications vs. pill-placebo and psychotherapy vs. care-as-usual). It is unclear how long Acute-Phase treatment benefits persist. Given the biases and large heterogeneity, it is not surprising that there is substantial disagreement about the clinical impact of treatments. It is not that treatments do not work, just that they do not work as well as the published literature suggests, or as is widely believed.
Hence, the a more accurate estimate of short-term efficacy is at best modest, which can explain in part the TPP.

Does research on maintenance of treatment gains and long-term efficacy over estimate beneficial effects? Are medication and psychotherapy interventions to prevent relapse-recurrence upwardly biased due to non-eligibility, insufficient response to Acute-Phase treatment, symptom return risk, and a variety of biases that need to be taken into account?

RCTs evaluating treatments aimed to reduce relapse-recurrence risk show substantial efficacy for preventive psychotherapy and for continued medication. However, these “effects” are rife with possible biases (misclassification, unblinding, allegiance effects, and differential mortality) complicating interpretation. In addition, many patients without sufficient response to acute-phase treatment are not eligible for relapse or recurrence prevention trials, and relapse-recurrence rates over 2 years after preventive treatment remain substantial, (though estimates vary greatly).
Hence, limited overall long-term efficacy also may help to explain the TPP.

Do RCTs generalize to real-world settings? How large is the gap between RCT-based efficacy and real-world effectiveness?

RCT-based efficacy does not generalize all that well to real-world practice, both for medication and for psychotherapy. Reasons: Large gaps in treatment choice and implementation quality exist in real-world practice; compared to the typical RCT patient, the real-world patient is somewhat less treatable (suicidal ideation; addiction, severe comorbidities). What the gaps, along with naturalistic follow-up studies, tell us is that treatment is not as effective long-term as we would like it to be.
This explanation appears to be one of the strongest candidates for understanding the TPP as it also amplifies the contribution of explanations 3 and 4.

Does treatment efficacy vary by different subtypes of depression? Specifically, could differential treatment benefits for chronic-recurrent versus non-recurrent cases dilute the potential beneficial effects of treatments for those most in clinical need? Further, chronic-recurrent cases are often very difficult to treat, or treatment-resistant.

The majority of people who initially become depressed have few if any recurrences, whereas recurrent and chronic cases become or remain depressed for much more time over the course of their lives. The availability of more and better treatments consequently has many more opportunities to benefit the smaller number of chronic-recurrent cases, while treatment effects at the population level for the many more non-recurrent cases most likely will be very limited. The resulting limited effects at the population level for the larger non-recurrent group could dilute more pronounced effects for the chronic-recurrent subgroup, obscuring a positive impact on prevalence for those in greatest need. However, it is unclear to what extent advances in preventive treatments specifically benefit the chronic-recurrent subgroup, or if these treatments are adequately transported into routine care for them (Explanations 3–5). Individually and combined, these subgroup considerations also provide potentially strong explanations for the TPP.

Can treatment sometimes also have counterproductive consequences?

Oppositional perturbation refers to a medication-induced state of built-up perturbation in homeostatic monoamine regulatory mechanisms that “bounces back” when medication is discontinued, and then overshoots the normal balance of monoamine storage and release, increasing the risk for symptom return compared to spontaneous remission. Loss of agency refers to the hypothesis that either medication or psychotherapy could be counterproductive if either or both reduce self-help activity and active coping and thereby interfere with natural recovery mechanisms. Although some indirect evidence exists for each possibility, both mechanisms are largely speculative.
The explanatory potential of this concern remains to be demonstrated for understanding the TPP, but is worthy of further investigation.

Similar Links:

Possible Explanations for the TPP	Evidence and (Preliminary) Conclusions
Have prevalence estimates been spuriously inflated due to increasing societal recognition of depression and associated diagnostic practices?	People have probably become more willing to admit depressive symptoms and to present for treatment, where they may receive false-positive diagnoses of MDD. But since epidemiologic surveys are conducted by well-trained interviewers using structured interviews to generate well-standardized diagnoses, it is unlikely that systematic drift at the population level of ‘caseness’ has occurred. Thus, an increase in “false positives” diagnoses would not mask a treatment-driven drop in “true” epidemiological prevalence.
Have first incidence rates increased and offset a “true” treatment-driven reduction in point-prevalence?	Post-1980 first incidence studies and information on trends in causal risk factors are few, and too inconsistent to provide a conclusive answer. The handful of incidence studies, though, do not hint at any substantial rise in incidence since the 1980’s, and some even suggest a decrease. The evidence is sparse and uncertain, though, and ends around 2010. It seems unlikely that an true increase in depression incidence offsets any treatment-driven prevalence reduction.
Do RCTs overestimate Acute-Phase treatment efficacy? Might biases both within the trials and across the larger literature on medication and psychotherapy inflate these short-term benefits of treatment? \| R	RCTs do yield inflated acute-phase efficacy estimates. Adjusted for bias, efficacy drops by a third to half to-modest effect-sizes at best (about 0.30 for medications vs. pill-placebo and psychotherapy vs. care-as-usual). It is unclear how long Acute-Phase treatment benefits persist. Given the biases and large heterogeneity, it is not surprising that there is substantial disagreement about the clinical impact of treatments. It is not that treatments do not work, just that they do not work as well as the published literature suggests, or as is widely believed. Hence, the a more accurate estimate of short-term efficacy is at best modest, which can explain in part the TPP.
Does research on maintenance of treatment gains and long-term efficacy over estimate beneficial effects? Are medication and psychotherapy interventions to prevent relapse-recurrence upwardly biased due to non-eligibility, insufficient response to Acute-Phase treatment, symptom return risk, and a variety of biases that need to be taken into account?	RCTs evaluating treatments aimed to reduce relapse-recurrence risk show substantial efficacy for preventive psychotherapy and for continued medication. However, these “effects” are rife with possible biases (misclassification, unblinding, allegiance effects, and differential mortality) complicating interpretation. In addition, many patients without sufficient response to acute-phase treatment are not eligible for relapse or recurrence prevention trials, and relapse-recurrence rates over 2 years after preventive treatment remain substantial, (though estimates vary greatly). Hence, limited overall long-term efficacy also may help to explain the TPP.
Do RCTs generalize to real-world settings? How large is the gap between RCT-based efficacy and real-world effectiveness?	RCT-based efficacy does not generalize all that well to real-world practice, both for medication and for psychotherapy. Reasons: Large gaps in treatment choice and implementation quality exist in real-world practice; compared to the typical RCT patient, the real-world patient is somewhat less treatable (suicidal ideation; addiction, severe comorbidities). What the gaps, along with naturalistic follow-up studies, tell us is that treatment is not as effective long-term as we would like it to be. This explanation appears to be one of the strongest candidates for understanding the TPP as it also amplifies the contribution of explanations 3 and 4.
Does treatment efficacy vary by different subtypes of depression? Specifically, could differential treatment benefits for chronic-recurrent versus non-recurrent cases dilute the potential beneficial effects of treatments for those most in clinical need? Further, chronic-recurrent cases are often very difficult to treat, or treatment-resistant.	The majority of people who initially become depressed have few if any recurrences, whereas recurrent and chronic cases become or remain depressed for much more time over the course of their lives. The availability of more and better treatments consequently has many more opportunities to benefit the smaller number of chronic-recurrent cases, while treatment effects at the population level for the many more non-recurrent cases most likely will be very limited. The resulting limited effects at the population level for the larger non-recurrent group could dilute more pronounced effects for the chronic-recurrent subgroup, obscuring a positive impact on prevalence for those in greatest need. However, it is unclear to what extent advances in preventive treatments specifically benefit the chronic-recurrent subgroup, or if these treatments are adequately transported into routine care for them (Explanations 3–5). Individually and combined, these subgroup considerations also provide potentially strong explanations for the TPP.
Can treatment sometimes also have counterproductive consequences?	Oppositional perturbation refers to a medication-induced state of built-up perturbation in homeostatic monoamine regulatory mechanisms that “bounces back” when medication is discontinued, and then overshoots the normal balance of monoamine storage and release, increasing the risk for symptom return compared to spontaneous remission. Loss of agency refers to the hypothesis that either medication or psychotherapy could be counterproductive if either or both reduce self-help activity and active coping and thereby interfere with natural recovery mechanisms. Although some indirect evidence exists for each possibility, both mechanisms are largely speculative. The explanatory potential of this concern remains to be demonstrated for understanding the TPP, but is worthy of further investigation.