Comments on AHRQ Technical Brief “Prevention, Diagnosis, and Management of Opioids, Opioid Misuse and Opioid Use Disorder in Older Adults"

On May 15, 2020, the Agency for Healthcare Research and Quality released a voluminous draft Technical Brief titled "Prevention, Diagnosis, and Management of Opioids, Opioid Misuse and Opioid Use Disorder in Older Adults" The Brief may be downloaded here:

https://effectivehealthcare.ahrq.gov/products/opioids-older-adults/draft-report

The brief is open for comments until Friday June 12, which must be submitted in an AHRQ form that isn't visible to the public. The AHRQ draft has not been submitted to the Federal Register. You may instead offer comments here:

https://effectivehealthcare.ahrq.gov/webform/products/opioids-older-adults/research-comments

I have not completed my own comments yet. However, I am releasing this interim "drop" on issues that I have detected so far. Others may also comment from their own lived experience if they choose. Please do not clip and drop from the text below. This is released as general guidance only on the substance of concerns that you may wish to address if you have time in the next week.

In broad general terms, I find the Technical Brief to be a little shop of horrors. It is fundamentally wrong on internal process, wrong on science, and wrong on medical ethics. It incorporates multiple medical mythologies, central among them the silly notion that opioids prescribed to seniors by physicians require "interventions" to ensure that patients are denied such assistance if "risks exceed benefits". However, it offers no useful demonstration that any presently available patient profiling instrument or protocol can usefully identify those who are presumably at risk. 

After acknowledging that there is relatively little published research on outcomes for either prescriptions or illegally obtained opioids in older adults, the document proceeds to selectively cherry pick results from the few available reports, and to combine these reports with un-referenced and unsupported assertions concerning dangers of prescribing and an asserted lack of evidence for "effectiveness" for opioids used over long periods. This is the same conceptual error published in the 2016 CDC Guidelines, and it is in my view disqualifying of the entire document. 

For my comments thus far, organized by sections in the AHRQ comments form, see below. 

=========

1.   Comments on the Evidence Summary of the draft report

From the Summary as drafted:  “We developed a Conceptual Framework outlining the stages of care for older adults who require or use opioids, and factors that have an impact on management decisions and patient outcomes (see Figure). The framework prioritizes three potential targets to determine factors associated with and interventions for: 1) reducing opioid prescriptions where harms outweigh benefits, 2) preventing opioid misuse and opioid use disorder (OUD), and 3) reducing other opioid-related harms.”

My comment: although the conceptual framework identifies many factors associated with long-term opioid use in all patient populations, it offers no useful evidence beyond outright surmise concerning reliable measures which justify restriction of opioid prescriptions on a basis of risk. We know from multiple published sources including the US CDC that there is no actual correlation between rates of opioid prescribing in any patient population versus risk of overdose-related mortality on a US State-by-State basis (Ref 1). We also know that risk of overdose-related mortality in persons age 62 and older is the lowest of any age cohort, while their rate of opioid prescribing is highest. (ibid)  This demographic inversion cannot be explained as an outgrowth of any factor directly relating to prescribing of opioid pain relievers to legitimate pain patients. 

Ref 1: Richard A Lawhern, PhD “Stop Persecuting Doctors for Legitimately Prescribing Opioids for Chronic Pain”, STAT News, June 28, 2019, https://www.statnews.com/2019/06/28/stop-persecuting-doctors-legitimately-prescribing-opioids-chronic-pain/

It is likewise unequivocally clear that rates of diagnosis for substance use disorder in otherwise opioid-na?ve post-surgical patients of all ages who are treated with opioid analgesics are on the order of a maximum of 0.6% (Ref 2). Rates of protracted opioid prescribing (prescriptions renewed longer than 120 days) are less than 1% for at least 11 common surgical procedures; these rates in all likelihood represent failure of surgical procedures and emergence of chronic pain, rather than any drug-seeking behavior on the part of patients themselves.  Rates of protracted prescribing in non-surgical patients are on the order of 0.136%. (Ref 3)

Ref 2: Gabriel A Brat, Denis Agniel, Andrew Beam, Brian Yorkgitis, Mark Bicket, Mark Homer, Kathe P Fox,  Daniel B Knecht,  Cheryl N McMahill-Walraven, Nathan Palmer, Isaac Kohane, “Postsurgical prescriptions for opioid naive patients and association with overdose and misuse: retrospective cohort study”, BMJ 2018;360:j5790 https://www.bmj.com/content/360/bmj.j5790.long 

Ref 3: Eric C. Sun, Beth D. Darnall, Laurence C. Baker, Sean Mackey, “Incidence of and Risk Factors for Chronic Opioid Use Among Opioid-Naive Patients in the Postoperative Period”, JAMA Internal Medicine 2016;176(9):1286-1293. https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/2532789

The combination of these data should inform us that quantifiable risk of addiction or mortality in patients of any age who are diagnosed and treated for chronic pain with prescription opioid analgesics, is very low.  In fact, it is so low that there is no possibility of usefully predicting individual patient risk. 

This fundamental principle is supported by no less an authority than Nora Volkow, MD, Director of the National Institutes on Drug Abuse:

“Unlike tolerance and physical dependence, addiction is not a predictable result of opioid prescribing. Addiction occurs in only a small percentage of persons who are exposed to opioids — even among those with pre-existing vulnerabilities...”  “Older medical texts and several versions of the Diagnostic and Statistical Manual of Mental Disorders (DSM) either overemphasized the role of tolerance and physical dependence in the definition of addiction or equated these processes (DSM-III and DSM-IV). However, more recent studies have shown that the molecular mechanisms underlying addiction are distinct from those responsible for tolerance and physical dependence, in that they evolve much more slowly, last much longer, and disrupt multiple brain processes.”

       Ref 4: Nora D Volkow, MD, and Thomas A McLellan, Ph.D., “Opioid Abuse in Chronic Pain — Misconceptions and Mitigation Strategies” . NEMJ 2016; 374:1253-1263 March 31, 2016]. https://www.nejm.org/doi/full/10.1056/NEJMra1507771

Also from the summary as drafted:

“35 studies assessed factors independently associated with opioid-related outcomes among older adults (≥60 years).

• While the 35 studies reported multivariable analyses, none of the analyzed models was designed or evaluated as a screening or prediction tool.
• 17 multivariable studies evaluated long-term opioid use, which may sometimes be a high-risk behavior, but is not necessarily evidence of problematic opioid use.
• All 8 studies that looked at prior or early postoperative opioid use found mostly strong associations (e.g., relative risk [RR] >2.0) with long-term opioid use.
• All 6 studies that examined greater amounts of prescribed opioids (higher number of opioid prescriptions or higher opioid dose) found mostly strong associations with long-term opioid use.”

My Comment: The assertion of long-term opioid use as a high-risk behavior is unsupported by any reference and is in fact wrong on the evidence. We should expect strong associations between prior post-operative opioid use and long term use. Likewise, we should expect associations between higher opioid dose levels and long term use. Neither is an issue of risk per se.

The typical course of progression for many underlying medical disorders and the typical course of dose titration is toward higher levels over time, to address more intense, intractable or multi-factorial pain.  We must also keep in mind the relatively tiny size of the population in which these effects have been documented. It is highly inappropriate to extrapolate rules of all medical practice or to infer “risk” in the general population of patients, based on this relatively small patient cohort (under 1%).  

Also from the summary as drafted:

“14 studies addressed interventions related to opioid use and opioid-related disorders in older adults.

• Only 1 study was a randomized trial. Each intervention was evaluated by only 1, or rarely, 2 studies.
• The most-studied interventions were screening tools to predict opioid-related harms but none of these tools has been tested in clinical practice to assess real-world results.
• 2 studies found that prescription drug monitoring programs have been associated with less opioid use (at the State level).”

My Comments: While it is true that no presently available screening tool has been tested in clinical practice, other AHRQ reports (Ref 5) go further to report that no two available screening tools produce mutually consistent results even in non-clinical settings. Very illuminating from this source is the realization that genomic testing also provides no reliable risk predictions for tolerance, addiction or mortality in individual patients (ibid).  Nor should we expect such predictions.

Ref 5: US Agency for Healthcare Research and Quality, “Opioid Treatments for Chronic Pain” - Draft Comparative Effectiveness Review, circulated October 2019 for public comment, pp 202-204.

           Key Messages: “No instrument has been shown to be associated with high accuracy for predicting opioid overdose, addiction, abuse, or misuse “

Even more concerning in the present report is that it makes no reference to genomic testing or opioid metabolism even as marginal concerns in modeling relationships between opioid exposure versus addiction. This omission appears to reflect a consistent and unreported anti-opioid bias among the (unnamed) writers group that produced the report. If the writers had acknowledged the existence of an extensive medical literature on genetically mediated polymorphism in P450 enzymes which govern opioid metabolism, they might also have been forced to assess whether metabolism is a significant confounding factor in any “one size fits all” criterion that might be proposed – as the 2016 CDC Guidelines did – as a basis for evaluating risks versus benefits. 

It is unsurprising that existence of PDMPs at State level is associated with reduced opioid prescribing. These databases are actively being exploited by State Medical Boards and regional prosecutors as a means of suppressing opioid prescribing by singling out and warning doctors who prescribe most often. Thus the tangential reference in this report can only be construed as a deliberate confusing of cause and effect to support a preexisting anti-opioid agenda.

Conceptual Framework (Figure)

From the Draft Document:

“We formed a 15-member panel comprising six individuals employed by federal agencies and nine individuals employed by nonfederal entities. These individuals included experts in the care of older adults, experts in pain treatment and opioid use, nationally and internationally recognized researchers, policy makers, and internationally recognized advocates for older adults with pain. We had discussions with the 15 Key Informants to help us revise the conceptual frame-work. We solicited the panel’s input in three teleconferences and over email until we deemed that we had sufficiently discussed all of the most relevant themes. The interactions with the Key Informant Panel were facilitated by the EPC and included several structured prompts based on all Guiding Questions. The Key informants were asked about the draft Conceptual Framework and to identify peer-reviewed publications or other relevant literature related to the topics of interest. In Appendix C we provide an overview of our discussions with Key Informants that helped to shape the Conceptual Framework and to evaluate the evidence base. Appendix C also includes specific themes identified during the discussion.”

My Comments: 

I find it very difficult to believe that the Key Informants group included any ”recognized advocates for older adults in pain.“ The draft report offers no evidence of even considering the lived experience of the 30 million US residents whose lives are constrained by daily severe pain which impacts quality of life and function. 

The conceptual framework proposed in this Technical Note incorporates a fundamentally faulty assumption unsupported by medical evidence: that interventions are needed to “reduce risk of opioid prescribing where harms outweigh benefits and increase access to non-opioid treatment.”   

As noted above, it is not at all clear either that (a) interventions involving restrictions on prescription opioid availability are actually needed, or (b) that non-opioid treatments are available and effective as replacements for opioids. A compelling case can be made for the finding that our US “opioid crisis” is not and never was the result of physicians prescribing to their patients -- and is not being sustained from that source. Indeed, the conceptual framework as introduced pointedly ignores the existence of multiple socio-economic factors which are in fact directly pertinent.  

The economic crisis of 2008 contributed directly to ongoing structural unemployment and the hollowing-out of small communities across the US rust belt, Appalachia, the deep south and rural west. Job loss from automation and international out-sourcing has caused family breakups in all of these areas, placing people of all ages – including some seniors – under enormous situational pressure. Street drugs become an attractive distraction from such factors. This phenomenon has been discussed under the designation “a Crisis of Despair.”

Even granting that recent years have seen somewhat increased rates of opioid related mortality in seniors, the increases seen have been small in absolute terms and in comparison to mortality in youth, young adults and people of middle age. The Figure below is a plot of mortality data by age cohort over a period of nearly 20 years [Ref 7]. It is based on direct data downloads from the CDC-Wonder database.

[See Figure 3 in Reference 7]

We should also note that behavioral and non-invasive therapies may play constructive roles as adjuncts to opioids, NSAIDs, off-label use of anti-seizure and anti-depressant medications. But there are simply no published trials of such therapies as replacements for opioids.  Moreover, medical evidence for nearly all “alternative” pain therapies is exceptionally weak.

          [Ref 6]   Richard A Lawhern PhD and Steven E Nadeau M.D., “Behind the AHRQ Report -- Understanding the limitations of “non-pharmacological, non-invasive” therapies for chronic pain.”. Practical Pain Management, V18 Issue 7, October 2018 .https://www.practicalpainmanagement.com/resources/practice-management/behind-ahrq-report

Original Publication of the AHRQ Report: “Non-Invasive Nonpharmacological Treatment for Chronic Pain”, June 11, 2018, https://effectivehealthcare.ahrq.gov/products/nonpharma-treatment-pain/research-2018

[Ref 7] Richard A Lawhern, PhD, “Over-Prescribing Did Not Create America’s Opioid Crisis”, Understanding Cronic Plain  - Online Blog of Lynn Webster, MD, April 6, 2019, https://www.lynnwebstermd.com/over-prescribing/

2.   Comments on the Introduction of the draft report.

From the Draft Document: 

“Opioid treatment is often indicated for older adults

Opioid medications are commonly used to treat pain; however their use at higher doses and concurrently with benzodiazepines is associated with increased risk of opioid-related harms, including overdoses.8-12 Furthermore, the effectiveness of long-term opioid therapy is unclear.13-15”

... and further along, “For example, combining opioids and benzodiazepines can result in respiratory depression and death.”

My Comments:

While there has been an historical bias against co-prescription of opioid analgesics and benzodiazepines, much of this bias has been grounded on postmortem blood toxicity screens. There are no published trials evaluating suppression of respiration in live patients administered these medications and then observed under controlled conditions.  Likewise, failure to evaluate and treat sleep interruption, depression and anxiety in chronic pain patients is associated with significantly elevated rates of medical collapse and suicide.

In large-cohort demographic studies, incidence of opioid overdose related mortality in patients prescribed opioids is on the order of 0.022% (22 deaths per hundred thousand).  In a definitive study of a full year of medical records for the entire State of North Carolina,  nearly 2 million people were found to have been prescribed both a benzodiazepine and an opioid.  But only 386 died of an overdose where both were implicated in the span of 1 year. Every overdose death is a heart break; but the numbers clearly speak to a general safe use of the meds together.   (Ref 8)

[Ref 8] Dasgupta N, Funk MJ, Proescholdbell S, et al. “Cohort study of the impact of high-dose opioid analgesics on overdose mortality”.[Erratum appears in Pain Med. 2016 Apr;17(4):797-8; PMID: 27025778]. Pain Med. 2016 Jan;17(1):85-98. PMID: 26333030

The assertion that effectiveness of long-term opioid therapy is “unclear” is strongly contradicted by literally hundreds of thousands of reports in social media, grounded upon patient lived experience. This assertion is largely an artifact of the rarity of long term double-blind trials due to patients dropping out of the placebo arms in such trials. The 2016 CDC guidelines on opioid prescribing in adults with chronic non-cancer pain deliberately conflated this rarity with lack of effectiveness. However had the same criteria been applied to behavioral and non-opioid analgesic trials, none of these therapeutic approaches would have been able to demonstrate strong evidence of effectiveness. (Ref 9)

[Ref 9] Baraa O. Tayeb, Ana E. Barreiro, Ylsabyth S Bradshaw, Kenneth K H Chui, Daniel B Carr, “Durations of Opioid, Nonopioid Drug, and Behavioral Clinical Trials for Chronic Pain: Adequate or Inadequate?” Pain Medicine, Volume 17, Issue 11, 1 November 2016, Pages 2036–2046.

https://academic.oup.com/painmedicine/article/17/11/2036/2447887

From the Draft Docment:

“It is plausible that many older adults misuse prescribed opioids by taking them in greater amounts, more often, or for longer than they were directed to by a prescriber, or even resort to illicit opioids to alleviate untreated or undertreated pain, increasing the risk of overdose.46”

My Comment:

It is indeed “plausible”, but the logic of this passage is profoundly outrageous: are we to presume that patients are at fault for the unwillingness of their physicians to prescribe and manage adequate pain control? And if they are at fault, are clinicians justified in refusing them adequate treatment? The withholding of treatment for pain when it is available is widely considered to be a violation of human rights.

From the Draft Document:

“As with younger individuals, opioid misuse may transition to OUD. Regardless of age, individuals may become physically dependent on opioids (i.e., the body adjusts its normal functioning around regular opioid use) and continue taking them to avoid uncomfortable withdrawal symptoms.48, 49 Long-term opioid use—use of opioids on most days for longer than 3 months— may predispose individuals to developing OUD; although, this connection has not been established in younger or older adults."

And later in the draft: 

“Opioid use in older adults may eventually result in opioid misuse or OUD, and a variety of factors may predict transition to misuse, OUD, or both (Octagon R2). Pharmaceutical, non-pharmaceutical (e.g., behavioral), nonmedical (e.g., educational, community-based), and other interventions could, at least conceptually, help older adults to safely use prescription opioids and prevent or reduce the risks of transition to opioid misuse and OUD (Triangle I2).”

My Comments:

If the connection has not been established between long term opioid use and OUD, then why are the writers at pains to introduce the idea at all? Is this not obvious evidence of disqualifying anti-opioid bias?

Likewise, the phrase “at least conceptually” reveals an uncritical and unsupported surmise that goes far beyond any real medical evidence. Reality on the ground, as summarized by Dr Nora Volkow (ibid Ref 4), is that prescription opioid misuse or addiction are not predictable outcomes of prescribing for any age group – and most certainly not for Seniors over age 62 who have the highest rates of prescription use and the lowest rates of overdose-related mortality. What is actually going on in this phrasing and in the construction of the conceptual model amounts to an exercise in hype and overstatement that is utterly unsupported by real evidence.  This deliberate distortion is by itself sufficient grounds for retraction of the Technical Note in total.   

3.   Comments on the Methods section of the draft report.

Although the process of developing a draft conceptual document is reviewed, the identities and professional positions of the Key Informants are withheld in the draft report. Thus is impossible for readers to assess the experience, biases, professional or financial self-interest of those who have shaped the methodology and conclusions of the report. This represents a fundamental and ethically unacceptable failure of public transparency, and should be grounds for rejection of the Technical Note out of hand.  

4.   Comments on the Findings section of the draft report.

From the Draft Document:

“While opioids are an option (Rectangle C), nonopioid medications could be used to manage pain. These medications include acetaminophen, nonsteroidal anti-inflammatory drugs (e.g., ibuprofen, naproxen), corticosteroids, antidepressants, antiepileptics, and others (e.g., topical capsaicin products). Nonpharmacological options are available as well and include a wide array of potential interventions, such as yoga, massage therapy, and acupuncture. Importantly, older adults may start “multimodal” treatment (of more than one intervention) that comprises a pain treatment approach that 1) combines medications from different pharmacologic classes and 2) combines pharmacologic and nonpharmacologic therapies or multiple nonpharmacologic therapies.”

My Comment:

As documented in a 2019 AHRQ systematic outcomes review, the state of medical literature on the effectiveness of non-pharmacological methods is abysmal. No trial for yoga, massage therapy, acupuncture or behavioral therapies has progressed beyond Phase II, and almost all published trials are at Phase I. Moreover, the strength of medical evidence for almost all of the so-called alternatives is assessed as weak, and the degree of improvement in pain levels or quality of life is strictly marginal and temporary [ibid Ref 6]. No trials have been published which directly compare opioid therapy with non-pharmacological techniques on an either-or basis. These documented outcomes give the lie to any assertion that Nonpharmacological alternatives are “available”.   They simply are not. 

Remaining Sections to Be Addressed:

5.   Comments on the Summary and Implications section of the draft report.

6.   Comments on the Appendixes of the draft report.

7.   General comments on the draft report.

8.   Does this report describe both the problem and the evidence in a way that you could understand?

9.   Did you find this report unnecessarily difficult to read?

10. Could you find and understand the results and conclusions?