antibiotic overprescribing, 2 more studies

Two recent articles came out documenting antibiotic overprescription, one from a large database of US claims data, the other on antibiotic prophylaxis in common urologic procedures. will do one today and one tomorrow.

The first study used massive data-mining to assess the appropriateness of antibiotic prescriptions in patients with private insurance (see antibiotic overprescribing bmj2019 in dropbox, or doi.org/10.1136/bmj.k5092)

Details:

--19.2 million enrollees in Truven MarketScan Commercial Claims and Encounters database (includes people aged 0-64 who have employer-based private health care), and includes a trove of info on many settings: primary care, specialty care, hospital outpatient, urgent care centers, retail clinics, ambulatory centers, emergency depts, home visits, community hospitals, and academic hospitals

    --this enrollee database includes 1/8^th of all who have employer-based coverage in this age range in the US

--they accessed ICD-10 coding for the specific diagnoses, as well as pharmacy data

--they meticulously reviewed 91,738 diagnostic codes in the 2016 version of ICD-10 and stratified them by the need for antibiotics (these were reviewed by one pediatrician and 2 general internists who were specialists in antibiotic overuse); they erred on the side of leniency, such as pneumonia being an antibiotic indication though many pneumonias are viral, or including acute otitis externa as “sometimes” needing antibiotics even though oral antibiotics are rarely needed [they published their diagnostic listings to elicit feedback and develop a more standard list. however, it is important to realize that their classification has not been externally validated].  their classification for antibiotic appropriateness:

    --“always”: eg pneumonia, UTI

    --“sometimes”: potential need for antibiotics, eg acute sinusitis, acute otitis media

    --“never”: almost never needs antibiotics, eg acute upper respiratory infection (URI), acute bronchitis, or non-infectious conditions

--they included 39 oral antibiotics, but not antiviral, anti-parasite, or anti-fungal agents

--19,203,264 enrollees evaluated: 76% adults and 24% kids

    --52% women, 12% from rural area, 16% from the Northeast/20% Midwest/46% South/17% West

Results:

--15,455,834 prescriptions filled in 2016: 805 prescriptions per 1000 enrollees

    --40% had an antibiotic filled: of these 52% had 1 antibiotic script, 24% had 2, 11% 3, and 13% 4 or more; mean of 2.0 antibiotic scripts for those getting antibiotics

    --18% of all adults filled at least one potentially appropriate antibiotic script vs 25% in kids

    --15% of all adults filled at least one potentially inappropriate antibiotic script vs 11% in kids

--antibiotics: most common azithromycin (19.0%), then amoxacillin (18.2%), then amoxacillin-clavulanate (11.6%); 5.9% were for refills of previous antibiotic scripts

-- by appropriateness categories:

    --13% “appropriate”

    --36% “potentially appropriate”

    --23% ”inappropriate”

    --29% “not associated with a recent diagnosis code" (99.8% of these did not have a claim in the prior 3 days)

--by the 3 most common conditions:

    --“appropriate” fills most frequently for UTIs, strep pharyngitis/tonsillitis, bacterial pneumonia

    --“potentially appropriate” fills were for acute sinusitis, acute suppurative otitis media, and acute phayrngitis

    --“inappropriate” fills were for acute bronchitis, acute URI, and respiratory symptoms such as cough

Commentary:

--one important advance in this study is that they used ICD-10 codes, which are not only more specific but also can be used in further international studies, since many countries use a modified-version of ICD-10’s. And they hope to develop a validated list of appropriate diagnoses with indications for antibiotic to inform further studies

--there are several concerns about the data in the above study:

     --in terms of patients getting antibiotics without having seen a provider, there are several reasonable scenarios where this appropriately takes place: eg, giving refills of minocycline or doxycycline, frequent longish-term treatments for acne; or for trimethoprim/sulfa, often called-in for women with typical urinary tract infection symptoms (or perhaps allowing refills in a prior prescription). And 58% of the patients getting prescriptions without an antecedent appointment received one of these 3 specific antibiotics

    --this study was done based on coded diagnoses, and may well be subject to error. Patients may be coded as “cough”, considered by them an inappropriate antibiotic indication, yet the treating clinician may have been concerned about acute sinusitis. But in the rush to treat patients appropriately, they may have decided to code just for the presenting complaint of cough, perhaps consciously deciding that taking care of patients was more important than spending yet more time on the profound time-sink of the electronic medical record….

    --no data on specific antibiotics per indication (ie, this aspect of antibiotic prescribing was not addressed: were the appropriate antibiotics given for a specific indication, with an emphasis on the most narrow-spectrum?? This is important both for effects on the microbiomes as well as the potential for drug resistance)

    --and, this is a select group of people, does not include (or represent) antibiotic prescribing in unemployed or uninsured people, or those over age 65.

--it would be useful to know if there were different aspects of the sites of care/services offered/numbers of patients clinicians saw per hour to see if those mattered: was there a difference by setting (academic vs non-academic, hospital vs community)?whether clinicians were seeing drug reps/going to drug company-sponsored events? if there was a relationship between the number of patients seen (it takes less time in a busy setting to write a script than discuss the issues of antibiotic overprescribing with patients, and even more time if the patient does not get a script and comes back for re-evaluation)? One of the potentially perverse incentives of the new ACOs is that it is more efficient and income-generating to see patients fewer times, which incentivizes the clinician to write more scripts in order to prevent revisits)

--I must admit that the broad conclusions inherent in these large data-mining studies are a bit jarring to me as a clinician, especially when they highlight terms such as “never appropriate”, since “never” is a pretty strong word, and:

    --they have no granular data about the actual condition of the patient. Were they labeled as a URI but were really sick and there was still clinical concern for a bacterial pneumonia?  Were they immunocoompromised and it was felt to be safer to treat even the unlikely prospect of a bacterial infection? Did they have other concerning comorbidities, perhaps severe underlying lung disease from continued smoking, or frailty that also swayed the clinician to prescribe antibiotics?

    --was the clinician concerned that the patient could not be followed well, in case they got worse? A long weekend without access to care? An upcoming vacation/travel plans? Difficulty for the patient to return if they get worse (financial difficulty?? Clinic far away from home? Demands at home making it hard for patient to leave? perhaps, a sick person at home they have to take care of? No transportation?)

    --was the clinician concerned that the patient did not have the appropriate judgment to return if getting worse? Psych issues? Cognitive issues? lots of denial?

    --and, using labels such as “never appropriate” may stigmatize clinicians. Maybe they will then start undertreating some of the above patients, their clinical judgement being undercut. Or be subject to undue scrutiny from peers or supervisors

    --the researchers did, however, try to compensate for the above concerns by skewing their interpretation of overprescribing to being quite lenient, as noted above

--and, on a macro scale, no doubt their broad conclusions about overprescribing are correct and my above concerns are unlikely to reflect a large number of people.

So, overall, this article does highlight the extremely concerning issue of antibiotic overprescribing, with its attendant increased risk of individual and societal harms:

--the individual is exposed immediate adverse effects:

        --severe allergies

        --even minor allergies or adverse effects that label them as allergic to an important or narrow-spectrum antibiotic, and therefore not having that antibiotic prescribed that in the future

    --And, longer-term effects:

        --adverse effects on the microbiome, leading to overgrowth/infection with potentially really bad bacteria (eg C. difficile)

        --and even the potential for bad non-infectious long-term effects, such as increased risk of colonic adenomas (see below)

--and, much more concerning, long-term changes in antibiotic resistance world-wide (see the following blogs, though there are lots more, if you search in my websites noted below)

blogs on antibiotic resistance:

    --a really shocking report by the WHO in 2014 of worldwide dramatic antibiotic resistance, with some organisms (like gonorrhea) being resistant to all drugs we currently use: http://gmodestmedblogs.blogspot.com/2014/05/who-report-on-antimicrobial-resistance.html 

    --a CDC report highlighting the increasing gonorrhea resistance: http://gmodestmedblogs.blogspot.com/2016/07/gonorrhea-resistance-increasing.html  

    --the development of more largely untreatable superbugs, eg e. coli and klebsiella: http://gmodestmedblogs.blogspot.com/2016/06/more-superbugs.html

    --an argument that we should mostly NOT be asking patients to take the full 10-day course of antibiotics, which in many cases leads to more antibiotic resistance without clear benefit: http://gmodestmedblogs.blogspot.com/2017/07/take-full-course-of-antibiotics.html

blogs on microbiome changes:

    --a study finding similar outcomes and fewer adverse events in kids on narrower-spectrum antibiotics: http://gmodestmedblogs.blogspot.com/2018/01/antibiotic-use-in-kids-narrow-spectrum.html

    --a small study showing that there can be long-term effects (12 month) on the gut microbiome even after a single exposure to antibiotics: http://gmodestmedblogs.blogspot.com/2015/11/longterm-microbiome-changes-with.html 

    --increased colorectal adenomas, including advanced ones, in women exposed to antibiotics: http://gmodestmedblogs.blogspot.com/2017/04/antibiotics-microbiome-changes-and.html

    --increased obesity and allergy in kids exposed to antibiotics: http://gmodestmedblogs.blogspot.com/2018/12/antibiotics-ppish2ras-increase-obesity.html

    --and a blog with 2 articles, one showing an unfortunate shift from prescribing narrow to broad-spectrum antibiotics and another showing that clinicians tend to prescribe more antibiotics late in a clinical session ("clinician fatigue"): http://gmodestmedblogs.blogspot.com/2015/07/antibiotic-overprescribing.html

other blogs on antibiotic overprescribing 

    --4 articles noting high prescription rates of antibiotics for URIs, pharyngitis and acute rhinosinusitis: http://gmodestmedblogs.blogspot.com/2016/01/antibiotic-overprescribing-and-acute.html 

    --another with a CDC report from 2013 finding huge overprescription of antibiotics for respiratory infections:http://gmodestmedblogs.blogspot.com/2015/07/antibiotic-overprescribing.html

a couple of hopeful notes: 

    --we are moving (slowly) in the right direction, see http://gmodestmedblogs.blogspot.com/2017/08/antibiotic-use-decreasing.html

    --antibiotic resistance can be decreased by hospital antibiotic stewardship programs: http://gmodestmedblogs.blogspot.com/2017/06/decreasing-antibiotic-resistance-by.html 

but a somewhat less hopeful one: 

    --drug companies are largely abandoning antibiotic development (it is much more profitable to develop a long-term drug for a chronic condition than even a high-priced one used for a few weeks): http://gmodestmedblogs.blogspot.com/2018/07/novartis-dumps-antibiotic-research.html . 

    --And, the above blog http://gmodestmedblogs.blogspot.com/2014/05/who-report-on-antimicrobial-resistance.html  also summarizes a pretty great BMJ article debunking the really exaggerated statement that R&D costs come close to justifying the astronomical cost of new drugs

geoff

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@uphams.org

to get access to blogs since 8/15/17:

1. go to http://gmodestmedblogs.blogspot.com/ to see them in reverse chronological order

2. click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category

3. or you can just type in a name in the search box and get all the blogs with that name in them

to access older blogs from the BMJ website, from October 2013 until 8/15/17: go to http://blogs.bmj.com/bmjebmspotlight/category/archive/