SDN members see fewer ads and full resolution images. Join our non-profit community!

Pharmacy Times Case Studies Question

Discussion in 'Pharmacy' started by BidingMyTime, Sep 10, 2017.

  1. SDN is made possible through member donations, sponsorships, and our volunteers. Learn about SDN's nonprofit mission.
  1. BidingMyTime

    BidingMyTime Lost Shaker Of Salt 10+ Year Member

    2,642
    1,501
    Oct 2, 2006
    Illinois
    So the latest Pharmacy Times has a case study concerning a woman who refuses vaccines due to religious beliefs, but wants a recommendation for prophylaxis treatment for meningitis. Interestingly, she is also on a BCP. The answer gives 3 choices, one of which is ceftriaxone, which Pharmacy Times says the woman can't take because her religion forbids all injections. Seriously? What religion forbids "all injections?" (I'm not a theologist, so maybe I'm ignorant, but I have never heard of any religion that forbids "all injections.")

    We all know there are religions that forbid vaccines, and I'll assume the woman is taking BCP for health reasons and not because she is engaging in hypocritical premarital intercourse which I'm pretty sure is forbidden by all religions that forbid vaccines, but where is this idea that she can't take an antibiotic injection because her religion forbids all injections coming from? I call BS. The real reason that Pharmacy Times should have given against ceftriaxone is that since there are non-injectable options, why in the world would you want to recommend a more expensive and intrusive option?

    I am in an annoyed mood today, and I found this stupid case study answer very annoying. I need coffee, I guess. I think the authors were lazy, so they completely made up this unnamed religion that forbids "all injections", because they couldn't be bothered to use common sense to rule out the ceftriaxone in the answer.
     
    Dalteparin likes this.
  2. SDN Members don't see this ad. About the ads.
  3. BidingMyTime

    BidingMyTime Lost Shaker Of Salt 10+ Year Member

    2,642
    1,501
    Oct 2, 2006
    Illinois
    And now I'm more annoyed because I just attempted to do that stupid crossword in the same issue. Bite? Seriously? I see what they are getting at, but the clue totally sucks.
     
  4. quickpic007

    quickpic007

    278
    133
    Sep 22, 2016
    just use the retail answer. "I recommend speaking to your physician"
     
  5. steveysmith54

    steveysmith54 Member 10+ Year Member

    1,819
    651
    Dec 28, 2005
    I tell this pt to weigh her pros and cons... you wanna get meningitis or you wanna listen to some silly religion. And I am out...
     
    BidingMyTime likes this.
  6. zelman

    zelman 7+ Year Member

    2,778
    1,850
    Nov 27, 2009
    Well, the meningitis will get you to heaven faster, so that's a pro.
     
  7. lord999

    lord999 Moderator Emeritus 10+ Year Member

    1,359
    1,060
    Feb 20, 2002
    DC
    Pharmacist
    Faculty
    No, coffee isn't necessary, you had the right answer intuitively.

    Proper answer: E: I don't give a ______(fill in the blank).

    Helpful answer: F. Kick this to a prescriber.

    Troll answer: G. Lookup in Sanford's for Ciprofloxacin, smile, get the script for Cipro 500mg, dispense, and hand the required PPI for the birth control and Cipro while saying nonspecifically that you ought to read the PPIs to the patient due to known interactions, and notate on the script that counseling was provided on the pregnancy risk and that medical advice was not followed for religious reasons (which is legally true but not sincerely so). Exchange a small problem by making it into a potential birth control disaster for wasting your time.
     
    Amicable Angora and BidingMyTime like this.
  8. lord999

    lord999 Moderator Emeritus 10+ Year Member

    1,359
    1,060
    Feb 20, 2002
    DC
    Pharmacist
    Faculty
    I also share your hatred for contrived "trivia" questions in CE where they obviously wouldn't be things that you'd need to know first-line and can look up later. There's easier ways of spicing up a case (just add diabetes, for instance).
     
    BidingMyTime likes this.
  9. Dalteparin

    Dalteparin 7+ Year Member

    647
    405
    Aug 17, 2009
    A religion that forbids all injections? GMAFB. They could have just said she's refusing injections due to an extreme needle phobia - that would have actually been believable!


    Sent from my iPhone using SDN mobile
     
    CetiAlphaFive and BidingMyTime like this.
  10. Carboxide

    Carboxide 7+ Year Member

    1,067
    140
    Dec 6, 2008
    It's not like ctx is the drug of choice anyways. Cipro, baby, cipro. And I don't care what laypeople say, it doesn't interfere with OCPs. That's rifaximin. Moms everywhere can just stop spreading that rumor thank you very much.

    Regardless, a x1 500mg cipro isn't going to cause any harm no matter what other drugs you're on. You don't need more than one dose, and she would only need it if the patient coughed/sneezed directly in her face or if she intubated the patient.
     
  11. BidingMyTime

    BidingMyTime Lost Shaker Of Salt 10+ Year Member

    2,642
    1,501
    Oct 2, 2006
    Illinois
    Well, she is scared of having caught it from her cough, cough "roommate". Since she's on BCP, we can assume she is having many "close" contacts with her "roommate."
     
  12. callMe_MrA

    callMe_MrA

    4
    0
    Sep 23, 2017
    Hi can you answer this question for me?

    How do you calculate this...

    what is the smallest number of samples to prove a 5% reduction?

    thank you in advance
     
  13. zelman

    zelman 7+ Year Member

    2,778
    1,850
    Nov 27, 2009
    That's easy. Percent comes from the Latin "centum" meaning one hundred. Thus to find a percent, you need a hundred data points. However, since you want a percent reduction you need to double it (100 points before and 100 points after to compare) and your answer is 200 samples.
     
  14. callMe_MrA

    callMe_MrA

    4
    0
    Sep 23, 2017
    wrong!
     
  15. owlegrad

    owlegrad Uncontrollable Sarcasm Machine Staff Member SDN Administrator 7+ Year Member

    Are you looking for a power calculation?
     
  16. lord999

    lord999 Moderator Emeritus 10+ Year Member

    1,359
    1,060
    Feb 20, 2002
    DC
    Pharmacist
    Faculty
    The question is not well constructed, attempting to redefine with a uniform manufacturing process that has a measurable fraility. What this question really is missing is the standard deviation, the estimated error, and if this were asked in the graduate context, that there is some distribution coming with this (tablet stability degradation is NOT usually a Normal process, it usually follows some power distribution or Weibull.)

    If you're talking about casualty statistics with Sampling with No Replacement, then:
    1. If the reduction is along normal lines and has a minimal standard error, no less than 15 through weak CLT. But, you HAVE to know that it's normal and that the standard error is minimal (<1). That's a really, really small number of samples, and guess what, pharmacy uses this all the time incorrectly. The proportional form would be 0.95(1-0.95) multiplied by the standard error (something of that nature, since it is at the far end of the proportion, is going to be big even with small standard errors).

    2. If the reduction is NOT along normal lines,
    a. But it can be defined as a parametric line, then it's Pocock's test at one cutoff or the O'Brien/Fleming at increasing cutoffs at each interval to a stopping point.
    b. Nonparametric lines, no answer, it's not known with certainty, and there's a lot of ongoing work on trying to quantify this.

    3. If this is Bayesian, then you would use 2a's answer and define some limit. That's actually the way industry does adaptive testing.

    If your Pharmaceutics or Calculations professor gave you this question with only that much, he/she is one of the idiots why I have such a job in industry.
     
    owlegrad likes this.
  17. lord999

    lord999 Moderator Emeritus 10+ Year Member

    1,359
    1,060
    Feb 20, 2002
    DC
    Pharmacist
    Faculty
    I can't tell, but if it is, there's key parts of the question that are missing that render it unanswerable. The parts you need to make a power or stability calculation work are straightforward irrespective of context, but require:
    1. Assumptions about the manufacturing process to figure out uniformity (Not given)

    2. A series of measurements of the actual process outputs and some baseline statistics (Not given, if this is a proportion, then it conforms to some rules that require you to know the standard error (SD/error), if this is a mean, then you would need the standard error and in context, the distance between the mean and difference)

    3. The testing plan or the assumed distribution around "failure" (strength decrease). If you don't know it, then you have to establish this from experimental runs of a full batch, and those are usually calculated at the parameter (the whole population) level. (Not given)

    I like asking my grad students these questions as the test is really what would you need to know to answer the clients' question. But, as presented, this is unsolvable. I'd like to see this guy/gal's solution without bringing in any other assumptions in. Otherwise, it's a badly stated problem.
     
    owlegrad likes this.
  18. zelman

    zelman 7+ Year Member

    2,778
    1,850
    Nov 27, 2009
    I'm pretty sure I'm right.
     
  19. owlegrad

    owlegrad Uncontrollable Sarcasm Machine Staff Member SDN Administrator 7+ Year Member

    Oh I am well aware, I was just trying to help him help us so we can help him. I am really curious what the context of the question is. My money, obviously, is on power. I wonder if this is a conceptual question so rather than the answer being "125" it is something like "a large enough sample size to disprove the null hypothesis while minimizing the chance of a type A error" or something like that (been a while since I took statistics so it is possible my example is basically nonsense).
     
  20. lord999

    lord999 Moderator Emeritus 10+ Year Member

    1,359
    1,060
    Feb 20, 2002
    DC
    Pharmacist
    Faculty
    Yeah, I agree. @zelman is correct under the USP friability/stability rules, which don't use statistics to deal with the matter, but require batches corresponding to either quantity or weight (it's a two full runs test, and more if it's sterile according to some GMPs) to establish variance (that's every single dosage unit tested individually for dose and recorded) which gives the standard error necessary to make a guess at the underling distribution. Most students don't work in real life conditions where sample statistics aren't trusted for industrial runs when establishing initial parameters. Inferential statistics are only used once those underlying procedures are worked out and a baseline control batch is prepared.

    And yes, 125 is far more than enough to establish whether weak CLT can apply (where t converges with z), but depending on the real underlying distribution and variance, could need quite a bit more. There's a provision in the strong form of CLT that states that despite the actual distributions, multiple samples taken from that underlying population will almost always be normal, the caveat that people don't read is that you just don't know where along that distribution the sample you actually took lies, and that you are making a very big assumption (and not justifiable under industrial conditions) that what predicts the dose from the process are purely individual contributors (the iid). Since iid doesn't really hold and that you want to be absolutely sure that if you're making a couple of thousand of these, you want a good idea of your quality control and variance measurement is key to figuring that out.

    The reason you were taught to a weight or a dosage unit standard was that statistics textbook assumptions do not always apply and drugs are usually expensive enough that destructive analytic chemistry needs to be minimized when possible. That's why the cGMPs and USP have those limits. But pharmacy students today are geeks with formulas that never cracked open a USP or Remington, much less actually made and analyzed enough stuff to understand where and why you would need different quality control statistics as well as a philosophy to deal with it. As an aside, I really wish we taught that part of statistics (statistical process control like run charts) rather than the frequentist crap that pharmacists get wrong all the time anyway and rarely applies.
     
    owlegrad likes this.

Share This Page