4 ms·

OpenPrescribing: England's medication prescription data

DanBC 11y ago
This is amazing. I do some work around suicide prevention. Having access to this data means I can ask why so much co-proxamol (a medication that has no evidence of effectiveness, but which is also very dangerous) is being prescribed in Gloucestershire.
- napoleoncomplex 11y ago
  Completely agree, I was blown away at the level of transparency this brings to the world of pharma prescriptions. Inefficient care, overprescribing, otherwise suspect prescriptions, all of this laid out very clearly. Ben Goldacre is moving mountains, stone by stone.
- shitlord 11y ago
  I was taking a look around some of the data for antidepressants and was wondering why there are sometimes such drastic decreases in spending. Spending for Escitalopram dropped over 80% in a single month[1]. There are a few others that had sharp drops in spending. Just curious, do you know why? In any case, it would be really interesting to see how legislation, lobbying, and current events (relative) may have affected this data. [1]: https://openprescribing.net/chemical/0403030X0 https://openprescribing.net/chemical/0403030X0
  - snthd 11y ago
    https://en.wikipedia.org/wiki/Escitalopram#History https://en.wikipedia.org/wiki/Escitalopram#History Maybe the patent expiring. (WP talks about a US patent and this is the UK, so it's a guess)
    - WeaselNo7 11y ago
      I read that as 'the patient expiring', which I believe would also have an effect on drug use.
  - DanBC 11y ago
    I'm not sure about escitalopram. Possibly the NICE guidance changed and people stopped or started prescribing it? It's mentioned here: https://www.nice.org.uk/advice/ktt8/chapter/evidence-context https://www.nice.org.uk/advice/ktt8/chapter/evidence-context (First choice is a good quality talking therapy, but meds may be useful) > The full guideline on depression concluded that antidepressants have largely equal efficacy and that choice should mainly depend on side‑effect profile, people's preference and previous experience of treatments, propensity to cause discontinuation symptoms, safety in overdose, interactions and cost. However, a generic SSRI is recommended as first‑choice because of its favourable risk–benefit ratio. Neither escitalopram nor any of the available 'dual action' antidepressants, such as venlafaxine and duloxetine, were judged to have any clinically important advantages over other antidepressants. Results from meta‑analyses (Comparative benefits and harms of second-generation antidepressants for treating major depressive disorder, Gartlehner et al. 2011 and 2 Cochrane reviews: Cipriani et al. 2012, CD006534 and Cipriani et al. 2012, CD006533) have provided no evidence to depart from NICE guidance when selecting antidepressants for people with depression. And similarly here: http://cks.nice.org.uk/depression#!scenario http://cks.nice.org.uk/depression#!scenario > Citalopram and escitalopram have been found to prolong the QT interval and are now contraindicated in people who are already taking medication known to prolong the QT interval as there may be an additive effect [Lundbeck Ltd, 2011a; Lundbeck Ltd, 2011b; MHRA, 2011]. Caution is advised in people with congenital long QT interval or who have pre-existing QT interval prolongation as they have a higher risk than average of developing a ventricular arrhythmia including Torsade de Pointes. But it's interesting.
tomjhill 11y ago
I searched for the medicines I get from my GP and it turns out I'm the only one getting them. You can see the blips in the graph from zero when I got the prescriptions!
- Symbiote 11y ago
  Isn't this a privacy issue? There are some very small GP offices. From the HSCIC FAQ: I am prescribed a drug for a rare condition; can I be identified in this dataset? All practice level information down to presentation level is being released, but no information about patients is contained in the data. It is not possible to identify individual patients in the data. In line with the recent High Court ruling on the release of abortion statistics, data can be released unless an individual can be identified from the data or from other data that is already in the public domain. The release of practice level prescribing data does not enable the identification of individual patients. If you are the only patient receiving a certain drug in your practice then the number of items prescribed and their cost for that medicine will be in this dataset but it will not show which patient received it. Note that information about the price of drugs is already available in the public domain. That last paragraph doesn't make sense. I can easily imagine someone who knows a friend/relative has a rare condition using a data source like this to see how often it's being prescribed, and (perhaps) whether they're taking the medicine as often as they're supposed to.
  - tomjhill 11y ago
    On your last point, you can actually tell the months when I didn't collect a prescription.
  - notahacker 11y ago
    That was my first thought. I think it's brilliant that people with background knowledge can assess unusual prescription patterns at GP practices, but this is bound to release data that can give you a pretty good idea of [some of] what some individuals with a particular rare condition are being prescribed.
stewhuk 11y ago
This is an amazing dataset. I did some work with it in 2012 at a hackathon, in particular linking it to wiki entities. I used the redirects as drug synonyms and automatically extracted and linked drug names to their Wikipedia pages, which then offered more structured info on the drugs (eg side effects etc). From that, I then did some basic NLP to identify diseases that mentioned those drugs, thereby linking the diseases each drugs can be used to treat for further analysis. The next step we didn't do much with was linking it to the British national formulary - the uk clinical guidelines on drug use (a digital version wasn't readily available then). You can see some of those wiki datasets on my website (http://www.stewh.com http://www.stewh.com - I'm on a train so can't get the exact url, but you should find it in the menu). I had more datasets and the MapReduce code for working with the dataset and wiki dumps if anyone could use it.
- doktor_k 11y ago
  you might find the MEDI dataset useful. It links rxnorm to the ICD-9 codes the medications are written for based on EHR and other data. https://medschool.vanderbilt.edu/cpm/center-precision-medicine-blog/medi-ensemble-medication-indication-resource https://medschool.vanderbilt.edu/cpm/center-precision-medici...

DanBC 11y ago

This is amazing. I do some work around suicide prevention. Having access to this data means I can ask why so much co-proxamol (a medication that has no evidence of effectiveness, but which is also very dangerous) is being prescribed in Gloucestershire.

napoleoncomplex 11y ago
Completely agree, I was blown away at the level of transparency this brings to the world of pharma prescriptions. Inefficient care, overprescribing, otherwise suspect prescriptions, all of this laid out very clearly. Ben Goldacre is moving mountains, stone by stone.
shitlord 11y ago
I was taking a look around some of the data for antidepressants and was wondering why there are sometimes such drastic decreases in spending. Spending for Escitalopram dropped over 80% in a single month[1]. There are a few others that had sharp drops in spending. Just curious, do you know why? In any case, it would be really interesting to see how legislation, lobbying, and current events (relative) may have affected this data. [1]: https://openprescribing.net/chemical/0403030X0 https://openprescribing.net/chemical/0403030X0
- snthd 11y ago
  https://en.wikipedia.org/wiki/Escitalopram#History https://en.wikipedia.org/wiki/Escitalopram#History Maybe the patent expiring. (WP talks about a US patent and this is the UK, so it's a guess)
  - WeaselNo7 11y ago
    I read that as 'the patient expiring', which I believe would also have an effect on drug use.
- DanBC 11y ago
  I'm not sure about escitalopram. Possibly the NICE guidance changed and people stopped or started prescribing it? It's mentioned here: https://www.nice.org.uk/advice/ktt8/chapter/evidence-context https://www.nice.org.uk/advice/ktt8/chapter/evidence-context (First choice is a good quality talking therapy, but meds may be useful) > The full guideline on depression concluded that antidepressants have largely equal efficacy and that choice should mainly depend on side‑effect profile, people's preference and previous experience of treatments, propensity to cause discontinuation symptoms, safety in overdose, interactions and cost. However, a generic SSRI is recommended as first‑choice because of its favourable risk–benefit ratio. Neither escitalopram nor any of the available 'dual action' antidepressants, such as venlafaxine and duloxetine, were judged to have any clinically important advantages over other antidepressants. Results from meta‑analyses (Comparative benefits and harms of second-generation antidepressants for treating major depressive disorder, Gartlehner et al. 2011 and 2 Cochrane reviews: Cipriani et al. 2012, CD006534 and Cipriani et al. 2012, CD006533) have provided no evidence to depart from NICE guidance when selecting antidepressants for people with depression. And similarly here: http://cks.nice.org.uk/depression#!scenario http://cks.nice.org.uk/depression#!scenario > Citalopram and escitalopram have been found to prolong the QT interval and are now contraindicated in people who are already taking medication known to prolong the QT interval as there may be an additive effect [Lundbeck Ltd, 2011a; Lundbeck Ltd, 2011b; MHRA, 2011]. Caution is advised in people with congenital long QT interval or who have pre-existing QT interval prolongation as they have a higher risk than average of developing a ventricular arrhythmia including Torsade de Pointes. But it's interesting.

tomjhill 11y ago

I searched for the medicines I get from my GP and it turns out I'm the only one getting them. You can see the blips in the graph from zero when I got the prescriptions!

Symbiote 11y ago
Isn't this a privacy issue? There are some very small GP offices. From the HSCIC FAQ: I am prescribed a drug for a rare condition; can I be identified in this dataset? All practice level information down to presentation level is being released, but no information about patients is contained in the data. It is not possible to identify individual patients in the data. In line with the recent High Court ruling on the release of abortion statistics, data can be released unless an individual can be identified from the data or from other data that is already in the public domain. The release of practice level prescribing data does not enable the identification of individual patients. If you are the only patient receiving a certain drug in your practice then the number of items prescribed and their cost for that medicine will be in this dataset but it will not show which patient received it. Note that information about the price of drugs is already available in the public domain. That last paragraph doesn't make sense. I can easily imagine someone who knows a friend/relative has a rare condition using a data source like this to see how often it's being prescribed, and (perhaps) whether they're taking the medicine as often as they're supposed to.
- tomjhill 11y ago
  On your last point, you can actually tell the months when I didn't collect a prescription.
- notahacker 11y ago
  That was my first thought. I think it's brilliant that people with background knowledge can assess unusual prescription patterns at GP practices, but this is bound to release data that can give you a pretty good idea of [some of] what some individuals with a particular rare condition are being prescribed.

stewhuk 11y ago

This is an amazing dataset. I did some work with it in 2012 at a hackathon, in particular linking it to wiki entities. I used the redirects as drug synonyms and automatically extracted and linked drug names to their Wikipedia pages, which then offered more structured info on the drugs (eg side effects etc). From that, I then did some basic NLP to identify diseases that mentioned those drugs, thereby linking the diseases each drugs can be used to treat for further analysis. The next step we didn't do much with was linking it to the British national formulary - the uk clinical guidelines on drug use (a digital version wasn't readily available then). You can see some of those wiki datasets on my website (http://www.stewh.com http://www.stewh.com - I'm on a train so can't get the exact url, but you should find it in the menu). I had more datasets and the MapReduce code for working with the dataset and wiki dumps if anyone could use it.

doktor_k 11y ago
you might find the MEDI dataset useful. It links rxnorm to the ICD-9 codes the medications are written for based on EHR and other data. https://medschool.vanderbilt.edu/cpm/center-precision-medicine-blog/medi-ensemble-medication-indication-resource https://medschool.vanderbilt.edu/cpm/center-precision-medici...