INCDocs request analysis
About an analysis of INCDocs requests over the period 1 August 2023 and 30 September 2024.
INCDocs started life in May 2023. You have told us that is has been transformational. It has streamlined document supply in the NHS in England.
The RD Team therefore decided to do some usage analysis. Here are the numbers for the 14 months between August 2023 and September 2024:
- Total number of requests: 83,653
- Total number of lookups: 169,010
Lookups tell you whether an article is available open access, via subscription in your library or at another NHS library, or not available at all. Any DOIs which cannot be reconciled with an article are recorded too.
Lookups
13% of lookups are open access or available via libraries’ own holdings
67% of lookups were available at other NHS libraries.
12% DOIs/PMIDs were not identified. This is most likely due to PMIDs not having associated DOIs. INCDocs only works for articles which have a DOI.
Only 8% of lookups were for articles from journals with no holdings in the NHS in England.
Of the 108,252 lookups available at another NHS library, 83,653 or 77.3% of them resulted in an INCDocs request being made. This may reflect continued use of legacy systems such as LAWMUNION and the PANDDA Union List (which have both been decommissioned now), but we’re keen to understand why this might be. If you have any other insight, please share.
13,683 lookups were not available in NHS library holdings. These would have had to be sourced via the INC scheme, subscription/paid for sources, or Reprints Desk.
Requests
Requests are made when you need to source articles from other NHS libraries.
- 80% or 66,736 were completed
- 15% or 12,696 were declined
- 5% of 4,221 had the status of pending
INCDocs requests may be declined for various reasons. Initially, this was mostly due to library holdings in HLM not being as accurate as they should be. In May 2023, the decline rate was 50%. Work on updating holdings led to a reduction in the number of requests being declined. In September 2024, it was 15% and when factoring in declines due to token access, duplicate requests and technical issues preventing download, the rate in nearer 10%. The reduction in the number of declines has improved the service to end-users and resulted in KLS staff spending less time on administration.
The pendings are mostly due to requests not being marked as complete in INCDocs. Don’t forget that you can go back to INCDocs to mark a request as complete. See How do I complete or decline an INCDocs request if I find it's still pending? on the service desk, for guidance.
The vast majority of requests are delivered to requesting libraries within 24 hours. This is due to the ease with which supplying libraries can locate, download and reply to request emails. This represents a step change in the timeliness of the service for end-users.
Holdings
We were keen to find out the impact of INCDocs requests on services. This depends on the total number of requests and the holdings available in the system.
There are 4,392 journals listed in INCDocs. Of these,106 (2.47%) had more than 100 requests for articles and 99 (2.31%) had no requests.
62.62% (2,988) of journals had between 1 and 10 requests, 22.62% (971) had between 10 and 50 and 5.31% (228) had between 50 and 100.
| Rank | Title | Number of requests |
|---|---|---|
| Rank 1 | Title JAMA | Number of requests 570 |
| Rank 2 | Title Nursing standard | Number of requests 508 |
| Rank 3 | Title New England journal of medicine | Number of requests 465 |
| Rank 4 | Title Radiograpics | Number of requests 439 |
| Rank 5 | Title The Lancet | Number of requests 410 |
A large number of requests with only small number of holdings places a bigger burden on services whereas, where there are small numbers of requests and many holdings, the opposite is true.
Initially we used the libraries supplying articles in response to requests as a proxy for the number of holdings in the NHS in England.
Where journals have the same number of requests and holdings, it’s not possible to tell how many holdings there might be in the system as a whole.
INCDocs randomises the allocation of requests to libraries with holdings, and so logically, the more requests, the more libraries involved in supply. On this basis, if there is more than one request, but the same library supplying, the more likely these journals are to be unique.
Since then we have been able to obtain a more accurate picture of e-journal holdings in NHS libraries and the number of unique titles is now 38 or 0.89% of all titles.
We divided the number of requests by the number of (holdings) to give a number. For example:
| Rank | Title |
Requests /holdings |
Requests | Holdings |
| 1 | JAMA oncology | 108.00 | 108 | 1 |
| 2 |
Journal of orthopaedic & sports physical therapy |
53.50 | 214 | 4 |
| 3 | NeoReviews | 53.00 | 53 | 1 |
| 4 | Reflective practice | 47.00 | 47 | 1 |
| 5 | The Lancet gastroenterology & hepatology | 46.33 | 278 | 6 |
| 6 | Physics in medicine and biology | 33.00 | 33 | 1 |
| 7 |
The Lancet rheumatology |
31.00 | 31 | 1 |
| 8 | The journal of clinical psychiatry | 23.80 | 119 | 5 |
| 9 | Clinical psychology review | 23.56 | 212 | 9 |
| 10 | Ophthalmology retina | 23.50 | 47 | 2 |
We can see that the request/holdings number for JAMA oncology is 108.00. That means that the libraries with holdings each handled 108 requests.
Journals with a higher request/holdings number are more important to the document supply system than ones which have a smaller number.
For example, Advances in Health Sciences Education with a requests/holdings number of 0.05 has considerably less impact on the document supply system than JAMA oncology
How can services use this information?
Find the INCDocs request analysis on the service desk.
Check the request/title figure for the titles to which you subscribe. The higher the number the greater their importance to document supply. Particularly, focus on journals where there is both only one holding and a high request/title number.
These 38 unique titles are coloured amber on the spreadsheet.
What else is happening with document supply?
We have commissioned PTFS Europe to integrate INCDocs within the Koha ILL module. This will streamline the ILL request workflow and make processing document supply requests even more efficient.
The ILL module will check for availability of articles within INCDocs and send the request to INCDocs automatically. It will check for the status of the request and let you know when it’s been completed or declined.
We are adopting a similar approach to PrintDOCs requests and will take advantage of the Reprints Desk plugin to automate this part of the workflow too. All NHS sources of supply will be queried and availability shown. If there are no NHS sources available, requests may be placed with other paid-for suppliers, up to and including the British Library, when they are back up and running.
For INCDocs itself we have commissioned several improvements. One to watch is the exclusion of holdings from INCDocs. This will be especially relevant to libraries with token access subscriptions and to manage supply for those services receiving a disproportionate number of requests due to the uniqueness of their holdings.
We are also looking to better understand PMIDs not found. This is likely due to articles not having a DOI. This will make it easier to identify these journals and add any holdings to PrintDOCs.
It’s a great time to be involved in document supply in the NHS.