Frequently Asked Questions - Services

1. Do you offer services besides thesaurus coding and versioning?

Return to top

2. How long does it take to get started?

Return to top

3. Do you have a standard contract, or do we need to provide one?

We have a standard contract, but are also happy to use one that you provide. Whichever option you (and your legal department) prefer is fine with us.

Return to top

4. I have some studies that need to be coded now, plus new studies and/or projects that may be starting up later. How will you arrange my contract to handle my immediate needs as well as future (but still to-be-determined) needs?

For example, some of our clients who have outsourced all their coding and versioning to us have a single statement of work that covers everything. This is usually the simplest approach as the legal arrangements only need to be handled once. Other clients prefer to execute an additional statement of work for each new study and/or project that arises. This approach is also fine with us. We will work with you to set up the arrangement that works best for you.

Return to top

5. MedDRA and WHO Drug licenses are very expensive. Do you provide the dictionaries?

The license agreements for both MedDRA and WHO Drug prevent us from giving coded data to another company unless that company also holds a valid dictionary license. So even if ThesIS is doing your coding and/or versioning, you will still need to have current licenses for MedDRA and/or WHO Drug. For MedDRA, the good news is that the license fees have been significantly reduced (especially for smaller companies), so the cost should no longer be an issue.

For WHO Drug, small sponsor companies (yearly revenue less than US$10 million) who are conducting only phase 1 or phase 2 clinical trials may qualify for a "Limited Study License for Sponsors." The price for a Limited Study License is significantly less than for a full WHO Drug license. Please contact us to learn more about this license option and the conditions needed to qualify for it.

Return to top

6. I am a CRO. Do I still need dictionary licenses if you and the sponsor companies already have them?

The answer to this question depends upon how the coded data is transferred. If we pass the coded data to you and then you pass the coded data to the sponsor(s), you will need dictionary licenses, since you will be a recipient of coded data. If we pass the coded data directly to the sponsor(s), then you will not need dictionary licenses (provided that the sponsors never pass that coded data to you later). In other words, if you will ever be in receipt of coded data, then you will need dictionary licenses.

As a provider of coded data, we are responsible for ensuring that we pass coded data only to current licensees of MedDRA and/or WHO Drug. What happens to the coded data beyond that is outside of our control and therefore not within the scope of our responsibility. If we pass the coded data to you, then we will need to verify that you have valid dictionary licenses, but we will not check your sponsors. On the other hand, if we pass the coded data directly to your sponsor companies, then we will need to verify that they have valid dictionary licenses, but we will not check you.

Return to top

7. Are rush jobs possible? What are the charges for a rush job?

Rush jobs are always possible. We fully understand that situations can unexpectedly arise where your timelines are tight. We have "bailed out" quite a few companies who were in a panic with seemingly impossible deadlines.

If you have something that you need done in a rush, please contact us to discuss the details. Quite often, what you may consider a rush job is something that we can accommodate without imposing any extra charges.

In those situations where your request does qualify as a true rush job, you will find that our rush charge is quite reasonable. We do not impose a rush charge as a means for inflating our revenues. A minimal surcharge is imposed only because of the impact that a rush job has on our existing workload and the increased pressures and challenges of turning around the data very quickly while maintaining the highest quality.

Return to top

8. I already have CROs performing my coding. Why should I use ThesIS?

Furthermore, if you have multiple CROs coding your data, a term that is reported to more than one CRO ends up coded by each CRO, rather than having that term coded once, with the coding shared by all your CROs. First, this redundant coding is highly inefficient and unnecessary. Second, having multiple sources of coding dramatically increases the likelihood that the same or like terms are not being coded consistently. Even if all of your CROs are diligently following the same coding conventions, there is no guarantee that the same term is coded exactly the same by all of your CROs. With multiple CROs coding your data, you will encounter difficulties when pooling that data.

By having ThesIS code your data:

• You will have a centralized source of coding done by some of the most experienced coders in the industry.

• You will pay only once for a term to be coded, regardless of the number of times that term has been reported across your studies (or across your entire company).

• You will eliminate difficulties with pooling data caused by inconsistent coding.

• You can be confident that your coding is being done accurately, consistently and cost-effectively across your entire company.

Return to top

9. Why should I use your services rather than those offered by another coding company?

Simply put, we offer better results. ThesIS stands out from other coding companies in many ways:

• All coding is done by our regular employees. We are not a "virtual" company that gets a coding job and then tries to find the personnel to fulfill the job. We never subcontract coding to someone else, nor do we ever use temporary employees. Your coding is always done by the same, established team of ThesIS coders.  Our team approach means that you get accurate and consistent results every time.

• Our coders average more than 16 years of coding expertise and are among the best in the industry. Every coder is a medical professional with years of practical medical experience working in hospitals and/or clinics. You will be hard pressed to find a better group of medically qualified, dictionary-proficient coders.

• We have used MedDRA in clinical trials since 1997 (the MedDRA 1.5 beta version). Most early adopters of MedDRA have used it only for drug safety. By using it early on for coding clinical trials, we have had many years of experience developing and refining practical techniques for effective MedDRA coding and versioning on a large scale.

• We offer a complete versioning service for both MedDRA and WHO Drug that is far and away the best in the industry.

• Our coders utilize arguably the best and most powerful thesaurus system available. Our proprietary software is not an autoencoder but a complete and extremely powerful system for coding and versioning. We combine expert staff with unparalleled technology to achieve the best possible results.

Return to top

10. To which dictionaries can you code?

We can code to any of the standard dictionaries such as MedDRA, WHO Drug, ICD, and SNOMED. We can also code to legacy dictionaries such as COSTART and WHO-ART, as well as to any custom dictionaries you have created.

Return to top

11. What are your turnaround times for coding?

This is one of the most frequent questions that we get from new clients, and also one that has no set answer. The turnaround time will depend upon the volume and, to a lesser extent, the quality of data submitted. We have turned around data as quickly as the same day. Very large jobs might, of course, require a few weeks or even months.

You can choose to send us data for coding on a scheduled basis (for example, weekly or monthly). We would typically send you back the coded data also on a scheduled basis, usually at the same frequency in which the data is sent to us so that each set of data is coded and sent back before a new set is submitted. The frequency of the data exchange can be adjusted as necessary. With a pre-defined schedule, you would always know your "worst-case" turnaround time, but would not be constrained by it. You always have the ability to send and/or receive data at any time in order to meet timelines that do not coincide with the pre-defined schedule.

A pre-defined schedule is not required. You can choose to send data whenever you feel it is "appropriate" and let us know when you would like the coded data returned to you. The turnaround time is discussed and agreed to at the time each set of data is submitted to us.

No matter which approach you choose, our goal is to ensure that your data is coded expediently while still maintaining the highest quality. We recognize that timely coding of your data is essential for meeting timelines. We partner closely with you to ensure that your timelines are never jeopardized because of a lack of coded data. We fully understand the nature of clinical trials and safety reporting, with their constantly changing priorities and timelines. Even if you need your data coded in a rush, we do our utmost to get your data completely coded when you need it.

If you are still concerned about our ability to meet your turnaround times, ask to talk to our existing clients. Ask them about their experiences working with us and how well we meet their timelines. Turnaround times are one of the most frequently asked questions from prospective clients, but are not an issue at all with our existing clients. Their testimonies are the best proof we can offer.

Return to top

12. Will you always have the same coder(s) coding my data?

Your coding may not always be done by the same person, but will always be done by a regular ThesIS employee. We do not subcontract coding out to others, nor do we ever use temporary employees. Your coding will always be done by the same group of individuals, all of whom are our regular employees.

Generally speaking, our coders work as a team, which means that any one of them can code your data without any variance in accuracy or consistency. We achieve this through constant discussion and communication among our coders, in addition to a thesaurus system that facilitates the use of client-specific coding conventions. This means all our coders become equally familiar with all of your coding conventions, and our proprietary system helps ensure adherence to those custom conventions.

While it may seem to be more beneficial to have your coding always done by the same person, our team approach offers a couple of distinct advantages. First, having more than one coder eliminates the potential for individual bias. Second, we can quickly and seamlessly assign as many coders as necessary to meet your timelines without sacrificing quality. In essence, your coding team contracts or expands as needed, but you continue to get the same great results.

Return to top

13. How do you handle terms that cannot be coded?

Terms that cannot be coded will be queried, i.e. tagged as uncodable with a reason why the term cannot be coded. In most cases, reported terms that are queried will need to be changed and the revised term re-submitted to us.

In those instances where the reported term cannot be queried and/or changed (for example, for legacy data, or when an imminent study deadline does not allow sufficient time), we can, as a last resort, still code the term using the best possible entry available in the corresponding dictionary. This will often be a generic, catch-all entry that may not fully or adequately represent the essence of the reported term, but will at least allow the term to be coded.

Return to top

14. How do I submit terms to be coded?

See the answer to this question in our Data Exchange FAQ.

Return to top

15. What will the coded data that I get back look like?

Return to top

16. What kind of reviews do you do on the coded data before sending it back?

Your coded data will be reviewed by a minimum of two of our expert staff, with at least one person who did not do the actual coding. In most cases, our entire coding team will review your data. This full team review not only achieves the highest quality results for you, but also ensures that our team is constantly benefiting from each other's findings and experience. The discussions and conclusions arising from these reviews are essential for our coders to maintain the highest level of expertise.

Return to top

17. What autoencoders do you use for coding?

We do not use any autoencoders for coding. In our opinion, there are no autoencoders currently available that are capable of critically analyzing and interpreting a reported term and then making a precise and correct selection from the myriad of possible choices within the highly specific dictionaries used today.

Coding within highly specific dictionaries such as MedDRA requires a tremendous understanding of the dictionary as well as a high degree of medical knowledge. Almost all autoencoders work using fairly simplistic methods (string and pattern matching, removal of punctuation and "stop" words, changing the order of words) without any provision or ability to distinguish the medically significant differences among similar entries in a dictionary. Embedding the requisite medical knowledge into an autoencoder would be quite difficult and require a tremendous number of extremely sophisticated algorithms. Using an autoencoder that lacks embedded medical knowledge will invariably deliver compromised results and will often require the data to be re-coded, costing more in the long run.

At ThesIS, a reported term will autoencode only if it is an exact string match to a dictionary entry (such as a MedDRA lower level term) or to a term that we have coded already. Instead of autoencoders, we use a proprietary thesaurus system with extremely powerful search capabilities that maximize our coders' dictionary expertise and medical knowledge. This combination of expert staff and powerful technology delivers the best results.

Return to top

18. What is the difference between accuracy vs. consistency?

Accurate coding means that the code chosen for a reported term fully and correctly represents the meaning contained with that reported term.

Consistent coding means that similar reported terms are all coded the same way.

It is possible to be accurate without being consistent (each reported term appears to be coded correctly when viewed by itself, but when grouped with like reported terms the coding differs among the like terms).

Likewise, it is possible to be consistent in coding but be inaccurate (like terms are all coded to the same dictionary entry, but the chosen dictionary entry is not the best possible choice).

For example, the following two similar reported terms may appear to be coded accurately when viewed individually. When viewed together, however, it becomes clear that they have not been coded consistently.

This inconsistency results from a failure to create and/or apply a proper coding guideline to handle a reported term with a symptom reported as "due to" an underlying medical condition. If the coding guideline specifies, for example, that a reported term with a "due to" should be linked to the underlying medical condition rather the symptom, then both of these terms should be linked to PORPHYRIA.

Another example of inconsistent coding:

Consistency without Accuracy

The two similar reported terms below have been coded consistently.

The coding for both is inaccurate, however, because a sarcoma is a very specific type of tumor and one cannot arbitrarily assume that the tumors being reported are in fact sarcomas. ENDOMETRIAL NEOPLASM NOS would be a better choice for these two terms (better to select a less specific but correct term than a more specific but incorrect term like ENDOMETRIAL STROMAL SARCOMA).

Accuracy and Consistency

At ThesIS, we strive to be both accurate and consistent in our coding. Accuracy is achieved because of our in-depth knowledge of the dictionaries and our medical qualifications. We achieve consistency by developing and rigorously applying comprehensive coding conventions. Both accuracy and consistency are further achieved from our thorough reviews of the coded data.

Return to top

19. How do you ensure consistency?

• Creation of good coding guidelines

• A thesaurus system that facilitates the adherence to those coding guidelines

• Coding of terms by first searching for similar reported terms that we have already coded so that new terms are coded the same as previously coded terms

• Comprehensive reviews using reports that highlight instances where similar reported terms have been coded differently

Return to top

20. Do I need to have my own coding guidelines?

You do not need to have your own coding guidelines. We will code your data using our "standard" coding guidelines if you do not have your own.

Return to top

21. What if I already have my own coding guidelines?

If you already have coding guidelines, we will use them instead of our ThesIS coding guidelines. We will first review your coding guidelines to ensure that we full understand them. Your coding guidelines will then be incorporated into our thesaurus system to ensure that we follow them when coding your data.

Return to top

22. What is meant by "out-of-context" coding?

"Out-of-context" coding means that we code using only the text as contained in the reported term. We do not use any information that is contained in other fields of a data collection form and/or that accompanies the reported term. All information necessary to classify the reported term must be part of the text of the reported term. No information may be assumed, nor may information be added.

For example, if the reported term is CHEST PAIN, we would code that in MedDRA to CHEST PAIN even if ANGINA PECTORIS is stated in an accompanying comment field. In order for us to code to ANGINA PECTORIS, it would have to be included in the reported term itself, e.g.

As another example, if the reported term is:

     ABDOMINAL PAIN, INCREASED SERUM AMYLASE, INCREASED SERUM LIPASE

it would be inappropriate for us to assign a diagnosis of PANCREATITIS. For us to code to PANCREATITIS, it would need to be included in the text of the reported term, e.g.

     ABDOMINAL PAIN, INCREASED SERUM AMYLASE, INCREASED SERUM LIPASE
     DUE TO PANCREATITIS

Likewise, a reported term of LOW HEMOGLOBIN would be coded to HEMOGLOBIN LOW and not to ANEMIA, even if LOW HEMOGLOBIN were reported in a clinical trial where anemia was an area of interest.

Return to top

23. When I submit terms for coding, can I ask for a certain set of them to be coded first?

Yes, you can specify which terms you want coded first. Please see the answer to this question in the Data Exchange FAQ for more information.

Return to top

24. What if I have a change in priority, for example a patient close-out, interim analysis, DSMB, etc.?

Let us know as soon as your priorities change, or if you believe they may be changing. We understand that priorities shift constantly in the drug development process and are quite flexible to meet your needs.

Return to top

25. How often can I send you data?

Generally speaking, you can send us data as frequently as you and we think is needed in order to meet your timelines. You do not need to wait until a clinical trial is nearing completion before sending us the data to be coded. Our data exchange process provides an easy way for you to send us data throughout the trial to allow for the coding to be done while the trial is in progress.

Some companies prefer to establish a regular schedule, e.g. weekly, monthly, or quarterly, in which they will send us data. Other companies work on a more ad-hoc basis, sending us data when they feel they have a significant enough number of new terms to be coded. Either approach works fine with us, as long as there is sufficient time allotted for us to meet your turnaround times.

Return to top

26. How often can I get back coded data? How quickly can I get my data?

In terms of getting all of your data completely coded, turnaround times will vary depending on the number of terms you submit, and, to a lesser extent, the complexity and quality of those terms. Click here for more information about turnaround times.

Return to top

27. What if I disagree with the coding of a reported term?

Return to top

28. Some of my studies are on one version of a dictionary, but others are on a different version. Can you code to more than one dictionary version?

We can code to as many versions of a dictionary as you need. Quite often, however, a company uses more than one version because they do not have a good versioning process in place. Our innovative versioning service offers a easy and cost-effective solution that can eliminate the need to code to multiple versions. We can assess whether it makes sense for you to use multiple versions, or whether it is better to migrate to a single version.

Return to top

29. You mention versioning services for MedDRA and WHO Drug. Do you offer versioning for other dictionaries?

We can version for dictionaries other than MedDRA and WHO Drug. The difference is in the method we use for versioning. New versions of MedDRA and WHO Drug are released on a regular, scheduled basis, and the need to continually upversion for these two dictionaries is well known. As a result, for MedDRA and WHO Drug, we developed special software tools that enable our coders to pinpoint only those reported terms that need to be re-coded due to changes between versions. We can therefore upversion on MedDRA and WHO Drug very quickly and cost-effectively. For other dictionaries, we would migrate your data to a newer version by re-coding all of your data, similar to the process we would use if you were migrating from one dictionary to another (e.g. from COSTART to MedDRA).

Return to top

30. Can you version my data even if someone else coded it?

Yes, we can version your data even if someone else coded it. When you provide us the existing coding, we will load it into our thesaurus system. From that point, our coders can use our special versioning software no differently than if we did the original coding.

Return to top

31. Do I have to migrate to every new version, or can I skip versions?

Our versioning process allows you to skip over as many versions as you want. We do not rely upon the change files provided by the dictionary vendors as they only show the changes between consecutive versions. Instead, our special versioning software automatically detects the differences between any two versions, so you can jump over as many versions as you want. For example, we can migrate you from MedDRA 8.0 directly to MedDRA 9.0 rather than migrating first to MedDRA 8.1 and then 9.0.

Return to top

32. How long will it take to version my data?

Return to top

33. Do I really need to version my data?

Many companies who refrain from versioning do so because of the perceived difficulties and cost in getting their data versioned, not because they lack good reasons to version. With our easy and cost-effective versioning service, whether or not to version becomes a much simpler decision.

Return to top

34. I currently version my data by re-coding only if a dictionary entry is no longer available. For example, for a new MedDRA version, all I do is to re-code any term that is coded to an LLT that has been set to non-current in the new version. I leave all my other data unchanged. Why isn't this approach sufficient?

The problem with this approach is that new LLTs are always being added to MedDRA. If you use these new LLTs for coding new data, but don't use them for your previously coded data, like terms will be coded differently over time. These inconsistencies can negatively impact your analysis and reporting. A proper versioning strategy for any dictionary must also determine if entries added to the new version are better choices for your previously coded data. The challenge is to do this without having to completely re-code all your data (which would be extremely costly and time consuming). Our unique versioning service employs a highly targeted approach so that your existing data is fully updated to the latest version with minimal re-coding.

Return to top

35. Do you use autoencoders to do the versioning?

We do not use autoencoders to do versioning for the same reasons we do not use them for coding. In addition, using an autoencoder to do versioning is essentially the same as re-coding your data from scratch. You end up having to review all of the autoencoded results as well from scratch, even though most of the original coding does not need to change between versions. None of the reviews, discussions, decisions, corrections, etc. that you conducted for the original coding is leveraged. Instead of re-coding all of your data, we employ a top-down approach that pinpoints only those instances where re-coding is needed. The quality of your original coding is therefore preserved, and only a small amount of review is needed.

Return to top

36. What makes your versioning service unique?

Our versioning service is unique in a number of ways:

• We have proprietary versioning software, developed from our long-term and extensive use of MedDRA and WHO Drug, to facilitate the versioning. Our coders utilize this special software to quickly and precisely pinpoint only those reported terms that need to be re-coded. You get the effect of completely re-coded data without our having to actually re-code all your data.

• You get more than just reports. You get actual data completely migrated to the new version.

• Your data is completely versioned, including instances where a newly added dictionary entry is a better choice for your previously coded data.

• The versioning is first done within our systems, so there is no impact on your ongoing operations.

• We offer versioning for both MedDRA and WHO Drug.

• We can version between any two versions.

Return to top

37. In theory, it sounds like a good practice to keep up-to-date with each new version. But it seems like there will be a huge impact on my data each time I version. Is it practical to version on a regular basis given the impact?

A common misconception about versioning is that a great deal of the previous coding will change. The reality is that only a small percentage (typically two to five percent) of reported terms will have a change in coding. Our versioning clients ultimately find that there is little overall impact as a result of versioning and that the small number of changes is very manageable. The advantages gained from having only one version of each dictionary in use and the ease by which data can be pooled usually far outweigh the minimal cost and effort required to version.

Return to top

38. Are there any startup costs?

We do not impose any startup costs, nor do we require a deposit or retainer. Since there is no up front investment required, there is no risk to try out our services.

Return to top

39. How do you charge for your services?

Coding is charged per unique term per dictionary and version, regardless of the number of times that term appears in your data and/or is submitted to us for coding. In other words, we only charge the first time we code a term. If that same term is submitted to us again (for example, reported multiple times in the same trial, or reported in other trials, or additionally reported as a serious adverse event in drug safety), we return the coding to you without charging you again.

Versioning is charged per unique term migrated between a pair of dictionary versions. The per term fee for versioning is substantially less than the per term fee for coding.

Our training classes are charged a flat fee per class, regardless of the number of people attending the class.

Consulting services are charged an hourly rate. The rate may vary depending upon the specific ThesIS employee performing the service.

Return to top

40. What is included in your per term charge for coding? What sort of "extra" charges should I anticipate?

Generally speaking, our per term charge for coding is all inclusive. The charge includes:

• setting up your company-specific area within our thesaurus system
• receiving your terms to-be-coded and loading them into our system
• the coding itself
• ongoing reviews of the coded terms for accuracy as well as consistency
• discussions required during the coding to address problems areas and coding guidelines
• sending the coded results to you
• follow-up questions you may have after the coding has been completed and returned to you

As you can see from the above list, there are usually no "extra" charges that are imposed. Unlike some companies who charge a deceptively low per term fee and then impose "hidden" charges later, we prefer to quote you a per term fee that truly reflects your actual cost.

Furthermore, we want you to fully understand and be completely comfortable with the results we provide you. You can ask questions and/or discuss the coding whenever you feel necessary without worrying about a charge being imposed each time you call or e-mail us.

Return to top

41. How much travel will be required to use your services? How much should I budget for travel expenses?

There is usually no travel required to use our coding and versioning services. Any discussions and/or questions can almost always be resolved using e-mail and/or over the phone.

Return to top

42. Are there any volume discounts available?

Yes, volume discounts are available. Please contact us with your specific details to determine the discounts that we can offer to you.

Return to top

43. How can ThesIS lower the cost of my thesaurus operations?

Using ThesIS for your coding and versioning can save you money because:

• With your own coding staff, you must always balance the ability to handle peak loads against the cost of your coding staff being idle at non-peak times. Outsourcing your coding to us means that you never pay for underutilized resources and yet never suffer the costs of not meeting your timelines.

• You pay only once per unique term, regardless of the number of times that term appears in your data and/or is submitted to us for coding. In other words, we only charge the first time we code a term. If that same term is submitted to us again, we return the coding to you without charging you again.

• You no longer pay for redundant coding due to multiple CROs coding the same terms.

• As your single source for coding and versioning, we provide consistent coding across your entire company, thereby reducing the effort and costs needed to pool data and/or to reconcile your clinical and safety data.

• Data is coded correctly the first time, meaning no costly delays in timelines or filings (and no need to pay again to have your data re-coded).

• You do not need to have your own thesaurus system. The cost for ThesIS to code and version your data is usually significantly less than the cost for you to license, install, validate, maintain and support a thesaurus system, let alone the cost of the staff needed to code within that system.

Return to top