Skip to main content

Table 3 Themes, subthemes and representative quotes regarding the primary theme ‘Working with data’

From: Clinical researchers’ lived experiences with data quality monitoring in clinical trials: a qualitative study

Theme

Subtheme

Exemplar quotes

 

Coping with data errors

“We’ve tried to minimise any bias, or you know introduction of any errors. So, we’ve always had the same training procedure” (P4).

 

“We picked 1% probably arbitrarily…how much error would you begin to feel a bit uncomfortable about in terms of the capacity for seriously changing the reported outcome from a study” (P2).

 

“Just because with excel there were quite a lot of ways things could go wrong like formulas that are set up in several spreadsheets or even jumping a line or just entering a wrong number. Just doing a typo which is not always visible straight away” (P5).

 

Data audits

“We also, under various funding arrangements were subject to external completely independent compliance checks... we would welcome those and work very closely with the people doing it. We didn’t like them.” (P2)

 

Coping with missing data

“I guess it makes it easier definitely the electronic way to see what’s missing…and I think it will save a lot of missing data.” (P5)

 

“we would give them a call and ask over the phone, and usually we tried to do it within the um, within a two week period from the time we were supposed to have received it.” (P5)

 

“I know that they did manage, that they managed to manipulate the data in such a way that they did get an outcome, but I know I remember we were struggling with that. I remember talking at meetings about how we were, how the statisticians were going to manage that to, to be able to provide an answer.” (P3)

Data monitoring

Monitoring approach

“doing some regular check, plotting the data, doing some simple stuff. So, descriptive stuff very regularly. Where I just got minimum maximum, you know approach and plotting the data to check it. Nothing was really um out of the ordinary.” (P5)

“So, I have been involved in project they have they are very fussy about the data monitoring they have to check every day… probably back in the day its paper based…they didn’t check until the very end of the trial” (P6).

“when we say monitoring, we are going to actually start implementing a lot more statistical compliance monitoring in house so we can save on travel because we are [name] funded so we don’t have a lot of funds to send people away.” (P3)

“It depended on whether it was academic, whether it was commercial…investigator-initiated study, or an investigator sponsored study or a commercially sponsored study and what the aims of the study were” (P3).

“I found that it varies from project to project and also ah even within the same setting ah you know different projects different research team um depending on their size may have different factors.” (P6)

“Well, look um I am going to be sort of bold here and say, it’s never really has been different [data monitoring in the academic environment].” (P2)

“I’ve always kinda taken the same approach in monitoring data quality” (P5).

“We had to submit…a monitoring plan…this actually should have been submitted with the protocol, but we didn’t know at the time” (P7).

Assumptions or opinions

“as opposed to a smaller um, experiments I guess where things, well things there’s no set date I guess and things can change at the start before you can do a lot of pilots I suppose before you start your ah your real data collection…there’s more freedom I suppose in changing things before you actually start” (P5).

“They’re not a complicated study it’s not like a drug trial. Drug trials are the ones we have all those sorts of trouble” (P1).

Data quality

Elements of quality

“Unless you could substantiate claims about data integrity and reliability you really might as well not bother” (P2).

“I remember even her [boss] saying ‘you know we don’t want to leave all this evidence around, sponsors to be looking at um and seeing that there’s of lots of dirty data sitting outside’ I don’t know if anyone else has said that to you but it’s something that has always stuck in my head, I always thought it was very interesting.” (P3)

Factors influencing data quality

“the first thing you’d realise then is that CRFs would often lay around uncompleted for considerable periods of time and then there’d be a rush to fill them in before people arrived or they were due to be sent and inevitably when you allowed time to elapse between a clinical assessment and the forms being filled in there’s much greater chance of there being mistakes and errors.” (P2)

Reporting data queries

“They [data collectors] are aware of the values they should be getting and what will be…outliers and they’re also I guess required to document everything [if] they think something strange happened during the visit and getting some odd values for an assessment. They take notes about what they think could be the cause for that at a later stage we can understand why this score will be an outlier.”(P5)