Only four of the GPs stated that they regularly discontinued medication. The majority of the respondents reported that discontinuation was rarely done, and as a specific practice they did not consider it to be very organized. As one GP noted:
We have always been bad at discontinuing something I think. I’ve, you know, been a doctor for 14 years now and I think that we become better at discontinuing some of it, but we often talk about that we don’t really do enough about it [discontinuing]. [GP17; p.3]
This description of discontinuing medication as something often talked about, but rarely done, was frequent among GPs.
There were at least two reasons for why discontinuing medication was rarely done. First, the routine of prescribing was so strong that a concerted effort is needed to even raise the possibility of discontinuing medication:
It’s clear that, when discontinuing medication, as soon as it’s in the system, you don’t stop any medicine, unless you yourself implement some systems that flag it up [as an issue to discuss]. [GP20; p.6]
And the GP sometimes forgot to discuss the possibility of discontinuation with the patients, especially if the patient did not mention it:
Sometimes you continue without thinking about it. The patient doesn’t come in and ask [for discontinuation]. They ask for the same medication. And you think they benefit from it. [GP6; p.4]
Second, it was difficult for the GPs to identify the right time to discuss discontinuation with the patient:
Because you can say she has been in the clinic for a lot of other things in between [the first prescription and the current check up] and you don’t sit and think, her cholesterol is looking good and therefore I suddenly want to look 12 years back in the journal to find out if the original indication was justified or not. [GP23; p.8]
Thus, a mixture of a strong culture of prescribing and the difficulty of identifying relevant patients for discontinuing medication led to a sense of discontinuing medication being challenging to do.
The role of cues and dissonance for discontinuing medication
Although deliberations about discontinuing a patient’s medication could come about in many different ways, such deliberations tended to occur more frequently in certain situations; e.g. consultations that involved a check up (annual chronic disease check up, cholesterol check up, nursing home check up) or consultations with new patients (from another practice). Such situations were characterized by being more likely to generate what can be called cues of discontinuation. A cue of discontinuation is something that attracts the GP’s attention toward the possibility of discontinuing medication. If enough attention is drawn to the possibility of discontinuation, then a dissonance situation is created, where a question mark is put by the necessity of the prescription.
Patient based cues are cues emerging from the patient that the GP would not be able to pick up on his/her own, without the patient. Examples include: a patient complains over side effects (muscle pain, depression) or a high drug burden, a patient makes an explicit wish to stop medication, a patient shows concern over media reports about problems with specific drugs:
I have discussed it as recently as today with a patient that takes a lot of medicine, who has also had some different symptoms and she was very tired of taking so much medicine. She didn’t ask about statins specifically, to Simvastatin, which she took, but the idea of reducing her statin was a good thing for her. Also because she had so many strange things with her body, that may actually be related to statins. [GP22; p.1]
Record based cues are data related cues that can be extracted by the GP without the patient being present. Examples include blood test results, patients’ age, number of drugs prescribed, history of CVD. For example, one GP said:
A healthy woman aged 87 came to me, she’s on a statin, and she’s changed her GP [to me]. She gets simvastatin and she has no diabetes, no ischemic heart disease and a cholesterol score of 5,2 [the threshold is 5]. [GP10; p.9]
Here a variety of record based cues, accessible from records independent of the patient, prompt the GP to consider discontinuing the statin.
Some cues might singlehandedly trigger considerations about discontinuing medication, e.g. patient changes GP practice, whereas others might act in conjunction with other cues, e.g. a patient mentions concern about side effects and the GP then looks at the record and sees no previous history of cardiovascular disease (Table 2).
Organising the discontinuation of medication
Despite the rarity and difficulty of discontinuing medication, there were some GPs who attempted to organise discontinuation more systematically, and they did so in two ways. The first was to actively schedule check-ups and monitoring for a variety of their patients with a need for reviewing their medication situation. GP 9’s strategies included visits after hospital discharge and planning hospital admissions (when possible):
I mean, when a patient is discharged from the hospital, then I get a discharge summary and then within a week I actually visit them. I mean, if I choose to come within a week, then I’m able to stop most of it [unnecessary medication] before the home nurse starts giving it for a long time. So that way I’ve got a grip on it. And I also think it’s pretty fun. [GP9; p.8]
Here the GP is proactive in identifying the patient’s medication needs and examining the necessity of any new prescriptions. No special reimbursement was available to the GP for this extra work in term of identifying and visiting relevant patients (though new tariffs have since been introduced for check-ups with vulnerable patients). However, the GP saw the process as a stimulating professional challenge, which presumably contributed to his willingness to perform these systematic checks.
The importance of having this routine systematisation of monitoring was emphasized by another GP:
It has something to do with the routines that we have [in the practice], where we have a clear agreement between us, that such and such prescription renewal MUST come in for a check-up or they get a bit now, and know that soon in the future they have to come in for a check-up. This is for select areas, where we know there are problems. [GP14; p.9]
If a practice does not proactively invite the patient in for a check-up (and often the responsibility of represcription lies with the secretary) then it is difficult to identify and monitor patients that may benefit from discontinuing medication. This shows the important role of other actors besides GPs and patients, as well as the level of organisation needed to systematically consider discontinuation.
Prescriptions are supposed to be renewed every three to four months in an effort to continually assess the drug’s relevance for the patient. However, renewing the represcriptions often occurs without the patient. So despite there being a theoretical control of medications one GP noted that the reality is not always as such:
They need to, you know, have their prescription renewed, so you get a hold of them at some point or another. I mean, people that get statins need to be checked. In principle they should be checked every fourth month, I think it says in the guidance, for their liver function. So they come to the check-up of the statin… But, I mean there are also a lot that continue to take them [statins] for years without being checked [GP24; p.16]
In other words, organising the discontinuation of medication by scheduling check-ups is necessary, but not sufficient in itself. It also requires social interaction - the patient must be there.
The second way of organising discontinuation was to elicit more explicitly the patient’s experience of taking the statin, as opposed to relying on the patient to self-report any important information.
It makes no sense, if you… can’t lift yourself and sit down, because you’re on a statin. It is incredibly different how people react, so it is so, so important that you follow up, especially after you’ve started something, to see how big a burden it might be. [GP1; p14]
A patient’s negative experience with taking statins was often an important factor in the GP’s decision to discontinue. However it was often difficult to predict which patients would suffer from harms when taking statins or which patients would experience their burden of drug treatment as ‘too high’. Therefore the patient’s experience of taking the drug was important information for the GP after treatment started. A difference between GPs was observed in how actively the patient’s experience was elicited. Some GPs preferred to rely on the patient to schedule their consultations after experiencing harms, or to report it themselves at check-ups. Other GPs were more systematic in following up and eliciting patient’s experience with medication, including the effects of individual drugs, the overall drug burden and patient’s fears or concerns about taking a drug.
Responding to ambiguity
Occasionally there were instances where discontinuing medication was the obvious thing to do and where the GP would simply discontinue the statin. For example, one GP had a new patient from another practice and saw her for a second consultation. She took a lot of drugs and they had been discussing her experiences with this. She felt the burden was high, and that there were side effects from the statins. She had no history of cardiovascular disease and her cholesterol scores were not high. The GP commented:
I mean in my world I would discontinue straightaway and see what happened. She has as I said not had any infarctions and so on. So it will be an obvious thing for me to do, to discontinue that. But then she is an easy case. [GP22; p.3]
As the final sentence indicates, such clear-cut situations were rare and the majority of decisions around discontinuing medication involved responding to some level of ambiguity.
A GP elaborates on the challenges of negotiating conflicting sources of information and evaluating the effects of the treatment on the patient:
It is, you know, like a jigsaw puzzle because I actually think that it is really hard [knowing when it is best to discontinue], because I’ve taken it up a lot of times with many different patients, exactly the issue of an isolated raised cholesterol level, and when you should begin treatment. And there’s no-one that agrees, is my sense… it’s a personal call, and yeah, we know about all that risk stuff, and that it is not completely OK [to use SCORE risk charts for persons over 65] … but we don’t have anything better and… [sigh].. I just think it’s hard. I don’t think it’s as black and white [as guidelines say], because who says that that is a good solution for the individual patient? [GP19; p.13]
For the GP, the lack of agreement and conflicting information about when to prescribe or discontinue the drug, made it hard to decide with confidence what was most appropriate for the patient. The GP noted that tools like the SCORE heart risk colour chart was often used to reduce the ambiguity with elderly patients, even though she knew it was not appropriate for people over 65. For the GP, these difficulties with determining the best course of action showed that the clinical reality was more ambiguous than suggested by the guidelines.
A common response was to see this ambiguity as a reason to continue the treatment. Because of the risk of an adverse event happening after discontinuing medication, the safer option was to simply continue prescribing, thinking that it was better to have tried giving treatment than not giving treatment, if something went wrong. There was a fear that:
…you overlook something or make a mistake, and so you’d better do as they say in all the recommendations [guidelines]. I know so many that… I mean if there is a guideline, and it says that we are in the yellow field here, and that treatment should get started, then you start [prescribing the drug]. I mean in that flowchart [for cardiovascular risk] it can be yellow or red. If you are in the yellow field, then you start [treatment], because then you at least haven’t done something wrong. Because if they [the patient] had a blood clot and you haven’t prescribed anything, then you can’t [protect yourself and] claim that they were at least being prescribed [the drug]. [GP19; p.25]
In contrast, there were GPs that saw ambiguity as a reason trying out discontinuation for a short period followed by a check-up consultation. One GP elaborated on the advantages of pausing, rather than discontinuing straight away:
So I say to the patient, I think we should just have a break, and making use of a break to see if there are any changes. But then it means that there is usually a new decision after the break, you have to decide whether to continue the drug again or whether it should be discontinued. So in that way, a break amounts to two decisions, where discontinuation demands just one, but that decision is just somewhat bigger. So a break is a good way to start discontinuing you could say. [GP20; p.11]
Trialling was a way of practising medication discontinuation without psychologically committing oneself indefinitely to the process. Rather than one, larger decision it became two smaller decisions. A pause of one to 6 months generated more information (e.g. does the patient’s muscle pain disappear or does the cholesterol level shoot up) and provided an opportunity to revisit the decision and reflect on its appropriateness. Thus, trialling allowed GPs to generate more cues if they felt they had a good reason to consider discontinuation, but not enough information to be certain. In sum, it was an acceptable way of trying out discontinuing medication without being ‘committed’ to the action, because it was just a ‘trial’.