In total, 37 interviews with 37 respondents and an average duration of 53 minutes were conducted between January and December 2018. Of the 37 interviews, 32 were held with one respondent, four with two respondents, and one with four respondents. Of the 37 respondents, 5 were interviewed on multiple cases (counted as separate interviews). Most interviews were administered by one interviewer (AR). In four interviews, a second interviewer joined (GdW or N.J.E. van Vooren) because of the number of respondents taking part in the interviews. A timeline displaying the reimbursement status and the policy reports published on the case was brought to the interviews to stimulate remembrance and to support the conversation (see Additional file E). Table 1 shows the number of interviews that was conducted per case and per respondent group. In four out of five cases, either data saturation was reached and/or all eligible actors were interviewed. In the quit smoking program case, not all eligible actors were interviewed as we did not succeed to include a politician who was involved.
Table 1
Number of interviews conducted per case and per respondent group
Case
|
Respondent group
|
Total number per case
|
Patients
|
Healthcare providers1
|
Health insurers
|
National Healthcare Institute (ZINL)
|
Ministry of health, welfare and sports
|
Other
|
Benzodiazepines
|
1
|
2
|
1
|
2
|
2
|
12
|
9
|
Medication for Fabry disease
|
1
|
1
|
1
|
1
|
1
|
13
|
6
|
Quit smoking program
|
1
|
04
|
1
|
1
|
1
|
35
|
7
|
Psychoanalytic therapy
|
1
|
36
|
1
|
1
|
NA7
|
12
|
7
|
Maternity care assistance
|
1
|
38
|
1
|
1
|
2
|
0
|
8
|
Total over all cases
|
5
|
9
|
5
|
6
|
6
|
6
|
37
|
1The respondent group healthcare providers includes both (leading) individual healthcare providers as well as healthcare provider organizations (e.g. physician associations)
2i.e. knowledge institute
3 i.e. pharmaceutical company
4 In this case, several interviewees are from interest groups that represent both patients and healthcare providers as well (i.e. different actor groups are united in one organization). Therefore, the healthcare provider perspective has been covered in the other interviews.
5 i.e. interest groups + advisor of interest groups
6 In this case, in the first interview with a healthcare provider, we identified two even more involved healthcare providers. Therefore, in total, three interviews with healthcare providers were necessary for this case.
7 Disinvestment decisions based on (lack of evidence of) effectiveness can be taken by the National Healthcare Institute and do not need involvement by the Ministry. For this reason, the Ministry was not involved in the disinvestment decision regarding psychoanalytic therapy, and was, hence, not interviewed.
8 Several healthcare provider organizations are involved in maternity care assistance. Therefore, we needed three interviews with healthcare providers to interview all relevant healthcare provider organizations.
|
The remaining part of the results section is structured around the five main aspects influencing the outcome of disinvestment processes, as identified in our analysis. Here, the differences and similarities of these themes between cases are discussed. Additional file F contains a description of these themes for each case, separately.
3.1 Theme 1: Support for active disinvestment and pressure exerted
In our interviews, respondents described that the degree of support for disinvestment from actors, including the public, was important for the outcome of the disinvestment process. However, the importance of support differed between actors and cases (see Additional file F and the sections below). Respondents from all cases described that the degree of support for disinvestment was related to actors’ views on the case (i.e. health problem and intervention/service concerned), and in some cases to the attention for and framing of the case in the media. Respondents described that for them, as well as for other actors, the degree of support influenced the actions undertaken, and, subsequently, the amount of pressure exerted. From a comparison of the interview data from the studied cases, we found that the pressure exerted had a large effect on the outcome of the disinvestment process: limited pressure was exerted in the cases that were actually disinvested, while much pressure was exerted in the cases for which reimbursement was maintained. In the following quote, a healthcare provider describes the effect of societal pressure on the outcome of the disinvestment process.
Healthcare provider, case D, reimbursement maintained:
“Yes, I believe that the social pressure, especially all media attention, including emotional arguments, helped a lot to reach this conclusion [not to stop reimbursement] in the end.”
3.1.1 Support from healthcare providers
Respondents described that especially the support from healthcare providers was important for the outcome of the disinvestment process (i.e. has an important effect on the decision that policymakers make):
Interest group, case C, first disinvested, later reimbursed again:
“Respondent: And [the minister] actually had less of a say in that, I think. I think that the social pressure, particularly on this item from the healthcare sector, which she ultimately needs much more, also for the rest of her policy, that it is higher.
Interviewer: In that respect, surely the lobby and the calls from the healthcare sector were very important in this?
Respondent: To reverse it [the decision], yes, definitely.”
By comparing the interview data from the studied cases, we found that the cases where healthcare providers successfully exerted pressure against disinvestment were not-disinvested, while cases where healthcare providers did not exert pressure or were unsuccessful in exerting pressure were disinvested. According to respondents, whether healthcare providers were successful in exerting pressure depended on 1) their willingness to exert pressure, subject to a) their level of support for disinvestment of the case, b) whether they feel it fits their role (e.g. whether they feel they have an advisory role which does not comply with exerting pressure) and c) their (financial) interests, and 2) the possibility they had to exert pressure, subject to a) their awareness on how to exert pressure and b) the opportunities provided to them to do so. In the first quote below, a healthcare provider from case B explains that they did not exert pressure because of their believe that disinvestment would improve healthcare delivery (i.e. other healthcare providers would better follow guidelines), suggesting their support for disinvestment. In the second quote a healthcare provider from case A describes his/her difficulty in convincing other actors to join forces (i.e. being successful) in exerting pressure.
Healthcare provider, case B, partially disinvested:
“And, and so in that, up to that point we were like yes, if it actually helps to have it paid for [by patients], […] and if that helps them [other health care providers] to follow our guideline better, and to give less [of the treatment], yes, then it is fine of course.”
Healthcare provider, case A, fully disinvested:
“Interviewer: Was there any party that you managed to get on board?
Respondent: No, not really. They did [...] I can tell it like it is. At least people paid lip service to this whole story. People understood and supported the arguments. But to say that [name] who was then president of [organization] was like 'will you state once and for all that [case] is indispensable in the Netherlands and at any cost... ', [name] never did that.”
3.1.2 Support from governmental institutions
From our comparison of the interviews, we found that the opportunity provided to actors, including healthcare providers, to exert pressure seemed to depend for a large part on the opportunity actors were given by governmental institutions, who are in charge of disinvestment processes, to be involved in the policy process. We observed large differences between cases in the opportunity respondents were provided to be involved in the policy process, which seemed to depend on the support from governmental institutions for the case at hand. This is illustrated by the quotes below describing the opportunity respondents felt they were provided to be involved in case E, where policymakers described they were not in favor of disinvestment, and case A, where respondents suggested that policymakers were in favor of disinvestment.
Healthcare provider, case E, reimbursement maintained:
“Interviewer: But they [ZiNL] were willing to listen to you in any case?
Respondent: Definitely.
Interviewer: And did they also accept them, um, did they agree with the arguments you put forward?
Respondent: Yes, they did. But I can still remember that they, and it's of course very good that they always wanted to investigate for themselves, were like “okay, we can't accept everything just like that”.”
Patient organization, case A, fully disinvested:
“And well, they [ZiNL] are always going to mention in their advice that they have spoken with – for example, right – in our case [organization name] with so and so. Yes, I can put that into perspective. Those five minutes that it took for us to quickly hand over a box of signatures, well, yes, even if it was a bit longer, that's not really a conversation, is it? There's no way you have really heard us.”
From our comparison of the interview data between cases, we found that the support from governmental institutions also seemed to affect the use of the formal assessment framework in the cases studied. We observed differences between cases in respondents’ descriptions of the criteria that were used in the disinvestment process, how these criteria were interpreted, and how these criteria were weighted against one another. For instance, as is illustrated by the quotes below, in case A, where respondents suggested policymakers were in favor of disinvestment, the focus was mainly on whether the treatment was evidence-based, while in case E, where policymakers described they were not in favor of disinvestment, more considerations than strictly related to the four package criteria were taken into account in the policy process.
Healthcare provider, case A, fully disinvested:
“The motive was: we only reimburse things that are evidence-based. Well, a lot of what's in the package just isn't evidence-based so that's kind of a weird reasoning.”
Policymakers from ZiNL, case E, reimbursement maintained:
“You could say it was a very heterogeneous set of effect measures and because of that diversity and heterogeneity, well, it is sort of, it is kind of a multi-factor analysis, right, of all those factors. Are they pointing in the same direction? Do they all contribute to the goal that you want to achieve with [the case]? Well, and that answer was yes in the end.”
3.1.3 Support from patients
Respondents from cases A and B, which were actually disinvested, described that patients in these cases were very vulnerable due to their disease and the societal problems associated with it. Because of this, according to respondents, patients were poorly organized, which made it difficult to get a large group of patients to exert pressure against disinvestment, as is illustrated by the quote below. This lack of pressure exerted may have facilitated disinvestment in these cases.
Patient organization, case A, fully disinvested:
“Yes, for most people it is a taboo, a long-term [treatment] like that. Especially when you're still in the middle of it. So, it was difficult to reach the patients. […] It was almost impossible, it was a real taboo, there is so much going on when you get such a long-term [treatment].”
3.2 Theme 2: Compassion for current users
Respondents described that the degree of support from actors for disinvestment was affected by whether measures could be taken to ease the effect of disinvestment for patients currently using the intervention at hand. This is illustrated by the quotes from policymakers below in which they describe that they consider disinvestment much more difficult than not to start reimbursement in the first place, especially if disinvestment results in current patients having to stop their treatment or switch to another treatment
Civil servant from the Ministry, case D, reimbursement maintained:
“But you know, you are dealing with people who have been under treatment for years, um, and who are confronted with a new reality overnight. So, that also played a role, also for the Ministry. Like "yeah, you know, what is wisdom, you can't stop people’s treatment just like that".”
Policymaker ZiNL, case A, fully disinvested:
“And we said, "People should be allowed to finish their treatment." That was actually, there was some discussion about that. What were the arguments? I don't know. I think, I don't know how explicit that was but, yes, a reliable government, you started [the treatment] after all, yes.”
Respondents described that if measures could be taken to relieve the effect of disinvestment on current patients, such as stopping reimbursement only for new patients or only restricting reimbursement, they considered disinvestment much more acceptable. This indicates that the possibility to take such mitigating measures impacts the support for disinvestment from actors and, subsequently, the outcome of the disinvestment process.
3.3 Theme 3: Role in the health insurance system
From the interviews, we observed that the role appointed to actors by the Dutch health insurance system also affected the outcome of the disinvestment process. In the Dutch health insurance system, two groups of policy makers do advice the Minister on reimbursement (policy formulation): ZiNL is involved with reimbursement aspects with regards to content, while policy makers of the Ministry itself both focus at the content and context of reimbursement. Once the Minister decides on reimbursement, health insurers subsequently implement these reimbursement decisions, and patients/healthcare providers are both consulted in reimbursement decisions and in the implementation of these decisions. Every actor can be involved in agenda-setting, although policymakers often have the most distinct role in this phase. In our interviews, respondents described that for the cases studied, actors acted in accordance with the role they have in the system. They described that actors tended to stay away from pursuing actions that they considered to be beyond their formal role. For instance, in the quote below, a health insurer explains that they generally refrained from trying to influence policy formulation and decision-making on reimbursement, as they only consider the implementation of reimbursement decisions to be part of their role.
Health insurer, case D, reimbursement maintained:
“Actually, over the years, we've had less and less of an opinion about this sort of thing. Because we feel, you know, everyone has their part to play in the system.”
Although this finding was most distinct for health insurers, other respondents also regularly discussed their tendency to stick to their formal role. As was described in the interviews, if actors felt that exerting pressure was beyond their formal role, they generally refrained from this. Hence, the formal role of actors affected the pressure that was exerted by these actors in the cases studied and, subsequently, the outcome of disinvestment processes.
3.4 Theme 4: Financial interest in disinvestment
From the interviews, we observed that the actions actors took and, subsequently, the outcome of the disinvestment process were also affected by the financial interest of actors in the outcome of the disinvestment process. Respondents described that actors who had a financial interest in maintaining reimbursement were more likely to take action and exert pressure against disinvestment, than actors who did not have a financial interest in maintaining reimbursement, as is demonstrated by the quote below:
Healthcare provider, case B, partially disinvested:
“No, because all those [case intervention] were no longer, there wasn't a patent on them anymore, so they were... It wasn't in those manufacturers interest anymore to interfere. Otherwise it would probably not have been possible [to stop reimbursement]. Then they would have pulled out all the stops, to undo that.”
3.5 Theme 5: Role of the formal package criteria
From our interviews, no consistent pattern of the influence of the formal Dutch package criteria (i.e. effectiveness, cost-effectiveness, necessity and feasibility) on the outcome of the disinvestment process was observed. This finding is illustrated by the fact that our study included two cases (i.e. case A and D) that, based on what respondents shared with us during the interviews, seemed to score similarly on the formal package criteria, while having a very different outcome (fully disinvested versus maintained), a different outcome that cannot be explained by the cases scoring differently on any other criterion as well.