This qualitative study generated a substantive theory about the development of MOH. The basic process leading to medication overuse was holding on to the indispensable medication. The acute medication was indispensable to the participants because they perceived it as the only thing that could prevent headaches from ruining their lives.
The perception of headaches as a threat to quality of life is consistent with previous research. Quantitative studies have shown reduced quality of life in MOH [3–5], and other qualitative studies have described similar patterns of disability in relation to work, family and social life to those found in this study [27, 36, 37]. The participants were not passive in relation to this threat. They struggled to uphold their preferred lifestyles despite headaches and invested substantial resources into finding strategies to deal with them. Being actively involved in the management of headache has been reported previously . Peters et al.  described active involvement through both decision-making and behaviours. In a study of functional coping, Lauwerier et al.  found that those who primarily focused on pain as a problem to solve were at higher risk of developing MOH than those who tried to disengage and focus on other areas of life. The participants in our study also focused on their headaches and invested a lot of effort searching for ways to manage them. In a recent paper, Lauwerier et al.  suggested that efforts to control pain may be regarded as attempts to protect valued life goals that are threatened by pain. This could explain why some engage excessively in pain control strategies, such as medication overuse, despite the costs associated with this, such as the development of MOH.
Choosing acute medication as the main strategy to master headache could be regarded as choosing an easy alternative. It requires less effort than many other strategies, e.g. lifestyle changes and therapy. However, the results of this study show that use of acute medication was not the participants’ first choice. They had put a lot of effort into trying to find other strategies. The range of strategies used was similar to that found in a previous qualitative study . Some of the treatments tried lacked scientific evidence but they had also tested treatments that are recommended in official guidelines, such as prophylactic medication and psychotherapy [39, 40], without experiencing improvement. In this study, we did not go into the clinical reasons as to why these strategies had not been effective. The participants perceived them as ineffective and consequently, they eventually became resigned, accepting acute medication as the only effective aid. Their reliance on acute medication was thus not a convenient quick solution to the problem; it was rather the only remaining alternative after having tried everything else.
The participants were sceptical about prophylactic medication because they did not like the idea of having to medicate daily. A reluctance to use daily medication has been observed in previous research, e.g. among asthma patients . In the case of MOH, this notion is particularly interesting since persons with MOH are already using acute medication more or less daily . When asked about this, the participants explained that the acute medication was indispensable. Since they were already using so much of this medicine, they felt it was not a good idea to add yet another medication, i.e. the prophylactic. This implies that they somehow viewed the prophylactic and the acute medication as the same thing, i.e. a medication that was harmful and ought to be used as little as possible. By regarding it that way, it is not surprising that they held on to the acute medication rather than the prophylactic. The acute medication had a more obvious effect and only had to be taken when needed. This is an example of how the perspectives of the individual medication user can differ from the traditional medical view. The first step towards successful use of headache medication is probably to bridge the gap between the perspectives of patients and health-care staff.
Lack of information may partly explain why some did not regard their medication use as a causative factor behind the increasing frequency of headaches. A few had never heard of MOH and this implies that there is an unmet need for information concerning this disorder among those at risk of developing it. Further, since we did not try to detoxify the participants, we cannot rule out the possibility that the chronic daily headache of some was indeed caused by factors other than medication overuse. However, despite these possibilities we found it noteworthy that participants who were aware of MOH did not necessarily view the medication as a cause of their own increasing headache frequency. This is somewhat surprising, considering the effort they reported having made in searching for explanations and strategies to treat their headaches. A few participants talked about the association, but our data did not explain why some were aware of it whereas others were not. Realising that medication overuse may be contributing to increased headache seems important for the successful treatment of MOH and thus more research on this stage is needed.
There is an on-going discussion as to whether MOH should be considered an addictive disorder or not [19–23]. The participants in this study expressed that they did not view themselves as addicts and that they felt offended if someone suggested that they were. However, as addiction is sometimes associated with denial, it is difficult to draw conclusions about addiction from this study. An important difference between those with MOH and those with addiction seems to be the reason for the overuse. Addiction is often characterized by a progressive neglect of alternative pleasures or interests because of drug use and may result in a reduction of social, occupational, and recreational activities [8, 21]. This is usually not the case in MOH. Instead, both this study and previous research suggest that persons with MOH are overusing the medication to live their lives as normally as possible and reduce the impact of their disorder on their daily lives [8, 21]. The participants held on to the medication to prevent the headaches from ruining their lives, not because they wanted the medication per se.
Several studies concerning decision-making among persons with headache suggest that they actively weigh the pros and cons of taking the medication before deciding whether to medicate [9–11]. This corresponds with the beliefs-about-medicines model, concerning chronic illness in general . It hypothesizes that patients engage in an implicit risk-benefit analysis in which beliefs about the necessity of a medication are weighed against concerns about its potential risks. In the case of headache, taking the acute medication is beneficial because the attack is aborted but it also leads to risks in terms of the potential development of MOH. If applying the model strictly, one would expect the risk-benefit analysis to lead to decreased medication use when such negative effects prevail. However, this is not the case in MOH, where many seem inclined to overuse despite being aware of the negative consequences [8, 42]. Even after successful withdrawal treatment, often consisting of thorough patient education, the relapse rate is around 30% [6, 43]. The model presented in this study provides possible explanations for this behaviour. The perception that headaches are threatening to ruin one’s life and that there are no available solutions other than the acute medication could tip the balance so that the benefits of taking acute medication outweigh the risks. Further, the fact that the participants did not necessarily keep track of their medication use nor think about it as something that contributed to increased headache, probably made it more difficult to conduct the clear-sighted kind of risk-benefit analysis, described by the decision-making models [9–12].
The participants had varied experience and insight into the phenomenon of MOH. Some did not know that MOH existed, others knew about it but did not think that it was the underlying cause of their own increasingly frequent headaches, and some acknowledged that their headache was indeed MOH. Despite this variation, the theoretical pattern relating to the core category applied to all participants. The variation added richness to the theory.
A limitation of the study is that all participants were recruited via advertisements and that we thus only recruited persons who had taken the initiative to talk about their situation. This may e.g. have led to a selection of MOH sufferers who were active and open and thus reinforced the impression that persons with MOH are actively searching for new treatments and new information about their disorder. Interviewing other persons with MOH may possibly have given another picture of the problem. The proportion of participants with a university education was higher than in the general population. However, the external validity in qualitative studies focuses of transferability rather than generalization [44, 45] and even though some examples in the data are specific to the participants’ context, they generally expressed the importance of the medication for preventing their disorder from disrupting their lives. This finding may be transferred to persons using headache medication in other settings and even to persons using medications for other disorders.
Analysing the data with another method, such as phenomenology, content analysis or narrative research, would most likely have given the results a different shape as the methods have different theoretical underpinnings and pose different types of questions. In this study, grounded theory was considered the most suitable method, as we wanted to analyse a process. A potential risk with grounded theory is that researchers may allow a preconceived theory to direct the sampling of data and the analysis, thus seeking to verify a preconceived theory rather than finding a new one. However, the structured procedures for data collection and analysis, including the constant comparisons and the asking of questions, are there to prevent such bias . In fact, the methodology emphasises that the theory should be grounded in the data and forces the researcher to constantly redefine the emerging theory as new data is included. In this study, the final theory was very different from the first embryos of the theory that were produced early on during the research process. The multidisciplinary group of co-authors had constant discussions throughout the analysis process in order to prevent preconceptions from affecting the developing theory. Further, the emerging theory was discussed in a multi-disciplinary research seminar and the regular peer scrutiny applied (by TH and AJ) throughout the analysis also adds to the credibility . With these procedures we have done our best to prevent the influence of preconceptions but, as in all qualitative research, the risk can never be entirely eliminated. The model is new and unique, but the essences of several categories are supported by other studies, and this strengthens the credibility of the findings .