Guidelines on uncomplicated urinary tract infections are difficult to follow: perceived barriers and suggested interventions

Urinary tract infection (UTI) is one of the most common health problems for which patients seek medical care. It is responsible for about 1% of all general practitioner (GP) consultations in the UK [1] and results in approximately 7.9 million physician visits in the United States annually [2]. In the Netherlands, UTIs rank 8^th on the list of most common reasons for visiting a GP and also account for 1% of all visits [3]. Most of these are uncomplicated UTIs, defined as cystitis in non-pregnant adult women. In the Dutch healthcare system, uncomplicated UTI is diagnosed and treated by GPs and rarely needs specialist care.

To optimize care concerning uncomplicated UTI, evidence-based clinical guidelines have been developed in several countries [4]. However, adherence to these guidelines has shown to be far from optimal. In a large study among a representative sample of general practices in the Netherlands, it was found that GPs followed the guideline with respect to the treatment of UTIs in 42% of the cases and that the level of adherence varied widely (0-95%) between practices [5]. A recent study showed that Dutch GPs treated UTIs according to the guideline in 50% of the cases [6]. In other countries similar levels of adherence regarding the treatment of UTIs were found [7‐9]. Reasons underlying GPs suboptimal behaviour are thus far poorly understood [6‐9].

An analysis of barriers to the implementation of a guideline is considered to be a first important step in improving guideline adherence [10, 11]. The barriers identified can subsequently be used to develop tailored implementation strategies. Little is known on how to translate the identified barriers into tailored interventions. Several studies have shown that the choice of a specific intervention in practice is not necessarily based on the analysis of barriers [12], but is often determined by personal preferences or familiarity with some types of interventions [10, 13]. Moreover, the target users of the guideline are usually not involved in selecting implementation strategies to improve adherence [14].

The aim of our study was to identify the perceived barriers to implementation of a national guideline on uncomplicated UTI among Dutch GPs and to explore interventions that could address these barriers. By conducting an in-depth analysis of barriers and possible interventions to achieving change among the target group, we aimed to provide useful suggestions for improving the care concerning uncomplicated UTI.

In this focus group study we identified the main barriers to the implementation of a national guideline on uncomplicated UTI perceived by Dutch GPs and explored interventions that could address these barriers. We found that the recommendations on both diagnosis and treatment were difficult to follow in practice and determined a specific set of barriers that needs to be addressed to improve adherence. Although GPs were aware of the recommendations, attitudinal and external barriers prevented them from following the recommendations consistently in practice. The care concerning UTI could be improved, if these barriers are sufficiently addressed. Several interventions for overcoming these barriers were suggested by the GPs, providing opportunities for guideline developers, implementers, and GPs in practice.

With regard to diagnosing uncomplicated UTI, one of the main barriers was that GPs disagreed with the recommendation because they argued the supporting evidence. Previous studies showed that adherence to recommendations based on scientific evidence is higher than to recommendations that are not supported with evidence [30, 31]. However, providing evidence-based recommendations in guidelines is not enough. More efforts are needed to raise awareness among GPs with the evidence supporting the recommendations and to convince them with strong arguments why they should change their current practice. Discussing the recommendations in peer review groups may be a useful method as the effectiveness of interactive small group education has been demonstrated [32, 33]. Since the barriers are mainly related to attitude, an educational program addressing GPs' attitudes in addition to knowledge transfer, may be particularly effective [34‐36].

Organisational constraints to performing dipslides in out of hour services were also mentioned as barriers. Some GPs perceived the use of dipslides in general as inconvenient and do not have a supply in practice. This is consistent with other Dutch studies showing that GPs hardly use the dipslide in case of a negative nitrite test, particularly in out of hours services [6, 37]. A suggested intervention is to adapt the recommendation to current practice, i.e. not using dipslides in out of hours services, which is more consistent with guidelines in other countries [4]. Another option, not mentioned in our focus group session, is to hand dipslides over to the patients and ask them to show it the next day in out of hours services (Saturday; Sunday) or to the own GP (Monday). Although the dipslide has high diagnostic accuracy, the guideline could also offer alternative options for diagnosis in specific circumstances. Improving the organisation and coordination in out of hours services by developing local protocols and agreements with hospitals was also suggested by the GPs.

One of the barriers to implementing the treatment recommendation on uncomplicated UTI in practice was a perceived lack of applicability due to local patterns of bacterial resistance. Bacterial resistance to commonly prescribed antibiotics in uncomplicated UTIs has been increasing in recent years [38‐40] and resistance patterns have been found to differ significantly between regions [41]. As a result, national guidelines may not always be regionally applicable. Although some regional variation in bacterial resistance in general practices in the Netherlands was reported in 2004 [42], up to date and conclusive evidence for the existence of such variation is not available. However, it seems useful to pilot test guidelines by systematically monitoring the regional resistance patterns. If there is strong variation, the recommendations in the guideline could be regionally adapted to specific patterns of resistance.

Another barrier perceived by GPs is that drug dosages recommended in the guideline are not always available at pharmacies. Some GPs did not want to prescribe drugs that need to be taken four times a day because of user inconvenience, and therefore do not prescribe the drug of first choice. It would be helpful if guideline developers consider the availability of drug dosages to optimize the implementability of recommendations. Negotiation with national pharmacy organisations may be helpful to reach these goals.

By focusing on the individual recommendations within the guideline, we were able to gain an in-depth understanding of the barriers and the interventions needed to address them. The use of a predefined framework of barriers to implementation triggered physicians to think about a broad range of barriers and potential interventions to improve guideline acceptance and guideline adherence [28]. Our approach appeared to be useful in exploring a wide range of barriers and potential interventions and had an educational effect as well [22]. Moreover, by involving the target group of GPs in exploring interventions to address these barriers, we expect that the feasibility and effectiveness of interventions will improve. These methods can be applied in implementation programs on a range of topics and in other settings as well.

A limitation of our study is that we organised only one focus group. The participants were motivated GPs and those with a positive attitude towards guidelines may be overrepresented. However, our sample of GPs does correspond quite well in terms of basic characteristics to the total population of Dutch GPs. In addition, by offering accreditation points to the GPs, creating an incentive to participate for less motivated GPs as well, we attempted to reduce this bias. Secondly, the number of participants in our study was limited, making it difficult to quantify our findings. However, our aim was to explore the relevant barriers qualitatively instead of quantifying their relative importance. As our sample seems to be representative in terms of basic characteristics, we assume having described a substantial variation in barriers and interventions perceived by Dutch GPs. An electronic survey among a larger sample of GPs will follow to quantify our findings.

We only included GPs in our focus group session, while guideline adherence often also depends on other staff members in general practice. Changing habits and routines of practice assistants may be as difficult as those of GPs. Specific protocols and educational sessions for practice assistants may be useful. Quality improvement programs, involving all practice staff, such as NHG Practice Accreditation (NPA), could facilitate this process [43].

Finally, our study was based on a Dutch guideline questioning the generalisability of our findings to other countries. Guidelines on UTI in different countries differ substantially, particularly concerning diagnosis recommendations [4]. For example, most guidelines do not recommend the use of dipslides. Interventions to address barriers regarding this method may therefore not be relevant. However, barriers regarding topics such as the organisation of care in out of hours services will be relevant to other countries as well, as management of UTI often happens out of hours. Problems with antibiotic resistance patterns and availability of drugs also apply to other countries. Regional pilot testing of the guideline may be useful in many countries, even in smaller ones. Moreover, our methods used to determine barriers to implementation among guideline users are applicable in other countries as well.

Despite GPs' awareness of the guideline recommendations, our study showed that several attitudinal and external barriers prevented them from consistently following the recommendations on uncomplicated UTI in practice. Guideline implementation could be improved if guideline developers and implementers are aware of the potential barriers and involve all relevant staff members in the implementation strategies. Educational programs addressing providers' attitudes in addition to knowledge transference, and improving the coordination and organisation of care, could improve adherence to the guideline on uncomplicated UTI. Involving the target group in selecting useful interventions to implement the guideline recommendations may improve the feasibility and success of implementation strategies.