Osteoarthritis (OA) is a chronic musculoskeletal disease of the joints that has a negative impact on the healthy aging of the population.1 This chronic condition produces disability resulting in significant costs to the society.2 It is estimated that in Mexico 10.5% of the people with musculoskeletal pain have OA,3 the knee being the most commonly affected joint.4–6 The high prevalence of OA within the Mexican population produces a high demand for healthcare, representing one of the 10 most common problems seen at the primary level of care in the Mexican Institute of Social Security (IMSS), one of the main public health institutions of the country.7

In Mexico, people with OA are usually managed by general and family physicians within the primary healthcare system,4 representing one of the principal sources of cost for this level of care.8 Consequently, there is an interest in containing the high social and economic costs produced by this chronic disease through development and implementation of clinical practice guidelines (CPG) that standardize management across the country.9 Clinical practice guidelines are systematically developed statements or recommendations that assist in the decision-making process within healthcare.10 In 2008, the Mexican Secretary of Health coordinated the development of a CPG for the management of knee and hip OA at the primary level of care with the main purpose of providing evidence-based recommendations to decrease the disabling effects of hip and knee OA in the Mexican population.11 The developers of this guideline used existing CPGs from other countries to structure their recommendations for practice.9

The CPGs used to create the “2008 Mexican Clinical Practice Guideline for the management of knee and hip OA at the primary level of care (2008-Mexican-CPG-OA)” have not been fully implemented within their own contexts.12–15 This situation raises questions about the direct transferability of the recommendations stated by the 2008-Mexican-CPG-OA to the Mexican context. According to the Knowledge-To-Action (KTA) framework, in order to successfully apply a knowledge tool such as a CPG, it is important to identify potential barriers to its implementation, considering the local context in which it will be utilized.16

The implementability of a CPG refers to a set of its recommendations’ characteristics that permit their successful conversion into actions.17 Only clear, concise, and actionable recommendations can be successfully implemented in clinical practice.18,19 In consequence, it is possible to assess the barriers to implementation of a CPG through analyzing the characteristics of its recommendations.

The Mexican Public Health system is formed by different institutions, such as the IMSS and the Secretary of Health (SS), each with its own government structures and procedural mechanisms. Even within each institution there are region-related structural and systemic differences. This situation underlies the complexity of the public Mexican Healthcare system, implying the presence of different healthcare contexts that could affect the implementability of the 2008-Mexican-CPG-OA.

As a result, the idea behind this study was to use the concept of “CPG implementability” to evaluate the recommendations proposed in the 2008-Mexican-CPG-OA. The main objectives were to evaluate the implementability of the 2008-Mexican-CPG-OA within different Mexican Healthcare institutions at the primary level of care in three Mexican regions (Northern, Central, and Southern) and to put forward some strategies to improve its successful implementation within clinical practice in Mexico. To accomplish this, we used the GLIA v2 instrument to: (a) identify implementation barriers for each recommendation of the guideline, (b) disclose differences on implementability issues among each institution and region, and (c) propose strategies to address the identified barriers.

The reviewers had been practicing family medicine for an average of 17 years (min – 7, max – 27) and dedicate 5–30% of their weekly clinical time to the management of OA. All reviewers were familiar with evidence-based practice concepts, and 4 reviewers were undertaking postgraduate studies in health research methodology. Two reviewers were aware of the existence of the 2008-Mexican-OA-CPG before starting this project; none of them had received training for implementing this guideline. Only one reviewer, from the Northern team, had taken formal training on OA management.

There was more moderate to near-complete agreement (n=81) than poor to fair agreement (n=41) in the assessment of each recommendation within all the teams (see Fig. 2). This trend did not change when considering the agreement of each team separately. Some calculated kappa coefficients were negative.

Fig. 2.

Frequencies of poor/fair and moderate/near-complete agreement between reviewers.

(0.15MB).

All teams agreed that the guideline clearly defined the target population and audience and has a format that allows for an easy identification of the recommendations. However, we detected important problems with the settings’ description, the guideline developers’ representativeness, and the recommendations’ sequence, internal consistency and conciseness (see Table 2).

The general and context-related barriers for implementation identified for each recommendation, along with their proposed solutions, are described in Tables 3 and 4. Thirty-one recommendations presented executability and/or decidability problems that make them difficult to be applied consistently to clinical practice. Eight recommendations were just statements which contained no actions, so they do not have an appropriate recommendation format. Only six recommendations were judged as appropriate to be applied consistently to clinical practice in their current format (see Table 5).

The three teams agreed that the guideline does not provide a clear description of how the strength of the recommendations was determined. Although the guideline mentions the strength for each recommendation, the criteria used to adjudicate this strength are not explained in the document. The nomenclature to describe recommendations’ strength is inconsistent; sometimes it is through letters and sometimes through numbers. In addition, information about how the guideline developers chose the recommendations from the different CPGs used to create this guideline is also not described in the document.

Twenty-five recommendations were identified with barriers related to the context of one or more of the institutions-regions represented in this project. The Northern team identified barriers mainly related to health providers’ beliefs and patients’ expectations. The Central team identified barriers related to the process of care and treatment availability. The Southern team identified barriers related to the process of care, and patients’ expectations. We also identified common barriers related to the context of the three institutions–regions represented in the study (see Tables 3 and 4).

This study uncovered some aspects of the 2008-Mexican-CPG-OA that can impede its successful implementation within the IMSS-Estado de México, Quintana-Roo and Yucatán, and in the SS-Morelos and Nuevo León. As a whole, the guideline presented problems of conciseness and clarity that negatively affect its credibility and application within the Mexican primary healthcare context. Only 6 of the 45 recommendations (14%) were considered to be implementable in a consistent way. We also detected differences and similarities in the identified context-related barriers for implementation among the three institutions and regions represented in this project. Finally, we were able to propose some strategies to increase the guideline's successful implementation within Mexican clinical practice.

Clinical practice guidelines synthesize and inform the best available knowledge to support decision-making processes in clinical healthcare. The developers of the 2008-Mexican-CPG-OA created this guideline not only to aid in the decision-making process of primary caregivers but also to “standardize national actions” regarding the management of knee and hip OA in Mexico.11 The standardization of clinical practice at a national level could be an ambitious goal for a CPG. The guideline presents some issues regarding its structure, recommendations’ clarity and developers’ credibility, all of which may affect its utility for supporting decisions, making the standardization of clinical practice within the primary care of Mexico difficult.

The guideline's structure does not allow for a clear understanding of the settings and the sequence in which recommendations should be applied. It has been observed that CPGs without comprehensible structure are not easy to implement during everyday practice.31 Furthermore, most of the recommendations stated in this guideline are neither consistent nor concise, and it is difficult to understand the required actions implicit within them. The literature shows that in practice, vague recommendations are significantly less utilized than recommendations that clearly state what to do.32 On the other hand, reviewers expressed concern about the guideline developers’ credibility, noting that some knowledge users and patients were not involved in the guideline creation. A literature review on OA management conducted in Mexico, concluded that management of people with OA should be multi-disciplinary, including rheumatologists, orthopedic surgeons and nutriotionists,33 none of whom were represented in the creation of the 2008-Mexican-CPG-OA.

Increasing evidence shows that involving patients and knowledge users in the development of a CPG improves its applicability to clinical practice.34,35 Consequently, we think it is necessary to restructure the 2008-Mexican-CPG-OA replacing ambiguous recommendations for behavior-specific statements that clarify the what, who, when, where and how of the intended actions.36,37 This restructuring process should also include representatives from patients, primary care clinicians, rheumatologists and orthopedic surgeons. We also suggest that the quick guide (see Table 2) should be restructured, ensuring that its content matches the contents of the complete guideline.

The vast majority of the guideline's recommendations (86%) presented executability and/or decidability problems, which implies great challenges for its consistent implementation to the Mexican context. Failed executability and decidability are often the result of vagueness in the description of the actions involved in a recommendation.38 This vagueness results in application inconsistencies that go against the standardization purposes of a guideline. Interestingly, the 2008-Mexican-CPG-OA recommendations preserved the original foreign CPGs’ format. This may suggest that problems observed in the implementation of some of these foreign CPGs in countries such as France,13,39 Canada,12 and the UK15 are due to executability and decidability issues.

Only two of the six recommendations judged to be implementable in their current form (see Table 5) address one of the eight actions recently proposed as the minimum standard of care for people with knee and hip OA, “to minimize individual risk potential for NSAIDs harms”.40 The Mexican CPG considers all of these standard actions of care. This suggests that a simple restructuring of the recommendations can potentially result in an improvement in the quality of care for people living with OA in México.

Apparently, the values used to define the strength of the recommendations in the 2008-Mexican-CPG-OA were based on study design. Nevertheless, it has been widely recognized in the literature that the strength of CPG recommendations should not only reflect the quality of supporting evidence, but also the values and beliefs of developers, considering the context.41–43 Moreover, it has been strongly recommended that the processes to determine the strength of a recommendation be described in a transparent way.19 Consequently, we urge the guideline developers to be more explicit about the procedures they followed to state the recommendations’ strength, so the audience can judge their validity.

The context-related barriers identified can be further categorized as individual, organizational or system related.18 The Northern team identified individual barriers, the Central team identified organizational and system-related barriers, and the Southern team identified individual, organizational and system-related barriers. This shows how the context could influence the implementability of a recommendation, supporting the argument that CPGs have to be adapted to local conditions.44 Furthermore, the identification of organizational and system-related barriers suggests that not considering local socioeconomic and political factors may result in the failure to implement the CPG.

We also found similarities regarding context-related barriers identified by the three teams, suggesting the existence of implementability issues in the whole Mexican public primary care health system. These barriers were related to individual factors (lack of family physicians’ skills for prescribing rehabilitation interventions or evaluating functional capacity, and patients’ negative beliefs about arthroplasty); organizational factors (insufficient time and inefficient referral processes); and system-related factors (absence of efficient primary care rehabilitation units, therapeutic pools, and therapeutic agents). These implementation barriers can be addressed through patient education and knowledge dissemination strategies for health providers. Moreover, organizational and system changes require negotiation with local, regional and federal health administration and policy representatives.

There were some issues during the application of the GLIA v2 methodology in this study. All reviewers had problems with the concept of flexibility. Flexibility is defined as “the degree to which a recommendation permits interpretation and allows for alternatives in its execution”.45 This seems to be opposed to the idea of “consistency” evaluated in the executability dimension. After some discussion, team members concluded that “flexibility” favors the implementability of a recommendation allowing for consistent tailoring and does not imply an “anything goes” approach. Moreover, all teams agreed that the recommendation's strength does not belong to “flexibility” but to the validity dimension of the GLIA v2. It was difficult for some reviewers to differentiate between conditional and imperative recommendations and for discerning which statements were not recommendations. These difficulties resulted in negative kappa coefficients during agreement analyses, meaning less agreement than expected by chance.29

In spite of this, all reviewers agreed that the application of the GLIA v2 instrument, although time-consuming, was very helpful to stimulate consideration about guideline implementation barriers, allowing them to reflect on their clinical practices. We think that the GLIA v2 methodology can be further improved by (1) clarifying the dimension of flexibility, (2) providing examples that demonstrate differences among dimensions, (3) providing examples that show the difference between conditional and imperative recommendations, and (4) adding criteria to allow for the identification of statements, which are not recommendations.

One of the limitations of this study is the small number of physicians involved in the guideline evaluation process, which affects the generalizability of the results. However, the 6 family physicians involved had experience managing people with knee or hip OA, representing the two largest public health institutions of Mexico across 3 large geographical regions. Moreover, involving physicians who were currently in research training ensured the quality of the evaluations and analyses. Another limitation is that we did not evaluate the actual success in the implementation of the recommendations at the primary level of care. Consequently, we deduced that the 2008-Mexican-CPG-OA is not consistently implemented in clinical practice due to problems with the executability and decidability of its recommendations. Obviously, this deduction needs to be further evaluated.

In conclusion, the 2008-Mexican-CPG-OA has important barriers to its consistent implementation within the IMSS of Estado de México, Quintana-Roo and Yucatán, and in the SS of Morelos and Nuevo León. We recommend that this guideline undergo thorough revision and restructuring to improve the clarity of the actions implied in each recommendation. We propose some strategies to facilitate this revision, along with some implementation activities to address individual, organizational and system-related barriers. These efforts will potentially increase the success in the 2008-Mexican-CPG-OA's implementation, improving the standard of care for Mexican people living with knee and hip osteoarthritis.

All authors declare no conflicts of interest.