Osteoarthritis is the most common type of arthritis in Western populations, and is a major cause of chronic musculoskeletal pain and mobility disability in elderly populations worldwide.1 The prevalence of osteoarthritis varies depending upon the diagnostic method used (clinical or radiological), the joint(s) studied, and the characteristics of the study population; nonetheless, it is considered one of the ten most disabling diseases in developed countries, affecting 10% of men and 18% of women over the age of 60.2 As with other chronic conditions, treatment of osteoarthritis is complex and involves a combination of pharmacological and non-pharmacological measures for its optimal management.3,4

Failure to implement recommendations contained in the available guidelines of any chronic disease is one of several factors contributing to clinical inertia. Clinical inertia is a leading cause of potentially preventable adverse events. According to Phillips et al.5 factors that could contribute to clinical inertia include the overestimation of the quality of the care delivered by the physician or the underestimation of the number of patients who need an intensification of pharmacotherapy. Additionally, other factors that could contribute are that some physicians lack the appropriate knowledge, tools and clinic facilities to deliver adequate care to patients with chronic diseases.5 As has been demonstrated in several randomised clinical trials and cohort studies, the deleterious consequences of clinical inertia for the patient can be significant, especially in certain chronic diseases such as diabetes mellitus or arterial hypertension.6–8 Unfortunately, information on the impact of clinical inertia on other diseases with low mortality but high morbidity, such as osteoarthritis, is limited or non-existent.

We carried out a cluster-based, multicentre, prospective, randomised, parallel-group study to evaluate whether patients with hip and/or knee osteoarthritis would benefit from a specific type of intervention received by their general practitioners, which was designed to reduce clinical inertia. Secondary objectives were to learn about the clinical characteristics of osteoarthritis patients attending primary care clinics, and to identify those factors related to clinical inertia that, when modified, may result in clinical benefit for the patient.

Between September and October 2007, a total of 1361 GP were selected: 403 (30%) were assigned to Group 1 and 958 (70%) to Group 2. No significant differences were observed between groups regarding the number of years practicing, scientific and academic activity, age and gender.

A total of 4076 patients with hip and/or knee osteoarthritis were included in the study, 1208 (30%) attended by GP assigned to Group 1 and 2868 (70%) attended by GP belonging to Group 2. Overall, 1007 (83%) group 1 patients and 2303 (80%) group 2 patients attended both study visits. Discontinuation rate was a slightly higher in the control group (20% vs 17%).

Baseline characteristics of the study groups are described in Table 1. There were small significant differences regarding the type of osteoarthritis (in Group 1, knee osteoarthritis was more frequently observed: 74% vs. 71%, P=.039; and hip osteoarthritis less frequently observed: 41% vs 45%, P=.041). There were small differences in some concomitant diseases: arterial hypertension, diabetes mellitus, hiatal hernia or gastroesophageal reflux, and dyspepsia or peptic ulcer. A large proportion of patients were being treated with non-pharmacological measures. Most of those who were treated with pharmacotherapy took: analgesics (>90%), oral and topical non-steroidal anti-inflammatory drugs (NSAIDs) (>70% and >50%, respectively), and SYSADOAS (Symptomatic Slow Acting Drugs for Osteoarthritis) (>40%). The only differences observed were patients receiving topical NSAIDs (P=.021) and opioids (P=.021).

No significant differences were observed in any of the items evaluated by the SF-12v2, the WOMAC or the VAS questionnaires in visits 1 and 2 (Table 2), except the vitality evaluated by SF-12v2 (P=.049). In the comparison of the incidence of relevant events occurring during the six-month study period (Table 3), Group 1 showed significantly more induced medical evaluations (P=.011) and less referrals to other specialties.

Comparing the results of the two visits for the whole study population, significant differences were observed in most of the items evaluated (Table 4).

The results of our study can be grouped into three major observations: (1) patients with osteoarthritis have a high degree of comorbidity and a low health-related quality of life (HRQoL), as reflected by the SF-12v2, WOMAC and VAS scores; (2) patients did not seem to obtain a clinical benefit from the intervention studied, as there were no significant differences between visit 2 and visit 1; and (3) patient-reported pain, functionality and HRQoL significantly improved in the overall study population.

Study patients seem to have a higher degree of impairment and a worse health status in relation to the general Spanish population. This was tested by Vilagut et al.,12 who observed that the mean SF-12v2 Physical Component Summary (PCS) and Mental Component Summary (MCS) scores of a representative sample of the Spanish population were 45.05 (SD: 10.17) and 47.81 (SD: 9.20), respectively, indicating a somewhat better quality of life than our study group.

One limitation of the study is the discontinuation rate, around 20%, higher than the 10% used to calculate the sample size, despite the motivation strategy, follow-up, phone and e-mail reminders. It was not possible to do a systematic analysis of discontinued patients, as many were due to job changes of GP, but were similar in number in both groups. The high prevalence of some diseases, like diabetes mellitus (33% and 37% in the control and intervention group) are consistent with the average age of the sample (68 years) and the existence of high rates of comorbidity.

As regards the clinical benefit obtained by patients treated by the GP who had received the specific intervention, our findings call for careful interpretation, and several aspects must be taken into consideration. Therefore, despite the efforts to modify clinical inertia in GP, it is likely that these were insufficient, probably, among other reasons, because they focus on a single factor.

The observation that HRQoL perception improved significantly in overall population from Visit 1 to Visit 2, regardless of the study group, is important. Since the majority of patients had one or more concomitant diseases, osteoarthritis may have been unimportant until enrolment in this study. Thus, a small increase in GP attention may have represented a significant change to the patient in the quality of medical care received, which in turn, had an effect on the HRQoL. We are aware that the mere fact of having been included in a study, which produces a psychological effect of being observed, may have generated a positive response in the physicians, which in turn affected the quality of healthcare delivered to patients, irrespective of the group assigned to. This Hawthorne effect is expected in all types of studies in which the investigator or subject are aware of their participation.15 Therefore, the long-term effects on HRQoL will need further evaluation.

The results of this study should be interpreted in the light of potential limitations. First, considering the chronic nature of osteoarthritis, the six months follow-up period might not have been long enough. Secondly, as discussed above, the type of intervention we designed in order to modify clinical inertia may have fallen short, or the factors we intended to modify were not the most appropriate. Nevertheless, since the actual recruitment was 1361 physicians and 4076 patients, we consider that the power of the study is enough to support the main conclusions. Also, although the fact that approximately 20% of patient data was lost on follow-up when an estimation of 10% was initially planned, does not invalidate our results. Moreover, it could be an additional factor contributing to clinical inertia due to a lack of an adequate follow-up of the disease.16 The search for factors that might contribute to clinical inertia through interventions on primary care physicians have been used in other cluster randomised trials,17,18 yielding positive results regarding the implementation of interventions carried out on physicians to reduce clinical inertia.18 Future educational interventions on practitioners could use the recent Clinical Practice Guidelines for Osteoarthritis of the Royal College of Physicians as a reference.19

In conclusion, the results suggest that minor interventions, such as an increased awareness of the patient's disease through the use of functionality indexes, as well as the mere fact of being observed, may be improving patient-reported pain, functionality and HRQoL. Nevertheless, because this specific intervention carried out on physicians to prevent clinical inertia did not provide an additional clinical benefit to our sample of patients with knee and/or hip osteoarthritis, efforts should be made to identify potentially modifiable factors that lead to clinical inertia other than physician-related factors.

AGC, ATV and FLV developed the original idea for the paper.

AGC, ATV and FLV contributed to the design of the study.

AGC, ATV, FLV, DLP, APM, FVN and JHH participated in the execution of the study.

The Clinical Epidemiology Unit of the Hospital 12 de Octubre performed the analysis and interpretation of data.

AGC, ATV, and FLV wrote the article, helped by Dr. Ximena Alvira from HealthCo SL (Madrid, Spain).

AGC, ATV, and FLV made the decision on submitting the manuscript for publication.

All authors approved the final version of the submitted article.

Researchers were independent from funders and sponsors, and they had access to all the study data.

The authors declare that they have no competing interests.