Osteoarthritis (OA) is considered to be the most common joint condition mainly affecting the older adult population.1 It is an anatomical and clinical syndrome characterised by mechanical pain, which is frequently associated with stiffness and which progressively leads to a loss of joint function.2 OA mainly affects weight-bearing joints, such as those of the lower limbs. The rate of incidence in knee joints is 240 cases per 100,000 people per year, followed by the hip with 88 cases per 100,000 people per year.3

Approximately 40% of adults over 70 have OA, and 80% of the people with this disease suffer from some type of limitation when carrying out their daily activities.4

In Colombia, one epidemiological marker of disease burden expressed in years of healthy life lost reports that 150 years healthy life are lost for every 1000 people with the disease.5 Years of health life lost to OA is among the first 20 causes of disability in women over 45 years of age.6

OA is one of the main causes of pain and disability in the world. It causes progressive impairment in functional capacity and consequently a diminishing quality of life.7 According to Neogi8 patients suffering from OA often present with chronic pain which increases with activity. As a result, people with OA lessen their participation in different activities to avoid triggering the pain. Pain therefore has a high impact on their quality of life, because it generates negative effects on their mood, their participation in social activities, recreation and even affects their sleep.

Patients with OA have to face changes in how they go about their daily life activities, bearing in mind that approximately 25% of people with this condition present with reduced functional capacity and therefore a diminished quality of life.9

Quality of life is considered to be a major trigger for measuring the effectiveness of interventions in these patients and it is therefore necessary to have tools which assess the impact of the different interventions in the patient's perception of their disease. Self-completion questionnaires are one of the tools used to assess the quality of life in patients with OA. They are a simple method, capable of evaluating the impact of the disease on the patient's life.4,10

The use of generic questionnaires such as the English Short Form Survey SF36 is common in patients with OA. This is a generic survey and one of its limitations is its low sensitivity to change.11 The Western Ontario and Master Universities Osteoarthritis Index (WOMAC) has been used worldwide to measure the quality of life of patients with OA; but it is a questionnaire which mainly measures functional capacity since it measures variables such as pain, stiffness and function and does not bear in mind aspects such as mental health and social functioning that may be affected by OA.12,13

The Osteoarthritis Knee and Hip Quality of Life (OAKHQOL) questionnaire is specifically for OA of the knee and hip but it has not been assessed in the Colombian context through interviews. This tool consists of 43 items distributed into 5 domains which include physical activity, mental health, pain, social support and social functioning. Each item is scored on an ordinal scale of 11 points, where 0 is the worst quality of life and 10 is the best quality of life. For each dimension the average score of the items contained in it are calculated.12

The aim of this study is to become aware of the reliability of the results from the tool in the Colombian context when it is applied through the use of interviews. Health professionals use interviews to attend to the users in research and clinical settings, since OA of the knee and hip is a major social and health problem in all societies, making it relevant and of general interest to adapt the existing assessment indexes to different cultures.

The results in reproducibility are similar to those found in the original research, which was determined by the French population,17 so the domains of activity and pain presented a ICC which was higher in comparison with the other domains. Regarding the domain of physical activity, reproducibility in both studies was almost perfect. The pain and social activities domains, for their part, were classified as substantial.

It may be observed that the similarity in ICC results, range between substantial and almost perfect, with the exception of the social support domain, which was classed with a moderate ICC in this study, whilst in the French version it was substantial. Another of the differences to the French study is that the mental health domain was classified as almost perfect, whilst the ICC result in this study was substantial. These differences may be due to the fact that in the French version the sample was larger compared to that of this study.

In the Spanish version12 of the questionnaire reproducibility was substantial in physical activity, pain and mental health domains, and light in social activities and social support domains. When the findings from this study were compared with those of the Spanish version some major differences were found in the social support and social activities domains which in our study were shown to have a higher reproducibility. This may be due to the form of application since in the Spanish study the questionnaire was sent via electronic mail and the participant was unable to air their doubts, whilst in this study they were able to be resolved.

Regarding internal consistency of the findings from this study, they were consistent with the French version, in that the dimensions of the social activities and the social support also present lower coefficients. However, in this study the consistency was moderate for this dimension, which may be explained by the way in which it was applied in this study, since the person in charge of explaining the questionnaire was present during its application and responded to doubts, which were able to improve the correlation between the items.17 The application using electronic mail may facilitate data collection and cover a large part of the sample more easily and quickly, but this form of application is inconsistent with the form of application of tools in the clinic or in research.

The level of agreement mainly describes the differences existing between measurements and offers an idea of the variability of the questionnaire between the different measurements. These variations between the different applications can relate to variations in the people at whom the questionnaire is conducted or the professional who explains the questionnaire, and may be due, among other factors, to the prevailing mood, fatigue, or time of day applied. The questionnaire also presents variability of measurements, and both those of the questionnaire and the people are the result of random situations.

The level of agreement enables an approach to be made to this variability determined by chance, which is important as a first step in establishing a cut-off point. From this cut-off point the variation of one questionnaire between 2 measurements will be the result of reasons other than chance, such as the research setting, an experimental design where changes would be expected to depend on intervention or on the clinical setting where changes would be expected to derive from the effectiveness of a therapeutic intervention.18

According to the graphic Bland–Altman analysis it may be proven that the 5 domains present an average of differences close to 0, which indicates that the difference in the score obtained in the first and second measurement was approximately 10 points; this is acceptable for a questionnaire with scores between 0 and 100.19 It should be considered that the maximum difference allowed between scores of both applications of the questionnaire is a clinical, not a statistical decision.20

The difficulty with the OAKHQOL questionnaire is that its agreement is determined in this case by the agreement limits, which show the maximum and minimum range of variation of the questionnaire and these were very extensive, with 5 domains between −50 and 50 points which means that the questionnaire in total may vary by around 100 points due to chance. In the terms previously mentioned this meant that changes under but close to 100 points in the research area could be due only to chance.20

Since the average of the differences is close to 0, this indicates that the error is approximately 10 points, which is acceptable to our interpretation, this questionnaire may be used in clinical practice, but in research more reliable measures are required, with fewer differences between the measurements to measure the impact of therapeutic interventions with quality of life as a dependent variable in older adults suffering from OA of the hip and knee. As a result this questionnaire could be the most appropriate one.

One limitation of this study is its small sample size and as a result all the participants included in this research may not be sufficiently representative of the adult population with OA of the hip and knee in Colombia.

The OAKHQOL questionnaire proved to have good psychometric properties in the domains of physical activity, pain and mental health, but in the domains of social activities and social support results were less reliable when it was applied through interview format. The variability of the questionnaire between measurements shows that it may be used in the clinical filed, but its use in research would have major implications.

None.