Background
Osteoarthritis (OA) is a leading cause of pain and morbidity, and globally, is the fastest increasing cause of disability [
1]. Despite this, a number of studies report low rates of consulting with a General Practitioner (GP) [
2‐
6], with 17% of patients with OA consulting annually [
3] and over 50% of those with severe pain not consulting over an 18 month period, in the UK [
2].
Understanding the precursors or antecedents to the consultation is important for researchers and clinicians who aim to enhance the care of patients with osteoarthritis in primary care for two reasons. Firstly, the reasons for consultation are intertwined with an individual’s ideas, concerns and expectations [
7], and their agenda for the consultation. Increasingly, the events in the consultation have become a focus of interest for researchers. Any consultation intervention to enhance care of OA needs to address patients’ agendas. Secondly, the reasons for seeking or using healthcare go hand in hand with reasons for not consulting, and hence understanding the drivers to consultation may further understanding of why many OA sufferers choose not to consult their doctor, described as the 'iceberg of morbidity’ [
8].
A number of studies have evaluated the influences on patients’ decision making around seeking healthcare for OA symptoms. This review aims to summarise what is currently understood about why patients with OA consult their GP, and was conducted within a wider programme of research aiming to enhance the care of patients with OA in primary care.
Methods
An initial literature search, performed as a scoping exercise identified relevant research using a wide range of both quantitative and qualitative research methods. Due to the diversity of studies, a narrative review was therefore felt to be most appropriate to confer the flexibility needed to review the relevant literature. A narrative review is described as a 'first generation 'traditional’ literature review’; narrative reviews have a useful place for identifying themes and gaps in the literature and for informing direction of further research [
9]. This review has been enhanced by a systematic literature search; combining narrative and systematic methods has value in enhancing transparency and rigour of narrative reviews [
9]. Studies have been identified by searching relevant databases (Medline, CINAHL, Psychinfo and Google scholar), in addition to reference checking, manual searching of the contents page of relevant journals and recommendations from experts. The search was divided under three headings: population (patients with osteoarthritis); setting (primary care) and consulting (consultations between GPs and patients). No search terms were used to limit the 'influences’ on consulting to avoid the risk of excluding relevant papers. Search terms used are shown in Table
1.
Table 1
Search terms used to identify influences on consulting behaviours in OA
Primary health care | Osteoarthritis | Consult* AND behavio* |
GP OR general practitioners | Osteoarthritis, knee | Consult* AND frequency |
Family physicians | Osteoarthritis, hip | Consult* and prevalence |
Family practice | Arthritis | Seek |
General practice | | Visit |
| | Utili*ation |
Inclusion of papers was deliberately not restricted. The aim was to include the broadest possible sample of papers, to determine all possible influences on consulting. Studies reporting consulting behaviour that have sampled older adults with joint pain, in contrast to clinically or radiologically diagnosed osteoarthritis have been included with two assumptions: that the factors affecting consulting rates are likely to be similar, and that patients with osteoarthritis are likely to account for the majority of respondents. Similarly, no exclusions were made on the basis of study design or patient population studied with the assumption that all studies may further understanding about consulting behaviours. 'Primary care consultations’ were defined as consultations between a GP and a patient for the purpose of this review, with consultations with other members of the primary care team excluded. To appraise the evidence, no single tool was appropriate for the diverse range of studies included. Particular attention has been paid to the characteristics of the sample and setting which may affect the transferability of the findings.
In a number of papers identified in this review, the Andersen- Newman model was used to describe influences on consulting [
2,
10‐
13]. This framework is used to describe factors that influence healthcare utilisation and is divided into three areas [
14]:
1.
Predisposing factors, the social and cultural characteristics of a person (including factors that may have existed prior to illness)
2.
Enabling factors, the logistical issues affecting accessing care
3.
Need factors, which are the most immediate cause for seeking healthcare and are usually related to the illness itself.
The influences on consulting behaviour have been classified under these headings in the results and discussion, in order to provide a framework for the narrative review and to organise discussion of similar themes.
Conclusions
We have reported the influences on consulting a General Practitioner with OA using Andersen and Newman’s model of healthcare utilisation which incorporates biological, psychological and social factors.
Health beliefs appear to be important predisposing factors in deciding whether or not to seek health care. The belief that OA is an inevitable part of ageing, about which little effective treatment exists and a perceived negative attitude of the GP are reported as disincentives to consulting. Health beliefs are also likely to interact with other identified themes; for example age, and the influence of previous healthcare use on consulting. Previous healthcare use has been associated with increased consulting, but could also result in fewer subsequent consultations if the patient perceived a negative response from the healthcare practitioner consulted. Other important health beliefs include perceiving severe consequences of pain and frustration, which are associated with increased likelihood of consulting. Depression is a further psychological variable for which the evidence is contradictory, and which is likely to be closely related to social context.
The 'need’ factors, in the context of OA are mostly represented by joint related symptoms, impact of the symptoms or comorbidities. Disruption of daily activities appears to be an important driver to seeking medical help. Severity of pain is higher in consulters compared to non-consulters although tests of statistical significance yield contradictory results; individual patients have reported pain as of importance in qualitative research and the lack of statistical evidence to support this may be related to limits of the quantitative measures used. Qualitative research has demonstrated a vast range of descriptors that patients use to describe pain which suggests the questionnaire tools used may be limited in ability to capture the full pain experience [
28], which may explain the discrepancy in findings. Again, need factors are likely to interact with an individual’s health beliefs. The physical factors such as severity, distribution and duration of pain may form a 'pattern’ of pain that patients perceive as normal or abnormal, which in turn will influence decision making to seek healthcare.
Patients with OA who consult their GP appear to have more comorbid conditions but how comorbidity affects consulting frequency about joint pain is not clear. Related to this is the finding that patients with severe pain are visiting their GP frequently about issues other than their joints. The literature would suggest these patients are not having their symptomatic joint pain managed, but this may be due to limitations in the various methods of estimating consultation frequency and content. Furthermore, the ways in which patients and doctors prioritise symptoms in the context of multi-morbidity is not well characterised in the literature.
In general, the social aspects and 'enabling’ factors are reported on less frequently than other variables in research in this area, although living alone and the area of GP practice appear to be important. The healthcare system is a further important contextual influence and some of the observed differences in findings may be explained by variation in healthcare access and availability, for example, the relationship between health insurance and financial status and consulting. Furthermore, differences in GP training across countries may impact on the consulting behaviours.
The predisposing, enabling and need factors are not mutually exclusive and there is some overlap between categories. For example, co-morbidities may be 'pre-disposing’ in the case of long term conditions that existed prior to the current illness, or 'need’ factors that are directly influencing the need for seeking health care. A further example are health beliefs, which may be classified as 'predisposing’, 'need’ or 'enabling’ factors. The model has been criticised for generally underplaying psychological factors [
10].
An alternative theoretical lens through which to consider access to healthcare is the notion of 'candidacy’ [
29], although this was not described in any of the papers in this review. A 'candidacy’ perspective suggests that an individual’s identification of his/her 'candidacy’ for healthcare may be constructed through their economic position as well as by dominant cultural values towards service use which may either increase or curtail health service use. It offers an alternative way of identifying what types of factors make patients 'better’ or 'worse’ candidates for consulting the GP’; based on their own lay beliefs about OA and wider cultural discourses around pain and disability, and their relationship to issues of fairness, equity and 'deservedness’. Candidacy draws together many of the factors discussed in this review, and is also a dynamic process, and this may go some way in explaining why the existing literature does not appear overly sophisticated in its ability to explain consulting behaviour.
One of the general methodological limitations of the studies included relates to estimates of consultation frequency. Consultation prevalence that is calculated using only diagnostic codes may underestimate consultation prevalence as there is evidence that GPs exercise caution when using diagnostic codes and may favour symptom descriptors [
30]. Coded data may also underestimate frequencies if not all aspects of the consultation are recorded [
31]. However, studies that identify consulters on the basis of all joint related medical record codes as well as a free text search may overestimate consultation frequency of OA specifically as alternative diagnoses will be included. Furthermore, overestimation of consultation rates in some of these studies may be attributable to selection bias due to the possibility that similar factors influence participation in research as those influencing decision-making to seek healthcare. Self-report is limited by recall bias which may over- or under-estimate consultation frequency.
A further limitation of the included studies that used quantitative measures to calculate influences on consulting are that these may underplay the interaction between variables. Depression for example, is an example where the evidence was weak and there may be variation in how this factor could influence consulting. One could argue that depression could both increase or decrease consultation frequency due to coping difficulties and lack of social support or due to isolation. Qualitative research may be better placed to explore complex influences, taking into account the social and environmental context.
This review highlights two particular aspects of the consultation itself that require further evaluation and research. Pain and disrupted activities appear to be important drivers to consulting and so there is therefore a need to establish how these issues are addressed within the consultation. Not all aspects of the consultation may be recorded or remembered and therefore observation research is well placed to explore this further. Secondly, perceived 'negative’ attitudes to osteoarthritis have emerged as disincentives to consulting and a further need exists to establish whether this is evident in consultations. The subjective issue of negativity is also a difficult topic to research using retrospective measures such as post-consultation interviews, and would require a research approach that incorporated multiple perspectives on the consultation.
Competing interests
The authors have no competing interests.
Authors’ contributions
ZP conceived the review, conducted literature searches, conducted the narrative review and drafted the manuscript. TS and ABH contributed to the narrative review and the manuscript. All authors read and approved the final manuscript.