Background
For patients, making an ineffective decision about seeking help can have profound consequences, not just for the health outcome of the patient, but for the use of limited healthcare resources. In 1979, Hannay described the ‘symptom iceberg’ where only a very few symptoms experienced (the tip of the iceberg) result in consultations with healthcare professionals (Hannay, Reference Hannay1979). The point at which the iceberg breaks the surface is the point where a wide range of underlying factors, such as anxiety and pain come together resulting in the patient's decision to consult. The factors that led patients to cross from the submerged part of Hannay's iceberg to the part above the water were considered by Zola (Reference Zola1973). A large body of literature has focused on these factors, identifying and examining their relative influence on patients’ decisions to seek help, for example the impact of rurality or loneliness (Ellaway et al., Reference Ellaway, Wood and MacIntyre1999; Farmer et al., Reference Farmer, Iverson, Campbell, Guest, Chesson, Deans and MacDonald2006).
The concept of help-seeking behaviour is not clearly understood and there has been an attempt by Cornally and McCarthy (Reference Cornally and McCarthy2011a) at clarification of the term to aid research and practice as well as theoretical development. Cornally concluded that help-seeking was essentially a response to a challenge of personal ability to resolve a health issue or problem (Cornally and McCarthy, Reference Cornally and McCarthy2011a). This definitional parameter does not differentiate between ‘appropriate’ and ‘inappropriate’ help-seeking. However, some studies have considered the ‘appropriateness’ with which patients seek help (Sempere-Selva et al., Reference Sempere-Selva, Peiro, Sendra-Pina, Martinez-Espin and Lopez-Aguilera2001; Carret et al., Reference Carret, Fassa and Kawachi2007; Adamson et al., Reference Adamson, Ben-Shlomo, Chaturvedi and Donovan2009). This has revolved around the patient's choice of which service to seek help from and has in part, informed the development of, for example, NHS Choices; a web site designed to help patients make more effective help-seeking choices.
Adamson et al. (Reference Adamson, Ben-Shlomo, Chaturvedi and Donovan2009) considered the notion of ‘appropriateness’ from the patients’ perspective. They found that patients did not want to be seen to be engaging in help-seeking behaviour that was considered inappropriate but considered that other help-seekers did consult inappropriately. Patients did seem to be comfortable with the term ‘inappropriate’ but the authors did not offer an explanation of the term. It is reasonable to assume that patients had their own subjective interpretation of the term and were happy with that. The study did not consider healthcare professional's perception of ‘inappropriateness’. In the same way that the term ‘help-seeking’ needs to be defined (Cornally and McCarthy, Reference Cornally and McCarthy2011a), so does ‘inappropriate’. Defining this term is likely to be challenging as it is subjective in nature.
The ‘inappropriate’ use of Accident and Emergency departments was investigated by Carret et al. (Reference Carret, Fassa and Kawachi2007). ‘Inappropriateness’ was measured using the Hospital Urgencies Appropriateness Scale (Sempere-Selva et al., Reference Sempere-Selva, Peiro, Sendra-Pina, Martinez-Espin and Lopez-Aguilera2001) but this scale only considers whether patients make visits to A&E departments rather than seeking help elsewhere or with another healthcare professional. The term ‘inappropriate’ has only been objectively defined in terms of seeking help from the wrong sort of healthcare professional.
The majority of studies consider help-seeking behaviour from three perspectives; help-seeking for specific conditions (Kooperman et al., Reference Kooperman, Mainous and Jeffcoat2004; Corner et al., Reference Corner, Hopkinson and Roffe2006; Adamson et al., Reference Adamson, Donovan, Ben-Shlomo, Chaturveedi and Bowling2007; Leydon et al., Reference Leydon, Turner, Smith and Little2008; Unger-Saldana and Infante-Castaneda, Reference Unger-Saldana and Infante-Castaneda2011; Cornally and McCarthy, Reference Cornally and McCarthy2011b), specific services (Sempere-Selva et al., Reference Sempere-Selva, Peiro, Sendra-Pina, Martinez-Espin and Lopez-Aguilera2001; Carret et al., Reference Carret, Fassa and Kawachi2007; Murray et al., Reference Murray, Majeed, Khan, Lee and Nelson2011) and help-seeking among specific patient cohorts (Baker et al., Reference Baker, Mead and Campbell2002; Galdas et al., Reference Galdas, Cheater and Marshall2005; Hippisley-Cox et al., 2009; Hale et al., Reference Hale, Grogan and Willott2010). The studies identified as investigating help-seeking for specific conditions were all concerned with delayed help-seeking. Various factors were identified as underpinning such as delay, but the delay itself may have resulted in poorer outcomes for the patients concerned. These studies, by their nature, did not examine help-seeking generally.
When examining help-seeking behaviour for specific service providers, it was found that patients often sought help from the wrong healthcare professional or service (Carret et al., Reference Carret, Fassa and Kawachi2007; Murray et al., Reference Murray, Majeed, Khan, Lee and Nelson2011). Again a variety of factors were thought to underpin this behaviour. No differentiation was made between day-to-day conditions and chronic or complex conditions that were presented to these service providers.
The literature considers the ‘appropriateness’ (or ‘inappropriateness’) of help-seeking for specific conditions, specific services and among specific patient cohorts (Sempere-Selva et al., Reference Sempere-Selva, Peiro, Sendra-Pina, Martinez-Espin and Lopez-Aguilera2001; Carret et al., Reference Carret, Fassa and Kawachi2007; Adamson et al., Reference Adamson, Ben-Shlomo, Chaturvedi and Donovan2009). These areas rarely deal with day-to-day symptoms that cause help-seekers to cross the water line to the part of Hannay's iceberg that is above the water. Authors have attempted to identify causal factors (Ellaway et al., Reference Ellaway, Wood and MacIntyre1999; Farmer et al., Reference Farmer, Iverson, Campbell, Guest, Chesson, Deans and MacDonald2006) but these are many and varied. What is known is that these and other factors come together to cause the individual to take Zola's path from person to patient (Zola, Reference Zola1973). When using the term ‘appropriate’ to describe the journey down Zola's path, we are dealing with a term that has limited definition and application (Sempere-Selva et al., Reference Sempere-Selva, Peiro, Sendra-Pina, Martinez-Espin and Lopez-Aguilera2001) and is open to subjective interpretation. A more objective set of terms is needed if objective research is to be conducted in this area.
To date the literature has only identified two types of ‘inappropriate’ help-seeking and these are: delayed help-seeking (ie, help sought at the wrong time; Sempere-Selva et al., Reference Sempere-Selva, Peiro, Sendra-Pina, Martinez-Espin and Lopez-Aguilera2001; Kooperman et al., Reference Kooperman, Mainous and Jeffcoat2004; Corner et al., Reference Corner, Hopkinson and Roffe2006; Hippisley-Cox et al., 2009; Cornally and McCarthy, Reference Cornally and McCarthy2011a) and help sought from the wrong service (Sempere-Selva et al., Reference Sempere-Selva, Peiro, Sendra-Pina, Martinez-Espin and Lopez-Aguilera2001; Carret et al., Reference Carret, Fassa and Kawachi2007). In addition, a third category should be considered; potentially avoidable help-seeking (Murray et al., Reference Murray, Majeed, Khan, Lee and Nelson2011).
We propose that the terms ‘appropriate’ and ‘inappropriate’ help-seeking are replaced by ‘effective’ and ‘ineffective’ help-seeking the latter of which falls into three categories as follows: help sought at the wrong time; help sought from the wrong service provider; and potentially avoidable help-seeking. For more information see Davies, G. (2009) available from the author.
Using a pilot scale survey, the aim of this study was to assess the proportion of consultations with GPs that are made ineffectively and which categories they fall into. This was done using a specifically developed data gathering tool (see Figure 1). The data gathering tool was based upon categories of ineffectively made consultations, which were informed by the literature and developed theoretically. It was not our intention to validate the tool at this stage. The questions we sought to answer in this study were: what proportion of self-referral consultations with GPs for day-to-day conditions are made ineffectively and what proportion of consultations fall into the three categories described above?
Data collection tool.
Methods
Design
A survey of four GP practices recording ineffectively made consultations made over a one-week period.
Participants and setting
All GP principals, salaried GPs, locum GPs and GP registrars (n = 22) at four North Wales primary care practices:
Practice 1: n = 7 (4 partners; 3 salaried doctors).
Practice 2: n = 4 (all partners).
Practice 3: n = 5 (all partners).
Practice 4: n = 7 (4 partners; 2 GP registrars; 1 locum).
Data collection took place between January and June.
See Table 1 for practice population information. Practice 4 was a split-site practice.
Summary of practice population profiles
Measures and procedure
All consultations for 1 week were assessed as to whether they were effective or ineffectively made. Out of hours consultations, home visits and clinic consultations (e.g. asthma clinics, blood pressure clinics, etc.) were not included. Data were not gathered from clinics where patients were asked to attend nor from home visits. See Figure 1 for an example of the data collection tool.
A presentation was given to GPs participating in the study during which a standardised explanation of the definitional parameters of an effectively targeted consultation and the three categories of ineffectively made consultation were given.
GPs recorded whether a consultation was effectively made or one of three types of ineffectively made consultation:
• In time (made too early, too late or in the wrong appointment slot);
• Made with the wrong healthcare professional;
• Potentially avoidable.
Data analysis
The proportion of consultations that were effectively made or ineffectively made for each category were calculated as percentages of total consultations and comparisons between practices were made using Fisher's exact statistic.
Results
A total of 1217 consultations were surveyed. Overall, across all practices, 28.59% of all consultations were recorded as ineffectively made in some way. Of all consultations made, 8.46% were made with the wrong healthcare professional, 5.01% were made at the wrong time and 15.12% were potentially avoidable. See Table 2.
Summary of results for all consultations by practice
IMC = ineffectively made consultation; HCP = healthcare professional.
The percentage of consultations that were ineffectively made varied across practices (see Table 2 for a breakdown of these figures by practice); however, a cross-tabulation of the rates of ineffectively made consultations between practices using Fisher's exact test revealed that the variation between practices 1, 2 and 3 (P > 0.7) was not significant. However, practice 4's rate of ineffectively made consultations was significantly greater than all other practices (P < 0.002). See Table 3.
Comparison of difference between practices’ rates of ineffectively made consultations
HCP = healthcare professional.
Practice 4 had significantly more avoidable consultations than practice 1 (P < 0.01) but was not significantly different from practices 2 and 3 (P > 0.1). Practice 4 had significantly more consultations made at the wrong time than the other practices (P < 0.01) but there was no significant difference between any of the practices for consultations made with the wrong healthcare professional (P > 0.1).
Discussion
The aim of this pilot scale survey was to assess the proportion of consultations with GPs that are made ineffectively and which categories they fall into. The rate of ineffectively made consultations for day-to-day conditions, over a one-week period was found to be significant.
When all GPs were considered together (regardless of practice) the mean ineffectively made consultation rate was 28.84% (SD = 15.6). Consultations made with the wrong healthcare professional accounted for 8.46% of all consultations made across the four practices and the difference between practices was not significant. This similarity between the four practices suggests that healthcare professionals had a reasonably consistent, objective understanding of consultations made with the wrong healthcare professional. However, consultations assessed by GPs to be made at the wrong time (5.01% of all consultations made) were not significantly different between practices one, two and three, but practice four recorded a significantly higher number of consultations made at the wrong time. In fifteen, 15.12% of all consultations made were considered to be potentially avoidable making this the largest of the three categories of ineffectively made consultation (wrong healthcare professional; wrong time; and potentially avoidable consultations).
The high standard deviation reflected the wide difference in the number of ineffectively made consultations recorded by individual GPs. This may be due to a combination of factors; the individual personality of GPs (Duberstein et al., Reference Duberstein, Meldrum, Fiscella, Shields and Epstein2007) and the potentially subjective nature of parts of the measurement tool (ie, ‘potentially avoidable’ consultations). This may reflect an issue with the term ‘potentially avoidable’ consultation. Despite a standardised explanation of the definition, this term still has the potential for subjective interpretation by GPs. Because of this, the category may have been used as a general category where GPs considered the consultation to be ineffectively made in some way but were unsure which category the consultation fell into. The term, or the definition of the term may need to be reconsidered in order to make it more objective. The significant difference between practice four and the three other practices in the potentially avoidable category may be partly explained by the timing of data gathering in practice four. This was during January while the other practices gathered data during the spring months (April, May and June). In the winter months, consultations for self-limiting seasonal conditions such as upper respiratory tract infections may increase, which may account for some of the difference between practice 4 and the other practices.
The examination of help-seeking for day-to-day conditions has not been examined before and the findings from this study would suggest that ineffective help-seeking for day-to-day conditions is a significant problem. This may detrimentally affect patients’ outcomes from diseases and conditions. Problems of delay in seeking help for specific conditions, by specific cohorts and from specific service providers is known to be a problem but this study shows that this issue extends beyond specific situations to more general day-to-day situations. The potential scale of the problem of ineffectively made consultations for day-to-day symptoms cannot be underestimated. In 2008, there were 21.9 million clinical consultations in England (Hippisley-Cox et al., 2009) 76% of which were made with GPs. If 15.12% (16.64 million) are potentially avoidable, this represents a cost to the NHS of some £87.85 million at £35 per consultation (Curtis, Reference Curtis2009). When consultations made with the wrong healthcare professional or made too early or late, the cost may be significantly higher.
The categories that describe the effectiveness with which patients make help-seeking decisions may need some refinement or development as there is still room for subjective interpretation by the health professionals that complete it. For example, the category of time used here includes consultations made too early, too late or in the wrong time slot (ie, routine versus emergency consultations). Additionally, the ‘potentially avoidable’ category may be vulnerable to subjective interpretation by GPs despite clear directions from the researchers regarding the nature of this category as discussed above.
While the categories used here require some development, it is nevertheless a useful method for gathering data of this type and can be used to further our investigation of effective help-seeking in self-referral situations. Future studies should seek to replicate this research with larger populations. More robust parameters of the three categories of ineffective help-seeking proposed here should be developed, particularly that of ‘potentially avoidable’ consultations.
This was a pilot scale study and thus generalisations should be made with caution. The study also only scrutinised patients’ help-seeking behaviour in general practice but there are many other self-referral situations where this type of general help-seeking is made. For example, patients will seek help from A&E departments, pharmacists and NHS Direct/24 to name but a few. An understanding of help-seeking behaviour across a range of self-referral situations is needed to broaden still further our understanding of this issue. The findings from this study suggest that a proportion of patients are seeking help from GPs ineffectively. The reasons underpinning this behaviour need to be further examined and understood in order to inform future health resource planning.
Acknowledgments
This paper benefited from the advice and guidance of Dr Clare Carolan who helped structure the final draft.
