Consultation charges in Ireland deter a large proportion of patients from seeing the GP: Results of a cross-sectional survey

Abstract

Objective: To estimate the effect of a consultation charge on the health-seeking behaviour of patients. Methods: Cross-sectional survey of patients carried out in Northern Ireland, where services are free at the point of delivery, and the Republic of Ireland, where 70% of the population are charged a consultation fee to see the general practitioner (GP). Results: There were 11 870 respondents to the survey (response rate 52%). In the Republic of Ireland, 18.9% of patients (4.4% of non-paying patients and 26.3% of paying patients) had a medical problem in the previous year but had not consulted the doctor because of cost; this compares with only 1.8% of patients in Northern Ireland. Because those in the Republic of Ireland on low income are entitled to free care, the effects of the consultation charge were most marked in the middle of the income distribution, with such patients being over four times as likely to have been deterred as those in the most affluent group. However, amongst paying patients, it was the poorest and those with the worst health who were most affected. Compared to the most affluent patients and those without depression, the likelihood of not having seen the GP due to cost was 6.75 (95% confidence interval [CI] 3.79, 11.09) for the poorest patients and 2.01 (95% CI 1.53, 2.52) for those with depression.

Conclusion: Even in countries with exemptions for the poor and more vulnerable, a consultation charge can deter a large proportion of poorer and less healthy patients from seeing their GP.

• Consultation charges
• co-payments
• cost avoidance
• general practice

Introduction

Many western countries, faced with increasing healthcare costs, are struggling with reform in order to give an enhanced role to market forces in organizing and paying for healthcare [1], [2]. Market regulation tries to ensure that the principles of equity and social solidarity are somewhat protected while permitting competition especially in the area of medical fees [3]. Primary care is currently receiving much attention from policy makers as it is seen as having a moderating effect on escalating secondary and tertiary care costs [4]. Most of the definitions of primary care in western countries are driven by the values of equity, justice, dignity, and solidarity, although in practice funding mechanisms usually “reflect and evolve from the economic conditions and socio-cultural and political characteristics of a country and its communities” [5]. These range from services being free at the point of delivery, as in the National Health Service in the UK, to cost sharing (where the patient contributes towards the cost of treatment), which is favoured by many other European countries [2]. Attempts to curb the increasing demand for care can be categorised into those operating on the supply side and those on the demand side. In supply side approaches, policy changes are aimed at reducing the incentive to provide services. It is known that the way in which doctors are paid may affect the propensity to order tests and the frequency of follow-up and that changes from capitation to fee-for-service payment usually result in an increase in all types of service including surgery and home visits along with laboratory tests [6]. Demand side approaches often take the form of user charges. The theoretical base for the excess use of medical care services induced by health insurance is the so-called ‘moral hazard’ problem. Essentially it is an extension of the Law of Demand, i.e. when the price goes down, consumption goes up. [7] User charges may be either co-payments (where a certain amount or proportion is paid by the patient) or a deductible (where the first amount of a payment of care is paid by the patient). Co-payments and deductibles are particularly attractive as they are an immediate deterrent to demand for care because they are imposed at the point that the decision is made to seek services. However, while there is evidence that this system of cost sharing reduces demand on services, there are concerns about adverse effects on poorer people and those with the poorest health [8]. This has led some countries to introduce payment exemptions to protect the most vulnerable [9], [10] but the effectiveness of such safety nets is unclear.

In order to measure the impact of cost sharing on the use of primary care services, we sought to compare the effect of a consultation charge on GP attendance. We compared two systems—one in which consultation charges predominate and one that is free at the point of use. We sought to characterize the types of people affected by the consultation charge in terms of socio-demographic and health characteristics. We have used the Republic of Ireland and the United Kingdom as exemplars of these two paradigms. In the Republic of Ireland, 70% of patients pay the GP directly for consultations, with the remainder being means tested for eligibility for General Medical Services (GMS), which entitles them to services that are free at the point of use. All patients aged 70 and over are automatically eligible for GMS, irrespective of income. The consultation charge for non-GMS patients is affected by market forces and ranges from approximately €35 to €55. Northern Ireland, like the rest of the UK, has a universal healthcare system that is free at the point of service delivery for all patients. Given that Northern Ireland and the Republic of Ireland are two parts of one island with very similar mortality and morbidity experiences they form a natural quasi-experimental resource for examining the effect of a consultation charge on attendance at the GP.

Methods

Twenty practices were purposefully selected in the Republic of Ireland to provide a good representative mix of national practices according to location (rural, small town, city) and practice size (one of three groups based on whole-time-equivalent GP principals) [9]. Practices in Northern Ireland were then classified according to these criteria, and a random selection of 20 practices drawn to match those in the Republic of Ireland. A questionnaire was sent to a random selection of 625 patients drawn from the patient lists of each practice using computer-generated random numbers. The survey was preceded by a personalized letter from the patient's GP, and non-responders were sent two reminders, the second containing another copy of the questionnaire. Parents/guardians were asked to complete questionnaires on behalf of patients aged less than 16 years old. The survey was conducted during October and November 2003.

In addition to the usual demographic variables, an array of socio-economic data was collected, including car ownership (categorized as no car, one car, two or more cars), tenure (dichotomized into renting and non-renting), and academic attainment (three levels of educational attainment [primary school only, secondary level, tertiary level]). The survey also captured data relating to gross annual household income, which was then equivalentized to adjust for the number and type of dependents in the household [10]. Health measures included limiting long-term illness (LLTI), general health over the preceding year, and a two-question instrument that has been recommended for use in a primary care setting as an aid to detecting patients at high risk of depression [11].

A question (Box 1), identical to one that had been included in the Commonwealth Fund International Survey of inequalities in healthcare [12], was asked about the effects of the costs on the propensity to visit the GP. As with the Commonwealth Fund survey, the percentage of patients affected by the co-payment was calculated by dividing the affirmative responders by the total number of responders, thus including the “not appropriate” responses (and the “not knowns”) as part of the denominator.

Box 1. Cost avoidance question.

Cost avoidance question:
“During the past 12 months, was there a time when you had a specific medical problem and did not visit the doctor due to the cost?”
Responses:
1. Yes	2. No	3. Not appropriate

All data were re-weighted to represent the national age and sex distributions within each jurisdiction. Chi-square tests were used to test for differences in the proportion of patients across the various demographic and socio-economic strata. Logistic regression analysis was used to determine which factors amongst the non-GMS (paying) patients were independently associated with the likelihood of being deterred by cost. Analysis was undertaken in STATA with robust estimation to account for the clustering of patients within practices.

Results

The overall response rate to the survey was 52% (11 870 respondents). Younger adults were underrepresented in both parts of the island, which probably reflects the difficulties GP registries face in maintaining accurate addresses for this more mobile age group. In Northern Ireland, where services are free at the point of delivery, 121 (1.8%) respondents had not consulted a GP in the preceding year because of the cost (Figure 1), and there was little variation according to age or sex. In the Republic of Ireland, 1745 (33.4%) respondents were in the GMS, and there were 989 (18.9%) patients who had not seen the GP because of cost (comprising 4.4% of the GMS patients and 26.3% of non-GMS patients; the latter representing 23.8% of males, 28.3% of females). Approximately 15% of children aged less than 10, and of those aged between 60 and 69 years, in the non-GMS group had been affected by the consultation charge. However, the effect was most pronounced amongst younger adults, over 40% of whom had a medical problem but did not see the GP because of cost. Table I shows that across the Republic of Ireland the deterrent effect of the consultation charge was most evident in patients in the middle of the income range. This pattern is to be expected given that, at higher incomes, the cost is a relatively weak disincentive while those with lower incomes are protected by the GMS safety net.

Figure 1.  Proportion of patients who, in the last year, had a medical problem but did not consult a GP because of cost; according to jurisdiction, age, sex and GMS status.

Table I.  The likelihood of patients in the Republic of Ireland with a health problem not having seen the GP in the previous year because of cost; variations according to household income.

	Income category
	Most affluent	2nd	3rd	4th	5th	6th	Least affluent
Respondents	440	674	827	599	965	834	629
Odds ratio*	1.0	1.82	2.58	4.14	3.06	2.49	1.98
(95% CIs)	—	(1.26, 2.61)	(1.83, 3.64)	(2.90, 5.90)	(2.17, 4.31)	(1.75, 3.55)	(1.34, 2.91)

* Adjusted for age and sex.

Logistic regression was undertaken to determine the characteristics of non-GMS patients most affected by the consultation charge. The analysis was limited to those aged 20 and over to maintain relevance of the socio-demographic factors, and also to patients aged less than 70, as all patients older than this are entitled to free medical consultations. The results (Table II) confirm that it is younger adults who are affected most by the consultation charge. Household income was a major predictor of the effects of a co-payment on cost avoidance within the non-GMS group, with those in the middle to lower income bands being around five or more times as likely to be affected as those in the upper band. The strong association between income and other measures of socio-economic status (SES) reduced the significance of the other SES variables in the model, so that although there was a tendency for the deterrent effects to be more evident amongst patients who rented, did not have access to a car, or had only primary level education, only tenure maintained significance in the final model. The figures suggest that females may be more vulnerable to the co-payment than their male counterparts, although sex did not maintain statistical significance at the recommended (p<0.05) level.

Table II.  Socio-economic and health characteristics of paying patients in Ireland who did not consult their GP because of cost; results of fully adjusted multivariate logistic regression.

	Non-GMS patients aged 20–69
Age, years	Adjusted odds ratio	P value*
20–29	1.00	—
30–39	0.84	0.255
40–49	0.52	<0.001
50–59	0.31	<0.001
60–69	0.19	<0.001
Sex
Male	1.00	—
Female	1.25	0.155
Tenure
Owner occupier	1.00	—
Renting	1.79	0.003
Income category
Highest	1.00	—
2^nd	2.07	<0.001
3^rd	2.68	<0.001
4^th	4.96	<0.001
5^th	4.48	<0.001
6^th	5.30	<0.001
Lowest	6.75	<0.001
LLTI
Absent	1.00	—
Present	1.38	0.047
General health
Excellent	1.00	—
Very good	1.24	0.159
Good	1.59	0.002
Fair/poor	2.08	<0.001
Depression
Absent	1.00	—
Present	2.01	<0.001

* Robust confidence intervals were calculated to take account of the clustering of patients within practices.

The likelihood of not having seen the doctor due to cost was almost 40% greater in those with an LLTI, and those who reported their general health as fair or poor were twice as likely to have been deterred from visiting the GP due to the cost. Patients with depression were also twice as likely as those without depression to have been deterred from seeing their GP because of the cost.

Discussion

Internationally, much attention has focused on the role of primary care in addressing the twin issues of rising demand and healthcare costs. In Northern Ireland, services are free at the point of delivery, and this study shows that very few patients (1.8%) are deterred by cost from seeing their GP. In the mixed system of the Republic of Ireland, more than one in four of the paying patients had a health problem in the year prior to the study but did not attend the GP because of cost. This contrasts with only 4.4% of non-paying patients for whom services are free at the point of delivery. Amongst the paying patients, it is those that are poorest and those in the worst health who are most affected by the consultation charge.

These findings are in line with those of the Commonwealth Fund Survey of Health Policy [13] from which the question on the deterrent effects of costs for care was drawn. That survey found evidence of inequalities in access to care in each of the countries where patients had to share part of the direct costs of their care (Australia, Canada, New Zealand, and the US). This was most pervasive in the US, where, despite worse health, adults with below average income were significantly more likely than those with above average income to report no visit to a doctor in the past year and no regular physician [13]. In the UK, where only 3% of respondents said that cost had deterred them from seeing their doctor, there was no evidence of inequality in access to care across the income spectrum. The slightly higher percentage of GMS patients in the Republic of Ireland affected by the cost may be due to the ability of these patients to consult a doctor other than the one they are registered with, if they are willing to pay.

The response rate to the survey was in line with large population-based studies, although the prevalence of GMS eligibility among respondents (33%) was a little higher than the national average of 30% in 2003. There is, however, no reason to believe that the effect of the consultation charge on the non-GMS patients in our survey differs significantly from the rest of the country. Another limitation is that the question about deterrence did not ask how often patients had been deterred from seeing their GP or about the severity of the presenting health problem. However, again the likelihood is that the analysis represents an underestimation of the true extent of the problem, given that those affected were some of the poorest patients with a greater prevalence of ill health.

Most of the evidence on the effects of co-payments comes from a series of natural experiments involving changes to the charging systems for patients and from a seminal randomized trial in the US, the RAND Health Insurance Experiment. Collectively, these show that the more patients have to pay out-of-pocket expenses, the fewer medical services they use [6], [14–17]. Even relatively small user fees can reduce the demand for medical services [14]. Despite there being no correlation between socio-demographic status and the overall decrease in usage of services under cost-sharing, the system was found to produce negative health effects in lower income groups [6]. Interestingly, the effects may be more marked amongst female patients [14], [15], and there is some suggestion that this may also be the case in the Republic of Ireland. This may be associated with the distribution of income within the family or the increased willingness of women to forgo their own needs when resources are stretched. It is possible that the consultation charge in the Republic of Ireland results in the delayed presentation of acute conditions and poorer control and management of chronic conditions such as asthma, diabetes, and hypertension for a large proportion of the population, again this is supported by evidence from the RAND HIE. Evidence from the US suggests the effects of consultation charges will be even greater on health-promoting and health-protecting activities [18], [19], and it has been suggested that eliminating cost sharing there would increase the uptake of preventative counselling, cervical screening, and blood pressure screening by up to 15% [20].

Proponents of consultation charges argue that they make consumers more cost conscious, and therefore discourage “unnecessary” utilization, which is a problem. However, it has been shown that a consultation charge is a rather blunt instrument which is as likely to reduce appropriate as inappropriate consultations [21]. Opponents of consultation charges point out that equity of access to healthcare is explicitly endorsed as one of the main objectives of health policies of most developed countries [22], [23] and should therefore be distributed primarily on the basis of need rather than ability to pay. Given that primary-care-orientated healthcare systems are associated with lower costs of care, better levels of health, and lower levels of medication [24], it seems reasonable that healthcare systems should now consider removing the financial obstacles that prevent a considerable proportion of patients from attending their general practitioner.

Declarations

Ethical approval was obtained from the relevant bodies in the Republic of Ireland (Irish College of General Practitioners) and Northern Ireland (Research Ethics Committee, Royal Group of Hospitals). Funding was provided by an Ireland–Northern Ireland Co-Operative Research Project Grant provided by both the Health Research Board in Dublin, Ireland, and the Research and Development Office in Belfast, Northern Ireland. The author reports no conflict of interest.