Socioeconomic gradients in access to services have been mitigated by the introduction of universal health insurance in Canada [3, 4], but persons with higher levels of education and household income continue to report relatively greater use of specialist services [3, 5]. American critics of Canadian health care have suggested that universal access is partly illusory, given the waiting lists for some diagnostic and therapeutic services that have emerged in all provinces [6-8]. Proponents of universal coverage argue that waiting occurs after patients have already accessed the health care system. If so, delayed care with prioritization according to need is theoretically more equitable than overt denial of access on the basis of income or insurance coverage [8].

For example, in the province of Ontario (population, 11.5 million persons), long waiting lists for coronary artery bypass graft surgery led to the generation of explicit clinical criteria for waiting list priority, along with a provincial queue management system. Analyses have since confirmed that waiting times for coronary artery bypass graft surgery are determined primarily by clinical factors, such as severity of angina and extent of coronary disease [9].

Most cardiovascular services (and almost all other specialized services) are not subject to similar management mechanisms, either in Ontario or in the other nine provinces. One might also expect some gaming of any system that imposes delays in access to services that patients value.

Accordingly, we undertook a survey to determine Ontario health care providers' perceptions and experiences of preferential access to cardiovascular services on the basis of factors other than clinical need.

Methods

Top Methods Results Discussion Author & Article Info References

Participants

All Ontario cardiologists (n = 268) and cardiovascular surgeons (n = 68) and random samples of general internists (300 [17.9% of all registrants]) and family physicians (300 [5.5% of all registrants]) were selected from the registers of the Royal College of Physicians and Surgeons of Canada and the Canadian College of Family Practice. Both national physician databases are updated regularly. We also included 218 chief executive officers, one from each Ontario acute care hospital in operation at the time.

Survey Development

Because preferential access is a sensitive issue, we aimed to keep the survey brief, general, and nonthreatening to respondents. We deliberately did not ask respondents about their exact role in provision of preferred access, nor did we seek to determine how often the phenomenon occurred. We developed prototypes for the questionnaire and pretested them with a convenience sample of 12 clinicians.

The actual wording of the survey and responses is provided in Table 1, Table 2, Table 3 and Table 4. The sequence of the questions proceeded from beliefs to experiences and personal involvement, along with increasing specificity in the definition of preferential access.

Survey Procedures

Participants were mailed a self-administered questionnaire in March 1997. We sent one reminder and one subsequent follow-up mailing of the questionnaire to nonrespondents. The survey consisted of the above-noted questions, each of which had several subquestions requiring a yes or no response.

Statistical Analysis

Results were analyzed as an aggregate of all respondents and by provider and administrator subgroups by using binary outcomes. Chi-square tests of independence were used to identify differences between respondent groups. If overall differences were confirmed (P < 0.05), multiple unpaired comparisons between respondent groups were made by using the Fisher exact test. Multiple paired comparisons were made as appropriate for detecting statistical differences for responses to items within and between questions by using 2 x 2 contingency tables with the McNemar test. Significance of results from multiple unpaired and pairwise comparisons was determined after using a Bonferroni correction: That is, the conventional P value of 0.05 was divided by the number of comparisons. We did not examine all possible comparisons within or between questions. The maximum number of unpaired or paired comparisons examined within any one question was 10; a corresponding P value of 0.005 for any single comparison was deemed significant. All data were analyzed by using the statistical software Stata (Stata Corp., College Station, Texas).

Results

Top Methods Results Discussion Author & Article Info References

Survey Response

Of 1154 mailed surveys, 7 were undeliverable. Eight hundred thirty of the 1147 remaining physicians and hospital administrators (72.4%) responded. Forty-two respondents declared themselves ineligible: Thirty-four physicians were not practicing or were not involved in cardiac care, and 8 chief executives were no longer associated with an acute care hospital. Thus, the eligible response rate was 71.3% (788 of 1105 respondents) (Figure 1). Item response rates averaged 96.8% (range, 92.0% to 98.7%). Results for all items are presented for valid responses.

View larger version (17K): [in this window] [in a new window]   Figure 1. Participant response rate. The eligible response rate is calculated by subtracting the number of respondents who deemed themselves ineligible from both the total respondents (numerator) and the potential recipients (denominator). CEO = chief executive officer.

Because of the sensitive nature of the topic, recipients were assured anonymity; there were no identifiers on responses beyond those demarcating provider specialties and administrators. Accordingly, we could not analyze responses by age, sex, or demographic status, nor could we compare respondents and nonrespondents in this respect. However, response rates of more than 60% were obtained for all subgroups of respondents, and postal code distributions did not differ significantly between respondents and nonrespondents.

Beliefs about External Pressure for Preferential Access

As shown in Table 1, only 26.2% of respondents believed that preferential access occurred without specific external pressures. More than 80% of respondents cited pressure from referring and consulting physicians or patients and their relatives as potential causes of preferential access. Significantly fewer respondents (57.4%) identified pressure from administrators as a potential factor (P < 0.001 compared with the preceding two categories). Hospital chief executives were less likely to indicate administrative intervention as a factor in preferential access (45.0% affirmative responses compared with 60.7% affirmative responses from responding physicians; P < 0.001). Otherwise, the proportions of affirmative responses for each item in this question did not differ significantly among groups of respondents.

Beliefs about Patient Factors That Cause Preferential Access

About 88% of respondents indicated that personal connections to the treating physician were likely to play a role in preferential access. As shown in Table 2, most respondents also believed that preferential access was more likely to occur in the presence of such patient factors as knowledge of the risks of delayed care (67.2%) and apparent propensity to litigate (63.8%). Only 44.3% of respondents believed that patient awareness of outcome differences among physicians and surgeons played a role in preferential access (P < 0.001 compared with the proportions of affirmative responses for any other item). About one third (33.1%) of respondents thought that patients who lived in a remote region with limited access to specialists would be given preferential access (P < 0.001 compared with any other item).

We found limited intergroup variability in responses. However, a significantly higher proportion of family practitioners than specialists and administrators believed that dealing with a potentially litigious patient or family was likely to promote preferential access (75.9% for family practitioners and 60.3% for all other physicians; P < 0.001).

Experience-Based Views on Types of Patients Receiving Preferential Access

Most respondents believed that a patient's community standing affected his or her likelihood of receiving preferential access (Table 2). Preferential access for physicians or their families was identified by a high proportion of respondents (93.2% [95% CI, 91% to 95%]; P < 0.001 compared with any other patient type); this finding was consistent with the responses to question 2 concerning personal connections to physicians. As shown in Table 3, high-profile patients, such as public figures (entertainers, professional athletes, media members, and others) and politicians, also generated frequent affirmative responses (84.8% [CI, 82% to 87%] and 83.0% [CI, 80% to 86%], respectively). These were followed, in order, by hospital board members, other health care professionals, patient donors to hospital foundations, business executives, and religious authorities. An open-ended question ("Other, please specify") elicited responses from 11.1% of the respondents. Respondents indicated that preferential access might also occur for patients enrolled in research protocols or those deemed to suffer intolerable stress from waiting for procedures. Proportionately fewer hospital chief executive officers than physicians rated hospital board members (42.3% and 67.5%) and patient donors to hospital foundations (62.9% and 80.9%) as likely to receive preferred access to care (P 0.001 for all comparisons).

Physician and Administrator Involvement in Preferential Access

We asked whether respondents themselves had ever been directly involved in the care of a patient who, for nonmedical reasons, received preferential access to various services (Table 4). A majority of respondents in each group answered affirmatively, with significantly more physicians than hospital chief executives acknowledging personal involvement in such cases (80.0% compared with 53.2%; P < 0.001). Medical specialty groups differed significantly for whether they had been involved in caring for a patient who received preferential access (P < 0.001), showing the expected trends according to degree of control over access to the listed cardiovascular services. After exclusion of tertiary procedures (coronary angiography, percutaneous transluminal coronary angioplasty, or coronary artery bypass graft surgery), most cardiologists (94.3%), general internists (75.5%), and family practitioners (53.1%) confirmed involvement in providing care to patients who received preferential referral for consultation or noninvasive tests.

We separately examined the responses of cardiologists and cardiovascular surgeons for major procedures. We hypothesized that preferential access would be less common with urgent procedures, for which clinical need might take precedence over nonclinical factors. Cardiologists reported involvement with patients receiving preferential access significantly less often in situations of urgent angiography (58.7% compared with 71.7% for elective angiography; P = 0.001), urgent angioplasty (48.1% compared with 58.6% for elective percutaneous transluminal coronary angioplasty; P = 0.005), and urgent bypass surgery (51.6% compared with 66.1% for elective coronary artery bypass surgery; P < 0.001). For cardiac surgeons, acknowledged involvement ranged from 27.9% of respondents for urgent percutaneous transluminal coronary angioplasty to 66.7% for elective coronary artery bypass graft surgery. Urgent-elective differences were of borderline significance only for bypass surgery (P = 0.005).

Discussion

Top Methods Results Discussion Author & Article Info References

In this exploratory survey of a large, representative sample of physicians and hospital chief executives in Ontario, most respondents acknowledged the existence of preferential access to cardiovascular services on the basis of factors other than clinical need. In addition, most respondents reported personal experience with the care of patients in such circumstances. Respondents reported that preferential access affected referrals to cardiologists or cardiovascular surgeons and use of specific cardiac procedures ranging from noninvasive tests to coronary surgery.

Respondents identified various nonclinical patient attributes that might influence care. Physicians and those with personal connections to them were deemed most likely to receive preferential access. A majority of respondents perceived that special treatment would also be accorded to persons with high social status, such as public figures and politicians. In addition, patients perceived to be knowledgeable, aggressive, or potentially litigious might receive preferential access.

Most of the intergroup differences in responses may be explained by respondents' political sensitivities (for example, the nonacknowledgment by chief executives of possible preferential access for donors or trustees), extent of involvement in cardiovascular care (such as differences in responses between specialists and family physicians), or degree of respondents' control over access to major procedures (differences in procedure-specific responses). Taken together, the responses are plausible and internally consistent. The responses also reflect positively on the honesty and integrity of several hundred practitioners and senior administrators who openly acknowledged potentially controversial issues and experiences.

We believe that these results are probably generalizable for cardiac care across Canada. The only salient way in which Ontario is distinct from the other provinces is that it has a well-established registry and reporting mechanism for cardiac surgery waiting lists. If anything, the existence of this registry might reduce, not increase, the extent of preferential access.

We studied cardiovascular care because the relevant group of specialty practitioners is well circumscribed, patients with heart disease are understandably anxious and might be more likely to seek preferential access, and waiting lists for major cardiac procedures have been heavily publicized in Ontario. We do not have similar survey evidence for other diseases. However, as noted above, analysis of a 1990 Ontario household survey showed that after controlling for self-reported health care needs, persons with higher socioeconomic status were more likely to visit specialists, whereas those with lower income or educational levels showed relatively greater use of general practitioners' services [3]. Education and income levels have also been associated with large disparities in screening rates for gynecologic cancer in Ontario [10]. Moreover, such phenomena are not unique to Canada. Health services researchers have generated considerable evidence that socioeconomic factors, including race or ethnicity, independently affect utilization of clinical services in many countries [10-14].

The key limitation of our study is that the survey gathered data only on the existence of preferential access; it provides no information on the absolute frequency with which preferential access occurs. Furthermore, it remains unknown whether preferential access leads to improved outcomes for persons who receive it or adversely affects those who do not. If very few persons benefit from preferential access and others are not harmed by it, our findings may not indicate a meaningful threat to the egalitarian aims of Canada's universal health insurance system.

If action to reduce preferential access were taken, several possible responses that are not mutually exclusive may occur. If the phenomenon is frequent, one option is to develop and implement ethical guidelines that define and circumscribe preferential access. Regardless of the frequency with which preferential access occurs, medical students and postgraduate trainees may benefit from education about these issues. Another option is to provide more generous funding for health care. If overall access to these services were improved, patients and physicians would have less reason to seek and provide preferential access and the impact of non-clinical factors on access could be reduced.

However, some preferential access is likely to persist. Our survey indicates that preferential access may occur if physicians perceive that patients or relatives are potentially litigious, demand preferential access, or are particularly well informed about relevant clinical issues. It is difficult to image a health care system, however egalitarian in its intentions, in which these factors would not have some influence on physician behavior.

Furthermore, if systems managers and providers were to address disparities in access to cardiovascular services, other foci of concern would become prominent, such as rural-urban differences in concentration of specialists and specialized facilities, wide variations in population-based rates of cardiac procedures, and differential utilization of angiography and revascularization in hospitals with and without such capacity on site [15-21]. Canadian researchers have also reported positive associations between measures of patient income and overall rates of cardiac surgery [22-24]. Implementing measures to provide better care for the socially or geographically disadvantaged may be more feasible and constructive than enforcing rigid rules to impede preferential access for advantaged persons.

Although our survey does not quantitate the frequency with which preferential access occurs, it explicitly confirms the existence of a phenomenon that is tacitly acknowledged by clinicians and administrators in many nations. Accordingly, our findings should provide a basis for closer study of the phenomenon of preferential access, an impetus for ethical debate, and a positive challenge to those who share our preference for a health care system in which access to care is predicated primarily on clinical need.