Background
Limited resources are a reality to which health care systems respond in very different ways. As physicians are confronted with scarcity and with the effects of cost-containment policies on clinical practice, they occupy a unique position from which to observe the impact of priorities set by health care systems.
Contradictory data exist as to whether physicians are aware of facing scarcity. In
The Painful Prescription: Rationing Hospital Care, Aaron and Schwartz noted that British physicians rationalized, or redefined health care standards to face scarcity more comfortably. [
1] Twenty years later, researchers conducting interviews with physicians regarding scarcity reported being struck with the strength with which scarcity was denied. [
2] US general internists, intensive care specialists, and oncologists, however, do report difficulties explicitly associated with resource scarcity. [
3] Data suggest that physicians accept prioritization decisions, both when faced with hypotheotical scenarios, [
4‐
13] and when reporting on their practice. [
3,
14‐
16] Physicians at the point of care are uniquely situated to observe the impact of priority setting decisions on patients in the form of scarcity, or less than equitable care. Their experience may thus yield useful insights and feedback about the impact of priorities on clinical care, which could contribute to evidence-based health policy. [
17] Despite this, insufficient attention is paid to their experience.
To examine the perceptions and attitudes of physicians regarding resource allocation in the European context, we conducted a three-part international survey of general physicians in Italy, Norway, Switzerland, and the UK. Results from the two other parts of this survey have been reported elsewhere. [
16,
18] In this paper, we report physicians' perception regarding lack of resource availability in their health care system and its adverse effects, their views regarding the equity of their health care system, and their attitudes towards various cost-containment policies.
Discussion
Scarcity, or resource unavailability, was reported by physicians in all four surveyed countries. Despite universal coverage, physicians reported underinsurance. Serious consequences of scarcity were reported in all countries. Resource availability was unevenly distributed: some interventions were more frequently unavailable, and some patients were identified as more likely than others to be denied care on the basis of cost. Physicians, however, accepted cost-containment policies. They reported willingness to participate in cost-containment, and did not want to be guided by prioritization decisions made at an administrative level.
Our study has several limitations. It has been suggested that physicians often deny scarcity [
2]. Although our results do not confirm this in the countries studied, physicians may still underestimate scarcity. There may also be pressures brought to bear on physicians, or expectations on the part of patients, but also physicians, that motivate them to think that more resources are necessary. This could lead to an overestimation of scarcity. However, as long as the interventions they consider to be indicated have at least marginal benefit, considering them to be unnecessary could be a matter of debate. As with all questionnaire studies, recall bias can be an issue. We used a conservative limit on the time we surveyed physicians about, however, they may still have remembered striking scarcity more than mundane everyday events [
26]. This could lead to an underreporting of scarcity, and a relative overreporting of the more serious kind of resource unavailability. Regarding the availability of specific resources, responses about mental health and chronic care bed shortages do seem to have face validity [
27,
28]. Asking about the most serious adverse event they had encountered in the previous six months, rather than the most frequent, may also mean that recall bias could be a lesser concern on that specific item. As we only surveyed general physicians, generalizations to other medical specialties, or to other health care systems should be cautious. Our results are also limited to the availability of resources to patients who have reached a physician in the first place. Finally, the response rate was modest, as is often the case for physicians [
29] and questionnaires addressing sensitive topics [
21]. Non-respondent bias is most likely to be associated either with lack of time, or with lack of interest with the topic. The latter could have led to an overestimation of scarcity, with a response bias in favor of physicians who were concerned with this problem. Reluctance to report an adverse impact on patient care could also have led to underreporting of scarcity and scarcity-related adverse events. However, extrapolating our results to a response rate of 100%, and considering all non-respondents to report no scarcity still results in a percentage of physicians reporting scarcity of 36%. One concern could be that the associations between variables could be affected by non-response bias. Variables independently associated with reported scarcity were reporting adverse events related to scarcity, and reporting less equity or more discrimination. If non-response were due primarily to lack of interest in the topic, then we could expect overestimation of adverse events related to scarcity, as well as overestimation of discrimination and lack of equity. As this would also likely be associated with overestimation of reported scarcity, however, the association between these two variables may not be affected.
Reports of scarcity in all the surveyed health care systems is not surprising. Every system in the world rations health care, some by wait times, some by availability of services, coverage decisions, or by ability to pay. There are thus good reasons for some resources to be unavailable, as choices will have to be made whenever demands exceed resources. Physicians are in a unique position to observe the impact of these choices, including when they may be unexpected. Our respondents' aggregate assessment of how various interventions were more or less sufficiently available differed across interventions, and between countries. Health care systems do not allocate their resources in identical ways; assessement of how existing services fit with perceived need, however, can be difficult. Despite growing research on variations in the distribution of resources in health care systems, and in utilization [
22], there is no gold standard on the proper availability of resources. Utilization is often used as a proxy outcome for availability, but making the distinction between utilization, need, and availability can be challenging [
30]. Availability is thus difficult to evaluate [
31]. In our study, we assessed unavailability of services based on physicians' assessment of need rather than on a measure derived from utilization. Physicians' situation at the point of care enables them to perceive discrepancies between need, and utilization, that may begin to serve as a more precise description of the actual availability of services. Their view may also contribute to an understanding of what a reasonable level of resources, or a more appropriate level, ought to be. Our results thus provide insights into the impact of different health care systems, with different structures and expenditures on health care. Mean scores on the scarcity scale were consistent with differences in national health expenditures. Where comparisons are possible, physician reports of scarcity based on our findings are supported by OECD mortality data, which yields identical rankings with regard to colon cancer screening and mental health services, and an almost identical ranking regarding rehabilitation for stroke (Table
4) [
32].
Table 4
Differences in reported unavailability is parallel to health outcomes
% respondents who reported unavailable rehabilitation for strokea
| 57 | 53 | 23 | 44 |
Potential years of life lost, cerebrovascular disease/100,000 p. >70 yearsb
| 89 | 74 | 58 | 121 |
% respondents who reported unavailable colon cancer screeninga
| 28 | 29 | 8 | 27 |
Potential years of life lost, malignant neoplasia of the colon/100,000 p. >70 yearsb
| 73 | 89 | 56 | 70 |
% respondents who reported unavailable mental health servicesa
| 37 | 80 | 58 | 53 |
Potential years of life lost, mental disorders/100,000 p. >70 yearsb
| 33 | 267 | 132 | 113 |
It would clearly be exaggerated to draw from this the conclusion that scarcity is the major cause of the differences in mortality reported here. For example, an alternative interpretation could be that physicians are more aware of problems related to diseases that are more prevalent. These comparisons, however, give construct validity to differences in the perception of scarcity between the four countries. If physicians were reporting different degrees of scarcity for, say, cultural reasons, we would not expect scarcity and disease-related mortality to be so parallel.
Although physicians's perception will be limited to situations where patients have reached them in the first place, their perception of scarcity may help to assess availability, a crucial element of access to health care [
31].
Despite universal coverage, physicians reported underinsurance. This should not be surprising. Universal health insurance means that coverage extends to all persons who legally reside in the country, as well as to foreigners in situations of emergency. It does not, however, necessarily mean that access to all interventions will be covered financially. For example, Switzerland and Norway mostly do not include coverage for dental care in health insurance. Neither does it mean that all included interventions will be covered without cost-sharing. As shown in table
1, this factor can vary extensively between the four systems. The extent of reported underinsurance was not related to the amount of national health care expenditure, suggesting that organizational factors and coverage decisions also contribute to apparent underinsurance.
Adverse outcomes attributed to scarcity were witnessed by most physicians, if infrequently. Some reported severe outcomes, such as death. This is concerning and warrants further research. However, it must be noted that we lack sufficient detail regarding the specific cases to formulate a judgment regarding the accuracy of this attribution, or its comparability across health care systems. The association between scarcity and reported adverse events may signify true lack of necessary resources. When extrapolated to the population served by general physicians, the estimate based on our respondents' report yields 0.15 scarcity-related deaths/1000 population [
33]. This is the same as the lower estimate, and 44% of the higher estimate, for deaths due to medical errors in the U.S [
34]. It may, however, also suggest greater sensitivity in the perception of scarcity by physicians who have been confronted with a possible adverse outcome. Either way, physicians reporting death as an outcome of scarcity are likely to be dissatisfied either with the level of resources in their health care system, with its distribution, or both.
Access was often reported as less than equal. More specifically, some patient groups were identified as more likely than others to be denied care on the basis of cost. Although the WHO distribution of responsiveness was identical in the four studies countries, Perceived equity was different in different countries, as was Perceived discrimination. Respondents thus perceived that access, viewed as a concern that "health care resources are mobilized to meet the needs of different groups in the population" [
35] was not fully realized. Respondents' views about equity did not vary in the same way as their views about discrimination did. Although physicians may be judging equity by standards different from the ones offered in our survey, a more likely explanation is that specific questions about patient groups were more likely to bring real cases to their minds. Thus, perceived discrimination may be a more sensitive tool to assess fairness in the distribution of health care resources. This finding also suggests that physicians, who are in a unique position to observe unequal access or discrimination in the health system, should be better equipped to address it. It is relatively easy for persons in a health care system to express a need for more resources but it is more difficult to develop an allocation process to ensure equitable distribution and resources allocated to a place to maximize benefit in terms of organizational or system objectives. Could physicians contribute to this? Data suggest that concerns for fairness are rarely explicit when physicians manage scarcity [
3]. More explicit thinking about fairness, and perhaps specific training, could enable physicians to make therapeutic decisions that enhance equitable access to medical resources. Concerns for fairness are applicable to clinical practice [
36]. In applying frameworks for fair resource allocation, implementing mechanisms for appeal and revisions [
37] would also give practitioners the opportunity to bring experience from clinical practice to bear on prioritization. Furthermore, our results suggest that efforts to measure a health system's equity might incorporate feedback from physicians about adverse events stemming from distributional decisions made at the system level. This feedback loop could be a way to connect the macro and meso levels of priority setting with the micro level.
Comparisons with other assessments of equity and utilization show some convergence. An OECD working paper evaluated General Practitioner care utilization to be pro-poor in all four countries included in our study, but specialist utilisation to be pro-rich in all of them [
38]. Reports by general physicians in our study that patients who cannot afford to pay for treatment are more likely to be denied care fits with those results. The degree of pro-rich inequity assessed by van Doorslaer and colleagues was highest in Italy, and lowest in the UK.
Our results suggest a link between perceived scarcity and perceived equity. Less equity was reported by physicians who attributed adverse events to scarcity, or more pressure to ration. More discrimination was perceived by those who reported more underinsutance or scarcity. This could mean that when there is less the most vulnerable are the first to get less. This view is both plausible and concerning.
Overall, however, physicians accepted cost-containment. Our results thus confirm that physicians are not fundamentally averse to such policies [
10,
14,
39]. Indeed, support was greater in our sample than in the study initially using the items we included [
14]. Respondents also indicated willingness to participate in these decisions: cost-containment policies close to the bedside were the most frequently approved. This suggests that physicians are not only ready to recognize that cost should play a role in allocating health care resources, but would rather participate in this sort of decision than not. If they are attentive to issues of fairness, they may be well situated to promote fair access to services in the face of resouce constraints.
Acknowledgements
The authors wish to thank Elizabeth Garrett-Mayer, for invaluable statistical support, Ole Norheim and Rein Vos for their review of the survey tool, Bruce Brinkley, Timothy Carey, Marie Neeser, Marc-Andre Raetzo, and Dan Sulmasy for their help in the pilot, the NIH Clinical Center Library for their outstanding translation service, MEDTAP international for their excellent data collection work, and Craig Mitton, Olli-Pekka P. Ryynänen, and Mark Messonnier for very thoughtfull comments in the review process. We also wish to thank all the physicians who took the time to complete the questionnaire.
This work was funded by the Department of Clinical Bioethics at the National Institutes of Health, and was conducted while SAH was a fellow at this Department. SAH was funded in part by the University Hospitals of Geneva, by the Oltramare Foundation, by the Centre Lémanique d'Ethique, and by the Swiss National Science Foundation. The views expressed here are the authors' own and do not reflect the position of the National Institutes of Health, of the Public Health Service, or of the Department of Health and Human Services. Nor are they necessarily those of the other funding sources. The funding sources were not involved in the study design, the collection, analysis, or interpretation of data, in the writing of the report or in the decision to submit the paper for publication.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All authors contributed to the conception of this paper, and to the acquisition of data. SAH wrote the first draft and all authors made important contributions to subsequent drafts. All authors have seen and approved the final version. SAH and MD had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.