Phantoms In The Snow: Canadians’ Use Of Health Care Services In The United States

Abstract

To examine the extent to which Canadian residents seek medical care across the border, we collected data about Canadians’ use of services from ambulatory care facilities and hospitals located in Michigan, New York State, and Washington State during 1994–1998. We also collected information from several Canadian sources, including the 1996 National Population Health Survey, the provincial Ministries of Health, and the Canadian Life and Health Insurance Association. Results from these sources do not support the widespread perception that Canadian residents seek care extensively in the United States. Indeed, the numbers found are so small as to be barely detectible relative to the use of care by Canadians at home.

PROLOGUE: Over the past three decades, particularly during periods when the U.S. Congress has flirted with the enactment of national health insurance legislation, the provincial health insurance plans of Canada have been a subject of fascination to many Americans. What caught their attention was the system’s universal coverage; its lower costs; and its public, nonprofit administration. The pluralistic U.S. system, considerably more costly and innovative, stands in many ways in sharp contrast to its Canadian counterpart. What has remained a constant in the dialogue between the countries is that their respective systems have remained subjects of condemnation or praise, depending on one’s perspective.

Throughout the 1990s, opponents of the Canadian system gained considerable political traction in the United States by pointing to Canada’s methods of rationing, its facility shortages, and its waiting lists for certain services. These same opponents also argued that “refugees” of Canada’s single-payer system routinely came across the border seeking necessary medical care not available at home because of either lack of resources or prohibitively long queues.

This paper by Steven Katz and colleagues depicts this popular perception as more myth than reality, as the number of Canadians routinely coming across the border seeking health care appears to be relatively small, indeed infinitesimal when compared with the amount of care provided by their own system. Katz is an associate professor in the Departments of Medicine and Health Policy and Management at the University of Michigan. Karen Cardiff is a research associate at the University of British Columbia’s Centre for Health Services and Policy Research. Also at the University of British Columbia are Morris Barer, professor and director at the Centre for Health Services and Policy Research’s Department of Health Care and Epidemiology, and Robert Evans, professor at the Centre for Health Services and Policy Research’s Department of Economics. Marina Pascali is a Dallas-based health care consultant.

Surprisingly few Canadians travel to the United States for health care, despite the persistence of the myth.

For more than a decade anecdotal reports of waiting lists for elective procedures in Canada and of hordes of Canadian “Medicare refugees” crossing the border in search of medical care in the United States have provided emotive fuel for critics of the Canadian health care system from both sides of the border.1 American opponents of universal public coverage have argued that global constraints on capacity and funding force many Canadians to cross the border in search of services that are unavailable or in short supply in their own country.2 Some have gone so far as to suggest that the widening health care spending gap between Canada and the United States is partly the result of counting expenditures by Canadian Medicare refugees in the U.S. rather than the Canadian expenditure totals, although there is an extensive body of evidence showing that the sources of the spending gap lie elsewhere.3

The Medicare refugee story is harnessed in Canada to promote the message that the Canadian health care system (known as Medicare) is chronically under-funded; the refugees are but one prominent symptom. The Canadian “under-fundists” are, however, divided as to the appropriate response. The many who support the fundamental principles on which Canadian Medicare is built argue that Canadian waiting lists and care seeking in the United States demonstrate the need for new public funds to increase capacity and services. While “evidence” in the form of Medicare refugees might be new, this debate about the level of public funding has been part of the dialogue between Canadian providers and provincial payers throughout Canadian Medicare’s history.4

But the putative refugees are also pawns in a debate driven by Canadian opponents of universal public funding, who wish to expand the role of private financing. This debate grew more intense during the 1990s as provincial payers increasingly constrained their health care budgets.5 News headlines suggesting that Canadians spend more than $1 billion annually south of the border have been cited to bolster the argument that private funding would reduce the pressure on the public system, thus reducing both public waiting lists and the flow of Canadians heading south for care. As a bonus, that $1 billion would stay at home.6

Unfortunately, this persuasive image of Canadian refugees survives in a virtual vacuum of evidence. How many Canadians actually head to the United States to seek medical care that they cannot obtain, or are unwilling to wait for, in Canada? What kinds of services do they receive? Where do they get these services, and how do they pay for them?

The paucity of answers to these questions is a result of large conceptual and empirical challenges facing researchers who attempt to fill in the gaps. Tens of thousands of Canadians enter the United States each year for a number of reasons unrelated to medical care seeking, such as holidays, business, education, or shopping. Any of these visitors might require medical care coincidentally while outside Canada. Thus, one must identify the context of Canadians’ medical care use in the United States to separate Medicare refugees from business travelers, “snowbirds,” and holiday seekers.

Paying for out-of-country medical care.

As part of a more widespread strategy to reduce public health care spending during much of the past decade, some provincial governments have imposed tighter limits on their financial liability for residents’ medical care received in the United States. Payment limits for emergency hospitalizations in 2000 varied somewhat across provinces: Per diem payments ranged from as little as Can$75 for residents of British Columbia to as much as Can$570 in Manitoba and Prince Edward Island. Outpatient emergency services are generally reimbursed at provincial fee-schedule rates, which are far below fees in the United States.7 But several provinces such as Ontario and Manitoba have also limited payments for outpatient emergency visits to as little as Can$50–$100. These restrictions have motivated more Canadians to obtain insurance for health care expenses incurred while traveling for extended periods in the United States.

In selected circumstances, more formal arrangements have been negotiated between provincial payers and U.S. providers. Provinces have always reimbursed individuals, subject to preapproval and negotiated payments, who are required to travel to the United States to obtain highly specialized services not available in their home province. More recently, several Canadian provincial payers have established temporary contracts with U.S. providers for specific services available but subject to unacceptable delay in Canada.

Research objectives.

In this study we attempt to quantify, across all sources of payment, the services provided to Canadians in U.S. regions located near the three most heavily populated Canadian provinces. Within these regions we examined data from two different types of sources: three states’ hospital discharge records and a survey of selected ambulatory care sites. In addition, we surveyed “America’s Best Hospitals” because they might serve as “magnets” for Canadians.8

Analytic framework.

Canadians might receive care in the United States for a number of reasons: (1) Services are available in Canada but often involve extensive wait times (wait-listed services). Examples often include magnetic resonance imaging (MRI), radiation oncology treatment, and selected surgical procedures such as total knee replacements, cataract surgery, and coronary artery bypass surgery.

(2) Leading-edge technology services are unavailable in Canada. Examples include gamma knife radiation and proton beam therapy for some cranial tumors and specialized programs to treat severe brain injuries.

(3) Services are available in Canada, but U.S. health care centers are more conveniently located for some Canadians (proximal services). Examples include some residents of rural border regions in Saskatchewan, Manitoba, New Brunswick, or western Ontario seeking primary care in U.S. settings; and some residents of urban centers such as Thunder Bay, Ontario, seeking secondary or tertiary care south of the border.

(4) Services are provided to Canadian snowbirds, who live in the United States during the winter months, or to other periodic business and leisure travelers to the United States (coincidental services).

(5) Services are available in Canada but are perceived by the patient to be of higher quality in specific U.S. medical centers such as those listed as one of “America’s Best Hospitals” (magnet services).

Across these categories, the sources of funding for care vary considerably. For example, patients in the fourth category will generally have their costs covered by varying combinations of provincial health insurance and private insurance. Services in the second category, approved by a provincial plan, would be paid in full by that plan at rates negotiated with the U.S. care center. Some services in the first and third categories may be provided under a contract between the provincial Ministry of Health and the U.S. providers. Other services in these two categories, as well as those in the fifth, require direct out-of-pocket payment by Canadian patients.

Sampling strategy and data collection.

From the American side.

Based on this framework, we developed a multiprong sampling and data collection strategy. We conducted a telephone survey in the fall and winter of 1998–99 of all ambulatory care clinical facilities located in specific heavily populated U.S. urban corridors bordering Canada (Buffalo, Detroit, and Seattle) that offered services that might be less available in Canada. These services included diagnostic radiology, ambulatory surgery, ambulatory eye surgery, cancer evaluation and treatment, and mental health and substance abuse treatment. Facilities performing these procedures were identified using a variety of federal, provincial, state, and local sources including local health care consultants and provider groups, the U.S. Federated Ambulatory Surgery Association, the American Hospital Association, the American College of Surgeons, and the SMG Marketing Group.

We performed a structured telephone interview of one or more key informants within the institution (typically senior personnel in billing, marketing, or public relations). Information collected included the number of Canadians who visited the institution in the prior year and whether there were any obvious trends, the nature of referral there, type of services provided, and methods of payment.

To examine inpatient care provided to Canadians, we acquired statewide hospital discharge data for 1994–1998 from Michigan, New York State, and Washington State. To differentiate care-seeking admissions from those related to coincidental activity, we categorized admissions according to admission status (emergency/urgent versus elective) and principal discharge diagnosis. Also, we attempted to contact key informants at each of “America’s Best Hospitals” to inquire about the number of Canadians seen in both inpatient and outpatient settings.

From the Canadian side.

We examined a number of different Canadian data sources to identify the extent of care seeking in the United States. We first analyzed data from the 1996–1997 National Population Health Survey (NPHS), a large survey representative of the Canadian noninstitutionalized population, that contained two questions pertaining to health care seeking in the United States. Respondents were asked: “In the past twelve months did you receive any health care services in the United States?” A positive response to the first question prompted a second one: “Did you go there primarily to get these services?”

An important potential source of Canadian patients for U.S. providers is formal contracts between them and provincial payers for specific diagnostic and treatment services. We identified the nature of these provincial contracts through personal contacts in the Ministries of Health of selected provinces. Finally, we spoke to the director of the Canadian Life and Health Insurance Association about the growth of out-of-country travelers’ emergency medical care insurance and insurance packages for services provided to Canadians in the United States on an elective basis. Unfortunately, one important source of Canadian data, provincial Ministry of Health expenditures specifically for out-of-country services, was insufficiently complete and comparable across provinces to be useable for this project. Remarkably, details such as patient demographics, types and dates of services, and location of U.S. providers are not being systematically tracked by most provincial Ministries of Health.

Study Findings, By Data Source

U.S. ambulatory facilities survey.

Almost 40 percent of the facilities we surveyed reported treating no Canadians, while an additional 40 percent had seen fewer than ten patients (Exhibit 1). Fifteen percent of respondent sites reported treating 10–25 Canadian patients, and only about 5 percent reported seeing more than 25 during the previous year (generally 25–75 patients; none reported more than 100). These findings were fairly consistent across the service categories. The overall response rate was 67 percent, and it varied across type of clinical facility from 56 percent for ambulatory surgery centers to 80 percent for cancer centers.

EXHIBIT 1

Number Of Ambulatory Health Care Facilities Reporting Having Treated Adult Canadian Residents In Michigan, New York State, And Washington State In The Prior Year, By Number Of Canadians Seen, 1997–1998

If we extrapolate these findings (assuming that nonrespondents show a pattern similar to that of respondents), these facilities in the three large metropolitan areas combined saw approximately 640 Canadian patients for diagnostic radiology services such as computed tomography (CT) scans or MRI and 270 patients for eye procedures such as cataract surgery over a one-year period. By comparison, the annual volume for CT scans and cataract extractions averaged about 80,000 and 25,000 procedures, respectively, in British Columbia alone during the mid-1990s.9 In Quebec the annual volume during the same period for CT scans and MRI averaged 375,000 procedures and 44,000 procedures, respectively.10

We also sought to examine Canadians’ use of mental health and substance abuse services in these same three U.S. catchment areas, because previous reports in the early 1990s suggested a cross-border flow of patients for these services.11 Because these regions have large networks of community mental health clinics, most of which do not regularly see patients from outside their community catchment area, we could not readily identify providers that would be the most likely targets for Canadian referrals. Therefore, we approached all such facilities that we could identify. Using the American Hospital Association’s guide to accredited freestanding substance abuse and mental health organizations, we identified thirty-two organizations in the Detroit area but only three in the Seattle area. We received responses to our telephone survey from twenty-three of the thirty-two organizations in Detroit (72 percent) and from all three of the Seattle sites. All but one reported seeing fewer than ten Canadian patients in the prior year, and none reported seeing more than twenty-five. In New York State the Office of Alcoholism and Substance Abuse collects data on treatment encounters at all centers in the state. From July 1997 through June 1998, 105,456 patients were seen, of which 246 were categorized as “other country.”

State hospital discharge data.

Over the five-year observation period from 1994 to 1998, 2,031 patients identified as Canadians were admitted to hospitals in Michigan; 1,689 to hospitals in New York State; and 825 to hospitals in Washington State. During the same period, annual inpatient admissions to hospitals within the bordering provinces of Ontario, Quebec, and British Columbia averaged about 1 million, 600,000, and 350,000, respectively.12 Thus, Canadian hospitalizations in the three U.S. states represented 2.3 per 1,000 total admissions in the three Canadian provinces. Furthermore, emergency/urgent admissions and admissions related to pregnancy and birth constituted about 80 percent of the stateside admissions. Elective admissions were a small proportion of total cases in all three states: 14 percent in Michigan; 20 percent in New York; and 17 percent in Washington.

Principal diagnostic categories.

The distribution of diagnostic categories varied by the type of admission (emergency/urgent versus elective) and by state. Diseases of the circulatory system and injury and poisoning accounted for 37 percent of all cases in Michigan, 39 percent in New York State, and 50 percent in Washington State (50 percent, 23 percent, and 21 percent, respectively, of all cases within the elective admission category) (Exhibit 2). Within the circulatory system category, the most common principal discharge diagnoses in all three states were acute myocardial infarction, cerebrovascular disorder, heart failure, and conduction disorders and arrhythmias. In New York State, admissions associated with digestive disorders (such as cholelithiasis, gastroenteritis/colitis, and appendicitis) represented 13 percent of emergency/urgent cases. In Michigan, admissions associated with mental disorders (schizophrenic disorders, affective/depressive disorders, and substance abuse) represented 20 percent of emergency/urgent cases, and the number of cases within this category was much greater than in either New York or Washington. However, we were unable to obtain further details from ministry or state sources. The remaining cases within the emergency/urgent category were distributed widely across principal diagnostic categories, and there was no consistent pattern across states. The distribution of elective cases across clinical categories was quite broad, with no consistent pattern across states.

EXHIBIT 2

Acute Care Hospital Discharges For Adult Canadian Residents In Three States, By State, Admission Type, And Principal Diagnostic Category, 1994–1998

America’s Best Hospitals.

Response from these institutions was low (eleven of twenty) and somewhat fragmentary. The numbers of Canadian patients seen in the prior year were generally very low: Six hospitals reported fifteen or fewer elective inpatients or outpatients; four hospitals reported 20–60 patients, and one hospital reported nearly 600 patients (90 percent outpatients and many related to proton beam radiation therapy for cancer).

Results from Canada.

Several sources of evidence from Canada reinforce the notion that Canadians seeking care in the United States were relatively rare during the study period. Only 90 of 18,000 respondents to the 1996 Canadian NPHS indicated that they had received health care in the United States during the previous twelve months, and only twenty indicated that they had gone to the United States expressly for the purpose of getting that care.13

Formal contracts.

Periodic formal contracts between provincial payers and U.S. providers have a long history, but a few such contracts have received considerable attention on both sides of the border.14 Most notable have been contracts for the provision of radiation therapy for cancer patients, in response to backlogs created by shortages of radiation technicians. For example, Quebec contracted with three radiation centers in Vermont and Maine in October 1999 for treatment of patients with breast and prostate cancer; 1,030 patients were treated during the subsequent year.15 Ontario contracted with three health care organizations in Michigan, New York, and Ohio in March 1999 to provide treatment for patients with breast and prostate cancer, and 1,416 patients had been referred as of 31 October 2000.16 This is equivalent to approximately 8.5 percent of all prostate and breast cancer patients treated with radiation therapy in Ontario during the same time frame.

Preapproval for stateside evaluation.

A relatively rare occurrence is preapproval for stateside evaluation of rare disorders or for experimental treatments not yet available in Canada. These treatments are often eventually adopted in Canada but diffuse less rapidly than in the United States. It is during that window between U.S. and Canadian adoption that occasional referral to the United States occurs. Examples of this include gamma knife therapy (a cobalt source is used to generate gamma rays that converge on a focal point) for treatment of cranial problems and brachytherapy (insertion of radioactive seed implants) for prostate cancer. Typically, a province the size of Quebec (approximately 7.3 million persons) may approve about 100 requests per year.17 Finally, in some provinces, contracts have been established between the provincial payer and U.S. primary care providers to provide primary care to residents of sparsely settled rural areas near the U.S. border. In New Brunswick (a province of 750,000 persons) this accounted for about 2,000 visits between 1996 and 1998.

Private insurance policies

Limits imposed since the early 1990s on out-of-province payments by provincial payers have motivated more Canadians to obtain travelers’ insurance for emergency out-of-province medical care. For example, the number of individual policies sold to Canadians increased from 700,000 to 2,800,000 from 1992 to 1999. However, we found no evidence that there is a demand in Canada for, or a supply of, insurance policies for elective medical care services.18 Some private insurance firms have expressed interest in offering policies that would provide service in the United States if one had to wait more than thirty days on a Canadian waiting list; however, there has been no apparent demand for such policies to date.

Discussion

A tip without an iceberg?

This study was undertaken to quantify the nature and extent of use by Canadians of medical services provided in the United States. It is frequently claimed, by critics of single-payer public health insurance on both sides of the border, that such use is large and that it reflects Canadian patients’ dissatisfaction with their inadequate health care system. All of the evidence we have, however, indicates that the anecdotal reports of Medicare refugees from Canada are not the tip of a southbound iceberg but a small number of scattered cubes. The cross-border flow of care-seeking patients appears to be very small.

Our telephone survey of likely U.S. providers of wait-listed services such as advanced imaging and eye procedures strongly suggested that very few Canadians sought care for these services south of the border. Relative to the large volume of these procedures provided to Canadians within adjacent provinces, the numbers are almost indetectable. Hospital administrative data from states bordering Canadian population centers reinforce this picture. State inpatient discharge data show that most Canadian admissions to these hospitals were unrelated to waiting time or to leading-edge-technology scenarios commonly associated with cross-border care-seeking arguments. The vast majority of services provided to Canadians were emergency or urgent care, presumably coincidental with travel to the United States for other purposes. They were clearly unrelated either to advanced technologies or to waiting times north of the border. This is consistent with the findings from our previous study in Ontario of provincial plan records of reimbursement for out-of-country use of care.19 Additional findings from the current study showed that a small amount of cross-border use was related to proximal services, primarily in rural or remote areas where provincial payers have made arrangements to reimburse nearby U.S. providers. Finally, information from a sample of “America’s Best Hospitals” revealed very few Canadians being seen for the magnet referral services they provide.

These findings from U.S. data are supported by responses to a large population-based health survey, the NPHS, in Canada undertaken during our study period (1996). As noted above, 0.5 percent of respondents indicated that they had received health care in the United States in the prior year, but only 0.11 percent (20 of 18,000 respondents) said that they had gone there for the purpose of obtaining any type of health care, whether or not covered by the public plans.

Was our net fine enough?

This study might have underestimated the number of Canadians seeking care in the United States, for several possible reasons. First, a number of institutions did not respond to our survey. Those institutions might have seen larger numbers of Canadian patients than did the institutions that responded. However, persons contacted at nonresponding sites suggested to us that in fact they simply had nothing much to report. Second, we may simply have asked the wrong institutions and collected hospital data from the wrong states. It is possible that Canadians found their way to more remote sites not identified as magnet institutions. Indeed, we know that many Canadians receive care in Florida and California, for example. However, these are predominantly coincidental services. We could determine no logical reason why Medicare refugees would go further afield or to less prominent sites. Finally, it is possible that surveyed providers and administrative data did not recognize Canadians because they were using local addresses. This would be a limitation on any study of U.S. providers, for which the only possible remedy would be a costly individual patient survey. However, we have no information that would suggest that Canadians who seek care in the United States are likely to have U.S. addresses.

On the Canadian side, the surprisingly poor quality of some of the provincial data leaves open the possibility that some patients heading south for contracted services reimbursed by the public plans may have been missed if they were cared for in facilities that did not participate in our stateside survey. However, earlier analysis of Ontario Health Insurance Plan (OHIP) data found that most spending for medical and hospital services received by Canadians in the United States during the early 1990s was related to the “coincidental” basic and emergency health care services typically used by Canadians traveling or temporarily residing in the United States.20 Although the possibility of underestimating cross-border care seeking can never be entirely eliminated, we do not believe that its magnitude would be sufficient to challenge our conclusions.

Why is cross-border care seeking so low?

Our results should probably not, on reflection, be surprising. Prices for U.S. health care services are extraordinarily high, compared with those in all other countries, and this financial barrier is magnified by the extraordinary strength of the U.S. dollar. Private insurance for elective services, being subject to very strong adverse selection, is, not surprisingly, nonexistent. Discussions with key informants in the Canadian private insurance industry indicated that carriers correspondingly confine themselves to the coincidental services market. Furthermore, provincial governments have been lowering their rates of reimbursement and tightening preapproval criteria for cross-border care. In the absence of either source of health insurance coverage, it would be somewhat surprising if large numbers of Canadians were choosing to head south and pay out of pocket for care. In fact, one recent survey found that Canadians were not even prepared to pay out of pocket in their own country to reduce their own waits.21

What about Canadian contracts with U.S. providers?

The numbers of true medical refugees—Canadians coming south with their own money to purchase U.S. health care—appear to be handfuls rather than hordes. But there are still the highly visible examples of Canadian provincial governments contracting with U.S. providers for specific services that are unavailable or in short supply in Canada. While these contracts have received extensive press coverage on both sides of the border, they have largely been short-term arrangements for a limited number of procedures for selected patients experiencing delays in several Canadian provinces.22 Do such purchases indicate that the Canadian health care system is inadequate to meet the needs of its citizens and is critically dependent on access to the better-resourced U.S. system?

Well, yes and no. In the case of highly specialized and leading-edge or experimental technologies, this contracting policy is obviously sensible. It would be impossible for a country one-tenth the size of the United States (much less individual provinces) to try to maintain the capability to offer every conceivable form of care, no matter how advanced or unusual. Purchasing such services from a small number of U.S. tertiary centers that offer them, as indeed many U.S. payers do, is the only reasonable option. As and if the technology matures and its range of applicability expands, it may be disseminated to Canadian centers.

Cross-border contracting for services to augment existing Canadian capacity for commonly used technologies raises somewhat different issues. An important cost containment strategy in Canada has been constraint on the capacity of diagnostic- and treatment-related technology. Tight capacity is particularly vulnerable to unexpected surges in demand for care or a sudden loss of supply attributable to, for example, a strike by critical support personnel. The consequence is increased waiting times that at some point may be perceived as excessive by providers, patients, or the public. Selective contracting with U.S. providers has been a response to these concerns.

A case for long-term contracts.

As long as Canadian capacity remains tight for selected medical technologies while at the same time the United States continues to generate excess capacity, cross-border contracting appears to be a perfectly sensible approach to dealing with patient queues. It also offers a way of delaying capital investments in response to shifts in patterns of clinical practice until these have had time to establish themselves. As a purely economically motivated “make or buy” decision, it might even make sense to enter into long-term contracts for the purchase of services in the United States, as long as these contracts were available at prices above U.S. marginal cost but below the Canadian average unit cost. Such contracts would reflect not a “failure” of the Canadian system but simply provincial governments’ behavior as a “prudent purchaser,” taking advantage of the opportunity to “buy” more cheaply than it could “make.” Americans would also benefit. As long as their health care system is organized to generate excess capacity, they are clearly better off if the excess capacity is sold to Canadians than if it is left to sit idle or used to generate unnecessary domestic servicing.

The case against.

But there are other important considerations that would be raised by a long-term Canadian policy of importing health care services from the United States, even at favorable prices. First, patients may resist absorbing the monetary and nonmonetary costs of travel to the United States. Second, Canadian purchasers of U.S. services may be most vulnerable to loss of a contract or increased prices if U.S. domestic demand surges or supply decreases. Third, solving the problem of Canadian waiting lists by sending a regular wave of patients south would imply a major loss of income for Canadian providers. For all three reasons, this policy would be largely unacceptable to providers and patients, and, as a result, politicians would likely face an ongoing chorus of accusations that the system fails to meet the medical needs of their constituency.

Phantoms in the snow.

Despite the evidence presented in our study, the Canadian border-crossing claims will probably persist. The tension between payers and providers is real, inevitable, and permanent, and claims that serve the interests of either party will continue to be independent of the evidentiary base. Debates over health policy furnish a number of examples of these “zombies”—ideas that, on logic or evidence, are intellectually dead—that can never be laid to rest because they are useful to some powerful interests.23 The phantom hordes of Canadian medical refugees are likely to remain among them.

Footnotes

  • The authors acknowledge financial support from the Canadian Institutes of Health Research (formerly the Medical Research Council of Canada) for this research.

NOTES

| Table of Contents