Home visits by family physicians during the end-of-life: Does patient income or residence play a role?

In the last ten years, more and more of those dying of cancer in Canada are doing so out of hospital[1]. In Nova Scotia, a Canadian Maritime province with a total population of approximately 950,000 people, the proportion of cancer deaths occurring out of hospital has recently grown by fifty per cent[2]. This trend appears to be associated with a number of factors. More individuals with cancer are choosing to remain in the home setting, hospitals have down-sized thus reducing the number of beds available for end-of-life care [3–5], and there is growing availability of services in the community such as homecare and community based palliative care programs [6–8].

As this trend has developed, it has become even more important for patients and families to have access to medical care in the community. Such first line medical care, in Canada, is generally provided by a family physician, usually previously known to the patient. Initially those with terminal illness will obtain their medical care in the office or out patient setting. As they become sicker, however, there will come a time when getting from the home to the clinic office will be too difficult. At such a time, access to home visiting by a physician becomes very important [9–11]. Research has shown that access to a supportive family physician willing to make home visits is associated with a greater likelihood of a home death [12–14], as is access to a comprehensive palliative care program (PCP)[2].

There is evidence that those better off financially live longer and are in better health than poorer individuals[15, 16]. Reasons for this have been postulated to include the fact that those with higher incomes have higher educational achievement, better living circumstances, and less risky health behaviours. Such better health may also be due, in part, to better access to services for those with fewer financial barriers. In Canada, such an association should not be the result of inadequacy of health services provided to those with lower incomes as our federal government has committed to the provision of a universal, accessible, comprehensive publicly administered health insurance system which aims to ensure that all residents have access to necessary hospital and physician services on a prepaid basis[17].

We wondered if access to terminal care home visiting by family physicians is better for those with higher incomes even in our publicly-funded health system. Therefore, the purpose of this study was to examine the association between income and the likelihood of receiving home visits by family physicians during the end-of-life among those with cancer. In addition, we examined the effect of regional residency, specifically residency in a major urban centre versus all other regions of NS (which are much smaller in size).

Subject characteristics and health service utilization by receipt of at least one home visit are displayed in Table 2. In addition to patients who received at least one FP home visit tending to reside in higher income neighbourhoods, they also were more likely to be female, have a breast cancer cause of death, survived at least 61 days from their initial cancer diagnosis date, did not receive a FP visit within a LTC facility, made more than 14 specialty visits, spent 13–31 days as a hospital inpatient during their survival time, received palliative radiotherapy, and were admitted into the PCP.

Regression results incorporating the total number of home visits using negative binomial regression and those derived from logistic techniques assessing the log odds of receiving at least one home visit compared to none proved similar. Therefore, for ease of presentation, we present the logistic regression results only.

Among all advanced cancer patients, additional factors predictive of receiving at least one FP home visit included a longer length of survival (for 121 to more than 180 days survival: adjusted OR 4.04; 95%CI 3.45, 4.72), admission to the QEII Palliative Care Program (PCP) (adjusted OR 2.25, 95%CI 1.97, 2.56), older age (for those 75 years and older: adjusted OR 1.41; 95%CI 1.24, 1.61) and being female (adjusted OR 1.15; 95%CI 1.04, 1.28). Patients who were in LTC at some point during their survival period (adjusted OR 0.55; 95%CI 0.46, 0.65) and those who spent 32 or more days in hospital compared to none (adjusted OR 0.80; 95%CI 0.67, 0.95) tended to be less likely to receive at least one FP home visit. Over time, the likelihood of receiving a home visit tended to decline. However, after accounting for all other predictors in the model, year of death was not a major factor.

Cancer cause of death, receipt of physician specialty visits and undergoing palliative radiation were not associated with home visit receipt in the final multivariate model.

The neighbourhood income of those dying of cancer is associated with the likelihood of receiving a home visit during the end-of-life by a family physician in Nova Scotia. The association found, however, appears to be modified by region of residence for those who died of cancer. It appears that income plays less of a role in predicting home visits by a family physician for those who live in the larger, urban centre of Halifax Regional Municipality. Given the finding that patients followed by the QEII Palliative Care Program are also more likely to receive family physician home visits, we speculate that the urban centre may provide a collaborative 'team-care' advantage to cancer patients. The publicly-funded PCP may act to equalize the opportunity to stay at home and facilitate family physician home involvement in ways rural locations may not be able to. In previous work, income was not associated with location of death but region of residency was[2]. We have also found that those who live in higher income areas tend to use the emergency department less[24].

In the United Kingdom, Aylin and colleagues[26] found, for the general population, those in social class 1 (highest income) received the fewest home visits. Their study also revealed a dose-response relationship in that as one moves to lower income class, the more likely one is to have received a home visit. Our study shows some gradient element, albeit in the opposite direction, but not as clearly. McNiece and Majeed found home visiting rates among patients with highest income to be half that of those with the lowest income[27]. Their study results and ours were adjusted for age and sex, such that the relationships between income, age and sex cannot be confounding the results. Aylin postulates that the reason for greater home visiting among those with lower income may be due to a number of factors including increased morbidity, poorer access to a car, and differing expectations of the services supplied by their general practitioners[26]. Such factors should also hold true for cancer patients. Nevertheless, our findings are opposite to those of Aylin.

We hypothesize that when it comes to routine home visits for brief, episodic illnesses, the home visiting trend may be as Aylin suggests. However, for those who are at home and looking to stay at home with advanced cancer, there are more substantial financial issues driving whether this is likely to happen or not. In Nova Scotia, as in many Canadian settings, there is access to home visiting nurses and some other health professionals through the publicly-funded health system. However, as disease progresses, the ability of a family to support death at home depends on many other factors. These include the presence of a family member who can stay at home, the ability of a family member to manage the medications and symptom assessment along with health professionals, the cost of drugs which are paid for in hospital but not in the home (unless the patient has private health insurance or is 65 or more years old), the cost of equipment in the home, and possibly the cost of additional nursing or personal care workers in the home (variably covered by the public system, and only sometimes covered by private insurance)[17]. All of these factors point to the fact that those with greater income would be more likely to succeed at staying at home[7]. In addition, in more rural settings there is less access to specialized services such as palliative care programs.

The home visits provided by family physicians may therefore be a direct response to the other capacities of patients and families to stay at home rather than being the critically enabling feature. Thus, the home visit is essential but not sufficient and without the rest of the support required, the patient will not be able to stay at home, thereby resulting in fewer home visits.

Another interpretation is that those with lower incomes may actually make different choices about how they wish to receive health care and where they would like to spend their last days. Depending on the study cited, up to 80% of those with advanced cancer wish to spend their last days at home[7, 28–30]. Grande and colleagues reported that those who lived in higher socioeconomic areas were more likely to die at home than their counterparts[31]. Sims found that those with cancer from social class IV and V (semi-skilled and unskilled occupations) were under-represented in deaths that occurred in hospices and homes when compared to those in social class I and II (professional occupations and managerial/technical occupations)[32]. All of this may be supported by any one or a combination of factors such as less desire to remain at home, less capacity (financial or otherwise) to remain at home or bias in the delivery of health service by professionals. Our study is the first we know of to show that the number of home visits made by family physicians to those at the end-of-life is also less for these individuals.

Home visiting has long been an element of continuity of care across settings (office, hospital, home, nursing home) provided by family physicians. Some would argue that home visits may be influenced by the geography in which the physician operates daily. As a result, physician travel patterns to and from the office (when home visiting often occurs) may not take them through low income neighbourhoods, thus reducing the likelihood of a visit. The work of Aylin[26] and McNiece[27], however, does not support this.

New initiatives are underway in Canada which may provide more opportunities for the enhanced presence of a range of health professionals in the homes of the dying. In response to the Romanow Commission[33], the federal government of Canada has initiated agreement with the provinces for them to provide coverage for enhanced home-based end-of-life care. In the future, we may see more nurses or nurse-practitioners making home visits as part of the community-based care team along with family physicians[34]. In rural or remote areas where there is a scarcity of family physicians, nurses and nurse practitioners with advanced assessment skills may play an even greater role.

Our study has limitations. As we are using routinely collected data used for administrative and billing purposes there may be biases operating. The data reflect those family physicians who bill for the services they provide. It should also reflect the "shadow-billing" of those on alternate payment mechanisms (estimated to be less than five per cent of family physicians at the time of the study) but in reality, these physicians may have less incentive to capture these fee codes and so may under-report home visiting slightly.

The data file used in this study was originally created for an alternate project looking at health service utilization among patients who died due to lung, colorectal, breast or prostate cancer. We were therefore limited to examining home visits provided to these patients only and are not able to report whether the use of family physician home services among those who died due to all other cancer causes is similar or different.

We are unable to adjust for homecare utilization (data did not exist for the study years), family member caregiving status (no data available) or account for additional insurance coverage (above provincial) which may have covered additional costs associated with drugs, home nursing, home equipment, etc. Our attempt to account for service availability by region is crude. HRM is more homogeneous with respect to services than our combination all other regions outside of HRM; however, the effect evidenced may, therefore, be a conservative estimate.

And so, in conclusion, it appears that even in death those with fewer financial resources may be less likely to achieve the same access to health services as those better off. What does this mean for our care of the dying? We must examine carefully which elements are the cause of this inequality. If there is less desire among those with more financial barriers, we need to examine the origins of these desires. Is it fear of caring for those at home? Is it an established culture of caring to move loved ones to hospital at the end-of-life? These issues need identification if we are to support such families. If there truly are financial barriers to such things as drugs, equipment and personnel then we must redefine health policy to make these more accessible to those with fewer financial resources. And finally, we should attempt to understand whether or not there is any bias operating on the part of health professionals. All of these need further research if we are to ensure patients receive the care where they wish to as they approach death.