Location of residence associated with the likelihood of patient visit to the preoperative assessment clinic

Data sources

Hospital discharge data were used to identify all patients who underwent surgery at the Foothills Hospital, in Calgary, Alberta, between July 01, 1996 to March 31, 1998. Among these, we included for study all patients residing in the province of Alberta (whether inside or outside of Calgary city limits), as well as patients living in the neigbouring provinces of British Columbia and Saskatchewan, because those provinces typically refer some patients to the Foothills Hospital in Calgary for tertiary care services. Individuals who lived beyond these areas were excluded since they were residents of other provinces and unlikely to be typical geographically proximal tertiary care referrals; rather, we suspect that such cases represent individuals having surgery in Calgary because of personal ties to the city (e.g., friends or relatives residing in the city, in the context of permissible out-of-province surgery through Canada's inter-provincial portability of health insurance).

The surgical specialties of general surgery, cardiovascular/thoracic, gynecology, neurosurgery, orthopedics, plastic surgery, otolaryngology, urology, and oral surgery were included. Given the small number of patients in some specialties, the divisions of plastic surgery, otolaryngology, urology and oral surgery were combined into one group.

Discharge data were used to capture patient demographics, assessment and discharge dates, urgency of surgery (i.e. surgery that is necessary to mitigate what would otherwise be an imminently threatening medical condition), International Classification of Diseases, 9^th Edition, Clinical Modification (ICD-9-CM) diagnostic and procedure codes[13], most responsible surgical specialty, and patient postal codes. Those who underwent emergent surgery, day procedures, or in the rare case had surgery cancelled after admission because of patient or surgeon decision, were excluded. Cases with a missing procedure or postal code were also excluded.

This study was approved by the Conjoint Health Research Ethics Board of the University of Calgary.

Variables

Analysis

The likelihood of a patient visiting the preoperative assessment clinic before surgery was examined for each patient as a function of geographical distance from place of residence to preoperative assessment clinic. Graphical examination of the proportion of patients who visited the preoperative assessment clinic was undertaken using 50 km categories. Prior to categorizing distance, we examined distance as a continuous variable and noted that the relationship between distance and visit to the preoperative clinic was not linear and hence violated the linearity assumption required for treating distance as a continuous variable. Categorical data were thus analyzed with chi-square tests and analysis of variance was used to compare age across distance strata. Multiple logistic regression models were used to calculate the adjusted odds of a patient visit to the preoperative assessment clinic as a function of distance, while controlling for other covariates. In order to calculate the crude odds of a preoperative visit, only distance categories were entered into the model. The adjusted odds of a preoperative visit were calculated by forcing all distance categories into the model and adding age, sex, major versus minor surgery, and urgency of surgery and all 21 comorbidity variables. As recommended by Sun [19], backward elimination was undertaken to remove non-significant (p > 0.05 Wald test) covariates one at a time. The adjusted odds of a preoperative assessment clinic visit for only non-cardiac procedures were also calculated using the same modeling procedures. This analysis was undertaken because it was suspected that cardiac surgery referral practices in the region studied may differ from other surgical specialties [14]. We also conducted an analysis stratified by surgical division, to determine the adjusted odds of referral by distance category for each surgical division.

Unadjusted utilization rates

Among patients who lived between 0 – 50 km from preoperative assessment clinic, 66% attended the clinic (Figure 1). The proportion of patients who attended the clinic decreased to 52% for those living between 51 – 100 km. A further reduction was observed for each consecutive 50 km increment, resulting in 39% for 101–150 km, 40% for 151–200 km, 30% for 201–250 km, and 34% for distances greater than 250 km.

Clinical characteristics and risk factors

Distance and adjusted utilization rates

Distance and utilization by surgical specialty

Implications

The significance of these findings has implications for a considerable portion of the population who rely on health services in the regional tertiary care centre that we studied. The province of Alberta has a population of approximately 3 million. The two metropolitan health regions in the province (i.e. the Calgary Health Region and the Capital Health Region in and near Edmonton) provide health services to approximately 1 million people each, constituting 67% of the total provincial population. The remainder of the population (i.e. 1 million) lives outside of these two immediate metropolitan areas. In our study, 26% (N = 2512) of the 9506 surgical patients lived further than 50 kilometers from the Foothills Hospital, which is the main tertiary care facility in the Calgary Health Region.

A sizable literature exists on the importance of preoperative assessment and potential benefits to both patients and the health care system. Further, preoperative assessment is recognized as an important discipline in medicine. Previous studies have identified that clinical and other factors are important in the referral and utilization of preoperative assessment clinics. In a study by Bugar et al. [20], clinical factors were strongly associated with patient referral and utilization of the preoperative assessment clinic. Further, surgical specialty and type of clinic consultation were also factors in patient referral and utilization. For example the overall utilization rate of the preoperative assessment clinic for general surgery patients was 72%, while the consultation rate for this patient group was 19% for general internists and 39% for anesthesiologists, whereas overall utilization for neurosurgery was 63%, while the consultation rate for this patient group was 24% for general internists and 19% for anesthesiologists. Our study was designed to identify whether patient distance from the preoperative clinic was also an important factor, independent of such clinical factors. The results of our study suggest that patient distance from the clinic is indeed an important factor, and is in fact as important as clinical factors. Further, Table 1 displays how specific clinical factors play out by distance categories

Inequitable geographic access has significant implications given the identified and potential benefits of preoperative assessment for patients and the health system. For instance, decreased costs due to a reduction in laboratory testing, and a decrease in delayed or cancelled surgical procedures were reported benefits to the health systems in the United Kingdom and United States [1–3]. Although a decrease in clinical outcomes such as perioperative complications has not been substantiated as yet, studies have indicated that this is a potential benefit but one that is difficult to confirm [1, 5]. Further, the preoperative assessment clinic provides patient-centred care. Patients and family members have an opportunity to discuss their concerns, medical risks, lessen anxiety, and obtain information about their surgery [5, 20].

A second benefit of the preoperative assessment clinic is improved clinical documentation. Enhanced reporting of clinical information in patient charts can assist health professionals in making optimal perioperative management decisions. Good clinical documentation in records also provides the essential data needed for ICD-10 coding, costing, and billing [5, 20]. Preoperative assessment clinics introduce a streamlined process for the patient and the health system through the use of standard assessment forms, provision of diagnostic services in one locale, and more complete medical records.

If we conclude that the inequitable access is problematic, what might be some of the solutions? One possibility is to ask all patients to travel for preoperative assessment clinic. This may or may not be acceptable to the general public because of personal, family, occupational, or financial reasons. Alternatively, satellite preoperative assessment clinics could be established, but this would obviously have staffing and health system cost implications. However, this may be a more appropriate option in a health system like Canada's that strives to achieve universal access, compared to the alternative above that would result in a shifting of costs from the health system to the patient. The burden of these extra costs on rural and remote residence can be significant, as demonstrated in an Australian study that examined the costs of accessing a surgical specialist [21]. Patients, who accessed a local as opposed to a metropolitan surgical specialist, were able to save an average of $1077 AU in out of pocket costs per specialist visit.

Another potential solution is the use of telemedicine technology. As telemedicine becomes more widespread, this alternative may be increasingly viable with time [22, 23]. Once again this would have staffing, training, funding, and physician compensation implications.

Another 'remote triage' solution that is used in some settings is to have an anesthesiologist contact the patient at home by telephone to get a sense of whether a patient needs specialist consultation or specific tests prior to their surgery. Alternatively we could simply accept that preoperative assessment clinic assessment is not feasible for remote patients. However, some patients from remote regions might object to this status quo. This option is also of concern given that patients are being encouraged to participate in their own medical management as health care moves toward patient-centered care. Also, many surgeons might prefer or demand preoperative assessment clinic consultation prior to performing surgery. Patients could be seen immediately prior to surgery, eliminating an additional trip for the patient, and gaining some benefits such as patient centred care and documentation. However, the benefit of avoiding the cancellation of surgery would not be attainable. The development of referral guidelines to assist surgeons in deciding which patients should be sent to the preoperative clinic prior to surgery would also be helpful. The surgeon would be better prepared to identify and consult patients living in remote areas regarding the need for further medical assessment regardless of the patient's distance from the clinic.

Our initial study identifies a need for further inquiry into this complex referral and utilization process, to gain insight into stakeholder decision-making. For example, a survey would be helpful to identify the general public's willingness for extra travel, as well as objections to being "passed over" for preoperative assessment clinic. Surgeons could be asked about their willingness to forego preoperative assessment clinic or their tendency to simply skip preoperative assessment clinic for remote patients where they otherwise might refer them to preoperative assessment clinic. Consulting internists are also important stakeholders who could be questioned about their willingness to participate in satellite preoperative assessment clinics or work through telehealth.

Study limitations

Our study has several limitations, the first of which is that we only studied one preoperative assessment clinic in a single health region in one province. Our findings may not apply to other health regions or provinces, although past studies have found similar distance decay effects for other health services [7–12]. Secondly, our study was undertaken in a single universal insurer health care system, and hence may not apply to other countries. However, studies from the United States and United Kingdom that examined geographical access also identified a decrease in utilization with increasing distance from the health service [7–12, 24, 25]. It is possible that patients at large distances from the preoperative assessment clinic received some form of preoperative assessment from their local physician or specialist outside of the clinic. Given that we used administrative date, we were limited for the most part in our ability to capture this information. However, the chance that preoperative assessment would have in fact taken place at the local level is likely low and would be atypical at this centre with the exception of patients scheduled for a cardiac procedure. Typically these patients see a cardiologist before attending the preoperative assessment clinic. Despite this however, we noted in our study that cardiac surgery patients did not appear to be as affected by distance as patients undergoing other surgical procedures.

There may also be other non-clinical confounders that we were unaware of and hence unable to capture or control for in this study that may have influenced referral to preoperative assessment clinic. For example, patients living in remote areas may have refused or were unable to travel to the preoperative assessment clinic.

Our use of straight-line distance to measure geographical access to the preoperative assessment clinic has some limitations. Research has shown that road network distance measures or travel times to a hospital more closely reflect 'true' distance because they take into account geographical and physical impeding structures such as roadways, mountains, rivers, etc. [26]. For these reasons, network and travel distances typically contain fewer errors and result in longer distance measures. As well, it should be noted that consideration of these features is likely less important in urban areas that are typically setup on a grid system, than in rural areas where geographic and physical structures are more prevalent [26]. However, the choice of whether to use a simple straight-line distance calculation versus network distance depends on the type of question under study. We were interested in the relative, rather than 'true' magnitude of distance and its effect on patient visits to the preoperative clinic. Further, our choice of distance measure was based on the assumption that straight-line distance is proportional to road network distance, as demonstrated in several studies [27, 28]. It should also be noted that the use of centroids to approximate the location of patients introduces an additional source of error since these do not refer to the actual address of the patient. The amount of error introduced can vary depending on whether the patient lives in an urban or rural location [29, 30].

Yet another caveat is that our study only examined actual utilization of the preoperative assessment clinic. Although it is possible that some of the referred patients may not have actually attended the clinic, the booking procedures for non-emergent surgery in the hospital studied are such that the vast majority of referrals actually lead to a clinic visit (– because planned surgical procedures are usually delayed or even cancelled if a patient does not attend the preoperative assessment clinic after a referral has occurred). On a final note, we grouped the specialties of urology, plastic, oral and otolaryngology surgery recognizing that relatively small number of surgical cases performed by each of these divisions would yield statistically unstable point estimates for the PAC clinic visit odds ratios that we present by division. By grouping these small divisions into a single combined grouping of "other" surgical divisions, we found that the relationship between patient visit and distance from the clinic still generally holds for these small surgical divisions.