This paper used Google Trends to explore search patterns for walk-in clinics and emergency departments (EDs) in Canada. High-quality, accessible primary care is central to the effectiveness and efficiency of health care systems [1,2]. In Canada, primary care is intended to be the first point of contact with the health care system and provides access to referred health care services and coordination of care [3]. However, access to primary care continues to be a challenge for many Canadians, with Canada ranking below average in cross-national surveys of primary care access. For example, according to the Commonwealth International Health Policy Survey for adults in 11 countries, Canada ranks below average on timely access to care with only 43% of Canadians reporting to have gotten a same- or next-day appointment at their regular place of care and only 34% reporting to have access to after-hours care without going to the emergency room [4]. In the absence of access to care with a regular provider, patients may turn to EDs or walk-in clinics as alternatives.

EDs provide crucial lifesaving health care and are designed to provide episodic care to patients with injuries or acute illnesses, though patients may also turn to them for primary care services that are otherwise inaccessible or inconvenient [5-7]. The ED setting is also a costly setting to deliver chronic and continuing care and may lead to mishandling or poor attention to upstream social and chronic health issues of frequent ED users within the episodic ED setting [8]. Walk-in clinics also provide primary care services to patients without an appointment [6]. They may be a more convenient alternative for patients with another regular source of primary care [9] or the only source of care for patients without a regular primary care provider. As with EDs, there are concerns that walk-in clinics may disrupt continuity of care or neglect preventive health care [6] and lead to duplication of services, failure to manage complex care needs, and higher costs [9]. Understanding patterns of walk-in clinic use is therefore important in planning primary care policy more broadly.

In Canada, primary care is financed and organized provincially, and provinces have undertaken quite varied approaches to primary health care reform, including changing models of care, physician remuneration, and after-hours access programs [10-12]. The province of Ontario has adopted the use of various models of interprofessional care, including family health networks and family health teams [13]. Western provinces and Quebec have implemented different organizational reforms, including family medicine groups, local services networks, medical homes, and primary care networks [14-17]. The proportion of clinical payments through capitated and salaried payment models is higher in Ontario, eastern provinces, and the territories, whereas the western provinces and Quebec have a higher proportion of fee-for-service payments [14,15,18]. Walk-in clinic practice may be more common under a fee-for-service model as physicians are paid per service without ongoing responsibility for specific patients. Provinces also have varying mechanisms for after-hours coverage. Most have a toll-free number that connects patients to a registered nurse who can provide advice and direct patients to after-hours care [10,14]. Under capitated models, physicians may receive negation if their patients access services from another primary care provider, including walk-in clinics. This variability in the organization and delivery of primary care may shape availability and use of walk-in clinic services, and also use of EDs for conditions that could be treated in primary care.

Despite the importance of monitoring the use of walk-in clinics and EDs, there are no comparable nationwide data sources to track their use. Although each province collects data on use of primary care services, there are no standard approaches to identify walk-in practices [19,20]. The Canadian Institute of Health Information collects some standardized data on ED use through the National Ambulatory Care Reporting System. Though coverage is increasing, this is currently available for only a subset of facilities in Canada [21].

Access to health care starts before the clinic door, and patients may use search engines like Google to navigate access to needed care, including walk-in clinics and EDs. They may use Google searches to identify sources of care, confirm location, check hours, or obtain other information such as wait times. Though not a direct measure of health care use, relative search frequency may provide insight into patterns of care seeking across provinces and over time.

The application of internet data provided by tools like Google Trends in health services research has potential to complement and extend the data that presently exist, collected primarily through survey and administrative sources. For example, internet data have been used quite extensively in epidemiology and public health research for the surveillance of infectious disease outbreaks, determining patterns and seasonality of disease incidence, and examining information seeking about health conditions [22,23]. Application of Google Trends in health care and health services research has been more limited but has included information seeking on addiction treatment programs and correlation between Google searches for dementia and Alzheimer diseases, and outpatient visits [21,24-29]. A systematic review of studies using Google Trends in health care research found a 7-fold increase in publications from 2009 to 2013, with 27% of the studies on infectious diseases, 24% on mental health and substance use, 16% on other noncommunicable diseases, and 33% on general population behavior [22]. However, to our knowledge, no studies have used Google Trends to examine searches for walk-in clinic or ED services.

In this paper, we used Google Trends to compare search frequencies for walk-in clinics and EDs across provinces in Canada and over time (2011-2018). We compared observed search patterns with survey data capturing access to primary care.

All search frequencies are expressed as a ratio of each monthly provincial value to the Canadian average over the study period (2011-2018). Plots of these data are included in Multimedia Appendices 1 and 2. To identify provinces with higher and lower search volumes, and those that experienced more dramatic changes over the study period, we used linear regression to calculate the intercept (model-predicted relative search frequency in January 2011), slope, average over the study period, and relative increase from January 2011 to December 2018 (based on model-predicted values at these time points). We report the Pearson correlation coefficients (r) and associated 95% CIs for the relationship between each province’s relative search frequencies for walk-in clinics and EDs over the 8-year study period, and provincial survey results (expressed as a percentage within each province). We interpret correlation coefficients between 0.7 and 1.0 as indicating variables that are highly correlated, between 0.5 and 0.7 as moderately correlated, and between 0.3 and 0.5 as being weakly correlated.

Walk-in clinic relative search frequency was highest in western provinces and is increasing over time across Canada. Provinces that had high walk-in clinic relative search frequencies reported greater same- or next-day access to care and lower use of ED for conditions treatable at the patient’s regular place of medical care. The percentage of patients with a regular medical doctor is also lower in provinces with high relative search frequency for walk-in clinics.

Findings may suggest that challenges accessing primary care have contributed to increasing patient use of Web-based tools to search for more convenient and easily accessible care through walk-in clinics [34]. The degree to which the rapid increase in searches for walk-in clinics reflects increased availability of this option relative to other sources of primary care [3], a preference for the convenience of walk-in style practice [9], or another explanation cannot be determined from our data.

The strong positive correlation between walk-in clinic searches and same- or next-day access and strong negative correlations with ED use for treatable condition and having a regular medical provider lend credibility to the Google Trend results. Walk-in clinics may make it easier to get timely services without going to ED [35]. However, this could raise concerns about discontinuity and poor coordination of care [6] as more patients also lack regular medical providers. It may be that people lack regular medical providers because physicians are not taking more regular patients but are instead choosing walk-in style practice [19].

ED search frequency varies by province, but there is no clear pattern east to west nor consistent pattern of increase or decrease over the study period. In the province of Manitoba, ED closures were a prominent topic in the media over the study period, which may explain high relative search frequencies in this province and would in no way correspond to service use. We did not observe correlations between ED search frequency and survey data, including ED use for condition treatable by provider at a regular place of care. It is likely that Google Trends less directly reflects ED service use than walk-in clinic use. Most people already know the location of major hospitals and attached EDs, which are open 24/7. As such, people may not need to search through Google to find their location or operating hours. Walk-in clinics are more numerous, may open and close in different locations, and may have fluctuating hours of operation and so service use may be better reflected in search data for walk-in clinics.

This study is subject to a number of limitations. First, Google searches can only partially reflect patterns of care seeking. Not all searches correspond to actual use, and not all use would be preceded by a search. It is plausible that the use of Google as a tool in accessing health care also differs across patients and provinces, independent of realized service use. We have no ability to validate Google Trends data against administrative data sources collected over time. Comparison with survey data adds credibility to search frequency for walk-in clinics, but this analysis is limited by the sample size for cross-provincial comparisons. This study builds on an existing body of literature that has confirmed the plausibility of Google Trends search results as public health indicators by correlating Google Trends data with other sources [36].

The correlations made are at the ecological level and require caution in interpretation. People who answered the surveys are not the same as those doing the searches. Survey data include only those who choose to respond, and search data only reflects Canadians who use the internet [37]. We were unable to disaggregate search data by demographic characteristics or other characteristics related to both search behavior and primary care access. It may be that internet users more commonly use walk-in clinics, but this would be true across provinces and over time and would not explain the province-level relationships observed.

Findings from Google Trends are consistent with survey information about the province-level access to primary care, and may offer some insight into how the organization of primary care differs across provinces. Findings suggest that patient use of Web-based tools to search for more convenient or accessible care through walk-in clinics is increasing. Further research is needed to validate Google Trends data with administrative information on service use.