Abstract
Objective To develop a better understanding of the current health status and health care use of the population of very elderly Newfoundlanders to inform policy makers, decision makers, and health care providers about aspects of the health care system that might be in higher demand in the near future.
Design Descriptive analysis using data from the Newfoundland and Labrador component of the Canadian Primary Care Sentinel Surveillance Network database for the 2013 calendar year.
Setting Newfoundland.
Participants A total of 1204 Newfoundlanders aged 80 years and older.
Main outcome measures Encounters with family physicians, medications used, hospitalizations, emergency department (ED) visits, laboratory tests, and mortality were described and compared by sex, age, and location (rural vs urban).
Results Compared with men, women were prescribed more medications (P = .01), were less likely to be hospitalized (P = .007), were more likely to visit an ED (P = .049), and died less frequently (P = .001). Compared with those aged 90 and older, those aged 80 to 89 made more visits to their family doctors (P = .001) and were prescribed more medications (P = .001). Predictably, those aged 90 and older died more frequently than their younger counterparts did (P = .001). Compared with those in rural communities, urban dwellers were prescribed more medications (P = .031), were hospitalized more often (P = .001), were more likely to visit the ED (P = .002), were more likely to have laboratory tests ordered (P = .001), died more frequently (P = .023), and visited their family physicians more frequently (P = .001).
Conclusion Octogenarian women living in urban areas are the subcohort using the most resources. This might be owing to movement of the elderly to urban locations as they age.
As a result of a decreasing birth rate and the aging baby boomer cohort, Newfoundland and Labrador (NL) is projected to have the highest proportion of seniors in the country, with an estimated 31% of the population expected to be older than 65 years of age by 2036.1 A higher proportion of seniors implies a higher demand for health care services, as it has been shown that seniors, primarily the very elderly (aged 80 years and older), generate the highest health care costs.2 Additionally, it is the very elderly population that is the fastest growing cohort in NL.3 Therefore, an increase in health care demand in NL communities is inevitable, given the projected rise in the senior population over the next 20 years. For this reason, it is increasingly important to ensure that there are policies and programs in place that will be responsive to the needs of this population in the years to come. Currently, we do not have a clear picture of the health and health service needs of the very elderly population and, until recently, there has not been a reliable, easily accessible source of primary care data available to develop that picture.
The Canadian Primary Care Sentinel Surveillance Network (CPCSSN) is an organization of primary care practitioners, originally funded by the Public Health Agency of Canada, that extracts de-identified, point-of-care data from the electronic medical records of participating family physicians and nurse practitioners across Canada. The database includes information on provider-patient encounters, diagnoses, chronic health conditions, medications, laboratory tests, diagnostic imaging, vaccinations, and procedures. The NL component of the CPCSSN database is referred to as NL-CPCSSN. When linked to other databases such as the hospitalization database and the mortality database held by the Newfoundland and Labrador Centre for Health Information (NLCHI), it can provide a broad picture of an individual’s health and health care use.
This study allows a better understanding of the health and health care service needs of the very elderly living in Newfoundland. (All family doctors who contribute to the NL-CPCSSN database are located in Newfoundland; therefore, no patients in the study cohort were from Labrador.) We have compiled information from a variety of health data sources to develop a health profile of the very elderly (aged 80 years and older) using key health and health care use indicators.
METHODS
All very elderly patients, who we defined as 80 years of age and older, who were registered in the NL-CPCSSN database as of December 31, 2012, formed the cohort for this study. We collected de-identified data on this cohort between January 1 and December 31, 2013, from the NL-CPCSSN database, the clinical database management system (hospitalization data), the laboratory requisition database, the emergency department (ED) visit database, and the NLCHI mortality database. The data were provided in 9 separate files for ease of analysis: 5 of these data tables were from the NL-CPCSSN database, and 1 table from each of the 4 other databases was available through NLCHI. We used SPSS, version 23.0, statistical software to perform quantitative descriptive and comparative analysis on the data in each table independently. Demographic comparisons within the cohort were made based on sex, age group (ages 80 to 89 years and 90 years and older), and geography (rural versus urban, based on the second character in the postal code). An α of .05 was used to assess the level of significance. We used
2 analysis to analyze categorical data and Student t tests to analyze continuous data. The study was approved by the NL Health Research Ethics Board.
Data files
There were a total of 9 data files, each of which contained a CPCSSN patient identification code, patient demographic characteristics (age, sex, and first 3 characters of the postal code), dates, and data related to the specific data file for the 2013 calendar year. Patient postal code was used to determine rurality based on the second digit of the code, where 0 indicates a rural location and all other numbers indicate urban locations.4 A full list of the variables included in each data file can be found in Table 1.
List of variables included in each data file
RESULTS
The NL-CPCSSN data set contained 1204 individuals aged 80 years and older at the end of December 2012. This represents 2.6% of the 46 194 individuals in the NL-CPCSSN database on that date. In 2014, NL had a population of 527 756, with 22 565 (4.3%) aged 80 years and older.5
As indicated in Table 2, the cohort contained 759 (63.0%) women and 445 (37.0%) men; 1030 (85.5%) were between age 80 and 89, while 174 (14.5%) were aged 90 years and older. We had postal code information for 1046 individuals; 824 (78.8%) of these lived in urban areas and 222 (21.2%) lived in rural areas.
Cohort description and comparison by demographic features
Number of encounters, medications, hospitalizations, ED visits, and laboratory tests, as well as mortality, are compared by sex, age, and location (urban vs rural) in Table 2. Compared with men, women were prescribed slightly more medications (P = .01), were less likely to be hospitalized (P = .007), were slightly more likely to have visited an ED (41.1% vs 35.0% had at least 1 visit in 2013; P = .049), and died less frequently in 2013 (5.3% vs 10.6%; P = .001).
Younger individuals (aged 80 to 89) made more visits to their family doctors compared with those aged 90 and older (P = .001) and were prescribed more medications than the older group were (4.5 vs 3.4, P = .001). Predictably, those aged 90 and older died more frequently in 2013 compared with their younger counterparts (14.9% vs 5.9%, P = .001). There was no difference between these age groups in the number of hospitalizations, ED visits, or laboratory tests.
Those living in urban areas visited their family physicians more frequently than those in rural areas did (mean of 6.1 visits per year in urban areas compared with 4.0 visits per year in rural areas, P = .001). Urban dwellers were also prescribed more medications (75.6% of urban patients were prescribed at least 1 medication compared with 68.5% of rural patients; P = .031). Urban dwellers were also more likely to be hospitalized (P = .001), visit the ED (P = .002), undergo laboratory tests (P = .001), and die (P = .023) than their rural counterparts were.
Table 3 presents the reasons for office visits based on the ICD-9 codes recorded by the physicians. Four broader categories accounted for slightly more than 50% of the reasons for visits; these included codes for cardiovascular disease (CVD); codes indicating symptoms and signs rather than specific diseases; V codes (codes related to vaccinations and prevention activities) and uncoded reasons; and codes for endocrine and metabolic conditions. Those visits classified as for CVD were primarily for hypertension, the symptoms and sign category mainly comprised concerns of general unwellness, the V codes were primarily for vaccinations (overwhelmingly the flu shot), and uncoded visits included reasons such as acupuncture, seen on call, blood pressure (BP) check, no show, driver’s medical, fall, and deceased (eg, a housecall or nusing home visit where a patient has died). The endocrine and metabolic conditions cited as reasons for visits were primarily diabetes and hypothyroidism. In terms of the specific diagnosis recorded, Table 3 shows that hypertension, diabetes, general symptoms (syncope, dizziness or vertigo, malaise, fatigue, etc), osteoarthritis, congestive heart failure, dementia, and depression make up more than one-quarter of all ICD-9 codes used during encounters with the very elderly.
Reasons for encounters: A) Top 10 categories and B) top 10 diagnoses.
Table 4 presents the “problem list” from patient charts. There were some interesting differences between what conditions were listed in the charts as ongoing problems and the codes given as reasons for the encounters. While hypertension was at the top of both lists, hyperlipidemia, ischemic heart disease, cardiac dysrhythmias, and chronic kidney disease appear on the top 10 ongoing problem list but do not make it onto the top 10 list of reasons for encounters.
The “problem list” from patient charts: A) Top 10 categories and B) top 10 diagnoses.
Table 5 lists the top 30 medications prescribed; they make up 45.9% of all medications prescribed in the very elderly in this study. Three of the drugs, prescribed to 237 (19.7%) of the 1204 people in the cohort, were statin drugs. We recognize that the exact medications used might reflect such things as physician preference, the influence of pharmaceutical companies, hospital and pharmacy inventories, and the proportion of patients with drug insurance.
Top 30 medications prescribed to the 1204 people in the cohort in 2013
Other data available to us from the data tables show that each patient’s BP was measured an average of 4 times during the 2013 calendar year. The mean (SD) systolic BP was 128 (17) mm Hg and the mean (SD) diastolic BP was 74 (10) mm Hg. The mean (SD) body mass index was 26 (6) kg/m2, with the lowest being 19 kg/m2 and the highest 41 kg/m2.
Of the 87 people who died in 2013, 12 died from cancer-related causes, 21 from CVD-related causes, 20 from respiratory-related causes, and 34 from other causes. The proportion of individuals in rural areas who died in 2013 was lower than the proportion in urban areas (Table 2).
DISCUSSION
We now know that this population aged 80 and older are mostly women (63.0%) in their 80s (85.5%) living in urban areas (78.8%). The women in the study cohort, while more likely to visit EDs, are less likely to be hospitalized or die than men are. The nonagenarians (90 years and older) did not use more resources than their younger octogenarian counterparts did. The older group saw their family physicians less often and were prescribed fewer medications (perhaps an indication of decreased mobility in the older group rather than decreased illness), but were no different from the octogenarians in terms of hospitalizations, ED visits, or number of laboratory tests. The nonagenarians were more likely to die, which was expected.
Of particular importance to policy makers and health care providers is the difference between urban and rural dwellers. The urban-dwelling very elderly used more resources on every parameter. They visited their family physicians more often, were prescribed more medications, visited the ED more often, were hospitalized more often, and were more likely to have laboratory investigations. There is a potential explanation for this. While, in general, the rural population in NL is moving toward urban centres, this move might not be random in the very elderly. Perhaps when the elderly become very old and more frail, they are more likely to move to the urban centres where their adult children are living and where there is better access to a higher level of health care. In contrast, those elderly who continue to be in good health and are robust remain in their rural homes. This would lead to a concentration of the more frail and resource-using elderly in urban centres. It is also possible that this trend is not related to a population shift of sicker elderly to the city, but rather, or in addition to, the greater availability and proximity of practitioners and health resources in the city, which generate greater numbers of diagnoses, prescriptions, and services. Whatever the reason, physicians and policy makers in urban centres need to prepare for a disproportionate increase in the number of heavy–resource-using very elderly over the next 2 decades as the baby boomer population becomes octogenarians and nonagenarians. This increase in the proportion of very elderly Newfoundlanders will likely happen in the rural areas as well, but our data suggest that these individuals are likely to use fewer health care resources.
It is important to note that the NL-CPCSSN database might miss many people in nursing homes who do not visit family physicians’ offices but are cared for in the homes. Our data should therefore be interpreted as representing people aged 80 and older who visit family physicians in their offices. With this in mind, regardless of age, sex, or location, in a 1-year period 99.3% of this cohort saw their family doctors at least once, 70.6% received a prescription, 28.7% were hospitalized, 37.4% visited the ED, and 72.0% had laboratory investigations. There is likely no other age cohort in the NL population with this level of health care resource use and, interestingly, it is those in their 80s who have higher usage, not those in their 90s.
The descriptive data around the illnesses and reasons for encounters in this population are not unexpected, although they could be useful in planning what health conditions we need to be prepared to manage as our population ages. The information on medications being prescribed is, perhaps, more useful. It is interesting to note that 3 of the drugs, which were prescribed to 19.7% of the study population, are statin drugs. The value of these drugs in the very elderly has been questioned. A continuing professional development program on the appropriate use of certain drugs like statins and on deprescribing the Beers list6 of medications, which are potentially harmful for the elderly, should be considered.
Limitations
The source of the data used to conduct this study is perhaps the main limitation. The data in the CPCSSN database was pulled from the practices of 50 family physicians on the island of Newfoundland. That constitutes only 12% of the family physicians practising on the island. A disproportionate number of these physicians work in an academic environment (about 35% of them) and they are all physicians who were early adopters of electronic medical records in their practices. They and their patients might not be fully representative of the provider and patient population of the island. Another limitation is the lack of contributing physicians, and hence patients, from Labrador.
Conclusion
This description of the current health and health service use of very elderly Newfoundlanders has detailed how this population currently uses, and will likely continue to use, the health care system. This information will be useful for policy makers who are responsible for implementing programs and policies to meet the needs of this growing segment of our population.
Acknowledgments
This study was funded by the Newfoundland and Labrador Healthy Aging Program.
Notes
Editor’s key points
▸ Most very elderly Newfoundlanders are women in their 80s who live in urban areas; they are more likely to visit emergency departments but less likely to be hospitalized or die than their male counterparts are.
▸ Overall, octogenarians (those aged 80 to 89 years) saw their physicians more often and were prescribed more medications than their nonagenarians (aged 90 years and older) counterparts were.
▸ Urban-dwelling very elderly Newfoundlanders used more resources on every parameter than their rural counterparts did.
Points de repère du rédacteur
▸ La plupart des Terre-Neuviens très âgés sont des femmes âgées de 80 à 89 ans qui vivent en milieu urbain; par rapport aux hommes du même âge, elles sont plus susceptibles d’utiliser les services d’urgence, mais ont moins de chances d’être hospitalisées ou de mourir.
▸ Dans l’ensemble, les octogénaires ont consulté leurs médecins plus souvent et se sont vu prescrire plus de médicaments que les nonagénaires (âgés de 90 ans et plus).
▸ Dans l’ensemble, les Terre-Neuviens très âgés qui vient en milieu urbain utilisent plus de services de santé que ceux du même âge qui vivent en milieu rural.
Footnotes
Contributors
Ms Hurd wrote the initial draft of the article and contributed substantially to acquisition and analysis of the data. Ms Pike contributed substantially to the study conception and design, and to interpretation of the data. Drs Knight, Tarrant, Aubrey-Bassler, and Asghari contributed to the study conception and design. Dr Godwin contributed substantially to conception and design of the study, acquisition of data, analysis and interpretation of the data, and the initial draft of the article. All authors revised the article for intellectual content, gave final approval of the version to be published, and agreed to act as guarantors.
Competing interests
None declared
This article has been peer reviewed.
Cet article a fait l’objet d’une révision par des pairs.
- Copyright© the College of Family Physicians of Canada
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.