Principal findings
This secondary analysis of the TOPICS-MDS data provided insights into differences in the level of frailty and healthcare use of older persons with or without out-of-hours GPC contact. This study showed that persons of 65 years and older who had contact with the GPC and received either an emergency consult at the GPC or an urgent GP home visit were significantly more frail compared to older persons who had no GPC contact during out-of-office hours. Furthermore, we observed an increased level of frailty in all domains (e.g., (co)morbidity, activities of daily living, social contacts, psychological wellbeing, and quality of life). In addition, older persons who visited the GPC were significantly more often admitted to the hospital and more often used home care compared to older persons who had no contact with the GPC. In summary, this study shows that older people who had contact with the GPC had a relevant problem with frailty. Previous studies on frailty of older persons receiving emergency care have usually focused on the setting of the emergency department (29-32) or the general practitioners’ practice during daytime hours (33-36). Studies in the last group did not differentiate between regular, chronic and emergency primary healthcare. To our knowledge, this is the first study that identifies the frailty of older persons after contact with the emergency primary healthcare setting during out-of-hours.
The identified differences between older persons with GPC contact and the reference group, all of which supported the findings. We showed statistically significantly increased levels of frailty in older persons who contacted the GPC. However, it is unclear whether the identified differences were clinically relevant for older persons. As the differences in frailty levels of older persons with(out) contact with the GPC were shown to be (more than) 5 percent, we also considered the identified levels of frailty to be clinically relevant (37).
This secondary data analysis was based on TOPICS-MDS data retrieved from the baseline measurements in studies that were performed in the period between 2009 and 2014. Since that time, several policy measures have been undertaken to reform the Dutch healthcare system, such as a reduction or abolishment of retirement homes and the introduction of increased thresholds for admission of older persons to nursing homes. Additionally, the financial reimbursement systems for long term care facilities (e.g., for older persons) changed in the Netherlands. As a result, more older people now remain at home longer with an accumulation of comorbidities, psychosocial problems, and at the same time a (temporary) loss of reserve capacity in one or more domains of functioning. Accumulation of frailty in community-dwelling older persons leads to a crisis where older people seek emergency (primary) healthcare. Additionally, in the (inter)national literature, the problem of emergency department crowding related to ageing (5, 38) and an increased number of vulnerable older persons requiring care by general practitioners (7) has been described. The effects of the earlier described healthcare system reform measures were not represented in the TOPICS-MDS database in the period between 2009 and 2014. Therefore, our findings could underestimate the actual problem of frailty in community-dwelling older persons who currently seek emergency healthcare.
We know that access to and use of (primary) healthcare is highly influenced by SES, and this study also showed a positive association between a low SES and higher GPC use. Furthermore, Shebehe et al. 2018 showed that older persons above 65 years of age who visited primary health care centres in Sweden localised in neighbourhoods with a low SES had higher rates of hospital readmission compared to older persons with a higher SES (39). This result suggests that interventions aimed at empowering older persons, reducing frailty in primary care, and reducing hospital readmissions for older persons should also take socioeconomic disparities into consideration.
Older persons rarely use the word ‘frail’ to describe their situation, because they do not think about themselves in terms of frailty (40). Older persons who are classified as frail according to medical criteria do not always feel frail (41, 42). This difference in interpretation between doctors and patients is also known as the ‘disability paradox’ (43). The discrepancy between the clinical understanding of frailty and the way people perceive frailty has important implications for older people’s wellbeing (44). While frailty classifications can be useful in guiding clinical care, it is also important to consider how individuals and others perceive and respond to their experiences of frailty. Older persons described that being labelled as frail – particularly against one’s will – was seen as damaging to their health, because it may lead to behavioural confirmation of the label. When labelling occurs, unless older persons possess strategies to resist this trend, their health and wellbeing may be compromised. Frail older persons are mainly concerned with their quality of life, asking questions such as ‘what is important’, ‘what do I value’ and ‘what gives meaning to my life’. At the same time, older persons consider it relevant to talk to healthcare professionals about aspects of physical decline, psychosocial reserve capacities and (emergency) primary healthcare treatment in terms of opportunities and threats (risks) (40). As frailty is a process involving an accumulation of physical, psychological and/or social deficits in functioning with the risk of adverse health outcomes (admission to a hospital or sometimes death) (45), advanced care planning between older persons and their GP is important (46). Furthermore, circumstances and preferences in the health care of older persons can rapidly change because of the occurrence of (adverse) events, such as a crisis or a hospital admission (47). The preference of older persons to what extent and how they prefer to be involved in shared decision making in primary healthcare varies (48), whereas GP characteristics, communication skills, GP consultation duration, and continuity of (emergency) healthcare were described as important factors in the enhancement of shared decision making (49, 50). GPs themselves showed different perspectives on their role in the management of complex health problems of older persons in primary care, varying from ‘manoeuvring along competence limits’, ‘Herculean task’, and ‘cooperation and networking’ (51).
In our opinion, the primary care GP plays an important role in the early identification of frailty, advanced care planning and the management of complex interventions in older persons. Where more adequate support of frail older persons can be provided, possibly (unnecessary) out-of-hours GPC contacts or ED admissions can be avoided. Brouwers et al. 2017 showed in an explorative study amongst emergency healthcare providers that early identification of frailty, improving the continuity between primary and home healthcare and hospital based (emergency) care and vice versa is recommended (52). In recent years, several instruments for the screening of frailty in older persons have been developed. Many of these instruments focus on the identification of physical, functional or cognitive aspects of frailty (34, 53), such as the ISAR (Identification of Seniors at Risk) (32), the SHARE-FI (Survey of Health, Ageing and Retirement in Europe Frailty Instrument) (31), and the APOP screener (54), while other instruments are focused on a multi-dimensional screening of frailty in the elderly such as the Groningen Frailty Indicator (55), the Frail-VIG (53) and Easy-Care TOS (33, 56). Based on our study results, we recommend developing building blocks and tools for a multi-dimensional screening of frailty in emergency primary healthcare settings such as the GPC. Early detection of frailty in primary care can improve the use of effective interventions in primary care that provide adequate biopsychosocial support in older persons and thus prevent frailty progression (57). A recent systematic review showed that group sessions, individual educational sessions by a geriatrician and cognitive training for older persons in primary care had a positive effect on improvements in frailty, physical activity and other outcomes (57). In addition to early detection and effective (preventive) interventions, follow-up care for older persons after a GPC visit also seems important.
If accessibility of older persons to primary care (including out-of-hours primary care) is high, than the rate of ED visits is significantly lower (58). Furthermore, structured and structural information exchange between healthcare providers in the emergency healthcare pathway and a more generalist approach of older persons in emergency healthcare is recommended in order to deliver appropriate emergency healthcare for older persons who are frail.
Strengths and limitations
An important strength of the TOPICS-MDS study is the large sample and therefore the ecological validity of the study results. Furthermore, differences between older persons who have (or have not) had contact with the GPC were shown to be statistically and clinically significant. However, the study sample contains a combination of studies with different designs, inclusion and exclusion criteria, sample sizes, and data collection methods. Therefore, the representativeness of this cross-sectional study sample for the whole Dutch population of persons 65 years or older could be questioned. Nevertheless, our results and associations are not less relevant (59). Furthermore, we corrected for the potential confounding demographic variables (gender, age and socioeconomic status) and for cluster effects between studies in the multilevel statistical analysis.
A disadvantage of secondary data analysis is that data quality control is challenging. For instance, for the ‘healthcare use’ outcome measures, we have chosen to include only information on hospital admissions and the use of home care in the analysis, and we omitted temporary admission of older persons to a retirement or nursing home as an outcome measurement due to the insufficient quality of these data.
The data analysis was based on baseline measurements of older persons after they had GPC contact. Therefore, it is unclear whether frailty was present before the GPC contact. It is possible that frailty predicts a GPC contact of older persons, or perhaps a visit of an older person to the GPC results in increasing levels of frailty. Additionally, the cross-sectional study design could be considered a limitation.
Finally, the assumption that data from the primary studies included in this secondary data analysis are robust is an important point of departure in the multilevel analysis. However, during cross data quality checks, data of some variables appeared to be incorrect, such as the date and time of the measurement at T0. Where possible, we corrected these data based on the original study publications. Despite these efforts, we cannot completely rule out some misclassifications.