INTRODUCTION

Population aging is a global phenomenon and Spain is one of the countries where this process is especially notable¹. In the face of this major demographic change and given that older adults are the main consumers of healthcare resources^2),(3, examining health services use by this particular group in terms of frequency and factors associated with it acquires especial relevance.

Most studies on healthcare utilization determinants are based on the model developed by Andersen⁴ which explains healthcare use patterns by distinguishing among three types of factors: predisposing, enabling, and need factors. Predisposing factors include sociodemographic characteristics such as age, sex, educational level, and marital status. Enabling factors refer to those aspects that facilitate the use of services such as accessibility and characteristics of the healthcare system, individual´s income level, or having insurance. Finally, need factors are related to the health-disease process such as self-perceived health, pain, specific diseases, comorbidity, mental health, and dependency.

Need factors are the main determinants of healthcare utilization^{5),(6),(7),(8),(9}. Nevertheless, the role assigned to predisposing and enabling factors vary across studies^9),(10),(11.The availability of a population-based cohort of one of Madrid´s districts, the Peñagrande study´s database^12),(13, is a great opportunity to examine healthcare utilization at the local level.

The main aim of this study is to assess the level of healthcare use at both the primary care and the specialized care level in individuals 65 years old or older residing in an urban area and identify the determining factors associated with it.

MATERIALS AND METHODS

This is a cross-sectional study of a population-based cohort of individuals ≥65 years of age. The study population included the 9200 individuals of 65 years of age or older residing in two neighborhoods in the northern part of Madrid (Peñagrande and Cuatro Caminos), known as the "Peñagrande Cohort." This cohort was started in 2008¹² and updated in 2011¹³.The study population was identified through the registry of state-subsidized healthcare card in those two districts. The cohort was comprised of a representative sample stratified by sex and six age groups (i.e., 65-69, 70-74, 75-79, 80-84, 85-89 and ≥90 years of age). Similar numbers were chosen for each age group in order to guarantee enough representation among the older groups. Data were collected during 2011.To be eligible individuals had to be a registered resident of those two neighborhoods at the time of the study¹³. The final sample of the Peñagrande Cohort included 1760 individuals (755 from Peñagrande and 1005 from Cuatro Caminos neighborhood) (Figure 1).

Figure 1 Flow Chart: Population-Based Cross-sectional study 

RESULTS

Response rate was 75.4% (n=1,327). The average age of participants was 77.4 ± 7.4, with a range of 65-105 years with women representing 53.2% of the sample. Figure 1 shows the distribution by age and sex. The 433 individuals refusing to participate or who could not be reached were of similar age (77.5 ± 7.4) as the actual participants; however, the proportion of individuals in the extremes of the distribution (60-64 and≥90 years) was greater while the female representation was smaller (45.3%). Table 1 shows the characteristics of the study population using weighed data.

In Table 2 we describe the indicators of healthcare utilization of interest here such as number of visits, hospital admissions, or prescribed drugs used (represented by the median and interquartile range) as well as the distribution of the population using the different healthcare services more than once a month or year (as previously defined). Among individuals 65 and older, 64.8% (95%CI: 62.3-67.4) saw a general practitioner and 44.6% (95%CI: 41.2-47.2) saw a nurse at least once a month. The percentage of people having a home visit by a general practitioner was 3.1% and 3.5% had a nurse pay a home visit. Over half (57%, 95%CI: 54.3-59.7) of the participants were polymedicated, 25.1% (95%CI: 22.7-7.4) went to see a rheumatologist at least once a year, 12.9% (95%CI: 11.1-14.7) went to see a physical therapist, and 30.6 (95%CI: 28.1-33.1) saw the podiatrist. The proportion of older adults being admitted to the hospital in the previous year was 16.4% (95%CI: 14.4-18.4).

Table 2 Indicators of health services utilization 

Weighed data by sex and age using the 2011population of the Northern Madrid district as the standard. *% (95%Confidence Interval) of the population with that characteristic. ¹ In the last month. ² In the last year. ³ Number of prescription drugs taken regularly at the time of the interview.

Table 3 shows the results of the multivariate analysis for each indicator. Those variables with no association at the bivariate level are left blank (except for age and sex). Comorbidity is the variable showing the strongest association with utilization of healthcare services, varying between an OR of 4.10 (95%CI: 3.07-5.49) and 1.39 (95%CI: 0.97-1.99) for polypharmacy and seeing the physical therapist, respectively. Age and being dependent are also associated to most of our indicators of healthcare use. The rest of variables are independently associated to fewer utilization indicators.

Table 3 Multivariate analyses of the indicators of health services utilization 

* All data presented as Odds Ratios (95% Confidence Interval). ** Reference category in parenthesis. These variables were not associated with the indicator at the bivariate level and were excluded from the multivariate analyses.

We analized all seven diseases grouped under "comorbidity" separately to identify any independent associations with the different healthcare use indicators. Figure 2 graphs the relationship between each of the 7 diseases with healthcare utilization in fully adjusted multivariate models. Among these, it is worth pointing out the independent association found between diabetes and cardiovascular diseases with the following variables: hospital admissions (OR 1.97; 95%CI: 1.44-2.71 and OR 3.31; 95%CI: 2.33-4.72, respectively), visits to a general practitioner (OR 1.34; 95%CI: 1.03-1.76 and OR 1.46; 95%CI: 1.05-2.02, respectively), visits to a nurse (OR 1.43; 95%CI: 1.11-1.85 and OR 1.97; 95%CI: 1.44-2.71, respectively) and polypharmacy (OR 4.45; 95%CI: 3.28-6.04 and OR 3.31; 95%CI: 2.33-4.72, respectively).

Figure 2 Association between the 7 diseases used to define comorbidity and the different indicators of health services utilization. Multivariate analysis adjusted only by the variables showing a bivariate association with the indicator 

DISCUSSION

Our results confirm the frequent use of healthcare services by the older sector of the population. In agreement with the literature^2),(7)-(9, need factors best explained the high consumption of these resources with comorbidity showing the strongest association to the various utilization indicators. Dependency was associated with home care and hospital admissions. Diabetes and cardiovascular diseases, individually, were the diseases with the highest associations with healthcare use, which underscores the importance of promoting the Chronic Patient Care Program¹⁹ being implemented in the Madrid region.

Among predisposing factors, age was related to service utilization but not in a homogeneous fashion across indicators. Patients between the ages of 65 and 74 were the ones visiting specialized care most frequently, whereas visits to primary providers were related to individuals older than that, again supporting previous work^{3),(5),(10),(20}. However, our multivariate adjusted analyses failed to detect an overall association between female sex and a greater use of health resources as found by others^{5),(8),(10),(11}. Women did see the specialists depicted here (rheumatologist/orthopaedist, physiotherapists and podiatrists) more often than men but no differences were found regarding seeing a primary provider, hospital admissions, or polypharmacy. Gender differences may have been explained away by adjusting the models for anxiety/depression, a condition more prevalent among women^5),(6. Finally, the rest of predisposing factors did not show an independent association with the healthcare use indicators.

Regarding enabling factors, physiotherapy and podiatry services did vary between the two neighborhoods which could be partly explained by the differentials in supply (i.e., elderly care centers offering low-cost podiatry services in the neighborhood of Cuatro Caminos) or different ease-of-access (direct referrals to physiotherapy by the general practitioner bypassing the specialist in physical therapy, as is the policy of the Peñagrande Health Center but not in the Cuatro Caminos Center).

A closer look at the healthcare use indicators related to Primary Care show a high demand for visits to general practitioners and nurses. Almost 2/3 of those over 65 see a general practitioner at least once a month and over 40% see a nurse. Palacios-Cena et al.⁸⁾ reported an increase in visits to a primary provider between 2001 and 2009 (from 39.9% to 53.2% in women and from 32.3% to 47.7% in men) which is also seen in our data.

The high volume of visits to a nurse reflect the growing role of nursing in chronic patient care. Patients suffering from diabetes, cardiovascular diseases, or cognitive deterioration, usually the older patients, are the ones seen most frequently by nurses²¹. It is worth mentioning that our results seem to challenge the previous conclusion put forward by others that the increase in visits to nurses is related to a reduction of visits to doctors ^22),(23.

Doctor and nurse home visits are below 5% of all visits per month and associated with older age and dependency, similar to previous findings^20),(24.

In our analyses, arthritis, used here as a marker for Chronic Non-Neoplastic pain, was not independently associated to primary care visits but, rather, to specialized care ones (rheumatology/traumatology and physical therapy). The relationship between arthritis and comorbidity in the older population may explain the lack of an association when those variables are adjusted for.

In our work we found that over half of older adults take more than 5 prescriptions daily, with the corresponding risk for drug interactions and secondary effects. At these ages, polypharmacy is a well-known fact linked to comorbidity^{8)-(10),(25),(26}. Our study shows its relationship to ageing and the presence of cardiovascular diseases, diabetes, and anxiety/depression.

This study took place in Spain in 2011, when physician-prescribed drugs for those 65 and over were 100% state-subsidized and the Madrid autonomous region had not adopted yet the electronic prescription system. These two factors may have contributed to the high volume of general practitioner visits since the main reason for a number of these visits is to get prescriptions renewed¹⁰.

There is already evidence linking the introduction of the 2012 prescription drug copayment for adults over 65²⁷ with the reduction of prescription drug expenditures²⁸ and with the reduction in the number of medications taken by chronic patients²⁹. Future studies should examine whether this copayment and the widespread use of the electronic prescription are affecting prescription drugs use³⁰ as well as the frequency of visits to a primary provider³¹ as suggested by previous work.³¹

Because ours is a cross-sectional study, our findings do not infer causal relationships; however, the data come from a population-based study with a large sample of older adults who were asked about a wide array of healthcare services. Thus, we believe our study provides a comprehensive overview of medical resources utilized by older adults. Although our data on utilization are self-reported, other authors have confirmed the high reliability and validity of self-reports of healthcare use.^32),(33

In sum, our work shows the high level of utilization of healthcare services by the older population. These high healthcare costs set against the background of scarce public resources highlights the need to develop chronic patient care programs focused on the empowerment of this population and the promotion of self-care¹⁹. The need for administrative improvement does not hide the urgent need for more healthcare resources to meet the new care needs of older adults.