Claims data analysis of medical specialist utilization among nursing home residents and community-dwelling older people

The number and proportion of older people in developed countries will increase rapidly in the coming decades [1]. This growing population group tends to have worse health and therefore has greater medical care requirements than younger population groups. Most older people have one or more chronic diseases, i.e. multimorbidity [2, 3]. Chronic diseases require regular medical care, not only by general practitioners (GPs) but also by medical specialists such as ophthalmologists and dentists.

Considerable numbers of older people become care-dependent at some point in later life. This means that they require a certain degree of support with activities of daily living. In Germany, the long-term care system provides support in cash and in kind for this population group. The actual degree of care dependency of a person is individually assessed according to § 18 of the German Social Code, Book XI, and reflected in the assigned “level of assessed need of long-term care”.

While especially older people in need of long-term care have a greater need for medical care, existing evidence is inconsistent on the utilization of medical care in this population group. Many older people in need of long-term care suffer from physical impairments such as vision and hearing loss, frailty and mental impairment, i.e. memory loss or dementia [4‐8]. Yet, while in some studies multimorbidity and the need for long-term care were associated with higher medical care utilization [9‐13], other studies showed that the picture is more complex: older people in need of long-term care, especially those living in nursing homes, tend to consult their GPs more often than older people who do not need long-term care, but they consult many specialists such as dentists, internists, orthopedists, and ophthalmologists less frequently [14‐16]. For other specialties (dermatology, otolaryngology (ENT medicine), gynecology/urology), no clear differences have been found so far [17, 18].

In Germany, these differences in medical specialist utilization may be explained by the particular conditions of medical care in nursing homes. Most nursing home residents are not capable of visiting a physician’s practice because of mobility or cognitive limitations. Yet, while GPs traditionally make home visits, most medical specialists do not [19]. Moreover, communication between nursing and medical actors may need to improve [20]. Recent reforms have tried to address this problem by increasing the remuneration of nursing home visits as well as by strengthening cooperation between nursing homes and physicians; however, an evaluation of these efforts is still outstanding [21]. Similar issues have also been found in other countries [22‐26], too. As a consequence, older people in need of long-term care may be at higher risk of inadequate medication, therapy and avoidable hospitalization [27‐34].

However, what is lacking in most of these studies is a thorough consideration of individual morbidity, as this information is often unavailable. Consequently, existing studies are unable to draw conclusions whether older people in need of long-term care under- or overutilize medical care. Moreover, existing studies differ considerably in their sampling strategies, i.e. whether cognitively impaired/institutionalized people and the oldest old are included or not [35, 36]. So far, it has not been investigated whether differences in specialist utilization persist when differences in morbidity and other explanatory factors such as age and the level of long-term care need are controlled. Our goal is to fill this gap and to investigate the medical specialist utilization of older people with and without a legally assessed need for long-term care while taking differences in morbidity into account.

We compared the medical specialist utilization of older people in need of long-term care (both nursing home residents and home care recipients) with older people who are not in need of long-term care. Our aim was to investigate whether differences in specialist utilization pertain when differences in morbidity between these groups are taken into account.

Our analyses revealed three main findings: First, when controlling for morbidity, older people in need of long-term care show a higher risk of not consulting a specialist than older people without the need for long-term care. This applies to nearly all of the medical specialties investigated. Psychiatry/neurology is the only specialty where older people in need of long-term care are less likely to have no specialist visit. The second finding was that the number of specialist visits does not differ between older people with and those without the need of long-term care. Our third finding was that the effects on specialist utilization vary by disease under study.

These findings provide a more specific picture of the medical care utilization of the heterogeneous group of older people than previous studies. While previous studies indicated that higher age and multimorbidity are associated with higher medical care utilization [10‐13, 36], our findings indicate that there are differences in specialist utilization between subgroups of the older people population that need closer inspection. Our findings indicate that older people in need of long-term care have a higher morbidity than older people not in need of long-term care. Inadequate or lacking control of morbidity may therefore lead to a misleading picture of medical specialist utilization among older people in need of long-term care. We controlled for these differences using 31 diagnosed disease groups. We find that older people in need of long-term care have a lower medical specialist utilization in most medical specialties. These findings support previous studies where morbidity was adequately controlled for [14, 16].

While we cannot conclude from our data whether these differences in specialist utilization represent unmet needs of older people, it seems plausible to assume that older people in need of long-term care may perceive access barriers. Many of them suffer from frailty and/or cognitive impairment [6], and are therefore limited in their mobility and more dependent on others in activities of daily living [38, 39]. This may also limit them in their ability to organize adequate health care and to visit a specialist practice, especially if the distance is perceived to be long [40]. Eventually this may lead to worse health outcomes [41, 42]. Furthermore, previous studies support the assumption that older people, especially those in need of long-term care, do not obtain adequate medical care [14, 32, 34, 43]. Another explanation might be that priority setting by nurses and relatives might prevent older people from consulting medical specialists. So far there is no existing research indicating these mechanisms, but a current study will investigate these mechanisms further [44].

Yet, although we found a differing visit probability the results indicate that the number of visits is similar for older people with and without the need for long-term care. This provides a more detailed picture than previous claims data analyses which investigated only the number of visits [17]. However, the data on the number of visits are to a certain extent “censored” because administrative claims data do not capture the exact number of visits. Instead, if a patient visits the same practitioner multiple times within one quarter of the year only the first visit is considered. Only if a patient visits different practitioners per quarter are multiple visits considered. Due to this data constraint, the factual number of visits per person as well as interpersonal differences in visit rates tend to be underestimated. Another limitation of our data is that sociodemographic and socioeconomic characteristics as well as morbidity of insurants vary significantly between statutory health insurances [45]. Consequently, despite the fact that the AOK covers a large part of the population in Germany, our sample may not be representative of the overall population. Moreover, it should be mentioned that the purpose of claims data is to justify revenues. Therefore, in terms of validity, the documented diagnoses may not always represent the actual prevalence of the respective health condition. In the same respect, long-term care need is defined on the basis of the legal assessment of care-dependency. Older people with actual care-dependency who have not been assessed in accordance with §18 of the Social Code, Book XI, are not correctly categorized into the group in need of long-term care.

Finally, the explanatory power of the models is relatively low. Pseudo R² measures range from 0.015 to 0.208 (see Additional file 3). Although the McFadden pseudo R² measure generally tends to be lower than the R² coefficient in linear regression analysis [46], crucial socioeconomic and sociodemographic covariates could not be included in the model. Moreover, the available data do not capture the complexity of the health care seeking process between patients, nursing staff, relatives, and physicians. The mechanisms of health care utilization have been shown to be multicausal, recursive, and asymmetric [47, 48].

Also, the samples were restricted to older people with a respective diagnosis; this already requires a former/previous practitioner visit. Consequently, ‘non-users’ were excluded from the samples because they do not have a chance of obtaining a diagnosis. We may also exclude older people with exclusion diagnoses because we only investigated the medical specialist utilization of older people with corresponding diagnoses.

Research on medical care in nursing homes and its health consequences is meanwhile addressed by several on-going studies [44, 49‐51]. Moreover, the German legislation on long-term care recently underwent extensive changes including strengthening the role of nursing homes in establishing cont(?)acts with physicians (§119b, German Social code Book V). This may influence the utilization of medical specialist care in nursing homes. While we cannot conclude from our findings or from the existing literature whether it is nursing home residents who obtain too little specialist care or whether it is older people not in need of long-term care who obtain too much specialist care, we can refer to national medical guidelines for specific medical conditions where a certain level of medical specialist care is advised and has been shown to reduce misdiagnosis [52, 53].

This study indicates that older people in need of long-term care may obtain inadequate specialist medical care. Future studies should assess the underlying mechanisms of the observed differences in specialist utilization to be able to draw conclusions about potential undersupply. Also, more research should be dedicated to the potential consequences of inadequate medical care on the quality of life of older people and on avoidable emergency/hospital admissions.