Patient- and Caregiver-Related Factors Associated with Caregiver Assessed Global Deterioration Scale Scoring in Demented Patients

Background Informant-based rating scales are widely used in dementia but patients' and caregivers' features influence the final scoring. We aimed to evaluate the role of patient- and caregiver-related factors in a caregiver rated Global Deterioration Scale (GDS) score in a sample of Greek patients with dementia. Methods We included 194 patients with dementia and 194 caregivers/family relatives; Mini-Mental State Examination (MMSE); Neuropsychiatric Inventory (NPI); Katz Index of Activities of Daily Living (K-IADL) were administered to (a) patients and Center for Epidemiologic Studies-Depression (CES-D) Scale; Zarit Burden Interview (ZBI) to (b) caregivers. Participants' demographics and patients' and caregivers' characteristics were entered into a 3-block regression analysis. Results The final model explained 55% of the total variance of the caregiver assessed GDS score. The following variables significantly contributed to the final model: MMSE (β=-0.524); K-IADL (β=-0.264); ZBI (β=0.145). Conclusion We herein confirm the contribution of patients' cognitive and functional status and caregivers' burden in caregiver rated GDS scoring irrespective of demographic-related characteristics.


Introduction
Global severity measures, including Global Deterioration Scale (GDS) [1], are commonly used in neurodegenerative research to categorize subjects in clinical trials and further evaluate the therapeutic efficacy of both pharmacological and nonpharmacological interventions in relation to the progression of dementia process [2]. It has been suggested that global severity scales are potentially "less influenced by factors such as education, occupation, practice effects, personal background, cultural factors and linguistic factors" that affect mental status and psychometric measures. However, these scales are based on "information about subjective complaints of memory deficit, objective observation of deficit on careful clinical interview, and on assessment of functional ability of the subject" [3] based in part on information from a knowledgeable informant. Herein, we investigated an informant-based version of the GDS, which does not use the direct patient interview procedures conventionally applied for this assessment, for example, in the worldwide approvals of currently used medications for the treatment of Alzheimer's disease (AD) (e.g., rivastigmine and memantine) and in worldwide clinical practice. In general, informant's based rating scales can be influenced by both patients' mental and neuropsychiatric status [4], as well as informant-/caregiver-related factors, including but not limited to depression and burden [5].
In the case of Greece, approximately 200.000 patients have been diagnosed with dementia of the AD type and 50.000 patients still remain undiagnosed [6] while the 2 Current Gerontology and Geriatrics Research majority of patients have a low educational level [7]. Apart from patients' neuropsychiatric status and caregivers' mood status and burden, it is still understudied whether patients' and informants' demographic factors including educational level influence GDS score and whether family relationships between patients' and their family caregivers are another influential factor. The latter could be of major importance considering family bonds and structure related to sociocultural values.
The aim of the present study was to evaluate patientand caregiver-related features in order to identify the predictors that explain the variability of an informant-based GDS score in a large sample of Greek patients with dementia.

Participants.
The study included 194 outpatients who had been referred to our memory clinic ("Nestor" Greek Psychogeriatric Association) during the previous 12 months, as well as their 194 responsible caregivers. Inclusion criteria were a diagnosis of AD, or vascular dementia [DSM-IV] [8], or Dementia with Lewy Bodies [9], or frontotemporal dementia [10]. Exclusion criteria were as follows: permanent or transient severely impaired consciousness or delirium during the previous month; newly started antipsychotic medication during the last week before test administration; absence of available data in psychometric scales ( Figure 1). All patients if possible and their caregivers were fully informed about the purpose of the study and provided written informed consent. The study was approved by the Local Ethical Committee of the Psychogeriatric Association and was conducted in accordance with the Declaration of Helsinki, as revised in 1983.

Testing Materials.
The Mini-Mental State Examination (MMSE) was administered to patients [11,12] and had a MMSE score ≤ 24. Dementia staging was defined based on the GDS [1] (translated and validated in Greek after permission by authors of the original scale) that assesses dementia severity over seven levels, i.e., GDS 1: absence of cognitive changes to GDS 7: very severe cognitive deficits. The score was derived from the caregiver interview [1]. Patients' neuropsychiatric symptoms and activities of daily living were evaluated using the Neuropsychiatric Inventory (NPI) [13,14] and the Katz Index of Independence in Activities of Daily Living (K-IADL) [15], respectively, completed by their caregivers. All caregivers were given the following self-administered scales: (a) the Center for Epidemiological Studies-Depression (CES-D) Scale [16,17] for the evaluation of depressive symptoms; (b) the Zarit Burden Inventory (ZBI) [18,19] for the assessment of the subjective burden experienced by dementia patients' caregivers. The following demographic characteristics for all patients were also included for the purposes of the present study: age, gender, years of education, and dementia type. Furthermore, the following characteristics were recorded with regard to caregiver/caregiving: age, gender, years of education, weekly caregiving time, caregivers being a spouse, and caregiver living with the patient.

Statistical Analysis.
Psychometric variables of interest were examined for normality (skewness, Q-Q plots, and Kolmogorov-Smirnov test) and assumptions for normality were not violated. A 3-block hierarchical regression analysis was conducted to identify the contribution of patients' and caregivers characteristics to GDS scoring. Multicollinearity was checked for regression analysis and assumptions (tolerance; VIF) were not violated. Patients' and caregivers' demographic features (age, gender,  years of education, and caregiver being a spouse) were entered in block 1, followed by patients' disease and clinical/psychometric characteristics (diagnosis of AD, MMSE, NPI, and K-IADL) in block 2, and caregiving characteristics (CES-D, ZBI, weekly hours of caregiving, presence of professional caregiver, and caregiver living with the patient) in block 3. All statistical analyses were conducted using IBM SPSS v. 22.0 and the level of significance was set at p<0.05.

Descriptive Characteristics of the Patients' and Caregivers'
Samples.  Figure 2 shows the distribution of GDS based on caregivers scoring.

Discussion
In this study, we investigated the contribution of demographic factors, patients' cognitive and behavioral characteristics, and caregivers' depression and burden status, as well as general caregiving features related to family relationships in the scoring of a global assessment scale such as the GDS. To the best of our knowledge, this is the first study that addresses the issue of caregiver scoring of the GDS as well 4 Current Gerontology and Geriatrics Research   as associated factors. We found that patients' MMSE and K-IADL and caregivers' ZBI scores were significant factors in the final model when demographics and patients' and caregivers' features were all included as prior variables which were accounted for. The GDS is one of the most commonly used global assessment scales for measuring patients' severity. It has been found to be advantageous to overall subject assessment in terms of utility in incipient and severe AD, reliability, sensitivity to AD course [2], and response to cognitive enhancers (e.g., rivastigmine [20]). Furthermore, as with other global assessment scales, it has been considered to be less influenced by demographic, cultural, and linguistic factors, patients' occupation, and personal background and practice effects [2]. These properties constitute a clear advantage when examining patients with low educational background, as it is the case in Greek patients with dementia [7]. In the present study, neither patients' nor caregivers' demographic characteristics (i.e., age, gender, and years of education) significantly explained the variance of GDS scoring, in accordance with previous studies using other informant-based rating scales for patients' functional abilities [21]. Others have found that caregivers' higher educational and sociocultural level is associated with more accurate reports of patients' functioning [22] or that gender may influence functional assessment of patients' basic and instrumental activities of daily living, behavioral and psychological symptoms, anosognosia, and quality of life [23].
Patients' clinical and psychometric data explained 53.5% of the variance on the GDS score in our study, with MMSE and K-IADL significantly contributing to the model. Higher MMSE and K-IADL scores were predictive of lower GDS scores. The significant contributions of MMSE and K-IADL are in accordance with other studies in patients with mild deficits; caregivers often overestimate functional abilities of patients whose cognitive impairment is mild [24]. Negative findings have also been detected [25].
Caregivers' information should always be complementary to patients' medical history and clinical/psychometric examination [26,27]. Thus, factors that could potentially influence caregivers' rating of patients' severity and functional status cannot be ignored in clinical practice or in research. In the present study, we found that caregivers'/caregiving features marginally increase the percent of GDS variance in the final regression analysis (Model 3) with the final model explaining 55% of the total variance. Caregiver burden [23,28,29], depression [23,30], and poor quality of life [31,32] have been often considered as confounders in the estimation of patients' clinical profile, including neuropsychiatric symptoms, cognitive severity, daily living functional activities, and quality of life. Even though we found a significant contribution of the caregivers' burden, we failed to identify a significant contribution with respect to caregiver depression in contrast with previous studies [28,33] but in accordance with others which did not report significant effects [21]. In contrast to other studies that report a relationship between the patient and the caregiver and the potential bias in rating of informant-based scales for patients [34,35], our findings did not support previous studies. None of the related factors (caregiver being as spouse, caregiver living with the patient) significantly contributed to regression models.
Limitations of the present study are the inclusion of a sample of patients/caregivers from a specialized outpatient memory clinic which might result in bias regarding higher awareness of cognitive and functional decline, as well as the inclusion of consecutive patients with the majority having a diagnosis of AD which did not permit the evaluation of the contribution of patients' and caregivers' characteristics to GDS scoring in other dementias with pronounced behavioral symptoms (e.g., frontotemporal dementia). We also acknowledge some methodological issues regarding the evaluation of caregivers' characteristics, including hours of caregiving. Hours of daily care may considerably vary within a week and this variability (which was out of the main aim of the present study) may affect caregiver burden and depression. Furthermore, we assume that several other factors that were not included in the current study might further contribute to the GDS score. For instance, we did not include psychometric questionnaires to assess caregivers' temperament and character features that might further contribute to their attitude towards patients' cognitive and functional status and their judgement regarding severity of patient's deterioration and the final GDS scoring. However, patients and caregivers characteristics as the ones examined in the current analysis are some of the most commonly recorded data in outpatient geriatric clinics. The fact that 55% of the GDS variance is explained by these factors and that caregivers' burden significantly contributes to the final model apart from patients' cognitive and functional status has significant clinical implications when considering GDS score based on caregivers' interview. Further endeavors to expand our understanding of the different factors contributing to informants' rating in different global assessment scales and how these factors influence or are associated with different outcomes at the time of diagnosis and after pharmacological and nonpharmacological interventions are warranted. Additionally, considering that in the past few years the enduring recession of Greece has been found to impinge on the mental health of the general population [36] and that one of the critical concerns of dementia caregivers is the financial strain they experience [37], future directions should include socioeconomic data regarding both patients' and their caregivers' .

Conclusions
In conclusion, we herein confirm the contribution of patients' cognitive and functional status and caregivers' burden to caregiver rated GDS scoring independent of demographicrelated characteristics. Considering that increased caregivers' psychological burden has been associated with a faster time to institutionalization and death in dementia patients [38,39], clinicians and researchers should consider caregivers' burden when interpreting dementia severity scales scored by caregivers.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors confirm that they do not have any conflicts of interest with regard to the present study.