**Spanish Validation of the "User Reported Measure of Care Coordination" Questionnaire for Older People with Complex, Chronic Conditions**

**Ester Risco <sup>1</sup> , Glòria Sauch 2,\*, Anna Albero <sup>3</sup> , Nihan Acar-Denizli <sup>4</sup> , Adelaida Zabalegui <sup>5</sup> , Belchin Kostov <sup>6</sup> , Paloma Amil <sup>7</sup> , Albert Alonso <sup>8</sup> , Ana Rios <sup>9</sup> , Jaume Martín <sup>10</sup> and Núria Fabrellas <sup>11</sup>**


Received: 22 June 2020; Accepted: 26 August 2020; Published: 11 September 2020

**Abstract:** Introduction: Older people with complex, chronic conditions often receive insufficient or inefficient care provision, and few instruments are able to measure their perception of care provision. The "User Reported Measure of Care Coordination" instrument has been satisfactorily used to evaluate chronic care provision and integration. The aim of this study is to validate this instrument in Spanish. Methods: The questionnaire was adapted and validated in two phases: translation and cultural adaptation of the questionnaire and psychometric property measurement. Study population were chronic care conditions patients. Results: A total of 332 participants completed test re-test as part of the questionnaire validation process. The final version of the questionnaire had 6 domains: Health and Well-being (D1), Health day to day (D2), Social Services (D3), Planned Care (D4), Urgent Care (D5), and Hospital Care (D6). Cronbach's alpha for the overall questionnaire was 0.86, indicating good internal consistency. When analyzing each domain, only Planned Care (D4) and Urgent Care (D5) had Cronbach's Alphas slightly lower than 0.7, although this could be related to the low number of items in each domain. A good temporal stability was observed for the distinct subscales and items, with intraclass correlation coefficients varying from 0.412 to 0.929 (*p* < 0.05). Conclusion: The adapted version of the "User Reported Measure of Care Coordination" into Spanish proved to be a practical tool for use in our daily practice and an efficient instrument for assessment of care coordination in chronic, complex conditions in older people across services and levels of care.

**Keywords:** integrated care; social care; health care; older people; comorbidity; person centered care

### **1. Introduction**

Healthcare provision to aging populations in the current economic slowdown represents a considerable challenge for major European countries, especially when readdressing policies to cover new demands related to health and social needs [1]. People with complex, chronic conditions often receive inadequate care due to poor coordination between health and social care providers [2]. This situation is even worse when dealing with people older than 65 years. Inefficient coordination is a severe limitation when trying to achieve the triple aim of optimizing health system performance: improvement of health outcomes, the patient's experience, and care cost-effectiveness [3,4].

Many integrated care initiatives responding to multiple health and social care needs have been introduced in European health systems in a variety of contexts [5,6]. Some countries, such as The Netherlands, Germany, or Austria have already worked to improve care coordination since the early 20000 s [2]. The United Kingdom, with extensive experience working towards integration of care, has also developed and evaluated different potential plans to improve chronic care [7,8]. Spain has done so, with numerous actions undertaken and a variety of state-wide and regional strategies targeting different chronic conditions [9,10]. However, the impact of these strategies has hardly been evaluated, with some exceptions, such as the strategy deployed in the Basque Country [11] or the regional implementation carried out in Spain, Norway, and Greece in the context of the NEXES project (Integrated Care Services supported by Information Technologies for chronic patients with obstructive pulmonary disease, cardiac failure, and/or type II diabetes mellitus [12]).

Despite the implementation of healthcare policies and the great commitment of professionals, there are still significant weaknesses in our healthcare system as it stands: organizational and economic fragmentation between different levels of care, lack of integration between social services and healthcare, lack of coordination with family associations, and long waiting lists for diagnostic and surgical procedures.

In the region of Catalonia, Spain, the Government is already working on a new personalized model based on Person Centered Care called Interdepartmental Plan for Social and Healthcare Interaction (PIAISS) 2017–2020 [13]. This new framework contributes to the development of new strategies to improve care integration. Instruments such as the one developed by Nuño-Solinís et al. [9] have shown very good ease-of-use in daily practice, but its application is limited to the organizational level. We need new instruments to measure care coordination across all levels of care, principally when registering the perspectives of the populations [14]. Through this approach, we aim to measure the impact on the health of the population. Moreover, understanding patients' experiences is a good way to determine whether our strategies meet the expectations of populations in care coordination for chronic, complex conditions [15]. Therefore, there is *a* need for new validated instruments for use with Spanish-speaking population. In order to guarantee the quality of their measurements, it is essential that the instruments should be subjected to a process of validation. This process consists in adapting the instrument culturally to the setting where its psychometric characteristics are to be administered.

The "User Reported Measure of Care Coordination" questionnaire [16] is a tool developed and validated in the United Kingdom to measure the experiences of patients accessing care across organizational and professional boundaries. This tool has the potential to identify weaknesses with the goal of improving care coordination in line with the expectations of the population. The survey provides some information about how well services support patients and help users to achieve their own life goals. This instrument allows evaluation of the provision of care as well the effectiveness of future interventions for care coordination. The questionnaire consists of 46 items, divided into 7 domains: Health and well-being (D1), Health day to day (D2), Social Services (D3), Planned Care (D4), Urgent access to healthcare (D5), Hospital care (D6), and Self-information (D7). The results are expressed in a single overall numeric variable

(0–100) that is expressed as a continuous variable, in which the patients with higher scores have a more positive perception.

The aim of this study is the translation and cross-cultural adaptation of the questionnaire "User Reported Measure of Care Coordination" and the analysis of its psychometric properties.

### **2. Methods**

### *2.1. Design*

Adaptation and validation of the questionnaire were conducted through a descriptive cross-sectional study. This was carried out in two phases: an initial phase in which the translation and cultural adaptation of the questionnaire was performed and a second phase in which the behavior of its psychometric properties was measured. All research procedures used in this study were established in accordance with the Declaration of Helsinki; participants gave their signed informed consent, and the study was approved by the Ethics Committee at Hospital Clinic, Barcelona (HCB/2017/0731).

### *2.2. Translation and Cultural Adaptation of the Questionnaire*

Authorization to translate the questionnaire was obtained from the copyright holders. The transcultural adaptation of the questionnaire was performed following the structured procedure described by Beaton et al. [17]. Two translations of the questionnaire from English to Spanish were done by 2 healthcare professionals working separately. A single, consolidated version of both questionnaires was then produced. This version was then back-translated into English by 2 qualified translators. Finally, both versions in English (the original version and the back-translation) were compared and found to contain no significant differences.

Content validity was performed on the pre-final version of the questionnaire by a panel of 8 expert judges, including 2 nurses and 1 physician from a primary care center, 2 social workers from social services and 1 nurse, 1 physician and 1 social worker from a tertiary hospital. Criteria for the selection of the experts were: (a) experience in evidence-based judgement and evidence-based decision-making (research, publications and experience), (b) availability and motivation to participate, and (c) impartiality. Members of the panel of experts were asked to conduct a qualitative evaluation of every item (degree of understanding, agreement with the text) and were also asked to conduct a quantitative assessment of each item following these criteria: (1) competence (items belong to theoretical established factors), (2) clarity (item is easily understood, its semantics and syntactics are suitable), (3) coherence (item is related to the factor being measured), and (4) relevance (item is essential and has to be included). Each point assessed for each item was quantified on a Likert- type scale of 3 points: (1: Agree; 2: Neither agree nor disagree; 3: Disagree). This was done in two rounds by email. In the first round, all suggestions were added to the pre-final version and the second was performed to obtain agreement among all the experts. None of the items was eliminated at this stage. Some items were adapted to our specific context, and some response categories were combined from the experts' group discussion. The self-information domain (D7) was removed from the validation process, considering that this was part of the sociodemographic data, and it could not be measured. Finally, the remaining 38 items were divided into 6 domains: Health and Well-being (D1) consisting of 6 items, Health Day to Day (D2) consisting of 14 items, Social Services (D3) consisting of 3 items, Planned Care (D4) consisting of 5 items, Urgent Care (D5) consisting of 3 items, and Hospital Care (D6) consisting of 7 items. Scored items were created by converting the individual responses to questions into scores on a scale of 0 to 100, where a score of 100 represents the best possible response, and a score of 0 represents the worst possible response. Where response options were provided that do not have any bearing on performance in terms of patient experience (such as the 'Don't know/Can't remember' option), the responses were classified as "not applicable" and no score was given. Questions without answers were not considered too. In order to compute the score for domains, a numeric value (0–100) was also established for each domain. According to this method, at least 2 questions should be answered in the Social Services (D3), Planned Care (D4), and Urgent Care (D5) domains; at least 4 questions in the Health

and Well-being (D1) and Hospital Care (D6) domains; and at least 10 questions in the *Health day to day* (D2) domain. Again, higher scores mean better perception.

In order to check viability in daily practice, we conducted a pilot test administering the questionnaire to a sample of 18 patients. Patients answered the overall questionnaire and were asked to review the items, identifying words or concepts they did not understand. No relevant modifications were made.

### *2.3. Psychometric Properties*

The characteristics of the participants were similar to those from the sample used by Crump et al. (2017) in the development and validation of the original questionnaire. To ensure a representative sample, registered nurses from across Catalonia, divided into 9 health management areas, were involved in data collection. Forty experienced registered nurses working in different primary care centers received specific training and guidelines on eligibility criteria, clinical tests, and instructions for the administration of the questionnaire. The participation of nurses was voluntary, and each nurse had to include a minimum of 5 participants. The registered nurses recruited participants during a routine visit at the health care center or at home. The participants were selected for convenience in accordance with the following inclusion and exclusion criteria. The participants should be older than 65 years old. Multimorbidity or unique condition difficult to manage according to Complex Chronic Patient (PCC) criteria or advanced chronic illness (limited prognosis, palliative orientation, advanced decisions planned) was required. Understanding oral and written Spanish was necessary. People living in long-term care institutions were excluded. All study participants signed informed consent, and the implications of participating in the study were explained through an information sheet.

A total of 40 experienced registered nurses were responsible for collecting data across Catalonia. Although all of them received specific training, some variations could appear on collecting data.

Sample size was calculated following the criteria established by Kline et al. 2010 [18], which recommend a ratio of 5–10 subjects per item.

Data analysis was performed using the R statistical software package V.3.5.0 for Windows. Descriptive statistics of the quantitative variables are presented as mean and standard deviation. Categorical variables were described according to their frequency. The descriptive analysis on the scores obtained in each domain is presented with the mean and standard deviation.

The intraclass correlation coefficient (ICC) was used to study test–retest reliability of each item between two similar assessments with an interval of 7–10 days between measurements.

The internal consistency of the questionnaire was evaluated using Cronbach's alpha, which was considered acceptable if the value was greater than 0.7. The "alpha if item deleted" index was computed to study the increase or decrease in the sample value of alpha according to the deletion of items. Correlations between the different items and domains were studied using the Spearman correlation coefficient, assuming that the various domains gather distinct information with moderate-low correlation coefficients. Correlation was considered weak when the *r* value was less than 0.3, moderate between 0.3 and 0.6, and strong when greater than 0.6. A *p* value ≤ 0.05 was considered significant.

### **3. Results**

### *3.1. Patient Demographics*

A total of 332 patients across Catalonia completed the study. The characteristics of these participants are shown in Table 1. The mean age of the study population was 82.1 ± 8.0 years, and 204 (56.4%) participants were women. From the overall sample, 269 (74.3%) participants lived with a family member and 13 (3.6%) with some other carer such as a friend or neighbor. The rest of the participants lived alone. Only 6 participants (1.7%) described their health status as excellent, 75 (20.7%) said it was good, 198 (54.7%) participants defined it as regular, and 83 (22.9%) described it as bad. From the overall sample, 247 (68.2%) participants did not have any help or additional assistance from a private service or institution

for their daily care. The most prevalent health problems in our study sample were hypertension (79%), cardiac diseases (51.4%), diabetes (45.4%), and arthrosis or arthritis (15.7%).


**Table 1.** Sample characteristics (n = 362).

The data is shown as a number and a percentage (%) unless otherwise indicated.

### *3.2. Descriptive Analysis*

All questions were answered by the patients that participated in the study except those Conditional questions that could be skipped depending on previous answers. Table 2 shows the frequency distribution of the different questionnaire items. If we look at the non-response pattern for the items answered in the questionnaire, *Health and Well-being* (D1), *Health day to day* (D2), and *Urgent Care* (D5) had practically 99.7% of the items answered. Only in *Health Day to Day* (D2) did we find an item that had a lower score (99.7%), and item (Q8) that was 90.9%. In *Social Services* (D3), *Planned Care* (D4), and *Hospital Care* (D6) there was a lower response rate as participants skipped some questions because they were not receiving this specific care. Therefore, the answer on the perception of Social Services, Planned Care, and Hospital Care cannot be described from the population point of view.


**Table 2.** Frequencies of the categories for the items.

*Int. J. Environ. Res. Public Health* **2020**, *17*, 6608


*Int. J. Environ. Res. Public Health* **2020**, *17*, 6608

Some 15% of respondents took question P7 to be a multiple-choice question. 1 the person who lives with me helps me as much as individuals who don't live with me; (4) I don't need help; (5) I need help, but I don't receive help from anyone. 2 important tasks for myself; (2) I have some difficulty in carrying out some important tasks; (3) I'm not able to carry out tasks which are important for me; (4) I don't know/I'm not sure. 3 no contact with anyone. I feel completely alone. 4 sure/I disagree. 6 (1) Yes, the Primary Care Physician; (2) Yes, the Primary Care Nurse; (3) Yes, the Primary Care social worker; (4) If another healthcare professional, who?; (5) No, I do it myself; (6) No, a family member or care giver does it; (7) I don't know/I'm not sure. 7 (1) Yes, completely; (2) Yes, to some degree; (3) No; (4) I don't know/I'm not sure. 8 (1) Yes; (2) No; (3) I don't know/I'm not sure. 9 (1) Never work together; (2) They sometimes work together; (3) They always work together; (4) I don't know/I'm not sure. 10 (1) None of my needs have been assessed; (2) Some of my needs have been assessed; (3) All of my needs have been assessed; (4) I don't know/I'm not sure. 11 (1) Yes, they are totally involved; (2) Yes, they are quite involved; (3) They are not sufficiently involved; (4) No, not at all involved; (5) There is no family member or care worker involved; (6) I don't want my family or a care worker to be involved. 12 (1) Yes, they receive the help they need; (2) They get help, but not as much as they need; (3) No, they get little or no help; (4) They don't need/want help; (5) There's no family member or care worker involved. 13 (1) Yes; (2) No. 14 (1) Yes, the right frequency; (2) No, I need fewer visits; (3) No, I need more visits; (4) Not applicable. 15 (1) Yes, they last the right amount of time; (2) No, I need a little more time; (3) No, I need a lot more time; (4) No, I have longer than I need; (5) Not applicable. 16 (1) Yes; (2) No; (3) I don't know/I can't remember. 17 (1) A family member; (2) A friend or neighbor; (3) Yes, the Primary Care Nurse; (3) Yes, the Primary Care social worker; (4) Yes, another healthcare professional; (5) No, the Primary Care worker; (6) The emergency services (061); (7) Another healthcare professional (not on this list) who?; (8) Another person (not on this list); (9) I don't contact anyone. 18 (1) Yes, totally; (2) Yes, partially; (3) Disagree; (4) Totally disagree; (5) I don't know. 19 (1) Not sufficiently; (2) Sufficiently; (3) Too much; (4) They didn't provide me with any information. 20 (1) Yes; (2) No; (3) I don't know/I'm not sure.

### *3.3. Psychometric Properties*

When calculating the scores, we left out 6 items (Q1, Q7, Q21, Q24, Q31, and Q32) because they were considered additional information questions only and could not be measured. Once these items were excluded from the analysis, the Cronbach's alpha for the overall questionnaire remained at 0.86, indicating good internal consistency. When analyzing each domain, only Planned Care (D4) and Urgent Care (D5) had Cronbach's alphas slightly lower than 0.7, although this could be related to the low number of items in each domain. Furthermore, for Planned Care (D4), the Cronbach's alpha value suggested eliminating item Q28, moving internal consistency from 0.35 to 0.58. The internal consistency of each domain and correlations across domains are presented in Table 3. The scores calculated in each domain indicated moderate correlations between Health and Well-being (D1) and Health Day to Day (D2) with the subtraction of the domains. The remaining 4 domains were clearly divided into two blocks: Social Services (D3) with Planned Care (D4), and Urgent Care (D5) with Hospital Care (D6). Social Services (D3) and Planned Care (D4) were poorly correlated with Urgent Care (D5) and with Hospital Care (D6). Table 4 represents the intraclass correlation coefficient (ICC) values for test–retest results. A good temporal stability was observed for the distinct subscales and items, with intraclass correlation coefficients varying from 0.412 to 0.929 (*p* < 0.05).

In total, 9 items were excluded when calculating the scores for different dimensions. Items Q1, Q7, Q21, Q24, Q31, and Q32 were excluded as they were not considered appropriate for a score. The aforementioned questions did not evaluate performance but rather the context or background information. Item Q38 was also excluded due to a low response rate (21.8%) as well as item Q8 since it directly depended on item 7, which was also was excluded due to a low response rate. Finally, "Cronbach's Alpha if item deleted" suggested removing item Q28, raising the score for internal consistency from 0.35 to 0.58. A total of 24 items achieved one or more positive correlations greater than 0.4 with other items in the survey.


**Table 3.** Intercorrelation and internal consistency.

\* Spearman correlation coefficients statistically significant at the 0.05 level. SD: Standard Deviation. Mean (SD)—TOTAL: 71.1 (16.3) Cronbach's Alfa (TOTAL) = 0.86.


**Table 4.** Test–retest reliability.

ICC2,1: Intraclass Correlation Coefficient. \* A value of *p* < 0.05 indicates the temporal stability between two consecutive evaluations.

### **4. Discussion**

In this study, we translated and validated the "User Reported Measured of Care Coordination" questionnaire into Spanish. This tool aims to capture the perceptions of older people with chronic complex conditions accessing care across organizational and professional boundaries, to identify facilitators and barriers to care coordination activities in our context.

Correlation analysis indicated, on one hand, moderate correlations between Health and Well-being (D1) and Health Day to Day (D2) with other domains and, on the other hand, Social Services (D3) moderately correlated with Planned Care (D4) and Urgent Care (D5) moderately correlated with Hospital Care (D6). However, a high non-response rate was found in 3 specific domains, Social Services (D3), Planned Care (D4) and Hospital Care (D6), as these services had not recently been used by many participants in our study. The non-use of Hospital Care (D6) may be due to the health-related circumstances of each individual but access to Social Services (D3) or Planned Care (D4), considering their specific complexity very much influenced by the characteristics of healthcare and social care systems. Poor access to social services, mostly due to economic issues and system fragmentation, could directly affect the health status of the population [18]. A high percentage of the

population are not aware of the existence of care planning [6]. The care plan should be a right of the population so that they can cope better with their chronic complex conditions, and this could be ensured through a formal agreement with our patients [19]. Again, there was a high percentage of missing data when studying the temporal stability in Social Services (D3), Planned Care (D4), and Hospital Care (D6) domains, as described in the original article [16]. According to the results on internal consistency, the "alpha if item deleted" index suggested to remove Q28. Authors finally decided to modify the original scale excluding this item. This was the only difference compared with the original instrument.

This instrument is adapted to our Spanish context following current policies to improve health and social care coordination with a variety of specific strategies across the country. Only a few European countries (e.g., Denmark, Portugal, and Ireland) enjoy a system that integrates social and health care horizontally. In most countries (Austria, Belgium, Spain, Italy, United Kingdom, among others), horizontal fragmentation between the social and health sectors accompanies a vertical division of responsibilities in which three levels of administration (national, regional, and local) with different allocation of power [6,20]. Thus, it is important to develop instruments to measure and improve coordination for each specific context [19]. Access, suitability, and financial sustainability as measurement indicators related to care provision are available in most countries. The quality of care coordination in chronic, complex conditions, however, is a multidimensional phenomenon that is difficult to measure; sometimes, the available data are only numeric measurements, and these do not cover the quality of care from population's point of view. It is, therefore, almost impossible to make comparisons between countries [21].

In terms of the care coordination process, the following factors appear to be important design features. At a personal level, a holistic focus that allows service users and carers to become more functional, independent, and resilient, and to live well by managing conditions in the home environment. At a service level, it is important to encourage multiple referrals to a single-entry point where care coordination can be supported. At a community level, the role of members of the local community should be seen as integral to the caregiving process. At a functional level, effective communication between members of the multidisciplinary team is essential. At an organizational level, effective targeting of service users is required to prioritize care provision. At a system level, integrated health and social care commissioning can support longer-term strategies and provide a greater degree of stability. Therefore, a political narrative that supports person-centered care coordination provides credibility when developing new ways of working [22,23].

Although the need for care coordination is clear, we have seen there are obstacles within most health care systems [24,25]. This must be overcome and has to be considered a priority. Health care systems remain disjointed and processes vary among and between primary care sites and specialty sites [26]. Patients are often unclear about why they are being referred from primary care to a specialist, and what to do after specialist visits. Referral staff deal with many different processes and lost information, which means that care is less efficient.

Addressing patient health from a broader perspective includes addressing the social determinants that influence health outcomes and engaging patients as collaborative partners in their own wellness. Health systems are beginning to recognize this dynamic interplay between an individual's social needs, healthy lifestyles, and behaviors and their corresponding health status [24,25].

Previous strategies gave us some recommendations to improve care coordination strategies: Ensure patient—and family—centeredness by identifying family needs and goals that can be supported by a care coordinator [24]. Ensure open lines of communication with the designated care coordinator [27]. Develop a shared understanding of community services and local supports for patients and families. Support interprofessional team training on implementing and measuring care coordination initiatives [28]. Use tools and measures to assess the added value of care coordination for patients and families, including patient/family experience as an outcome measure [24,29].

Our study has some limitations. The results on internal consistency should be interpreted with caution as assessing the internal consistency required omitting conditional items or self-information items. Finally, we believe it would have been preferable to determine the correlation with another tool specifically validated for measuring population perception of care coordination. However, this was impossible, since no other tool was available in the Spanish language. Moreover, although the sample was selected from the whole of Catalonia, there was a particular health management area that could not perform data collection due to a new, recently implemented research intervention.

### **5. Conclusions**

Successful deployment of care-coordination is a big challenge that should benefit of tools capturing the experiences of patients as they access care across organizational and professional boundaries. The adapted version of the "User Reported Measure of Care Coordination" into Spanish proved to be a practical tool for use in our daily practice and an efficient instrument for the evaluation of care coordination in chronic, complex conditions in older people across services and levels of care.

Further research is required to assess the extent to which survey data offer information that providers and purchasers can use to make specific changes to improve the quality of their services.

**Author Contributions:** Conceptualization, E.R., A.Z., P.A., A.A. (Anna Albero), G.S., A.R., J.M. and N.F.; methodology, E.R., A.Z., N.A.-D., B.K., A.A. (Albert Alonso) and A.A. (Anna Albero); software, N.A.-D. and B.K.; validation, N.A.-D., B.A. and E.R.; formal analysis, N.A.-D., B.K. and E.R.; investigation, E.R., G.S., A.Z., A.A. (Anna Albero), J.M., A.R. and A.A. (Albert Alonso); resources, E.R., G.S., A.A. (Albert Alonso), N.F., A.A. (Anna Albero), N.A.-D., P.A. and B.K.; data curation, N.A.-D., B.K. and E.R.; writing—original draft preparation, E.R., A.Z., B.K., N.A.-D. and G.S.; writing—review and editing, E.R., A.Z., B.K., N.A.-D., A.A. (Albert Alonso), J.M., A.R., N.F., G.S. and A.A. (Anna Albero); visualization, N.A.-D., A.Z., E.R., G.S., N.F., J.M. and B.K.; supervision, E.R., A.Z., P.A., A.A. (Albert Alonso), G.S., A.Z., A.R., J.M. and N.F.; project administration, E.R., G.S., A.Z., J.M. and A.A. (Albert Alonso); funding acquisition, E.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was funded by the Agència de Qualitat i Avaluació Sanitàries de Catalunya, Departament de Salut, Generalitat de Catalunya and Fundació Infermeria i Societat del Col·legi Oficial d'Infermeres i Infermers de Catalunya.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

International Journal of *Environmental Research and Public Health*

### *Article* **Sex Di**ff**erences in Frail Older Adults with Foot Pain in a Spanish Population: An Observational Study**

**Emmanuel Navarro-Flores <sup>1</sup> , Carlos Romero-Morales 2,\* ,**

**Ricardo Becerro de Bengoa-Vallejo <sup>3</sup> , David Rodríguez-Sanz <sup>3</sup> , Patricia Palomo-López <sup>4</sup> , Daniel López-López <sup>5</sup> , Marta Elena Losa-Iglesias <sup>6</sup> and César Calvo-Lobo <sup>3</sup>**


Received: 28 July 2020; Accepted: 19 August 2020; Published: 24 August 2020

**Abstract:** Frailty is a condition that can increase the risk of falls. In addition, foot pain can influence older adults and affect their frail condition. The main objective was to measure the frailty degree in older adults in a Spanish population with foot pain from moderate to severe. Method: This is a cross-sectional descriptive study. A sample of people older than 60 years (*n* = 52), including 26 males and 26 females, were recruited, and frailty disability was measured using the 5-Frailty scale and the Edmonton Frailty scale (EFS). Results: Spearman's correlation coefficients were categorized as weak (rs ≤ 0.40), moderate (0.41 ≤ rs ≥ 0.69), or strong (0.70 ≤ rs ≥ 1.00). There was a statistically significant correlation for the total score (*p* < 0.001) and most of the subscales of the 5-Frailty scale compared with the EFS, except for Mood (*p* > 0.05). In addition, females and males showed similar 5-Frailty and Edmonton Frail scales scores with no difference (*p* > 0.05). Conclusion: Foot pain above 5 points, i.e., from moderate to severe, does not affect the fragility more in one sex than another.

**Keywords:** frailty; older adults; foot deformities; foot diseases; foot pain

### **1. Introduction**

Aging and chronical illness processes, like hyperglycemic disease, as well as muscle skeletal and heart processes can produce frailty syndrome, and as a consequence of this, be degenerative and exhibit some alterations that can affect one's mental and general health [1]. For example, aging and frailty can affect gait speed and increase fall risk due to balance alterations [2–4]. Furthermore, the presence of frailty symptoms affects the health-related quality of life (HQoL) [5] in this population group.

We can define frailty syndrome as a group of health alterations that can affect several aspects, such as the psychological, biological and social aspects, as it is a consequence of a dynamic process that reduces a person's health status [6]. Regarding foot conditions in older adult populations, foot disorders and diseases are present most frequently in the frailty population group, comprising approximately 25% [7,8].

The frequency of a frail state in people older than 65 years has been estimated between 4 and 59.1% [9].

Consultations at general practitioners related to ankle and foot conditions of osteoarticular pain origin account for more than 8% [10]. Accordingly, suffering may raise this predominance in older adults who have characteristic foot requirements that can be akin to bigger disorders [11], foot health-related quality of life (HQoL) [12], and risk of falls [13,14]

The 5-Frailty scale is a questionnaire of 5 items, which was established using an auto-administered dimension [6]. Respondents can provide affirmative or negative answers, with one punctuation to the positive response. For measuring the degree of frailty, respondents can score between zero and five points, and subjects are qualified as robust (zero point), pre-frail (one to two points), or frail (three points). The level represents their respective tiredness, resistance, ambulation, disease, and loss of weight.

Tiredness is evaluated by asking subjects if they felt tired; resistance was determined by each subject's report on their capacity to climb stairs; ambulation is represented by each subject's information on their ability to move around; illness is determined by the presence of more than five of a total of eleven pathologies, including cardiovascular diseases, diabetes, and loss of weight by a reduction of five percent during the last year [15].

The Edmonton Frailty scale (EFS) assesses nine subscales: (1) cognitive, (2) general health status, (3) independence, (4) social support, (5) pharmacologic treatment, (6) feeding, (7) mood, (8) continence, and (9) functional performance, using eleven questions. The maximum score is 17 and represents the highest degree of frailty [16]. In this case, a frailty score between zero and four does not present frailty, scores of five to six represent apparently vulnerable, scores of seven to eight represent fair frailty, scores of nine to ten represent frailty moderate, and scores of eleven or more represent severe frailty [17].

No study has yet correlated the scores of the EFS and the 5-Frailty scale. Therefore, the goal of this research was to correlate the subscales of the EFS and 5-Frailty scale in older adults and those with related foot pain.

In the literature, no references have been found regarding the frailty of older adults with foot pain related to sex, and therefore our hypothesis is that there are differences in the levels of frailty in older adults with foot pain related to sex. The objective of this study is to determine if sex can influence a greater degree of frailty.

### **2. Materials and Methods**

The study was developed in Spain. We recruited older adult patients in a medical center, a rehabilitation service, and podiatry clinics in the Generalitat of Valencia, and all survey data were collected between September 2019 and January 2020. Researchers obtained signed informed consent from all subjects, and observational research was carried out using STROBE [18].

Considering sex distribution as the independent variable to calculate the sample size, G\* Power 3.1.9.2 software (Heinrich-Heine-Universität Düsseldorf; Düsseldorf, Germany) was used after considering testing the sex differences between two independent means about sex, comparing their frailty scores. A two-tailed hypothesis, a large effect size of 0.8, an error of α = 0.05 with a 95% confidence interval, an error β = 20%, and a power analysis of 1 − β = 0.80 were considered. Consequently, a total sample size of 52 subjects with 26 in each group was included in this study.

Prior to beginning the research, approval for conducting this study was obtained from the Ethics Committee of the University of Extremadura, 1/2020.

Informed consent was obtained from each participant after the purpose and process of the study were explained and the privacy of the participants' information was assured. The fact that their participation was entirely voluntary was also highlighted.

The inclusion criteria comprised adult patients older than 60 years who presented foot pain during the last 6 months due to toe or foot deformities, regardless of its origin or cause; higher than five points in the VAS score, excluding wounds; and able to communicate orally and provide written informed consent. VAS scores above five points, i.e., from moderate to severe, showed an intraclass correlation coefficient reliability of 0.870 [19].

The exclusion criteria were major neurocognitive disorder; patients who did not answer initial identification questions or who did not understand the rules of participation; and those who refused to participate in the research, i.e., no signed consent.

To recruit volunteer participants, we posted recruitment flyers in senior centers' gathering places. We also addressed groups of older adults in the center to invite them to contact us if they were willing to participate in the study. Once a potential participant expressed interest, a cognitive function evaluation was performed by a gerontological nurse practitioner (GNP) to establish the cognitive eligibility of the participant. Following evaluation by the GNP, the investigators explained the study procedures in detail to the participants.

The interview was composed of general questions of general health status, socio demographic characteristics (sex, age, BMI, height, and weight), comorbidities (e.g., anxiety, depression, diabetes, obesity, osteoarticular diseases, vascular disorders, and kidney illness) collected from medical records. Furthermore, specific items related to foot pain, such as the actual treatment or foot deformities, were assessed by a senior podiatry physician (ENF).

In this study, a total of 65 older adults expressed interest in study participation, and all met the cognitive requirements. The participants all attempted to complete the survey questionnaires. Ultimately all surveys were analyzed for the study; 14 surveys were excluded due to incomplete answers. For participants who were not able to read the questionnaires due to vision problems, the investigators read the questions aloud and marked the participants' answers on the questionnaires. Participants took about 15 min to complete the questionnaires. Participants did not get any compensation for their participation in the study.

### *2.1. Evaluation of Frailty*

The EFS was designed to measure frailty in nine subscales: cognitive, general health status, independence, social support, pharmacologic treatment, feeding, mood, continence, and functional performance [16,20]. Total scores range from 0 to 17, and higher scores indicate more frailty, ranging from 1 = not frail; 2 = ostensibly susceptible frail; 3 = almost never; to 4 = almost always. The EFS ranged by 3 degrees. The Cronbach's α for the EFS was 0.93 [21].

First, a strong frailty mark and subjects without frailty were classified using the EFS. Subjects who scored less than five points were designated as not frail. Secondly, ostensibly susceptible frail subjects were designated as those who obtained six to eleven points. The third group included those who scored between twelve and seventeen points. The questionnaire administration required only 15 min to complete.

Participants also completed the 5-item Frailty scale [22], which has been previously validated into Spanish with an ICC = 0.82 [6,23]. This scale measured five subscales: tiredness, resistance, go around, disease, and loss of weight. Frailty scores ranged from 0 to 5, and higher scores indicate more frailty. Those participants that scored between three and five were considered frail, those that scored one or two were considered to be pre-frail, and those that obtained zero points were considered to be not frail.

The authors obtained permission from the original authors of the EFS and 5-Frailty scale to use their clinimetric tool to measure the frailty degree.

### *2.2. Statistical Analysis*

All the variables were normally distributed, as determined by the Kolmogorov–Smirnov test (*p* > 0.05).

Regarding the results of the quantitative variables, the non-parametric data were described in terms of their median, interquartile range (IR), and minimum and maximum (range) values. Parametric data were described using means, the standard deviation (SD), and minimum and maximum (range) values.

A comparison of the quantitative data between males and females for the different questionnaire subscales of the EFS and 5-Frail scale was conducted, and significant differences were checked using an independent Student's *t*-test. Non-normal data were analyzed using Mann–Whitney U tests.

Spearman's correlation coefficients (rs) were determined between tests and qualified as low rs ≤ 0.40, moderate 0.41 ≤ rs ≥ 0.69, or solid 0.70 ≤ rs ≥1.00.

All analyses were considered statistically significant when the *p*-value < 0.05 with a 95% confidence interval (CI). Statistical analyses were made in SPSS (V.26.0, Chicago, IL, USA).

### **3. Results**

### *3.1. Descriptive Data and Socio-Demographic Data*

A normal distribution for age, height, weight, and BMI was shown (*p* > 0.05), and all items from the 5-Frailty test and the EFS showed no normal distribution (*p* < 0.05).

The size sample included 52 subjects whose mean age was 77.47 ± 10.69 years. The study subjects included 26 (50.00%) females and 26 (50.00%) males. Table 1 shows the socio-demographic characteristics. Males and females did not show statistically significant socio-demographic differences (*p* > 0.05) for age or BMI, although a higher weight and height (*p* < 0.05) were shown for males compared with females. There was no difference in the intensity of pain between sexes (*p* = 0.561).


**Table 1.** Descriptive and socio-demographic data of the sample.

BMI: body mass index; Mean-standard deviation, range (min–max), and Student's *t*-test for independent samples were applied. In all the analyses, *p* < 0.05 (with a 95% confidence interval) was considered statistically significant.

### *3.2. Edmonton Frail Scale and 5-Frailty Scale Sex Distribution*

As shown in Table 2, the 5-Frailty scale scores did not manifest any statistically significant difference (*p* > 0.05) for the subscales nor for the total scores between females and males with foot pain. Furthermore, the EFS scores by sex distribution are shown in Table 3, whose subscales did not show statistically significant differences (*p* > 0.05). The results of both test show that frailty in males and females with foot pain is similar.


**Table 2.** Comparisons of the 5-Frailty scale scores between males and females.

CI, Confidence Interval; IR: interquartile range. Mann–Whitney U tests were used. In all the analyses, *p* < 0.05 (with a 95% confidence interval) was considered statistically significant.

**Table 3.** Comparisons of the Edmonton Frail Scale scores between males and females.


CI, Confidence Interval; IR: interquartile range. Mann–Whitney U tests were used. In all the analyses, *p* < 0.05 (with a 95% confidence interval) was considered statistically significant.

Table 4 shows a good correlation between items of both tests, except for the subscale nutrition (*p* = 0.490).


**Table 4.** Spearman's correlations between the 5-Frailty scale and Edmonton Frail Scale score domains and totals.

Spearman correlation coefficients (r) and *p*-values were applied. In all the analyses, *p* < 0.05 (with a 95% confidence interval) was considered statistically significant.

### **4. Discussion**

This investigation aimed to compare the frailty degree differences between males and females with foot pain from moderate to severe using the 5-Frailty Scale and EFS. Indeed, frailty did not show statistically significant differences between the sexes. Although the total frailty score Spearman correlation was significant in Table 4, it is worthy to note that the correlation was modest, even when comparing domains, which were expected to correlate with functional independence and performance from the EFS and with resistance and ambulation from the 5-Frailty scale.

Both scales are correlated, which confers concurrent validity of each subscale to recent studies and sustains the application of the 5-Frailty score as an acceptable measurement related to frailty aspects such as ambulation, illness, or loss of weight. This aspect can be considered as an advantage with respect to other frailty scales adapted to Spanish to evaluate specific frailty aspects, like the Frailty Trait Scale (FTS) [24].

Moreover, the results from this study are different from those obtained by other researchers in functional performance related to foot pain [25,26]. Our results did not show statistically significant differences related to foot pain [5,8]. These findings are different, with similar research reported on subjects with foot pain [1], as well as with foot pain [27] as measured by employing the visual analogic scale (VAS) to determine the sex differences in healthy subjects with foot problems.

There was a statistically significant correlation in the total score and most of the subscales of the 5-Frailty scale compared with the EFS, except for Mood (*p* > 0.05).

Furthermore, prior studies have identified females who suffered from fibromyalgia and had developed foot problems as a consequence had an increased frailty degree [28].

Future studies should incorporate all other foot risk factors related with frailty syndrome. Despite the fact that the EFS has determined the frailty score [29,30].

Several limitations of this research should be taken into account. A population from different territories may be useful to ameliorate the strength of this research.

This research has only determined if foot pain can influence a greater degree of frailty by sex and we found that foot pain does not affect the frailty by sex.

Although ambulation and functional performance and risk of fall are very common in frail people [2,4], this research should also be developed for other population groups to determine the degree of frailty, for example, in widows who usually have higher frailty scores due to psychosocial aspects [29,31,32].

Furthermore, selective sampling can cause bias; for this reason, randomized sampling should be considered in future studies.

Ultimately, the correlation impact between the different foot disorders, including several genetics and acquired or traumatic alterations and chronic illness, was not studied in our research because the studied population was not suitably adjusted to develop these comparisons. The researchers therefore suggest that future research should be conducted considering different foot pathologies.

### **5. Conclusions**

Foot pain above 5 points, i.e., from moderate to severe, does not affect the fragility more in one sex than another. Further research is needed considering different foot pathologies.

**Author Contributions:** Conceptualization, E.N.-F., P.P.-L., D.R.-S., R.B.d.B.-V., M.E.L.-I., C.C.-L. and D.L.-L.; methodology, E.N.-F., P.P.-L., C.R.-M., C.C.-L. and D.L.-L.; software, R.B.d.B.-V. and C.C.-L.; formal analysis, R.B.d.B.-V. and C.C.-L.; investigation, E.N.-F., P.P.-L., C.R.-M., D.R.-S. and D.L.-L.; resources, E.N.-F., P.P.-L., C.R-M., C.C.-L. and D.L.-L.; data curation, C.C.-L.; writing—original draft preparation, E.N.-F., P.P.-L., C.R.-M., C.C.-L. and D.L.-L.; writing—review and editing, E.N.-F., P.P.-L., E.N.-F., C.C.-L. and D.L.-L.; supervision, R.B.d.B.-V., M.E.L.-I., D.R.-S., C.C.-L. and D.L.-L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

International Journal of *Environmental Research and Public Health*
