Identification of frailty in primary care: accuracy of electronically derived measures

Karin MA Swart; Amber AWA van der Heijden; Marieke T Blom; Jetty A Overbeek; Giel Nijpels; Hein PJ van Hout; Petra JM Elders; Ron MC Herings

doi:10.3399/BJGP.2022.0574

Abstract

Background Routinely collected clinical data based on electronic medical records could be used to define frailty.

Aim To estimate the ability of four potential frailty measures that use electronic medical record data to identify older patients who were frail according to their GP.

Design and setting This retrospective cohort study used data from 36 GP practices in the Dutch PHARMO Data Network.

Method The measures were the Dutch Polypharmacy Index, Charlson Comorbidity Index (CCI), Chronic Disease Score (CDS), and Frailty Index. GPs’ clinical judgement of patients’ frailty status was considered the reference standard. Performance of the measures was assessed with the area under the receiver operating characteristic curve (AUC). Analyses were done in the total population and stratified by age and sex.

Results Of 31 511 patients aged ≥65 years, 3735 (11.9%) patients were classified as frail by their GP. The CCI showed the highest AUC (0.79, 95% confidence interval [CI] = 0.78 to 0.80), followed by the CDS (0.69, 95% CI = 0.68 to 0.70). Overall, the measures showed poorer performance in males and females aged ≥85 years than younger age groups (AUC 0.55–0.58 in females and 0.57–0.60 in males).

Conclusion This study showed that of four frailty measures based on electronic medical records in primary care only the CCI had an acceptable performance to assess frailty compared with frailty assessments done by professionals. In the youngest age groups diagnostic performance was acceptable for all measures. However, performance declined with older age and was least accurate in the oldest age group, thereby limiting the use in patients of most interest.

INTRODUCTION

Frailty is a common condition at older ages, characterised by loss of biological reserves across multiple organ systems and vulnerability to physiological decompensation after a stressor event.¹ Frailty is associated with poor health outcomes, including falls, disability, admission to hospital, and mortality.²^–⁶ Given the increased numbers of older people with frail health, care models should include frailty to focus on optimising health and avoiding the admission to hospital of frail and well older adults alike. Detection of frail older people can support timely management to maintain or improve functioning.⁷ Screening tools, such as frailty scales, and an understanding of a patient’s cognitive condition, physical function, and functional reserve, might alert the physician to start frailty management.⁸

Many frailty measures have been developed to identify patients with frail health in clinical practice.⁹ The most commonly used method to identify frailty in research settings combines questionnaires and functional measures.¹⁰ Alternatively, frailty has been operationalised, among other measures, as a risk index by counting the number of impairments accumulated over time, including disability, diseases, physical and cognitive impairments, psychosocial risk factors, and geriatric syndromes.¹¹ Furthermore, methods have been developed to use routinely collected clinical data based on electronic medical records to define frailty. A significant advantage of these measures for clinicians is that no additional data collection is needed. They can be easily applied, thereby increasing their applicability in research and care settings, and they might make the identification process of frail older people more efficient.

However, hardly any studies have validated frailty measures against a diagnostic reference standard such as clinical judgement. Most validations reported associations with future adverse events. Varying results regarding the strength of the associations with mortality might be caused by varying distributions of the age and sex of the validation populations.¹² In addition, Clegg et al created categories from fit to severe frailty purely on statistical distribution in an adult population between 65 and 95 years of age. It is still not clear how this categorisation relates to the clinical judgement of professionals.¹³

Routinely collected clinical data might aid healthcare professionals in identifying frail older people. However, there is a lack of studies that have validated frailty measures against a diagnostic reference standard such as clinical judgement. In this study, it was found that, among the four measures evaluated, only the Charlson Comorbidity Index had an acceptable level of performance for assessing frailty, regardless of age. Although all four measures can be used to identify frailty in young older people (65‒74 years), their performance declined with increasing age.

How this fits in

The current study aims to estimate and compare the ability of four potential frailty measures used in research and clinical practice that make use of electronic medical records to identify older patients who were actually considered frail according to their GP. The four measures are the Dutch Polypharmacy Index (DPI), the Charlson Comorbidity Index (CCI), the Chronic Disease Score (CDS), and the Frailty Index (FI). In addition, the study aimed to compare the diagnostic performance of these measures across sex and age groups. The hypothesis was that electronic medical records can be used to identify frail older people, and the diagnostic performance might differ across age groups.

METHOD

Study population

Data for this retrospective cohort study were obtained from 36 GP practices from the PHARMO Data Network in the Netherlands that routinely coded frailty as part of older care programmes in 2019.¹⁴ These practices served a total population of 31 511 patients aged ≥65 years. The electronic medical records of the GP practices include information on diagnoses and symptoms, laboratory test results, and referrals to specialists and healthcare product/drug prescriptions. Diagnoses and symptoms were coded according to the International Classification of Primary Care (ICPC),¹⁵ and prescription drugs were coded according to the World Health Organization Anatomical Therapeutic Chemical (ATC) Classification System.¹⁶

All data from patients and practices were anonymised.

Reference standard

GPs’ clinical judgement of patients’ frailty status was considered the reference standard (ICPC diagnosis code A05, derived from episodes). Within the older care programme, no strict definition of frailty was used, as the group of frail older people is heterogeneous by definition. Instead, a pragmatic definition was applied with loss of autonomy as a core manifestation and starting point for frail older people. This was judged by the GP. The GP’s clinical judgement of frailty has been shown to be an accurate indicator of frailty and a strong predictor of future mortality and long-term care admission.¹⁰^,¹⁷

Frailty measures

Four measures used in research and clinical practice to distinguish patients who are frail from those who are not are the DPI, the CCI, the CDS, and the FI (Box 1). All four multimorbidity measures are widely used in epidemiological studies and, especially the DPI and FI, in clinical practice.¹⁸

Characteristic	DPI	CCI	CDS	FI
Definition	Regular use of ≥5 medicines	Sum of weighted morbidity scores, based on its mortality risk	Comorbidity score based on the aggregate number of prescription medications	Number of health deficits divided by the total number of 50 deficits
Initial purpose	To identify polypharmacy	To predict mortality risk attributable to comorbidity	To predict health outcomes	To predict adverse health outcomes in older people
Setting where it has been developed	Primary care	Hospital	Pharmacy	Primary care
Updates	NA	ICPC codes mapped to ICD codes	Including novel pharmacotherapies	NA
Input data	Medication records — number of ATC codes (third level)	Comorbidity records — ICPC codes for comorbidity conditions	Medication records, age, sex — ATC classes of medication for treatment of different chronic diseases	Health deficits — ICPC codes of 50 health deficits
Setting in which it is mainly applied	Research + clinical practice	Research	Research	Research + clinical practice

ATC = Anatomical Therapeutic Chemical. CCI = Charlson Comorbidity Index. CDS = Chronic Disease Score. DPI = Dutch Polypharmacy Index. FI = Frailty Index. ICD = International Classification of Disease. ICPC = International Classification of Primary Care. NA = Not applicable.

Box 1.

Global characteristics of the four frailty measures

The DPI is based on the concurrent regular use of medications, based on medication prescriptions. It is defined as the concurrent regular use (at least three single prescriptions, including at least one prescription in the preceding 6 months) of five or more medicines.¹⁹ The use of several medicines within one pharmacological subgroup (ATC third level) is counted as one.

The CCI was initially developed to measure the risk of 1-year mortality attributable to comorbidity and is based on diagnoses registered in the GP medical records. The CCI included 19 conditions that are weighted based on the severity of the condition. The CCI is calculated by summation of the weighted comorbidity scores.²⁰

The CDS is a comorbidity measure based on 1 year of medication prescription data and age and sex. Classes of medication are weighted to correspond to disease complexity and severity.²¹ The CDS was adapted by the research group to also include additional ATC codes of newly developed drugs to the medication classes.

The FI is based on a predefined list of 50 health deficits. The FI (range 0 to 1) is calculated by dividing the number of present deficits in a patient by all 50 deficits. The lookback period is 6 months (for instance, for mood symptoms) or 5 years (for instance, for fractures), depending on the clinical relevance.²²

Statistical analysis

The characteristics of the study population are presented for the total study population and stratified by age (65–74, 75–84, and ≥85 years). Categorical variables are presented as numbers and proportions, and continuous variables were presented as mean (standard deviation [SD]) or median (interquartile range [IQR]) based on their distributions. In the total population and in subgroups of age and sex the ability of the index to distinguish between patients who were frail and those who were not according to the GP was assessed by calculating the area under the receiver operating characteristic curve (AUC). An AUC was considered excellent for values between 0.9 and 1.0, good for values between 0.8 and 0.9, acceptable for values between 0.7 and 0.8, poor for values between 0.6 and 0.7, and failed for values between 0.5 and 0.6.

The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for each index. The optimal cut-off values for identifying the frailty of each index within each subgroup were based on the Youden index (sensitivity + specificity −1) maximising the sum of sensitivity and specificity of each index with an equal weight of the two measures. Calibration (that is agreement between predicted and observed frailty incidence) was assessed by visual inspection of calibration plots and the observed and expected frailty incidence ratio. Calibration plots were created by plotting the observed mean incidence of frailty against the expected mean frailty incidence within deciles of the predicted probability of frailty.

Differences between GPs in their opinion on frailty may result in discrepancies in the assignment of frailty status in patients with similar comorbidity profiles but in different GP practices. To take these potential differences in frailty assignment into account, a sensitivity analysis was performed by testing the performance of the measures stratified by GP practice. Moreover, the performance of the measures was tested in the subgroup of GPs with an age- and sex-standardised frailty prevalence within the IQR of the total population. All analyses were performed using R (version 4.2.2).

As the underlying data represent attended medical care, it was assumed that the absence of a record meant no occurrence, for example, if an indicator of disease was missing for a patient, it was assumed that the patient did not have the disease.

RESULTS

Of the total study population of 31 511 patients (mean age 75.0 years, 45.9% males), 3735 (11.9%) patients were classified as frail by their GP (Table 1). As expected, with increasing age groups a higher proportion of the people was classified as frail by their GP (Table 1).

View this table:

Table 1.

Characteristics of the study population, according to age

Frailty measures in total study population

In the total study population, the CCI showed the highest AUC (0.79, 95% confidence interval [CI] = 0.78 to 0.80) followed by the CDS (0.69, 95% CI = 0.68 to 0.70) (Table 2). Sensitivity to identify frailty was highest when using the DPI or the CCI (0.72, 95% CI = 0.71 to 0.73 and 0.74, 95% CI = 0.73 to 0.76, respectively). Specificity was highest for the CCI (0.70, 95% CI = 0.69 to 0.70). Calibration plots of the four measures are presented in Supplementary Figure S1. There was a good agreement between predicted incidence and observed incidence in different deciles of the predicted probability of frailty of the four measures.

View this table:

Table 2.

Diagnostic performance of four measures to identify frailty^a

Frailty measures in subgroups of age and sex

The performance of the four frailty measures showed different results when applied in subgroups of age categories and sex (Figure 1). The ability of the measures to discriminate between frail and non-frail decreased with increasing age. In patients aged 65–74 years, the AUCs ranged from 0.70 to 0.76 in males and 0.73 to 0.78 in females. In the 75–84 age group, the AUCs decreased to 0.63 to 0.70 in males and 0.60 to 0.67 in females. In the oldest age groups (≥85 years), the ability to identify frailty further decreased in males (0.57 to 0.60) and females (0.55 to 0.58). In all age groups, the CCI showed the most favourable results.

Figure 1.

Area under the receiver operating characteristic curve (AUC) for each index according to age group (65–74 years, top; 75–84 years, middle; ≥85 years, bottom) in (a) males (b) and females. CCI = Charlson Comorbidity Index. CDS = Chronic Disease Score. DPI = Dutch Polypharmacy Index. FI = Frailty Index.

The diagnostic performance of the four measures was expressed as sensitivity, specificity, PPV, and NPV, with calculations based on the optimal cut-off value specific for each subgroup. Overall, the four measures showed poorer performance in males and females aged ≥85 years than younger age groups (Tables 3 and 4, respectively). A large decrease in the negative predictive value was seen in the oldest age group, which was more pronounced in females than in males, meaning that a larger proportion of patients classified as non-frail by the measures were considered frail according to the GP.

View this table:

Table 3.

Diagnostic performance of the four measures identify frailty according to age groups in males^a

View this table:

Table 4.

Diagnostic performance of the four measures identify frailty according to age groups in females^a

Sensitivity analyses

When the performance of the frailty measures was calculated for each GP separately, the performance was again best for the CCI, and ranged from AUC 0.71 (95% CI = 0.65 to 0.78) to AUC 0.88 (95% CI = 0.83 to 0.93) (data not shown in table). For the majority of the GPs, the performance decreased with the older age subgroup, with an AUC below 0.7 in the oldest age group, in approximately 90% of the GPs for the four frailty measures.

The median age- and sex-standardised prevalence of frailty per GP practice, as classified by the GP, was 11.0% (IQR 9.5–14.0). The performance of the frailty measures was tested when only including GP practices with age- and sex-standardised frailty prevalence within the IQR. The measures’ performance was similar compared with the total population (Supplementary Figure S2).

DISCUSSION

Summary

This study investigated the diagnostic performance of the DPI, CCI, CDS, and FI for the identification of frail older adults with the use of electronic medical records of GPs compared with the clinical judgement of GPs. An acceptable performance, based on the AUC, was found for the CCI in the total sample, and a poor performance for the DPI, CDS, and FI. When stratifying the results according to age, the diagnostic performance was acceptable for all indexes in the youngest age group (65–74 years). However, the performance decreased for the higher age groups, showing a poor to failed performance in patients aged ≥85 years, and worse performance in females than males.

Strength and limitations

A limitation of the current study includes the clinical judgement of the GP as a dichotomous definition, thereby ignoring the complexity of frailty. The use of more than two frailty categories has been suggested. For instance, the electronic FI as implemented in the UK uses four frailty categories.¹³ A strength of the current study was the use of GPs’ clinical judgement as a reference standard. Most previous studies validated the FI by prognostically reporting associations with future adverse events or based on statistical distributions. Although their judgement of frailty will show within- and between-GP variation, the GPs’ judgement on the presence and absence of frailty was found to be the best predictor of mortality.¹⁰ The sensitivity analysis among GP practices with a frailty prevalence within the IQR showed that the AUCs were similar compared with the AUCs in the total population. This indicates that the indexes were robust. Furthermore, it should be noted that data were derived from GP practices that routinely coded frailty as part of older care programmes. This ensures that frailty was registered. Another strength was the large sample size that made it possible to stratify the population to compare the performance of the frailty measures across sex and age groups.

Comparison with existing literature

A previous study that evaluated a polypharmacy score and the FI in a primary care sample against Fried’s frailty criteria and clinical judgement by an expert panel showed similar performance as observed in the current study.³^,¹⁷ Furthermore, in a systematic review, the psychometric properties of various frailty measures were investigated, and an association between the FI and several adverse health outcomes was consistently present. However, the ability of the measures to discriminate between people who will experience such an event and those who will not was poor to moderate, with the lowest AUCs in studies consisting of relatively older people.²²^–²⁴ Adjustment for age and sex and consultation gap resulted in an improved AUC.²²

The performance of the multimorbidity-driven measures decreased with increasing age. An explanation might be that, with increasing age, a survival bias of people with relatively few comorbidities occurs.²⁵ This was reflected by a larger proportion of the patients being classified as non-frail by the measures while considered frail by the GP in the oldest age group. In these patients, other factors then multimorbidity may be more important in deciding whether patients are judged as being frail. When classifying patients’ frailty status, GPs typically use a broader definition, also taking into account functional, cognitive, emotional, and social aspects, and the type and number of complications and medicines.⁹^,²⁶^,²⁷ Characteristics related to more sudden changes might be useful to consider specifically, such as acute hospital admissions, falls, or more specific conditions such as dementia, forgetfulness, or incontinence.

Regarding the observed gender differences, previous studies already showed that females accumulate more deficits,²³^,²⁶^,²⁷ resulting in higher scores on a comorbidity index than males. Despite this higher proportion of deficits, the risk of mortality in females is lower than in males because of the higher tolerability of deficits in females, specifically at an older age.²⁸ This may call for the development of sex-specific cut-off scores to increase the diagnostic accuracy of measures.

Implications for research and practice

The results suggest that the current electronically derived measures of frailty are applicable for identifying frailty in individuals up to a certain age in clinical practice and research. However, for the oldest old, it may be necessary to consider additional information to identify frail people.

Future research should explore factors beyond multimorbidity measures, such as characteristics related to more sudden changes including acute hospital admissions, falls, or specific conditions such as dementia, forgetfulness, or incontinence.

To advance research in this field, it would be beneficial to combine data that are available in GP practices with more comprehensive data sources. For example, data from nursing homes could provide valuable additional insights.

Acknowledgments

Prior presentation: poster presentation at the International Conference of Pharmacoepidemiology (ICPE) 2022, 24–28 August 2022, Copenhagen, The authors would like to thank all the healthcare providers contributing information to the PHARMO Data Network.

Notes

Funding

This work has been supported in kind by Stichting Informatievoorziening voor Zorg en ONderzoek (STIZON).

Ethical approval

The study was approved by the Institutional Review Board of Stichting Informatievoorziening voor Zorg en Onderzoek (STIZON, reference number CC2022-09).

Data

The dataset generated and analysed during the current study are not publicly available but are available from the corresponding author on reasonable request and after approval of the Compliance Committee of Stichting Informatievoorziening voor Zorg en Onderzoek.

Provenance

Freely submitted; externally peer reviewed.

Competing interests

Karin MA Swart and Jetty A Overbeek are employees and Ron MC Herings is scientific director of the PHARMO Institute for Drug Outcomes Research. This independent research institute performs financially supported studies for government and related healthcare authorities and several pharmaceutical companies. The other authors have declared no competing interests.

Discuss this article

Contribute and read comments about this article: bjgp.org/letters

Received November 18, 2022.
Revision requested January 9, 2023.
Accepted April 4, 2023.

http://creativecommons.org/licenses/by/4.0/

This article is Open Access: CC BY 4.0 licence (http://creativecommons.org/licences/by/4.0/).

REFERENCES

1.↵
1. Clegg A,
2. Young J,
3. Iliffe S,
4. et al.
(2013) Frailty in elderly people. Lancet 381, 9868, 752–762.
OpenUrl CrossRef PubMed
2.↵
1. Ensrud KE,
2. Ewing SK,
3. Cawthon PM,
4. et al.
(2009) A comparison of frailty indexes for the prediction of falls, disability, fractures, and mortality in older men. J Am Geriatr Soc 57, 3, 492–498.
OpenUrl CrossRef PubMed
3.↵
1. Fried LP,
2. Tangen CM,
3. Walston J,
4. et al.
(2001) Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci 56, 3, M146–M156.
OpenUrl CrossRef PubMed
4.
1. Gill TM,
2. Gahbauer EA,
3. Allore HG,
4. Han L
(2006) Transitions between frailty states among community-living older persons. Arch Intern Med 166, 4, 418–423.
OpenUrl CrossRef PubMed
5.
1. Graham JE,
2. Snih SA,
3. Berges IM,
4. et al.
(2009) Frailty and 10-year mortality in community-living Mexican American older adults. Gerontology 55, 6, 644–651.
OpenUrl CrossRef PubMed
6.↵
1. Xue QL
(2011) The frailty syndrome: definition and natural history. Clin Geriatr Med 27, 1, 1–15.
OpenUrl CrossRef PubMed
7.↵
1. De Lepeleire J,
2. Iliffe S,
3. Mann E,
4. Degryse JM
(2009) Frailty: an emerging concept for general practice. Br J Gen Pract, DOI:https://doi.org/10.3399/bjgp09X420653.
8.↵
1. Boeckxstaens P,
2. De Graaf P
(2011) Primary care and care for older persons: position paper of the European Forum for Primary Care. Qual Prim Care 19, 6, 369–389.
OpenUrl PubMed
9.↵
1. Hoogendijk EO,
2. Afilalo J,
3. Ensrud KE,
4. et al.
(2019) Frailty: implications for clinical practice and public health. Lancet 394, 10206, 1365–1375.
OpenUrl CrossRef PubMed
10.↵
1. Sutorius FL,
2. Hoogendijk EO,
3. Prins BA,
4. van Hout HP
(2016) Comparison of 10 single and stepped methods to identify frail older persons in primary care: diagnostic and prognostic accuracy. BMC Fam Pract 17, 102.
OpenUrl CrossRef PubMed
11.↵
1. Mitnitski AB,
2. Mogilner AJ,
3. Rockwood K
(2001) Accumulation of deficits as a proxy measure of aging. Sci World J 1, 323–336.
OpenUrl
12.↵
1. Gordon EH,
2. Peel NM,
3. Samanta M,
4. et al.
(2017) Sex differences in frailty: a systematic review and meta-analysis. Exp Gerontol 89, 30–40.
OpenUrl CrossRef PubMed
13.↵
1. Clegg A,
2. Bates C,
3. Young J,
4. et al.
(2016) Development and validation of an electronic frailty index using routine primary care electronic health record data. Age Ageing 45, 3, 353–360.
OpenUrl CrossRef PubMed
14.↵
1. Kuiper JG,
2. Bakker M,
3. Penning-van Beest FJA,
4. Herings RMC
(2020) existing data sources for clinical epidemiology: the PHARMO database network. Clin Epidemiol 12, 415–422.
OpenUrl CrossRef PubMed
15.↵
1. Nederlands Huisartsen Genootschap
International classification of primary care, https://www.nhg.org/praktijkvoering/informatisering/registratie-adviezen-icpc-nhg-standaarden/ (accessed 7 Jul 2023).
16.↵
1. WHO Collaborating Centre for Drug Statistics
(2023) ATC/DDD Index 2023, http://www.whocc.no/atc_ddd_index (accessed 19 May 2023).
17.↵
1. Hoogendijk EO,
2. van der Horst HE,
3. Deeg DJ,
4. et al.
(2013) The identification of frail older adults in primary care: comparing the accuracy of five simple instruments. Age Ageing 42, 2, 262–265.
OpenUrl CrossRef PubMed
18.↵
1. Bleijenberg N,
2. ten Dam VH,
3. Drubbel I,
4. et al.
(2013) Development of a proactive care program (U-CARE) to preserve physical functioning of frail older people in primary care. J Nurs Scholarsh 45, 3, 230–237.
OpenUrl
19.↵
1. Nederlandse Vereniging voor Geriatrie
(2020) Richtlijn Polyfarmacie bij ouderen (NVKG, Utrecht) [Guideline polypharmacy in the elderly].
20.↵
1. Charlson ME,
2. Pompei P,
3. Ales KL,
4. MacKenzie CR
(1987) A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 40, 5, 373–383.
OpenUrl CrossRef PubMed
21.↵
1. Von Korff M,
2. Wagner EH,
3. Saunders K
(1992) A chronic disease score from automated pharmacy data. J Clin Epidemiol 45, 2, 197–203.
OpenUrl CrossRef PubMed
22.↵
1. Drubbel I,
2. de Wit NJ,
3. Bleijenberg N,
4. et al.
(2013) Prediction of adverse health outcomes in older people using a frailty index based on routine primary care data. J Gerontol A Biol Sci Med Sci 68, 3, 301–308.
OpenUrl CrossRef PubMed
23.↵
1. Drubbel I,
2. Numans ME,
3. Kranenburg G,
4. et al.
(2014) Screening for frailty in primary care: a systematic review of the psychometric properties of the frailty index in community-dwelling older people. BMC Geriatr 14, 27.
OpenUrl CrossRef PubMed
24.↵
1. Song X,
2. Mitnitski A,
3. Rockwood K
(2010) Prevalence and 10-year outcomes of frailty in older adults in relation to deficit accumulation. J Am Geriatr Soc 58, 4, 681–687.
OpenUrl CrossRef PubMed
25.↵
1. Rothman KJ,
2. Greenland S,
3. Lash TL
(2008) Modern epidemiology (Lippincott Williams & Wilkins, Philadelphia, PA).
26.↵
1. van Kempen JA,
2. Schers HJ,
3. Jacobs A,
4. et al.
(2013) Development of an instrument for the identification of frail older people as a target population for integrated care. Br J Gen Pract, DOI:https://doi.org/10.3399/bjgp13X664289.
27.↵
1. Morley JE,
2. Vellas B,
3. van Kan GA,
4. et al.
(2013) Frailty consensus: a call to action. J Am Med Dir Assoc 14, 6, 392–397.
OpenUrl CrossRef PubMed
28.↵
1. Shi J,
2. Yang Z,
3. Song X,
4. et al.
(2014) Sex differences in the limit to deficit accumulation in late middle-aged and older Chinese people: results from the Beijing Longitudinal Study of Aging. J Gerontol A Biol Sci Med Sci 69, 6, 702–709.
OpenUrl CrossRef PubMed

In this issue

Download PDF

Download PowerPoint

Email Article

Citation Tools

Keywords

More in this TOC Section

Show more Research

Cited By...

Intended for Healthcare Professionals

Tweets by @BJGPjournal

[1] 1.↵
Clegg A,
Young J,
Iliffe S,
et al.
(2013) Frailty in elderly people. Lancet 381, 9868, 752–762.
OpenUrl CrossRef PubMed

[2] Clegg A,

[3] Young J,

[4] Iliffe S,

[5] et al.

[6] 2.↵
Ensrud KE,
Ewing SK,
Cawthon PM,
et al.
(2009) A comparison of frailty indexes for the prediction of falls, disability, fractures, and mortality in older men. J Am Geriatr Soc 57, 3, 492–498.
OpenUrl CrossRef PubMed

[7] Ensrud KE,

[8] Ewing SK,

[9] Cawthon PM,

[10] et al.

[11] 3.↵
Fried LP,
Tangen CM,
Walston J,
et al.
(2001) Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci 56, 3, M146–M156.
OpenUrl CrossRef PubMed

[12] Fried LP,

[13] Tangen CM,

[14] Walston J,

[15] et al.

[16] 4.
Gill TM,
Gahbauer EA,
Allore HG,
Han L
(2006) Transitions between frailty states among community-living older persons. Arch Intern Med 166, 4, 418–423.
OpenUrl CrossRef PubMed

[17] Gill TM,

[18] Gahbauer EA,

[19] Allore HG,

[20] Han L

[21] 5.
Graham JE,
Snih SA,
Berges IM,
et al.
(2009) Frailty and 10-year mortality in community-living Mexican American older adults. Gerontology 55, 6, 644–651.
OpenUrl CrossRef PubMed

[22] Graham JE,

[23] Snih SA,

[24] Berges IM,

[25] et al.

[26] 6.↵
Xue QL
(2011) The frailty syndrome: definition and natural history. Clin Geriatr Med 27, 1, 1–15.
OpenUrl CrossRef PubMed

[27] Xue QL

[28] 7.↵
De Lepeleire J,
Iliffe S,
Mann E,
Degryse JM
(2009) Frailty: an emerging concept for general practice. Br J Gen Pract, DOI:https://doi.org/10.3399/bjgp09X420653.

[29] De Lepeleire J,

[30] Iliffe S,

[31] Mann E,

[32] Degryse JM

[33] 8.↵
Boeckxstaens P,
De Graaf P
(2011) Primary care and care for older persons: position paper of the European Forum for Primary Care. Qual Prim Care 19, 6, 369–389.
OpenUrl PubMed

[34] Boeckxstaens P,

[35] De Graaf P

[36] 9.↵
Hoogendijk EO,
Afilalo J,
Ensrud KE,
et al.
(2019) Frailty: implications for clinical practice and public health. Lancet 394, 10206, 1365–1375.
OpenUrl CrossRef PubMed

[37] Hoogendijk EO,

[38] Afilalo J,

[39] Ensrud KE,

[40] et al.

[41] 10.↵
Sutorius FL,
Hoogendijk EO,
Prins BA,
van Hout HP
(2016) Comparison of 10 single and stepped methods to identify frail older persons in primary care: diagnostic and prognostic accuracy. BMC Fam Pract 17, 102.
OpenUrl CrossRef PubMed

[42] Sutorius FL,

[43] Hoogendijk EO,

[44] Prins BA,

[45] van Hout HP

[46] 11.↵
Mitnitski AB,
Mogilner AJ,
Rockwood K
(2001) Accumulation of deficits as a proxy measure of aging. Sci World J 1, 323–336.
OpenUrl

[47] Mitnitski AB,

[48] Mogilner AJ,

[49] Rockwood K

[50] 12.↵
Gordon EH,
Peel NM,
Samanta M,
et al.
(2017) Sex differences in frailty: a systematic review and meta-analysis. Exp Gerontol 89, 30–40.
OpenUrl CrossRef PubMed

[51] Gordon EH,

[52] Peel NM,

[53] Samanta M,

[54] et al.

[55] 13.↵
Clegg A,
Bates C,
Young J,
et al.
(2016) Development and validation of an electronic frailty index using routine primary care electronic health record data. Age Ageing 45, 3, 353–360.
OpenUrl CrossRef PubMed

[56] Clegg A,

[57] Bates C,

[58] Young J,

[59] et al.

[60] 14.↵
Kuiper JG,
Bakker M,
Penning-van Beest FJA,
Herings RMC
(2020) existing data sources for clinical epidemiology: the PHARMO database network. Clin Epidemiol 12, 415–422.
OpenUrl CrossRef PubMed

[61] Kuiper JG,

[62] Bakker M,

[63] Penning-van Beest FJA,

[64] Herings RMC

[65] 15.↵
Nederlands Huisartsen Genootschap
International classification of primary care, https://www.nhg.org/praktijkvoering/informatisering/registratie-adviezen-icpc-nhg-standaarden/ (accessed 7 Jul 2023).

[66] Nederlands Huisartsen Genootschap

[67] 16.↵
WHO Collaborating Centre for Drug Statistics
(2023) ATC/DDD Index 2023, http://www.whocc.no/atc_ddd_index (accessed 19 May 2023).

[68] WHO Collaborating Centre for Drug Statistics

[69] 17.↵
Hoogendijk EO,
van der Horst HE,
Deeg DJ,
et al.
(2013) The identification of frail older adults in primary care: comparing the accuracy of five simple instruments. Age Ageing 42, 2, 262–265.
OpenUrl CrossRef PubMed

[70] Hoogendijk EO,

[71] van der Horst HE,

[72] Deeg DJ,

[73] et al.

[74] 18.↵
Bleijenberg N,
ten Dam VH,
Drubbel I,
et al.
(2013) Development of a proactive care program (U-CARE) to preserve physical functioning of frail older people in primary care. J Nurs Scholarsh 45, 3, 230–237.
OpenUrl

[75] Bleijenberg N,

[76] ten Dam VH,

[77] Drubbel I,

[78] et al.

[79] 19.↵
Nederlandse Vereniging voor Geriatrie
(2020) Richtlijn Polyfarmacie bij ouderen (NVKG, Utrecht) [Guideline polypharmacy in the elderly].

[80] Nederlandse Vereniging voor Geriatrie

[81] 20.↵
Charlson ME,
Pompei P,
Ales KL,
MacKenzie CR
(1987) A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 40, 5, 373–383.
OpenUrl CrossRef PubMed

[82] Charlson ME,

[83] Pompei P,

[84] Ales KL,

[85] MacKenzie CR

[86] 21.↵
Von Korff M,
Wagner EH,
Saunders K
(1992) A chronic disease score from automated pharmacy data. J Clin Epidemiol 45, 2, 197–203.
OpenUrl CrossRef PubMed

[87] Von Korff M,

[88] Wagner EH,

[89] Saunders K

[90] 22.↵
Drubbel I,
de Wit NJ,
Bleijenberg N,
et al.
(2013) Prediction of adverse health outcomes in older people using a frailty index based on routine primary care data. J Gerontol A Biol Sci Med Sci 68, 3, 301–308.
OpenUrl CrossRef PubMed

[91] Drubbel I,

[92] de Wit NJ,

[93] Bleijenberg N,

[94] et al.

[95] 23.↵
Drubbel I,
Numans ME,
Kranenburg G,
et al.
(2014) Screening for frailty in primary care: a systematic review of the psychometric properties of the frailty index in community-dwelling older people. BMC Geriatr 14, 27.
OpenUrl CrossRef PubMed

[96] Drubbel I,

[97] Numans ME,

[98] Kranenburg G,

[99] et al.

[100] 24.↵
Song X,
Mitnitski A,
Rockwood K
(2010) Prevalence and 10-year outcomes of frailty in older adults in relation to deficit accumulation. J Am Geriatr Soc 58, 4, 681–687.
OpenUrl CrossRef PubMed

[101] Song X,

[102] Mitnitski A,

[103] Rockwood K

[104] 25.↵
Rothman KJ,
Greenland S,
Lash TL
(2008) Modern epidemiology (Lippincott Williams & Wilkins, Philadelphia, PA).

[105] Rothman KJ,

[106] Greenland S,

[107] Lash TL

[108] 26.↵
van Kempen JA,
Schers HJ,
Jacobs A,
et al.
(2013) Development of an instrument for the identification of frail older people as a target population for integrated care. Br J Gen Pract, DOI:https://doi.org/10.3399/bjgp13X664289.

[109] van Kempen JA,

[110] Schers HJ,

[111] Jacobs A,

[112] et al.

[113] 27.↵
Morley JE,
Vellas B,
van Kan GA,
et al.
(2013) Frailty consensus: a call to action. J Am Med Dir Assoc 14, 6, 392–397.
OpenUrl CrossRef PubMed

[114] Morley JE,

[115] Vellas B,

[116] van Kan GA,

[117] et al.

[118] 28.↵
Shi J,
Yang Z,
Song X,
et al.
(2014) Sex differences in the limit to deficit accumulation in late middle-aged and older Chinese people: results from the Beijing Longitudinal Study of Aging. J Gerontol A Biol Sci Med Sci 69, 6, 702–709.
OpenUrl CrossRef PubMed

[119] Shi J,

[120] Yang Z,

[121] Song X,

[122] et al.

Main menu

User menu

Search

Identification of frailty in primary care: accuracy of electronically derived measures

Abstract

INTRODUCTION

METHOD

Study population

Reference standard

Frailty measures

Statistical analysis

RESULTS

Frailty measures in total study population

Frailty measures in subgroups of age and sex

Sensitivity analyses

DISCUSSION

Summary

Strength and limitations

Comparison with existing literature

Implications for research and practice

Acknowledgments

Notes

Funding

Ethical approval

Data

Provenance

Competing interests

Discuss this article

REFERENCES

In this issue

Citation Manager Formats

Jump to section

Keywords

More in this TOC Section

Related Articles

Cited By...

NAVIGATE

RCGP

MY ACCOUNT

NEWS AND UPDATES

AUTHORS & REVIEWERS

CUSTOMER SERVICES

CONTRIBUTE

CONTACT US