jfa journal

AND option

OR option



A. Campeau Calfat, C. Sirois


Université Laval, Faculté de pharmacie, Québec, Canada

Corresponding Author: Alexandre Campeau Calfat, Université Laval, Faculté de pharmacie, Québec, Canada, alcac7@ulaval.ca

J Frailty Aging 2024;13(2)179-183
Published online February 21, 2024, http://dx.doi.org/10.14283/jfa.2024.18



Frailty is a state of increased vulnerability that can lead to premature death. While various clinical tools effectively measure frailty in individual care, their applicability at the population-level is limited. However, in the era of big-data, administrative databases serve as valuable sources for medication-based research and population surveillance. This narrative scoping review synthesizes the literature on tools used within administrative databases to detect frailty in community-dwelling older adults. The 17 identified publications explore four tools that meet the criteria of the Rockwood & Mitnitski frailty index model. Despite variations in the deficits they incorporate, all tools appear to be valuable for identifying frailty and predicting the risk of adverse events. Using those tools within administrative databases can be useful for research and surveillance purposes.

Key words: Frailty, administrative database, older adults.



In the era of big-data, administrative databases serve as valuable sources for medication-based research and population surveillance (1). Administrative databases provide information on many health determinants such as healthcare utilization, chronic disease diagnoses (often using the International classification of diseases (ICD)), sociodemographic information (age, sex, region of living, etc.), and medication claims (1). However, administrative databases often lack clinical information. Many individual characteristics and laboratory data are not included in claim databases or are not consistently recorded during medical visits or hospitalizations. Frailty, characterized by increased vulnerability, is an important determinant of medication use, outcomes, and healthcare utilization in older adults. However, this variable is seldom present in administrative data (2).
The definition of the concept of frailty is not unanimous (2–4). Frailty describes a dynamic state of vulnerability to environmental stressors, whether they are physical or psychological (5). While the aging process is currently irreversible, the state of frailty is modifiable (5, 6). Two prominent conceptual models allow for the operationalization of frailty (3–5, 7): 1) frailty defined by the accumulation of deficits (7); 2) frailty defined by a phenotype (4).
Rockwood & Mitnitski present frailty as an accumulation of deficits related to aging (7). The operationalization proceeds with criteria for defining and measuring deficits, with the idea that deficits lead to a loss of essential physiological reserves for an adequate response to stressors (7). The deficit accumulation model led to the creation of the Frailty Index (FI). This index allows for standardization of the definition of frailty in clinical settings (3, 7, 8). The variables considered as deficits and included in the index must be acquired, age-related, have a negative impact on health, and be uncommon (i.e., the proportion of individuals with the deficit should not be close to 100%) (7, 9, 10). Different frailty indices can be designed based on the context and population observed, as long as the deficits included meet the aforementioned criteria (7, 8, 11–14). The score is determined by dividing the number of deficits present by the number of deficits considered in the index (7, 8, 11). The FI appears to predict mortality better than an individual’s chronological age, demonstrating the importance of integrating frailty in research models (15, 16).
Unlike the FI, the frailty phenotype emphasizes the nature of deficits rather than their number. The phenotype model developed by Fried & al. describes frailty as a clinical syndrome characterized by strict criteria (3, 4). Frailty is present in an individual who exhibits at least three of the following criteria (4): 1) Involuntary weight loss >4.5kg or at least 5% of body weight in the last year; 2) Weak muscle strength, characterized by grip strength in the lowest quintile after adjusting for sex and body mass index; 3) Low endurance/energy, characterized by a state of exhaustion as measured by the Center for Epidemiologic Studies-Depression (CES-D) questionnaire; 4) Slow walking speed, characterized by the time required to walk 15 feet (4.5m) in the lowest quintile after adjusting for sex and height; 5) Low level of physical activity, characterized by energy expenditure in the lowest quintile after adjusting for sex. An individual with none of these criteria is considered robust, and one with one or two criteria is considered pre-frail (4).
The prevalence and incidence of frailty are increasing worldwide (17, 18). A systematic review, incorporating data up to 2020, revealed that the prevalence of frailty among community-dwelling individuals aged 50 and over was 24% and 12% using the accumulation deficit and the phenotype model, respectively (18). There is a pressing need for improvement in assessment of frailty (17, 18). While the two frailty models are relatively easy to implement in clinical settings, integrating them with administrative data proves more challenging. Indeed, administrative data lack much of the clinical information necessary to conduct a frailty evaluation. Nonetheless, given the high prevalence of frailty and its association with negative health outcomes, it would be relevant to conduct studies employing large available administrative databases to address this public health concern and provide population-based information. It is unclear how the frailty models can be operationalized in administrative databases based on the available information. The aim of this narrative review is to describe how administrative databases can be used for identifying frailty in community-dwelling older adults.



The authors (ACC, CS) devised the systematic search strategy, which was subsequently reviewed and approved by a specialized health sciences librarian (F. Bergeron). The PRISMA method was followed (Figure 1) (19). We searched Medline on March 5th, 2022 from its inception until March 2022 for articles discussing frailty in administrative databases. Key search terms included notably frail*, clinical data, health data, administrative data. The complete search strategy is described in Table A1.

Figure 1. PRISMA flowchart describing the inclusion process of publications


The PICO model (20) was employed to formulate the search question and, subsequently, to define inclusion and exclusion criteria. Specifically, the inclusion criteria were defined as follows: 1) Research articles identifying frailty using administrative databases. We included all types of research articles that explored various definitions of frailty, without limiting them to accepted conceptual models, to ensure a comprehensive evaluation. Administrative databases encompassed claims data, electronic medical records, or other large databases allowing frailty assessment; 2) Articles including older community-dwelling individuals. Our search was confined to the older population since frailty is predominantly studied in this demographic. Since the age of 65 is often the administrative threshold to define older age, we applied this age limit for inclusion. We limited the search to community settings, reflecting the predominant residence of the older population. Moreover, administrative databases are often more restricted in other settings (hospitals, long-term care), potentially impacting the tools developed. Additionally, individuals in such settings exhibit a higher degree of frailty, which would have introduced heterogeneity and complicated interpretation; 3) Articles written in English or French, available in full. Articles that could not be retrieved were excluded. Inclusion was not restricted based on specified outcomes. Meta-analyses and systematic reviews were considered for discussion purposes.
An Excel spreadsheet adapted to our objectives was used for data extraction and synthesis. It comprised title, authors, country, database, objective(s), tools, variables included in the tool. One reviewer (ACC) conducted the full-paper evaluation. In case of uncertainty, a second reviewer (CS) was invited to make a final decision. The completed PRISMA checklist is available as a supplementary file. No protocol for this study has been published.



From the initial 1,193 publications identified, 17 met the inclusion criteria (Table 1). We excluded five articles associated with non-community settings (hospital =4, long-term care=1). Included articles used data from the United States (n=8), United Kingdom (n=7), and Taiwan (n=2). The administrative databases used in included articles were electronic medical records (21–28), which contain information about health services provided to an individual, and third-party claims data (6, 13, 29–35), which contain billing information.

Table 1. Summary of publications discussing frailty in relation to administrative databases (n=17)


None of the included articles discussed the frailty phenotype. The four different tools retrieved, referred to as frailty index (FI), identified frailty based on the accumulation of deficits model: the Electronic Frailty Index (27), the Claims-based Frailty Index (13), the Multimorbidity Frailty Index (33), and the Veterans Health Administration Frailty Index (35). Table 2 shows the comparisons of the four tools.

Table 2. Characteristics of four frailty index that uses administrative databases



We identified four tools that use administrative databases to identify frailty. Those tools are based on Rockwood & Mitnitski’s FI model which seems best suited for integration with administrative data. Indeed, the frailty phenotype can rarely be identified by variables drawn from administrative databases since its characteristics have not yet been translated into specific diagnostic codes.
In line with our results, two systematic reviews attest the abundance of tools created to measure frailty in different contexts (14, 36). They indicate that these FIs vary depending on their context of use (e.g., community or hospital setting), the data source used (claims or electronic medical records), and the type of variables included (diagnoses, health services, pharmacy, others) (14, 36). All the identified FIs meet the criteria of the Rockwood & Mitnitski model which could explain why they all seem useful to monitor frailty and to predict the risk of adverse events despite their differences in included deficits (13, 26, 29, 33, 36). Searle & al. indicates that an FI developed according to Rockwood & Mitnitski’s model characteristics would share the same strengths and weaknesses whichever administrative database is used (11).
Most FIs perform better than comorbidity indices for predicting disability, mobility disorders, falls and mortality (14, 36). The Multimorbidity Frailty Index uses only diagnostic codes, making it less distinct from comorbidity indices (33). Indeed, the FIs that are constructed using diagnostic codes as well as codes for medical service utilization appear to be proxies for assessing functional capacity and physical performance (29, 36). The inclusion of these variables allows these FIs to better predict institutionalization and disability than comorbidity indices (13, 29). However, only the Claims-based Frailty Index has been directly compared to a comorbidity index (37).
The identified FIs have some limits. Electronic medical records, which contain administrative and clinical variables, are well-suited for the development of FIs (27). However, the lack of uniformity in record completion can affect their validity, particularly with regards to medication use (38). Additionally, sicker individuals, who may be more frail, appear to have more complete medical records, which could introduce bias in frailty assessment (38). Electronic medical records frequently include information that is unavailable in claims databases (e.g., smoking status, alcohol consumption), which enables the creation of FIs using a broader range of variables compared to claim codes (38). The use of electronic medical records is limited by their lack of availability for research in many jurisdictions (38). Claim data could be more accessible for research purposes, with the downside that they might not contain all meaningful clinical data. Also, the presence of a variable depends on its payment by a third-party payer which could also introduce bias in the identification of frailty. Variability in coding practices can introduce information bias which will influence the interpretation of the FI. Also, some diagnostic codes such as dementia or urinary incontinence may be less well assessed by clinicians, unlike myocardial infarction for example (38). One must be aware of the possible underestimation of the FI since the tool is generally not very sensitive (36).
A comparison among the various identified tools would provide additional insights, especially in determining which one is best suited for predicting events or adjusting in epidemiological research. Ideally, comparing the different tools within the same database would yield the most accurate answers. However, the availability of information in databases remains a limitation; tools can hardly be applied to all databases due to this constraint. Nevertheless, research could be conducted to identify which variables, within the same tool, have the most significant predictive capabilities. If adjustments need to be made to the tool based on available data, it would then be possible to roughly estimate the impact that the lack of information may have on the tool’s quality.
Our review has some limitations. First, we searched only one database and restricted our inquiry to two languages, which may result in potential omissions of relevant articles. Second, our review lacks a critical appraisal of the included articles. Third, broader inclusion criteria beyond community-dwelling individuals could have enhanced the representativeness of our review. Nevertheless, we believe our review effectively highlights the challenges associated with identifying frailty using administrative data. Given the increasing use of big data and machine learning, our review provides insights into potential solutions for incorporating frailty into epidemiological studies.



Our review indicates the presence of several tools for identifying frailty in administrative data. Grounded in the conceptual basis of the Rockwood & Mitnitski’s deficit accumulation model, these tools appear to be valid and robust FI measures. They can be employed to investigate frailty as an outcome or as a potentially confounding variable. However, operationalizing these tools remains challenging due to different limitations in the availability of variables within administrative databases. There is a pressing need to compare and contrast these different tools to gain a deeper understanding of the methodological and conceptual issues associated with measuring frailty in administrative data. This research will aid in tailoring the FI to individual databases.


Acknowledgements: We thank Frederic Bergeron for his help with the bibliographic search.

Conflicts of Interest: The authors have no conflict of interest. This work was supported by VITAM- Centre de recherche en santé durable [2003]; and the Fonds de Recherche du Québec en Santé [Junior 2 Salary Grant].

Ethical standards: No ethical approval was needed to conduct this review.





1. Johnson EK, Nelson CP. Values and pitfalls of the use of administrative databases for outcomes assessment. J Urol. 2013;190(1):17‑8. doi: 10.1016/j.juro.2013.04.048.
2. Rolfson D. Successful aging and frailty: A systematic review. Geriatrics (Basel). 15 nov 2018;3(4):79. doi: 10.3390/geriatrics3040079.
3. Rockwood K, Andrew M, Mitnitski A. A Comparison of two approaches to measuring frailty in elderly people. J Gerontol A Biol Sci Med Sci. 1 juill 2007;62(7):738‑43. doi: 10.1093/gerona/62.7.738
4. Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J, et al. Frailty in Older Adults: Evidence for a Phenotype. J Gerontol A Biol Sci Med Sci. 2001;56(3):M146‑56. doi: 10.1093/gerona/56.3.m146.
5. Morley JE, Vellas B, Abellan van Kan G, Anker SD, Bauer JM, Bernabei R, et al. Frailty consensus: A call to action. J Am Med Dir Assoc. 2013;14(6):392‑7. doi: 10.1016/j.jamda.2013.03.022.
6. Ward RE, Orkaby AR, Dumontier C, Charest B, Hawley CE, Yaksic E, et al. Trajectories of frailty in the 5 years prior to death among U.S. veterans born 1927–1934. J Gerontol A Biol Sci Med Sci, 2021;76(11):e347‑53. doi: 10.1093/gerona/glab196.
7. Rockwood K, Mitnitski A. Frailty defined by deficit accumulation and geriatric medicine defined by frailty. Clin Geriatr Med. 2011;27(1):17‑26. doi: 10.1016/j.cger.2010.08.008
8. Rockwood K, Mitnitski A. Frailty in relation to the accumulation of deficits. J Gerontol A Biol Sci Med Sci. 2007;62(7):722‑7. doi: 10.1093/gerona/62.7.722
9. Mitnitski A, Song X, Skoog I, Broe G, Cox JL, Grunfeld E, et al. Relative fitness and frailty of elderly men and women in developed countries and their relationship with mortality. J Am Geriatr Soc. 2005;53(12):2184‑9. doi: 10.1111/j.1532-5415.2005.00506.x
10. Blodgett J, Theou O, Kirkland S, Andreou P, Rockwood K. Frailty in NHANES: comparing the frailty index and phenotype. Arch Gerontol Geriatr. 2015;60(3):464‑70. doi: 10.1016/j.archger.2015.01.016
11. Searle SD, Mitnitski A, Gahbauer EA, Gill TM, Rockwood K. A standard procedure for creating a frailty index. BMC Geriatr. 2008;8(1):24. doi: doi: 10.1186/1471-2318-8-24
12. Rockwood K, Mitnitski A, Song X, Steen B, Skoog I. Long-term risks of death and institutionalization of elderly people in relation to deficit accumulation at age 70. J Am Geriatr Soc. 2006;54(6):975‑9. doi: 10.1111/j.1532-5415.2006.00738.x
13. Kim DH, Schneeweiss S, Glynn RJ, Lipsitz LA, Rockwood K, Avorn J. Measuring Frailty in Medicare Data: Development and Validation of a Claims-Based Frailty Index. J Gerontol A Biol Sci Med Sci. 2018;73(7):980‑7.
14. Lim A, Choi J, Ji H, Lee H. Frailty assessment using routine clinical data: An integrative review. Arch Gerontol Geriatr. 2022;99:104612. doi: 10.1093/gerona/glx229
15. Mitnitski AB, Mogilner AJ, Rockwood K. Accumulation of deficits as a proxy measure of aging. Sci World J. 2001;1:323‑36. doi: 10.1100/tsw.2001.58
16. Mitnitski AB, Song X, Rockwood K. The Estimation of relative fitness and frailty in community-dwelling older adults using self-report data. J Gerontol A Biol Sci Med Sci. 2004;59(6):M627‑32. doi: 10.1093/gerona/59.6.m627
17. Ofori-Asenso R, Chin KL, Mazidi M, Zomer E, Ilomaki J, Zullo AR, et al. Global Incidence of Frailty and Prefrailty Among Community-Dwelling Older Adults: A Systematic Review and Meta-analysis. JAMA Netw Open. 2019;2(8):e198398. doi: 10.1001/jamanetworkopen.2019.8398
18. O’Caoimh R, Sezgin D, O’Donovan MR, Molloy DW, Clegg A, Rockwood K, et al. Prevalence of frailty in 62 countries across the world: a systematic review and meta-analysis of population-level studies. Age Ageing. 2021;50(1):96‑104. doi: 10.1093/ageing/afaa219.
19. Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 372:n71. doi: 10.1136/bmj.n71.
20. Huang X, Lin J, Demner-Fushman D. Evaluation of PICO as a knowledge representation for clinical questions. AMIA Annu Symp Pro. 2006;2006:359‑63. PMID: 17238363
21. Hollinghurst J, Fry R, Akbari A, Clegg A, Lyons RA, Watkins A, et al. External validation of the electronic Frailty Index using the population of Wales within the Secure Anonymised Information Linkage Databank. Age Ageing. 2019;48(6):922‑6. doi: 10.1093/ageing/afz110
22. Pajewski NM, Lenoir K, Wells BJ, Williamson JD, Callahan KE. Frailty screening using the electronic health record within a medicare accountable care organization. Newman A, éditeur. J Gerontol A Biol Sci Med Sci 2019;74(11):1771‑7. doi: 10.1093/gerona/glz017.
23. Fogg C, Fraser SDS, Roderick P, de Lusignan S, Clegg A, Brailsford S, et al. The dynamics of frailty development and progression in older adults in primary care in England (2006–2017): a retrospective cohort profile. BMC Geriatr. 2022;22(1):30. doi: 10.1186/s12877-021-02684-y
24. Pradhananga S, Regmi K, Razzaq N, Ettefaghian A, Dey AB, Hewson D. Ethnic differences in the prevalence of frailty in the United Kingdom assessed using the electronic Frailty Index. AGING Med. 2019;2(3):168‑73. doi: 10.1002/agm2.12083
25. Stow D, Matthews FE, Hanratty B. Frailty trajectories to identify end of life: a longitudinal population-based study. BMC Med. 2018;16(1):171. doi: 10.1186/s12916-018-1148-x
26. Stow D, Matthews FE, Barclay S, Iliffe S, Clegg A, De Biase S, et al. Evaluating frailty scores to predict mortality in older adults using data from population based electronic health records: case control study. Age Ageing. 2018;47(4):564‑9. doi: 10.1093/ageing/afy022
27. Clegg A, Bates C, Young J, Ryan R, Nichols L, Ann Teale E, et al. Development and validation of an electronic frailty index using routine primary care electronic health record data. Age Ageing. 2016;45(3):353‑60.
28. Boyd PJ, Nevard M, Ford JA, Khondoker M, Cross JL, Fox C. The electronic frailty index as an indicator of community healthcare service utilisation in the older population. Age Ageing. 2019;48(2):273‑7. doi: 10.1093/ageing/afx001
29. Kim DH, Glynn RJ, Avorn J, Lipsitz LA, Rockwood K, Pawar A, et al. Validation of a claims-based frailty index against physical performance and adverse health outcomes in the health and retirement study. J Gerontol A Biol Sci Med Sci, 2019;74(8):1271‑6. doi: 10.1093/gerona/gly197
30. Johnston KJ, Wen H, Joynt Maddox KE. Relationship of a claims-based frailty index to annualized medicare costs: A cohort study. Ann Intern Med. 2020;172(8):533-540. doi: 10.7326/M19-3261
31. Cuthbertson CC, Kucharska-Newton A, Faurot KR, Stürmer T, Jonsson Funk M, Palta P, et al. Controlling for frailty in pharmacoepidemiologic studies of older adults: Validation of an existing medicare claims-based algorithm. Epidemiol 2018;29(4):556‑61. doi: 10.1097/EDE.0000000000000833
32. Festa N, Shi SM, Kim DH. Accuracy of diagnosis and health service codes in identifying frailty in Medicare data. BMC Geriatr. 2020;20(1):329. doi: 10.1186/s12877-020-01739-w
33. Wen YC, Chen LK, Hsiao FY. Predicting mortality and hospitalization of older adults by the multimorbidity frailty index. PloS One. 2017;12(11):e0187825. doi: 10.1371/journal.pone.0187825.
34. Lai HY, Huang ST, Chen LK, Hsiao FY. Development of frailty index using ICD-10 codes to predict mortality and rehospitalization of older adults: An update of the multimorbidity frailty index. Arch Gerontol Geriatr. 2022;100:104646. doi: 10.1016/j.archger.2022.104646
35. Cheng D, DuMontier C, Yildirim C, Charest B, Hawley CE, Zhuo M, et al. Updating and validating the U.S. veterans affairs frailty index: Transitioning from ICD-9 to ICD-10. Newman AB, éditeur. J Gerontol A Biol Sci Med Sci. 2021;76(7):1318‑25. doi: 10.1093/gerona/glab071.
36. Kim DH. Measuring frailty in health care databases for clinical care and research. Ann Geriatr Med Res. 2020;24(2):62‑74. doi: 10.4235/agmr.20.0002.
37. Gagne JJ, Glynn RJ, Avorn J, Levin R, Schneeweiss S. A combined comorbidity score predicted mortality in elderly patients better than existing scores. J Clin Epidemiol. 2011;64(7):749‑59. doi: 10.1016/j.jclinepi.2010.10.004.
38. Eriksson I, Ibáñez L. Secondary data sources for drug utilization research. In: Elseviers M, Wettermark B, Almarsdóttir AB, Andersen M, Benko R, Bennie M, et al., éditeurs. Drug utilization research [Internet]. Chichester, UK: John Wiley & Sons, Ltd; 2016, p. 39‑48. Available: https://onlinelibrary.wiley.com/doi/10.1002/9781118949740.ch4

© Serdi 2024