Medication adherence tools and measures in chronic kidney disease: a systematic review

Abstract

Background

There is no gold standard for assessment of medication adherence. This study aimed to systematically review the literature to identify validated medication adherence measurement tools and methods used in clinical practice and research settings in the context of patients with chronic kidney disease (CKD) and to synthesize key features of the identified medication adherence tools.

Methods

We systematically reviewed MEDLINE via Ovid, Embase via Ovid, Cochrane Central Register of Controlled Trials (CENTRAL), and CINAHL (via EBSCO) from inception to September 25, 2025. All abstracts were screened by pairs of reviewers independently, followed by a full-text review identifying the tool/method used for measuring medication adherence. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed to conduct this systematic review. General study and medication adherence method/tool-specific characteristics were summarized. The quality criteria for measurement properties were applied across the included studies to synthesize and assess the strength of the evidence.

Results

The 43 included articles originated from 25 countries. The most common measures used for evaluating medication adherence were the Eight-Item Morisky Medication Adherence Scale (MMAS-8) (n = 11 [25.6%]), Medication Possession Ratio (MPR) (n = 10 [23.3%]), Proportion of Days Covered (PDC) (n = 8 [18.6%]), and Medication Events Monitoring System (MEMS) (n = 6 [14.0%]). Five (15.6%) studies used multiple methods to measure medication adherence.

Conclusion

No accepted reference tool is available to measure CKD patients’ medication adherence. Some tools, however, were used more frequently in the context of patients with CKD. Choosing an appropriate method/tool or a combination of methods depends on the clinician/researcher’s goals, study setting, availability of data and other resources, and patients’ characteristics.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12882-025-04734-8.

Introduction

Chronic kidney disease (CKD), with a prevalence of 10% (9.1% to 13.4%) is a frequently treated condition within health care systems, both universally and within Canada [1]. Globally, the number of patients with all CKD categories reached approximately 850 million [2], which is reported to be more than those with diabetes, chronic obstructive pulmonary disease, asthma, osteoarthritis, or even depressive disorders [3]. Recent data suggests that one in ten adults in Canada likely have CKD [1].

Poor medication adherence is a significant issue across healthcare systems, affecting individual’s health outcomes and the performance of healthcare systems [4]. In patients with CKD, medication adherence is a key component of effective disease management [5, 6]. The main goal of pharmacotherapy in CKD is to slow disease progression and manage disease-associated complications and comorbidities while treating the underlying etiology [6]. Left unmanaged or poorly managed, CKD can lead to kidney failure. Patients with CKD are on multiple pharmacological treatments often starting with hypoglycemic and anti-hypertensive agents and depending on the CKD stage, phosphate binders, calcimimetics, vitamin D preparations, erythropoiesis-stimulating agents, anti-nauseants, and iron supplements might be added [7]. The drug classes may require multiple doses per day, therefore leading to a very high pill burden. Thus, CKD patients are on 10 drugs on average [7, 8] and reports include counts as high as >20 pills/day [5]. It has been estimated that as much as 50% of patients across the globe are non-adherent to their prescribed medication [9]. There is a growing body of evidence on the prevalence of medication adherence/non-adherence among patients with CKD [10]. Medication non-adherence was reported to vary from 17% to 74% among CKD patients [5, 6] and from 3% to 80% among hemodialysis patients [11] based on the tools used to measure adherence.

No study has evaluated the rate of adherence/non-adherence to multiple medications in the Canadian CKD, non-dialysis population. According to a study that assessed compliance with newly initiated anti-hypertensives in CKD patients in Canada, a year after initiating anti-hypertensives, almost one-third of CKD patients were either not taking the drug (no persistence) or had not been compliant [12]. The high rate of non-adherence in CKD patients presents a significant barrier to adequate disease management, thereby increasing the risk of disease progression, hospitalization, and mortality [13]. It is crucial to properly understand and measure patients’ medication adherence levels due to adverse health effects [14] and the economic burden of non-adherence to pharmacotherapy in CKD patients [15]. In fact, optimizing medication adherence is a priority for healthcare providers worldwide [10, 16].

Despite the decades of research on medication adherence, no widely accepted standard or method for assessing adherence to prescribed medication exists [17–20]. Consequently, the true extent of adherence among CKD patients remains uncertain. The variability in study results is partly due to differing measurement tools and the absence of standardized methods tailored to specific patient populations or clinical settings. Furthermore, there are methodological differences in validating self-reported adherence tools, highlighting the need to examine the validity of these tools for CKD patients. In general, clinician recognition of medication non-adherence is also low [19] and in order to develop interventions to improve medication adherence, it’s first crucial to have a sensitive and precise adherence measurement tool and understand the extent of medication adherence in patients with CKD. In sum, there is a need for more structured studies to guide clinicians and researchers in distinguishing which tool might have the most relevant features and consequently might be most useful and valuable in their clinical/research setting and patient population.

The present study aimed to address the gap in this area to both raise awareness of the issues and to facilitate work to develop and implement a standard medication adherence measurement tool specific to patients with CKD. Ultimately, the goal is to establish a set of tools appropriate for both research and clinical care that may aid clinicians and policymakers in adopting these tools into daily practice and electronic health records.

This systematic review aimed to 1) Identify validated medication adherence measurement tools/methods used in clinical practice and research in the context of patients with CKD. 2) Identify and synthesize key features of medication adherence methods and validated tools.

Methods

A systematic review protocol was developed based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [21]. The protocol included instructions for search strategy, database selection, inclusion and exclusion criteria, data extraction, risk of bias assessment, and data synthesis. The protocol of the study is available in PROSPERO with the registration number CRD42023469602 and can be accessed via the link below: https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42023469602

PRISMA guidelines were followed to conduct and report this systematic review.

Search strategy

To determine a candidate list of medication adherence measurement methods and tools applicable to CKD patients, a systematic literature search was conducted to identify reports of such measurement methods and tools. The following databases were searched for relevant English language studies from inception to September 25, 2025: MEDLINE via Ovid, Embase via Ovid, Cochrane Central Register of Controlled Trials (CENTRAL), and CINAHL (via EBSCO). A medical librarian was consulted as an advisor. Multiple comprehensive search strategies were used to avoid bias resulting from narrow searches. The search strategy used both Medical Subject Headings (MeSH) terms and keywords to create search strings for medication adherence, chronic kidney disease, and measurement methods and tools. Search strategies in all databases are presented in Appendix 1 in Supplemental Materials.

Study selection criteria

Studies reporting medication adherence/non-adherence using subjective and/or objective tools in CKD patients aged >14 years were considered for inclusion. Eligible studies must have been reported as original articles and full-text publications in English language. Review papers, editorials, letters to the editor, non-peer-reviewed articles and abstract-only publications such as conference proceedings were excluded. Regarding study design, randomized controlled trials, observational studies (including cohort and cross-sectional), and other experimental studies investigating medication adherence were included. Studies that reported an algorithm for testing or validating adherence measurement tools were not considered. If studies had a sample size < 25 or did not mention the evaluation method for medication adherence, they were deemed ineligible. Validity measures (e.g. correlation, kappa, sensitivity) were required to be mentioned for subjective tools (e.g. self-report/clinician-administered questionnaires).”

Studies evaluating adherence to guidelines or non-pharmacological treatments (e.g., diet, exercise, hemodialysis session, etc.) were also excluded. Additionally, the tool was excluded if researchers couldn’t locate the complete tool published in the peer-reviewed literature.

Selection of studies

Search results were imported to Covidence [22], a screening and data extraction software for systematic reviews. After removing the duplicates, the studies were divided evenly among five pairs of reviewers (renal pharmacists) to complete the screening process. Each article was reviewed by two reviewers independently. Before beginning the review, we piloted reviewing 100 papers to guarantee an inter-reviewer agreement. During the first screening phase, each reviewer independently assessed the content from titles and abstracts against the common eligibility and pre-defined selection criteria to decide whether the study qualifies to go to the following screening phase (i.e., full-text review). During the full-text review, each reviewer independently evaluated the full-texts against the inclusion criteria, extracting pre-determined data elements. At all stages, the primary researcher resolved any discrepancies via discussion between the reviewers and the primary researcher. The selection process was documented with a PRISMA flowchart (Fig. 1).

Fig. 1.

Preferred reporting items for systematic reviews and meta-analyses flow diagram

Data extraction

A data extraction template was developed using Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia). The primary researcher piloted the template on four studies from each adherence measure type and finalized the template thereafter. The included studies were then screened, and data extraction was done by three pairs of reviewers independently. The data elements in the finalized standardized template included general study and medication adherence method/tool-specific characteristics. Data were extracted on the general study characteristics as follows: Corresponding author, affiliation, email, address, source of funding, publication year, country of study, country income status (low-income, lower middle-income, upper middle-income, high-income) as defined by The World Bank Atlas method [23], study setting, study design, inclusion criteria, exclusion criteria, total sample size, mean age of the participants, sex, and whether the population included dialysis patients or not. The subjective tool-specific characteristics included the following: Tool’s name/version, definition, number of questions/items, type of question/items (e.g. Likert scale, yes/no, etc.), scoring system, scoring interpretation method, adherence threshold (whether or not threshold was specified), serial/individual assessment, recall period, original language, whether barriers/facilitators of adherence were included in the tool, mode of administration (self/clinician-administered), and psychometric properties of the tool including reliability and validity. The objective method-specific characteristics included the following: Name of the method, description (calculation methodology), serial/individual assessment, and threshold (whether or not threshold was specified).

Quality assessment of the included studies

QUADAS-2 tool was used to assess the studies’ quality and appraise the risk of bias [24]. This tool is recommended for use in systematic reviews of diagnostic accuracy by the Agency for Healthcare Research and Quality, Cochrane Collaboration, and the UK National Institute for Health and Clinical Excellence [24]. This risk of bias assessment tool comprises four domains: patient selection, index test, reference standard, and flow and timing. It includes 14 items assessing the risk of bias, sources of variation (applicability), and reporting quality with each item being rated as “yes,” “no,” or “unclear” [24].

The quality assessment template was prepared in Covidence. A risk of bias assessment was completed for each included study independently by three pairs of reviewers and in duplicate. Discrepancies were resolved by the primary researcher and by consultation with the project lead if needed. The final quality assessment template is presented in Appendix 2.

Data analysis and synthesis

Descriptive statistics were reported on general study characteristics, including study design, country of origin, sample size, age group, specific medication adherence measurement methods and tools. In addition, different parameters of the included medication adherence measurement tools and methods (e.g. psychometric properties of questionnaire tools, scoring methods, thresholds, etc.) were analyzed. The quality criteria for measurement properties were applied across the included studies to synthesize and assess the strength of the evidence.

Results

Literature search and screening results

The literature search from January 1, 1946, to September 25^th, 2025, resulted in a total of 64,156 records across the four databases. The PRISMA flow diagram in Figure 1 outlines the review and screening process. 12,667 studies were identified from Ovid MEDLINE, 39,570 studies were identified from Ovid EMBASE, 6,816 studies were identified from CINAHL via EBSCO, and 5,103 studies were identified from CENTRAL. 38,078 of these studies were removed as duplicates, leaving 26,078 remaining for initial title/abstract screening. Following the title/abstract review, 932 studies were potentially eligible for full-text review. After reviewing the relevant full texts, 889 articles were excluded, mainly because no full text was available (supplement issue) or a wrong tool type was included. Thus, 43 studies met the criteria to be included in the review and data were extracted from these remaining articles (Fig. 1).

Study characteristics

The general characteristics of the included articles are presented in Table 1. The first article was published in 1996, and almost 81% of the studies were published in the last 10 years (2015 or later). Nine articles were published between 2004 and 2015; no other article was published in the 1990s or before. The articles originated from 25 countries that were grouped geographically into seven regions, with most of the studies conducted in North America and Europe (n = 21). Eight studies were conducted in Asia, six in the Middle East, two studies in Africa, two in Australia and one study was conducted in the UK. Regarding country income status, 64.4% (n = 29) of the studies were conducted in high-income countries, with the remaining sixteen studies conducted in upper-middle-income (n = 9), lower-middle-income (n = 6) and low-income (n = 1) countries (as defined by The World Bank Atlas method at the time of data extraction) [23]. Most articles were observational studies (n = 40, 88.9%), including 21 retrospective cohort studies, 15 cross-sectional studies, and 4 prospective cohort studies. The remaining three studies were a randomized controlled trial, a cluster randomized controlled trial and a pre-post trial. The median sample size was 220 (2110) participants, ranging from 29 to 31,688 participants, with the age range varying between 14 and 88 years old. There were few studies conducted on children and adolescents (n = 3, 7%), with the majority of the studies having adult participants (n = 40, 93%). Of the included studies (n = 43), thirteen studies were conducted on patients on dialysis (hemodialysis (n = 10) and peritoneal dialysis (n = 3)). Other studies were conducted on CKD patients, or they didn’t mention if dialysis patients were included in the study population.

Table 1.

General characteristics of the studies included in the systematic review (n = 43)

Characteristic	No. of studies, n (%)
Types of studies
Observational	40 (88.9%)
Retrospective Cohort	21 (46.7%)
Prospective Cohort	4 (8.9%)
Cross Sectional	15 (33.3%)
Randomized controlled trial	3 (6.7%)
Cluster randomized controlled trial	1 (2.2%)
Pre-post study	1 (2.2%)
Geography
North America	12 (26.7%)
Europe	9 (20.0%)
United Kingdom	1 (2.2%)
Australia	2 (4.4%)
Asia	8 (17.8%)
Middle East	6 (13.3%)
Africa	2 (4.4%)
Sample Size (CKD patients)
<100	11 (25.6%)
100–999	16 (37.2%)
999–4999	7 (16.3%)
>5000	8 (18.6%)
Not mentioned	1 (2.3%)
Age Groups
Mean age < 18 y	3 (7.0%)
Mean age < 65 y	22 (51.2%)
Mean age > 65 y	15 (34.9%)
Not reported	3 (7.0%)
Adherence Measurement Methods*
Administrative datasets	19 (44.2%)
Self-report questionnaires	21 (48.8%)
Electronic Medication Monitoring	6 (14.0%)
Prescription record review	2 (4.7%)
Pill counting	1 (2.3%)
Patient appointment records	2 (4.7%)
Multiple Measurement Methods	5 (11.6%)
No known tool	1 (2.3%)

*Some studies included more than one adherence measurement methods

Study-specific details of the included studies are presented in Table 2. The details include country, sample size, setting, data collection method and specific methods/tools used to calculate medication adherence.

Table 2.

Study-specific characteristics of the studies included in the systematic review

Study	Country	Sample Size	Population	Setting	Data Collection methods	Medication adherence Measurement Methods
Ada and Ozcan [25]	Turkey	192	HD + PD	Dialysis department of a tertiary care hospital	Self-report questionnaire	MMAS-8 (Turkish version)
Afshari et al. [26]	Iran	405	Non-dialysis CKD + HTN	Health Service Centers located in suburban area	Self-report questionnaire	MMAS-8
Alkatheri et al. [27]	Saudi Arabia	89	HD	HD units in hospital	Self-report questionnaire	MMAS-8 (Arabic version)
Amado et al. [28]	Portugal	122	HD	Patients undergoing OL-HDF	Self-report questionnaire	MTA (adapted for Portugal)
Bandiera et al. [29]	Switzerland	73	DKD	Primary care clinic of university hospital	Electronic Medication Monitoring	MEMS
Bandiera et al. [30]	Switzerland	31	DKD	Databases of several prospective studies and cohorts from the IMAP	Electronic Medication Monitoring	MEMS
Bao et al. [31]	China	29	PD	Patients switching from rhEPO to roxadustat in dialysis unit of hospital	Self-report questionnaire	MMAS-8
Bonikowska et al. [32]	Poland	35	T2DM + CKD	Primary healthcare centers	Self-report questionnaire	ACDS
Chang et al. [33]	Taiwan	1,695	T2DM + ESKD	Taiwan National Health Insurance Research Database (NHIRD)	Administrative datasets	MPR
Cohen-Glickman et al. [34]	Israel	75	HD	Outpatient HD units	Prescription record review	Adherence Ratio
Cooke and Fatodu [35]	USA	1,698	T2DM + HTN and/or kidney disease	Medical claims from a Medicaid managed care organization in Maryland	1) Administrative datasets	1) MPR
					2) Administrative datasets	2) Median gap between prescription refills
Gamal et al. [36]	Egypt	88	Non-dialysis CKD	A private nephrology clinic	Self-report questionnaire	SMAQ
Gincherman et al. [37]	USA	101	HD	Patients in Missouri Kidney Program	Administrative datasets	MPR
Gor et al. [38]	USA	9,019	T2DM + non-dialysis CKD who initiated with a DPP-4 Inh or pioglitazone	Truven MarketScan administrative claims databases	Administrative datasets	PDC
Hamza et al. [39]	Pakistan	390	HD	Hemodialysis units at multisite hospitals	Self-report questionnaire	MMAS-8
Hsu et al. [8]	Taiwan	1,271	Childhood CKD	Taiwan national health insurance research database	Administrative datasets	PDC
Janse et al. [40]	Sweden	31,688	CKD (7% were dialysis patients) + HF	Swedish Heart Failure Registry (SwedeHF)	Administrative datasets	PDC
Kanakubo1 et al. [41]	Japan	483	HD	Six Outpatient Dialysis Centers	Self-report questionnaire	ASK-12 Scale (Japanese)
Kaul et al. [42]	India	120	Non-dialysis CKD + HD	Outpatients	Self-report questionnaire	MMAS-8
Khokhar et al. [43]	Pakistan	120	CKD (Non dialysis)	Nephrology outpatient department of tertiary healthcare hospital	Self-report questionnaire	9-item modified Morisky medication adherence questionnaire
Kronish et al. [44]	USA	30	CKD + HTN	2 hospital based primary care practices in New York City	Electronic Medication Monitoring	MEMS
Lassen et al. [45]	Denmark	2,199	T2DM + CKD	Danish national health registries	Administrative datasets	MPR
Lee et al. [46]	USA	91	CKD (25 < eGFR < 70 mL/min/1.73 rn) + HTN	African American Study of Kidney Disease and Hypertension Pilot Study (AASK)	1) Electronic Medication Monitoring	MEMS
					2) Pill counting	Pill Counting
Ling et al. [47]	China	226	PD	Outpatient follow-up clinic	Self-report questionnaire	MMAS-8
Li et al. [48]	China	2,445*	HTN + Multimorbidity (including CKD)	General outpatient clinics	Self-report questionnaire	MMAS-8
Machnicki et al. [49]	USA	11,105*	HTN + Multimorbidity (including CKD)	Patients covered by commercial and Medicare Supplemental insurance in the Truven MarketScan database	1) Administrative datasets	1) PDC
					2) Administrative datasets	2) MPR
Mailani et al. [50]	Indonesia	164	HD	HD units in two tertiary hospitals	Self-report questionnaire	MMAS-8
Mohammadnezhad et al. [51]	Iran	100	CKD stage 3–5 (GFR < 60 mL/min/1.73 m2)	Outpatient nephrology clinic in a hospital	Self-report questionnaire	MMAS-8
Muntner et al. [52]	USA	3,936	CKD + HTN	Dataset from Reasons for Geographic and Racial Differences in Stroke (REGARDS) study	Self-report questionnaire	MMAS-4
Okoro et al. [53]	Nigeria	220	Non-dialysis patients with CKD stages 1–4	Medical and nephrology outpatient clinics	Self-report questionnaire	MGL MAQ (MMAS-4)
Park et al. [54]	USA	11,732	HD	Dataset from US Renal Data System (USRDS)	Administrative datasets	MPR
Pruette et al. [55]	USA	87	Adolescents and young adults with stage 1–5 CKD/ESRD + HTN	3 academic medical centers in the Mid- Atlantic region of the USA.	1) Electronic Medication Monitoring	1) MEMS
					2) Administrative datasets	2) MPR
					3) Medication records (Reference Standard)	3) Medication records (Reference Standard)
					4) Self-report questionnaire	4) MMAS-8
					5) Patient appointment records	5) Provider reported adherence
Qin et al. [56]	Australia	1,496	HF + CKD	Hospital Morbidity Data Collection (HMDC) from the WA Data Linkage System and Pharmaceutical Benefits Scheme (PBS) claims data	1) Administrative datasets	1) MPR
					2) Administrative datasets	2) Modified MPR
					3) Administrative datasets	3) PDC
Roggeri et al. [57]	Italy	994	Dialysis patients with secondary hyperparathyroidism	Administrative database of the Lombardy region	No known tool	% of treatable days out of the total number of days of treatment with the drug
Santoro et al. [58]	Italy	4,451	CKD patients on RAASi	Administrative and laboratory databases of five Local Health Units	Administrative datasets	PDC
Schmitt et al. [59]	USA	7,227	CKD patients being on at least one antihypertensive medication	Ambulatory population at Veteran Affairs Medical Centre (Veteran Affairs CKD cohort)	Administrative datasets	MPR
Shayakul et al. [60]	Thailand	220	Non-dialysis DKD	Outpatients’ clinics of Hospital (university-based tertiary care center)	Self-report questionnaire	MTB-Thai
Tang et al. [61]	UK	20,967	DKD on metformin	Clinical Practice Research Datalink (CPRD) database	Administrative datasets	PDC
Tesfaye et al. [62]	Australia	101	Pre-dialysis CKD (eGFR < 30 mL/min/1.73 m2)	Dataset from the Tasmanian CKD study	1) Self-report questionnaire	1) MGL MAQ (MMAS-4)
					2) Self-report questionnaire	2) TABS
Tilea et al. [63]	Romania	525*	HTN + multimorbidity (including CKD)	Outpatient setting (family medicine clinic)	Prescription records review	Prescription records review
Truong et al. [12]	Canada	7,119	CKD	Quebec Health Insurance Board (RAMQ) database	Administrative datasets	PDC
Van Camp et al. [64]	Belgium	135	HD	Multi-center randomized clinical trial to enhance adherence	Electronic Medication Monitoring	MEMS
Vasylyeva et al. [65]	USA	34	Pediatric CKD	Texas Tech University Health Sciences Center (TTUHSC) pediatric nephrology clinic	Self-report questionnaire	CAAMQ

HD, hemodialysis; PD, Peritoneal Dialysis; MMAS-8, Morisky Medication Adherence Scale- 8 item; OL-HDF, online-haemodiafiltration; MTA, Measure Treatment Adherence; DKD, diabetic kidney disease; IMAP, interprofessional medication adherence program; MEMS, Medication Event Monitoring System; rhEPO, recombinant human erythropoietin; T2DM: Type 2 diabetes; SMAQ, Simplified Medication Adherence Questionnaire; ASK-12 Scale, 12-item Adherence Starts Knowledge; ACDS, Adherence in chronic disease scale questionnaire; MPR, Medication Possession Ratio; PDC, Proportion of Days Covered; HTN, Hypertension; CKD, Chronic Kidney Disease; DPP-4 Inh, Dipeptidyl peptidase-4 Inhibitor; HF, Heart Failure; eGFR, estimated glomerular filtration rate; MMAS-4, 4-item Morisky’s scale; MGL MAQ, Morisky, Green, Levine Medication Assessment Questionnaire; ESRD, end-stage renal disease; RAASi, renin-angiotensin-aldosterone-system inhibitors; MTB-Thai, Medication Taking Behavior in Thai; TABS, Tool for Adherence Behavior Screening; CAAMQ, The Child & Adolescent Adherence to Medication Questionnaire

*Entire patient population (not just CKD patients)

Medication adherence measurement methods and tools

Of 43 qualifying articles, 6 (14.0%) studies used multiple methods to measure medication adherence. The most common measures used for assessing medication adherence were MMAS-8 (n = 11[25.6%], MPR (including MPR modified) (n = 10 [23.3%]), PDC (n = 8 [18.6%]), and MEMS (n = 6 [14.0%]).

Subjective tools (self-reported adherence measures)

Tool-specific characteristics of the subjective tools are presented in Table 3. The number of questions/items in each subjective tool ranged from 4 [4-item Morisky’s scale; (MMAS-4)] to 19 [The Child & Adolescent Adherence to Medication Questionnaire (CAAMQ)]. On average, the tools had eight items, with response types varying: four tools were Likert scales [28, 41, 60, 62], two tools were yes/no answers [43, 52, 53, 62], one tool had a combination of yes/no and Likert scale [25–27, 31, 33, 39, 42, 47, 50, 51, 55], one tool required close-ended and open-ended answers [65], one tool had five answers to choose from [32], one tool had both yes/no and open ended answers [36], one tool consisted of open-ended answers, yes/no answer, Likert scale and rating as a percentage [43].

Table 3.

Tool-specific characteristics of the subjective tools

Tool Name	Items/questions count	Answer type	Scoring	Adherence threshold	Recall period	Language	Psychometric properties	Barriers of adherence explored	Self- vs. clinician administered
Adherence in chronic disease scale questionnaire (ACDS) [32]	7	5 possible answers to choose from	0–28 (each answer: 0–4 points)	3 levels of adherence: ≥26: high 21–26: medium 0–20: non-adherence	Not specified	English	-) Acceptable Cronbach’s alpha -) Validated in CVD diseases.	Yes	Self-administered
Simplified Medication Adherence Questionnaire (SMAQ) [36]	6	4: Yes/No response 2: Quantitative/frequency type	Not specified	Nonadherent: -) A positive response to any of the first four qualitative questions -) A response of more than two doses missed over the past week to question 5 -) A response of not taking any prescribed medication over 2 days during the past month to question 6	-) 1 week -) 3 months	Arabic	Acceptable Cronbach’s alpha	Yes	Self-administered
Measure Treatment Adherence (MTA) [28]	7	6-point Likert scale	Not specified	≥5: Adherent  < 5: Non-adherent	Not specified	Portuguese	-) Cronbach’s alpha: 0.74 -) Acceptable concurrent validity	Yes	Self-administered
Morisky Medication Adherence Scale, 4 items (MMAS-4) Or Morisky, Green, Levine Medication Assessment Questionnaire (MGL MAQ) [52, 53, 62]	4	Yes/No	Yes: 1 No: 0	3 levels of adherence: 0: high adherence 1–2: medium adherence 3–4: low adherence	Not specified	English	Acceptable reliability and validity	Yes	Self-administered
Morisky Medication Adherence Scale, 8 items (MMAS-8) [25–27, 31, 33, 39, 42, 47, 50, 51, 55]	8	7: Yes/no 1: 5-pt Likert	0–8	8 = high adherence 6 - 7 = medium adherence  < 6 = low adherence	2 weeks	English Arabic Chinese	Acceptable reliability and validity	Yes	Self-administered
Adherence Starts with Knowledge‑12 scale (ASK-12) [41]	12	5-point Likert	12–60 (items 4–7: reverse scored)	Total score cut-off of 23 discriminates non-adherence (refill rate < 80%) vs adherent (refill adherence ≥80%)	-) 1 week -) 1 month	Japanese	-) Cronbach’s alpha: 0.75 -) Acceptable Validity	Yes	Self-administered
9-item modified Morisky medication adherence questionnaire [43]	9	3: Yes/No 4: open ended 1: Likert 1: rating as%	Each correct answer: 1 mark	3 levels of adherence: 8–9: high (If most answers are no) 6–7: medium 0–5: low	3 months	English	Acceptable reliability and validity	Yes	Self-administered
Medication Taking Behavior in Thai (MTB Thai) [60]	6	4-point Likert	0–24	3 levels of adherence: 24: high 22–23: medium  < 22: low	2 weeks	Thai	-) Cronbach’s alpha: 0.76 -) Acceptable validity	Yes	Self-administered
The Child & Adolescent Adherence to Medication Questionnaire (CAAMQ) [65]	19	9 close-ended & 7 open-ended & 3 demographics questions.	Not specified	Not specified	Not specified	English	-) Reliability: not mentioned. -) Acceptable face, construct and content validity	Yes	Clinician- administered
Tool for Adherence Behavior Screening (TABS) [62]	8 Items (2 subcomponents each 4 items)	Likert Scale	differential score	Total for ‘adherence’ minus total for ‘nonadherence’ of ≥ 15 reflects good adherence and of ≤ 14 indicates suboptimal adherence	Not specified	English	Acceptable reliability and validity	Yes	Clinician-administered

The tools’ scoring and scoring interpretations are described as a measure of the actionability of the tool in practice and vary across different tools (Table 3). Regarding the threshold for adherence/non-adherence, nine tools (90.0%) specified a threshold for adherence or different levels of adherence (high, medium and low adherence) based on the scores. One tool (10.0%) didn’t mention any threshold for identifying adherent and non-adherent patients [62].

In assessing patient recall of adherence within a specified time period, five tools (50%) did not specify a recall period; the remaining five (50%) had recall periods ranging from a week to three months. In terms of the method of administration used in the included studies, eight tools were described as self-administered by the patients [27, 28, 31–33, 36, 41, 43, 52, 53, 55, 60, 62]; whereas two were clinician-administered [62, 65].

All the included subjective tools explored barriers to adherence, such as perceived concerns, forgetfulness, costs/affordability, feeling better/worse, and worries about taking medications long-term as well as the extent of taking medication prescribed. Five tools were available in languages other than English [27, 28, 36, 41, 48, 60], while five others were available in English [31–33, 43, 52, 53, 55, 62, 65].

In terms of psychometric properties, only one study reported validity measures based on data obtained from MEMS (which was used as a reference standard) [55]. Others cited the articles that documented the psychometric properties of that specific tool. All tools have been reported to have acceptable reliability and validity measures except one for which no reliability measure was found [65].

Objective tools

One third of the studies (n = 15 [34.9%]) used administrative datasets as a type of objective measure to assess medication adherence. MPR and PDC were reported to be the most commonly used measures among objective tools. Pill counting and the median gap between the prescription refills were among the least frequently used measures. However, formulas for derivations of MPR and PDC varied between studies. Regarding the adherence threshold or the identification of different levels of adherence, some variations were reported for MPR, PDC, and MEMS across the included studies. Specific characteristics of the objective measures are presented in Table 4.

Table 4.

Specific characteristics of the objective measures

Tool Name	Study	Calculation Method/Tool Definition	Adherence Threshold
MPR	Chang et al. [33]	The total number of days for which medication was prescribed were divided by the number of days in a year (365).	≥80% = good adherence
MPR	Cooke & Fatodu [35]	The total days supply for all medication were divided by the total possible days supply from the date when the Rx was originally dispensed. The denominator of total possible days supply consisted of the number of days from the first fill until the end of the study period to account for Rxs filled later in the study period.	≥0.8 = good
			0.5 ≤ MPR < 0.8 = poor
			MPR < 0.5 = very poor
MPR	Gincherman et al. [37]	Sum of medications supplied divided by the number of days between first and last fill.	≥80% medication possession ratio over the 12 months = good adherence
MPR	Lassen et al. [45]	Sum of days’ supply of all GLP1RA dispensations divided by total days in observation period (first fill to followup)	≥80% = adherent
			<80% = non-adherent
MPR	Machnicki et al. [49]	Number of days of medication supplied within the refill interval divided by the number of days in the refill interval.	Not specified.
MPR	Park et al. [54]	The sum of total days’ supply for all fills divided by the number of days between first and last fills plus days supply of the last fill or the number of days between first fill and December 31, whichever ends later.	≥80% = adherent
			<80% = non-adherent
MPR	Pruette et al. [55]	Ratio of the sum of days’ supply between interval start and end dates (days’ supply/study interval).	≥75% = adherent
			<75% = non-adherent
MPR	Qin et al. [56]	Total days supplied including excess divided by the landmark date − 1st supply date capped at 1.	≥80% = adherent
			<80% = non-adherent
MPR	Schmitt et al. [59]	Actual treatment days divided by total possible treatment days (truncated for the study period or death).	≥80% = good adherence
			<80% = poor adherence
Modified MPR	Qin et al. [56]	Total days supplied including excess were divided by the last supply date minus 1st supply date.	≥80% = good adherence
			<80% = poor adherence
Median gap between prescription refills	Cooke & Fatodu [35]	The measure was calculated as the median of the number of days a Rx was refilled in relation to the end of the days supply of the previous Rx. A positive median gap represents the No. of days the patient was late in filling subsequent refills, with larger numbers denoting lower adherence.	N/A
			Larger numbers denoting lower adherence
PDC	Gor et al. [38]	Sum of days supply from the index date through the last dispensing divided by the total days of the follow-up period (i.e., 365 days).	≥80% = adherent
			<80% = non-adherent
PDC	Hsu et al. [8]	Number of follow-up days covered with medication divided by the total number of days in follow-up. Any oversupply during a specific observation period is truncated.	≥80% = adherent
			<80% = non-adherent
PDC	Janse et al. [40]	The number of consecutive fills and the distance (in days) between them were considered. The expected duration of the fill (i.e. the number of days that one package of medication would last) was estimated through the average frequency of dispensing for each single formulation in the study population.	≥80% = good adherence
PDC	Machnicki et al. [49]	Proportion of days the patient had the index drug within the 12 month follow-up period.	≥80% = adherent
			<80% = non-adherent
PDC	Qin et al. [56]	Total days covered by drug supply divided by the landmark date minus the 1st supply date.	≥80% = good adherence
			<80% = poor adherence
PDC	Santoro et al. [58]	The ratio between the no. of days of medication and days of observation (365 days), multiplied by 100.	≥80% = adherent
			<80% = non-adherent
PDC	Tang et al. [63]	Sum of days’ supply of drug was divided by the number of days between the first fill of the prescription in the follow-up period and the end of the follow-up period.	≥80% = adherent
			<80% = non-adherent
PDC	Truong et al. [12]	The number of days covered by at least one drug class was divided by during the year after treatment initiation by 365 days.	≥80% = adherent
			<80% = non-adherent
MEMS	Bandiera et al. [29, 30]	It registers the date and hour of each opening in real time as a proxy of drug intake.	If all MEMS used by a patient were opened at least once daily, medication implementation was considered as active (=1) and no implementation was considered (=0) otherwise.
MEMS	Kronish et al. [44]	The device records the date and time when each compartment is opened. Adherence data are transmitted from the pillbox to the MedSignals server via landline or, for devices used in the final year of the study, via a cell phone in the device.	≥80% = adherent
			<80% = non-adherent
MEMS	Lee et al. [46]	An electronic monitor which records the date and time of bottle cap openings.	Adherence by MEMS to a once-a-day drug dosing schedule was acceptable if 80% of the time intervals between MEMS openings were within 24 ± 6 h.
MEMS	Pruette et al. [55]	The MEMS TrackCap is a child-resistant medication bottle cap that records the date and time of each bottle opening and closure, which is later downloaded.	≥75% = adherent
			<75% = nonadherent
MEMS	Van Camp et al. [64]	The microchip in the MEMS cap registered date and time of each opening (i.e. the presumed intake).	-) Being adherent was defined as missing < 1 total daily dose/week.
			-) being totally adherent as missing < 1 total daily dose/week, every week.
Adherence Ratio	Cohen-Glickman et al. [34]	Drug practices were abstracted from patients’ records and compared to electronic pharmacy data. The discrepancy between drug intake reports and the actual purchase was measured to estimate adherence.	N/A
Pill counting	Lee et al. [46]	Adherence as measured by pill count can only be determined when patients return their med bottles and is inflated if pills that are not returned to the clinic are not taken by the patient.	An acceptable level of adherence by pill count was achieved if 80% to 100% of the prescribed pills were not returned to the clinic.

MPR, Medication Possession Ratio; Rx, prescription; PDC, Proportion of Days covered; MEMS, Medication Event Monitoring System

Quality assessment

The quality assessment found that all included studies were at “low” concern regarding applicability for this review. Among studies that used subjective tools, eleven studies were deemed to be at risk of bias due to “unclear” assessments in the QUADAS-2 domains of “patient selection” [25, 26, 28, 41–43, 47, 51, 55, 60, 65] four studies had “unclear” assessments in the domain of “index test” [10, 41, 43, 65] and ten studies were found to have “unclear” assessments in the domain of “flow and timing” [25, 27, 28, 39, 41–43, 47, 48, 50]. On the other hand, among studies that utilized objective measures for assessing medication adherence, one study was judged to have “unclear” assessments in the QUADAS-2 domains of “patient selection” [55], and three studies had “unclear” assessments in the “index test” domain [30, 34, 57], and three studies were found to have “unclear” judgment in the domain of “flow and timing” [37, 46, 56]. Moreover, it’s noteworthy that the domain of “reference test” was not relevant to the study as there is no reference standard for measuring medication adherence in CKD patients. However, among the included studies, only one used MEMS as the reference standard [55].

Discussion

This systematic literature review summarizes the medication adherence measures and tools used in the context of CKD patients from articles published from January 1, 1946, to September 25, 2025. To the best of our knowledge, this is the first systematic review conducted on medication adherence measures and tools used in patients with CKD. There has been one systematic review of qualitative studies of patients’ experiences of factors that facilitate and hinder adherence to medication in CKD patients [13]. However, there has been no study on methods and tools that are used to assess medication adherence in patients with CKD. The present review fills this gap in literature. Moreover, we found that there is no gold standard method or tool for measuring medication adherence in the CKD population.

Studies yielded different results regarding the most frequently used medication adherence measurement methods and tools for various disease states and conditions. For instance, we found that MMAS-8 and MPR, followed by PDC and MEMS, were the most frequently used methods in the context of CKD patients. However, other studies showed that MPR, PDC, and patient-reported outcome questionnaires were used in almost equal numbers in the context of patients with cardiovascular disorders [4]. Additionally, studies on adherence to medication in psychiatric and mental disorders revealed the dominant use of patient-reported questionnaires [4]. This may be due to the practicality and ease of administration of these latter measures to patients with psychiatric disorders. Prevalent use of MPR measures in the context of CKD patients could be due to the large number of this group of patients in both primary and secondary care which allows for the assessment of adherence on a large scale. Moreover, polypharmacy is common in CKD patients, with each patient taking an average of 10 medications [7, 8]. Measuring adherence to multiple medications is generally more feasible and straightforward with administrative databases than with multiple isolated questionnaires. However, many studies still relied on MMAS-8 as a self-reported measure of adherence, reflecting the practicality of conducting smaller-scale studies compared with large-scale administrative data analyses. Subjective measures based on patient self-reported questionnaires are user-friendly, clear, and cost-effective, making them convenient for administration in clinical practice and research settings [4, 66]. This may explain why they were found to be the most used measure in our review. Within the self-reported questionnaires, ten different tools were administered in CKD patients, with MMAS (MMAS-4 and MMAS-8) being the most widely used tool (n = 11[25.6%]). The dominance of MMAS measures is likely due to MMAS being a well-established measure with good psychometric properties (reliability, concurrent and predictive validity measures) across many disease states [67–70].

To help clinicians choose the most suitable tool(s) by which to measure adherence in their clinical setting, one needs to consider several factors, including patient population, literacy, availability of resources, and burden of data collection. One of the major aims of using subjective tools (self-reported questionnaires) is to understand barriers to medication adherence and to elicit patients’ beliefs and perceptions about their pharmacotherapy [17, 66]. Patients’ beliefs and perceived barriers to adherence are not captured by objective measures of adherence (e.g. administrative databases, MEMS, etc.). Thus, subjective tools do allow an improved understanding of patients’ beliefs, perceptions, and correlates of medication-taking behavior. This is a major advantage of using these methods and tools in any clinical/research setting, as interventions may then be suggested based on the results [17, 66]. Reported reasons for non-adherence and exploration of patients’ perception of their therapeutic regimen [27, 28, 31–33, 43, 52, 53, 55, 60, 62, 65] may be helpful for clinicians and healthcare professionals to address the identified barriers and implement interventions tailored to each patient to resolve specific barriers and improve medication adherence. However, clinicians should consider that these self-report tools have the potential to underreport non-adherence due to social desirability concerns (e.g. to avoid disapproval from the healthcare providers), recall bias, missing data, or fault in self-observation. Patients’ psychological state may also impact the responses [4, 17, 66]. Healthcare providers who aim to use self-reported adherence tools in clinical practice should ideally use tools that allow for both short- and long-term assessment of medication adherence. This may require modifying existing tools to include both short and long recall periods [17]. In this way, clinicians and researchers will be well equipped to evaluate medication-taking behavior over multiple periods and not solely rely on recall surrounding one period of time, and potentially also recognize that adherence may change over time, as the clinical or personal circumstances of the individual changes.

Some self-reported tools were not correlated with objective measures (e.g., those using administrative datasets and electronic monitoring devices), which may complicate their validation. The adherence questionnaires that were correlated with objective measures employed a range of objective measures for this purpose, making interpretation of correlations across different self-reported tools difficult. It’s also noteworthy that some self-reported medication adherence questionnaires are designed for specific diseases. For instance, the “Adult Acquired immunodeficiency syndrome (AIDS) Clinical Trial Group Adherence to Combination Therapy Guide” and “Community Programs for Clinical Research on AIDS Adherence Form” are specific questionnaires designed for patients with HIV [71]. Moreover, in the context of osteoporosis pharmacotherapy, the “Adherence Evaluation of Osteoporosis Treatment (ADEOS) Questionnaire” and “Osteoporosis Specific Morisky Medication Adherence Scale (OS-MMAS)” are available [72]. However, the present review found no specific questionnaire designed for evaluation of medication adherence in the context of CKD patients. There is a validated questionnaire named “The End-Stage Renal Disease Adherence Questionnaire (ESRD-AQ)” which measures adherence to dialysis treatment in patients on dialysis. However, it is not specific to assessment of medication, and it evaluates adherence to dialysis treatment schedule, fluid intake, diet, and medication altogether [73].

When clinicians intend to choose the most suitable tool(s) to measure adherence in their setting, they should consider that both subjective and objective measurement methods have their own specific advantages and disadvantages. The use of administrative databases (i.e., curated primary data in systems such as electronic prescription services or pharmacy insurance claims) as one of the most popular objective measures allows for analyzing a large population. This has led to the popularity of this method in research settings [66]. It is essential to recognize that use of these methods is dependent upon centralized computerized systems and consistency between prescribers and dispensers to collect a complete dataset for the specific period of the study. Some prescriptions may also be missed if obtained outside the insurance plan. Another drawback of using administrative data sets is that it is assumed that medication is taken exactly as prescribed. Therefore, partial adherence (i.e. patients taking only part of a treatment regimen in the study period) cannot be determined [66].

Our work has several strengths, including our strict adherence to the methods outlined in the guidelines for designing and conducting systematic reviews. We searched multiple platforms for relevant studies. Our search strategy was reviewed by an experienced librarian specialized in the area to aid in locating all available evidence related to the topic and minimize the chances of inappropriately excluding studies. Our study selection, data extraction, and quality assessment were all duplicated by five pairs of specialized reviewers (all the reviewers were community and clinical pharmacists or nephrologists) to minimize bias and improve reliability. We included international representation of studies that assessed medication adherence in CKD patients to prioritize generalizability and understand the intricacies and various features of different methods and tools used to measure medication adherence in CKD patients. Unlike most studies that are conducted on adults, our study had an inclusive age range, thus being more extensive. It’s also noteworthy to mention that in order to enhance the methodological rigor and comparability of the included studies, we applied additional exclusion criteria beyond the original protocol. Studies that focused only on developing/validating algorithms for adherence measurement were excluded to ensure that the review captured the practical application of medication adherence tools in real-world or clinical settings. Similarly, studies with small sample sizes ( < 25) were excluded to reduce the potential impact of underpowered results. For subjective measures, we required that validity measures be reported, allowing for a more accurate assessment of the quality and interpretability of the findings. While these criteria were not specified in the original protocol, their implementation strengthened the review by focusing on studies that provide robust evidence regarding medication adherence measurement.

The present systematic review did not aim to find the best or favored tool(s) for assessment of medication adherence in the targeted population. We do, however, provide clinicians and researchers with a comprehensive description and interpretation of key characteristics of the available methods and tools that could guide decisions on which method/tool might fit best in each setting and the clinicians’/researchers’ goals in eliciting medication-taking behavior.

Therefore, we believe the results of the present systematic review would be of value to nephrologists, general practitioners, and pharmacists, and any other healthcare provider involved in the care of patients with kidney diseases.

Future research might focus on developing standardized tools for measuring medication adherence, specifically in patients with CKD and across the spectrum of the condition (from early diagnosis to later stages). Additionally, research could focus on assessment of acceptability of the use of those tools by the patients. Depending on purpose and perspective (individual, health system, clinical, research), using multiple instruments simultaneously to estimate medication adherence in CKD populations may be beneficial. Another area of future research could be an investigation of differences in the level of adherence of CKD patients to medications that treat signs and symptoms of the disease versus those prescribed for ‘asymptomatic’ conditions. We also suggest the development of a checklist/guideline to address the minimum criteria required in medication adherence measurement studies (similar to Peterson et al.’s study) [74]. In this way, future studies could be easily replicated and compared systematically, paving the way for establishment of policy-making guidelines. Moreover, a checklist could potentially ensure that an established minimum quality level is met and maintained across all future medication adherence studies, which will be a great improvement for the field of medication adherence research. We also suggest that developing tools specifically for administration via mobile device-based applications or patient electronic health record portals may be critical to improve uptake and relevance in clinical practice. Such tools may also lend themselves to incorporate possible interventions to improve adherence (e.g. reminders and prompts).

Conclusion

The findings of this systematic review suggest that there is no accepted consistent reference tool available to measure medication adherence in patients with CKD. Given the importance of medication adherence and reports of poor sustained adherence across multiple conditions, including CKD, we would advocate for the development of practical and robust tools adapted for this patient population. With current data, the use of a multi-measure approach is recommended for the assessment of medication adherence in patients with CKD. Given the strengths and weaknesses of both objective and subjective tools identified to date, some combination of methods is likely the best to serve both individual clinicians and patients, as well as health care systems and researchers.

In conclusion, out of subjective tools used to measure adherence in patients with CKD, the MMAS-8 appears to have the best psychometric properties and is currently in widespread use. Out of the objective tools, using MEMS would be recommended (although not cost-effective) as this tool has been regarded as a reference standard by several researchers in the field. If access to administrative datasets is feasible, utilization of PDC would be suggested.

Finally, clinicians who care for patients with CKD will need to carefully review the characteristics and features of each available tool to identify the most meaningful ones for their patients’ population and clinical workflow.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

ADEOS

Adherence Evaluation of Osteoporosis Treatment Questionnaire

AIDS

Acquired immunodeficiency syndrome

CAAMQ

The Child & Adolescent Adherence to Medication Questionnaire

CKD

Chronic Kidney Disease

ESRD-AQ

End-Stage Renal Disease Adherence Questionnaire

MEMS

Medication Events Monitoring System

MeSH

Medical Subject Headings

MMAS-4

4-item Morisky’s scale

MMAS-8

Eight-Item Morisky Medication Adherence Scale

MPR

Medication Possession Ratio

OS-MMAS

Osteoporosis Specific Morisky Medication Adherence Scale

PDC

Proportion of Days Covered

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

Author contributions

ER designed the project and wrote the manuscript. ER, DR, MB, NB, CH, HW, KH, AP, SG, & MA collected data and reviewed the manuscript. ER did data analysis. AL contributed to conception of the study, provided academic supervision, and revised the manuscript. All authors read and approved the final manuscript.

Funding

The authors received no financial support for this research.

Data availability

No datasets were generated or analyzed during the current study.

Declarations

Ethics approval and consent to participate

Ethical approval was not required for this type of study at the host institution. Consent to participate is not applicable here.

Consent for publication

Consent for publication is not applicable for this study type.

Competing interests

The authors declare no competing interests.