Introduction

Tablets and capsules are the most commonly used oral dosage forms because they are portable and less invasive. Ensuring patient adherence is essential for effective medication use and successful pharmacotherapy. However, medication use problems, such as difficulties in handling, can lead to decreased adherence, particularly when these issues impact the patients’ self-management abilities1,2. The term “handling” in this context refers to the physical actions required to manage oral medications, including removing tablets or capsules from packaging, picking them up with fingers, and placing them into the mouth. These tasks can be affected by impairments in hand function, such as tremor, reduced grip strength, or joint stiffness. Several studies have been conducted to identify the factors associated with difficulties in handling tablets and capsules. On the patients’ side, impaired hand and finger function3,visual impairments, and hand disabilities caused by conditions such as rheumatoid arthritis4 often lead to difficulties in handling tablets and capsules. In terms of tablet size, smaller tablets are easier to swallow; however, they can also present a “handling difficulty.”5,6 This difficulty arises from challenges in picking them up, holding them securely, or extracting them from packaging, especially for individuals with reduced hand function or coordination. These handling-related issues are distinct from swallowing difficulty and may independently affect medication adherence. While swallowing difficulty refers to challenges in safely ingesting medications once placed in the mouth, handling difficulty encompasses the preceding steps required to access and manipulate tablets or capsules. Moreover, excessively small or flat tablets have been reported to affect usability negatively7. However, these studies come with several limitations. Some of these studies include evaluations based on handling single doses of tablets and capsules, which may not reflect real-world clinical scenarios where multiple tablets are taken simultaneously5,8,9,10,11. Other issues include a lack of evaluation on extracting medication from packaging sheets5,11, small sample size, ranging from a few cases to several dozen cases11,12, and the use of model formulations with limited medication options5,10,11. These limitations reduce the generalizability of the findings and highlight the need for real-world, patient-centered data that reflect actual medication use in daily life.

To overcome these methodological limitations, we focused on the electronic medication notebook as a data source and survey platform. The electronic medication notebook system (harmo®) is a smartphone-based application that stores patients’ dispensing records, images of dosage forms, and personal health information such as allergies and medical history. This system allows linkage between questionnaire responses and dispensing data, enabling analysis that combines patient-reported experiences with real-world prescription information. Conducting a questionnaire survey among users of the electronic medication notebook enables the analysis of healthcare information based on survey results. Our previous study examined patient preferences for oral medications using electronic medication notebooks and personal health records (PHRs) and revealed contributing factors, including size, that lead to difficulties in swallowing tablets or capsules13 and extracting them from press-through-package sheets14. Based on the preference of a large number of patients, a combined size of 21 mm (length + width + depth) was determined as the concrete threshold for medical tablets and capsules that patients found too large to swallow13.

In this study, we aimed to explore the factors affecting the “handling difficulty” of oral medications using PHRs and questionnaires to obtain patients’ preferences directly. We sought to identify (1) the size of medications perceived as “small and difficult-to-handle tablets/capsules” and (2) the patient backgrounds more likely to perceive oral medications as “small and difficult-to-handle.”

Results

The questionnaire was electronically sent to 49,488 electronic medication notebook harmo® users, of whom 1,554 (3.1%) accessed the questionnaire website, and 1,230 (2.5%) responded to the questionnaire. In response to the question about having “small and difficult-to-handle tablets/capsules,” 295 (24.0%) reported having these types of medications. Regarding the condition of their hands and fingers, 1,029 (83.6%) responded that they had no problems. The age and sex characteristics of the respondents and their responses to the question about having “small and difficult-to-handle tablets/capsules” are listed in Table 1.

Table 1 Characteristics of respondents (n = 1,230).
Full size table

Examination of indicators and calculation of threshold values

The flowchart of the analysis is shown in Fig. 1. Of the 286,831 dispensing records (for 1,230 respondents), we extracted the data from 295 respondents who answered having “small and difficult-to-handle tablets/capsules (self-reported).” Of these, 53,484 dispensing data, including 42,648 tablets or capsules, were obtained from the 192 respondents whose “small and difficult-to-handle tablets/capsules” could be identified by the prescription records. The remaining 103 patients were excluded from the main analysis because the “small and difficult-to-handle tablets/capsules” they reported could not be identified in the prescription data due to vague or inaccurate descriptions in the questionnaire (e.g., responses such as “don’t remember,” “unclear,” or just “tablets”.) For reference, 210 medications were identified as “small and difficult-to-handle tablets/capsules,” and 859 as “concomitant medications,” based on the prescription data of the 192 respondents with verifiable dispensing records. Figure 2 shows the ROC curves, while Table 2 shows the ROC-AUC and various parameters that differentiate between 210 “small and difficult-to-handle tablets/capsules” and 859 “concomitant medications.” The ROC-AUC for “long diameter + short diameter” reached a maximum of 0.834, and the threshold for “long diameter + short diameter” was calculated to be 13.3 mm.

Fig. 1
figure 1

Flowchart of analyzed medications.

Full size image
Fig. 2
figure 2

Receiver-operator characteristic curves for identifying “small and difficult-to-handle tablets/capsules” based on 8 indicators and their Area Under the Curve (AUC).

Full size image
Table 2 Size thresholds for “small and difficult-to-handle tablets/capsules”.
Full size table

Exploring patient factors

A univariable analysis was conducted (Supplemental Table 3), and several factors were identified as statistically significant: conditions of respondents’ hands and fingers (“having any problems” or “hard to apply force”), the prescription of soft capsules, the prescription of medications with long diameter + short diameter ≤ 13.3 mm, the prescription of medications with long diameter + short diameter + thickness ≤ 17.45 mm, the number of oral medications taken per day, the number of tablets or capsules taken per day, the highest number of tablets or capsules taken at the same time, and certain medical histories (hypertension, arthritis or rheumatoid arthritis, and osteoporosis).

Binomial logistic regression analysis

Table 3 shows the results of the binomial logistic regression analysis using statistically significant variables in the univariable analysis, adding age and sex. The following variables were identified as statistically significant: difficulty in applying force with the hands and fingers, prescription of tablets or capsules with a long diameter + short diameter ≤ 13.3 mm, history of hypertension, and history of osteoporosis. As the highest number of tablets or capsules taken at the same time, the number of oral medications per day, and the number of tablets or capsules taken per day showed multicollinearity among the variables with a P-value < 0.05 in the univariable analysis, the latter two factors were not used as exploratory factors. After eliminating the two factors that indicated multicollinearity, the calculated variance inflation factor was < 3.000, confirming the absence of multicollinearity among the explanatory variables (Supplemental Table 2).

Table 3 Binomial logistic regression analysis for identifying the factors related to the prescription of “small and difficult-to-handle tablets/capsules”.
Full size table

Decision tree analysis

As a result of the decision tree analysis (Fig. 3), the patients who were prescribed tablets or capsules with the sum of the long and short diameter ≤ 13.3 mm and with a history of osteoporosis were judged to have the highest risk (61.1%) of having “small and difficult-to-handle tablets/capsules.” The second highest risk was the combination of (1) prescription of tablets or capsules with the sum of the long and short diameter ≤ 13.3 mm, (2) no history of osteoporosis, and (3) hand and finger condition: hard to apply force (36.4%).

Fig. 3
figure 3

Decision tree analysis with the prescription of “small and difficult-to-handle tablets/capsules”.

Full size image

Discussion

In this study, we utilized an electronic medication notebook-based PHR and conducted a survey regarding “small and difficult-to-handle tablets/capsules.” We discovered the threshold value for “small and difficult-to-handle tablets/capsules” and identified patient factors that affected “handling difficulty”, such as hand and finger conditions that make it hard to apply force, and a medical history of hypertension and osteoporosis. Our study is noteworthy because we were able to combine patient preference data (typically presented as narrative data) with dispensing data, and calculate clear threshold values for “small and difficult-to-handle tablets/capsules” using actual formulation data.

We found that a sum of 13.3 mm for the long and short diameters of tablets and capsules was the threshold indicator for “small and difficult-to-handle tablets/capsules.” (ROC-AUC: 0.834) Taking tablets or capsules smaller than the threshold value was identified as a related factor in the patient factor analysis, demonstrating the robustness of the threshold value. In a previous study that explored the size of tablets, easy-to-grasp size was determined to be more than 7–8 mm in diameter for circular tablets (the sum of the longer and shorter diameters must be 12–14 mm), as well as 7–8 mm for the longer axis and 5.4–5.6 mm for the shorter axis for oval tablets (12.4–13.6 mm in total) in older ages ranging from 60 to 89 years5. It may be difficult to directly compare our results with those of the previous study, but our calculated thresholds align with previous studies, indicating the robustness of our results.

The binomial logistic regression analysis identified four significant and independent factors associated with the prescription of “small and difficult-to-handle tablets/capsules”: (1) the prescription of tablets or capsules with long diameter + short diameter ≤ 13.3 mm, (2) difficulties in applying force (“hard to apply force” condition) with the hands and fingers, (3) a medical history of hypertension, and (4) a medical history of osteoporosis. Notably, the subjective symptom “hard to apply force” in the hands and fingers was more strongly associated with difficulty than diagnoses such as arthritis or rheumatoid arthritis. Although patients with arthritis or rheumatoid arthritis are known to have difficulty pushing tablets and capsules out of packaging and picking small tablets15,16,17, the current study did not find a significant relationship. This may be due to the limitations of self-reported medical histories recorded at the time of registering for the electronic medication notebooks, which may have lacked clinical accuracy or completeness. Although age was not a statistically significant independent factor in our logistic regression analysis, it may still play an indirect role in handling difficulty. Older adults are more likely to experience physical conditions such as frailty, reduced manual dexterity, or comorbidities like osteoporosis, which were found to be significant in our study, suggesting that age may act as a proxy for these underlying functional limitations, even if not directly associated with the outcome in multivariable analysis. This interpretation is consistent with prior research indicating that age-related physical decline contributes to tablet handling difficulty, even when age itself is not a direct predictor5. Based on the results of this study, it is recommended that highly individualized dosage form selection be made, taking into consideration the factors specific to each patient and the size of the tablet or capsule formulation. Our findings suggest that direct, patient-reported symptoms may provide a more sensitive measure of physical limitations in daily medication use than diagnostic labels alone.

In addition, significant associations with a medical history of hypertension and osteoporosis highlight potential risk groups who may benefit from more careful dosage form selection. Hypertension is a common disease among middle-aged and older adults, who are often prescribed multiple medications. The rise of lifestyle-related diseases and age-related syndromes increases the need for medications in older adults, often resulting in polypharmacy18. Given that hyperlipidemia showed a similar trend, it is possible that the observed association with hypertension reflects broader patterns of long-term medication use and lifestyle-related conditions, rather than a direct causal effect of hypertension. In particular, evidence from a recent cohort of older inpatients demonstrated that polypharmacy was significantly associated with worsening frailty and increased mortality over two years, even after adjusting for multimorbidity and baseline frailty, including measurements of handgrip strength19. This supports our hypothesis that the burden of multiple medications, rather than hypertension per se, may underlie the observed handling difficulties. A medical history of osteoporosis could be related to age and sex. Osteoporosis is more prevalent among postmenopausal women, who are also more likely to experience conditions that cause stiffness in the fingers, such as rheumatoid arthritis, osteoarthritis, and Heberden’s nodes20. A medical history of osteoporosis might be a surrogate factor for having these characteristics. Further, the formulation characteristics of osteoporosis medications may directly contribute to handling difficulty. Vitamin D preparations, such as eldecalcitol and alfacalcidol capsules, are often prescribed as spherical soft capsules. These are known to roll after being removed from blister packs, making them harder to handle3. In a previous study, the round shape of Alfarol® (alfacalcidol) capsules caused them to roll after being taken out, and patients preferred the medication in tablet form rather than as spherical soft capsules3. Although age, sex, and arthritis-related conditions were adjusted for in the regression model and were not statistically significant, osteoporosis remained independently associated with handling difficulty. This suggests that osteoporosis may reflect unmeasured functional or anatomical factors, such as subclinical impairment in dexterity or chronic musculoskeletal burden, that were not fully captured in our dataset. Therefore, we acknowledge the possibility of residual confounding influencing this association. Although previous studies3,4 have identified arthritis as a factor impairing medication handling, our analysis did not find a statistically significant association. One possible explanation is the reliance on self-reported medical history, which may have led to underreporting or misclassification of musculoskeletal conditions. Additionally, mild or undiagnosed joint problems may not have been captured, potentially attenuating the observed association.

In our previous study, we reported a threshold for “difficult to swallow due to large size” as the sum of the long diameter, short diameter, and thickness equaling 21.5 mm13. Although direct comparison is difficult due to differences in indicators and outcomes, the present study also demonstrated relatively good discriminatory performance for the “long diameter + short diameter + thickness” metric (threshold: 17.45 mm, ROC-AUC: 0.818). Furthermore, “hard-to-push-out” drugs were found to have characteristics such as a spherical shape or small major axis or small major axis drug-to-pocket- size ratio by using the same research framework14. These findings offer valuable insights into the physical characteristics that affect the ease of swallowing and handling of oral medications, potentially contributing to the future design of more patient-friendly dosage forms.

This study has several strengths. Compared to previous studies, this study was able to determine a threshold for the size of oral medications in a setting closer to real-world clinical practice. The evaluation was conducted on medications that patients actually take daily, allowing for a more detailed determination of size-related thresholds. Relying solely on questionnaires made it difficult to accurately identify the specific medications taken by patients, and the influence of patients’ recall bias made it difficult to accurately identify the correct medication names in the free-text responses. By integrating the questionnaire data with the dispensing information through PHR, we were able to obtain more detailed information on the patient’s medication history and on medications (concomitant medications) that were not answered.

This study has several limitations. First, regarding the study participants, we included only patients who used an electronic medication notebook. This selection may have led to a bias toward individuals who are capable of using smartphone applications, potentially excluding older adults and individuals who, for economic reasons or visual impairments, are unable to use smartphones. Moreover, as both the electronic medication notebook and the questionnaire required smartphone operation, our study may have inadvertently excluded patients with significant manual dexterity impairments. In addition, the overall response rate to the questionnaire was low (2.5%), which may introduce response bias and limit the generalizability of the findings. Further validation in broader and more diverse patient populations is warranted to confirm the applicability of our results. Second, concerning the questionnaire content, the study relied on self-reported responses regarding medical history, making the reliability of the information uncertain. Additionally, the questionnaire itself was not formally validated, which may impact the consistency and reproducibility of responses. Although it was based on prior surveys and expert input, this lack of validation should be considered when interpreting the findings. Regarding medical history, participants selected conditions from a predefined list, which may not have covered all possible conditions. To improve the reliability of information on medical histories, it may be beneficial to combine diagnosis data from medical institutions or health claims.

Despite these limitations, this study has provided new insights into difficulties in handling tablets and capsules among patients using an electronic medication notebook. These results could meaningfully contribute to the manufacturing of formulations that do not compromise medication adherence.

This study identified a threshold for “small and difficult-to-handle tablets/capsules,” as “the long diameter + short diameter = 13.3 mm.” The prescription of small tablets/capsules, the condition of the respondents’ hands and fingers where they find it hard to apply force, a medical history of hypertension, and a medical history of osteoporosis were additionally found to be factors related to difficulty in handling tablets and capsules. These findings could serve as a reference for dosage form selection in the future development of new and generic medications.

Furthermore, our results may help guide pharmaceutical manufacturers in designing dosage forms that are easier for patients with limited dexterity to handle. Formulations that avoid overly small or spherical shapes could improve usability, especially for older adults and patients with musculoskeletal conditions. In addition, integrating formulation-based risk alerts into electronic prescribing or medication management systems may help healthcare providers identify patients at risk of medication handling difficulties and select more suitable alternatives.

Methods

Questionnaire survey to the electronic medication notebook user

The questionnaire survey was conducted from January 26 to February 1, 2023. The survey was sent to patients aged 20 years or older who were prescribed either capsules or tablets, through harmo®, an electronic medication notebook smartphone app operated by harmo Co., Ltd. The questionnaire included both closed- and open-ended items, covering topics such as difficulties in handling medications, hand and finger function, and medical history. Survey items are listed in Supplemental Table 1. While no formal validation study was conducted, the questionnaire was developed based on previous related surveys and expert consultation, and was internally reviewed for clarity and relevance. PHR data (dispensing history, age, sex, medical histories, allergies, side effects, etc.) of the respondents were collected and analyzed in combination with the questionnaire responses. Information about the dosage forms of the oral medications was evaluated based on the pharmaceutical formulation data obtained from the package inserts and pharmaceutical interview forms.

Identification of “small and difficult-to-handle tablets/capsules”

We identified the names of the prescription medications reported by patients as “small and difficult to handle” based on their responses in the questionnaire, in combination with dispensing histories. Of these, tablets and capsules for which the data of “long diameter,” “short diameter,” “thickness,” and “weight” could be identified from the package inserts or pharmaceutical interview forms were defined as “small and difficult-to-handle capsules/tablets,” and each parameter was used to calculate the threshold value. Tablets and capsules, for which parameter information was unavailable, were also referred to in the patient factor analysis. Tablets and capsules that were identified from the dispensing histories as being taken simultaneously as the “small and difficult-to-handle tablets/capsules” were defined as “concomitant medications.” Size parameters were also identified for the “concomitant medications,” and these were also analyzed subsequently.

Threshold determination using ROC analysis

We examined indicators that could identify “small and difficult-to-handle tablets/capsules” by comparing “small and difficult-to-handle tablets/capsules” and “concomitant medications,” and calculated threshold values. For each of the “small and difficult-to-handle tablets/capsules” and “concomitant medications,” the measures of a total of eight indicators were calculated: “long diameter,” “short diameter,” “thickness,” “long diameter + short diameter,” “long diameter + short diameter + thickness,” “long diameter × short diameter,” “long diameter × short diameter × thickness,” and “weight.” Among them, a “threshold value” refers to a cutoff point in a size-related indicator that best distinguishes between “small and difficult-to-handle tablets/capsules” and concomitant medications. The receiver operating characteristic (ROC) curve was plotted for each indicator, and the largest area under the curve (AUC) was used as the representative indicator of the size of “small and difficult-to-handle tablets/capsules.” The point at which the Youden index was the highest was defined as the threshold for the size of “small and difficult-to-handle tablets/capsules.”

Exploring patient factors

Univariable analysis was conducted to compare differences between the two groups of patients: those who reported “small and difficult-to-handle tablets/capsules” and for whom corresponding prescriptions were identified, and those for whom no such combination was observed. The explanatory variables were sex, age (divided into decades), condition of respondent’s hands and fingers (multiple choices from “no problem,” “trembling,” “numbness,” “hard to apply force,” and “difficulty in bending fingers”), with or without soft capsules prescription, with or without OD tablets prescription, with or without powder medicine prescriptions, with or without liquid medicine prescription, with or without single packaged medicine prescriptions, with or without “small and difficult-to-handle tablets/capsules” prescription, the number of all oral medications taken per day, the number of tablets or capsules taken per day, the highest number of tablets or capsules taken at the same time, and medical histories. For patients who reported having “small and difficult-to-handle tablets/capsules,” the “small and difficult to handle tablets/capsules” and other medications taken regularly at the same period were analyzed to identify the patient factors. For patients who answered that they did not have any “small and difficult-to-handle tablets/capsules,” we defined the most recently dispensed medications as the “medications for analysis.” The ROC-derived threshold value was used for characterizing individual formulations, but not for grouping patients.

Multivariable analysis and statistical considerations

For categorical variables, Pearson’s chi-square test was used when expected cell counts were sufficient, and Fisher’s exact test was used when the expected cell counts were small. The Mann–Whitney U test was used for continuous variables. These analyses were followed by binomial logistic regression analysis. In addition to the variables with significant differences in the univariable analysis, age group (organized by decade), and sex were used as explanatory variables in the binomial logistic regression analysis, as they have previously been identified as factors associated with difficulties in handling medications. Variables with high intercorrelation were screened for multicollinearity using the variance inflation factor (VIF), and only non-collinear variables were retained. In all analyses, a P-value < 0.05 was considered statistically significant.

Decision tree analysis

A decision tree analysis was conducted to determine the factors related to the prescription of “small and difficult-to-handle tablets/capsules,” incorporating as explanatory variables the factors for which P-values < 0.05 were obtained in the binomial logistic regression analysis. The chi-square automatic interaction detection algorithm, which divides features using a chi-square test, was employed as the algorithm for decision tree analysis. The algorithm branches the tree by selecting the variable that is most significant in the chi-square test. Based on a previous report using the same method, the analysis conditions were set as follows: the minimum number of cases required to split a node was 20, the minimum number of cases in a child node was 10, and the significance (α-merge, α-split, P-value) was < 0.0513.

Microsoft Excel 2016 (Microsoft Corp., Redmond, WA, USA) was used for descriptive statistics, and SPSS 29.0.2.0 (IBM Corp., Armonk, NY, USA) was used for indicator exploration, threshold calculation, univariable analysis, binomial logistic regression analysis, and decision tree analysis.