Schizophrenia is a mental disorder that imposes a heavy economic burden and causes severe functional disabilities, including cognitive impairment [1,2]. Its clinical manifestations are divided into positive and negative symptoms. Positive symptoms include delusions, hallucinations, and disorganized behavior, which are often accompanied by impaired cognitive and socio-occupational functions, while in contrast, negative symptoms manifest as apathy and decreased motivation, which gradually become apparent as the disorder progresses [3]. Many factors accelerate aging in patients with schizophrenia, making their life span shorter than that of individuals without schizophrenia, with impacts such as deterioration of life functioning, the presence of chronic illnesses, obesity, and an unhealthy lifestyle [4]. Male and female patients with schizophrenia have an average life expectancy of 61.48 years and 60.81 years, respectively [5].

Schizophrenia is a treatable disorder, and early intervention can prevent further deterioration. Antipsychotic drugs are the first choice for alleviating the psychiatric symptoms of these patients, and the goal of regular medication with the right drugs, dosages, and frequency is to prevent recurrence and overall functional deterioration [6,7]. Although it is common for patients with schizophrenia to not adhere to their prescriptions for a variety of reasons, including the stigma of mental illness, side effects of drugs, forgetting to take medicine, and lack of insight [8,9].

The medication adherence rate among patients diagnosed with schizophrenia for the first time is only 50% in the first year [4,5], and the rate of relapse within 5 years is up to 80% [10]. Similar studies have reported that 50% to 80% of patients with schizophrenia do not take their medications regularly [6,7] or take the wrong medications or dosages [8]. Nonadherence may lead to the deterioration of symptoms, a decline in daily life functions, and a poor prognosis; it may also result in hospitalization and the escalation of medical expenses [9,11].

The COVID-19 pandemic has affected social contact patterns in various situations, including in the field of medical services. Individuals with mental disorders exhibited higher levels of concern regarding COVID-19 and fear of getting infected [12], with the result that 64% of them had worsening of their symptoms, 39% were unable to receive treatment, and 38% missed their medications [13]. Subsequently, wider use of telehealth and digital tools was promoted, such as smartphone apps, in providing medical services for the reduction of person-to-person contact to lower the likelihood of COVID-19 transmission to patients [14]. Poor medication adherence was one of the major causes of the increasing relapse rate among patients with schizophrenia during the COVID-19 pandemic [15], and it is crucial to use remote monitoring to track oral medication adherence to lower relapse rates or the deterioration of psychiatric symptoms.

An increasing number of smartphone apps are being developed to care for and train patients with schizophrenia [16,17], which include apps for improving patients’ psychiatric symptoms, medication adherence, and social and cognitive functions [17-19]. Most of the apps for medication adherence in patients with schizophrenia are designed to remind patients to take their medications on time [19,20]. A smartphone app named “MedAdhere” was developed by Chen et al [21] to observe the medication adherence of these patients with recognition of the appearance of the prescribed antipsychotic drugs as well as drug-taking behavior through the smartphone camera [21].

This study aimed to explore the effectiveness of intervention with the MedAdhere app on medication adherence and accuracy in patients with schizophrenia and to measure their psychiatric symptoms and cognitive functions during the period of the COVID-19 pandemic. We hypothesized that using the MedAdhere app would improve medication adherence, psychiatric symptoms, and cognitive function in patients with schizophrenia compared to the control group.

The protocol was registered with ClinicalTrials.gov (NCT05892120) at the time of submission. We used an experimental design, using random sampling with each daycare center as a unit, and assigned patients in these centers to either the experimental or control group using a random number generator. This study is single-blinded, with research assistants being the blinded party. All participants received the usual care, while participants in the experimental group additionally downloaded and used the MedAdhere app on their personal smartphones during the 12-week study.

The sample size was calculated using G*Power (version 3.1.1; Axel Buchner) for the F₁ test and its repeated measures with an effect size of 0.25, a significance level of .05, and a power (1 – β) of 0.80; the estimated minimum sample size was 68 participants [22], with all participants being recruited from a psychiatric daycare center across 2 medical centers and 1 regional hospital.

There are various treatment modalities for patients with schizophrenia based on each patient’s psychiatric condition, including acute psychiatric units, chronic wards, daycare centers, and community services, and they may be transferred from one treatment modality to another based on their condition. Patients in daycare were transferred from acute wards where they received initial treatment, and after a period of time, they were treated well with a relatively stabilized mental status under a suitable medication and dosage. The care in the daycare center mainly focuses on regular medication and psychosocial rehabilitation. Daycare for psychiatric patients can be effective in returning the patient to a normal life in the community or relieving the strain on a family [23]. The daycare center is an important psychiatric rehabilitation unit for patients with stable psychiatric symptoms who are suitable users of this app.

A total of 105 patients participated in this study, and they were randomly assigned to the experimental or control group, although 11 patients in the experimental group did not complete this study; finally, 94 participants completed all processes of the study protocol (Multimedia Appendix 1).

The inclusion criteria were as follows: (1) patients diagnosed with schizophrenia by a psychiatrist according to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition [3]; (2) aged between 20 and 65 years; (3) able to read traditional Chinese; (4) owned a smartphone; and (5) currently admitted to a psychiatric daycare center. The exclusion criteria were as follows: patients with intellectual disability or severe cognitive function impairment (Mini-Mental State Examination [MMSE] score <17) [24].

The data were collected between June 2021 and September 2022. Demographic characteristics, including age, gender, marital status, years of education, living status, and religious beliefs, were gathered. The symptoms and cognitive functions of all participants were measured using the Positive and Negative Syndrome Scale (PANSS) and MMSE at baseline and at the end of the intervention, respectively.

Descriptive analyses by chi-square and Mann-Whitney U test were performed using SPSS software (version 21.0; SPSS Inc). Descriptive analysis was based on participants’ demographic characteristics, such as age, gender, education, marital status, and religious beliefs. After controlling the confounding factor of age, the generalized estimating equation was used to assess psychotic symptoms (positive, negative, and general psychopathological symptoms) and cognitive functions (orientation, memory, attention and calculation, language, reasoning and problem-solving, and executive function) from baseline to the end of intervention in the 2 groups. All tests were 2-tailed, and P<.05 was considered significant.

This study was approved by the institutional review board of the Kaohsiung Medical University Hospital: KMUHIRB-SV (I)–20200096. The principal investigator visited each psychiatric daycare center, explained the study’s purpose and process to all participants, and inquired regarding their willingness to participate. Written informed consent was obtained from all participants, and they were informed that they could withdraw from the study at any time without providing a reason. Each participant was offered NT $400 (US $13-14) as compensation for their participation.

All participants had access to their private mobile networks (4G LTE). A total of 105 participants were randomly assigned to the experimental or control group, and 94 completed all the processes of our protocol. The demographic characteristics of the respondents are presented in Table 1.

The medication adherence rates were 94.72% (2785/2940; total doses taken by all patients over 12 weeks / [35 person ×12 weeks × 7 days × doses] × 100%) and 64.43% (3193/4956; total doses taken by all patients over 12 weeks / [59 person ×12 weeks × 7 days × doses] × 100%) in the experimental group and the control group, respectively. The medication adherence rate of the experimental group was higher than the control group during the intervention (Table 2).

The changes in psychiatric symptoms and cognitive functions in the 2 groups from baseline to the end of the intervention in 12 weeks are shown in Tables 3 and 4.

After controlling the confounding factor of age, we found significant improvements in positive, negative, and general psychopathology symptoms in the experimental group, with a decrease of 0.143 in the positive (P=.02), 0.137 in the negative (P=.007), and 0.013 in the general psychopathological symptoms (P<.001) following an 8-week intervention with the MedAdhere app (Table 3). Additionally, there were significant improvements in the cognitive domains of memory, language, and executive function in the experimental group, with an increase of 1.51 in language (P=.012) and 1.39 in executive function (P<.001) after intervention with the MedAdhere app for 12 weeks (T2; Table 4).

The policy of the government in Taiwan during the COVID-19 period was to reduce interpersonal contact against the transmission of coronavirus. In addition to the medication adherence of patients, this app was designed to reduce interpersonal contacts during the COVID-19 pandemic and to reduce the burden on health care providers.

In this study, the MedAdhere app contributed to the medication adherence rate during the intervention, and this study also found that patients’ positive and negative symptoms and general psychopathology symptoms significantly improved after the intervention, even though the experimental group had more severe general psychopathology symptoms than the control group at baseline. In addition, there were significant differences in cognitive functions between the 2 groups.

The estimated medication adherence rate was 50% among patients with schizophrenia who did not receive any intervention [29]. Another study reported an adherence rate of 65% with different interventions for the same purpose of enhancing medication adherence in patients [30]. A higher medication adherence rate essentially contributes to stabilizing the psychiatric symptoms of patients and lowering their readmission rates [31]. The medication adherence rate in the experimental group was 94.72% during the intervention using the MedAdhere app, higher than the control group’s 64.43% in this study. Our results suggest that the utility of the MedAdhere app for artificial intelligence verification of medication ingestion behavior plays an important role in enhancing patients’ medication adherence, and this is an easily implemented intervention [32].

Our results showed that intervention with the MedAdhere app could significantly improve patients’ positive and negative symptoms in 12 weeks compared to the control group. This result was similar to that of a previous study showing that using the medication adherence app could improve positive and negative symptoms in patients with schizophrenia [33].

Negative symptoms should be treated by other rehabilitation modalities to improve pleasure, motivation, and cognitive skills relevant to apathy and anhedonia [34,35]. In addition to verifying medication ingestion behavior using the MedAdhere app, patients were able to ask questions and interact with researchers on the web during the intervention through the Line app, which is one of the related links on the MedAdhere app’s home page. Interactions between researchers and patients can be regarded as a form of social skill training that has been documented to have medium effects on reducing negative symptoms in individuals with schizophrenia [36], and such interactions might be one of the reasons for alleviating our participants’ negative symptoms of social withdrawal, blunted affect, emotional withdrawal, and a lack of spontaneity and flow of conversation.

Cognitive functions declined in the control group, whereas the cognitive functions of the experimental group improved significantly in the domains of memory, language, and executive function after the 12-week intervention. This could be attributed to the use of the MedAdhere app with novel technology, which provides repeated multisensory stimuli and rewards after each medication ingestion, which in turn enhances their motivation to continue using the app.

Cognitive function in patients with schizophrenia can be improved through exercise [37], quality sleep [38], and symptomatic stability [39]. Our results may have been affected by the patient’s sleep patterns [38] and symptomatic stability [39], both of which are considered to be improved by antipsychotics. Compared to conventional methods for medication adherence that require routine monitoring by family members, telephone visits by home health care providers, or monthly visits by clinicians, this innovative MedAdhere app effectively empowers patients’ medication adherence and reduces the burden on health care providers. Such an artificial intelligence-assisted adherence-enhancing app can be extended to all patients who require regular medication but may forget to take their medication on time, for the benefits of better adherence and to prevent the intake of the wrong medication or duplicating doses.

The MedAdhere app requires minimal training for implementation. To ensure that patients can continue using this app, we initially designed the app interface to be user-friendly and easy to operate, with the added incentive of earning a digital gold coin upon completing medication ingestion as a reward. In addition, we have established alliances with clinical professionals and family members, providing the app and cloud access to health care teams and allowing them to keep monitoring patients’ medication adherence. With the support of patients’ families, each participant can keep using this MedAdhere app after the intervention.

This study had some limitations. First, there might be selection bias because the MedAdhere app can only be installed on the Android platform. Patients who did not own a smartphone or whose smartphones were not Android system based were excluded from this study. Second, this was a single-blinded design with a 12-week study duration, and a double-blinded study design with long-term follow-up is indicated to reduce potential bias and confirm the effects of the MedAdhere app on improving psychiatric symptoms and cognitive functions in patients with schizophrenia. Third, we did not calculate how many times or the frequency of communication between participants and our team. In the future, it will be possible to calculate the number of bidirectional interactions. Fourth, the MedAdhere app has only been used in Taiwan thus far and can be translated into other languages for use in other countries in the future. Finally, there were other psychotropic medications prescribed for our participants, and these adjunctive drugs were not included nor identified by our MedAdhere app.

The novelty of this study is the development of the MedAdhere app using a smartphone’s camera to recognize patients’ faces, drugs, and medication-taking behavior. This accessible tool has the potential to significantly improve medication adherence in patients with schizophrenia. The additional value of this MedAdhere app is to assist associated health care providers or partners’ families in confirming patients’ medication status to reduce the risk of relapse or repeated hospitalization.