The introduction of antipsychotics has led to major breakthroughs in the treatment of schizophrenia, but 30% to 60% of patients still lack or only partially respond to adequate drug therapy [1]. Data show that even among patients with first-episode schizophrenia, about 14% have difficulty achieving the desired outcome [1]. It was found that up to 60% of patients with refractory schizophrenia who were treated with standard antipsychotics for a full time and full dose without significant effect and with severe disease were even treated [2]. Currently, the treatment of refractory patients has become one of the popular research topics in the field of schizophrenia therapeutics. As the name implies, refractory treatment refers to the failure to achieve the desired outcome with treatment according to the generic approach. In the field of clinical psychopharmacology, refractory treatment remains a rather ambiguous concept, due to the fact that antipsychotic treatment follows the principle of individualization. Is there a universal approach that fits all patients? What kind of efficacy is achieved to be considered ideal? Different people have different interpretations. Refractory schizophrenia (treatment-resistant schizophrenia) has been well documented in China [2, 3], but unfortunately, there is still no widely accepted and well-operated definition. Although this phenomenon cannot simply be considered a bad thing, it has many adverse effects from the perspective of clinical research. Chief among them is the lack of good comparability between study data due to the different understanding of refractory treatment by different investigators. Thus, a good operational definition of refractory treatment is of great importance to facilitate the communication of research results. According to Morrison (1996), refractory treatment refers to the failure to achieve satisfactory results in patients with a correct diagnosis who are treated with different types of antipsychotics at full doses and courses of treatment by different routes of administration [2]. In contrast, the US Food and Drug Administration (FDA) defines refractory as a patient with severe schizophrenia who has been treated with an adequate course of standard antipsychotics without significant effect [2]. However, the question of how large a dose and how long a course of treatment is considered adequate and what kind of efficacy is satisfactory cannot be found in these descriptions. Recently, there appear to be some operational and relatively good definitions of refractory schizophrenia.Juarez-Reyes et al. (1997) defined refractory patients in a flow-through study to include patients who received at least two antipsychotics equivalent to at least 600 mg of chlorpromazine daily for at least 4 weeks of treatment that did not work; patients with TD [2].Kane et al. (1988, 1992. 1995) defined refractory cases to include: poor response to treatment with 3 antipsychotics of appropriate dose and duration (at least 2 of the 3 drugs are chemically different) in the last 5 years; patients who cannot tolerate the side effects of antipsychotics; and patients who relapse or worsen even with adequate maintenance or prophylactic treatment [1, 4, 5]. Clearly, to define refractory treatment, at least the following 4 questions must be determined: (1) How long does medication actually need to be maintained before a course of treatment is considered adequate? (2) What dose is used, i.e., what constitutes an adequate dose? (3) How many drugs have been used in the past? (4) What criteria are used to judge the efficacy, i.e., what is considered satisfactory? First of all, how long does a drug treatment need to be maintained before it is considered adequate? From the perspective of clinical treatment, to determine whether a particular antipsychotic is effective, a general impression can be obtained by observing it for about 6 weeks. However, if the highest and final level of treatment response is to be observed, 6 weeks would appear to be insufficient. It was observed that the median time required for complete elimination of psychiatric symptoms by medication was 11 weeks, while the average amounted to 35 weeks [1]. Apparently, the lack of adequate response of patients to medication during the initial 6 weeks of treatment does not mean that continued adherence to treatment will not have the desired effect. Thus, to determine a patient’s lack of therapeutic response to a drug, the duration of treatment should generally not be less than 12 treatment weeks (Liu Tie-bong et al., 1994) [6]. Second, what is the adequate dose of a drug? There is no certainty. Currently, the average daily therapeutic doses of antipsychotics such as chlorpromazine, clozapine, and haloperidol are gradually decreasing in psychiatric clinics. In general, if there is some degree of improvement with previous medication, it is expected that a better response may be obtained by increasing the therapeutic dose; whereas for those who have responded poorly or not to medication in the past, there is no advantage of high dose treatment. Several studies have shown [1] that high doses of typical antipsychotics are not superior to conventional doses; however, despite this, some clinical psychiatrists still prefer to increase the dose further if conventional doses are ineffective. We tend to believe that two indicators can be combined to determine whether an antipsychotic is adequate: whether it is an effective blood level and whether it is a regular effective therapeutic dose, without overemphasizing whether it has ever received an over-regular dose. Third, how many antipsychotic drugs lack therapeutic response to be called refractory? There is also no consensus. Theoretically, there are differences in the drug action characteristics of different types of typical antipsychotic drugs; lack of response to one drug does not preclude a therapeutic response after switching to another class of drugs, and most clinicians have similar practical experience. However, it has been found that a patient’s response to one typical antipsychotic is broadly predictive of his response to other typical drugs; patients who do not respond to one typical antipsychotic often lack a therapeutic response to another typical drug as well. Therefore, some scholars doubt the effectiveness of switching to other therapeutic drugs [2]. However, there is no consensus on whether the lack of therapeutic response to a typical drug can be considered refractory, and most scholars disagree with this view. We believe that, from the perspective of clinical practice, instead of observing one drug for 12 weeks and multiple drugs consecutively before finally deciding whether it is refractory, it is better to systematically observe two drugs with different chemical structures and treat them as refractory cases if they are ineffective. This is because this treatment appears more aggressive and allows for a quicker possible benefit. Finally, the criteria used to rate the effectiveness of the treatment is also an important parameter to determine whether it is refractory or not. The current common practice is to use psychopathological indicators, such as the use of psychiatric symptom scales for rating before and after treatment, and to observe whether a statistically significant improvement in psychiatric symptom scale scores occurs after a certain period of treatment; or to use the reduction rate as an indicator. The reduction rate is calculated by dividing the difference between the pre- and post-treatment scale scores by the pre-treatment scale score, and the resulting quotient is multiplied by 100%. In a study by Kane et al. (1988) [4], validity was defined as a 20% or greater reduction in BPRS score after treatment, a clinical global impression (CGI) scale less than or equal to mild, or a BPRS score ≤35 points. Recently, some authors (Bondolfi et al., 1998) [8] used the efficacy index (effect size) as an index to assess efficacy, which is calculated by taking the difference between the pre- and post-treatment scale scores as a divisor and dividing it by the standard deviation of that difference. It is generally accepted [8] that if the efficacy index is less than 0.2 it should be considered ineffective, 0.2 to 0.5 is minimally effective, 0.5 to 0.8 is moderately effective, and more than 0.8 is strongly effective. In the above-mentioned studies in which improvement in psychiatric symptoms was used as an indicator of efficacy, the severity of symptoms at the beginning of treatment had a significant impact on the judgment of symptom improvement, and it was clear that different observers had different interpretations of whether a reduction in BPRS total score from 70 to 56 (20% reduction) was equivalent to a reduction in total score from 40 to 32 (also 20% reduction). Special attention should be paid to the improvement of psychotic or negative symptoms when observing the efficacy. If the improvement in psychiatric symptom scale scores is minimal, but meaningful improvements in subjective health status, ability to care for oneself, psychosocial functioning, or effort to participate in treatment occur, the therapist should not ignore their potential clinical importance [5]. Thus, examining the efficacy of treatment based solely on the degree of improvement in psychopathological symptoms is inevitably biased. We believe that a comprehensive efficacy assessment should include information on psychopathology, social functioning and quality of life, medication adherence, subjective patient experience, and risk/benefit ratio. In terms of treatment it is generally accepted [1] that clozapine is the current antipsychotic with a positive effect in refractory schizophrenia. Data suggest that clozapine can result in significant clinical improvement in 30% to 60% of patients who have failed previous treatment [4, 5, 9, 10]. The outstanding advantage of clozapine is that it is less likely to produce extrapyramidal side effects, but there is a risk of causing granulocytopenia. Data from the United States show that the cumulative incidence of granulocytopenia after 1 year of clozapine treatment is about 0.8%, a risk that prevents the widespread use of clozapine [5]. To date, the effectiveness of risperidone in refractory schizophrenia has not been as clearly established as that of clozapine. However, for those refractory cases before applying clozapine treatment, it is advisable to give risperidone with a better safety profile and fewer side effects; if risperidone is not effective, then switch to clozapine. However, a comparative study found that when clozapine was used instead of risperidone, the effective rate was up to 40%, while when clozapine was used instead of risperidone, the effective rate was only 15%. It is generally believed that the most appropriate time to evaluate the efficacy of risperidone is 6 to 8 weeks after the drug is administered. A small number of refractory cases may show efficacy after several months of risperidone use; therefore, the time to observe efficacy should be extended appropriately. It can be expected that the problem of treatment of refractory schizophrenia will be a major issue for psychopharmacologists and clinical psychiatrists at present and for quite some time in the future.What is refractory? Different researchers have different understandings, and in order to facilitate academic exchange, it is necessary to develop a well-operated diagnostic criterion. To develop such criteria, there must be agreement on, at a minimum, how long treatment must be maintained (what is considered adequate), what dosage is used (what is considered adequate), how many medications have been used, and what criteria are used to judge efficacy (what is considered satisfactory).