This is a chapter from Design and Implementation of N-of-1 Trials: A User’s Guide. The full report can be downloaded from the Overview page.
Traditional research approaches do not always provide the type of personalized evidence necessary for patients to make fully informed decisions. In the absence of individualizable evidence, defined as evidence that is directly applicable to an individual patient, clinicians and patients use imperfect approaches to personalize care. Since it is often difficult to apply evidence from randomized controlled trials (RCTs) directly to the care of individual patients, clinicians use clinical judgment to decide whether trial results are applicable to an individual patient. When treating chronic or intermittently symptomatic conditions, they may conduct informal trials of therapy whereby a clinician will try an intervention for a particular patient to “see if it works.”1,2 Many interventions of interest to patients, such as those that facilitate the management of side effects or those addressing lifestyle changes, have never been tested by RCTs, making evidence-based decisions among these alternatives impossible. Accordingly, patients often use self-experimentation, whereby various interventions are tested in an ad hoc fashion without any underlying scientific rigor.3,4 It is likely that the informal approaches being used by clinicians and patients to personalize care are flawed and can be improved.5
N-of-1 trials are a useful method to personalize care but are not widely used (see Chapter 1). While n-of-1 trials were introduced to the medical community in the late 1980s and early 1990s, inspiring the creation of n-of-1 trial services at several academic medical centers,6,7 the burden of conducting the trials eventually led to a drop-off in popularity. Recently, however, there has been a resurgence of interest in n-of-1 trials, both independent of and in cooperation with n-of-1 consult services.2,8 However, this resurgence may be doomed to fail if specific attention is not paid to engaging, training, and supporting patients and providers interested in conducting n-of-1 trials. This includes addressing problems encountered by both patients and clinicians as well as designing systems that support training and execution of trials with fewer burdens.
This chapter will discuss methods of increasing patient and provider engagement, training, and support, with the aim of presenting a framework that will facilitate the ease and conduct of n-of-1 trials in a sustainable way.
Eliciting interest, understanding, and cooperation from patients is of utmost importance when conducting n-of-1 trials. According to Guyatt,9 "an n-of-1 RCT is indicated only when patients can fully understand the experiment and are enthusiastic about participating." Generating buy-in from patients for n-of-1 trial participation occurs both through initial engagement (approaching the patient and demonstrating the benefits of the trial) and through ongoing training and support (helping patients to understand the disease, trial procedures, data collection, analysis, and decisionmaking).
Initial Engagement and Motivation
When approaching patients to discuss initial participation in an n-of-1 trial, it is important to recognize that patients who are good candidates for an n-of-1 trial are those whose disease is easily monitored, those with disease that has been unresponsive to standard therapy, and/or those patients who are quick to respond to treatment (please also see the section “Indications, Contraindications, and Limitations” in Chapter 1).10 It is not important that the patient meet traditional inclusion criteria for a parallel group trial; patients who are older, have comorbidities, or have lower levels of education or income are all potential candidates for n-of-1 studies.11 However, some patient characteristics have been shown to be associated with successful trial participation, including a positive attitude, high level of motivation, intact cognitive capacity, and willingness to be proactive following poor treatment results.10 In addition, patients who are responsible and open to both novel therapies and experimentation make ideal candidates.10 Patients must be willing to undergo multiple treatment periods and, if important to trial design, to take a blinded medication.6 Finally, patient recruitment works best when there is a strong and trusting relationship between the provider and patient.10 Some patients may be unwilling to participate in an n-of-1 trial if they believe the results may lead to a recommendation to discontinue a medication they believe is effective, and this should also be addressed at the time of recruitment.12 It is important that patients have a genuine uncertainty (equipoise) regarding which treatment is superior and a willingness to use the information from the n-of-1 trial to determine future treatment. Patients who do not express this willingness may not be suitable for an n-of-1 trial.
Partnering with patients and agreeing that the benefits of the n-of-1 trial outweigh the burdens are crucial. The following benefits of the n-of-1 trial should be highlighted when engaging patients:
Patients who participate in an n-of-1 trial will learn more about their own disease process and treatment than patients who receive usual care or patients who participate in parallel group trials. When patients participate in n-of-1 trials they undergo a rigorous and personalized process utilizing careful monitoring and frequent outcome reporting that can offer unique insight into their disease process10 and patterns of symptoms related to daily activities.13 Finally, patients will learn which treatment or drug dose maximizes benefits while minimizing adverse side effects (recognizing that treatments may affect individuals differently). Furthermore, they will be able to apply that information immediately toward their own treatment decisions, and also contribute that information to the general pool of scientific knowledge about that disease, potentially helping others.13
Unlike traditional parallel or crossover group trials, n-of-1 trials are tailored to the individual circumstances of the patient. First, the intervention tested using n-of-1 methods can be individualized for a particular patient. For trials of medication effectiveness, patients often receive individualized dosing regimens.8 In addition, n-of-1 methods can be used to test interventions tailored to the patient’s interests outside of traditional medication effectiveness, including trials of behavioral therapy or complementary and alternative medicine (CAM).13 Because they have the potential for more side effects than behavioral or CAM trials, drug trials might run into more resistance from patients;13 however, the n-of-1 method can be used effectively to study both medications and lifestyle modifications, offering maximal flexibility. Second, there is flexibility in the ways outcomes are selected and measured. Patients participating in n-of-1 trials can choose the outcomes that are most important to them, and the manner in which an outcome is assessed.6 Guyatt recommends measuring a patient’s symptoms or quality of life directly,9 and all outcomes should be patient centered.10 Data can be collected via self-administered questionnaires with Likert scales, daily diaries, and various other formats (see Chapters 1, 4, and 5). In particular, daily diaries have been shown to be highly successful with infrequent missing data.13
Low risk to participation
N-of-1 trials pose low risks to patients, since the patients are already likely to be familiar with the treatments from prior use.13 Furthermore, patients can withdraw at any point if they feel the trial is clearly ineffective or not beneficial, or if the treatment leads to undesirable side effects.
Increased collaboration between providers and patients
Patients participating in n-of-1 trials require increased monitoring and may have more appointments than other patients. This may lead to increased communication between patient and provider, ultimately better supporting shared decisionmaking.10 This heightened communication may not only increase adherence10 but also afford the patient greater autonomy than is typically observed in clinical practice.
N-of-1 trials have been successful in the past
Prior consult services have reported between 62 percent14 and 84 percent7 of trials providing a definite clinical answer, and 79 percent of patients participating in n-of-1 trials considered them useful.6 Between 44 percent and 65 percent of patients8,14,15 reported treatment change as a result of the trial, with between 84 percent and 100 percent11,14,15 continuing with therapy consistent with definitive n-of-1 trial results. However, the high level of treatment continuation may not be universal across all patients and treatments.
Information about the public’s motivation and interest in n-of-1 trials has been gleaned from a small number of peer-reviewed publications that conducted interviews and focus groups with patients.
Further exploration of social marketing methods, patient focus groups, and patient communities that are engaged in e-science (e.g., PatientsLikeMe, CureTogether, and DIY Genomics) may assist researchers in designing better strategies and approaches to engage patients.
In addition to approaching patients through providers, direct social marketing may have a useful role in patient recruitment. Patients outside the clinic may be interested in n-of-1 trials for many of the same reasons as patients in a clinical setting: there is value in knowing that the benefits of a particular medicine are “worth it” compared with the costs and side effects. This knowledge becomes even more valuable as patients become responsible for greater and greater portions of drug costs out of pocket. If combined with other campaigns (such as health literacy), this kind of social marketing could be even more successful than going through individual physicians.1 In Australia, Nikles and colleagues have shown the benefits of utilizing a central administrative support structure to reach an entire country.15 They used mainly print, TV, and radio campaigns for reaching patients, although other possibilities could include support groups, brochures in doctors’ waiting rooms, and Web sites.8 In the Australian studies, patients were able to contact the consult service directly and then provide their physicians with trial materials, including packets of medications and instructions.15,16 It is also possible for patients to participate in n-of-1 trials completely independent of providers, although we recommend this only when the n-of-1 trial does not involve a prescription drug. In all instances, patients should discuss the results of the n-of-1 trial with their physician.
Ongoing Training and Support
Because most patients will not be familiar with the n-of-1 trial approach, researchers hoping to engage and support patients must offer education on all aspects of such trials. First and foremost, patients should have a basic understanding regarding their disease or condition as part of good clinical care. Patients will also need education and training in n-of-1 study design, especially since decisions regarding therapies, outcomes, and stopping rules are often made collaboratively among researchers, providers, and patients.10 In patient interviews, Brookes et al.13 found that patients see n-of-1 trials as being similar to the self-experimentation that occurs in everyday life. One patient commented, “Well yeah, well it’s just like anything isn’t it? You try cabbage, you don’t like that, so you try broccoli.”13 Other patients found n-of-1 trials similar to conventional RCTs, but preferred n-of-1, since they would be more likely to receive two active treatments as opposed to a placebo.10,13 Furthermore, emphasizing that variation exists between individuals and that making comparisons against oneself as opposed to others is more likely to result in a true answer for an individual patient will further differentiate the n-of-1 from more traditional trial designs.13
Patients are also likely to have concerns regarding the potential hazards or consequences of the trial, and these should be addressed when discussing study design. Specifically, patients may be worried about adverse drug interactions, possible suboptimal treatment for some period of time, and whether the medication will be prepared in a safe and efficient manner.10 Finally, researcher, provider, and patient should discuss the interpretation of results prior to beginning the trial. Patients may have a desire to continue a medication despite unclear trial results or results clearly in favor of an alternative medication. Options regarding treatment after trial completion should include treatment cessation, treatment continuation, and (when the trial results are unclear) the extension of the n-of-1 trial to include more crossover periods in order to minimize uncertainty.17
As previously stated, one of the most important advantages of n-of-1 trials is that patients have the ability to tailor the trial to their needs and can address the outcomes (e.g., symptoms or predictors of future health) that are most important to them. Patients must be supported in identifying and defining outcome measures that clearly determine effectiveness of an intervention. Patients may be interested in exploring outcomes that have been established as effective for other patients, or they may prefer to explore outcomes on which there is little information in the literature. Regardless, researchers must emphasize the importance of consistency and accuracy in data reporting irrespective of the type of outcome measure used. As discussed in Chapters 1, 4, and 5, possible outcome measures may include quality-of-life assessments, symptom diaries, or objective outcomes (such as blood pressure measures). It is recommended that any measurement assessment be brief and easy; prior research has shown success with daily diaries.13 Recording these outcomes can be facilitated through Web sites and other electronic technology. Some examples include using smartphones, tablet computers, and personal digital assistants (PDAs) to collect and transmit the data (mHealth), or utilizing devices and sensors (e.g., a cell phone’s accelerometer) to capture data passively.18 The role of these devices and other technology is discussed in more detail in Chapter 5. Patients must also be supported in identifying appropriate study questions. For example, it is possible that patients may prefer trials of medical devices as opposed to drug trials due to potential side effects and discomfort.13 Furthermore, patients are unlikely to agree to participate in a lot of experiments that do not show the clear clinical benefit of one treatment versus another, so it is important to work with patients to identify possible interventions based on existing scientific evidence, the individual’s prior treatment history, knowledge gained from other individual n-of-1 trials, and even anecdotal experience.
Patients will need support and training in completing the n-of-1 trial and interpreting the results. It is possible that providing support via ongoing phone, Short Message Service (SMS), and/or email from the provider or researcher may minimize dropouts and encourage adherence to data collection and trial protocols. In discussing potential trial results, the difference between clinical significance and statistical significance should be explained. As discussed in Chapter 4, significance testing might be less pertinent for n-of-1 trials; instead, the statistical methods for n-of-1 trials should provide the decisionmaker with all necessary information in a format that facilitates decisionmaking. Some experienced researchers6,7,19 advocate using an a priori difference in outcomes as indication of clinical significance, in lieu of the exclusive use of statistical significance for decisionmaking. For example, they considered a 0.5 mean difference on a 5-point Likert scale to indicate effectiveness, since that corresponds to a meaningful improvement in well-being.20 While we agree with the principle of emphasizing clinical significance in decisionmaking, we would like to note that it is also important to take uncertainty into consideration, to ensure that the observed outcome difference is reliable enough for decisionmaking.
Trial results should be discussed with the patient, and decisions regarding future treatment should take these results into account. There are numerous ways to present trial results to patients, and it is important to recognize that all patients will not benefit equally from the same presentation method. Providing the same information both graphically and numerically will allow the patient to explore the results in the manner that is most meaningful to him/her.21 The researcher should always be responsive to the patient’s needs, preferences, and goals, and promote shared decisionmaking. At the same time, the cumulative gain of knowledge from participating in n-of-1 trials can be emphasized. For example, even if a small effect is seen with one intervention, this may be a stepping stone to greater improvement. Scenarios for how to handle ambiguous results or results that do not favor treatment continuation should have been discussed during the trial planning phase. Please see Chapter 4 for more detailed information regarding the analysis and presentation of n-of-1 trial results.
Finally, all support and training must be provided in a way that minimizes the time commitment necessary for a patient to participate in an n-of-1 trial. The demands on a patient’s time need to be realistic.18
Though researchers interested in conducting n-of-1 studies may interact directly with patients in conducting trials, researchers will often be partnering with both patients and providers in experimentation. Historically clinicians have played a central role in the execution of n-of-1 trials, either by carrying out the n-of-1 trial directly or through working with patients to interpret and implement trial results in cases where the trial was conducted by an n-of-1 consult service (either locally or remotely).9 Therefore, it is important that researchers address the needs of the clinician when designing systems to support n-of-1 trials.
In order to successfully execute n-of-1 trials at any scale, researchers must convince providers that the benefits outweigh the inconvenience. In addition, researchers must provide adequate training and support to integrate n-of-1 trials into the clinical workflow at the lowest possible transactional costs for providers.
Engagement and Motivation
When attempting to engage clinicians in conducting n-of-1 trials, researchers should target clinics and specialties that will find them the most useful. Traditionally, primary care providers in fields such as family medicine, general internal medicine, and pediatrics as well as specialty clinicians who manage patients with chronic diseases have participated in n-of-1 trials.6,10 However, researchers are likely to be successful engaging providers from a broader range of fields as long as those clinicians feel that they have a management problem that needs to be solved and that n-of-1 methods will help in that solution (see Chapter 1).
The key to motivating clinicians to participate in n-of-1 trials is to persuade them of the benefits. Researchers are encouraged to highlight the following advantages of n-of-1 trials:
Enables truly personalized evidence-based medicine (EBM)
According to the Institute of Medicine, the practice of EBM means that to the greatest extent possible, health care decisions are grounded on a reliable evidence base, account for individual variation in patient needs, and support creation of new knowledge regarding clinical effectiveness.22 N-of-1 trials have the potential to provide the highest strength of evidence for making individualized treatment decisions, in that they provide truly personalized evidence of clinical effectiveness.10,23 While clinical guidelines and standardized care algorithms have emerged as methods to facilitate consistent application of evidence in practice and to eliminate variation among providers, these shared baselines are still meant to allow for variation to accommodate differences in patient needs and preferences.24 N-of-1 trials provide a means for supporting such expected individual variation with sound evidence.
Improves relationships between patients and providers
Engaging in an n-of-1 trial facilitates collaboration between patients and providers. N-of-1 trials allow patients to participate more comprehensively in their own care—promoting self-management, greater insight into the disease, and personal engagement in their own health.4,13 N-of-1 trials increase communication between the patient and the clinician and support shared decisionmaking above and beyond what traditionally exists in current care models.10 The process of custom designing n-of-1 trials—selecting interventions, determining which outcomes to measure, specifying stopping rules, and agreeing on the desired effect sizes—establishes a genuine two-way partnership between patients and providers.10 In addition, n-of-1 trials can be beneficial when there is disagreement between patient and provider regarding the best approach to treatment.6,25
Provides more precise answers about how to select among treatment options
N-of-1 trials increase the precision of clinical decisionmaking in a number of ways. Current methods of trial-and-error prescribing among clinicians and self-experimentation among patients have little rigor and may provide misleading results. In most instances, insufficient data are collected to provide clear evidence of effectiveness, and even when data are collected, the apparent effectiveness of a treatment in the short term may only be the result of random variation in the patient’s symptoms or the effects of uncontrolled external factors. Additionally, treatments that initially produce subtle improvements may be abandoned before their efficacy is ever appreciated. The more comprehensive, concrete, personalized information that surfaces from n-of-1 trials (e.g., from daily symptom diaries) provides a better understanding of symptom patterns and frequency that allows for deeper insight into the condition and overall better management.13 By making clinical uncertainty explicit and using a rigorous design that includes randomization or counterbalancing, multiple crossover treatment periods, systematic outcome assessment, and blinding (if possible), n-of-1 trials enable providers to make informed decisions about the effects of various treatments in a way that reduces cognitive bias, one of the main threats of informal experimentation.10 Studies of n-of-1 trials in medicine show increased provider confidence in their treatment decisions. In one of the original series of n-of-1 trials published by Guyatt et al., 84 percent of completed trials provided a definitive clinical answer, and physicians reported a high level of confidence in their treatment decisions in over 80 percent of trials.7
A range of approaches can be used to market these benefits of n-of-1 trials to providers. Strategies that have previously been used to promote n-of-1 trials among care providers include newsletters, professional media, Web sites, and presentations at clinical meetings.8 Engaging a local physician champion within the clinic or unit may also be a helpful strategy for engaging other clinicians. However, researchers are encouraged to think broadly and creatively when identifying forums and methods to engage with care providers around n-of-1 trials.
Training and Support
Although most clinicians have had formal exposure to the concepts of EBM as well as research and trial design, many providers may not be familiar with n-of-1 methods. Researchers who hope to engage providers in n-of-1 trials need to educate clinicians about their basic features, emphasizing the validity and safety of the approach as well as specific issues around analysis such as display of data and how patients and clinicians determine whether a particular intervention has resulted in improvement (see Chapter 4).10 Providers may also be particularly interested in more generalizable results, which can be achieved by aggregating data across n-of-1 trials (see Chapter 4 for additional information about aggregating n-of-1 trials). Researchers who wish to create large-scale n-of-1 trial systems should consider developing a scalable n-of-1 curriculum (e.g., online tutorials) to educate providers and other potential end-users about the basics of n-of-1 methods.
In addition to providing adequate training, researchers must also offer tools to support the shared decisionmaking that is central to n-of-1 trials. It is also important for researchers to support the major paradigm shift that accompanies n-of-1 trials in the current care delivery systems, including intensive patient-provider collaboration, explicit recognition of clinical uncertainty, and a more formal structure for experimentation beyond current models of informal experimentation.10 In addition, researchers should help clinicians engage more intensively with patients, particularly with the growing number of self-experimenters (e.g., the Quantified Self community) who have developed their own expertise in data tracking and hypothesis testing.4 Development of tools to support shared decisionmaking in n-of-1 trials is an area in need of additional research.
Researchers must also provide support to providers to allow them to conduct n-of-1 trials effectively within the setting of their typical clinical workflow and with minimal demands on time; setting realistic expectations regarding time and resources is also important. This support should be flexible enough to cover a range of engagement needs, from those clinicians who would like a great deal of autonomy and flexibility in designing and conducting trials to those who want a more prescriptive approach. Less intensive support could be provided by researchers in the form of a tutorial service that serves mainly educational needs.7 More intensive support to guide clinicians through study design and execution, including identifying a study question, selecting outcome measures, designing the trial, and analyzing the data, could be provided in numerous ways; however, to date, this type of support has typically been provided by n-of-1 consult services.6,9 For example, the n-of-1 consult service described by Larson et al. consisted of a core research group including a general internist, clinical pharmacist, family practitioner, and biostatistician. This service provided the key support functions for n-of-1 trials, including assistance with randomization and blinding as necessary, as well as with Institutional Review Board application preparation (see Chapter 2), general study design, and analysis. Having been offered this type of support, 85 percent of physicians indicated they experienced little or no inconvenience in referring patients to the n-of-1 service, and 77 percent reported that they spent no extra time or effort on their patients’ participation in the trials.6 Nikles et al. have also published their experience with a successful n-of-1 consult service delivered remotely across Australia.8 This service consisted of centralized administrative support facilitated by use of mail, telephone, and electronic communication as well as standardized kits containing all the necessary supplies and information for an n-of-1 trial, including randomized doses, symptom diaries, and instructions. Kits were mailed directly to treating physicians, who were provided with trial results at the completion of the trial.8,15 Nikles et al. found that >80 percent of physicians reported that they would order more n-of-1 trial kits and believed that n-of-1 trials were useful and worth the time commitment.8 There has also been one model of a commercial n-of-1 trial service called Opt-e-Script. Although this venture was unsuccessful, it used the same combination of prefabricated kits (including blinded treatments and questionnaires) and analytic support.1 The key to these successful n-of-1 consult services has been dynamic leadership, a multidisciplinary team, and a focused investment of resources.1 While support for conducting n-of-1 trials has typically been offered through consult services, researchers are encouraged to identify other (perhaps more sustainable) ways of providing similar support to providers who are interested in executing n-of-1 trials.
The support offered by researchers to enable providers to conduct n-of-1 trials must address what has persistently been one of the most prominent barriers—increased time demands associated with n-of-1 trials. When discussing barriers to n-of-1 trials, physicians have mainly reported logistical concerns, particularly the administrative time demands in addition to time already spent on patient care.10 Time-tracking data from some of the original n-of-1 consult services revealed that an average of 16.75 hours was spent on any one individual trial (total of the time spent by the entire team), half of which was spent on trial preparation.6 Though this is not an estimate of the time required from an individual provider, and it is likely an overestimate for today’s researcher, it underscores the need to develop systems that take advantage of new technologies to conduct n-of-1 trials in a way that minimizes the time required to participate (see Chapter 5). In addition, n-of-1 trials represent a different way of engaging and collaborating with patients—an approach that is not currently easily supported within the setting of the traditional clinical encounter. In fact, time-tracking data reflect that only one-third of the time spent on n-of-1 trials involved patient visits.6 Researchers need to develop methods to support other types of patient-provider encounters that may result from engagement in design and execution of an n-of-1 trial, including out-of-office encounters and email correspondence. Creating methods that allow for compensation of provider time spent conducting n-of-1 trials is another important aspect of reducing transactional barriers that could potentially be achieved through Medicare or traditional insurance reimbursement (see Chapter 3 for additional details on financing of n-of-1 trials).
Development of new strategies and tools to educate and support clinicians in the design, execution, and analysis in n-of-1 trials that are fast, flexible, and inexpensive is one of the key areas of research necessary to move n-of-1 trials further into mainstream clinical practice.
Collaboration Among Users
Researchers must also design systems to allow for active collaboration among all types of users, including providers, patients, and researchers. This type of active collaboration across user groups is necessary to optimally facilitate engagement and provide adequate support for users. Collaboration among patients (e.g., through online communities) can allow them to learn from the experiences of other patients and provide opportunities to improve both the tools used to conduct n-of-1 trials (outcome measures, tracking tools, study protocols, etc.) as well as their own decisionmaking by incorporating results of prior trials either informally through review of available records, or formally through Bayesian analyses. Collaboration between various provider types such as physicians, pharmacists, dietitians, and psychologists is necessary to fully execute all aspects of an n-of-1 trial. For example, pharmacists can assist with packaging blinded medications, preparing randomization schedules, and providing information on the time of onset to action and washout periods of a particular drug in order to inform study design.9 Psychologists can help researchers with aspects from patient adherence to data reporting and also with interventions targeting behavioral modifications.
Collaboration among user groups can bridge the gap between independent efforts by patients to engage in self-experimentation and efforts by researchers and health care providers to generate and apply evidence within the traditional health care delivery system. Examples of leveraging new types of technology-enabled collaborations across user groups to advance research, and useful models for researchers interested in advancing n-of-1 methods, include DIYGenomics, which is a not-for-profit research organization that focuses on crowd-sourced health studies and mHealth development, along with Genomera, which handles online study operations and engagement with patient communities.26,27 These types of multidisciplinary online communities could ultimately become self-sustaining repositories for n-of-1 protocols and results in a way that advances the field of n-of-1 methods in health care.
N-of-1 trials have the potential to provide the highest level of evidence-based medicine for the individual, but are currently underutilized due to barriers at both the patient and provider levels. In this chapter, we offer methods to address these barriers with patients and providers in order to better engage, train, and support both parties. Reducing transactional costs for all participants may help increase the use of n-of-1 trials in clinical practice. It is recommended that researchers utilize the checklist below when addressing these barriers with patients and providers.
|Engage patients through emphasizing the purpose and potential of n-of-1 trials||
|Provide patients with basic education about n-of-1 methods||
|Engage providers by emphasizing the purpose and potential of n-of-1 trials||
|Provide clinicians with a basic education about n-of-1 methods||
|Provide user-friendly support tools to facilitate n-of-1 study execution and decrease time demands||
|Design systems that encourage collaboration among all user types||Explore the potential of online communities.|
- Kravitz RL, Duan N, Niedzinski EJ, et al. What ever happened to N-of-1 trials? Insiders' perspectives and a look to the future. Milbank Q. 2008 Dec;86(4):533-555.
- Nikles CJ, Clavarino AM, Del Mar CB. Using n-of-1 trials as a clinical tool to improve prescribing. Br J Gen Pract. 2005 Mar;55(512):175-180.
- Quantified Self. Self Knowledge Through Numbers. http://quantifiedself.com/. Accessed November 4, 2011.
- Mehta R. The Self-Quantification Movement–Implications for Healthcare Professionals. SelfCare. 2011;2(3):87-92.
- Sackett DL. Clinician-trialist rounds: 4. why not do an N-of-1 RCT? Clin Trials. 2011 Jun;8(3):350-352.
- Larson EB, Ellsworth AJ, Oas J. Randomized clinical trials in single patients during a 2-year period. JAMA. 1993 Dec 8;270(22):2708-2712.
- Guyatt GH, Heyting A, Jaeschke R, et al. N of 1 randomized trials for investigating new drugs. Control Clin Trials. 1990 Apr;11(2):88-100.
- Nikles CJ, Mitchell GK, Del Mar CB, et al. An n-of-1 trial service in clinical practice: testing the effectiveness of stimulants for attention-deficit/hyperactivity disorder. Pediatrics. Jun 2006;117(6):2040-2046.
- Guyatt G, Sackett D, Adachi J, et al. A clinician's guide for conducting randomized trials in individual patients. CMAJ. 1988 Sep 15;139(6):497-503.
- Kravitz RL, Paterniti DA, Hay MC, et al. Marketing therapeutic precision: Potential facilitators and barriers to adoption of n-of-1 trials. Contemp Clin Trials. 2009 Sep;30(5):436-445.
- Kent MA, Camfield CS, Camfield PR. Double-blind methylphenidate trials: practical, useful, and highly endorsed by families. Arch Pediatr Adolesc Med. 1999 Dec;153(12):1292-1296.
- Scuffham PA, Yelland MJ, Nikles J, et al. Are N-of-1 trials an economically viable option to improve access to selected high cost medications? The Australian experience. Value Health. 2008 Jan-Feb;11(1):97-109.
- Brookes ST, Biddle L, Paterson C, et al. "Me's me and you's you": Exploring patients' perspectives of single patient (n-of-1) trials in the UK. Trials. 2007;8:10.
- Gabler NB, Duan N, Vohra S, et al. N-of-1 trials in the medical literature: a systematic review. Med Care. 2011 Aug;49(8):761-768.
- Nikles CJ, Yelland M, Glasziou PP, et al. Do individualized medication effectiveness tests (n-of-1 trials) change clinical decisions about which drugs to use for osteoarthritis and chronic pain? Am J Ther. 2005 Jan-Feb;12(1):92-97.
- Yelland MJ, Nikles CJ, McNairn N, et al. Celecoxib compared with sustained-release paracetamol for osteoarthritis: a series of n-of-1 trials. Rheumatology. 2007 Jan;46(1):135-140.
- Woodfield R, Goodyear-Smith F, Arroll B. N-of-1 trials of quinine efficacy in skeletal muscle cramps of the leg. Br J Gen Pract. 2005 Mar;55(512):181-185.
- Lillie EO, Patay B, Diamant J, et al. The n-of-1 clinical trial: the ultimate strategy for individualizing medicine? Per Med. 2011 Mar;8(2):161-173.
- Jaeschke R, Adachi J, Guyatt G, et al. Clinical usefulness of amitriptyline in fibromyalgia: the results of 23 N-of-1 randomized controlled trials. J Rheumatol. 1991 Mar;18(3):447-451.
- Jaeschke R, Singer J, Guyatt GH. A comparison of seven-point and visual analogue scales. Data from a randomized trial. Control Clin Trials. 1990 Feb;11(1):43-51.
- Fagerlin A, Zikmund-Fisher BJ, Ubel PA. Helping patients decide: ten steps to better risk communication. J Natl Cancer Inst. 2011 Oct 5;103(19):1436-1443.
- The Learning Healthcare System: Workshop Summary. Paper presented at: Institute of Medicine; 2007; Washington, D.C.
- Guyatt GH, Haynes RB, Jaeschke RZ, et al. Users' Guides to the Medical Literature: XXV. Evidence-based medicine: principles for applying the Users' Guides to patient care. Evidence-Based Medicine Working Group. JAMA. 2000 Sep 13;284(10):1290-1296.
- James B. Quality Improvement Opportunities in Health Care–Making It Easy to Do It Right. JMCP. 2002;8(5):394-399.
- Kamien M. The use of an N-of-1 randomised clinical trial in resolving therapeutic doubt. The case of a patient with an 'attention disorder.' Aust Fam Physician. 1998 Jul;27 Suppl 2:S103-105.
Kaplan HC, Gabler NB, the DEcIDE Methods Center N-of-1 Guidance Panel. User Engagement, Training, and Support for Conducting N-of-1 Trials. In: Kravitz RL, Duan N, eds, and the DEcIDE Methods Center N-of-1 Guidance Panel (Duan N, Eslick I, Gabler NB, Kaplan HC, Kravitz RL, Larson EB, Pace WD, Schmid CH, Sim I, Vohra S). Design and Implementation of N-of-1 Trials: A User's Guide. AHRQ Publication No. 13(14)-EHC122-EF. Rockville, MD: Agency for Healthcare Research and Quality; January 2014: Chapter 6, pp. 71-81. http://www.effectivehealthcare.ahrq.gov/N-1-Trials.cfm