This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
There is growing evidence that many patients and caregivers innovate by developing new solutions to cope with their health disorders. Given the easy access to vast internet resources and peers globally, it is increasingly important to understand what may influence user innovation and its adoption in health for improving individual well-being and ensuring their safety, in particular, how interactions with peers and physicians or search behavior, along with sociodemographics, may influence the decision to develop a solution or adopt one developed by a peer.
The aim of this paper was to study the development and peer-to-peer adoption of user innovations in health care and identify individual-level factors associated with these processes.
Data were collected via computer-assisted phone survey from a large, random, and representative sample of adult residents in Portugal (N=6204). User innovation questions were added to 1 wave of an ongoing observational, longitudinal, population-based epidemiological study. By asking about individual innovation activity, the sample was split into 3 groups: (1) the developers of health-related solutions for own use (developers), (2) the adopters of solutions developed by other patients or caregivers (peer-to-peer adopters), and (3) the rest of the population. Within the last group,
In the population considered in this paper, an estimated 1.3% (75/6008) reported having developed a solution for own use and 3.3% reported to have adopted a solution developed by peers. The 3 groups (developers, adopters, and remaining population) have distinctive characteristics. Gender plays an important role in the solution development, as women are less likely to develop one (odds ratio [OR] 0.4, 95% CI 0.20-0.81;
This paper demonstrates the importance of the peer community, doctor-patient relationship, citizen’s search for information on innovation, and individual attitudes toward peer-to-peer adoption in health care. It stresses the need for a reliable Web-based health-related information and the necessity to deeper understand complex relationships between the need to improve health and fulfill the need and the perception of the health care system.
User innovators are the ones who have developed a new good or service or modified an existing good or service for own use; they differ from producer innovators for whom profit is the dominant motivation to innovate [
The largest group of user innovators in health care are
When the value of patient-developed solutions is considered, there is evidence of patients reporting significant improvements in the quality of life after using their self-developed solutions [
Innovation and adoption activity by patients and caregivers in health care may be strongly influenced by health care–related and sociotechnological contextual factors. However, no prior work systematically explored such relationships. For example, we know that people invest significant efforts to search for health-related information (online and offline) [
The data used in this paper are survey responses from a random and representative sample of adult residents in Portugal (N=6204), collected via computer-assisted phone survey conducted by NOVA Medical School. Professional interviewers were additionally trained by a psychologist to communicate the innovation questions, and 2 of the authors trained them to fill in the survey responses in a computer program during the conversation.
The innovation section of the survey, which is the focus of this paper, was integrated into a larger project, the second wave of a longitudinal, prospective, observational, population-based study named Epidemiology of Chronic Diseases (EpiDoC) (
Participants were selected through a process of multistage random sampling. The sample was stratified according to the Portuguese statistic regions in the 2001 Census and the size of the population (<2000; 2000-9999; 10,000-19,999; 20,000-99,999; and ≥100,000 inhabitants). The number of participants of each stratum was proportional to the actual distribution of the population. In Madeira and the Azores, the sample size was increased (oversampling) to allow separate analyses in these regions. Candidate households were selected through a random route process; sampling points were randomly selected on the maps of each locality, where the interviewer began a systematic step count (defined for each locality according to its size), granting each household and everyone an equal probability of being chosen [
Most of the EpiDoc wave participants (10,153) also agreed to integrate into a prospective cohort and be contacted in the next round of surveying (EpiDoc 2)—the cohort of rheumatic diseases (CoReumaPt) wave (2013-2015). The Portuguese National Commission for Data Protection and the NOVA Medical School Ethics Committee have approved both EpiReumaPt and CoReumaPt [
To guarantee the representativeness of the sample in relation to the Portuguese population (Mainland and
Flowchart of the population-based study named Epidemiology of Chronic Diseases (EpiDoC). The first wave (2011-2013) is entitled EpiDoC 1 - Epidemiology of rheumatic diseases study (EpiReumaPt). The second wage (2013-2015) is entitled EpiDoc 2 - Cohort of rheumatic diseases (CoReumaPt).
Scales, sources, and factor loadings.
Construct | Items | Factor loadings (N=6204) |
Self-responsibility for health (2 items, adapted from a scale by Hibbard et al [ |
“It is me, more than any other person, who is responsible for my health and well-being.” (5-point Likert scale: 1=totally disagree, 5=totally agree) | 0.9 |
Self-responsibility for health (2 items, adapted from a scale by Hibbard et al [ |
“The most important factor that influences my well-being and health is my active role and responsibility for my health.” (5-point Likert scale: 1=totally disagree, 5=totally agree) | 0.9 |
Trust in medical doctor (Reduced scale proposed by Anderson et al [ |
“I trust my doctor so much that I always try to follow his/her advice.” (5-point Likert scale: 1=totally disagree, 5=totally agree) | 0.9 |
Trust in medical doctor (Reduced scale proposed by Anderson et al [ |
“If my doctor tells me something is so, then it must be true.” (5-point Likert scale: 1=totally disagree, 5=totally agree) | 0.9 |
Trust in medical doctor (Reduced scale proposed by Anderson et al [ |
“I feel my doctor does everything he/she should for my medical care.” (5-point Likert scale: 1=totally disagree, 5=totally agree) | 0.8 |
Perceptions of medical science frontiers (new) | Do you believe that the medical science can treat your disease? (5-point Likert scale: 1=not at all, 5=completely trust) | 0.9 |
Perceptions of medical science frontiers (new) | How likely is it that the medical science can successfully treat you for your disease? (5-point Likert scale: 1=not at all, 5=it certainly can) | 0.9 |
Intention to adopt a patient-developed solution [ |
How likely is it that you would use a solution developed by another patient to help you cope with your ailment? (5-point Likert scale: 1=very unlikely, 5=I would definitely) | 0.9 |
Intention to adopt a patient-developed solution [ |
Do you intend to use a solution developed by another patient to help you cope with your ailment? (5-point Likert scale: 1=I do not intend to use, 5=I definitely intend to use one) | 0.9 |
The first group of questions in the innovation section of the survey measured contextual factors, self-responsibility for health management, search for health information, the frequency of online and face-to-face interactions with peers, and trust in medical doctors and medical science. Next, respondents answered a question that split the sample into 3 groups: (1) the developers of solutions to cope with their health disorders, (2) adopters of health-related solutions developed by other patients or caregivers, and (3) the remaining population.
The question asked if an individual had developed a health-related solution or adopted a health-related solution developed by other patients. In the case of an affirmative response, the survey continued with sections that focused on the details about solution development or adoption, dividing the population into the developer or adopter groups. The third group, those who neither innovated nor adopted a solution, were asked whether they have ideas about potential solutions for health-related problems they so far encountered. Those who neither developed nor adopted a patient-developed solution were also asked about their intentions to adopt a patient-developed solution. For all the respondents, the survey started with sociodemographic questions and ended with life habits, functional, and quality of life questionnaires.
Questions regarding user innovation were built upon a questionnaire used in user innovation measurement surveys [
The question about creative activity is formulated to ask about newly developed or modified solutions and about the adoption of a patient-developed solution for personal use or for someone close to the respondent. From the survey responses on this question, 2 dependent (binary) variables were created, a solution development variable (developer) and a solution adoption variable (adopter). Furthermore, the origin of the advice for the adopter was asked for, to ensure that the source of the solution is a patient/caregiver. The objective was to identify the characteristics of those who had engaged in creative activity, regardless of the artifact’s quality that is developed or adopted.
A large share of the population is likely to be neither solution developers nor adopters of peer-developed solutions, as not everyone has a need for a solution to cope with health-related issues. For this group, and to study the drivers of the attitude toward peer-developed solutions in a general population, the theory of planned behavior [
Considering the earlier stated goal of this paper, learning about individual creative activity in the health care context, a set of questions was added to the survey. To learn about individual search efforts, the survey measured the depth of search for health information and health-related solutions as the average weekly time spent searching. As social interactions among patients may influence adoption or intentions to adopt a patient-developed solution, the interviewees were asked about the frequency of their interactions with individuals who are afflicted with the same health disorder or who share interests in the disorder; 5-point Likert scales were used to represent different levels of frequency of interactions.
To measure the perceptions of the scientific frontier, trust in medical doctors, and the attitude toward personal health management, where possible, existing scales were used (reported in
The medical part of the survey, pertinent to the epidemiological study, included standard measurement instruments that assessed health and quality of life. In this paper, the EuroQoL-5D (EQ-5D) score [
Given that measurement instruments were used, the survey was pretested on 106 randomly selected interviewees. In this step, exploratory factor analysis is conducted to test whether theoretically constructed 4 factors could be identified and if there is a sufficient level of internal consistency.
Descriptive statistics are reported for the full sample (N=6204) after applying probability weights to obtain the population estimates [
The scales for self-responsibility for health, trust in doctors, and perceptions of medical science frontiers are included as standardized values, and the interpretation of the coefficients should be in terms of the SDs from the population mean.
The results of exploratory factor analysis on the initial sample of 106 individuals suggest that the items load well on the 4 factors, and all the factors have high internal consistency (alpha≥.7) [
Descriptive statistics are reported in
The results show that 1.3% of the population reported being developers and 3.3% peer-to-peer adopters. The respondents have on average 9 years of formal education (SD 4 years), and 49% reported being diagnosed with at least one noncommunicable chronic disease.
There are notable differences among the 3 groups along several dimensions. Within the developers’ group, males represent the majority (66%). Furthermore, unemployment or temporal disability/retirement among the developers (54%) is higher than that among adopters (39%) or the remaining population (35%). The developers have, on average, 1 more year of education than the adopters. For all 3 groups, interacting with peers (patients/caregivers) via the internet is rare, and the remaining population (neither developers nor adopters) are more active in that regard, with 2% more active people than that in the other 2 groups. Majority of the developers and adopters have frequent in-person interactions, 65% and 53%, respectively. The adopters have a higher number of comorbidities, 2.1 compared with 1.9 for developers and 1.5 for the rest of the population, on average. Although all 3 groups have left-skewed self-responsibility for health (4.9 out of 5), developers are more active than others, as 64% exercise regularly compared with around 40% in the other 2 groups. All 3 groups have high trust in doctors, with a marginally higher value for the remaining population, 4.5 compared with 4.3 out of 5. Perception of medical science frontier is also left-skewed, with average values of 3.5 for the adopters and 3.8 for the other 2 groups.
Absolute values of correlations between independent variables (correlation matrix available upon request) were below .35, with 4 exceptions. These exceptions were (1) age and education (
Descriptive statistics for the 3 groups (Innovator, Adopter, Remaining Population).
Population characteristics | Innovator (n=75) | Adopter (n=210) | Remaining population (n=5723) | |
Gender (female), n (%) | 40 (34.0) | 172 (58.4) | 3146 (47.6) | |
Age (years), mean (SD) | 44.62 (14.40) | 49.53 (17.09) | 46.41 (17.85) | |
Years of education, mean (SD) | 9.36 (3.28) | 8.33 (3.93) | 8.92 (3.81) | |
Employed full-time, part-time, or domestic worker | 38 (45.8) | 116 (60.9) | 3193 (65.2) | |
Temporally work disabled/retired | 24 (22.4) | 70 (30) | 1834 (24.9) | |
Unemployed | 8 (31.7) | 23 (9.1) | 479 (9.9) | |
Norte | 17 (33.1) | 65 (32.5) | 1883 (37.8) | |
Centro | 23 (36.4) | 49 (25.4) | 1210 (23.7) | |
Lisboa | 15 (17.2) | 46 (28.2) | 1135 (24.3) | |
Alentejo | 5 (8.2) | 12 (7.4) | 264 (6.3) | |
Algarve | 1 (2.2) | 4 (3) | 170 (3.6) | |
Azores | 7 (1.4) | 17 (1.5) | 521 (2.0) | |
Madeira | 7 (1.6) | 17 (2) | 540 (2.3) | |
Physical exercise at least once per week, n (%) | 37 (63.9) | 85 (40.0) | 2390 (44.6) | |
Quality of life, EQ-5Da score—CoReumaPtb, mean (SD) | 0.71 (0.23) | 0.73 (0.27) | 0.80 (0.26) | |
EQ-5D score difference CoReumaPtb-EpiReumaPtc, mean (SD) | −0.05 (0.26) | −0.09 (0.27) | −0.05 (0.24) | |
Number of chronic diseases, mean (SD) | 1.81 (1.64) | 2.07 (2.55) | 1.51 (1.71) | |
Never | 30 (36.2) | 88 (48.8) | 3181 (58.3) | |
Less than once a week | 12 (10.1) | 61 (26.9) | 1298 (22.3) | |
At least once a week | 33 (53.7) | 60 (24.3) | 1207 (19.4) | |
No | 69 (92) | 248 (92) | 5450 (96.1) | |
Yes | 6 (8) | 16 (8) | 179 (3.9) | |
Depth of search: search time on health (hours per week), mean (SD) | 1.97 (4.2) | 1.09 (2.08) | 0.47 (1.64) | |
Self-responsibility for health, mean (SD) | 4.82 (0.28) | 4.90 (0.30) | 4.86 (0.41) | |
Trust in doctors scale, mean (SD) | 4.31 (0.73) | 4.35 (0.86) | 4.47 (0.79) | |
Intentions to adopt, mean (SD) | —d | — | 2.63 (0.3) | |
Perceptions of medical science frontier, mean (SD) | 3.81 (0.85) | 3.53 (1.12) | 3.85 (0.94) |
aEQ-5D: EuroQoL-5D.
bCoReumaPt: Cohort of rheumatic diseases (EpiDoC 2).
cEpiReumaPt: Epidemiology of rheumatic diseases study (EpiDoC 1).
dNot applicable.
Results of the multivariable analysis are shown in
Considering the solution development for own use (model 1), the results show that women are less likely to develop a solution for own use (OR 0.40, 95% CI 0.20-0.81;
The population of adopters is significantly different from the population of developers. Adoption (model 2) is weakly positively associated with female gender (OR 1.54, 95% CI 0.94-2.52;
In models 3 and 4, the dependent variable is the intention to adopt a solution developed by a patient or a nonprofessional caregiver. Intentions, according to the theory of planned behavior, are a proxy for actual behavior. In this paper, they are interpreted as attitudes toward peer-developed solutions.
The results for the subsample of individuals with at least one chronic noncommunicable disease (model 3) suggest a distinct combination of statistically significant associations. Intentions to adopt are negatively associated with age (beta=−.01; 95% CI −0.02 to −0.01;
Within the remaining population, the subsample of individuals without a chronic disease is very similar to the subsample of those with a chronic disease. Distinctive characteristic of the former subgroup is a negative association between retirement/temporary work disability and the intentions to adopt a peer-developed solution (beta=−.23; 95% CI −0.41 to −0.04;
Multivariable analysis with population estimates.
Population characteristics | Developer (versus all the others); Model 1, odds ratio (95% CI) | Adopter (versus all the others); Model 2, odds ratio (95% CI) | Intentions to adopt (remaining population with a chronic disease); Model 3, beta estimates (95% CI) | Intentions to adopt (remaining population without a chronic disease); Model 4, beta estimates (95% CI) |
Gender: female versus male | 1.54b (0.94 to 2.52) | − |
0.01 (−0.10 to 0.12) | |
Age (years) | 0.99 (0.96 to 1.02) | 1.00 (0.99 to 1.02) | − |
− |
Education (in years) | 0.95 (0.87 to 1.03) | |||
Employment: temporarily work disabled/retired versus employed (full- or part-time) | 1.41 (0.53 to 3.79) | 0.95 (0.53 to 1.72) | −.02 (−0.16 to 0.13) | − |
Employment: unemployed versus employed (full- or part-time) | 1.02 (0.55 to 1.90) | .04 (−0.13 to 0.21) | −0.14 (−0.34 to 0.06) | |
Marital status (married or union versus single or widow or divorced) | 2.46b (0.98 to 6.14) | 1.20 (0.66 to 2.19) | −.00 (−0.13 to 0.12) | −0.02 (−0.14 to 0.09) |
Regular physical exercise | 0.85 (0.52 to 1.38) | .08 (−0.06 to 0.21) | 0.03 (−0.08 to 0.14) | |
Health state, EQ-5Dd score—CoReumaPte | 0.63 (0.25 to 1.62) | 1.09 (0.30 to 4.03) | −.02 (−0.26 to 0.23) | −0.23 (−0.61 to 0.15) |
Score Difference EQ-5D score: CoReumaPte-EpiReumaPtf | —g | — | −.09 (−0.31 to 0.14) | −0.03 (−0.37 to 0.32) |
Face-to-face Interaction with other patients/caregivers: less than once a month versus no interactions | 0.68 (0.27 to 1.71) | 1.24 (0.74 to 2.08) | — | |
F2F Interaction with other patients/caregivers: once a month or more versus no interactions | 1.21 (0.72 to 2.02) | — | ||
Online interactions with other patients | 0.62 (0.12 to 3.14) | 2.12b (0.95 to 4.74) | −.09 (−0.38 to 0.19) | — |
Health information search depth (hours per week) | −0.00 (−0.04 to 0.04) | |||
Personal responsibility for health (standardized) | 1.06 (0.73 to 1.56) | 1.14 (0.90 to 1.44) | −.04b (−0.09 to 0.00) | 0.04 (−0.04 to 0.11) |
Trust in physician (standardized) | 0.92 (0.72 to 1.18) | 0.94 (0.80 to 1.09) | − |
− |
Perceptions of medical science frontier (standardized) | 0.99 (0.71 to 1.37) | 0.81 (0.63 to 1.05) | −.01 (−0.07 to 0.04) | −0.02 (−0.08 to 0.04) |
With at least one disease versus no disease | 1.02 (0.60 to 1.73) | — | — |
a
b
c
dEQ-5D: EuroQoL-5D.
eCoReumaPt - Cohort of rheumatic diseases (EpiDoC 2).
fEpiReumaPt – Epidemiology of rheumatic diseases study (EpiDoC 1).
gNot applicable.
The analysis suggests that solution development and adoption of peer-developed solutions are relatively infrequent but significant phenomena and that the 3 groups have distinctive characteristics. Population estimates of the share of solution developers, 1.3%, is over 2 times higher than the estimate of the share of health care–related innovation by citizens in the United Kingdom [
A series of results are aligned with the extant academic literature in user innovation. Regression results suggest that solution developers are more often men and educated individuals, confirming the findings from the study of user innovations by consumers in the United Kingdom [
Our study showed that developers are more likely to be unemployed than employed. A plausible explanation may be that they have more time to reflect upon needs and solutions or that unemployment is associated to a higher likelihood of suffering from health disorders and to having financial difficulties [
Group-mean comparison suggests that the average time spent searching for health-related information by the developers (2 hours/week) is almost twice the time spent by the adopters (1.1 hours/week) and 4 times higher than that for the rest of the population (0.5 hours/week). This result emphasizes the importance of the provision of curated and accurate information, especially when, following an advice of a peer without consulting a health professional may be quite dangerous. For example, applying a plant extract without understanding side effects or permitted dosages may provoke serious health issues.
Considering adoption of solutions developed by patients or caregivers, the regression results do not suggest any stark characteristic of the group of adopters. However, the application of the intentions to adopt, a concept from the theory of planned behavior, reveals an important association. The attitude of those who did not engage in neither developing a solution nor adopting one may be influenced by the doctor-patient relationship. In light of the safety concerns regarding the diffusion of (self-made) health solutions in informal communities of patients and caregivers, doctors are a vital element of the health care system that helps patients to establish safety and efficacy of the available solutions. A negative association between the intentions to adopt and age possibly reflects the generational change in the perception of the role of the conventional health care system. In particular, older individuals may be used to the paternalistic doctor-patient relationship, and they may put a higher value on the official source of health-related solutions. Education is positively associated with the intentions to adopt a solution, which is also potentially linked to the paradigm shift in health care from paternalistic to more egalitarian relationships between patients and health professionals.
In this paper, data have been collected from a prospective cohort; as we worked with cross-sectional data, only associations may be claimed. Recollection and interpretation bias may be present in the data. Although some people may have developed or adopted a patient-developed solution without being aware of it, the focus of this work was to explore the characteristics of those who are aware and have had chosen to develop or adopt a peer-developed solution. Hence, these biases are likely not to influence the results significantly.
A set of preemptive steps were taken before administering the survey to control for item-related (common method) bias, as suggested by Podsakoff et al [
The advantage of the study is the size of the sample and the sampling design. As this study is conducted in 1 country, it casts doubt whether the results are generalizable to other cultural and health care policy settings.
This paper is the first-of-type exploratory analysis of creative activities in the general population that focuses on health care and takes into consideration health care contextual factors. It demonstrates distinctive characteristics of: (1) the patients and caregivers who are developers of solutions, (2) the adopters of peer-developed solutions, and (3) the attitudes of the remaining population. Two actionable takeaways from the study are the importance of supplying reliable health-related information to patients who are searching and of the investment in good doctor-patient relationships.
Treating patients as equals is becoming the new mantra in organized health care systems [
Data description.
Cohort of Rheumatic Diseases
Epidemiology of Chronic Diseases
Epidemiology of Rheumatic Diseases Study
EuroQoL-5D
Odds ratio
The authors are grateful to the Public Health Initiatives Programme (PT06), funded by EEA Grants Financial Mechanism 2009-2014. In addition, they are grateful for the funding provided by the Fundação para a Ciência e a Tecnologia (FCT) and the Carnegie Mellon Portugal Program through project Technology, Entrepreneurship and Innovation Policy Lab, CMUP-ERI/TPE/0028/2013. They are also thankful for help of the EpiReumaPt study group and the invaluable comments and wisdom of Professor Eric von Hippel (Massachusetts Institute of Technology).
None declared.