Explainable analytics: Understanding causes, correcting errors, and increasingly achieving perfect accuracy from nature of distinguishable patterns

In addition to pursue accurate analytics, it is invaluable to clarify how and why inaccuracy exists. We propose a transparent classification method (TC). In training, we discover patterns from positive and negative observations respectively; next, patterns are excluded if they appear in both types. In testing, observations are scored by the pure patterns and connected like social networks. Based on set theory, pure patterns have explanatory power for distinguishing tangled relationship between negative and positive observations. Experimental results demonstrate that TC can identify all positive (e.g., malignant) observations at low ratios of training to testing, e.g., 1:9 in Breast Cancer Wisconsin (Original) and 3:7 in Contraceptive Method Choice dataset. Without fine-tuned parameters and random selection, TC eliminates uncertainty of the methodology. TC can visualize causes, and therefore, prediction errors are traceable and can be corrected. Further, TC shows potential of identifying whether the ground truth is incorrect (e.g., diagnostic errors).

The relationship between shared-decision making in contraceptive counseling and satisfaction in Accra and Kumasi, Ghana

Background: Current use of modern methods of contraception remain low in many parts of sub-Saharan Africa, including Ghana. One way to improve both satisfaction with and continuation of contraceptive usage is to increase the level of shared decision-making around method choice. In this study, we sought to evaluate the extent to which patients in urban Ghana experienced shared decision-making and if this was associated with method chosen, satisfaction, or continuation of the method at three-months post-visit. Methods: We conducted a longitudinal survey. Women were recruited when they were starting a new method of contraception and followed-up with at three-months post-initiation from five family planning clinics in Accra and Kumasi, Ghana. Participants were asked who made the decision about their method choice, the patient herself, the provider, or the patient and provider together. Our outcomes included measures of satisfaction and three months’ continuation. Results: Fifty-eight percent of our participants reported making the decision of which method to use themselves, and eighty percent reported being satisfied to be leaving with their chosen method. At three months, those who reported they engaged in shared decision-making were more likely to report they would choose the same method again (p=.003), a measure of satisfaction. Patients who reported they made the decision of which method to use (p=.002) and those who left with an injection or pill (p=.019) rated their provider less favorably, while participants who had used a method before (p=.024) and those who reported they received their method of choice (p=.000) rated their providers more favorably. Conclusions: Measured in multiple ways, women who made the decision of which method to use were less satisfied. These results show the importance of providers engaging with patients during the contraceptive decision-making process.

Explainable analytics: Understanding causes, correcting errors, and increasingly achieving perfect accuracy from nature of distinguishable patterns

In addition to pursue accurate analytics, it is invaluable to clarify how and why inaccuracy exists. We propose a transparent classification method (TC). In training, we discover patterns from positive and negative observations respectively; next, patterns are excluded if they appear in both types. In testing, observations are scored by the pure patterns and connected like social networks. Based on set theory, pure patterns have explanatory power for distinguishing tangled relationship between negative and positive observations. Experimental results demonstrate that TC can identify all positive (e.g., malignant) observations at low ratios of training to testing, e.g., 1:9 in Breast Cancer Wisconsin (Original) and 3:7 in Contraceptive Method Choice dataset. Without fine-tuned parameters and random selection, TC eliminates uncertainty of the methodology. TC can visualize causes, and therefore, prediction errors are traceable and can be corrected. Further, TC shows potential of identifying whether the ground truth is incorrect (e.g., diagnostic errors).

Factors affecting uptake of the levonorgestrel-releasing intrauterine device: A mixed-method study of social franchise clients in Nigeria

Background Despite the positive characteristics of the levonorgestrel-releasing intrauterine device (IUD)–a long-acting, highly effective contraceptive with important non-contraceptive attributes–the method has not been widely available in low- and middle-income countries. This study of hormonal IUD, copper IUD, implant and injectable users in Nigeria compares their characteristics, reasons for method choice, and experiences obtaining their method. Methods We conducted a phone survey with 888 women who received a hormonal IUD, copper IUD, contraceptive implant or injectable from 40 social franchise clinics across 18 states in Nigeria. We analyzed survey data descriptively by method and assessed factors associated with hormonal IUD use through multivariate logistic regression models. Follow-up in-depth interviews conducted with 32 women were analyzed thematically. Results There were few differences by method used in the socio-demographic profiles and contraceptive history of participants. Among users choosing a long-acting, reversible method, the top reasons for method choice included perceptions that the method was “right for my body,” long duration, recommended by provider, recommended by friends/family, few or manageable side effects, and high effectiveness. Among hormonal IUD users, 17% mentioned reduced bleeding (inclusive of lighter, shorter, or no period), and 16% mentioned treatment of heavy or painful periods. Qualitative data supported these findings. Among survey respondents, between 25% and 33% said they would have chosen no method if the method they received had not been available. Both quantitative and qualitative data indicated that partner support can affect contraceptive use, with in-depth interviews revealing that women typically needed partner permission to use contraception, but men were less influential in method choice. Conclusions Expanding access to the hormonal IUD as part of a full method mix provides an opportunity to expand contraceptive choice for women in Nigeria. Findings are timely as the government is poised to introduce the method on a wider scale.

The Effect of Implant Surface Design and Their Decontamination Methods in Peri-Implantitis Treatment

Different titanium implant surfaces are prone to microbial colonization and dental plaque accumulation contributing to peri-implantitis pathogens adherence and growth. In conjunction with systemic, local, and implant-based factors such as micro- and macro-designs, implant location, and region, these pathogens can cause a complex inflammatory response resulting in peri-implantitis and deleterious bone loss. Implant surface decontamination plays a crucial and important step in peri-implantitis therapy. The primary goal of implant surface decontamination is to eradicate bacteria and their products outside of implant pits and grooves reducing inflammation and promoting tissue regeneration and/or reparation. Various implant surface decontamination methods such as mechanical, chemical or physical methods have been proposed to prevent bacterial resistance development or/and surface damage. The chapter aimed to assess if implant microdesign could influence the decontamination method choice.

“In the Moment I Wanted to Kill Myself, and Then After I Didn’t”: A Qualitative Investigation of Suicide Planning, Method Choice, and Method Substitution Among Adolescents Hospitalized Following a Suicide Attempt

Research on planning, method choice, and method substitution in adolescents’ suicide attempts is limited. We conducted in-depth interviews with 20 adolescents following their suicide attempt to learn the extent to which the attempt was planned, why they used the method they did, and whether they would have substituted another method if the one they used had been unavailable. Applied Thematic Analysis was used to identify codes and develop themes. Attempts were largely unplanned, and planned attempts were often haphazard, as urgency to escape immediate pain was a main impetus for the attempt. Method choice was driven by easy access. Half of participants said they would not have attempted suicide if the method they used were inaccessible, but 7 of 20 said they would have, and 3 were unsure. Not all suicide attempts would be prevented by blocking access to methods that adolescent attempters would otherwise use. To understand whether restricting access to low-lethality methods could harm some attempters, future research should examine in-the-moment method substitution.

Method-specific beliefs and subsequent contraceptive method choice: Results from a longitudinal study in urban and rural Kenya

Introduction Evidence from sub-Saharan Africa, including Kenya, shows that negative beliefs about contraceptive methods are associated with non-use. However, little is known about the relationship between contraceptive beliefs and subsequent method choice. Methodology We used data from a two-year longitudinal survey of married women aged 15–39 years at enrollment from one urban site (Nairobi) and one rural site (Homa Bay) in Kenya. Analysis entails descriptive statistics and estimation of a conditional logit analysis to examine associations between method-specific beliefs and choice of injectables, implants or pills among women who were not using any method or were pregnant at baseline (round 1) but adopted these methods at 12-month follow-up (Nairobi, n = 221; Homa Bay n = 197). Results Beliefs about pills, injectables and implants among non-users were generally negative. With the partial exception of the pill in Nairobi, the majority thought that each method was likely to cause serious health problems, unpleasant side effects, menstrual disruption, and would be unsafe for long-term use. In both sites, satisfied past use of a method and the perception that a method is easy to use had a major influence on method choice. Concerns about menstrual disruption and safety for long-term use were unimportant in both sites. There were some marked differences between the two sites. Beliefs about long-term fertility impairment and perceived husband approval had strong influences on choice of injectables, implants or pills in the urban site but not in the rural site. Conclusion The relative importance of beliefs, some erroneous, in predisposing women to choose one method over another appears to be conditioned by the social context. There is need for family planning counseling programmes to pay attention to erroneous beliefs and misconceptions about contraceptives.

Perceptions of Peer Contraceptive Use and its Influence on Contraceptive Method Use and Choice among Young Women and Men in Kenya: A Quantitative Cross-Sectional Study

Background Prior research has established that an individual’s social environment may influence his or her reproductive behaviors, yet less is known about peer influence on contraceptive use among young people (ages 15–24). In Kenya, the site of this study, 15% of adolescents ages 15–19 have begun childbearing and 45% of sexually active young women report current use of a modern contraceptive method. This highlights the need to better understand what factors influence young people to use contraception. The objective of this study is to explore the relationship between the perception of peers’ use of contraceptives and contraceptive use and method choice among young men and women in Kenya. Methods This study utilizes a nationally representative sample of women and men aged 15–24 years from the 2018 and 2019 cross sectional Shujaaz State of the Kenyan Youth annual surveys. Among the sample of sexually experienced young people (59%), multivariable multinomial logistic regression was used to explore the association between the perception of peers’ use of contraceptives and the respondent’s contraceptive method choice: non-user, condom use or use of any other modern method. Results are presented separately for young men and young women. Results Our results show that sexually experienced young men and women who perceive that their peers are using contraceptives are more likely to report current use of modern contraception. Among women and men, the perception that their peers use contraceptives is associated with higher use of condoms compared to being a nonuser; they are also more likely to use condoms than another modern method of contraception. Young women are more likely to use another modern method (not including condoms) than be a nonuser when they perceive that their peers use contraceptives. Conclusions The results of this study highlight the important role of peer influence on young people’s contraceptive choices. These findings can be used to develop programs that encourage behavior change communication activities in Kenya that focus on providing information on the full range of contraceptive methods as well as normalizing contraceptive use among peer groups of sexually experienced young people.