Evaluating the effectiveness of a mobile application to improve the quality, collection, and usability of forensic documentation of sexual violence

Background Survivors of sexual violence deserve timely and high-quality forensic examination, evidence collection, and documentation as part of comprehensive care. However, in many countries, the quality of medical-legal documentation is severely limited. MediCapt is an innovative digital application that enables clinicians to document forensic medical evidence as well as capture and securely store forensic photographs of injuries. This study evaluated the effectiveness and usability of MediCapt to document forensic medical evidence of sexual violence. Methods This mixed-methods evaluation involved key-informant interviews, usability questionnaires, and forensic record reviews. Participants included clinicians, medical records personnel, information technology personnel, and health facility administrators, as well as law enforcement and legal professionals in Kenya. Results The Physicians for Human Rights (PHR) data quality checklist found that using MediCapt led to significantly higher data-quality scores compared to paper-based forms. MediCapt forms scored higher on 23 of 26 checklist items. While a wide difference in quality was seen among paper-based forms, MediCapt appeared to both standardize and improve quality of documentation across sites. MediCapt strengths included data security and confidentiality, accuracy and efficiency, and supplemental documentation with photography. Weaknesses included infrastructure challenges, required technological proficiencies, and time to learn the new system. Although it is early to assess the impact of MediCapt on prosecutions, providers and law and justice sector professionals were optimistic about its usefulness. They identified MediCapt as appropriate for use with survivors of sexual violence and reported MediCapt’s legibility and photography features had already been commended by the court. Conclusion MediCapt was well-received across all sectors, its use was perceived as feasible and sustainable, and it significantly improved the quality of collected forensic data. It is anticipated this improvement in forensic documentation will increase successful prosecutions, strengthen accountability for perpetrators, and improve justice for survivors.

Introduction improper management of forensic evidence due to lack of consistent chain of custody mechanisms, remain some of the major factors in the significantly low number of convictions for SGBV, contributing to continued impunity [23]. Physicians for Human Rights (PHR), a global non-governmental organization, has been collaborating with partners in Kenya since 2011 to enhance local capacities for forensic documentation of sexual violence through training on forensic documentation, institutional assessment to strengthen post-rape care, and multisectoral network building to enhance collaboration amongst professionals. To help address persistent shortcomings in forensic documentation and to improve access to justice for survivors, PHR-in close partnership with colleagues in the medical, law enforcement, and justice sectors in Kenya, the Democratic Republic of Congo, and beyond-collaboratively designed and developed MediCapt, an award-winning application that enables clinicians to document medical evidence of sexual violence cases on a digital medical certificate (such as the PRC form in Kenya), capture forensic photographs of the injuries sustained, and store this crucial evidence securely in the cloud [24]. MediCapt was developed using a rights-based collaborative design approach that was guided by the Principles for Digital Development [25].
MediCapt was first introduced in Kenya in October 2017 with an initial roundtable with key stakeholders, a training in January 2018, and then the pilot testing period [26]. During the MediCapt pilot, the application development team received recommendations from the end users (health care professionals) on ways to make improvements to the application. Updates were made progressively during the pilot to enhance the usability of the application, based on end-user input, as well as to ensure that it was effective as a forensic documentation tool for this context.
This study evaluated the effectiveness and usability of the MediCapt mobile application as a tool to document forensic medical evidence of sexual violence. In particular, the authors sought to understand whether digital forensic documentation tools improve the evidence collection process and documentation quality compared to paper-based forensic documentation. As MediCapt is intended to be used in low-resource environments this study also endeavored to identify obstacles to uptake of digital forensic evidence tools as well as strategies to overcome them.

Materials and methods
This mixed-methods study involved baseline and endline key-informant interviews, usability questionnaires, and forensic record reviews related to MediCapt implementation in Kenya. The evaluation was conducted at two hospitals in Kenya's Nakuru County: Naivasha County Referral Hospital in Naivasha (trained in MediCapt October-December 2018) and Rift Valley Provincial General Hospital (RVPGH) in Nakuru (trained in MediCapt October 2020 -January 2021).
The study was conducted by a multidisciplinary team of professionals based in Kenya and the United States with experience in sexual violence research, forensic medicine, and assessment of health interventions in resource-limited settings.

A. Data-quality assessments of forensic records
The Kenyan Ministry of Health's PRC form is a two-page, triplicate form used by clinicians in Kenya to document survivor-reported sexual violence. The PRC form is divided into two sections: Part A, the description of the incident, the physical examination findings, and the documentation of the clinical management and forensic evidence; and Part B, the psychological assessment.
The MediCapt application (S1 and S2 Figs) collects the same data as the paper-based PRC forms. The application also allows providers to obtain (with survivor consent) forensic photographic evidence alongside the traditional history-and-physical data.
To objectively assess the data quality of the paper-based and MediCapt forensic records, a data quality checklist tool was iteratively developed by the study team with local clinicians, law enforcement, and legal professionals in Kenya. The draft checklist was piloted and underwent an inter-rater reliability study in which two independent researchers scored 31 de-identified MediCapt records [27]. For nearly all 26 checklist items, there was very strong inter-rater agreement; the one item with "moderate" agreement was the checklist item related to chief complaints, suggesting greater subjectivity in assessing these [27].
Kenyan data collectors were recruited and trained on use of the finalized checklist tool. Paper-based records were selected and reviewed at each of the two facilities. At each facility, five paper records per month (or all records in the rare months that had less than five paper records) were randomly selected for the 20 months preceding MediCapt training, for an estimated total of 100 paper-based records per site. This purposive stratified sampling by month helped ensure the sample was representative and minimize sample and selection biases.
For quality assessment of MediCapt records, a spreadsheet of de-identified MediCapt output for all completed MediCapt forms from both facilities was obtained from PHR.
Data quality scores for paper-based and MediCapt records were then compared. Given its ability for offline data collection and the PHR team's familiarity with the software, Kobo Toolbox (Cambridge, MA) was selected for data collection [28]. Data were analyzed using traditional descriptive analysis (e.g., means, standard deviations) and non-paired two-tailed ttests comparing data quality checklist scores for paper-based versus MediCapt forensic records. Statistical significance was set at p<0.05.

B. MediCapt usability questionnaires
To evaluate the usability and feasibility of the MediCapt application in these settings, a 10-page usability questionnaire (Appendix 2) was administered to local health professionals who had been trained on and had used MediCapt for forensic documentation [27].
The questionnaire consisted of both open-and closed-response questions. Open-response answers were organized into general themes and closed-response questions were analyzed using traditional descriptive analyses. The software Dedoose Version 4.12 (Los Angeles, CA) was used for coding and code analysis [29].

C. Key-informant interviews
Participants included clinicians, law enforcement, judiciary professionals, information technology professionals, and medical records personnel who interact with PRC and MediCapt forms and had been trained (from October 2018 to January 2021) or oriented on the MediCapt application.
Semi-structured interviews were conducted at each of the two sites at the evaluation's baseline (October 19-23, 2020) and endline (June 21-July 8, 2021) with clinicians, medical records, administration staff, law enforcement professionals, and legal professionals. Law enforcement and legal professionals, such as police officers, prosecutors, and magistrates, who participated in previous MediCapt sensitization sessions were also interviewed.
A semi-structured interview guide (Appendix 3) was developed and iteratively revised by the PHR team for use during the interviews. This guide was used for both the baseline and endline assessments. Consent was obtained from participants to take part in the study and for evaluators to digitally record all interviews. Verbatim transcriptions of these recordings were made and de-identified of any participant or survivor names or other identifying information. These transcripts subsequently underwent theme analysis with Dedoose [29] by two independent researchers using an inductive approach to identify emergent themes.

A. Data quality assessment of forensic records
The quality of 197 paper-based PRC forms and 139 MediCapt forms was evaluated using the validated data quality checklist. When compared, MediCapt forms more frequently had higher data-quality scores than paper-based forms (Table 1). MediCapt was associated with higher scores in 88.5% (23 of 26) of checklist items. The mean score for paper-based forms (n = 197) was 42.1 (SD 6.2, range 5-52), with 81 forms (41.1%) achieving the target score of 44 out of 54 (>80%). The mean score for MediCapt forms (n = 139) was 48.2 (SD 6.8, range 36-53), with 133 forms (95.7%) achieving the target score of 44 out of 54 (>80%). This difference in quality between the two form types was statistically significant (t-value -11.0, p-value <0.00001).
Overall, there was a total of 336 forms, with a mean data-quality score of 44.7 (SD 5.8).
Comparing the two sites, the quality of the paper-based data was statistically different. The paper-based data in Naivasha (n = 97) had an average data quality score of 40.1 (SD 4.0), while the paper-based data in Nakuru (n = 100) had a statistically higher average data quality score of 44.1 (SD 7.2) (t-value -4.82, p-value <0.00001).
The quality of the MediCapt data was statistically the same across sites. The MediCapt data in Naivasha (n = 91) had an average data quality score of 48.3 (SD 2.7), while the MediCapt data in Nakuru (n = 48) had an average data quality score of 48.1 (SD 2.4) (t-value 0.39, pvalue 0.70). Table 1 shows the quality of documentation in paper-based versus MediCapt forms. The checklist items for which MediCapt forms scored particularly high relative to the paper-based forms were clothing information (#12), statement in "Comments" summarizing the genital examination (#16), list of chain-of-custody samples (#20), and psychological assessment (Part B) (#24).
The three items for which the paper-based forms scored higher than MediCapt were the orphan-vulnerable children (OVC) status of the survivor (#4), date of last consensual intercourse (#11), and police officer signature and date (#22). The police officer signature and date prompt was also one of the two checklist items that tended to have the lowest data-quality scores across both types of forms, the other being the circumstances surrounding the incident (#7).

B. MediCapt usability questionnaires
Fourteen individuals completed the MediCapt usability questionnaire. These included clinical officers (n = 8), nurses (n = 5), and a gender officer / social worker (n = 1). Table 2 shows the results of the closed-response portion of the MediCapt usability questionnaire.
On average, respondents had 5.4 years (range 2-11) of experience conducting sexual assault examinations. Theyconduct 11.4 examinations (range 1-45) each month. They had each used MediCapt for an average of 11.2 times (range . Respondents reported that, at least initially, using the MediCapt form may take slightly longer than using the paper form. Questionnaire participants reported spending an average of 32.7 minutes (range 10-120) documenting with the paper PRC form and 36.8 minutes (range 5-60) documenting with MediCapt. Subsequent interview data presented in the next section suggest that this trend is reversed as providers became more familiar with the MediCapt application. For the examination of the survivor, participants reported spending an average of 25.7 minutes (range 10-60) using either of the forms. Listen to music 79% Experience using applications, or "apps," on smart phones 100% Ever used the camera function on a mobile phone or smart phone 100% Ever used a digital camera (not on a mobile or smart phone) to take a photograph 100% Normally take forensic photographs when you conduct sexual assault examinations (before using MediCapt) 0%

Agreement scores for various statements (listed from strongest disagreement to strongest agreement)
The use of MediCapt with a sexual violence patient is culturally unacceptable.
I have to rely on a generator to charge smart phones or tablets on a daily basis.
The tablet should be smaller. -0.9 I currently complete a paper-based medical certificate for examinations of all sexual violence patients.
-0.9 I find the text size on the screens too small. -0.9 The process of obtaining consent to use MediCapt for data collection is too cumbersome. -0.9 I am likely to lose my device.
-0.7 I think connection to the internet is a major problem for uploading files. -0.7 It is difficult to charge the smart phones or tablets on a daily basis. -0.7 The printer is likely to be stolen.
It is difficult to maintain printer supplies (ink, paper, etc.).

Position Clinical Officer (8), Nurse (5), Gender Officer/SW (1)
It is difficult to get reliable Wi-Fi or internet access to transmit the files. -0.4 The device is likely to get stolen. The printer is likely to be used by others for purposes UNRELATED to MediCapt.

0.4
It is easy to use MediCapt while I am conducting a sexual assault examination on a patient.

0.5
MediCapt helps me save time in documentation. 0.6 It is easy to troubleshoot problems that I encounter with MediCapt. 0.7 It is easy to connect the Bluetooth keyboard to the tablet. 0.8 It is easy to type using the keyboard. 0.8 I think that sexual violence patients accept my use of MediCapt during their examination.

0.8
It is easy for me to hold the tablet. 0.9 It is easier for me to take forensic photographs using MediCapt. 0.9 Printing the MediCapt document serves the patient well. 0.9 My colleagues will be happy using MediCapt. 0.9 Patients readily provide consent for MediCapt to be used in documenting their cases.

0.9
Printing the MediCapt document is more acceptable to me than sending the data electronically.

0.9
I think that the forensic photography function on MediCapt makes it more comfortable for my patients to be photographed, versus using a separate camera to take photographs.
1.0 I believe that my patients understand the risks and benefits of using MediCapt.

1.0
Additional measures will need to be put into place to make sure this device gets used.

1.0
Health care professionals who use MediCapt will take better forensic photographs because they are using MediCapt. I find that multiple places to enter information on a single screen makes data entry on MediCapt easy to use.

1.1
The different screens all made sense to me. 1.1 I like the prompts to take forensic photographs that were built into MediCapt.
1.1 I feel confident in my ability to explain to the patient the purpose and risks involving the use of MediCapt to obtain and record their information.

1.1
MediCapt is intuitive to my needs when documenting sexual assault examinations.

1.1
I am able to take forensic photographs easily using MediCapt.

1.1
It is easy to type on the tablet.

1.2
It is easy to take photographs using the tablet.
All participants were familiar with smart phones and typically use them for communication, looking up information, checking email, and taking pictures. No respondents had taken forensic photographs prior to using MediCapt.
Usability questionnaire respondents generally agreed that the tablets are useful and Medi-Capt is easy to use. Participants agreed that MediCapt is appropriate for use with survivors of sexual violence, MediCapt is acceptable to providers and survivors, and using MediCapt in these settings is both feasible and sustainable.

C. Key-informant interviews
Semi-structured key-informant interviews (n = 57) were conducted during this evaluation of the MediCapt program. The mean years of work experience among interviewees was approximately 10.5 years (range 0. [25][26][27][28][29][30]. Additional demographic information about the interview participants is presented in Table 3. Of the interviewees who participated in the baseline assessment (n = 24), 19 (79%) were available to participate in the endline assessment (n = 33). Of the 19, there were 15 (79%) clinicians and four (21%) non-clinicians (i.e., medical records/administration staff). The baseline did not have participants from the judiciary, prosecution, and law enforcement officers, the endline assessment engaged six individuals from these sectors. MediCapt will ensure that sexual assault records are transferred to the appropriate law enforcement and legal personnel.

1.3
MediCapt is better than what I am currently using to document sexual assaults.

1.3
The MediCapt app "made sense." 1.3 It is easy to use the touchscreen on the tablet. 1.4 The risks to the patient of lost personal information are greater with the paper form than with MediCapt.

1.4
I could one day train my colleagues on how to use MediCapt to document sexual assault examinations.

1.4
MediCapt offers a useful way to take forensic photography. 1.5 The screens appear to be straightforward and easy to use. 1.5 The tablet itself appears to be suitable to document sexual assault examinations.

1.6
MediCapt helps me do a better job of documenting sexual assault examinations. By design, the baseline assessment interview guide largely focused on the paper-based PRC forms. The endline assessment interview guide largely focused on the MediCapt application and how its use compared to that of the paper-based PRC forms. The codes that emerged from the interview transcripts were organized into five themes summarized in Table 4.
There were multiple features of the paper-based forms that interview participants appreciated (Table 4: Theme 1). The leading reported strengths of the paper form were its triplicate design and that it simply "works well." Participants also indicated that there was existing familiarity with the forms among people across all sectors. Respondents stated that paper forms capture most of the necessary details, which made them easy to review and navigate when pursuing cases.
Interview participants cited several weaknesses of paper forms (Table 4: Theme 2). The most frequently reported weaknesses were that the paper forms can be lost, changed or destroyed without permission and are not confidential. Logistically, participants also noted that the paper-based forms have multiple features that complicate their use, including limited space, requiring a lot of time to complete, being challenging to correct, often having missing or incorrect data, and not having a survivor consent section.
Many interview respondents felt the digital documentation tool had significant strengths and could address many of the weaknesses of paper-based documentation. The most frequently reported strengths of MediCapt (Table 4: Theme 3) included elements related to data security and completeness of the data collected. Respondents appreciated MediCapt's secure storage, ease of accessing stored data across sectors, confidentiality, and that tampering with the documentation was difficult. MediCapt assisted in the completeness of data collection with its greater time efficiency and convenience, ability to take photos, inability to skip questions    MediCapt improves services for survivors 0 16 16 All medical records should or will be digital 7 7 14 MediCapt stores more data / more room for answers 6 8 14 MediCapt reports are more legible 0 11 11 Copies of completed form can be printed 0 10 10 MediCapt is more readily available than paper 4 4 8 MediCapt data can be stored easily 5 2 7 Printer works well 0 3 3 MediCapt allows errors to be corrected 2 0 2   Never used or not currently using MediCapt application 0 11 11 Providers need to use MediCapt more 0 9 9 Long delays before court hearing 9 0 9 Multi-sector collaboration 4 2 6 More tablets needed 0

PLOS ONE
Effectiveness of a mobile application to improve forensic documentation of sexual violence (hence, enhancing completeness), required consent process, and overall comprehensiveness and accuracy.
Respondents were asked what areas of weaknesses they had noted with the MediCapt application (Table 4: Theme 4). The most frequently reported MediCapt weaknesses related to information technology issues included internet or logging-in difficulties, printer troubles (such as being inaccessible, slow, or jamming), and the requirement for typing proficiency to use MediCapt. Respondents also noted challenges with the length and format of the form printed from MediCapt, leading to difficulty presenting the form in court. Participants also shared that completion of MediCapt required more time and effort, and specific issues with the MediCapt form itself, such as small font, unclear format, and some repetition. Some respondents also noted difficulty obtaining consent for photography with MediCapt.
In addition to strengths and weaknesses of the paper forms and MediCapt, interview participants discussed several other relevant issues (Table 4: Theme 5). Most frequently, this included a request for more PHR/MediCapt training. Also, in contrast with some of the above interview data, but consistent with the usability questionnaire findings, most providers reported they were able to obtain consent from survivors using the MediCapt applicationeven consent for photography.

Discussion
This study demonstrates that MediCapt-a mobile application for the forensic documentation of sexual violence-more accurately, consistently, comprehensively, and securely collects the same forensic medical certificate data collected by the paper-based PRC forms in Kenya. In addition, the application has the unique ability to collect forensic photographic evidence that is not possible with the paper-based form. This study also shows that this mobile application is acceptable to the end users in the health, law enforcement, and legal sectors in Kenya. There is optimism amongst these groups that the forensic evidence collected through this application will be useful in the pursuit of justice and accountability for crimes of sexual violence. This study is unique in comparing the quality of data collected via a mobile application with data collected via paper-based forms and establishes important methodologies for evaluating these types of interventions going forward.
MediCapt data more frequently had higher data-quality scores than the paper-based forms, indicating that digital data collection of a standardized forensic medical certificate (in this case, the Kenyan PRC form) was of higher quality than data collected using a paper-based version of the same medical certificate. While end-user uptake is often presented as an obstacle to the implementation of mobile or digital technology over paper-based options, especially in low-resource environments, the results here suggest that uptake of these technologies was not, in fact, prohibitive in this context [30,31].
These results identify the use of a mobile application as a strategy that standardizes the collection of forensic evidence of sexual violence across different health facilities in the same jurisdiction. When comparing the quality of data collected using the paper-based forms, it was statistically higher in Nakuru compared to Naivasha. However, there was no statistical difference in quality in the MediCapt data across sites. An important feature of MediCapt that may support the increase in data quality is that incomplete fields are indicated to the user in the interface and on a summary page before completion of data collection. This feature serves as a reminder and an automatic data quality check built into the application. This is consistent with other studies, which have shown that routine data quality checks were associated with meaningful improvements in data quality of both paper and electronic medical records and increased concordance across multiple sites [32][33][34].
Although it is early to assess the impact of MediCapt on survivor outcomes, users were optimistic about its usefulness and reported that its legibility and photography features had already been appreciated by the court. The application's required consent process also greatly increased the frequency of providers seeking consent, helped empower the survivors in their care, and improved the quality of documentation and survivor-centeredness of the sexual violence services. This finding is aligned with other literature that has shown that focusing on survivor empowerment and feminist-informed models of engaging with survivors for documentation of sexual violence can both improve survivor experiences and documentation outcomes [17,18]. The process for obtaining informed consent should be interpreted as a survivor-centered strength of MediCapt, as the application prompts clinicians to engage with survivors, empowers survivors to make decisions about the provision of their information to law enforcement and to decide about the collection of forensic photographs for their case, and requires clinicians to take great care in obtaining consent for every step in the evaluation, including the photographs.
It is important to note that informants did share some conflicting feedback related to the ease of use and acceptability of informed consent and consent for photography processes in MediCapt. Respondents spoke both about challenges with collecting consent on paper-based forms and concerns regarding obtaining survivor consent in MediCapt. Concerns for the consent process in MediCapt seem to more often relate to the time that it takes to obtain consent in a meaningful way through the application and the different steps in the examination and documentation process.
While, at least initially, using MediCapt may overall take slightly longer than using the paper form, the usability questionnaire and interviews suggested that this trend is reversed as providers became more familiar with the use of mobile technology and the MediCapt application specifically. This indicates that not only does MediCapt improve the quality of documentation collected but, over time, it improves the efficiency of forensic documentation, giving clinicians more time for interaction with patients to ensure survivor-centered care and appropriate referrals to different services and resources. Additionally, data from the usability questionnaire and semi-structured interviews suggest that use of MediCapt is acceptable to both clinicians and survivors, which can likely be linked to the improved efficiency in documentation. The use of mobile health applications has been shown to be time-saving and to increase the efficiency of patient information collection [35]. Technology can have a mixed impact on clinicians' connection to or time spent with patients; while it increases efficiency, it should not be seen as synonymous with a reduction in time spent on patient care, particularly in cases of sexual violence [36]. This is why an emphasis is placed in MediCapt training on how clinicians can use technology with survivors in a way that is empowering and transparent.
Data security and privacy are issues that have been raised by end-users and stakeholders, especially because of the sensitive nature of the information and evidence collected. Among its key features, MediCapt includes sophisticated data encryption, high fidelity to chain of custody standards, and tamper-proof metadata. At the hardware level, end-users have unique credentials, passwords and PIN number that grant them access to the application and the forms they fill out. PHR works also with security experts to run audits to expose any potential gaps in the system ensuring that we proactively test and maintain the code. Currently the practice at most health facilities is that sensitive documentation, including paper PRC forms, are kept in the open, on desks, or stored in unlocked cabinets where any unauthorized individual can access them. MediCapt introduces additional levels of security and protection to ensure that these records are stored in a secured, encrypted space, allowing only those key individuals with passwords to access the sensitive information.
Major costs for MediCapt so far have included technology development for maintenance of the app and upgrading of the code and core implementation staff. Potential MediCapt costs for those implementing at the facility level will include technology maintenance expenses to facilitate the integration of MediCapt with existing systems for medical records, hospital costs for implementation materials (tables, printers, upgraded WiFi etc.) and capacity development expenses for training staff and administrators to engage with and effectively use MediCapt. There may be additional and ongoing resource needs including WiFi expenses and hardware such as tablets, printers, and ink. The results of this pilot are being used to understand how to deliver a cost-effective product at scale.
The results of this study show that MediCapt has the potential to dramatically improve how forensic evidence is collected in Kenya, and potentially in other low-resource settings at multiple levels: completeness, standardization, consent, and survivor-centeredness. Research has shown the need for strengthened tools and protocols to improve forensic documentation of sexual violence globally [37]. While few studies have documented the impact of forensic medical evidence in resource-constrained settings, one study found an association between documentation of anogenital injuries and convictions in South Africa [16]. Others have noted that the methods and practices for forensic evaluations more generally vary greatly amongst experts and centers [38]. MediCapt's reliance on a digitized standardized forensic medical certificate helps to improve the quality of forensic evidence and to ensure that the data the clinicians document and share with police is focused on forensic medical evidence useful for police investigations and the legal process. Furthermore, the standardization of forensic evidence collection at the core of MediCapt can help support efforts to provide more survivor-centered care by ensuring evidence is complete and not requiring repeat and potentially retraumatizing interviews with survivors [39].
The data from this study are stored in the Physicians for Human Rights' password-protected database which can only be accessed by the PHR Sexual Violence team. Due to the scope of participant consent, we request that the data not be available publicly, but can be accessed by PLOS One or other researchers, upon reasonable request, to validate our findings. To request for the data, please contact Dr. Michele Heisler at mheisler@phr.org, Medical Director and head of the ethics committee at Physicians for Human Rights.
MediCapt supports a broader range of trained health professionals in documenting high quality forensic evidence of sexual violence in contexts where those services are harder to access due to shortages in trained personnel [40]. MediCapt can contribute to survivors better realizing their right to health by ensuring that medical-legal examinations are more accessible, acceptable, available, and of higher quality [41]. Finally, survivors' access to justice may also be improved in the sense that, with MediCapt, there is higher-quality, probative evidence that is gathered to support their cases. Against this background, MediCapt is poised to become a robust tool to address impunity for crimes of SGBV in low-resource settings. For MediCapt and other similar applications to achieve these intended outcomes, it is important that there is a continued collaborative design approach with multisectoral actors to achieve maximum functionality and responsiveness of the intervention for all stakeholders [24].
There were limitations to this study. Clinicians were asked about their experiences before and after MediCapt introduction and, as some time may have passed since the initial Medi-Capt training, this may have introduced recall or response bias. As a result of any reliance on self-reporting, there is also the possibility of social-desirability bias. However, this potential bias was minimized by using external evaluators, encouraging participants to share both the good and the bad, and informing participants that responses will remain anonymous and only reported in aggregate. This study was conducted at just two health care facilities-both located in one county in Kenya. While we do not see obvious differences between these facilities and others in Kenya or beyond, they may not be representative and, therefore, their findings cannot confidently be generalized to other settings. Future evaluations can be conducted in other areas where MediCapt is in use, such as in the Democratic Republic of the Congo. Future research could explore acceptability of MediCapt technology to survivors of sexual and gender-based violence, what factors eased uptake of the mobile application, and how added time with survivors was spent, given the increased efficiency gained from use of the application. Finally, it is critical to understand what effects improved forensic medical evidence of sexual violence have on prosecutions and case outcomes.

Conclusions
MediCapt has been well received in Kenya across all sectors and has been shown to significantly improve the quality and standardization of forensic data collected across sites. It is anticipated that this improvement in forensic documentation will, most importantly, increase the likelihood of successful prosecutions, resulting in strengthening accountability for alleged perpetrators and improving access to redress and justice for survivors. An application like MediCapt has the ability to reduce impunity for crimes of sexual violence and ensure that a survivor-centered approach remains at the core of sexual violence response.

Inclusivity in global research
Additional information regarding the ethical, cultural, and scientific considerations specific to inclusivity in global research is included in the Supporting Information (SX Checklist) in S4 File.