Use of Telemedicine to Diagnose Tinea in Kenyan Schoolchildren

Introduction

In Kenya and other East African countries there is approximately one doctor for every 50,000 people compared with one doctor for every 390 people in the United States. ¹ Because of the extreme shortage of physicians in Kenya, patients often have to travel far to reach remote clinics and consequently delay their visit while their condition worsens. At hospitals and clinics patients may have to wait hours or days to see a physician. Increasing the number of healthcare professionals is essential to improve healthcare access in many developing nations, but this requires considerable time and resources. Other approaches should be explored and implemented alongside these goals to improve healthcare access and effectiveness.

Technological and communication innovations promise to address some of the gaps and inequalities in healthcare faced by much of the world's population. One such innovation is telemedicine. The rapid expansion of the Internet, cell phones, and electricity sources in developing nations increasingly connects rural populations to trained healthcare providers. Students and faculty at The Pennsylvania State University have developed and field-tested a telemedicine system called Mashavu to provide increased access to healthcare for patients in underdeveloped countries.

A Mashavu kiosk is a portable, user-friendly station operated by community health workers in rural areas without medical clinics. It allows patients to be evaluated by medical professionals without having to travel long distances. The kiosk operator logs onto the Mashavu server to create an account for a new patient or to update a returning patient's profile. Data, including vital signs and exam findings, are transmitted to a clinician for evaluation.

This study investigated the usefulness of a new addition to the Mashavu kiosk (a digital Web camera) to diagnose common skin conditions such as tinea. ² Tinea, also known as ringworm, is caused by a group of dermatophytic fungi of the genuses Trichophyton and Microsporum. Tinea may affect the body, scalp, groin, or feet (Tinea corporis, Tinea capitis, Tinea cruris, and Tinea pedis, respectively). Tinea is very common among children in East Africa; the prevalence in Kenyan primary schoolchildren has been estimated at 9.3–33.3%. ^3,4 It is spread by direct contact with affected people or animals and by contaminated objects. It usually presents as itchy, raised, scaly patches with hair loss. The lesions often have dark-red, clearly defined edges that surround a central area of normal skin tone, often appearing as a ring. ⁵ Although tinea is not fatal, it can lead to secondary bacterial infections as well as social stigmatization. In addition, affected individuals serve as reservoirs of infection. ⁶ The gold standard for the diagnosis of tinea is a potassium hydroxide (KOH) preparation. In this technique, skin scrapings of the leading edge of the affected area are obtained using a scalpel blade and examined via light microscopy.

This study served as the first clinically relevant test of the Mashavu system by determining whether Kenyan clinicians could reliably diagnose tinea infections using photographs of lesions taken with a digital Web camera. Although the utility of telemedicine to evaluate skin conditions has been demonstrated in several studies, most of these studies applied it as a referral method for physicians and not as access for primary care. ^{7 –9}

Subjects and Methods

The research took place in Nyeri District, Kenya, from May 16 to June 7, 2011. Approval was obtained from The Pennsylvania State University Hershey Medical Center Institutional Review Board and from the Kenyan Ministry of Health. Children were recruited from a local primary school and a local church and by visiting homes in the town of Thungoma in Central Kenya. At Thungoma Primary School, the teacher identified children who had suspicious lesions for recruitment. At the church and homes, parents were asked if their child had skin conditions and would like to be part of a research study. All parents signed a consent form, and the children gave verbal assent. A total of 32 children participated in the study. The ages ranged from 6 to 17 years with a mean age of 9 years. Seventeen were boys, and 15 were girls. Each child received a brief physical examination supervised by a Kenyan nurse to identify the skin lesion. A digital picture was taken of the affected area with a low-resolution 640×480 pixel image Web camera. Pictures of skin lesions on faces were limited to the area of concern, and the photographs were altered by covering the eyes with a black bar to prevent identification. The photographs were uploaded to a computer and saved with unique and anonymous identifier. Each subject was assigned a unique ID number documented in a log book with the file information. Some volunteers had more than one lesion photographed. Forty pictures were taken of skin lesions. In each case a skin sample was taken, and the diagnosis of tinea was confirmed or refuted by KOH prep.

In total, 40 pictures were taken, of which 20 were of T. capitis and 2 were of T. corporis; presumed diagnoses for the remaining (negative) images were eight infected wounds, four scars, one mouth sore, one case of scabies, one of vitiligo, two nonspecific rashes, and one healthy scalp.

Using these pictures, we assembled a multiple choice test containing 15 pictures of tinea and 15 pictures of other lesions, wounds, or normal anatomy. Each question on the test asked to identify the lesion in the photograph as (A) scabies, (B) leishmaniasis, (C) fungus/dermatophyte, (D) bacterial infection, or (E) other/undecided. Test subjects were recruited from Nyeri Provincial Hospital. Following informed consent, each participant was issued a unique ID number and asked to complete a questionnaire to obtain demographics and information about their level of training. In total, six physicians (locally known as medical officers with a 6-year degree), 6 physician assistants (known as clinical officers with a 2-year diploma), and five nurses (2-year diploma) were tested. The study was powered to detect a sensitivity and specificity of 75% for each group with a 95% confidence interval of 53–97%. Clinicians were told they would take a test where they would be shown a series of 30 pictures and asked to diagnose different skin conditions. The specific focus of the study was omitted in order to avoid bias. No time limitation was imposed, but most clinicians finished within 20 min. The answers were recorded on a standardized answer sheet with the participant's unique ID number printed on the top. A binomial regression analysis was applied using a logistic regression model to estimate sensitivity and specificity for each of the three groups and to compare the three groups. A Bonferroni correction factor of 3 was imposed to account for the multiple comparisons. In addition, we calculated kappa statistics to determine inter-observer agreement within each group. All calculations were performed in SAS^® version 9.3 (SAS Institute Inc., Cary, NC).

Results

Table 1 summarizes the results. The mean (±standard deviation) values of the sensitivity and specificity for the whole cohort were 73% (±19%) and 83% (±11%), respectively. Physicians had the highest sensitivity and specificity of all three groups, although only sensitivity reached statistical significance when compared with nurses and physician assistants and after correction for multiple comparisons. Although physicians also had the greatest specificity, there was no statistical difference in this parameter among groups. The kappa statistics for each group were as follows: physicians, −0.17 to 0.47; nurses, −0.14 to 0.52; and physician assistants, −0.06 to 0.55. These wide ranges suggest that the level of agreement within each group ranged from nonexistent to modest.

Discussion

The use of telemedicine in primary care has been explored in various settings. ^10,11 However, the application in which it stands to make the most impact is in bringing health resources to the underserved. In this study we explored the diagnostic utility of a key component for the overall health evaluation of a person: a Web cam for the evaluation of common skin conditions. Our results demonstrate that Kenyan clinicians were able to correctly diagnose most cases of tinea infections from photographs taken with a low-cost digital Web camera. The high prevalence of T. capitis in our study population could be considered a limitation because clinicians may have been biased toward identifying tinea upon seeing a photograph of a scalp lesion. However, most individuals failed to accurately identify all cases of T. capitis. The lower sensitivity of physician assistants may be due to the fact that a few of these were recent graduates. Regardless, it is likely that the performance can be improved with training.

Our results have important implications for the future of telemedicine systems in areas of limited access to healthcare because they suggest that such a system can accurately diagnose certain medical conditions without requiring a visit in person. Therefore, this system could serve as a gateway for primary care for straightforward medical conditions. Dermatological conditions are ideal for telemedicine evaluation because in many cases the diagnosis can be made by direct visualization of the lesion. In the future, the technology could also be applied to other diseases with greater morbidity and mortality. This can potentially benefit many people in areas where access to healthcare is limited, from developing nations to rural Appalachia.

A 1-year pilot of the Mashavu system is currently underway with two kiosks distributed to six locations. There are many challenges that need to be overcome prior to the successful large-scale implementation of Mashavu. Many of the physicians in the study, although excited about the idea from a clinical perspective, expressed reservation due to infrastructural challenges such as lack of reliable electricity. Mashavu is partnering with the United National Industrial Development Office to initiate telemedicine at their renewable energy kiosks in remote areas. It is important that our study suggests that the use of nurses can be a viable alternative to physicians or clinical officers. Further studies should be carried out with increased sample size to confirm our preliminary results. Despite challenges, we believe that systems such as Mashavu will bring greater access to healthcare for underserved populations.