Hair Loss Evaluation Efficiency Study
Vanderbilt University
Department of Biomedical Engineering
April 22, 2003
Peter Claise
Barb Visher
Advisors:
Dr. Paul King
Dr. Lloyd King
Dr. Sara Whitehead
Dr. Jennifer Dempsey
ABSTRACT
Alopecia, baldness occuring in both males and females of all ages and races is still largely a mystery to physicians and researchers. Genetic correlations have been attributed to androgenetic alopecia while autoimmune diseases have some linkage to alopecia areata. Unfortunately, positive treatment outcomes are still sporadic with little few breaking developments in the fields.
Although hair loss is a significant problem, research efforts in the US have been sporadic and incoherent due to lack of clinical communication and a centralized information system for collected data. Therefore, efforts for learning more about and stopping hair loss are slow and tedious. The goal of this project is to assist in the overview, containment and efficiency of a useful hair loss study by creating a database with a wide variety of fields that are easily sortable for a large number of records in order to allow researchers the ability of developing genotypic correlations regarding hair loss.
After attempting several database formats, the final database project is a complex interlinking with defined relationships consisting of tables, forms and queries. The data is acquired through a series of forms that are connected through linking relationships. The majority of the fields in the database are stored as either “yes/no” or “text” fields, thus minimizing user error and maintaining consistency.
Efficiency in patient meetings was noted with significant implications on the possible increase of patients seen overall. Not only was patient waiting time decreased, but time spent between doctor and patient was optimized by focusing discussion on current problems, changes and treatments rather than medical history. Future possibilities of a larger scale trial would yield large amounts of data necessary to determine correlations in the etiology and pathology of alopecia.
INTRODUCTION
Hair loss affects millions of Americans annually. Loss of hair may be related to a hormonal imbalance, poor receptor biomechanics, follicular detriment, malnutrition, severe stress, disease, genetics, allergies and/or medication. Types of hair loss range as vastly as the causes to include alopecia areata, androgenetic alopecia, alopecia totalis and alopecia universalis.
Alopecia areata, an autoimmune disorder of unknown cause, affects 4 million people in the United States alone, and 1.7% of people globally (www.alopeciaareata.com). With a prevalence equivalent in males and females, alopecia areata is characterized by patchy, circular, bald spots through out the scalp. Progression of the disease may lead to alopecia totalis, complete scalp hair loss, or eventually alopecia universalis, total body hair loss. However, diagnosis of alopecia areata does not have to precede the latter two. Together alopecia totalis and alopecia universalis affect 800,000 people in the US. While no gender, race or age group is immune, occurrence is most often diagnosed in children. Heredity influences the likelihood of alopecia areata as 20% of those afflicted have a relative who also suffers from the disease (www.alopeciaareata.com).
Hair loss is often temporary with alopecia areata. However the course of the disease fluctuates and recurrent episodes may be detected in up to one-third of alopecia areata patients (Goldstein et. Al.). Atopy, a prepubertal state, widespread development (alopecia totalis or alopecia universalis), pathology duration greater than five years, or peripheral scalp involvement may all decrease the potential for regrowth (Goldstein 5). Regardless of the extent of hair loss, follicles lay dormant until appropriate signals are received to stimulate normal hair production (www.alopeciaareata.com).
Approximately 95% of hair loss is caused by androgenetic alopecia, commonly known as male or female pattern baldness (www.hairlossfyi.com). Known for its genetic linkage, androgenetic alopecia is characterized by a receding hairline and centered bald spot (which eventually may meet) in men and a general thinning at the top of the head in women. Common patterns of hair loss for men and women with male and female pattern baldness respective is shown if figures 1 and 2.
However, some argue that female pattern baldness may not necessarily be the equivalent counterpart of male pattern baldness. “The majority of women with pattern hair loss do not have the degree or the synchronization of miniaturization in a given region of the central scalp as men do and, thus, do not have the same recognizable patterns of hair loss as in effected men” (Olsen 2003). Research is currently studying ovarian and adrenal hormones to better understand the etiology of female androgenetic alopecia in fear that development of new effective therapies are being hindered due to an over generalization of defining characteristics in male/female pattern baldness (Olsen, September 2001).
Diagnosis of alopecia is variable; literature on the topic fluctuates in the diagnostic symptoms used to categorize patients. Universally, smooth and discrete areas of hair loss in round patches are noted for diagnosis of alopecia areata. Examination of hair roots and follicles also yield pertinent information for diagnosis and pathology. Biopsies are not standard protocol for general diagnosis but do assist in more specific categorization.
Available treatments currently offered to reduce the visibility of alopecia include topical and injectable medications including minoxidil, propecia, and rogaine. Laser assisted hair transplantation and hair prostheses are also common tools in dealing with alopecia symptoms. Current research strongly supports gene profiling and hair cycle control to diminish alopecia’s results. Unfortunately, findings are still non specific and small scale with limited expectation of near future solutions. While some reported studies show that certain patients do benefit from treatment, often times statistical analysis is not done to truly represent the likelihood of causation between treatment and regrowth and reported data is therefore skewed.
Although hair loss is not terminal, results are profound due to the unpredictable nature and recurrence of alopecia. Society places a high value on personal appearance and hair is a large source of beauty and self-confidence for many. The sudden theft of this visible trait can be a difficult and painful process for anyone to endure, especially a young child. Emotional consequences are often times larger than medical consequences resulting in years of therapy, and low self esteem, in addition to countless treatments for the disorder. Additionally much of society is unaware of the disorder and its symptoms. However, the problem is not just cosmetic, much speculation and study has linked alopecia to a range of autoimmune diseases including rheumatoid arthritis, diabetes, lupus, and multiple sclerosis.
Currently no universally effective treatments exist. However if the etiology of alopecia is better understood, great strides can be made in understanding, localizing and preventing pathological recurrence (Brzezińska-Wcisło et. Al.). Despite the significance of the hair loss problem, research efforts in the US have been sporadic and incoherent due to lack of clinical communication and a centralized information system for collected data. Therefore, efforts for learning more about and stopping hair loss are slow and tedious. Etiology, pathology and personal patient history must be presented on a clinical scale to facilitate a wide pool of patient information for further study. Therefore, the goal of this project is to assist in the overview, containment and efficiency of a useful hair loss study by creating a database with a wide variety of fields that are easily sortable for a large number of records in order to allow researchers the ability of developing genotypic correlations regarding hair loss. Because of the large volume of data and the possibility of future widespread use efficiency in data collection, requested information and storage is essential for an effective study.
METHODS & RESULTS
Several possibilities were proposed for creation and set up of a hair loss database. However, in choosing parameters and an overall organizational structure of the intake information many trials were attempted. Originally, examples of phone screenings and arbitrary trial questionnaires were studied to gain a sense of understanding regarding the organizational structure and general information necessary. Personal patient information is now separated from all other relative medical screening to insure patient privacy protection in accordance with the Health Insurance Portability and Accountability Act of 1996 (HIPPA).
The first database trial consisted of mostly open ended questions, which to date had been recorded by hand. Although the general fields and overview were a valid first attempt, this first trial allowed for far too much variability in the data within a given field. Additionally, because the attempt was made to encompass all relative topics in regard to hair loss from total family history to a description of all discernable phenotypic characteristics organization was muddled.
Further attempts led to drop down boxes, lists, yes/no choices and check boxes, leaving little to no room for severe variability in answer types. While answers may vary significantly (with over 50 past medication possibilities, and any combination of the choices), human error is drastically decreased because the clinicians are no longer responsible for typing in the answer. Therefore spelling errors and compatibility errors (1 vs. one) are negated.
In order to increase efficiency and organization a branching structure was proposed as seen in figure 3. The purpose of this format was to direct data acquisition to only those topics associated with the patient’s condition and minimize extraneous lines of questioning as well as time spent per visit. However, in assembling such a structure the overall purpose of obtaining all information that would later be useful in determining genetic correlations was lost. This method reverted back to storing only that information which appeared to have a direct relationship with the prevalent disorder rather than storing all possible data factors to be later sorted and evaluated.
The patient evaluation is now reformatted to encompass all characteristics associated with hair loss in an organized and convenient manner. The final database project is a complex interlinking with defined relationships consisting of tables, forms and queries. The data is acquired through a series of forms that are connected through linking relationships. There is a series of 12 tables consisting of “General Information”, “Characterization of Main Complaint”, “Past Medical History”, “Family Medical History”, a series of general and specific symptoms, “Physical Exam”, “Labs” and “Common Drugs.” These sections could not be separated perfectly into one table each because of the large quantity of information and the limit to the number of fields in a given table by Microsoft Access. Each table can only store a finite number of fields so it is essential to have multiple tables to store the immense amount of information.
Each table is maintained and organized for later use with specific “Primary Keys.” The “primary key” is a single field in the table, which is used to automatically sort all fields within the table. “Primary keys” must be unique so these were artificially created in most cases due to the reproducibility of the information stored in the database. Additionally, “primary keys” are essential in linking different tables to one another. The linking of tables is referred to as “relationships” and is necessary to combine multiple forms and tables in searches and analyze the data inputted.
Each field in the table is associated with a specific data type such as “yes/no”, “number”, “text”, “memo”, “date/time”, “currency” and “hyperlink.” The majority of the fields in the database are stored as either “yes/no” or “text” fields, thus minimizing user error. Additionally, human error is further minimized by using “combo” and “list boxes”, also known as drop boxes that store preset text answers for text inputs in a given field, such as race (see figure 4).
A series of forms were also created to assist in user input. Multiple forms allowed for an artificial tracking number to be created using “primary keys” for each patient to comply with HIPPA regulations. The multiple forms also make it more efficient for the user to enter information into the database because they do not have to scroll down (see figure 5). Multiple forms allow for quick movement onto the next page while still looking at the work they are doing. In addition, on the forms themselves, precautions were taken to ensure the correct and useful data is entered into the database. Access 2000 is a very helpful program in that if the field is set to “numbers” and the user inputs characters then an error message immediately comes on the screen and will not let the user move on until the error is solved. All values entered into the database that are not “toggle buttons” or combo boxes have some degree of freedom which can create bad data and produce an inefficient database.
However, these errors are minimized with a series of specific rules set in Access such as “validation rules”, “default values” and “input masks.” The “Input Mask” property is used to display literal display characters in the field with blanks to fill in. For example, if all phone numbers you enter in a field have the same format, you can create an input mask such as a spacing (___) ___-____. This minimizes the likelihood for the user to misinterpret the information wanted. “Validation rules” block the user from entering data that does not correspond to the field. These can be set by altering the properties of individual fields to fit the specifications. For instance, someone’s age is stored as a number in the field “age.” A number can be anything from negative infinity to positive infinity but by setting reasonable “validation rules” the user is limited to entering a number from 0 to 110 assuming that no one over 110 is going to be participating in the study. “Default values” also assist in this process so the user knows what type of entry is expected, for example number of nails affected is set at 0 until an amount is added thereby assuming the patient has no nail problems until otherwise specified. (See Appendix A for a complete patient visitation evaluation form used to create the prototype database).
With good, helpful data stored under fields in tables within the database it is now possible to use the abundance of information. Access 2000 has the ability to enter a number of queries to sort and filter the data in a useful manner. The user sets queries by selecting tables and fields within the tables and isolates data stored in the field that matches the criteria the user sets. Some sample queries that were created examine all individuals that have had symptoms since they were younger than 10, have a family history of asthma, have a family history of psoriasis, experience leg swelling and have been on the medication Amantadine. The results received were high due to the artificial data inputted for testing purposes.