OptoSearch is a forensic identification tool that allows the user to quickly calculate the frequency of occurrence for specific or generalized eye or eyeglass prescriptions. The search program allows the investigator to determine the frequency of a prescription in a population and therefore, determine the strength of match between an unknown and known sample. Please see Berg and Collins (2006) for a detailed description of the use of OptoSearch.

The investigator can simply input a complete prescription into the search function and OptoSearch will calculate the frequency of match within the selected database or population. A second option is the investigator can input partial or incomplete prescription information and search solely on selected variables. Third, the investigator can structure queries that accommodate biological information such as age, sex, or ethnicity. Again, the program calculates the frequency of match using the parameters of the investigator’s guidelines. The frequency of match can then be incorporated into the identification process or be used as a means to create a “short-list” of possible individuals based on the strength of match.

The lens fragment or complete lens must be analyzed prior to the use of these databases. Lenses should be analyzed using a lensometer, such as the Humphrey® 350 Lens Analyzer (a three point laser lensometer) though any manual lensometer can be used to determine refractive errors from optical materials (it must be used by a trained person). The lens side determination should be made, if possible. Refractive errors are measured using three variables: sphere power (sphere), cylinder power (cylinder), and the axis of the cylinder power (axis). Sphere and cylinder powers are measured in quarter diopter increments. Though prescriptions can occur with very high numbers, the sphere correction rarely exceeds -15 diopters (myopic) to +15 diopters (hyperopic) in each eye. Cylinder corrections (for astigmatism) are measured from 0 to -10 in each eye. While the typical correction ranges are given for these variables, some corrections fall outside of these parameters. The axis is measured in single degree increments from 0 to 180. Also, bifocal corrections can be measured and used in one database. Bifocal corrections are typically between 0.75 and 3.5 diopters.

An additional technique can be employed, which is that of a tolerance match. A tolerance match is an inherently a more conservative estimate of the frequency of a given prescription because it takes into account potential manufacturing errors as well as slight changes to the vision of a given patient. A tolerance match includes ±0.25 diopter magnification variation for both sphere and cylinder categories, as well as ±3 degree variation for the axis category. The variations are from the acceptable limits as published in the ANZII standards for the manufacturing of prescription eyewear.

The Humphrey® lensometer can reliably read the refractive error from a fragment of optical glass smaller than 1 cm2. While the axis variable can not be obtained from loose fragments that do not have a mountable edge portion, sphere and cylinder variables are unaffected by this constraint. Heavily scarred or pitted glass fragments may not produce results, though the laser lensometer often can mitigate these problems. The measuring procedure is quick and reliable. The refraction strength is determined by placing the lens into the lensometer and observing its digitally-measured strength. Typically, three trials are performed, ensuring an accurate determination.

After a prescription has been read from an eyeglass, a search for its frequency can be undertaken. On the OptoSearch entry page, enter a unique case number or other identifier in the box provided (this step is optional and is only used for display and printing purposes). Three databases can be used (see Databases). Select the appropriate database from the drop-down menu, or use the combined database for general queries.

Input the prescription in the line blocks provided, using a negative cylinder format [all cylinder entries need to be prefaced with a negative (-) sign]. The prescription should be input by right or left eye, or both if known. If only one side is known, input it in the appropriate block; if the side is unknown, use the unknown line. One or all variables can be used in the analysis. The axis variable can only be used when the side is known. Bifocal corrections can be used in conjunction with either eye or the unknown line. At this time, a tolerance match can be selected by checking the Tolerance Match box (a tolerance match adds ±0.25 diopters to the sphere and cylinder variables and ±3 degrees to the axis variables – see the Introduction). After the data is entered, press the Process button, which will then return the result in a pop-up window. The data can be cleared from the input boxes using the Reset button.

All searches will be conducted using an exact match format unless a Tolerance Match is requested; this means that the databases will only be searched for the exact criteria entered. A Tolerance match will return a larger set of results (see above). Returned results will include the following data: the upper block contains the case number, search date, and selected database. The next line includes the calculated frequency of match and the formula which it was derived. The final output contains the search parameters (input values), and whether or not a Tolerance Match search was conducted. The results page can be printed for inclusion into a case file or for the investigators’ convenience.

Databases:

The OptoSearch program allows the user to search three different databases or a combined database. The largest database utilized in this program is prescription data from the Naval Ophthalmic Support and Training Activity (NOSTRA) located in Yorktown, Virginia. This database represents approximately 40% of all Department of Defense ordered eyeglasses from across the United States. The database includes approximately a decade of refractive error information, from ~1992-2002. It contains over 370,000 individual prescriptions representing 750,000 individual eyes. While the database contains no biological information regarding sex, age, or ethnicity, it does contain information pertaining to an individual’s military or civilian rank, branch of service, and lens style.

A cautionary note on the use of the NOSTRA database: this information source was compiled from orders placed to NOSTRA from various military optometrists and ophthalmologists around the world. Each prescription was given a unique order number, regardless of the patient. Thus, if a patient ordered regular glasses and sunglasses, two unique orders were filled at NOSTRA, thereby creating duplicate information in their system. We have chosen to conservatively minimize duplicate data by removing one or more consecutive orders if all categories of information were exactly matching. By undertaking this process, approximately 400,000 prescriptions of the original 1.2 million were removed from the database. Duplicate prescriptions still occur in the database primarily due to a delay in ordering another set of glasses, (e.g. the original order was followed by another order three weeks or three years later). Since we cannot realistically remove all duplicate data, some remain present. Thus, when frequencies are calculated using the NOSTRA database, a degree of interpretation must accompany the results. The frequency of a common prescription in the NOSTRA database may actually be slightly less common than the results indicate due to duplicate data. For rare prescriptions, it is highly likely that the calculated frequencies are accurate. Any unique prescriptions are, obviously, unique.

The second largest database is derived from a recent multi-year study conducted by the National Center for Health Statistics on a U.S. population sample. The National Health and Nutrition Examination Survey (NHANES) compiled biological information on approximately 20,000 study participants of which we have selected approximately 8,000 prescriptions for use in our database. Please see Berg and Collins (2006) for criteria regarding individual prescriptions in our database. Our NHANES database includes males and females from 12 to 84 years in age, all with self-reported ethnicity. Forty-seven percent of the data base is male and fifty-three percent is female. The self-reported ethnicity is 41.3% White, 21.5% Black, 28.6% Mexican-American, 5.1% Other Hispanic, and 3.5% other ethnicity.

The CILEPI database is a continuously open survey being conducted at Lackland AFB Optometry clinic in Texas and the 15th Airwing Optometry clinic in Hawaii. The survey participants are chiefly active duty military personnel, but their dependants and other individuals eligible for treatment are included. Currently, the database contains slightly over 4500 individuals and all participants have recorded sex, age, and self-reported race. The majority of the study participants are male, and the participants range in age from 4 to 95 years. For this study, the CILEPI database recognizes the following racial categories: White (62.4%), Black (17.1%), Hispanic (13.1%), Asian (4.5%), Native American (1.6%), and Pacific Islander (1.3%).  

This study presents a web-based tool that can be used to assist in identification of unknown individuals using spectacle prescriptions. Currently, when lens prescriptions are employed in forensic identifications, investigators are constrained to a simple “match” or “no match” judgment with an antemortem prescription. It is not possible to evaluate the strength of the conclusion, or rather, the potential or real error rates associated with the conclusion. Three databases totaling over 385,000 individual prescriptions are utilized in this study to allow forensic analysts to easily determine the strength of individuation of a spectacle match to antemortem records by calculating the frequency at which the observed prescription occurs in various U.S. populations. Optical refractive errors are explained, potential states and combinations of refractive errors are described, measuring lens corrections is discussed, and a detailed description of the databases is presented. The practical application of this system is demonstrated using two recent forensic identifications. This research provides a valuable personal identification tool that can be used in cases where eyeglass portions are recovered in forensic contexts.

Prescription eyewear can be used to aid in forensic investigations. Until now, investigators and consulted eye professionals have been limited to a simple “match” or “no match” judgment. This article introduces to optometry a web-based tool that can be used to assess the strength of a match between spectacle prescriptions and recorded patient information. Three databases with over 385,000 individual prescriptions were used to create the web tool that calculates the frequency with which a prescription occurs in various U.S. populations. A search for any prescription in the tool’s database will result in a report of the number of times a given prescription occurred in the queried database(s) as well as the calculated frequency with which the combination of the given sphere power, cylinder power and axis are likely to occur. Practical application of the web tool in two published cases has shown matches with frequency of occurrence of 5.33 x 10^-6 and 2.66 x 10^-6, respectively. This application is currently being used by the Joint POW/MIA Accounting Command Central Identification Laboratory (JPAC-CIL) and other agencies when optical materials are available in forensic settings. Further, this application is currently contributing evidence in a high profile murder case. The creation of this easy-to-use web tool allows optometrists and other eye care professionals to provide strong statistical assessments when serving as consultants to forensic investigators.

In addition to active duty military members, retired military members and the members of their immediate family member are eligible for eye care in military medical treatment facilities (MTFs). We recorded refractive errors, age, sex and race for 4,595 individual beneficiaries visiting optometry clinics at two U.S. Air Force MTFs during 2005-2006. Evaluation revealed that most patients requiring optical correction were myopic, or near-sighted, and that there is an increase in the degree of myopia between ages four and 23. That trend is reversed at age 30 and, by age 60, most patients are hyperopic, or far-sighted. Both trends were true for both sexes and all ethnicities studied. The degree of astigmatism was distributed similarly between races and age groups. Presbyopia occurred at similar ages and progressed at similar rates in all ethnicities and both sexes.

The identification of human remains is a primary focus of forensic specialists. In many instances, the results from medicolegal examination, odontology, anthropology, and nuclear or mitochondrial DNA analysis can identify unknown individuals. Alternate lines of non-biological evidence, such as identification cards, clothing, and shoe wear are often used as corroborating evidence. Using spectacle prescription data is not a new idea to law enforcement and forensic specialists, but opticians and doctors are usually constrained to a simple 'match' or 'no match' conclusion with the prescriptions listed in medical records. The web-based tool introduced in this paper will let doctors, analysts, and investigators easily determine the strength of individuation by calculating the frequency at which the observed prescription occurs in various U.S. populations.

The available databases draw from both military and civilian sectors of the U.S. population and currently contain more than 1.2 million individual eye prescriptions. An additional dataset contains approximately 4000 individuals with self-reported biological data (sex, age, and ethnicity). While the bulk of the prescription data is linked with individuals of military service, civilians from the Department of Defense and dependants of military service members are also included. Additional information available in the largest database includes rank or grade, job type, and type of glasses. General population information is available for the smaller database, which contains approximately 65% males and 35% females. Reported ages cluster around the late-teens to mid twenties, though every age is represented from 4 to 95 years. Self-reported ethnicity is largely White (60%), with other major ethnicities present (Black 15%, Hispanic 15%, Asian 4%, Native American 3%, Pacific Islander 1%, mixed 1%).

The web-based tool introduced here allows the user to search for matching prescription information within each database. The databases can be queried for any combination of the corrective states including sphere, cylinder, cylinder axis and bifocal powers for each eye. The sphere and cylinder corrective powers are typically measured in increments of .25 diopters, while the axis correction is on a 180 degree scale. These variables have a respective minimum of 80, 72, and 180 possible conditions, giving a total of 1,036,800 possible combinations per eye, or 1x1012 combinations for both eyes (exclusive of bifocal corrections). As with many other types of biological data, some corrections are more common than others; common single eye corrections (using sphere and cylinder corrections only) may occur in about 12 per 1000 individuals though common dual eye corrections drop to approximately 2 per 1000. If the axis correction is added to the query, the frequency can drop to 1 per 10,000 or greater.

As eyewear is typically directly related to the genetic make-up of an individual (trauma and surgery are the major exceptions), prescription data is highly individualized. Further, the frequency of a given prescription can be combined with frequency data for other independent biological information such as dental or mtDNA data, to provide extremely strong statistical estimates of the likelihood of individual identification. We will demonstrate these applications through several cases, particularly those dealing with the identification of fallen U.S. service personnel, as conducted by the Joint POW/MIA Accounting Command in Hawaii.