Assessment Of Amputee Rehabilitation Using A Text-Generating Data Processing System
Peter H. Stern, M.D.
In 1974, The Burke Rehabilitation Center began a major effort to computerize its medical record system in an endeavor to simplify the ever increasing demands for documentation in health care. Previous experience with a terminal-oriented, time-sharing computer system called APL convinced us of the practicality of using computer-generated English-text discharge summaries for major disease categories that can be described in a relatively finite number of variables. In a previous publication the procedures required for stroke rehabilitation discharge summaries were described.
This paper is concerned with discharge summaries for lower-limb amputees that were referred to the Burke Rehabilitation Center during the period between 1/1/74 and 9/20/75 and with the concurrent establishment and analysis of a data base accumulated during this period.
The computer system, APL (A Programming Language), consists of a terminal connected by telephone equipment to a remotely located central IBM-370 computer (Fig. 1 ). It is a time sharing system; that is, many terminals are connected simultaneously. A new general purpose program, APG (A Program Generator), is added for our purposes. The APL/APG system is highly interactive, user oriented, and does not require any special knowledge in computer sciences or mathematics. At the Burke Rehabilitation Center medical, nursing, and clerical personnel are able to operate the terminals with very little instruction.
The user, in this case the physician in charge of the Amputee Service, constructs a questionnaire type discharge summary work sheet as shown in Appendix A. The encircled numbers are used to generate the English text discharge summary. The programmer is provided with a sample prose; upper and lower bounds for queries such as laboratory values for the inclusion of validity checks, and a general idea of the data that might be subject to calculations, correlations, or tabulations.
In order to generate a discharge summary the attending physician simply encircles the appropriate answers to the prepared series of queries and if necessary completes the free text provisions which are of fixed character length. The data is then entered by a terminal operator or the physician himself. It is retrieved either as an English text discharge summary (Fig. 2 ) or as part of a statistical report, the format of which are predetermined in the preprogramming and programming phases (Table 1).
Table 1 (cont.), Table 1 (cont.)
During the period between January 1, 1974 and September 20, 1975, 127 amputee patients were discharged from the Burke Rehabilitation Center. The tabulated results follow.
Equal sex distribution and an average of 65 years (range 17-90) shows that the elderly "vascular" amputee is the major public health problem in amputee rehabilitation.
Length of Stay (L.O.S.)
The mean L.O.S. of the entire group was 45 days (median 39). If bilateral and asymmetrical amputees are separated out, the mean length of stay dropped to 35.8 days. The L.O.S. of the bilateral below-knee amputees was 56 days and of the asymmetrical amputees 96.4 days.
The median L.O.S. of 39 days on the Amputee Service of a rehabilitation hospital compares favorably with national P.S.R.O. standards. Our data was influenced by a long-term (284 days) stay of a bilateral traumatic amputee. The mode of 28 days signifies a trend towards shorter L.O.S.
The achievement scale of Russek was selected to assess outcome. The results appear gratifying since 13 bilateral and 5 asymmetrical amputees are included. The majority (more than 90%) of the patients were discharged with a temporary prosthesis with a plaster-of-Paris socket (Fig. 3 & Fig. 4), applied almost immediately following admission. These devices are worn for an average of 6 8 weeks. It can be assumed that most patients will achieve a higher rating once supplied with the permanent device. Only about half of the bilateral below-knee amputees achieved a classification III rating. The rest remained Class IV. Of the asymmetrical amputees only two achieved Classification III.
The results, not encouraging, are attributable to the retirement age of most male patients and the presence of a variety of associated medical conditions which are listed in section II.
Level of Amputation
The classification recommended by the task force on standardization of prosthetic-orthotic terminology was used . Over two-thirds of the patients had either short or standard length below-knee (B/K) amputations as opposed to above knee(A/K).
This signifies a laudable trend for surgeons to carry out B/K amputations in preference to A/K amputations, which only 15 years ago was the preferred operative site if popliteal pulses were absent. Knee disarticulation, thought to be a suitable alternative to long A/K or a very short B/K amputation, was encountered only once.
Description of Amputation
The slightly higher incidence of right versus left amputation is probably statistically insignificant. There were 11 bilateral B/K amputees, 5 asymmetrical, but only 2 bilateral A/K amputees. The admission of this category of patients is generally discouraged as successful prosthetic application is usually not possible for older persons.
Interval of Amputation to Walking
It takes about two months after amputation before patients can walk again with a prosthesis. The mode of 48 days indicates a trend towards a much shorter interval.
Reason for Amputation
As expected, diabetic arteriosclerosis obliterans (ASO) is the most frequently encountered reason for amputation, followed by other vascular conditions.
Preceding Surgical Procedures
The data show a trend in surgery away from the once popular sympathectomies towards vascular surgical efforts such as by-pass procedures or embolectomies to restore failing circulation. Twenty-seven patients had previous amputations such as partial foot, conversions or amputations on the other side.
Complications at Acute Hospital
Reported complications which occurred at the referring hospital were frequent and ranged from mild (blood loss anemia) to pulmonary or myocardial infarctions. Only 41 patients had no complications. Wound infections occurred in 24 patients.
Only 36 patients had optimal stump conditions. Bulbous (47), not healed (54), edematous (70), or infected (24) stumps were noted. These significantly affected L.O.S. data.
This tabulation shows that practically all patients have one or more significant associated disorders, including 27 patients who had mild to moderate dementia.
Interval of Amputation to Admission
Mean and median values show that this interval is between 6-8 weeks with a trend towards a shorter interval (mode 21 days). Some of the reasons for delay in transfer to a rehabilitation hospital can be explained by the data presented in Sections 8, 9, 10, and 11.
The utilization of the described APG/APL application is not only time-saving and convenient for the physician, but has a direct, beneficial effect on health care delivery.
The patient's summary is available at the time of discharge and contains vital information concerning his medication schedule and appointment place and time for outpatient re-evaluation.
The interactive questionnaire type program will remind the physician and allied health personnel of possible omissions in record keeping or care.
The periodic exploration of a cumulative data base allows not only the detection of trends but the constant monitoring of the amputee service activities for the purposes of quality control.
This is a description of an APL/APG system oriented towards use by medical personnel essentially unskilled in computer sciences. An interactive questionnaire type input allows the generation of English-text summaries of patients discharged from The Burke Rehabilitation Center. Variables contained in the summary are stored to form a data base for concurrent statistical analysis.
Appendix A, cont., cont., cont.
- Stern, P.H. and Miller, J.M. Quantitative evaluation of long term L-DOPA treated patients with Parkinsonism using a computer terminal, 6° Congreso International de Medicina Fisica, 2-6 July 1972, Vol. II.
- Stern, P.H., Lincoln, J.C. and Robinson, M.B. Data base for stroke rehabilitation using computerized English text discharge summaries, Stroke, Vol. 6, March-April 1975.
- Russek, A.S. Management of lower extremity amputees. Archives of Physical Medicine and Rehabilitation, 42:687, 1961.
- Kay, Hector W. Assistant Executive Director, National Research Council. Committee on Prosthetics Research and Development, Washington, D.C.-personal communication.