Computerized Progress Notes for Chronic Pain Patients Receiving Opioids; the Prescription Opioid Documentation System (PODS)

Abstract

Objective

We herein provide a description of a health information technology tool using computer-assisted survey instruments as a methodology for documentation during long-term opioid therapy.

Design

We report our experience using the Prescription Opioid Documentation and Surveillance (PODS) System, a medical informatics tool that utilizes validated questionnaires to automate the assessment of opioid prescribing for chronic non-malignant pain.

Setting and Patients

Chronic pain patients answered questions that were presented on a computer terminal prior to each appointment in a Department of Veterans Affairs Pain Clinic.

Measures

Pain levels, activities of daily living, and screening for common psychological disorders were sought at each visit. Results were tabulated with some information gathered sequentially permitting evaluation of progress. Following a face-to-face interview, the clinician added additional comments to the medical record.

Results

By deploying a systematic series of questions that are recalled by the computer, PODS assures a comprehensive assessment.

Conclusions

The PODS fulfills medico-legal requirements for documentation and provides a systematic means of determining outcomes. This process facilitates the determination of the appropriate intervals between clinic visits by stratifying patients into high, moderate, and low risk.

INTRODUCTION

Documentation is an essential component of safe and effective opioid prescribing. Guidelines and polices for the use of opioids in treating pain uniformly stress the importance of complete and current documentation that is readily available for review. According to the Federation of State Medical Boards Guidelines, documentation of risk management is a key part of what medical boards are instructed to look for in judging clinical practice.¹ Yet the demands of documentation are challenging. Fear of punitive consequences for not meeting such requirements and the large effort required to conform to these guidelines may deter some physicians from prescribing opioids, resulting in continued and pervasive inadequacies of pain management.²

In a review of 300 medical records at a Veteran’s Administration primary care clinic, Clark found low rates of adequate documentation of pain complaints.³ Comments pertaining to a treatment plan or the efficacy of prescribed opioids in reducing pain, improving function, or otherwise benefiting the patient were found in only 39% of records. Only 41% of patients had received a physical exam within the preceding 6 months directed at the painful area, and only 17% of patients receiving opioids had received both comments on a treatment plan or follow-up and a pain-related physical exam within 6 months of the review. The presence or absence of side effects was documented in only 9% of charts reviewed.

Another medical record review in an academic family medicine center evaluated the adherence to state opioid prescribing laws regarding patient assessment for treatment of intractable pain with opioids.⁴ Opioid agreements were present in only 39% of records, and a pain evaluation and functional evaluation were documented in only 67% and 54% of records, respectively. These studies indicate the need to systematically collect information supporting the prescribing of opioids. Such a system should prompt physicians to collect surveillance data, make this data available at each office visit, document risk management, adherence as well as outcomes, and support the physician should they need to justify their practice to regulatory agencies or legal entities.

Faced with the need to prescribe opioids with substantial requirements for screening and ongoing documentation, physicians need systems that aid in the process of prescribing opioids. This includes methods for identifying patients at risk for opioid abuse and addiction and for complying with the requirements of state medical boards, law enforcement agencies, third-party payers, and evidence-based medicine. The Prescription Opioid Documentation System (PODS) is designed to help fill this need with pertinent information that is readily accessible. Designed using a Microsoft Access database, it provides a methodology for input from both patient and clinician that standardizes input while individualizing the medical information obtained.

PODS facilitates the systematic collection of data for monitoring the progression of patients taking opioids, and meets or exceeds medico-legal requirements for documentation devised by regulatory agencies and endorsed by professional societies.^5,6 PODS provides sequential data so that a time line of outcomes can be tracked for clinical decision making. This informatics tool utilizes validated instruments providing current and previously attained levels of analgesia, levels of activity, and depression. Health information technology tools such as PODS can translate clinical information into a format that promotes appropriate clinical exchange.⁷ The results are readily available during the clinic visit in a consistent format that promotes comprehension and integration. Clinicians gain insight into the patient’s status and thus configure the interview to provide optimal management. The results of the survey instruments and subsequent dialogue are entered into the medical record as a progress note. Physicians can explore the progress (or lack thereof) from the validated instruments compiled by PODS in order to plan therapies as well as intervals between return visits. Our previous manuscript described the use of PODS as a tool for the evaluation of initiation of opioid therapy; ⁸ this paper will its utility following deployment during follow-up visits in a Veterans Affairs Pain Clinic.

METHODS

Long-term treatment with opioids requires initial and ongoing assessments that are recommended to include several outcomes that should be documented in the medical record.^9–11 PODS utilizes validated instruments to accomplish these goals.

Description of Patient Participation at Each Visit

Patients were escorted by pain clinic staff into a separate room with several tables outfitted with desktop computers. Inserting their social security number, the computer identified them as a previous user of PODS and provided screen prompts that allowed them to either click a mouse or use a touch screen device (Model 370SD Light Pen, Interactive Computer Products, Irvine, CA) on a cathode ray tube monitor to input fixed choice answers to the questionnaires to be described below. Staff members left the room but would return if requested to do so by the patient or family. Friends and family members were allowed to help patients if they agreed not to improvise answers themselves. In general, this process was undertaken with enthusiasm often prompting patients to later ask questions that the computer-assisted survey instruments identified as an issue. Some patients expressed minor dissatisfaction often announcing that they were not computer literate. Quite commonly, this responded to verbal reassurances from staff that many other Veterans were in a similar predicament and completed PODS without struggling with the technology. In a very small minority of cases, patients were not cooperative and declined to participate in the computer-assisted interview. Patients were directed to the interview room after completing PODS on the computer. The clinician copied and pasted the Progress Note (Appendix) by locating the patients name among an alphabetized list in the PODS directory of completed reports. The clinician would then amalgamate the self-reports from PODS with the face-to-face interview and complete the progress note in the Computerized Patient Record System, the Veterans Affairs electronic medical record.

Serialized Instruments Provided at Each Visit

The Brief Pain Inventory,¹² is used to report both pain intensity (best, worst and average pain intensity in the previous 24 hours) and a global pain relief score produced by the current analgesic regiment. Figures 1A and 1B illustrate the computer screens presented to patients to garner this information. In addition, seven items concerning functional impairment are queried as illustrated in Figure 1C. Analgesic side effects including sedation, constipation, nausea, vomiting, cognitive impairment, loss of libido, and driving impairment are also examined using the computer screen depicted in Figure 1D. All of this information is placed into a Microsoft Access report that is copied and pasted into the electronic medical record (Appendix). To permit analysis of patient progress, sequential scores of percentage overall improvement and functional impairment are presented in descending order over time (shaded areas in Appendix).

FIGURE 1.

A–F. Sample PODS Electronic Forms

PODS also submits questions from the Current Opioid Misuse Measure (COMM), ^♣13 a 17-item self-report measure designed to identify aberrant drug related behavior for pain patients already on long-term opioid therapy. These items were composed to capture a 30-day time period (i.e., “in the past 30 days”), and only behaviors that could change from time to time were included (i.e., historical items were purposefully omitted). The intent was to capture signs and symptoms of drug misuse, (e.g., problems with thinking), emotional/psychiatric issues, evidence of lying, appointment patterns (e.g., doctor shopping), and medication misuse/noncompliance (e.g., borrowing pain medications from others, taking more than prescribed or self escalation of dose).¹⁴ A COMM score of 9 or greater indicates that the patient may be misusing opioids. A low cut-off score was purposefully selected to over-identify misuse, rather than to mislabel someone as responsible when they are not.¹³

To help clarify aberrant drug related behavior (and exclude false positives detected by the COMM), physician observations are recorded using a modified version of the Prescription Opiate Abuse Criteria.¹⁵ This instrument has been shown to have good reliability and was devised to be applied during normal clinic interactions permitting observed aberrant behaviors to be organized in a coherent and readily available fashion. The occurrence of the abuse criteria are determined by the clinician at each encounter as well as between visits (e.g., phone calls). As a reminder to input data, PODS presents a questionnaire (Figure 1E) after the clinician requests the automated progress note (Appendix). The physician chooses one of several types of aberrant behaviors (Table 1) and provides further comment in a text box (Figure 1E). The accumulated aberrant behaviors are presented in a list at every clinic visit (Appendix).

TABLE 1.

Prescription Opiate Abuse Criteria^*

1	Focusing on opiate issues during pain clinic visits which impedes progress with other issues
2	Pattern of early refills or escalating drug use in the absence of a medical need
3	Multiple telephone calls or visits to the office to request more opiates, get early refills, or has problems associated with the opioid prescription
4	A pattern of prescription problems for a variety of reasons that may include lost medications, spilled medications, or stolen medications
5	Supplemental sources of opiates obtained from other sources

^*Adopted from Chabal¹⁴

Inasmuch as a predictor of prescription opioid abuse may include a personal history of illicit drug and alcohol abuse,¹⁶ the evaluation of patients by PODS includes questions from the Addiction Severity Index (ASI). At the initial visit, PODS queries 16 specific items from this instrument (D 1–13, 17, 19, and 20) concerned with the quantity of abuse of alcohol and drugs consumed over two time periods; 1) the past 30 days and 2) lifetime experience. In patients in whom the ASI disclosed positive lifetime exposure to substances of abuse, dialogue boxes are presented at every subsequent visit to inquire about activities during the previous 30 days. This tracking system automates follow-up of alcohol and/or drug intake as well as queries patients about attendance at support groups (i.e., AA or NA) if lifetime exposure to an abusable substance was initially endorsed.

Comorbid major depressive disorder is provided special attention as the medical costs of combined treatment of somatic and mental suffering are intertwined.¹⁷ PODS evaluates the degree of depression each visit utilizing the 9-item Patient Health Questionnaire (PHQ-9) depression module.¹⁸ The PHQ-9 is a self-administered questionnaire derived from the Primary Care Evaluation of Mental Disorders (PRIME-MD) set of diagnostic instruments for common mental disorders. The PHQ-9 scores each of the 9 DSM-IV criteria as "0" (not at all) to "3" (nearly every day). A PHQ-9 score equal to or greater than 10 has a sensitivity of 88% and a specificity of 88% for major depression.¹⁸ In addition to making criteria-based diagnoses of depressive disorders, the PHQ-9 is also a reliable and valid measure of depression severity. PHQ-9 scores of 5, 10, 15, and 20 represented mild, moderate, moderately severe, and severe depression, respectively.¹⁸ These characteristics enable serialized assessments in the DSM-IV criteria of depression over time (shaded areas in Appendix).

Instruments Provided at a Single Visit

To provide additional mental health screening, self-report measures of anxiety disorders are administered. To insure that the patient does not develop frustration or experience boredom from excessive questioning at a single visit, these anxiety screens are presented as separate units as the patient returns over several months for opioid prescription refills. The initial instrument is the Post-traumatic Stress Disorder Checklist.^{19, 20} This instrument is an 18-item tool used to screen patients for post-traumatic stress disorder (PTSD). The patient is asked to indicate the level of distress caused by symptoms on a 5-point Likert-scale (1=not at all to 5=extremely). A score equal to or greater than 44 is indicative of Post-traumatic Stress Disorder. The Panic Disorder Self-Report,²¹ a screening tool modeled after the Anxiety Disorders Interview Schedule for DSM-IV (ADIS-IV), is administered next. Patients answer “yes” or “no” questions regarding panic attack symptomotology such as excessive worry, change in behaviors, interference in daily life, and emotional distress with verification that they experience the peak of the panic attack within 10 minutes of onset. Patients with a score equal to or greater than 8.75 on this screening tool meet the criteria for Panic Disorder; the test has a specificity of 100% and a sensitivity of 89%. On the next visit, the Generalized Anxiety Disorder Questionnaire-IV²² is presented by PODS. This diagnostic instrument is based on symptoms from the forth edition of the Diagnostic and Statistical Manual (DSM-IV). This screening tool has a specificity of 89% and a sensitivity of 83% in identifying patients with generalized anxiety disorder (GAD). The self-report includes questions evaluating the amount of excessive and uncontrollable worry experienced by a patient, the topics they worry about most frequently, whether they have experienced these symptoms for at least 6 months. Patients with a score equal to or greater than 5.7 meet the criteria for generalized anxiety disorder. The serial responses to the PHQ-9 and the responses to specific anxiety disorder screens are provided at each visit. In this manner, clinicians may be alerted to the presence of common psychopathologies which may be interfering with medical therapy for chronic pain that might indicate the need for referral to a mental health practitioner.

The rationale for being alerted to the presence of a mental health or substance problem is becoming more evident as epidemiologists study long term opioid therapy. Chronic use of prescription opioids for non-malignant chronic pain is much higher and growing faster in patients with mental health and substance use disorders than in those devoid of these diagnoses.²³ Clinicians should monitor the use of prescription opioids in these susceptible patients to determine whether opioids are replacing or interfering with appropriate mental and/or substance abuse treatment. Close association with a mental health or addiction provider may be warranted; the use of the above screening instruments may serve the purpose of notifying an otherwise unsuspecting clinician.

Urine Toxicology Screening

PODS randomly requests that the patient obtain a Urine Toxicology Screen deploying a 30% probability that this will be asked at each clinic visit. A dialogue box appears on the computer (Figure 1F) alerting the patient that the clinician will be ordering this test and informing the patient that the selection was random (i.e., not related to any answers they provided). The clinician is then made aware of the selection when he/she is attempting to obtain the PODS report that will be pasted into the electronic medical record. This urine toxicology screen may be forgone if the clinician determines that the patient has already undergone this test recently or if clinical judgment deems the testing unnecessary.

Prescription Database

A Prescription Database is an integral part of PODS, offering a serialized view of previously prescribed opioids, as well as instructions, dosages and amounts (Figure 2). At each encounter, the provider may consider adjusting the dose or switching opioids using the results of analgesia levels attained, side effects, and other variables provided through the PODS report. The prescription is printed on a laser printer and the provider signs the paper copy. The information in the database cannot be altered (after the date that the prescription is written) so that a medicolegal record is maintained of controlled substances provided.

FIGURE 2.

Dosages and Amounts of Opioids in Prescription Database

RESULTS

PODS has been used to provide objective yardsticks for clinical decision making during opioid prescribing. Global pain relief, a measure commonly used in research and clinical practice, was used as the primary outcome.^24,25 A target of 50% or greater overall relief was sought by increasing the opioid dose gradually and/or adding co-analgesics at sequential visits while simultaneously scrutinizing side effects (Appendix). PODS presented a serial view of overall improvement at each visit, sorted by date in descending order (shaded areas in Appendix). In this manner, a readily interpretable decision point for dose titration was utilized. Of course, clinical judgment was always used and overrode the results of screening instruments in the case of a conflict.

Despite its inclusion in the Brief Pain Inventory and wide spread utility, the global pain relief metric occasionally was a source of confusion to the patient. This was recognized by the clinician when the answer was incongruent with the patient’s previous responses from earlier visits. Because of its significance as the primary outcome measure, extra care was devoted to insuring a valid response. The clinician and patient would jointly view this Brief Pain Inventory item (Figure 1B) on the computer screen and arrive at an accurate response. If warranted, the clinician would comment upon this discussion in the medical record.

If overall improvement was not improving within a reasonable time frame (e.g., by the third or fourth visit), or if irremediable side effects intervened, opioid rotation was performed. If necessary, a short-acting opioid was added for breakthrough pain to help achieve the targeted outcome. Upward dosing was curtailed after either 50% or greater overall improvement was noted or a maximum dose of 180 mg of morphine equivalents per day was reached. The later restriction stems from concerns about toxicity, especially when opioids are utilized in high doses for prolonged periods, related to hyperalgesia, hormonal and/or immune function.²⁶

At the same time, evaluation of the serialized PHQ-9 scores was evaluated to ascertain if the reason for lack of improvement could possibly be explained by the presence of mood disorder which might have benefited from medication or biopsychosocial interventions.^27–32 In addition, the Post-traumatic Stress Disorder Checklist, Panic Disorder Self-Report and Generalized Anxiety Disorder Questionnaire-IV were also evaluated for input because of the strong association between anxiety and physical disorders.^{33, 34}

PODS also investigated the extent to which pain interfered with function by calculating an average value from the Brief Pain Inventory Pain Interference Subscale. This was assimilated into the PODS report that was cut and pasted into the progress note in sequential descending order by date in the progress note (shaded areas in Appendix). As support for improvement in function and quality of life is mixed in opioid clinical trials, and is less convincing than evidence supporting subjective pain reduction, only the later was used as a target for dose escalation or opioid rotation. However, patients with persistent disability were referred for multidisciplinary treatment with physical therapy and individual or group psychotherapy.

As it is clear from current evidence that many patients abandon chronic opioid therapy because of the unacceptability of side effects,²⁶ the presence of adverse effects was queried at every visit (Appendix). Symptoms such as constipation, nausea, itching, fatigue, drowsiness, and loss of libido are scored as “none,” “mild,” “moderate,” “severe.” “Mild” and “moderate” scores triggered the physician to prescribe specific remedies (e.g., a laxative protocol, psychostimulants for sedation, benztropine for diaphoresis, etc.). Higher scores drew more attention, so that follow-up visits were scheduled at more frequent intervals and opioid rotation was performed if the side effect(s) did not improve.

Demographics of the pain clinic population was typical of a VA population (Table 2). The average (SE) number of follow-up visits was 6.85 (0.17) with a positively (left) skewed distribution (Figure 3). The time to complete these assessments varied according to the additional anxiety screens presented at each visit (Table 3). In general, the patient spent 20 to 25 minutes completing the follow-up evaluation plus one of the anxiety screening instruments.

TABLE 2.

Demographics and characteristics of patients (N =1861)

Gender
	Male	1683 (90%)
	Female	178 (10%)
Age
	Mean (SE)	56.4 years (1.6)
Education
	less then 8 years of grade school	20 (1%)
	8 to 11 years of school	138 (7%)
	graduated from high school	312 (17%)
	some college	759 (41%)
	trade school	215 (12%)
	graduated from college	299 (16%)
	postgraduate studies	118 (6%)
Ethnicity
	White	1451 (78%)
	Black	239 (13%)
	Hispanic	57 (3%)
	American Indian	31 (2%)
	Asian American	23 (1%)
	Other	30 (2%)
Marital Status
	Married	871 (47%)
	Divorced	511 (27%)
	Separated	95 (5%)
	Widowed	105 (6%)
	Living with Significant Other	127 (7%)
	Never Married	152 (8%)
Employment Status
	Working part time	150 (8%)
	Working full time	209 (11%)
	Disabled	1108 (60%)
	Retired	394 (21%)
Average VAS Pain Intensity During Past Week (SE)		6.8 (0.04)
Duration of pain (SE)		14.6 years (0.28)

FIGURE 3.

Frequency Distribution of Number of Visits

TABLE 3.

Timing of Evaluations from Computerized Time Log

	Mean (minutes)	SE
Follow-up Evaluation	11.0	0.35
Posttraumatic Stress Disorder	12.16	0.78
Panic Disorder	12.32	1.44
Generalized Anxiety Disorder	9.43	0.18

DISCUSSION

William Osler opined, “It is much more important to know what sort of a patient has a disease than what sort of a disease a patient has.” Anachronistically, it is conceivable that today’s equivalent of “what sort of a patient has a disease” might be the psychological distinctiveness of the patient. Assuming that this is what he had in mind, Osler’s penetrating insight would be greatly enhanced by computer-assisted survey instruments. While self-report may not be a substitute for a clinical interview, its utilization supplements a physician’s diagnostic abilities in a time saving manner. But beyond the capability of validated instruments to provide insight into psychological conditions and a computer to calculate summary scores of screening instruments, PODS real strength is the collection of data through structured interviews at consistent intervals that enables a sequential, temporal analysis and documentation of vital clinical outcomes. The automation of a bio-behavioral chronicle could not have been foreseen by Osler who composed the above aphorism almost 100 years ago at the beginning of the last century.

Given the problems inherent in predicting who will develop misuse of opioids, the concept of universal precautions has been increasingly tied to pharmacovigilance with controlled substances.³⁵ Taken from the infectious disease literature, the application of this principle mandates that treatment be applied in a uniform and structured manner using high levels of scrutiny and a similar assessment for every patient. As the monitoring of pain severity, functional limitations, adverse events, and progress is performed in an efficient and standardized manner at each visit, PODS allows a practitioner to apply this standard using time sensitive data that is easily interpreted. In general, the risk factors for prescription opioid abuse or misuse include male gender, young age, positive family history for abuse, or psychiatric history or extensive past substance abuse.^{16, 36, 37} In addition to evaluating these factors prior to implementing opioid therapy, monitoring aberrant behaviors after initiation of therapy is warranted. PODS provides a means of quantifying both self-reported aberrancies (COMM) and witnessed events in a systematized manner. In addition, this computerized tool provides a non-accusatory method of selecting patients for random urine toxicology screening; the latter being necessary to detect unexpected substance abuse.³⁸

Using the above factors, PODS facilitates the determination of the appropriate interval between clinic visits by stratifying patients into high, moderate, and low risk. For example, a low risk patient may be one who has been seen for at least 6 months, is on a stable opioid regime, has a low COMM score, exhibits no aberrant behaviors, has a negative urine toxicology screening and has no evidence of receiving opioids from multiple providers. The latter could be evaluated by querying a state prescription monitoring program, if feasible. This type of patient is neither likely to have co-morbid psychiatric disease of a substantial nature nor a history of recent substance abuse. They can probably be seen every 3 months as deemed acceptable by the DEA³⁹ or even every 6 months as advocated by recent guidelines.⁴⁰ Moderate risk patients would have followed their opioid agreement but have had a lapse perhaps demonstrating one or two aberrant behaviors before curtailing this type of activity. They might also have problems reaching the 50% overall improvement milestone. These patients might be seen at closer intervals such as monthly. At the other end of the spectrum are the high risk patients. They will have recent histories of substance abuse, exhibited multiple aberrant behaviors, had positive urine toxicology screens with illicit substances, or found to have utilized multiple providers. The risk-to-benefit analysis will be far less in favor of treating with opioids and these patients may warrant weekly or biweekly visits. If unabated, referral for substance abuse therapy and reliance on non-opioid treatments might be necessary.⁴⁰

The collection of data with the use of computer-assisted, self-administered interviews has the potential to reduce interview bias. In addition, there is the potential for eliciting reporting of "stigmatized behaviors" more frequently than during face-to-face interviews.^{42, 43} But the use of computer assisted survey instruments may have important effects on instrument reliability⁴⁴ Future evaluations of PODS should compare the results of the computer-assisted survey instruments comparing PODS against paper and pencil versions of the same validated instruments to see if the automated format is as reliable as the older method. If it can be shown that instrument reliability is not hampered, more widespread implementation should be encouraged.