Tag Archives: patient safety

Baby Breastfed by Wrong Mom

By ThinkReliability Staff

After a newborn baby at a Minneapolis hospital was placed in the wrong bassinette, he was delivered to the wrong mother and breastfed.  Because breastfeeding can carry risks of transmission of communicable diseases the CDC recommends HIV and hepatitis testing after such events.

We can examine this incident – and what went wrong – in a visual root cause analysis, or Cause Map.  The Cause Mapping procedure begins by determining the impact to the organization’s goals.  In this case, the patient safety goal is impacted due to the risk of transmissible disease.  The hospital involved has stated there will be consequences to staff for not following hospital procedure.  This is an impact to the hospital’s employee impact goal.  The patient services goal is impacted because babies were switched (and apparently misplaced for some period of time) and because of the testing that the baby who was breastfed by the wrong mother will require.  The hospital will pay for the testing, which can be considered an organizational goal impact.

The analysis step of the Cause Mapping process begins with the impacted goals.  To continue the analysis, we ask “why” questions.  The patient safety goal is impacted because of the risk of disease.  The risk of disease is caused by being breastfed by the wrong mother.  This occurred because the wrong baby was brought to the mother,  the mother was breastfeeding, and the infant’s bands were not matched to the mother’s bands, although this was hospital procedure.  According to the hospital’s statement, “While hospital procedures require staff to match codes on the infant’s and mother’s identification bands in   order to prevent incidents like this, it appears these procedures were not followed in this case.”

The wrong baby was brought to the mother because multiple babies were kept in bassinettes in the nursery, and the baby had been placed in the wrong bassinette.  It is unclear what procedure was used to determine which bassinette the baby should be placed in, but the procedure was obviously ineffective.

The hospital has stated that its procedures will be reviewed.  Certainly the procedure to verify a baby’s wristband to a mother’s will be emphasized and retrained.  Additionally, matching of the baby’s wristband with a tag on the bassinette would reduce these types of issues.  Some hospitals have gone so far as to stop using nurseries where multiple babies are placed and instead keep the newborn in the mother’s room.  This also would reduce the risk of baby switching incidents.

To view the Outline, Cause Map and potential solutions, please click “Download PDF” above.

A Tongue Tie Release Wrongly Performed in Case of Tongue Lesion Resection

By ThinkReliability Staff

A California hospital has been fined $50,000 – its fifth administrative penalty from the State since 2009 – for performing the wrong procedure on a 6-year-old boy.  The boy was supposed to receive a tongue lesion resection, but instead a tongue tie release was performed.

We can examine the issues that resulted in this incident within a Cause Map, or visual root cause analysis.  The first step in any analysis is to define what you are analyzing.  We begin with impacts to the organization’s goals.  In this case, we look at the impacted goals from the respect of the hospital.  First, the patient safety goal  was impacted due to an increased risk of bleeding, infection, and complications from anesthesia.  The compliance goal is impacted because performing the wrong surgical procedure on a patient is a “Never Event” (events that should never happen).  The organizational goal is impacted because of the $50,000 fine levied by the State of California.  The patient services goal is impacted because the wrong procedure was performed and the labor goal was impacted due to the additional procedure that was required to be performed.

The second step of our analysis is to develop the cause-and-effect relationships that describe how the incident occurred.  We can develop these relationships by beginning with the Impacted Goals and asking “why” questions.  For example, the patient safety goal was impacted because of the additional risk to the patient. The patient received additional risk because of the performance of an additional procedure.  An additional procedure was necessary because the wrong procedure was initially performed.

There are many causes that contributed to the wrong surgery being performed.  These causes are outlined in the  report provided by the California Department of Public Health.  In this case, there were several causes that likely resulted in the wrong procedure.  The Operative Report had the incorrect diagnosis – tongue tie – which would suggest that a release would be the appropriate procedure.  Additionally, the Anesthesia Record contained the wrong procedure (tongue tie release), possibly because the Pre-Anesthesia Evaluation originally noted that a tongue tie release was to be performed and was later corrected (by crossing out the incorrect procedure and writing in the actual procedure).

The type and site of surgery was not verified.  The surgeon who performed the surgery could not remember if a time-out had been performed, although there was a record of a time-out performed immediately prior to the surgical procedure.  Since the time-out was performed immediately prior to the procedure and the surgeon was unable to remember the proper procedure, the time-out was obviously ineffective.

The surgeon stated after the surgery that he believed that the tongue tie release surgery which was performed was indicated based on scar tissue that was found under the tongue.  The surgeon did not notice the lesion on the tongue during the surgery and no pre-surgical exam was performed by the surgeon.  Additionally, the surgical site was not marked (as the site of the correct, as well as the incorrect, surgeries were both within the patient’s mouth).

During the procedure, none of the other staff stopped the surgery as it was occurring.  However, given the proximity of the “correct” site to the “incorrect” site, it may have been difficult for the other staff to see what was going on.  The surgeon did note that the lesion removal should have created a sample, the lack of which was not noted by staff.

The surgeon involved in this case has indicated that he will be examining his patients prior to surgery in the future. Hopefully this incident will also serve as a reminder to all medical staff that in the case of a site that cannot be marked as per procedure, extra care should be taken to ensure the correct site is operated on and the correct procedure is performed.

To view the Outline and Cause Map, please click “Download PDF” above.  Or click here to read more

NYC Hospital Unexpectedly Evacuated During Sandy

By Kim Smiley

On October 30, 2012, power outages forced evacuation of a New York City hospital amidst the onslaught of Hurricane Sandy.   All 217 patients in the hospital to had moved, including 20 infants staying in the neonatal intensive care unit.

This incident can be analyzed by building a Cause Map, an intuitive format for performing a root cause analysis.  The first step in the process is to fill in an Outline that lays out the basic background information and also identifies the impact to the goals.  In this example, the safety goal is clearly impacted because it is risky to evacuate patients during a hurricane.  Although the potential for injury was very real, no one was hurt during the evacuation and the hospital staff did an amazing job of carrying patients down darkened stairwells and ensuring basic life support remained stable.  The customer service goal is also worth considering since the unexpected evacuation received a large amount of negative publicity.

After the Outline is completed, the next step is to ask “why” questions to add Causes to the Cause Map.  Why were patients at risk?  This occurred because the hospital had to be evacuated because it lost power and the backup power generators failed.  Why the generators failed hasn’t been identified yet, but there is speculation that the age of the equipment may have played a role. It’s also possible that the location of the generators might be factor since a number of hospital building were flooded by ten feet of water.  Electrical service was lost because New York City was hit hard by Hurricane Sandy and saw unprecedented flooding and strong winds.  This wasn’t an unexpected impact of the storm, but the hospital did not expect the generators to fail, especially so quickly.  The patients were also at risk because the hospital had many patients in critical care units that required life support systems and the patients were evacuated under dangerous conditions, both inside and outside the hospital.  At the time of the evacuation the hospital had lost power and patients were being carried down stairs lit by flashlights.  Some patients were bought down 16 flights of stairs.  The evacuation also occurred during the hurricane so the conditions during the drive to a new faculty were potentially dangerous.  The evacuation occurred during the hurricane, as opposed to before the storm hit, because the hospital assured the city that it was prepared and could ride out the hurricane.

This issue is still being investigated, but once all the facts are known solutions can be developed and implemented to help ensure that patients aren’t forced to evacuate under similar adverse conditions.

Click on “Download PDF” above to see a high level Cause Map of this issue.

Manifestation of Poor Glycemic Control Part 2

By ThinkReliability Staff

In a previous blog, we discussed how poor glycemic control can result in hyperglycemia which could lead to nonketotic hyperosmolar coma.  Diabetic ketoacidosis, if resulting from poor glycemic control within a hospital setting, is another hospital-acquired condition as determined by Medicare & Medicaid, meaning that hospitals will not receive additional payment for cases when this condition is acquired during hospitalization.  Like nonketotic hyperosmolar coma, diabetic ketoacidosis can have a significant impact on patient safety and can be investigated within a Cause Map, or a visual root cause analysis.

The impacted goals for a hospital resulting from hospital-acquired diabetic ketoacidosis are very similar to those for nonketotic hyperosmolar coma.  Patient safety is impacted due to an increased risk of death, which can also result in a provider being a “second victim.  This is a “no-pay” hospital acquired condition, which is estimated to cost $42,974 per case.  According to the Centers for Medicare & Medicaid Services (CMS), in 2007 there were 11,469 cases of hospital-acquired diabetic ketoacidosis, resulting in a total cost to the healthcare system of almost half a billion dollars.

According to a study  published in the International Journal for Quality in Health Care, diabetic emergencies, including nonketotic hyperosmolar coma,  increases the risk of patient death (from 9% to 16%),  length of patient stay (from 7 to 14 days) and treatment requirements.  The costs associated with nonketotic hyperosmolar coma (greater than $114 million in the US in 2007, according to CMS) are no longer reimbursable when the condition is acquired in the hospital.  Additionally, patient death due to hospital-acquired conditions can result in a second   victim – the healthcare provider(s).  Additionally, this diagnosis results in increased stay and treatment requirements.

Beginning with the impacted goals and asking “Why” questions, we quickly determine that diabetic ketoacidosis, like nonketotic hyperosmolar coma, results from uncontrolled hyperglycemia.  Rather than perform the same analysis of causes of hyperglycemia (which, if we’re doing our job right, should result in the same cause-and-effect relationships), we can link to the analysis shown in our previous blog.   However, for diabetic ketoacidosis, we also have a cause of dehydration.  Since this was not a cause previously analyzed, we will add to this portion of the Cause Map.

Patient dehydration can result from a medication that increases fluid loss, an underlying medical condition, or inadequate water intake.  Inadequate water intake can result from a patient’s limited access to water, such as a patient who is bedridden and is not provided adequate water from a caregiver, or the patient feels too ill to drink, or the patient is unable to drink, due to incapacitation, confusion, restraints or sedation.  A combination of these causes may also occur.

Because of the importance of preventing these conditions resulting from hyperglycemia and dehydration, every effort should be made to prevent these outcomes from occurring.

Two other conditions are considered hospital-acquired manifestations of poor glycemic control, diabetic ketoacidosis and hypoglycemic coma.  In future blogs, we will discuss the causes of these issues, and suggested solutions to reduce the risk of these types of incidents.  It is recommended that an individualized insulin plan be used, rather than a sliding scale, to ensure blood glucose levels are kept at or below 110 mg/dL.  A specific glycemic management team, which carefully coordinates medical nutritional therapy with insulin control, can also reduce the risk of glycemic events.  Patients who are found to have an insulin deficiency should be treated with intravenous insulin.

Because 20-30% of diabetic ketoacidosis cases are estimated to be the initial presentation of previously undiagnosed diabetes, some experts recommend testing the glucose levels of all children who have not been diagnosed with diabetes, and all patients who are vomiting or require intravenous hydration.  To reduce the risk of dehydration, patient’s fluid intake should be tracked and any patients who are unable to  drink should have intravenous fluids.

Nonketotic hyperosmolar coma and diabetic ketoacidosis are two hospital-acquired events that result from hyperglycemia.  The remaining hospital-acquired manifestation of poor glycemic control, hypoglycemic coma, will be covered in a future blog.

To view the Outline, Cause Map, and Solutions please click “Download PDF” above.  Or click here to read our previous blog.

More Known About Why A Donated Kidney Was Trashed

By ThinkReliability Staff

In a previous blog, we wrote about a donated kidney that was accidentally thrown out rather than being transplanted.  We began the root cause analysis investigation with the information that was available, but there were still a lot of open questions.

The Centers for Medicare & Medicaid Services (CMS) has released a report on the incident, which provides additional information we can use to update our Cause Map.  We can update all areas of the investigation, including updating any additional goals that were found to be impacted.  In this case, three employees had been placed on administrative leave.  Since the time of the previous blog, four employees have had their careers impacted – one has resigned, one has been fired, one has had a title removed, and another has since returned from paid administrative leave.  Additionally, there is a risk that the hospital may be removed from the Medicare program, another impact to the compliance goal.

The report provides more specific causes, and evidence, regarding the incident.  We know now that the kidney, which was to be transplanted, was instead thrown in a hopper by the circulating nurse.  We can ask “Why” questions to add more detail.  The kidney was thrown in the hopper because the contents of the slush machine were thrown in the hopper and the kidney was in the slush machine.  It still isn’t clear why the kidney was in the slush machine in the donor’s operating room (rather than being transferred immediately to the recipient’s room), but more information regarding the disposal is now available.

The nurse disposed of the hopper because she was unaware that  the slush machine contained the kidney.  The nurse had been on lunch break when the location of the kidney was announced and was not briefed on the status of the operation upon her return.  There was no documentation on where the kidney was located, and the nurse assumed that it was in the recipient’s room.  For reasons that are unclear (as it is usually the job of the technician who is responsible for the machine), the nurse decided to empty the slush machine while the operation was still ongoing.  This appeared to be against procedure, but the procedure had “exceptions” according to staff, and was ineffective in this case.  The technician that was responsible for the slush machine was exerting inadequate control, as the staff members have stated that no one noticed the nurse empting the slush machine.  This also demonstrates inadequate control of the kidney, since there appeared to be no staff person responsible for the kidney itself.

Since the incident, the hospital has developed a procedure for intra-operative hand-off, which includes a briefing requirement for staff members who enter an operating room mid-procedure.  Additionally, clarification has been provided that nothing will leave an operating room until the patient has left, post-procedure.  Although the transplant program is still shutdown pending investigation, a recommendation that might reduce this type of problem in the future would be to ensure that a staff member is designated as responsible for any donated organs from removal to transplant.

To view the updated Cause Map and potential solutions, please click “Download PDF” above

Teen Impersonates a Physician’s Assistant

By Kim Smiley

A teen, who was 17 at the time, was arrested on September 2, 2012 for impersonating a physician’s assistant in a Florida hospital.  The young man worked at a hospital, treating patients and performing duties typical of a physician’s assistant, for about a week before anyone became suspicious of his lack of credentials.  Investigation into the case found that he examined patients, removed an IV and even performed CPR without any medical training.

How could this possibly happen?  A Cause Map, or visual root cause analysis, of this situation can be built to help understand the different causes that contributed to a young man successful impersonating a medical professional.  The first step in building a Cause Map is to determine how the issue impacted the overall organization goals.  In this example, the safety goal is clearly impacted since an unlicensed individual treated patients.  The customer service goal was also impacted because of the negative publicity for the hospital involved in the scandal.

Causes are added to the Cause Map by asking “why” questions.  Why did this happen?  How did a teen end up performing the duties of a physician’s assistant?  Statements by the teen indicated that he was interested in learning more about the profession so he decided to work at the hospital.  He was able to pull this off because he was incorrectly given a physician’s assistant identification badge and nobody initially questioned his credentials because he acted the part well.

The teen worked as a clerk in a doctor’s office near the hospital and when he went to the ID office to get a badge, he was somehow given the wrong one.  His credentials were never checked and personnel at the ID office have stated that this was because the office was very busy at the time.  The teen also never told anybody he had the wrong badge and decided to use it.

The masquerade was also successful for a time because the teen played the role of physician’s assistant well.  He wore scrubs and a stethoscope and used the correct terminology.

This case went to trial in August 2012.  The teen was found guilty on two counts of impersonating a physician assistant and two counts of practicing medicine without a license.  His sentencing is scheduled for November 14 and he faces up to 25 years in prisons.

To view a high level Cause Map, click “Download PDF” above.

Safe Use of Opioids in Inpatient Hospitals

By ThinkReliability Staff

The use of opioids for pain relief in inpatient hospitals can lead to serious potential adverse effects, including respiratory depression and drug interaction.  On August 8, 2012, The Joint Commission published a Sentinel Event Alert: “Safe use of opioids in hospitals”.  The alert contains information about potential causes of the adverse effects possible with the use of opioids as well as solutions that, if implemented by healthcare facilities, can reduce the risk of patient safety impacts from the use of opioids.

We can present the information provided by The Joint Commission in a Cause Map, or visual root cause analysis.  We begin with the impacts to the goals.  In this case, we look specifically at two potential impacts to the patient safety goal – the risk of drug-drug interactions and respiratory depression involving opioids.

Drug-drug interactions can result when a patient is taking another drug that interacts with opioids. In this case, the provider prescribing the opioid is unaware of the potential interaction between the drugs prescribed or is unaware of the patient’s drug history, because a complete history is unavailable and a patient is either unable or unwilling to provide a compete list. While drug-drug interactions are possible with any level of opioid, the over-use of opioids for pain relief is a particular concern.  Opioids can be effectively used for pain relief, but over-use can occur when a high dose is needed to manage pain, either due to tolerance from chronic conditions or patient abuse, or obesity.  Studies have shown that obese patients may require more opioids for pain relief than would be suggested by their weight alone.  A patient receiving the wrong dose of opioids (besides being an issue in itself) can also contribute.  Issues have been raised regarding the difficulty in calculating doses with drugs of different potency, especially as patients move from one drug to another.  Additionally, prescribing dose based on weight alone can result in a higher or lower dose than needed as the proper dose of opioids is subject to patient weight, age, sex, and tolerance level.

Issues with prescribing the wrong dose or wrong type of medication can occur when a patient or family member is responsible for the administration.  Problems with medication administered by a provider typically occur around changes of the type or delivery method of the pain killer.  Special care should be taken to recalculate the dose  corresponding to any change in the drug dosage, type or delivery method.  Similar-looking bottles and similar-sounding names are always a potential pitfall in proper drug administration and special care should always be taken in these cases.

Opioids reduce respiratory rate, which can result in respiratory depression.  Respiratory depression can be impacted by other factors, such as a patient who is sleeping (most respiratory depression occurs during typical sleeping hours), or who is already pre-disposed to respiratory depression.  This most commonly occurs with post-surgical patients (who may have residual anesthesia), old or young patients (who may be affected more greatly by the respiratory effects), patients who have abnormal respiratory control due to obstructive sleep apnea or morbid obesity, patients with supplemental oxygen and patients who have a self-administered drug delivery system, such as a fentanyl patch.  Special care and monitoring should be taken with patients who have a higher risk level for respiratory depression.

However, monitoring for respiratory depression is difficult.  Visually assessing respiratory depression (especially while a patient is sleeping or on supplemental oxygen) is extremely difficult.  Using pulse oximetry can result in misleading values (including normal values while a patient is suffering from respiratory depression) and high false alarms.  Because respiratory depression occurs gradually, intermittent monitoring may not be sufficient to pick up on a patient’s decline.

There is no one-size-fits-all solution for reducing respiratory depression.  Rather, an individualized plan based on patient pain requirements and risk factors is shown to be the recommended way to reduce the risk of respiratory depression and ensure proper pain control for patients.

To view the Cause Map and recommended solutions, please click “Download PDF” above.  Or learn more from The Joint Commission Sentinel Event Alert.

Donated Kidney Trashed

By ThinkReliability Staff

On August 10, 2012, a living donor’s kidney was thrown out, instead of being transplanted as planned.  The incident was chalked up to “human error”, which is almost certainly part of the problem . . . but definitely not all of it.

This extremely rare, but serious, event is being analyzed by several oversight agencies, as well as a contractor hired by the medical center in Ohio where the event took place, to ensure that needed improvements are identified and put into place so this type of incident doesn’t happen again.  We can examine the currently known information in a visual root cause analysis, or Cause Map.  To do so, we begin with the impacted goals.

There are many goals that were impacted as a result of this error.  Firstly, the patient safety goal was impacted because the patient did not receive the transplanted kidney.  This can also be considered an impact to the patient services goal.  Three personnel from the hospital were placed on administrative leave as a result of the incident.  This results in an impact to employees.  The compliance goal is impacted because this event has resulted in a review by several oversight agencies.  The living kidney donor program is currently shut down for review, which can be considered an impact to the organization goal.  The kidney was disposed of improperly, which is an impact to the environmental goal.  (Medical waste has strict requirements for disposal.)   The loss of the donated kidney can be considered an impact to the property goal.  Personnel time was taken both to attempt to resuscitate the kidney and to participate in an independent review of the donor program.  These can both be considered impacts to the labor/time goal.

Once we have determined the impacts to the goals, we can ask “Why” questions to develop the cause-and-effect relationships that led to these impacts.  In this case, the patient did not receive a kidney transplant because the kidney was thrown out and because of concern about the kidney’s viability.  Part of this concern was the delay in actually finding the kidney, likely due to the fact that it was disposed of improperly.  The reason given by the medical center for the disposal of the kidney is “human error”.  However, there is ordinarily a support system involved in organ transplants that would minimize these types of errors.  Certainly the fact that the program has been stopped and three employees – at least one of whom was not directly involved in the transplant operation – were placed on administrative leave suggest that the organization is looking at more than just a screw-up by one person acting alone.

Specifically, the investigation should look at communication – was the nurse who disposed of the organ told it was destined for transplant?  Was there a surgical time-out immediately prior to the removal with the entire operating team that discussed the plan for the kidney?  Also the training and preparation of the surgical team should be investigated.  Had the team been properly trained and prepped for this type of surgery?  The fact that it was done frequently at this facility doesn’t mean that adequate training was in place.  What about the procedure for treatment and supervision of donated organs?  Donated organs have to be treated in a very particular way to ensure their viability for the transplant patient.  Who, if anyone, was responsible for ensuring that the organ was prepared in a proper way for transplant?  Were they involved in the surgical time-out?  Lastly, because an error was made with the disposal procedure, the procedure, training and communication regarding disposal of medical waste needs to be analyzed to ensure it is adequate. The hope is that by doing a thorough review – and improvement – of policies, procedures, training and communication at the facility, it will not only reduce the risk of this type of error, but provide improvement in many other aspects of the care provided as well.

To view the Outline and Cause Map, please click “Download PDF” above.

At Least 31 Patients Contracted Hepatitis C

by Kim Smiley

Testing is still ongoing, but at least 31 people have contracted hepatitis C from contaminated syringes at a New Hampshire cardiac catheterization lab.  A previous blog discussed the outbreak when it was initially announced that four patients who had used the same cardiac catheterization lab had tested positive for the same strain of hepatitis C, but more information has been released and the Cause Map should be updated to incorporate all the relevant details.  One of the strengths of a Cause Map, a visual root cause analysis, is that it can be updated relatively quickly to document important information as it becomes available.  In this example, investigators are continuing to work to understand the issues involved, but two new significant pieces of information should be added to the Cause Map.

The source of the hepatitis C has been determined by investigators.  Investigators found that a medical technician with hepatitis C contaminated syringes that were then used on patients.  The medical technician is a drug addict who used the syringes because they were filled with Fentanyl, an anesthetic far more powerful than morphine.  Hepatitis C is spread through blood to blood contact so syringes contained with hepatitis C are a major health hazard that are capable of spreading the disease. The syringes were not secured so he was able to attain them.  He then used them, refilled them with saline or another liquid and replaced them without any other member of the staff noticing.

Investigators have also learned that the medical technician responsible for the contamination has worked in 18 hospitals in seven other states during the last 10 years.  It’s not known when the medical technician contracted hepatitis C, but investigators believe he had a positive test for hepatitis C in June 2010.  This means that the investigation needs to be expanded and that many more people may need to be tested.

This article contains information about what facilities the medical technician worked at and the timeline for his employment.  To view an updated high level “Cause Map”, click here.

The Low Survival Rate of Pancreatic Cancer

By Kim Smiley

Sally Ride, the first US woman in space and a national hero, died after a 17-month battle against pancreatic cancer on July 21, 2012.  Pancreatic cancer is a particularly deadly cancer with only a 6% five-year survival rate.  This disease also affects many people.  In 2010 alone, an estimated 43,000 people in the US were diagnosed with pancreatic cancer.

The reasons that pancreatic cancer is so deadly can be explored by building a Cause Map, a visual root cause analysis.  The first step in building a Cause Map is outlining the problem which includes defining how the problem impacts the organizational goals.  In this example, the primary goal considered is the impact to the safely goal since pancreatic cancer has such a low survival rate.

In order to build the Cause Map, “why” questions are asked and the answers are added to the Cause Map.  Why does pancreatic cancer have such a low survival rate?  The survival rate is low because the cancer has usually spread beyond the pancreas by the time it is detected and pancreatic cancer is difficult to treat.  The cancer has typically spread before detection because there are very few symptoms in the early stages of the cancer and any symptoms that do exist are usually vague, like aches and pains that could easily be attributed to other illnesses.  There is also no screening test like there are for breast or prostate cancer to detect pancreatic cancer at this time.

Pancreatic cancers are difficult to treat for several reasons.  First, pancreatic cancers are resistant to chemotherapy.  The best course of treatment is typically removal of the tumor, but many cases are caught too late for the tumor to be removed because the cancer has usually spread by the time it is detected.

Researchers are working on improving the survival rate for pancreatic cancer.  There are some promising studies that show it may be possible to develop a screening test that could detect pancreatic cancer at earlier stages, which could significantly improve the chances for survival.

To view a high level Cause Map of this issue, click on “Download PDF” above.